lish11 Endocrine and Metabolic Disorders

Page 507 Chapter 11: Endocrine Diseases and Metabolic DisordersThe relationship between endocrinology and psychiatry has attracted a good deal of attention for obvious reasons. Endocrine disorder can be accompanied by prominent mental abnormalities, as for example in myxoedema and hyperthyroidism, and epochs of life marked by endocrine change such as pregnancy and the menopause appear to be associated with special liability to mental disturbance. In the reverse direction it is now clear that primary emotional disturbance is accompanied by changes in neuroendocrine regulatory functions of a complex nature.

Historically it is interesting to note that treatment by means of hormones has often been viewed as a possibility in psychiatry. Kraepelin (1896) at one time proposed that dementia praecox was basically an endocrine disorder. Others have speculated on the role of hormones in regulating the ’biological background of psychic life’, noting their influence on such matters as impulsivity, attention, arousal and their role in numerous drives in animals as well as humans (Bleuler 1967). Patients on substitution therapy are said to ’lose something in their personal profile’ due to lack of the complex interplay between emotions and hormonal levels. On this general question, however, there is as yet no clear evidence.

More recent interest centres on the way hormonal influences during intrauterine life and immediately after birth can come to alter fundamental aspects of behaviour and brain development. Experimental work in animals has clarified the morphological basis by which thyroxine lack during early development impairs the maturation of behaviour (Eayrs 1968). And it is now apparent that prenatal steroid hormones have a decisive influence in animals on sexual differentiation and a wide range of sexual and social behaviours (McCarthy 1994; Signoret &• Balthazart 1994). In human beings. Money and Ehrhardt (1968) reported that girls affected by androgens in utero, either due to spontaneous hyperadrenocortical activity or as a result of progesterone given to pregnant mothers, showed an increased tendency towards tomboyish behaviour later in their development. Such findings, if well confirmed, may open a further chapter in the relationship between endocrinology and psychological functions in human beings.

During the past two decades there has been an explosion of interest in peptide hormones, their regulation and their possible relevance to psychiatry. Thyrotropinreleasing hormone (TRH), corticotrophin (adrenocorticotrophic hormone, ACTH) and other hypothalamic peptides have proved to be under delicate control from neural as well as endocrine feedback processes. The activity of the hypothalamic-pituitary axis has come under detailed scrutiny in mental disorder, with evidence of, for example, impaired cortisol production in response to dexamethasone (Carroll 1976a; Carroll et al. 1981) and impaired growth hormone response to donidine and other drugs in endogenous depressive illness (Checkley 1992). Such disturbances appear to reflect central, presumably hypothalamic, alterations accompanying the emotional disorder. Pituitary neuropeptides related to ACTH and vasopressin are now known to affect learned behaviour in animals (De Wied etal. 1976). Through them it can be expected that the pituitary plays an important role in motivational, learning and memory processes. The opiate peptides (the endorphins and enkephalins) are obviously an immensely important discovery with potential relevance to addictive behaviour and the control of chronic pain (Rees 1981).

The purpose in what follows is not to explore these many aspects in detail, but rather to concentrate on the clinical psychiatric manifestations of primary endocrine disorder. It may be said that in all of the conditions considered below, occasional individuals will react in such a way that psychiatric manifestations gain prior attention and the endocrine disturbance goes unnoticed. With some conditions, for example myxoedema and Addison’^ disease, the psychiatric abnormalities are regularly intrusive to such a degree that there is a constant risk of mistaken diagnoses.

The other metabolic disorders to be discussed hypoglycaemia, anoxia, uraemia, electrolyte disturbance and hepatic disorder-similarly illustrate the importance

Page 508 CHAPTER 11

of the correct biochemical milieu for the proper functioning oi the central nervous system. They are quite often encountered by psychiatrists working in general hospitals and their psychological manifestations have accordingly received increasing attention in recent years. Porphyria is included as a rare but striking example of an inborn error of metabolism with important psychiatric features.


Hyperthyroidism affects females much more commonly than males in a ratio of approximately 6 :1. It is commonest in the second and third decades of life but the range is wide. The cause may lie in a hyperplastic nodule or secreting adenoma of the thyroid gland, but more commonly the gland is diffusely overactive (Graves’ disease). Goitrous enlargement may or may not be present.

Attention has been directed to the role of stress and emotional disturbance in precipitating hyperthyroidism, also to the psychological constitution of those who develop the disorder. The onset is often abrupt and may be seen to follow directly upon some stressful event or emotional crisis. It is hard, however, to exclude the possibility that such emotional traumas may themselves have been the by-products of early and unsuspected thyroid overactivity. Jadresic (1990) reviews the evidence from several studies, including Gray and Hoffenberg’s (1985) failure to find an association between stressful life events and the onset of thyrotoxicosis. More recently, Wisna etal. (1991) have reported a significant excess of negative life events compared with controls in the 12 months preceding the diagnosis of Graves’ disease, but this study has been criticised on a number of methodological grounds (Harris et al. 1992). The issue would therefore appear to remain subjudice.

With regard to predisposition, opinions have varied widely. Mandelbrote and Wittkower (1955) emphasised instability of the premorbid personality in persons subject to hyperthyroidism. Robbins and Vinson (1960), however, considered that the role of personality factors had been overstressed. Gurney et al. (1967) found that hyperthyroid patients as a group fell somewhere between neurotics and normals in terms of previous stability, with a similar family and personal history of psychiatric disorder to the neurotics but with greater previous stability on a number of other indices. It is clearly difficult to make accurate assessments of personality from the patient’s own account since this will be biased by the changes of mood that occur in the course of the illness (Jadresic


Common psychological accompaniments

Psychological disturbance of some degree is universal with thyroid overactivity. The patient becomes restless, overactive and irritable, sometimes with hyperacuity of perception and over-reaction to noise. Heightened tension leads to impatience and intolerance of frustration, and emotional lability may develop with unreasonable or histrionic behaviour. Depression can be prominent, though unaccompanied by retardation. Kathol et al. (1986) found that almost a third of 29 consecutive patients seen in an endocrine clinic met DSM-IH criteria for major depression. Eighty per cent had generalised anxiety. The picture of ’apathetic hyperthyroidism’ with anergia and mental slowing occurs mainly in elderly patients and is discussed on p. 511.

The over-arousal leads to distractibility so that concentration is impaired and effort cannot be sustained. In addition, careful examination may reveal definite cognitive impairments of which the patient is unaware, in the form of difficulty with simple arithmetic or difficulty with recent memory (Whybrow etal. 1969).

The emotional disturbance can reach a degree which leads to difficulty in clinical management, though modern antithyroid drugs have proved invaluable in circumventing the problems which arose when urgent thyroidectomy was the treatment of choice. States of extreme anxiety or irritability may emerge as a direct extension of the heightened emotional tension, or paranoid features may appear as an aspect of the disturbed mental state. Whybrow etal. (1969) found that seven out of 10 consecutive hyperthyroid patients in a general hospital showed psychiatric abnormalities severe enough to constitute a ’psychiatric illness’, even though none had been referred or considered as ’psychiatric problems’. Nevertheless hyperthyroidism does not appear to be exceptionally frequent among hospitalised psychiatric patients. McLarty et al. (1978) found eight patients with thyrotoxicosis after surveying the entire population of two psychiatric hospitals, a total of over 1200 persons in all. In six the hyperthyroidism had been unsuspected prior to the survey, and in five it seemed to be contributing to the mental illness.

Psychoses V

Other developments include the organic, affective and schizophrenic psychoses which sometimes accompany hyperthyroidism. Occasionally these are the presenting feature and lead directly to psychiatric referral.

Acute organic reactions

accompany ’thyroid crises’ and , show the typical picture of delirium, usually accompaniedPage 509 ENDOCRINE DISEASES AND METABOLIC DISORDERS

by fever. They were formerly one of the commonest forms of major mental illness encountered in the disease, but are now relatively rare owing to modern methods of treatment. They constitute ,a grave emergency and warrant urgent intervention. Diagnostic confusion is unlikely to arise on account of the abundant evidence of hypenhyroidism accompanying their development. The rare ’apathetic hypenhyroidism’ (p. 511), however, can sometimes progress to stupor or coma, and here diagnostic difficulties may be encountered.

Affective and schizophrenic psychoses

It is no longer believed that a specific ’thyroid psychosis’ exists, but it is generally agreed that a distinctive colouring may be lent by the hyperthyroidism. Thus a manic component may accompany otherwise typical schizophrenic symptomatology, and agitation is often profound in the presence of depression. Most observers are also agreed that paranoid features are especially common whatever form the psychosis may take.

The diagnostic distinctions between the affective and schizophrenic reactions are often blurred, and an admixture of organic psychiatric features is relatively common. A seemingly schizophrenic psychosis may sometimes represent covert organic disorder, the essential evidence for which may be overlooked as in the patient reported by Greer and Parsons (1968):

are sometimes indistinguishable from their naturally occurring counterparts. Mania is said to be more frequent than depression, and often the progression to mania can be seen as a direct outgrowth from the characteristic mental changes of the endocrine disorder. Schizophrenic illnesses of all types have been reported – hebephrenic, catatonic and paranoid – and have sometimes been found to outnumber affective psychoses.A man of 28 developed a short-lived schizophrenia-like illness with a paranoid delusional system, ideas of reference and influence, and auditory and visual hallucinations. Orientation and memory were apparently intact, but a contribution due to organic cerebral disorder was suggested by the presence of deja vu and panoramic memory at the height of the illness.

Psychotic developments have been reported in up to 20% of cases, though this may reflect special selection and the inclusion of acute organic reactions in earlier series. Johnson (1928) found only 24 examples of psychosis among over 2000 patients referred for thyroidectomy when patients with obvious confusion or delirium were excluded. Most were depressive states with hallucinations and delusions. The majority had a personal or family history of mental disorder, and the psychosis had usually been in evidence long before symptoms of hyperthyroidism appeared. This prevalence, at little more than 1 %, would suggest a chance association in most instances, with the hyperthyroidism merely aggravating an established mental disorder. In a careful survey of a number of patients with manic-depressive psychosis who had also had thyrotoxicosis, Checkley (1978) was unable to detect clear time relationships suggestive of a link between the two. He argued that if the hyperthyroidism had so little effect on the course of the affective disorder in patients long subject to manic-depressive episodes, it would seem unlikely to serve as a precipitant in patients without such liability.

Dunlap and Moersch (1935) reported 143 patients with mental disturbance accompanying hyperthyroidism; over 70% were organic psychosyndromes, but 26 patients showed manic-depressive psychoses (mostly depressions), two had dementia praecox and two were paranoid. Bursten (1961) found 10 examples of psychosis among 54 hyperthyroid patients seen in a general hospital during a 4-year period. Five were schizophrenic, three were organic reactions, one was a depressive illness and one a psychosis of undetermined type. From the same source of referral six examples of psychosis were observed among an equivalent number of patients with diabetes matched for age and sex, and only two among an equivalent number of patients with cholecystitis.

Neurological accompaniments

The commonly associated tremor at 8-12 Hz is best seen in the outstretched hands; it persists during movement and is absent at rest. The well-known signs of lid-lag and upper lid retraction may be accompanied by exophthalmos due to swelling of the retro-orbital tissues. There may then be impairment of convergence (Moebius’ sign). The tendon reflexes are usually brisk, some patients showing clonus and positive Babinski responses.

Epileptic fits are occasionally precipitated, thyrotoxicosis being a contributory factor in about 1 % of new-onset fits in adults (Kaminski & Ruff 1989). Abend and Tyler (1989) review the electroencephalographic changes seen in up to 50% of patients, these usually being reversible when the euthyroid state is attained. A proximal myopathy may develop with weakness, wasting and myalgia, and a significant association exists between thyrotoxicosis and myasthenia gravis. Occasional thyrotoxic patients suffer from recurrent attacks of ’periodic paralysis’ (p. 721).


Laboratory investigations are essential for confirming the

diagnosis of hyperthyroidism. In the majority of cases they give unequivocal results: raised serum thyroxine

Page 510 CHAPTER 11

(T4) and triiodothyronine (T3) as measured by radioimmunoassay, coupled with a raised ’free-thyroxine index’. The latter is calculated when the free T4 cannot be measured directly, since total T4 is much influenced by the levels of thyroxine-binding proteins in the serum.

In a small number of cases, however, the results may be borderline or even self contradictory. The clinical features then require careful appraisal, and referral for specialist investigation is usually indicated. Such borderline cases will often prove to be suffering from primary emotional disorder rather than hyperthyroidism, but further tests are nonetheless essential.

The estimation of serum thyroid-stimulating hormone (thyrotropin, TSH) is invaluable for confirmation of the diagnosis. With recent sensitive methods using monoclonal immunoassay, levels below 0.01 mU/1 are almost invariably found with symptomatic hyperthyroidism (McGregor 1996). On rare occasions it may be necessary to test for an impairment of the TSH response to an injection of thyrotropin-releasing hormone (TRH), though this is seldom now required.

Special difficulty may be encountered with psychiatric patients in that thyroid function tests can be transiently abnormal after admission to hospital. Among 480 newly admitted psychiatric patients, Cohen and Swigar (1979) found abnormalities in total thyroxin, thyroxine-binding capacity and free T4 in 9% of the patients, usually returning to normal within a few weeks. They labelled this ’acute stress hyperthyroidism’. Low T4 levels were found in a similar proportion, also often resolving spontaneously.

In the syndrome of T3 toxicosis’ there is an isolated excess of triiodothyronine which can produce clinical signs of thyrotoxicosis despite normal levels of total and free serum T4.

Differential diagnosis

The differential diagnosis between hyperthyroidism and anxiety neurosis is a classic and often difficult exercise. Physicians and psychiatrists need to be aware of the pitfalls. The presenting mental symptoms can be virtually identical in both conditions; both show tachycardia, fine finger tremor, palpitations and loss of weight, and both may appear to have been precipitated by stressful events. The presence of previous neurotic symptomatology in thyrotoxic patients and their families can lead to further blurring of the diagnostic criteria between the two conditions.

Careful analysis of the features shown by large numbers of patients with hyperthyroidism has clarified the physical symptoms and signs which are of most

importance in indicating this disorder (Wayne 1960) and it is useful to refer to such data in doubtful cases. The symptoms, in descending order of discriminating value, were sensitivity to heat and preference for cold, increased appetite, loss of weight, sweating, palpitations, tiredness, ’nervousness’ and dyspnoea on effort. The signs, in order of importance, were cardiac dysrhythmias (chiefly auricular fibrillation), hyperkinetic movements, tachycardia exceeding 90 beats per minute, a palpable thyroid gland, a bruit audible over the thyroid, exophthalmos, lid retraction, hot hands, lid-lag and fine finger tremor. These lists show how closely anxiety neurosis may be simulated.

Gurney et al. (1967) focused more closely on the problem by reviewing the features found in euthyroid patients with psychiatric disorder but who were initially referred with suspected thyrotoxicosis. Such patients, when compared with thyrotoxics, had an increased frequency of psychological precipitants for the illness, a lower age of onset, more frequent hysterical symptoms and panic attacks, and more neurotic features in the personality.

Thus hyperthyroidism will usually be readily suspected when the patient gives a clear history of sensitivity to heat and a preference for cold, and this deserves careful specific enquiry. Similarly the classic signs of exophthalmos, lid retraction and lid-lag will clarify the situation when such are present. Precipitation by stress will be found more commonly and more impressively in anxiety neurosis. But perhaps the most decisive feature in differentiating the two conditions is the preservation or otherwise of appetite in face of steady loss of weight; in hyperthyroidism appetite is characteristically increased whereas in anxiety states it is reduced.

In the presence of frank psychosis, diagnostic difficulties are liable to be increased and the hyperthyroidism may sometimes go unrecognised for a considerable time. It is necessary to beware of the occasional case of hyperthyroidism in which fluctuations occur with periods of spontaneous resolution. Repeated episodes of affective disorder may be particularly misleading:

A man of 40 was admitted to hospital with a typical attack of hypomania which responded satisfactorily to chlorpromazine during the next 3 weeks. Ten days later he was re-admitted with a relapse after discontinuing the medication, but once again he responded rapidly to chlorpromazine. Three ’months later he developed marked weakness and depression, and for the first time appeared to be physically unwell. It was noted that he had a persistent tachycardia, a warm moist skin and possibly an enlarged thyroid gland. Investigations confirmed hyperthyroidism, and retrospective enquiry revealed steady loss of weight and increased appetite since shortly before the first episode of hypomania. The admission notes on the two previous occasions


had shown a tachycardia which had been overlooked at the time.

Special diagnostic difficulty is likely to be encountered when thyrotoxicosis is accompanied by depression. ’Apathetic hyperthyroidism’, though rare, may easily be overlooked. The typical picture is of a middle-aged or elderly patient with considerable weight loss and apathy or depression (Lahey 1931; Thomas etal 1970). Cardiovascular symptoms may overshadow other evidence of thyrotoxicosis, and eye signs in particular tend to be absent. The physical appearance may resemble senility. In younger patients, too, depression can be the presenting feature. Folks and Petrie (1982) describe a woman of 23 presenting with depression, insomnia and early morning waking, who after an overdose of amitriptyline was found to be thyrotoxic. The affective disorder resolved as the hyperthyroidism came under control. Taylor (1975) reported a patient who was found to be thyrotoxic during a second attack of psychotic depression, the first having responded to electroconvulsive therapy; here the depression likewise abated when the thyroid disorder was treated.

Alcoholism may be wrongly blamed for the tremulousness and emotional lability of hyperthyroid patients. Davis et al. (1971) reported three men with previously stable records who were found to have been indicted for larceny shortly after the onset of thyrotoxicosis. In two of them alcoholism had been suspected by the employers on account of tremulousness, weakness and inattention at work, and loss or threatened loss of employment had precipitated their crimes.

Aetiology of mental disturbances

The common psychological accompaniments of hyperthyroidism are probably the direct result of increased T4 levels and subsequent metabolic derangements within the central nervous system. This is supported by the uniformity of the common mental changes from case to case, their fluctuations with exacerbations of the disorder, and the rapid subsidence of symptoms with antithyroid treatment. In Kathol et al.’s (1986) investigation generalised anxiety disorder, but not depression, showed a strong relationship with measures of the free-thyroxine index. Cerebral catecholamines may be intimately involved, particularly increased sensitivity of beta adrenergic receptors. Thus beta adrenergic blocking agents can be shown to improve psychiatric symptoms in hyperthyroidism despite unaltered levels of thyroid hormones (Trzepacz et al. 1988). The acute organic reactions are likely similarly to have an origin in brain metabolic disturbance and usually appear only at peaks of thyrotoxicosis. ,

The precise aetiology of psychotic developments is incompletely understood. Constitutional predisposition is usually invoked to explain severe affective or schizophrenic disorders, but even so the situation may be complex. The psychosis may be precipitated by the metabolic derangement or by the resulting emotional turmoil; alternatively, an ingravescent psychosis may have served to precipitate the thyrotoxicosis. Simple coincidence may account for the two developments, but the parallel course which they sometimes pursue suggests that a causal relationship of some sort is likely to exist quite commonly. At all events, once the processes are under way they doubtless augment one another.

During treatment with antithyroid drugs, such as carbimazole, an acute organic psychosis may make its first appearance, presumably in response to the toxic effects of the drug or a period of drug-induced hypothyroidism (Brewer 1969; Herridge & Abey-Wickrama 1969). Other cases are reported in which schizophrenia-like psychoses make an appearance at such a time. Bewsher et al.’s (1971) case illustrates the difficulties that can be encountered in deciding on the precise aetiological factors at work:

During the fifth week of treatment with carbimazole a woman became acutely psychotic with paranoid delusions, auditory hallucinations and marked overactivity, producing bursts of song and whoops of excitement. Initially she was febrile (38

̊C, 100̊F) and showed a tachycardia, but there was no evidence of recurrence of thyrotoxicosis. Memory and orientation were normal throughout and the level of consciousness was unimpaired. The paranoid delusions slowly resolved over several months during which chlorpromazine was given and carbimazole continued. The patient was euthyroid at the time the psychosis developed and remained so throughout its resolution. Follow-up over the next 2 years showed no recurrence. It was suggested that the psychosis had been precipitated by the fairly rapid alteration in the level of circulating T4 from severe excess to normality over the preceding 4 weeks, and perhaps by virtue of the effects of this transition on cerebral catecholamines.Outcome of mental disturbances

The result of treatment is in general satisfactory, with resolution of emotional disorder as the patient is rendered euthyroid. Sometimes, however, emotional instability persists, and in most cases is probably attributable to premorbid tendencies in this direction. Kathol et al. (1986) found that depression and anxiety resolved in the great majority of cases with antithyroid treatment alone. Occasionally, however, additional psychotropic medication will prove to be necessary. The acute organic psychoses usually also respond rapidly as the thyrotoxicosis comes under control. Schizophrenic psychoses may run a more

Page 512 CHAPTER 11

variable course, but as with other psychoses in which a precipitating cause is apparent the prognosis will usually be better than for schizophrenias which arise spontaneously.


(myxoedema)Myxoedema is of great importance in psychiatric practice and notorious for leading to mistakes in diagnosis. It is liable to be overlooked on account of its insidious development, and the minor and diffuse nature of the early complaints. Mental symptoms are almost universally present by the time the patient seeks advice, and many examples come before the psychiatrist. It is only by keeping the disorder in mind that early cases, or sometimes even advanced examples, will be detected.

As with hyperthyroidism, myxoedema is very much commoner in females than males, in a ratio of approximately 8:1. It presents most frequently in middle age though the range is wide. The physical accompaniments deserve first consideration since these will usually prove to be the features which raise suspicion.

Physical features

The appearance in classic examples is characteristic, with a pale puffy complexion and baggy eyelids. The skin is dry and rough, with a non-pitting oedematous appearance over the face and limbs and in the supraclavicular fossae. The patient may have noticed increased loss of hair, which has become lank and dry in texture. Speech is slow, and the voice often coarse, thick and toneless. The whole disposition of the patient tends to be sluggish and inert. It must be appreciated, however, that all such features may not be apparent in early stages of the disorder.

The pulse is usually slowed and angina is not infrequent. Appetite is diminished, the patient is often constipated, and hearing, taste and smell may be impaired due to deposits of mucoid material. Intolerance of cold is often a prominent early complaint. Menorrhagia is common in females, and impotence in males. Vague generalised aches and pains of a rheumatic nature may occur. Very occasionally muscular weakness may be the initial manifestation. On examination the ankle reflexes may be slowed with marked delay in the relaxation phase.

Common psychological accompaniments

The typical picture is of mental lethargy, general dulling of the personality and slowing of all cognitive functions. In the earlier stages the patient is subjectively aware of such changes, and complains of a thickness in the head or of ’feeling in a fog’. Ready fatigue may be a conspicuous feature, and relatives may have noted increasing psychomotor retardation with the patient taking progressively longer to perform routine tasks. Memory is often affected from an early stage, with failure to register events and forgetfulness for day-to-day happenings.

With further progression there is marked inability to sustain mental exertion and increasing slowness of uptake and grasp. The profound loss of interest and initiative carries the risk of delaying medical attention, since the patient may cease to complain and come to spend her time in a state of sluggish indifference.

The typical mood change is towards apathy rather than depression, though the distinction is not clear-cut. Irritability is a frequent feature, and some patients become markedly agitated and aggressive.

Psychoses and dementia with myxoedema

It is against the background of these changes that the more severe psychiatric disturbances of myxoedema occur. An organic psychosis may develop acutely or run a subacute course over several weeks or months. More commonly, memory impairment develops insidiously and may progress over several years to advanced dementia. A severe depressive psychosis may emerge, or a typical schizophrenia may be precipitated. It is generally agreed that there is no form of psychosis specific to myxoedema, but rather a variety of ’reaction types’ which may be called forth differently in different individuals. The only unifying feature, upheld by many observers, is the frequency of a paranoid colouring whatever form the psychosis may take.

In Asher’s (1949) classic paper on ’myxoedematous madness’ five patients showed an organic reaction with hallucinations and persecutory ideas, five showed the picture of schizophrenia with a marked paranoid colouring, two presented as advanced dementia and two with depressive features.

Organic psychoses


develops as an extension of the mental impairment characteristic of the condition generally. It progresses insidiously in a manner indistinguishable from a primary presenile dementia, and may have reached anusually show the features of delirium, with florid delusions and hallucinations, mental confusion and impairment of consciousness. Delusions of persecution may be gross and bizarre. Auditory hallucinations appear to be particularly common. The condition may run a fluctuating course, but even, when clouding of consciousness cannot be established there is usually evidence of impairment of cognitive function, particularly of recent memory.Page 513 ENDOCRINE DISEASES AND METABOLIC DISORDERS

advanced degree by the time the diagnosis is made. Olivarus and Roder (1970) considered myxoedema to be the most important, and the most frequently overlooked, of the metabolic causes of reversible organic intellectual impairment.

A 53-year-old woman developed dizziness and a constant diffuse headache after a mild head injury. Four months later she was admitted to hospital for repair of a rectal prolapse, but the operation was deferred because she was found to be confused and deluded. The EEG revealed diffuse slow activity with occasional sharp waves in the frontotemporal regions.

Full examination of the mental state showed her to be mentally sluggish, slightly depressed and with marked intellectual impairment. This was confirmed on psychological testing. There was slight left facial weakness and incoordination of the left arm. Her husband explained that in recent years she had had increasing difficulties with her job as a teacher, mainly because the pupils made fun of her lapses of memory. In the last 4 months she had become slowed and sluggish, with increasing inability to concentrate or remember.

Angiography failed to reveal a cerebral tumour and a diagnosis of presenile dementia was seriously considered. Air encephalography, however, showed littje evidence of cortical atrophy. Signs of myxoedema were then noted and the diagnosis was confirmed.

Within 3 weeks of starting replacement therapy she reported improvement, with decreased fatigue and improved memory. She was able to do crossword puzzles which she had given up several years before. Two months later psychometric testing confirmed marked intellectual improvement and showed only slight residual impairment of memory. She resumed her work without difficulty and remained well on follow-up.

(Olivarus & Roder 1970)

Depressive and schizophrenic psychoses

Neurological accompaniments

The slowing of the tendon reflexes has already been mentioned. This may serve as a useful confirmatory sign, and with careful measurement has been used to monitor the progress of treatment.

Jellinek (1962) drew attention to the occurrence of fits, faints and cerebrovascular accidents in myxoedematous patients. Four patients were reported with grand mal

attacks which responded to thyroid replacement, others with attacks of syncope, and others with unusual confusional episodes which suggested temporal lobe dysfunction. Millichap (1974) has reported further examples.

may or may not be accompanied by organic mental features, though these are usually found when sought out with care. Paranoid symptoms again figure prominently. The depressive psychoses are often severe, with agitation or bizarre hypochondriasis, and may prove to be particularly resistant to treatment until the myxoedema is discovered. Schizophrenic psychoses will in general be coloured by mental slowing, and often include features indicative of organic cerebral impairment.A man of 57 had sudden attacks which lasted a few minutes and consisted of varying sensations of familiarity and unfamiliarity: ’You know where you are but things face the wrong way’. On one occasion, when quite near home, he crossed the road away from his intended route and walked straight into a flow of moving traffic. He came to, in a state of panic, in the centre of the road. The attacks stopped when thyroxine was commenced.

Another patient had ’blackouts’ in which she would become briefly confused and talk nonsense. On one occasion she was accused of being drunk by a taxi driver when she refused to pay her fare. These attacks also ceased when she started thyroid treatment. (Jellinek 1962)

Cerebrovascular accidents were found to have occurred in several of Jellinek’s patients, and evidence of attacks of transient cerebral ischaemia in several more.

Jellinek and Kelly (1960) described other cases of myxoedema presenting with cerebellar disturbance in the form of ataxia, tremor, dysarthria and nystagmus, and remitting promptly with replacement therapy. Wise et al. (1995) have confirmed such pictures and reported patients with other unusual presentations. One presented with shock-like pains in the feet and an altered sleep pattern, another with myalgia and weakness in the calf muscles. The myopathy of myxoedema typically affects the proximal musculature with weakness, cramps and sometimes wasting. Sleep apnoea may occur as a result of obesity or upper airway obstruction.

Myxoedema coma is a grave condition which carries a high mortality. It typically develops in association with superimposed infection, surgery or trauma (Kaminski & Ruff 1989). It should be suspected in any patient with severe impairment of consciousness and hypothermia. The skin feels icy cold, and a low-reading rectal thermometer is required to confirm the hypothermia. Respiration may be sluggish, and cardiac failure or dysrhythmia are features of serious significance.


As with hyperthyroidism it is essential to confirm the diagnosis by laboratory tests before starting treatment. The total and free T4 levels are low, though the serum T3 may be normal. Elevated plasma TSH will indicate primary thyroid failure, whereas normal basal levels will suggest a pituitary origin. The latter is confirmed by an absent, subnormal or delayed response to TRH (McGregor 1996). The estimation of circulating autoantibodies to

Page 514 CHAPTER 11

thyroglobulin and thyroid peroxidase may help to establish the cause, for example Hashimoto’s thyroiditis.

The serum cholesterol is elevated, the heart is usually enlarged and the electrocardiogram shows a low-voltage tracing with flattened or inverted T waves. The EEG shows lowered voltage and slowing of the dominant frequencies; occasionally it is normal despite severe myxoedema but this is rare. The protein in the cerebrospinal fluid may be moderately raised.

Differential diagnosis

Not uncommonly the myxoedema is first recognised only after a considerable lapse of time. Early dementia or intractable depression are probably the diagnoses most frequently entertained, or the patient may have been labelled as neurotic, hypochondriacal or personality disordered. Occasionally such patients are found to have spent some time in a vain quest for medical help before the correct diagnosis is made.

The suspicion of myxoedema is usually derived from the characteristic facial appearance or other physical symptoms and signs, but unless the disorder is specifically considered these may easily be overlooked. In the 14 cases reported by Asher (1949), all with florid mental illnesses, the myxoedema had been missed by the referring doctor in every case. In the more severe psychotic illnesses organic features are usually evident in the mental state but not invariably so.

Even without overt evidence of myxoedema, psychiatric patients may warrant investigation if they have a history of thyroidectomy or of having required thyroid medication in the past. These were the factors that prompted investigation in five of 18 myxoedematous patients surveyed by Tonks (1964) in a psychiatric hospital. Patients on long-term lithium therapy are also at increased risk of developing hypothyroidism and require periodic checks of serum thyroxine levels.

Aetiology of mental disturbances

The mental symptomatology in myxoedema can be largely ascribed to changes in cerebral metabolism. Such changes are reflected in the electroencephalographic findings described above, and these can be observed to improve with substitution therapy. Cerebral blood flow has been shown to be considerably reduced as a result of diminished cardiac output, while the cerebral metabolic demands for oxygen and glucose are unaltered (Scheinberg etal 1950; Sensenbach etal. 1954; O’Brien & Harris

1968). The relative cerebral hypoxia that results is worsened by the anaemia which frequently coexists. The cerebral changes therefore appear to be largely secondary to the effects of thyroxine deprivation on other organs such as the heart, rather than the direct result of thyroxine lack on the brain itself.

Such abnormalities probably go much of the way towards explaining the mental slowing and dulling which form an integral part of the disorder. The acute organic psychoses are likely to be due to some additional aspect of the intracerebral metabolic disturbance. Distinctive colouring by way of mood disorder or paranoia will perhaps be derived from premorbid personality factors.

Cases of major affective disorder and schizophrenia are likely to owe a good deal both to organic factors and to matters of constitutional vulnerability. In the rare examples where organic features are entirely absent from the mental state the cerebral metabolic defect has probably served merely as a precipitant. But the situation is not entirely straightforward, since patients with purely depressive symptomatology have been found to respond to thyroxine after other forms of treatment have failed entirely (Michael &• Gibbons 1963).

’Non-myxoedematous hypometabolism’

Kurland etal. (1955) reported a syndrome of ’metabolic insufficiency’, in which the basal metabolic rate was low in spite of a normal serum protein-bound iodine and radioactive iodine uptake. Such patients frequently cornplained of fatigue, lethargy, sensitivity to cold, musculoskeletal pain and diminished sexual potency. Treatment with thyroxine was ineffective, but triiodothyronine was said to raise the basal metabolic rate and lead to striking clinical improvement. The condition gained popularity and came to be rather commonly diagnosed.

Levin (1960), however, performed a careful doubleblind trial of thyroxine, triiodothyronine and placebo in patients fulfilling the diagnostic criteria, and neither treatment could be shown to have definite effects on symptoms or on the basal metabolic rates. The patients’ complaints were more typical of neurotic disorder than of myxoedema, and Minesota Multiphasic Personality Inventory (MMPI) scores showed a high degree of psychological maladjustment in the individuals concerned. It would seem therefore that the syndrome is unlikely to be a clinical reality, and that triiodothyronine js not effective treatment. Some neurotic patients clearly have low basal metabolic rates, but so also do some members of the general population.

Outcome of mental disturbances

The treatment of myxoedema is usually highly rewarding.


The patient gradually regains vitality, physical symptoms diminish and mental processes return to their usual speed and efficiency.

The great majority of patients with serious psychiatric developments can also be expected to respond, even those with overt dementia, provided too long an interval has not elapsed. Jellinek (1962), however, stressed that several of his cases were left with measurable defects of intellect and memory after being rendered euthyroid, mostly those who • had remained undiagnosed for very long periods of time or where treatment had been inadequate.

Where response to thyroxine is concerned, the frankly organic psychoses can in general be expected to do better than psychoses with predominantly ’functional’ symptomatology. This was confirmed by Tonks (1964), who surveyed 18 hypothyroid patients in a psychiatric hospital during a period of treatment with thyroid preparations alone; the proportion who made complete and lasting recoveries was much higher among patients who showed evidence of disturbance of consciousness, by way of disorientation or confusion, than among those who did not. The duration of the illnesses was also important, in that no patient with a mental illness exceeding 2 years had a satisfactory response to the trial of thyroid replacement therapy alone. The only clearly organic condition which failed to respond was a patient with chronic progressive dementia and aphasia of 7 years’ duration.

Additional measures in the form of phenothiazines, antidepressant medication or electroconvulsive therapy may be necessary in severe psychotic disorders, and particularly so when organic features are absent from the mental state. It must be borne in mind, however, that phenothiazines carry some risk of precipitating hypothermic coma in hypothyroid patients, as in the case reported by Mitchell etal. (1959).

It is necessary to introduce thyroxine with caution at the beginning of treatment, especially in the elderly, because of the possibility of myocardial damage. The starting dose of 1-thyroxine sodium should not exceed 50 ug per day. If there is no evidence of cardiac failure or angina this may be increased by 25-50 pg per day every 2-3 weeks until the maintenance dose of 100-150 pg per day is reached.

Very occasionally the initiation of treatment is accompanied by the emergence of psychotic disorder, which interestingly usually takes the form of mania. Josephson and Mackenzie (1980) refer to 18 examples in the literature, 12 being manic illnesses and the others mixed affective or depressive disorders. The symptoms usually began within 4-7 days of starting thyroxine treatment, resolving over 1-2 weeks irrespective of further therapeutic intervention. All recovered completely. Such patients often had a personal or family history of psychiatric disorder and had frequently been depressed or delusional prior to starting treatment.

Cushing’s syndrome

Cushing’s syndrome is commoner in women than in men, usually starting in young middle age though the range of onset is wide. A tendency has been noted for the disorder to start during pregnancy, at the menopause or at puberty, or while the subject is undergoing a prolonged period of psychological stress.

In the majority of cases, perhaps some 80%, it is due to pituitary overproduction of ACTH, resulting in secondary bilateral hyperplasia of the adrenal cortices. How cornmonly this in turn is due to a primary abnormality of the pituitary, or a primary hypothalamic disturbance affecting the mechanisms of corticotrophin release, is still not clear, though small pituitary microadenomas are demonstrable in a high proportion of cases (Burke 1983). Occasionally there may be a radiologically demonstrable pituitary basophil adenoma, though this is rarely large enough to cause chiasmatic compression or raised intracranial pressure. Rarer causes include adrenal tumours-benign adenomas or malignant carcinomas – or ectopic ACTH production from malignant tumours elsewhere. Apart from differing levels of circulating ACTH, the endocrine abnormalities are the same whether due to adrenal or pituitary disease – a sustained excessive production of cortisol, obliterating the normal diurnal rhythm, and usually excessive production of adrenal androgens as well. Chronic alcoholics may occasionally develop a typical Cushing’s syndrome which resolves within days or weeks when the alcohol intake stops (Smals et al. 1976; Morgan 1982); it can therefore be wise to rule out alcoholism as a cause.

The great majority of cases present for medical attention on account of the physical disorder which develops, but psychiatric features are strikingly frequent and can be severe. Moreover occasional cases have been reported to present with psychiatric illnesses from the outset as discussed below, and the endocrine disorder may then be recognised only after a considerable delay.

Physical features

The physical changes include the well-known moon face, buffalo hump and purple striae on the abdomen and thighs. Truncal obesity is almost always present, and insidious weight gain is often the earliest sign. The cornplexion is plethoric and hirsutism may be marked. Excessive

Page 516 CHAPTER 11

bruising is common. Skin pigmentation may develop from the direct action of excessive ACTH on melanocytes. Hypertension is often severe and mild glycosuria may appear. Amenorrhoea is usual in the female, and impotence, testicular atrophy or gynaecomastia in the male. Other noteworthy features include liability to intercurrent infections, osteoporosis leading to backache or vertebral collapse, and muscular weakness particularly involving the legs which can sometimes be extreme.

Psychiatric features

Among patients reported from general hospitals, psychiatric disturbance has often been found in more than 50% of cases (Michael & Gibbons 1963). Trethowan and Cobb’s (1952) series of 25 consecutive patients seen in a general hospital is typical – four were severely disturbed and psychotic, six moderately disturbed and eight mildly disturbed; three had relatively insignificant psychiatric symptoms but only four could be declared mentally normal. Jeffcoate et al. (1979) surveyed 40 patients of whom 22 were depressed, five severely, and four who showed other psychiatric disorders (mania, chronic anxiety and an acute organic reaction). Only a third were judged to be free from mental disorder. Whybrow and Hurwitz (1976), in a review of the literature up to that time, suggested that some 35% of patients develop depression, 16% disturbed cognition and 9% psychotic illness. Less than 4% appeared to show euphoria, in contrast to the situation when exogenous steroids are administered for therapeutic purposes (p. 628). Starkman et al. (1981) reported a particularly high frequency of psychiatric symptoms in 35 patients examined prospectively before the start of treatment. All 35 were fatigued, 30 showed irritability, 29 impairment of memory, 26 depression, 24 decreased libido, 24 insomnia, 23 anxiety and 23 impairment of concentration. Irritability was the earliest symptom in most cases, often antedating the physical manifestations. Depression could be of sudden onset and was usually intermittent rather than sustained, rarely lasting for longer than 3 days at a time. Social withdrawal was common, seemingly often due to feelings of shame at the physical appearance.

Depression is widely reported as the most frequent psychiatric symptom, and paranoid features are also very common. A range of other mental abnormalities is seenemotional lability with gross overreaction to emotional stimuli, uncooperative behaviour or sudden outbursts of restless hyperactivity. These may be noted from very early in the development of the illness. Acute anxiety may also figure prominently, or states of apathy verging on stupor.

Fatigue and asthenia derived from the physical disorder often colour the psychiatric picture.

Cohen’s (1980) study was important in that a consecutive and unselected series of 29 patients with Cushing’s syndrome were examined closely from the psychiatric point of view. Twenty-five of them (86%) showed a significant degree of depression, this being mild in seven, moderate in 13 and severe in five. Almost half of the series had a family history of depression or suicide, or a past history of early bereavement or separation; six had had a major emotional disturbance shortly preceding the onset of the endocrine disorder, and in five this had consisted of a loss (bereavement, separation or broken engagement). These are all factors of known importance in the genesis of depression, raising at least the possibility of an aetiological link between Cushing’s syndrome and depressive illness. Moreover, depression was particularly common among the 21 patients with a pituitary origin for their Cushing’s syndrome; and all six patients with a disturbing life event preceding it fell into this group. It was, by contrast, uncommon to find severe psychiatric disturbance in the eight patients with adrenal adenomas or carcinomas.

The severe psychoses accompanying Cushing’s syndrome are again mostly depressive in nature. Typically they are florid illnesses with delusions and auditory hallucinations and often with paranoid symptoms. Retardation tendj to be severe, sometimes bordering on stupor. Anxious agitation may replace the retardation in other cases, or there may be acute brief episodes of grossly disturbed behaviour. Marked fluctuations in the severity of the condition appear to be characteristic.

Acute organic reactions are rare, but an element of disorientation or transient impairment of consciousness may be detected in severe examples of the disease. Classic schizophrenic psychoses are also rather uncommon though a schizophrenic colouring may be lent to the total picture. Johnson (1975) reviewed the occasional cases of schizophrenia in the literature and presented an unusual example of his own:

A woman of 50 had had a chronic schizophrenic illness with first-rank symptoms for 25 years. This had been extensively treated with insulin comas and electroconvulsive therapy. Signs perhaps suggestive of Cushing’s syndrome – excessive bruising and pigmentation – had been noted some 13 years before the florid endocrine illness was declared. Despite the length of the psychiatric history, bilateral adrenalectomy led to a dramatic and sustained improvement in her mental state. It was concluded that in all likelihood she had had a primary schizophrenic illness, partly in remission as a result of earlier treatment, then exacerbated by the developing Cushing’s syndrome. (Johnson 1975)


The chief diagnostic hazard lies with those patients who develop psychotic features early in the illness. These may dominate the picture to such an extent that the endocrine disorder goes unnoticed. Two of Spillane’s (1951) patients were apparently psychotic from the outset and long before the physical changes were sufficiently marked to suggest Cushing’s syndrome.

One of Spillane’s patients, a man of 26, had developed a paranoid psychosis which was treated with electroconvulsive therapy and continuous narcosis. It was not until 2 years after the first hospitalisation that Cushing’s syndrome was diagnosed. It was on his return home after being invalided from the army that his mother noted a pronounced change in his appearance, with obesity, a bullneck and a plethoric complexion.

Trethowan and Cobb (1952) similarly reported a woman of 31 who had developed obesity, marked muscular weakness and amenorrhoea for a year before becoming excited, overactive and disorientated. Two further acute psychotic episodes occurred before Cushing’s syndrome was diagnosed. Another of their patients was diagnosed as schizophrenic for several months, and another as hysterical for a year before the physical changes led to investigation of Cushing’s syndrome.

Cognitive impairments have rarely been investigated systematically in Cushing’s syndrome. A report by Whelan et al (1980), involving neuropsychological testing of 35 unselected patients before treatment, indicated some degree of diffuse cerebral dysfunction in almost two-thirds of the sample. Thirteen patients showed essentially normal results on an extensive battery of tests, 10 showed mild impairments, eight moderate and four severe and frequent deficits. No aspect of cognitive functioning was spared, though impairments tended to be more marked on non-verbal tests (visual-ideational, spatial-constructional and visual memory tasks) than in the field of language and verbal reasoning. Patients may therefore be more impaired in cerebral functioning than is evident in conversation or on purely verbal assessments. The deficits sometimes extended beyond purely cognitive functions, with poor performance on tests of manual dexterity and somatosensory discrimination.

Mauri et al. (1993) carried out a carefully controlled investigation of 25 patients, and emphasised a selective disturbance of memory functions. Patients with depression or other psychiatric disorders had been excluded, also patients with a past history of head injury or drug abuse. Significant impairments were revealed on a number of memory tests, particularly in older subjects, and approximately half reported mild impairment of memory and attention in everyday-life tasks. Other cognitive functions were, by contrast, spared, apart from impairment on the Digit Symbol Substitution Test (DSST) reflecting problems with attention and/or visuomotor functions. In a small subgroup who were retested 6 months after removal of their pituitary adenomas significant improvement was noted in both memory and DSST performance.

Radiological evidence of cerebral atrophy has also been found in a considerable proportion of patients (Momose et al 1971), occurring as commonly in those below 40 as in older groups. Cerebellar atrophy was often conspicuous despite a lack of clinical signs. Computerised tomography (CT) scanning has shown atrophy, reversible with treatment and perhaps attributable to electrolyte and fluid changes or protein loss (Heinz et al. 1977). Ventricular enlargement and cortical atrophy have also been reported at autopsy (Soffer etal. 1961).

Aetiology of mental disturbances

The depression so characteristic of Cushing’s syndrome is doubtless partly reactive to the physical disfigurements and discomforts produced by the disease. But a more direct connection is suggested in those cases where affective disorder is an early or even presenting feature, and by the frequency with which the depression reaches ’psychotic’ intensity.

Cohen’s (1980) observations (p. 516) are particularly interesting in this regard. The common finding of factors predisposing to depression in the histories of his patients may merely illustrate their vulnerability to depression in the face of physical illness; or it may indicate something more – a close pathophysiological link between the genesis of depression and the genesis of some forms of Cushing’s syndrome. Thus it is noteworthy in his series that depression was significantly more common in primary pituitary than primary adrenal forms of the syndrome, a difference that had already been discerned by Carroll (1976b) from cases in the literature. Disturbing life events, antecedent to the development of the endocrine disorder, were confined to this form of the disease. More discriminating controlled studies will be necessary, however, before concluding that Cushing’s syndrome may sometimes be stress induced.

The depression of Cushing’s syndrome has often been contrasted with the elevation of mood characteristically seen when steroids or ACTH are administered for therapeutic purposes. Whether the difference is due to differing plasma levels of biologically active steroids, or to the longcontinued chronic elevation of steroids in Cushing’s syndrome is not known. Neither Cohen (1980) nor Kelly et al. (1983) could relate the severity of depression to the levels of circulating cortisol in their patients; yet its alleviation

Page 518 CHAPTER 11

after surgical removal of the hyperplastic adrenals suggests that it must owe a good deal to some substance they produce. Hypothalamic factors may also be presumed to play a pan in view of the complex neuroendocrine relationships now apparent in the control and regulation of the hypothalamic-pituitary axis (p. 507). In cases of pituitary origin additional factors could be the increased levels of beta endorphin and methionine enkephalin which are secreted along with ACTH (Fava et al. 1987).

The mechanism whereby hypercortisolaemia leads to cognitive impairment is largely unknown, but could conceivably depend on hippocampal damage. Thus Starkman et al. (1992) found indications on magnetic resonance imaging (MRI) of reduced hippocampal volume in three of 11 patients with Cushing’s disease, the reduction in volume of the hippocampal formation correlating significantly with measures of verbal memory impairment and negatively with the levels of plasma cortisol. The hippocampal formation contains the highest proportion of corticosteroid-binding sites in the brain, which may make it particularly vulnerable in Cushing’s disease.


A successful psychiatric outcome can be expected when the endocrine disorder is effectively treated. The physical and mental symptoms usually improve in parallel until the patient regains her former stability. Depression is regularly observed to recede after adrenalectomy, pituitary operation or treatment with metyrapone, often starting to abate within days or weeks though sometimes taking as long as a year to clear completely (Jeffcoate et al. 1979; Cohenl980; Kelly et al. 198 3). Needless to say, when psychiatric disturbance has long antedated the Cushing’s syndrome there may be little or no change when the latter is remedied.

With florid psychotic illnesses the results can be dramatic, as in the following examples:

A woman presented initially with physical symptoms of Cushing’s syndrome, but on admission to hospital developed an acute psychotic picture with auditory and visual hallucinations and delusions about changing her sex. This was thought to have been precipitated partly by the mounting anxiety surrounding her admission to hospital. She became markedly paranoid and agitated, developed confusional episodes and showed bizarre catatonic motor phenomena. The entire condition responded well to bilateral extirpation of the hyperplastic adrenal glands and her mental state returned to normal within a few days of the operation. Follow-up 3 years later showed that she remained entirely well. (Hickman et al. 1961)

A soldier of 23 with a good service record became abruptly confused and hallucinated, and showed severely disturbed behaviour with grandiose and religious delusions. He was diagnosed as schizophrenic and treated extensively with electroconvulsive therapy. It was not until 1 year from the start that Cushing’s syndrome was diagnosed. He continued to be severely disturbed, but pituitary irradiation 18 months and 2 years after onset led to transitory amelioration of the psychotic symptoms. Two and a half years after onset bilateral adrenalectomy was performed, and thereafter there was steady and gradual improvement until full premorbid stability was regained. (Hertz et al. 1955)

Details of management of the endocrine disorder, and of the distinction between pituitary and adrenal causes, will not be dealt with here. There are rival claims and special indications for bilateral adrenalectomy, pituitary operation and pituitary irradiation. Metyrapone can be useful in suppressing cortisol production.

Addison’s disease

Addison’s disease usually presents in early adult or middle life. It results from autoimmune destruction of the adrenal cortices or from diseases such as tuberculosis which involve the glands bilaterally. The output of all adrenal steroids is low – glucocorticoids, mineralocorticoids and androgens. Loss of sodium is accompanied by retention of potassium and extracellular dehydration. The blood sugar is usually low.

Physical features

The onset of symptoms is gradual and the usual presentation is with general weakness, loss of appetite and loss of weight. Tiredness is an almost universal complaint. Pigmentation develops mainly on exposed skin surfaces. The voice is often soft and whining. Loss of libido is common, with impotence in the male and amenorrhoea in the female. Resistance to stress is lowered and sensitivity to infections increased. There is often pronounced intolerance of cold and the body temperature is usually subnormal. Hypotension is almost always present, syncope is common and symptoms of hypoglycaemia may appear at higher levels of blood sugar than is usual. There is an increased liability to convulsions, and the EEG is often abnormal with diffuse high amplitude slow activity. Potassium retention may lead to hyperkalaemic periodic paralysis (p. 721).

These several features deserve emphasis because the correct diagnosis is often delayed, sometimes until a severe ’Addisonian crisis’ has occurred with considerable threat to life. The Addisonian crisis consists of a sudden exacerbation of symptoms with pyrexia, vomiting, epigastric pain, dehydration and profound hypotension. It may occur spontaneously or in response to infection, chilling or drugs such as morphine or anaesthetic agents.


Psychiatric features

Psychiatric abnormalities are present almost without exception in patients with Addison’s disease. The cornmonest changes are those which might be expected in persons suffering from chronic physical exhaustion depression, emotional withdrawal, apathy and loss of drive and initiative. There are sometimes sudden fluctuations of mood, or episodes of marked anxiety and irritability. Based on his own experience and on cases from the literature, Cleghom (1965) described the mental symptoms as apathy and negativism in 80% of cases, depressive withdrawal and irritability in 50%, suspiciousness in 15%, agitated behaviour in 10% and paranoia with delusions in 5%.

Difficulties with memory are a major feature in up to three-quarters of cases (Michael & Gibbons 1963). Mild dementia may be simulated on account of the mental anergia, poverty of thought and general air of indifference. Considerable perceptual impairment may be seen as well, with increased thresholds to tactile, auditory and olfactory stimuli (Leigh & Kramer 1984). Drowsiness can be conspicuous though some patients show restlessness and insomnia. The severity of the changes may fluctuate from time to time, varying directly with the severity of the endocrine disorder.

Addisonian crises are sometimes preceded by increasing irritability and apprehension. Nightmares and episodes of panic lead on to acute organic reactions with clouding of consciousness, delirium, stupor and epileptic fits. In Addisonian stupor the patient is obviously unwell, lies curled up in bed resenting interference, and is collapsed and cold with dehydration, falling blood pressure and peripheral circulatory failure.

Psychotic pictures of a depressive or schizophrenic nature are rare, in contrast to the situation in Cushing’s disease. Cleghorn (1951), however, reported examples of acute and chronic paranoia, hallucinatory states and schizophreniform psychoses. Such disturbances may be evanescent. They are sometimes intimately related to impending crises. McFarland (1963) reviewed reports of 10 patients with schizophrenia, six with affective psychosis and one with organic psychosis, concluding that the form of psychotic development is unpredictable. One of his patients presented with hypomania; this masked the adrenal disorder until the patient lapsed into coma after electroconvulsive treatment, when severe hyponatraemia was discovered.

Differential diagnosis

Addison’s disease must be differentiated from hypopituitarism and from other chronic debilitating diseases. Weight loss, hypotension and pigmentation may all be seen, for example, in carcinoma, tuberculosis, malabsorption or malnutrition. It is therefore essential to investigate adrenal function adequately before making the diagnosis.

Hyponatraemia is present in about 90% of cases of primary adrenal insufficiency, and hyperkalaemia in 65% (Edwards 1996). Normal blood levels of sodium and potassium therefore do not preclude the diagnosis entirely. Similarly, basal plasma cortisol levels and urinary free cortisol levels cannot be relied upon to exclude the diagnosis since these often lie within the low normal range (Edwards 1996). The definitive tests for primary adrenal failure are either the simultaneous measurement of plasma cortisol and plasma ACTH, showing that the latter is disproportionately elevated in comparison to the former, or measurement of the plasma cortisol response to the ’synacthen test’. This involves the intramuscular injection of tetracosactrin, an ACTH analogue, then examining the rise in plasma cortisol that follows. In primary hypoadrenalism there is little or no response; in hypoadrenalism secondary to pituitary failure the response is delayed with much higher values at 24 hours than at 4 hours. The estimation of plasma renin activity can also be of value, levels being raised in primary adrenal failure in consequence of mineralocorticoid deficiency.

From the psychiatric point of view an erroneous diagnosis of neurosis or early dementia may easily be made. The depression and generalised weakness is often attributed to neurasthenia, especially when pigmentation is slight and the serum electrolytes are normal. The impression of neurosis is strengthened by the anorexia, irritability and diminished libido, and by the fluctuations which occur from time to time. Dementia, or a chronic amnesic syndrome, is suggested when memory difficulties are in the forefront of the picture.


Adequate replacement therapy is usually highly successful in alleviating both physical and mental disturbances. The patient’s sense of well being is quickly restored, and appetite and energy gradually return to normal. It has been observed that Addisonian patients are unusually sensitive to the mood-elevating effect of steroids (Cleghorn 1965). Glucocorticoids appear to be more important than mineralocorticoids for reversing the mental symptoms and abolishing the electroencephalographic abnormalities, indicating that these do not rest entirely on disturbances of electrolyte and water balance (Reichlin 1968). Cleghorn (1951) found that apathy,

Page 520 CHAPTER 11

depression and irritability often persisted on treatment with desoxycorticosterone acetate and salt alone, but could resolve when cortisone was added later. Further treatment with androgens appeared to give no additional benefit.


Phaeochromocytomas are tumours of the chromaffin cells of the adrenal medulla. Occasionally they are found ectopically in relation to the sympathetic ganglia lying along the aorta or in the cervical and thoracic chains. Most are benign and some occur familially. Some association with neurofibromatosis has been noted; also with hyperparathyroidism and medullary carcinoma of the thyroid (multiple endocrine adenomatosis, MEA type 2) (p. 528). In Hutchison et al.’s (1958) series the age of presentation varied from 9 to 51 years.

Clinical features

The tumours secrete an excess of’adrenaline and noradrenaline, the relative proportions differing in different cases. The output may be paroxysmal or continuous. Accordingly the clinical features are subject to great variation.

Hypertension is always present during attacks and cornmonly persists in between (Ross 1972a, 1972b). The paroxysms that occur are usually the presenting feature. They last anything from 5 minutes to several hours at a time, and consist usually of severe palpitations, flushing or blanching, sweating, dizziness and tremulousness. A violent tachycardia is common, sometimes with substernal pain and acute dyspnoea. Nausea and vomiting may occur. The acute rise of blood pressure can be accompanied by agonising headache and may precipitate a cerebrovascular accident, epileptic fit or myocardial infarction. Death may result from ventricular fibrillation. After a severe attack the patient is left exhausted for hours or sometimes days.

Marked mental symptoms regularly accompany attacks. Intense fear is often present at the start and the patient may be overwhelmed with a feeling of impending death. Anxiety usually remains severe throughout the attack, and a period of excitability and confusion can follow. Attacks are precipitated by physical exertion, change of posture or raised intra-abdominal pressure, but also sometimes by emotional factors. Quite commonly they are triggered by a recognisable stimulus such as excitement, shock or panic. Sometimes, however, there are no discernible precipitants.

While the above is the classic picture, with well-marked episodes, cases may also present surreptitiously. Attacks are sometimes minor in nature, or mentioned only in passing as feelings of faintness, palpitations or episodes of sudden anxiety. Hence the great importance of carrying out appropriate investigations whenever the disorder is remotely suspected.

Examination reveals marked hypertension during attacks and usually also in between. Papilloedema may very occasionally be present, with haemorrhages and exudates in the retina. Transient glycosuria may accompany the attacks, and a considerable proportion of patients show diabetes mellitus (Hutchison et al. 1958).


The essential investigation is the demonstration of greatly increased levels of catecholamines in the plasma or urine, or of their metabolites in 24-hour samples of urine (metadrenaline, metnoradrenaline and vanilmandelic acid). When the index of suspicion is high, repeat estimations may have to be undertaken. Bravo and Gifford (1984) discuss the relative merits of urinary and plasma estimations. Twenty-four-hour urinary specimens can occasionally be misleading, since the amount of free catecholamines and their metabolites varies according to the levels of synthesising and metabolising enzymes within the tumour. On the other hand, a single plasma estimation may miss the occasional patient with truly episodic secretion. Plasma must be withdrawn with the patient fasting and resting supine for at least 30 minutes, also with the cannula in place well beforehand. In borderline cases a suppression test is useful: pentolinium or clonidine cause an immediate fall in plasma catecholamines and blood pressure when a phaeochromocytoma is present. Provocation tests formerly used various procedures to demonstrate an abrupt rise of blood pressure; these have now been largely abandoned since they can give misleading results and are potentially dangerous.

Differential diagnosis

Many patients referred as possibly having phaeochromocytomas prove to be suffering from some other condition. In Evans etal.’s (1951) series, 10 out of 20 suspected cases were suffering from anxiety or hysteria, often with episodes of hyperventilation. Other cases prove to be suffering from vascular headache, epilepsy, agitated depression or alcoholism (Hutchison etal. 1958).

More serious mistakes may occur in the reverse direction in that the phaeochromocytoma is missed. Essential or renal hypertension are probably the commonest misdiagnoses. Hyperthyroidism is often suggested by the


patient’s hypermetabolic state and associated heat intolerance. Any patient with hypertension in whom hyperthyroidism is suspected should immediately be screened for phaeochromocytoma. Other misdiagnoses include temporal lobe epilepsy, hypoglycaemic attacks and paroxysmal cardiac dysrhythmias (Ross 1972b).

From the psychiatric point of view an anxiety state may be very closely simulated, especially when emotional factors are known to trigger attacks. In two patients reported by Doust (1958), anxiety states of considerable duration had been attributed to psychological factors alone. One showed no obvious acute episodes and was normotensive. In a patient described by Gillmer (1972) a diagnosis of endogenous depression was made initially and the true condition was revealed in an unusual manner:

A woman of 61 with a strong family history of affective disorder complained of depression and anxiety for 18 months which had recently intensified greatly. She had severe insomnia and marked psychomotor retardation alternating with periods of acute anxiety and agitation. Blood pressure was 180/100 and there were minor hypertensive retinal changes. Treatment with antidepressants, chlordiazepoxide and electroconvulsive therapy was commenced. After the first electroconvulsive treatment she complained of severe headache associated with, sweating and tachycardia, and the blood pressure was found to be 120/60. In view of the drop in blood pressure 6-hourly recordings were instituted before further electroconvulsive therapy was given. During the period of observation it was found that bouts of severe headache, dizziness and sweating were associated with peaks of greatly elevated blood pressure, for example to 300/170. A phaeochromocytoma was confirmed and removed successfully.

A patient seen personally illustrates another unusual mode of presentation, and underlines the importance of screening tests:

A 57-year-old man presented with a 1-year history of decline in work performance, loss of confidence and a change towards becoming quiet and subdued. This appeared to follow an accident at work when, as a senior ship’s pilot, he had grounded a large vessel. After a second similar accident he was referred for investigation.

There was a 13-year history of diabetes mellitus, currently being treated with 44 units of insulin per day. Two years previously he had been treated briefly for hypertension with propanolol. Searching questions revealed two possible episodes of transient neurological dysfunction; for a few days after the first grounding he had appeared disorientated and had shown problems with direction when driving, and a year before that there had been tingling in the left arm and dragging of the left foot for a few hours.

On examination he was found to be depressed, apathetic and poorly informed about recent items of news. He was fully orientated but performed poorly on tests of memory and showed word finding difficulties. The blood pressure varied from 140/95 to 180/130. There were no neurological abnormalities.

The CT scan showed some diffuse cerebral atrophy, generalised white matter low attenuation, and two small cerebral infarctions-in the head of the right caudate nucleus and in the right cerebellar hemisphere It was considered that the likely diagnosis was an early multi-infarct dementia in the setting of hypertension and diabetes mellitus. However, as part of a thorough screening procedure 24-hour urine collections were obtained and showed greatly elevated levels of vanilmandelic acid, metadrenahne and metnoradrenaline. A body scan showed bilateral suprarenal masses.

At operation bilateral phaeochromocytomas were removed. Evidence of diabetic retinopathy rapidly receded thereafter and he was soon able to dispense with hypoglycaemic agents. The blood pressure remained within normal limits or only mildly elevated without antihypertensive treatment. Repeat psychometric testing showed steady gains in general intellectual competence, the full-scale IQ rising from 99 to 114 during the first postoperative year. Word finding difficulties resolved completely, though difficulties with new learning persisted. He regained a good deal of his former vitality though remaining quieter than formerly. Repetition of the CT scan showed persistence of the white matter low attenuation, and some dilatation of the anterior horn of the right lateral ventricle adjacent to the infarct in the caudate nucleus.

In this patient the phaeochromocytomas had clearly been responsible for diabetes of many years standing, and for the hypertension that had more recently been discovered. The diagnosis was only made, however, after episodes of silent cerebral infarction had led to difficulties in a demanding work situation. There had at no point been indications of episodic changes in blood pressure or anything remotely resembling anxiety attacks.


Overproduction of the pituitary growth hormone results from an adenoma, or rarely simple hyperplasia, of the eosinophil cells of the anterior pituitary gland. Skeletal overgrowth develops insidiously, affecting mainly the hands, feet, skull and lower jaw. Headache is often marked and incapacitating, kyphosis is common and joint pain may be severe. Hypertension, hypogonadism and diabetes mellitus usually occur. Loss of libido is a frequent early sign. Chiasmatic compression is relatively rare.

With regard to psychological accompaniments Bleuler (1951b) set the stage by a study of 22 patients from ttye Burghoelzli clinic in Zurich. He described alterations of personality by way of lack of initiative and spontaneity, sometimes interrupted by brief periods of impulsive behaviour, also changes of mood towards cheerfulness and self satisfaction. Brief swings of mood were regarded as characteristic, sometimes with spells of anxiety.

Page 522 CHAPTER 11

Bradyphrenia was observed in advanced cases and somnolence was common. Egocentricity and lack of consideration of others could be a problem for members of the family.

These impressions have rarely been checked in systematic studies. Anecdotal reports have emphasised depression and anxiety, coupled with loss of self confidence and concern over body size (Avery 1973). Margo (1981) reported a patient with a chronic depressive illness beginning 12 years before the acromegaly was diagnosed and with prominent psychomotor retardation from the outset. Sivakumar and Williams (1991) described a patient who showed pronounced depression and marked behavioural changes by way of pathological gambling out of keeping with his previous personality. This appeared to represent a loss of impulse control.

Psychotic disorders appear to be rare. Pye and Abbott (1983) described a patient who developed ideas of reference, visual hallucinations, delusions that her food was poisoned and voices urging her to harm her children, shortly before her acromegaly was diagnosed. Spence (1995) reported a patient with persecutory delusions and visual and auditory hallucinations in the context of depression. This was accompanied by an episode of impulsive stealing.

It is hard in such examples to know how far simple coincidence may be responsible. Abed et al.’s (1987) survey of 51 acromegalic patients used standardised assessments, and failed to find any general increase in psychiatric morbidity, nor a specific increase in depression when compared with the rates in other population studies. Indeed the scores were significantly lower than in certain other samples, and the interviewer was impressed with the optimism and even elation shown by many of the subjects. No overall relationship could be found between growth hormone levels and psychiatric morbidity; however, females showed significantly higher morbidity rates than males, and 10 of the 11 subjects who scored above the cut-off points for psychiatric illness were women.

The instruments used in Abed et al.’s study were unsuited to the measurement of personality change so this aspect remains to be elucidated. Patients are described as sometimes reserved, touchy and irritable, with emotional lability and traits of obstinacy. How far such features may be understandable as reactions to disfigurement, headache and limb pain, or whether they depend on metabolic changes or basal brain compression has not been clarified.

Neurological complications include mild proximal myopathy, and occasionally carpal tunnel syndrome due to hyperplasia of tendons and ligaments coupled with synovial oedema (Pickett et al. 1975). Hypersomnia may be a consequence of sleep apnoea occasioned by airway obstruction due to macroglossia and hypertrophy of the pharyngeal soft tissues (Perks et al 1980, Seggev et al.1986). In some cases, however, the daytime somnolence remits rapidly on treatment of the acromegaly, suggesting that it is not solely due to airway obstruction.

The mainstay of treatment is the removal of the responsible pituitary adenoma, usually by the transsphenoidal route (Wass 1993). With large tumours there is a risk of loss of pituitary function and other complications, and radiotherapy may then be employed. Bromocriptine improves symptoms in many patients, but carries a small but definite risk of precipitating a schizophrenic or hypomanic psychosis (LeFeuvreetfl/ 1982; Turner et al. 1984). Octreotide, a long-acting analogue of somatostatin, has proved to be even more effective in suppressing growth hormone secretion.


The commonest secreting tumour of the pituitary is a prolactin-secreting adenoma, leading to amenorrhoea and more rarely galactorrhoea in women, and to impotence and infertility in men. There is often a very considerable delay in recognising the tumour as the cause (Franks et al. 1977). In women it is usually a microadenoma, less than

1 cm in diameter, whereas in males it is commonly larger and may present with headache or visual field disturbance (Abrams & Schipper 1989). Even microadenomas can now be identified in a high proportion of cases by magnetic resonance imaging (Naheedy et al. 1987). Short of direct visualisation the tumour may be revealed by displacement of the infundibulum, focal bulging of the pituitary gland or focal erosion of the floor of the sella turcica.

Prolactin secretion is normally under tonic inhibition from the hypothalamus, the major inhibitory factor being dopamine secreted by neurones in the tuberoinfundibular region. Loss of such inhibition can be an important indicator of hypothalamic disease. Other important causes of hyperprolactinaemia include drugs which block dopamine receptors, notably neuroleptics. It may also accompany Cushing’s disease, perhaps through stimulation of prolactin secretion by beta endorphin.

Fava etal (1987) review studies showing high levels of anxiety and depression in women with the disorder, this being greater than in controls suffering from amenorrhoea but with normal prolactin levels. Up to a third of patients have been found to satisfy DSM-III criteria for major depression. In males the evidence for a relationship with affective disorder is less clear, suggesting that the


effects of prolactin may in part depend on its interaction with gonadal hormones.

In general the depression responds poorly to antidepressants, but is ameliorated by the fall of prolactin induced by bromocriptine. Tumour size is also decreased by the drug. Transsphenoidal adenomectomy can be effective when a tumour is identified, likewise radiotherapy. Treatment with bromocriptine may very occasionally result in the development of a psychosis as in the treatment of acromegaly (Turner et al. 1984). Such reactions may occur in the absence of a personal or family history of psychiatric disorder, and usually remit on stopping the drug or reducing the dosage:

A 38-year-old woman with hyperprolactinaemia due to prolactinoma was treated with 7.5 mg bromocriptine daily for 7 months. During this time she became increasingly depressed, anxious and tearful and became convinced that she was being followed. She described people talking about her in the street and attempting to try keys in the lock of her door. She stopped taking bromocriptine because of her mood state and lost her schizophreniform symptoms within a month.

(Turner et al. 1984)


The commonest cause of hypopituitarism was formerly ischaemic necrosis of the anterior pituitary gland as a result of postpartum haemorrhage (Sheehan’s syndrome). The cause is now usually a pituitary tumour, in particular a prolactin-secreting tumour in adults or a craniopharyngioma in children. A rare cause is head injury with fracture of the base of the skull.

Physical features

The condition is commonly of long duration, sometimes extending over many years when first presenting for attention. Leading symptoms include weakness, ready fatigue and marked sensitivity to cold. There is loss of libido, with amenorrhoea in females and impotence in males. Loss of weight is common, but despite the earlier name of ’pituitary cachexia’ it is not universal. Nor is it extreme until the terminal stages of the disorder (Sheehan & Summers 1949). In cases with pituitary neoplasms, weight may actually be gained if hypothalamic function is disturbed. Anorexia is common but in some cases appetite is well preserved.

Cardinal signs on examination are a thin dry skin, which fails to tan normally and may become wrinkled as in premature ageing; a dull expressionless face and loss of pubic and axillary hair. The body temperature is often subnormal, the pulse slow and the blood pressure low.

Psychiatric features

The mental picture can be equally striking. The frequency of psychiatric disorder was shown by Kind’s (1958) survey of cases from the literature and from his own experience. Ninety per cent showed psychiatric symptoms and in half these were severe.

Depression may be marked, sometimes with outbursts of irritability. Drive and initiative are impaired, and the patient comes to spend progressively longer in bed. Virtually all patients show apathy, inertia and somnolence in some degree. Ultimately most are dull and drowsy, prone to self neglect and indifferent about their state. The degree of psychological change commonly seen is greater than in other chronic debilitating diseases, and the patient’s poor physical condition is therefore unlikely to be the complete explanation.

Impairment of memory may occasionally figure prominently and give rise to an impression of a dementing process. Other severe psychiatric complications include episodes of delirium in relation to impending metabolic crises, or more rarely chronic paranoid hallucinatory psychoses.

Metabolic crises may lead on from delirium to hypopituitary stupor or coma which is always a grave complication. Such severe developments usually set in only several years after the physical disorder has made its first appearance. The following case reported by Blau and Hinton (1960) illustrates the problems that may arise:

(Simmond’s disease)A woman of 46 was admitted with drowsiness, neck stiffness and a moderate pyrexia of 2 days’ duration. She opened her eyes to her name but would not obey commands and resisted examination. Meningitis was diagnosed at first, but scanty pubic and axillary hair soon led to the diagnosis of Simmond’s disease.

Falling blood pressure required intravenous noradrenaline in addition to intravenous glucose and hydrocortisone. She emerged from the semicomatose state but remained incontinent and uncooperative and proved to be deluded about her attendants. Violent behaviour necessitated transfer to a psychiatric hospital 2 weeks later, and over the next month she fluctuated from apathy to outbursts of restlessness with aggressive shouting. Memory and orientation were faulty and she was unable to concentrate for long. During the next few weeks she settled into a calm rather foolish euphoria and was correctly orientated for most of the time.

She was followed up with regular treatment with cortisone. Three months after recovery from the coma her mental state and intellectual functions were back to normal. She had recovered her libido, which had deteriorated along with her general health since the birth of her child 10 years previously.

In this case the episode of coma appeared to result from a combination of intercurrent infection, hypoglycaemia, hypotension and hypocorticoidism. The transient organic

Page 524 CHAPTER 11

psychosis which followed it was probably due to reversible cortical damage resulting from some of the latter factors.

A more typical example was reported by Khanna et al. (1988):

A 28-year-old woman presented with a history of abnormal behaviour and irrelevant talk during the preceding 5 days. This had been sudden in onset and progressive. She suspected others of taking away her belongings and heard bizarre threatening sound and voices ’from inside her abdomen’. She was neglectful of personal care and incontinent of urine. On examination she was dehydrated, markedly pale, mildly hypotensive and showed loss of axillary and pubic hair She was disorientated in time, place and person and showed disturbance of immediate and recent memory. The provisional diagnosis was of delirium.

Her last child had been born 18 months previously along with excessive blood loss requiring transfusion. Failure of lactation and persistent amenorrhoea had followed the delivery. Thereafter she became slowed in household work, complained of weakness and fatigue, intolerance of cold, poor appetite and almost complete loss of libido. For 2 weeks prior to hospitalisation she had had frequent vomiting.

She was treated with rapid intravenous infusions to correct her dehydration and hyponatraemia. Within 48 hours she improved markedly and became fully orientated. Memory and intellectual function improved gradually on prednisolone and thyroxine and she regained her physical strength. She remained well on follow-up 3 years later.

Thus, as with adrenal cortex hypofunction, hypopituitarism rarely leads to functional psychoses, but cornmonly to acute organic reactions in association with crises of metabolic disturbance. In both endocrine conditions alterations of mood form an integral part of the clinical picture and take the form of apathy, anergia and indifference. In all these respects the psychiatric accompaniments and complications are in contrast to those of Cushing’s disease, where functional psychoses are common and where the usual mood change is towards depression and emotional lability.

Differential diagnosis

Hypopituitarism must be differentiated from myxoedema, in which the facial appearance of the patient is very different, and from Addison’s disease in which pigmentation is a prominent feature. In questionable cases full endocrine assessment is essential before embarking on the appropriate replacement therapy.

From the psychiatric point of view, neurosis and dementia may sometimes be closely simulated, but the principal differential diagnosis is from anorexia nervosa. Many of the early reported cases of hypopituitarism seem in retrospect to have been anorexia nervosa and vice versa (Sheehan fr Summers 1949). Now that the situation has been clarified, however, there is rarely clinical doubt, even though both share the cardinal feature of amenorrhoea. Severe weight loss is rare except terminally in hypopituitarism, whereas it is usually a presenting feature in anorexia nervosa. Similarly appetite may sometimes be well preserved in hypopituitarism. Loss of pubic and axillary hair is unusual in anorexia nervosa, and the fine downy facial hair of anorexia nervosa is rare in hypopituitarism. The psychological features of the two conditions are also very different: in hypopituitarism the patient is dull, apathetic and somnolent, whereas in anorexia nervosa the patient is typically restless and surprisingly active; distinctive attitudes to food and to the body image are lacking in hypopituitarism, whereas they form an important constellation of symptoms in anorexia nervosa. When serious doubt exists full endocrine assessment will clarify the differential diagnosis.


Response to replacement therapy is usually good. Within a few days the patient experiences return of interest and energy, and most lose their symptoms entirely. In cases of very long duration, however, apathy and lack of drive may persist in some degree. Cortisol or prednisolone alone may suffice, though thyroxine is sometimes given in addition. Gonadal steroids may be required to restore libido and potency in the male. Textbooks of medicine should be consulted for further details.

Diabetes insipidus

The syndrome of diabetes insipidus consists of polyuria with secondary polydipsia, resulting either from a deficiency of circulating antidiuretic hormone (ADH, vasopressin) or a lack of action of the hormone on the kidney. ADH is synthesised in the supraoptic and paraventricular nuclei of the hypothalamus, whence it is transported to the posterior lobe of the pituitary and then gains access to the circulation. It acts to increase the reabsorption of water by the distal convoluted tubules of the kidney, resulting in the production of a more concentrated urine. In cranial (neurogenic) diabetes insipidus ADH is produced in insufficient quantity. In the nephrogenic form the kidney fails to respond normally to, that available. Both result in the production of large volumes of dilute urine, normally accompanied by thirst. The urine osmolality is low, but the plasma osmolality is usually only slightly raised provided the thirst mechanisms are intact and the patient drinks adequately. If thirst does not occur, or if fluid intake is prevented, a dangerous degree of hypernatraemia and dehydration may develop.


Cranial diabetes insipidus can set in at any age without apparent cause, usually as an isolated abnormality but occasionally with other indications of hypothalamic disorder. The onset is typically abrupt. In very rare examples the condition is familial, being inherited as a Mendelian dominant. Other cases result from head injury with damage to the pituitary stalk, then often being transient, or follow pituitary surgery or yttrium implantations. Primary or secondary tumours involving the hypothalamus may be responsible.

Nephrogenic diabetes insipidus can occur as a rare sexlinked recessive disorder affecting males, and usually presenting soon after birth. Causes in adults include hypercalcaemia, potassium depletion and the prolonged intake of excessive amounts of water, all of which can impair the action of ADH on the nephron. A variety of drugs, including lithium, may also be responsible. Polyuria from lithium treatment can develop when plasma levels are within the therapeutic range; some 40% of patients on lithium experience thirst, with perhaps 12% developing polyuria (Ledingham 1983). In most cases this resolves within several weeks of withdrawing the drug.

The differential diagnosis of diabetes insipidus must include primary renal disease, diabetes mellitus, and the polydipsias induced by drugs such as chlorpromazine or thioridazine which may stimulate drinking by a direct action on the hypothalamus. The major diagnostic problem, however, is to distinguish diabetes insipidus from compulsive water drinking as described below.

The treatment of neurogenic diabetes insipidus consists of administering vasopressin. For transient states, as after head injury, aqueous vasopressin may be given subcutaneously. In the chronic condition the synthetic analogue desmopressin (desamino-D’-arginine vasopressin, DDAVP) is preferable on account of its longer duration of action and diminished pressor activity. This can be administered as a nasal spray. Nephrogenic diabetes insipidus is treated by thiazide or amiloride diuretics combined with indomethacin (Baylis 1996).

Compulsive water drinking

(psychogenic polydipsia)

Compulsive water drinking may be associated with a wide range of psychopathology-neurosis, personality disorder or psychosis. In psychotic patients it is frequently delusionally motivated. Among the nine examples described by Barlow and De Wardener (1959), long-standing personality disorder was common, often with hypochondriasis and depression. Six of the patients had had hysterical conversion episodes and some had histories of compulsive eating. Denial and evasion were sometimes a prominent part of the picture. Illowsky and Kirch (1988) estimate that polydipsia and polyuria without identifiable medical cause occur in 6-17% of psychiatric patients, especially those with chronic schizophrenia.

The clinical syndrome that results can simulate diabetes insipidus closely. In both conditions the fluid intake and output are raised and the urine osmolality is low. With compulsive water drinking, however, the plasma osmolality is also likely to be low. Hyponatraemia may develop when the water intake is so excessive that it exceeds the kidneys’ ability to excrete it. There may be other evidence of psychiatric disorder to give the clue, or the onset may be clearly related to a depressive phase or period of emotional stress. The onset will often be gradual rather than abrupt, and consumption may tend to fluctuate from hour to hour or day to day in contrast to the steadily increased intake of diabetes insipidus. Nocturnal polyuria will often prove to be absent.

Not infrequently, however, the distinction can be difficult, and such difficulty can persist during fluiddeprivation studies. In normal subjects fluid deprivation over an 8-hour period leaves the plasma osmolality unchanged, while the urine osmolality rises to twice that of the plasma (Hall 1983). In diabetes insipidus the plasma osmolality rises, but that of the urine remains relatively low. The test may indeed have to be discontinued if the patient loses more than 3% of body weight. In cornpulsive water drinking the initial plasma and urine osmolality are low, and the plasma osmolality rises to normal at the end of the test. However, the urine osmolality may fail to rise to twice that of the plasma, since the prolonged excessive water intake may have led to a secondary nephrogenic diabetes insipidus. Prolonged water deprivation (carefully monitored) for 2-4 days may be necessary to allow the return of normal renal function, or this may be even longer delayed. Hypertonic saline infusion with measurement of the plasma ADH response may ultimately help towards clarifying the diagnosis, likewise a carefully supervised therapeutic trial of desmopressin (Baylis 1996).

When compulsive water drinking is mistaken for diabetes insipidus and treated with ADH, hyponatraemia and symptoms of water intoxication (p. 559) can develop. Water intoxication also seems to be a special hazard in compulsive water drinking associated with psychosis numerous examples of such a complication have been described in schizophrenic patients, sometimes presenting acutely with vomiting, impairment of consciousness or fits (Jose et al. 1979; Khamnei 1984; Singh et al. 1985; Grainger 1992). Fatalities have been reported from time to time (Vieweg et al. 1985).

Page 526 CHAPTER 11

Both Jose et al. (1979) and Khamnei (1984) noted that a high proportion of patients with water intoxication secondary to polydipsia were psychotic, sometimes with evidence of inappropriate ADH secretion. In other cases there have been indications of enhanced renal sensitivity to vasopressin (Goldman et al. 1988; Emsley et al. 1989). How far such features may reflect hypothalamic or other disorders intrinsic to the psychosis is, however, uncertain. Multiple factors may often be at work, including the effects of medical illnesses or drugs (Fowler et al. 1977; Illowsky & Kirch 1988).

Special attention has been drawn to the role of psychotropic medications in leading to the syndrome of inappropriate ADH secretion (SIADH) (p. 559), including tricyclic antidepressants, tranylcypromine, fluoxetine, paroxetine, phenothiazines, haloperidol, thiothixine and carbamazepine (Sandifer 1983; Grainger 1992; Committee on Safety of Medicines 1994). In half of the cases reviewed by Sandifer, hyponatraemia had developed within a week of starting medication, usually but not always in association with SIADH. When hyponatraemia ’ is found in a patient taking psychotropic medication it is important to test the response to a water load while on and off the drug, after ensuring that the serum sodium has been restored to normal. The drug may then be exonerated as the cause.

It is clearly important to enquire for a history of polydipsia in any psychotic patient who presents with seizures or lowering of consciousness. Similarly the discovery of polyuria with a low urinary specific gravity should always lead to careful observation of the patient’s water intake. This may on occasion be skilfully concealed. Treatment must aim at restricting fluid intake, along with attempts to obtain maximal control of the patient’s psychiatric disorder. Where long-term fluid restriction proves to be impractical, or when an offending medication must be continued, demeclocycline has proved to be beneficial in controlling both the hyponatraemia and the polydipsia, even in the absence of classic SIADH (Illowsky & Kirch 1988).

Klinefelter’s syndrome

Klinefelter’s syndrome results from the presence of at least one additional X chromosome in the nucleus in the male. It may present with infertility or delayed sexual maturation, or the hypogonadism may be discovered on routine examination. The usual karyotype is 47 XXY, revealed on buccal smear examination. Other variants occur, however, and in mosaicism the abnormal cell line may be restricted to testicular or other tissue. Examination shows small testes and a variable degree of androgen deficiency, manifest as gynaecomastia or scanty beard growth. Azoospermia or oligospermia are always present and irreversible. The urinary gonadotrophin levels are raised.

Psychiatric features

A high prevalence of psychiatric disorder has emerged in this condition. Intelligence is often low, personality and behaviour are frequently abnormal, and there is a probable excess of psychotic illness. Some psychiatric features appear to be attributable to the endocrine disorder, others to be more directly related to the chromosomal abnormality.

Early reports of an excess of Klinefelter’s syndrome among patients in mental subnormality hospitals led to the view that severe impairment of intellect was characteristic. It is now appreciated, however, that even superior intelligence may occasionally be encountered (Swanson 8- Stipes 1969). The usual picture is of mild impairment only, though perhaps a quarter of patients presenting at infertility clinics fall within the subnormal range. The greater the number of additional X chromosomes in the karyotype the more severe the mental retardation (Forssman 1970).

It seems clear that genetic rather than hormonal factors are operative in reducing the level of intelligence. No relationship has emerged between the degree of hypogonadism and the IQ levels obtained; and patients with hypogonadism due to other causes tend to show intelligence within the normal range. Thus Pasqualini et al. (1957) andWakeling (1972) found a mean IQ of approximately 80 among their Klinefelter patients, compared with a mean of 100 among hypogonadal patients, despite a tendency towards more severe endocrine disorder in the latter.

Interesting findings have emerged from more detailed psychometric studies, which show impaired verbal abilities in comparison with non-verbal abilities. Thus Netley and Rovet (1982a) found mean verbal and performance WISC IQs of 85 and 101, respectively, in their sample of 33 children with XXY karyotypes. The verbal IQs were significantly lower than in unaffected sibling controls, and a ’sentence verification test’ confirmed the presence of specific language impairments. Similar results were reported by Graham et al. (1988), who also found expressive language disorder with problems in word finding and the production of syntax. Achievement on a variety of reading and spelling tasks was impaired. In many respects this pattern is the reverse of what is found in subjects with Turner’s syndrome (p. 528).

The altered intellectual functioning may be due at least


in part to impaired cerebral maturation or other brain abnormality consequent upon the genetic defect. A high prevalence of EEC abnormalities has been reported in the condition, chiefly slowed alpha frequencies but also slow wave dysrhythmias and paroxysmal features (Hambert & Frey 1964). Epilepsy is also commoner than chance expectation. Netley and Royet (1982b) found that almost a quarter of 33 children with the condition were lefthanded.

The personality in Klinefelter patients is sometimes abnormal. A variety of pictures has been described, ranging from markedh antisocial conduct to passivity and social withdrawal. Common descriptions are of patients lacking in drive and initiative, with severe restriction of interests and generally indolent, insecure and dependent. At the same time tolerance of frustration tends to be impaired, with explosive irritability and outbursts of aggression. Poor school and work records, marital instability and impoverished social relationships are said to be common.

Nielsen’s (1969) review showed histories of alcoholism in 6% and of criminal behaviour in 12% of patients. A small excess of XXY patients has emerged in surveys of institutions caring for severely disturbed criminals, along with the more usual excess of XYY or XXYY karyotypes (Swanson & Stipes 1969). Schiavi eta!. (1984), however, failed to support any excess of violent or aggressive behaviour among Klinefelter (or XYY) subjects in their comprehensive survey from Copenhagen. The prevalence of criminal convictions was slightly higher than for XY men, but this difference disappeared on controlling for intelligence and parental socioeconomic status. Moreover the great majority of the offences committed did not involve personal violence.

The endocrine disorder may play a part in hindering personality maturation and contributing to some aspects of personality difficulties. Thus patients with hypogonadism due to other causes are typically shy, timid and markedly lacking in drive. They tend, however, to show more stable histories and temperaments than Klinefelter patients, and lack anv excess of criminal behaviours. Wakeling (1972) compared 11 Klinefelter patients and nine other hypogonadal patients seen in a psychiatric hospital; both groups showed insecurity and low tolerance of frustration, but passivity was more marked in the hypogonadal patients and impulsive erratic behaviour in the Klinefelter patients. The latter, moreover, frequently had histories of prepubertal maladjustment, with a higher incidence of unsettled schooling, neurosis and behaviour disorder in childhood. It appears, therefore, that in Klinefelter’s syndrome dela\ ed cerebral maturation may make a contribution, over and above the androgen deficiency, in leading to poor social adjustment and disturbed personality functioning.

Sexual problems, as might be expected, are not uncommon. Potency tends to be low and to show an early decline, especially when features of hypogonadism are marked (Pasqualini et al. 1957). Androgen treatment can be successful in restoring libido and potency (Beumont et al. 1972). Occasional reports have described homosexuality, transvestism, exhibitionism and paedophilia in Klinefelter patients, but there is little to suggest that deviation is characteristic of the syndrome (Orwin et al 1974). Sexual pathology, when it occurs, probably again reflects the restricted personality development and incapacity for deep interpersonal relationships. The more extreme examples of deviant sexual practice have usually occurred in severely antisocial or psychotic individuals.

Mental hospital surveys, reviewed by Forssman (1970), have indicated a threefold increase in Klinefelter patients compared with the general population. This appears mainly to be due to psychotic illnesses of a schizophrenic nature. Nielsen (1969) found that 6% of patients recorded in the psychiatric literature had been given a diagnosis of schizophrenia, and another 7% had psychoses of an uncertain type but almost all with paranoid delusions. Well-documented examples of schizophrenia in association with Klinefelter’s syndrome are provided by Pomeroy (1980) and Roy (1981). The increased risk of mental illness may represent another facet of the increased vulnerability to stress of the Klinefelter patient, or may have more direct genetic determinants.

There is little to suggest an increased incidence of organic psychiatric illness. Jablensky et al. (1970) have described a patient who demented rapidly in his early forties, showing diffuse white matter degeneration and adrenal cortical atrophy at autopsy; this, however, may well have represented a chance association with Klinefelter’s syndrome.

Turner’s syndrome

The XO karyotype is associated with oestrogen deficiency and hence failure of sexual maturation. Primary amenorrhoea is accompanied by short stature and a variety of skeletal abnormalities including cubitus valgus and arching of the palate. The facial appearance is often characteristic with a small jaw, fish-like mouth and low set ears. The neck is short and may be webbed. Congenital renal abnormalities and coarctation of the aorta are common. The classic case shows a 45 XO karyotype, though other X chromosome abnormalities occur. The concentration of follicle-stimulating hormone (FSH) in the urine is raised.

Page 528 CHAPTER 11

Psychiatric interest in the condition has largely centred on the cognitive functioning of such patients. Mental retardation, chiefly mild in degree, was once regarded as common, but it now seems probable that verbal intelligence, at least, is normally distributed. Superior intelligence can certainly be encountered.

Money (1963, 1964) drew attention to the common finding of lower performance than verbal intelligence in patients with the syndrome, the means in his sample being 88 and 105 respectively (on the WISC or WAIS). This was paralleled by inferior scores on tests of perceptual organisation when compared with scores on verbal comprehension – discrepancies which became more marked at the higher intelligence levels. Visuospatial ability appeared to be particularly poor.

In some respects this disparity can be seen as an accentuation of the usual female as opposed to male pattern of differential cognitive abilities, which brings interest to Money’s suggestion that it might be a specific cognitive consequence of the abnormal chromosomal condition. Garron and Vander Stoep (1969) review further evidence that Turner’s patients have poor ability at drawing geometrical designs from memory, at drawing human figures and in certain aspects of left-right orientation.

The more recent study of 18 patients by Murphy et al. (1993) has confirmed significantly greater verbalperformance discrepancies than in controls, with lower scores on most aspects of the WAIS except for verbal comprehension. The greatest difference was in visuospatial ability. Among a subset of patients with XO mosaicism, visuospatial ability was negatively correlated with the percentage of lymphocytes containing the XO karyotype.

It remains uncertain how far such deficits reflect impairment of cerebral maturation due to genetic or endocrine factors, but quantitative MRI analysis of specific brain regions has shown interesting differences from controls (Murphy et al. 1993). Even after allowing for differences in head size the Turner’s syndrome patients showed significantly smaller cerebral hemisphere volumes, parieto-octipital brain matter volumes, and smaller hippocampi, lenticular nuclei and thalamic nuclei bilaterally. On many of these measures the mosaic patients occupied an intermediate position. Among the group as a whole there was a significant right-left asymmetry in parieto-occipital brain matter volumes, that on the right being reduced. Occasional reports have suggested the presence of neuropathological abnormalities at autopsy, including atrophy, white matter heterotopias or small focal infarcts, chiefly affecting the right side of the brain (Reske-Nielsen et al. 1982).

With regard to personality, feminine sexual identification and interests are usually normal, and gross psychopathology appears to be rare (Garron & Vander Stoep

1969). Libido tends to be low, however, and many descriptions stress childish, meek and overcompliant behaviour. Traits of passivity and immaturity in the personality are congruent with failure of sexual maturation (Kihlbom 1969; Nielsen 1970b). Such features may also be partly determined by the infantilising response called forth by the patients’ short stature and child-like appearance.

Neurotic traits may emerge, but severe emotional disorder appears to be rare despite the handicaps imposed by the physical defects and sterility. In contrast to the situation in Klinefelter’s syndrome, there is no suggestion of an increased prevalence of antisocial behaviour or of psychotic illness. It would seem, therefore, that the lack of an X chromosome has substantially less effect than the possession of an additional X chromosome where mental health is concerned (Forssman 1970).


Hyperparathyroidism has gained recognition as a rather rare but important cause of psychiatric morbidity. It is important because the diagnosis may be missed, resulting in many years of chronic mental ill health, yet treatment of the endocrine disorder can bring prompt relief. The cause is usually a benign adenoma of one of the parathyroid glands. Sometimes multiple tumours are present, and occasionally the condition may occur familially. More rarely there may be diffuse hyperplasia of all parathyroid tissue. In ’multiple endocrine adenomatosis’ (’pluriglandular syndrome’), parathyroid adenomas are accompanied by endocrine tumours of the pancreas and pituitary (MEA type 1) or by phaeochromocytomas and medullary carcinoma of the thyroid (MEA type 2). Secondary hyperparathyroidism can result from renal failure due to elevated parathormone levels and impaired activation of vitamin D.

Women are affected more often than men. Cases usually present in middle age though the range of onset is wide. Calcium and phosphorus are mobilised from the bones and excreted in excess in the urine.

Physical features

Renal calcification is present in about two-thirds of cases in the form of renal calculi or diffuse nephrocalcinosis. The typical X-ray changes of osteitis fibrosa are present in many of the remainder.


In the great majority physical complaints are the predominant feature, with pain, fracture or deformity of bones, renal colic or profound muscular weakness. The myopathic syndrome consists of proximal muscular weakness and wasting, hypotonia and discomfort on movement. Other common symptoms which may suggest the condition are increased thirst, polyuria, dull diffuse headache, anorexia and nausea. On examination corneal calcification may be seen close to the corneoscleral junction as linear aggregations of granular material.

Psychiatric features

Mental symptoms are also common and were found in two-thirds of Petersen’s (1968) series, even after excluding patients who had been referred specifically because of psychiatric disturbance. In a third the mental abnormalities were severe. Watson (1968) found that a very small but important group presented with mental symptoms alone, and showed neither renal stones nor bone disease. In Karpati and Frame’s (1964) series four out of 33 patienfs had psychiatric complaints which dominated the picture to the extent that they had been referred initially for psychiatric or neurological consultation:

A woman of 40 presented with depression which had proved resistant to drugs and psychotherapy for several years before hyperparathyroidism was diagnosed. A woman of 64 had a 2year history of agitated depression with tremulousness, disorientation, confusion and severe headache. A man of 43 presented with increasing nervousness and obsessive-compulsive features which subsided after operation. The fourth patient presented with a conf usional state accompanied by severe headache.

Gatewood etal. (1975) reported five further examples, all presenting with problems that seemed to be mainly psychiatric, and four of them showing no evidence of bone or renal pathology:

A 63-year-old man developed persistent confusion following a cholecystectomy, which subsided within 2 weeks of discovering and removing a parathyroid adenoma. A 65-year-old man presented with a 14-monrfi history of progressive depression, fatigue, lethargy and periods of confusion, and was similarly cured. A 56-year-old woman had been treated for 4 months for catatonic schizophrenia and improved gradually without medication after operation. A 74-year-old woman with endogenous depression similarly recovered without antidepressants. And a 75-year-old diabetic with a recent history of syncope, confusion and drowsiness made a remarkable recovery from what had initially been thought to be a cerebrovascular accident.

The commonest mental change is depression with anergia. The patient gradually becomes tired, listless and dull, with marked lack of initiative and spontaneity. In

Petersen’s (1968) series 36% of patients showed such changes, and almost all of these had been unable to work on account of lack of energy during the months preceding operation. Tension and irritability sometimes accompanied the depression, and explosive outbursts were occasionally seen.

Even among patients who do not complain of such symptoms at the time, they can often be recognised retrospectively when operation has restored the metabolic state to normal. Anderson (1968) found that threequarters of patients reported that they felt better postoperatively, with higher spirits and greater energy than for many years before. It seemed that the chronicity of the disorder, which had often been present for 10 or’more years when diagnosed, had made it difficult for the patient to appreciate the mental changes subjectively at the time.

Organic mental symptoms were present in 12% of Petersen’s cases, chiefly impairment of memory or general mental slowing. This may be an insidious and chronic development, or may herald an acute organic reaction as part of a ’parathyroid crisis’. Such acute organic psychoses occurred in 5 % of cases, with spells of mental confusion, or acute delirious episodes with hallucinations, paranoia and aggressive behaviour. Stuporose states may also occur, or recurrent convulsions leading to coma. Hockaday et al. (1966) described a patient who presented with stupor, and Cooper and Schapira (1973) reported a patient in whom stupor supervened during the course of a depressive illness. Both of these showed flexibilitas cerea at some point in their course.

Numann et al. (1984) have shown that even in patients without specific neuropsychiatric complaints operation can be followed by significant improvement on tests of verbal memory and cognition. Very occasionally the degree of intellectual impairment can give rise to a mistaken diagnosis of presenile dementia.

Non-organic psychoses appear to be rare in hyperparathyroidism, and when present are probably coincidental. However, a clear example of paranoid psychosis was reported by Alarcon and Franceschini (1984); the florid psychosis was unaccompanied by organic features in the mental state and subsided rapidly after removal of a parathyroid adenoma. Kleinfeld et al. (1984) refer to a case with mania as the sole clinical manifestation.


Confirmation of the disease is usually obtained by finding a raised serum calcium. Repeat estimations may sometimes be required. The serum phosphate may be low but is

Page 530 CHAPTER 11

sometimes normal. Blood must be taken while the patient is fasting and without venestasis, and account must be taken of the serum albumin level. Radioimmunoassay for parathyroid hormone levels can provide further confirmatory evidence, though a normal result does not exclude the condition. A hydrocortisone suppression test can also be useful -in hyperparathyroidism the administration of steroids usually fails to lower the plasma calcium, whereas this occurs in hypercalcaemias of other origin.

The serum alkaline phosphatase is raised when the bones are involved. Renal stones or calcification may be detected on X-ray, and typical changes may be seen in the bones. The hand X-ray can be particularly informative. Radiography of the skull may occasionally show calcification in the caudate nuclei and frontal lobes, though this is very much less common than in hypoparathyroidism since the calcium deposits are usually finely distributed. The EEG shows widespread slow activity, sometimes with paroxysms of frontal delta waves at high levels of serum calcium.

Differential diagnosis

The disorder should be borne in mind in patients who show chronic affective disorder, neurotic disability or minor intellectual impairment in association with suspicious physical symptoms. Neurotic ill health together with polydipsia and polyuria is a not uncommon mode of presentation. Petersen (1968) suggested that hyperparathyroidism should always be considered ’when lack of initiative, depression and thirst appear during a prolonged, insidiously developing and diagnostically unclear change of personality’.

Patients with fluctuating confusion or delirium may cause special diagnostic difficulties. When the acute organic reaction is recognised the main hazard is concentrating on a fruitless search for some intracranial cause (Henson 1966). In the presence of stupor, the electroencephalographic finding of widespread slow waves may provide the important clue to the metabolic derangement (Moure 1967; Cooper & Schapira 1973).

Neurological disorders are more frequently simulated. Headache, vomiting, fits and drowsiness can lead to a suspicion of cerebral tumour, or profound muscular weakness may suggest a primary muscular disorder. Cerebral arteriosclerosis, subdural haematoma, uraemia and phaeochromocytoma are other misdiagnoses that have been reported. Patten and Pages (1984) describe two patients who presented with severe muscle weakness, atrophy and fasciculation, one being first diagnosed as having amyotrophic lateral sclerosis. They suggest that all patients with this condition should be checked for hyperparathyroidism.

Aetiology of mental disturbances

The cause of the psychiatric disturbance appears to lie chiefly or even exclusively with the elevation of serum calcium. Unlike the other endocrinopathies a relatively straightforward quantitative relationship is found between the severity of the psychological disturbance and this simple measure of serum chemistry. In Petersen’s (1968) careful review, affective disorder and disturbances of drive corresponded to a serum calcium of 12-16 mg/ 100ml, acute organic reactions with florid delirium appeared at 16-19 mg/100 ml, and somnolence and coma were found with levels exceeding 19 mg/100 ml. Such a sequence of changes could sometimes be traced in the single patient. Bleuler (1967) pointed out that the psychiatric disorders of hyperparathyroidism are more constant from person to person, and less dependent on the dynamics of the personality, than are the psychiatric pictures seen with most other endocrine disorders. This is no doubt because they depend upon a widespread ionic intermediary, and not upon the direct cerebral effects of a hormone which can influence brain functions in a more complex manner.

The level of circulating parathormone does not appear to be directly responsible, since mental symptoms can improve rapidly when the serum calcium is lowered by peritoneal dialysis (Petersen 1968). Nor can a relationship be discerned with the level of serum phosphorus or serum alkaline phosphatase, or with the duration of the disorder. The possible role of hypomagnesaemia remains unclear. Other factors may make a contribution, such as hypertension or renal failure due to nephrocalcinosis, but in the majority of cases these are clearly of subsidiary importance (Karpati & Frame 1964).


Removal of the parathyroid adenoma usually brings relief to disorders of affect and drive, also to acute organic psychoses. The mental disorder is commonly found to be wholly reversible, with rapid resumption of former mental health. Headache is abolished and muscular strength increased. The time to recovery has been found to be independent of the duration of the disease and of the severity of the mental changes, and to parallel closely the fall in serum calcium. The rare psychotic states of long duration, with thought disorder and paranoia, may respond less satisfactorily, but probably owe a good deal to premorbid vulnerability. With severe depressive illness antidepressant


medication may be required to obtain complete resolution, as in a patient described by Noble (1974):

A woman of 50 developed a severe depressive illness for the first time in her life. This responded well to electroconvulsive therapy. Eighteen months later she became apathetic and retarded, failed to respond to antidepressants, and during a course of electroconvulsive therapy became dehydrated and incontinent. A parathyroid adenoma was discovered while this was being investigated. Its removal, however, left her profoundly apathetic and unwell, despite the return of the plasma calcium to normal. After 4 weeks she was started again on tricyclic antidepressants which now led to progressive and full recovery.

Postoperatively, care is needed to guard against hypocalcaemia and plasma calcium should be monitored daily for the first few days. Preoperative preparation is required in patients who have significant bone disease, with the administration of vitamin D and calcium supplements (Kanis 1996). Acute anxiety may herald tetany, usually between the 10th and 14th postoperative days. Sometimes short-lived psychiatric disturbance sets in within a few days of the operation even though hypocalcaemia cannot be demonstrated, presumably occasioned by the abrupt drop in extracellular calcium which has occurred. Karpati and Frame (1964) reported such examples showing catatonia, acute agitation or mental confusion. Mikkelson and Reider (1979) described patients with fluctuating paranoid psychoses, hallucinations and stupor, with onset 4-12 days after operation and subsiding within a few days or weeks.


Hypoparathyroidism, like hyperparathyroidism, has come to be recognised as a cause of remediable psychiatric disorder, especially since the comprehensive survey of the literature by Denko and Kaelbling (1962). This is replete with examples of failure to diagnose the condition, sometimes over very many years, yet treatment offers an excellent chance of reversing both the physical and psychiatric changes.

The commonest cause is removal of the parathyroid glands at thyroidectomy, or interference with their blood supply in the course of other operations on the neck. In other cases the aetiology is obscure; the parathyroids are found to be absent or degenerated, sometimes in more than one member of a family and occasionally in association with Addison’s disease (idiopathic hypoparathyroidism). An autoimmune basis is suspected. The deficiency of parathormone leads to a low serum calcium and a raised serum phosphate. Calcium deposits may occur in the skin and the brain. ,

Two allied conditions occur more rarely:

In pseudohypoparathyroidism the parathyroid glands function normally but the tissues are resistant to the effects of parathormone. The same abnormalities are found in the serum chemistry despite elevated plasma hormone levels, and calcium deposits are found similarly in the soft tissues including the brain. It is due to loss of the actions of parathormone on calcium and phosphate, and the reduced formation of 1,25-dihydroxyvitamin D with consequent defects in mobilisation of calcium from bone and reduced calcium absorption from the gut (Spiegel 1989). Various subtypes are recognised according to the associated physical features and the location of the putative defect leading to hormonal resistance. Type la (Albright’s osteodystrophy) is usually inherited as an autosomal dominant, showing additional traits of short stature, shortening of some of the metacarpals and a characteristic rounded fades. This normally presents early in life and may show resistance to multiple hormones. Type Ib usually shows a normal physical appearance, and resistance is limited to parathormone. It can be both sporadic and familial. Type n is rarely if ever familial and shows a different form of resistance. Parathyroid hormone infusion produces a normal rise in urinary cyclic adenosine monophosphate (cAMP) but a blunted phosphaturic response.


is identical to Albright’s osteodystrophy in the associated traits, but hormonal resistance is lacking so the serum calcium and phosphate are normal. Spiegel (1989) discusses the confusion that has surrounded the term. It is best restricted to the relatives of patients with type la pseudohypoparathyroidism who show the associated physical features of Albright’s osteodystrophy but produce a normal rise in urinary cAMP excretion in response to parathyroid hormone infusion. Diagnosis on the basis of serum calcium alone can be misleading since patients with hormonal resistance are sometimes normocalcaemic. Moreover the severity of hormonal resistance can vary widely within a family and may represent a polygenic trait.Physical features

Hypoparathyroidism should be suspected in patients with symptoms of chronic tetany, or when ocular cataracts develop at an unusually young age. A history of operation on the neck should bring the possibility of the condition to mind. Symptoms can sometimes be present for many years before it is diagnosed, as in the remarkable example reported by Bellamy and Kendall-Taylor (1995). Thyroidectomy in childhood had been followed by epilepsy from the age of 27 and episodes of nocturnal stridor 10 years later, yet the patient’s hypocalcaemia was detected only at the age of 47 after some years of progressive weakness accompanied by paraesthesiae and spasms in the hands and feet. ,

Tetany occurs in the form of numbness and tingling in the hands and feet or around the mouth. With more severe degrees the patient experiences muscular cramps

Page 532 CHAPTER 11

and stiffness in the limbs, carpopedal spasms or laryngeal stridor. In carpopedal spasms the metacarpophalangeal joints are flexed and the interphalangeal joints of the thumb and fingers are extended to produce a characteristic posture of the hand (’main d’accoucheur’). Epilepsy can be the first and sometimes the only manifestation. In addition to cataracts the patients may have a dry coarse skin, scanty hair, trophic changes in the nails and poor dental development. Calcium deposits may be detected in the skin, or appear on skull X-ray as calcification in the region of the basal ganglia. Clinically useful signs include twitching of the facial muscles on tapping the facial nerve below the zygoma (Chvostek’s sign), and the production of carpopedal spasm by temporarily occluding the circulation to the arm (Trousseau’s sign). Very occasionally papilloedema may be observed.

Psychiatric features

A wide variety and a high incidence of psychiatric disturbances have emerged in hypoparathyroidism. Denko and Kaelbling (1962) estimate that at least half of the cases attributable to surgery have psychiatric symptoms, and that the frequency is probably higher still in idiopathic hypoparathyroidism.

The most frequent disturbances are organic psychiatric syndromes. Acute organic reactions with features typical of delirium are prone to develop in surgical cases, where the biochemical changes are likely to be abrupt. More chronic and insidious developments are not uncommon in idiopathic hypoparathyroidism, where the biochemical changes have developed gradually and have been much longer in operation. Such patients may show sustained difficulty with concentration, emotional lability and impairment of intellectual functions. Robinson et al. (1954) reported a patient with idiopathic hypoparathyroidism who presented first with status epilepticus and seeming presenile dementia, which illustrates the clinical problems involved:

For 7 years a woman of 61 had suffered from depression, cornmencing shortly after the death of her husband, and had gradually lost interest in her appearance and surroundings. Sometime after the onset bilateral cataracts had been removed. For several years she had experienced occasional numbness and tingling in the legs, and some 3 years previously skull X-ray had shown calcification in the basal ganglia. However, she had not reattended hospital for follow-up at this time. For 2 years there had been episodes of urinary incontinence, and for 6 months ’fainting spells’ in some of which twitching of the limbs had been observed. For 5 weeks there had been considerable mental deterioration with confusion and loss of memory.

She was admitted to hospital with status epilepticus which subsided with treatment, and she was then found to be disorientated, apathetic and doubly incontinent. Evidence of self neglect was extreme. She showed dysarthria, fine lateral nystagmus, diminished tendon reflexes and feeble extensor plantar responses. On the 10th day there were attacks of tetany and carpopedal spasm and Chvostek’s sign was positive. The electrocardiogram showed prolonged Q-T intervals and low T waves. She was treated with intravenous calcium gluconate, oral dihydrotachysterol (A.T.10) and calcium lactate. Within a few days she had improved, becoming continent, orientated and taking an intelligent interest in her surroundings.

She remained well and her mental state did not deteriorate, but 3 months after treatment she developed choreiform jerks of the limbs and twitching in the face, presumably as a result of lesions in the calcified basal ganglia.

An important point stressed by the authors was that the patient showed no evidence of tetany on clinical examination until 10 days after her acute presentation.

Another remarkable example was reported by Eraut (1974):

An 80-year-old man was admitted to hospital on account of numerous falls at home. He could give no account of himself, but his wife reported that he had been deteriorating for many months and had been confused and forgetful for a long time. Cataracts had been extracted 14 and 20 months previously. While in hospital the previous year mild dementia had been noted. He was disorientated in time and place, could answer simple questions but could not respond to commands. There were no neurological abnormalities. A chest infection responded to antibiotics, but he remained demented. Three weeks after admission he had a grand mal fit, and when started on phenytoin became tremulous and totally unresponsive. This was followed by tetanic spasm of the left hand and laryngeal stridor, whereupon tests confirmed the presence of idiopathic hypoparathyroidism.

Treatment with dihydrotachysterol led to considerable improvement within a few days. He was discharged well recovered mentally, capable of lucid conversation and showing reasonable memory for recent and past events.

Tseudoneurosis’ is described as the next most common change, both in surgical and idiopathic hypoparathyroidism, and occurs in all age groups. Children show temper tantrums and night terrors, and adults become depressed, nervous and irritable with frequent crying spells and marked social withdrawal. The emotional disturbances may fluctuate in degree or show periods of spontaneous resolution.

In this connection the concept of ’partial parathyroid insufficiency’ is interesting. Fourman etal, (1963) showed that in a quarter of patients who had undergone thyroidectomy the plasma calcium was merely at the lower limit of normal, but could be provoked to fall to definitely subnormal values by calcium deprivation or intravenous administration of edetic acid. About half such patients had mental symptoms in the form of tension and anxiety, panic attacks, depression and lassitude. Often there were


no other pointers to parathyroid insufficiency and the symptoms were therefore indistinguishable from those commonly found in middle-aged neurotic women. Fourmanetal. (1967) later assessed the relevance of such symptoms by a double-blind trial of calcium citrate tablets and placebo, and confirmed that calcium was significantly effective in reducing psychiatric symptom scores. The most consistent changes were with regard to depression and diminution of appetite.

More rarely, psychotic illnesses of manic-depressive or schizophrenic type may be seen, particularly in cases due to surgery. Again spontaneous remissions or response to other forms of treatment may delay diagnosis of the underlying condition.

In pseudohypoparathyroidism and pseudopseudohypoparathyroidism intellectual impairment is by far the most frequent psychiatric abnormality, occurring in approximately half of reported cases. Several such patients have been discovered in hospitals for the feeble minded, and Denko and Kaelbling (1962) suggest that the serum calcium should be investigated in every patient with mental retardation. Pollard et al. (1994) have recently reported a 13-year-old girl of normal intelligence with type II pseudohypoparathyroidism who presented with an episode of apparent hysterical paralysis and rapid cycling bipolar mood disorder.


The serum calcium is low, the serum phosphate raised, and the urinary excretion of calcium and phosphate diminished. The serum alkaline phosphatase is normal. Skull X-ray frequently shows calcification in the region of the basal ganglia as symmetrical bilateral punctate opacities. Electroencephalographic abnormalities may be present even in the absence of epilepsy, usually generalised but sometimes surprisingly focal (Watson 1972).

In pseudohypoparathyroidism the same abnormalities of serum chemistry are found, but a distinction can be made by examining the effect of an infusion of parathyroid hormone on the excretion of phosphate and cAMP in the urine.

Differential diagnosis

The diagnoses which may be mistakenly entertained include mental retardation, presenile dementia, neurosis, hysteria, idiopathic epilepsy and cerebral tumour.

A diagnosis of neurosis or hysteria is suggested by the peculiar and intermittent nature of the symptoms, including bizarre paraesthesiae and muscular spasms. Moreover the patient may give a vague and perplexing account with obvious difficulty in observing and describing the symptoms. Attacks may be triggered by emotional influences, since hyperventilation will lead unusually readily to tetanic symptoms. Hypochondriasis is readily suggested by the generally heightened level of anxiety, the vagueness of the complaints and the occurrence of periods of spontaneous remission. As a result patients with hypoparathyroidism are sometimes found to have carried a label of psychogenic disorder for several years before the true diagnosis is made. In other cases well-defined mood swings have led to an initial diagnosis of manic-depressive disorder (Denko & Kaelbling 1962).

Epileptic attacks may be thought to be idiopathic in origin, and the serum calcium should be determined in every epileptic patient when the precise cause of the attacks remains uncertain. Occasionally increased intracranial pressure and papilloedema are encountered in patients with hypoparathyroidism, this reversing with correction of the serum calcium; cerebral tumour may be closely simulated, especially when fits are present and alteration of personality has occurred.


The response to correction of the serum biochemistry is usually gratifying. Neurotic symptoms are reported to clear in the majority of cases even though some weeks may elapse before the patient feels entirely well.

Acute organic reactions may be expected to improve promptly. Chronic cognitive impairments may also be completely reversed. Denko and Kaelbling (1962) noted that when adequate details were given, about half the cases of idiopathic hypoparathyroidism with intellectual impairment were reported to improve whereas very few cases were unchanged or worse. Patients with pseudohypoparathyroidism may also improve cognitively when the serum chemistry is corrected, but rarely to a spectacular extent. This is perhaps because such patients have been damaged intellectually while still immature, or perhaps because there is an associated genetic cause for their cognitive impairment.

Details of long-term management will not be dealt with here. This usually requires vitamin D or equivalent preparations such as dihydrotachysterol in addition to oral calcium.

Diabetes mellitus

Diabetes mellitus results from an absolute or relative deficiency of insulin production by the pancreas, causing disturbed carbohydrate metabolism with hyperglycaemia and glycosuria. Secondary changes are prone to occur in

Page 534 CHAPTER 11

the metabolism of protein and fat, the latter leading to ketosis and acidosis. It is a syndrome rather than a disease entity. The requirements of insulin may be found to be in excess of normal pancreatic production, owing to cellular resistance to the action of insulin or excessive gluconeogenesis. Hence it is probable that factors other than decreased insulin production contribute to the severity of diabetes in many patients, with the anterior pituitary and adrenal glands also playing a part.

The onset can be at any age from infancy to old age, with approximately half of cases appearing before 50 and another quarter between 50 and 60 years. The disorder tends to be rapid in onset in the young, but usually insidious in development and milder in older persons. A division is made into type I (insulin dependent) and type II (non-insulin dependent) diabetes. Type I typically occurs in young thin patients, shows absolute insulin deficiency and is ketosis prone; type n is usually found in older obese patients, shows relative insulin lack and tissue resistance to insulin effects, and is not ketosis prone (Windebank & McEvoy 1989).

It is well established that genetic mechanisms are operative in both forms of the disease, and interesting progress has recently been made in relation to type I diabetes (Bennett etal. 1995; Kennedy etal. 1995; Bell & Hockaday 1996). Multiple loci are clearly involved, including loci on chromosome 6 in the human leucocyte antigen (HLA) region, and on chromosome 11 in close relation to the insulin gene itself. The insulin gene is flanked upstream by multiple repeats of a 14 base pair sequence, variations in length of the sequence correlating with disease susceptibility, perhaps through a direct effect on transcription of the insulin gene. Genetic influences in type II diabetes are especially apparent hi patients who are not overweight, including, it seems, changes in mitochondrial DNA.

Textbooks of medicine should be consulted for the general clinical associations of the disorder and the principles of management by diet, insulin and oral hypoglycaemic agents.

Psychiatric attention to diabetes has been sporadic, and few systematic surveys have been made of the emotional and other mental complications. However, valuable reviews were produced by Wilkinson (1981) and Tattersail (1981). In clinical practice it is clear that the diabetic who is poorly endowed or emotionally unstable can pose a considerable therapeutic problem, since cooperation in treatment is essential if adequate control is to be achieved. Moreover, there are indications that psychological stresses can be important in aggravating the disorder or precipitating episodes of loss of control, and there are even suggestions that emotional factors may sometimes bring the disorder into being.

These issues will be discussed below, along with the question of brain damage in diabetic patients. When evidence of brain damage emerges this may be attributable to episodes of hypoglycaemia or diabetic coma, or alternatively to the high incidence of atherosclerosis which exists in diabetics. The picture of diabetic coma, and certain common neurological complications, will also be briefly described.

Emotional influences on the course of diabetes

A considerable body of evidence shows that emotionally stressful experiences can produce fluctuations in levels of blood glucose and ketone bodies, both in diabetic and non-diabetic persons. Early experimental observations in this area are summarised by Hinkle and Wolf (1952a, 1952b). The magnitude of such changes is much greater in diabetics, and if of long enough duration they appear capable of leading to ketosis and hyperglycaemia in some cases and to hypoglycaemia in others. It has proved difficult, however, to define the extent to which such factors may be operative in the actual disease.

Stress, either physical or emotional, has often been blamed as the initial cause. Examples of the sudden manifestation of diabetes in relation to dramatic stresses are scattered throughout the literature, also attempts at constructing a characteristic personality profile which has rendered the diabetic unusually susceptible to stress. Treuting (1962) reviews the theories that have been elaborated, suggesting for example that diabetes is a disorder of adaptation and that persons showing it have reacted to various life crises with a physiological response that is appropriate to starvation. However, no increased incidence of the disorder emerged in battle casualties from the First and Second World Wars, and it now seems most unlikely that stress can bring the disorder into being in people who would otherwise never have developed it. It remains possible, nonetheless, that stress may sometimes change a latent case of diabetes into an active one, i.e. that physical or emotional stresses may play a part in determining the time at which the disorder is declared.

A rather less controversial area is the effect of emotional influences on the course of the established disease. Diabetic patients may sometimes show a close relationship between disturbing life experiences, and episodes of loss of control, even to the extent of developing ketotic coma. Hinkle and Wolf (1952a, 1952b) observed 64 diabetics, giving special attention to their prevailing attitudes and the relationships important to their emotional security. In long-term studies extending over several years, periods of exacerbation and remission were correlated with events in the life situation. Events which were consciously


or unconsciously interpreted by the patients as threatening to their security appeared to be particularly liable to lead to loss of control. Case histories were presented to illustrate how admissions to hospital for coma could regularly follow stressful situations, as in the following example:

An adolescent girl from a disturbed home background had 12 admissions for diabetic acidosis and coma during 5 years, all following acutely stressful life situations. ’To each of these stressesfights between her parents, arguments with her mother, change to a new school, the departure of her sister (“the only one who loved me”)-she reacted as if it were a threatened deprivation of love and security. They aroused in her resentment, which she felt afraid to express, and were accompanied by the rapid development of thirst, polyuria, ketosis and coma. On several occasions . .. she expressed her hopelessness and rebellion by stopping her insulin when the ketosis developed. On other occasions she expressed her resentment and hopelessness by failing to sterilise her equipment, and the subsequent infections led to hospital admissions. In half of these instances, however, diabetic coma followed swiftly upon the onset of a stressful life situation, despite the fact that no infection was present and the insulin dose was not altered. It may be remarked in passing that this patient was typical of the group in that the exacerbations of the diabetic state were in all instances closely related to situational and interpersonal conflicts …’. Close supervision and help with her emotional problems led in this case to better control of the diabetes. (Hinkle & Wolf 1952b)

The uncertainty in such examples concerns the extent to which the direct metabolic consequences of the emotional upheavals are responsible, rather than secondary effects due to abandoning dietary regimes or insulin requirements. Thus some patients may overeat or resort to alcohol when under stress, and others may omit insulin or neglect sterile precautions. In occasional cases comas may be deliberately induced to secure attention or as a means of escape into the shelter of hospital. Indeed, Tattersall (1981) has termed insulin-dependent diabetes ’the manipulator’s delight’.

Hinkle and Wolf, however, argued strongly for a more direct influence of life experiences upon the metabolic disorder and hence on the course of the disease. When it appeared that a certain personal conflict was connected with variations in the diabetic state, they tested this in short-term experimental settings. Baseline observations were made, then the suspected topic of conflict was vigorously introduced into discussion. Control studies were made in which neutral topics were discussed. In this way psychological stress was shown to lead to ketonaemia and increased water, glucose and chloride excretion, in addition to alterations in fasting blood sugar levels.

Baker and Barcai (1970) felt able to demarcate a small number of ’super-labile’juvenile diabetics in whom emotional arousal led directly to ketoacidosis, mediated by an increased ketone response to endogenous catecholamines. Beta adrenergic blockade was apparently successful in inhibiting the metabolic decompensation in such patients, producing marked therapeutic benefit. Tattersall (1981), however, doubts whether organic a priori causes are common. In the usual ’brittle diabetic’, whose life is constantly disrupted by episodes of hyper- and hypoglycaemia, physiological, psychological and social problems come in time to be inextricably intertwined. Depression and reactions of frustration and futility breed carelessness in self management and distrust of prescribed routines. An emotional origin often appears to be paramount in leading to this vicious circle.

Whatever the mechanisms, many agree that life experiences and emotional factors can have an important bearing on the course of diabetes, and that this is particularly important in juveniles and adolescents (Treuting 1962). Attempts at demonstrating the influence of life events have indeed been partially successful despite difficulties in methodology. Grant et al. (1974) studied 37 adults over 8-18 months, and found a trend towards an association between important life events, particularly those of an unpleasant nature, and fluctuations in diabetic control. Bradley (1979) reviewed 114 patients retrospectively, and found a significant association between the number of life events experienced over a 12-month period and the incidence of glycosuria, changes in prescription and number of clinic attendances. Insulintreated patients appeared to be more vulnerable in this regard than those receiving oral hypoglycaemic agents. Accordingly it may sometimes be necessary to pay careful attention to psychological and psychosocial aspects of the patient’s situation if optimal control of the diabetes is to be achieved.

Psychological problems

In several ways the situation imposed by diabetes is unusual in comparison with other chronic diseases. The patient must often face interminable dietary restrictions and daily self-administered injections, yet is usually symptom-free so gets no perceptible reward. He has responsibility, which is rare in other illnesses, for judging unusual situations and adjusting the dose of insulin required. Repeated hospitalisations can stigmatise him> from an early age. Such factors can contribute to neurotic developments or disturbed family relationships, and hypochondriacal attitudes may come to be established.

Certain types of behaviour are said to be common in diabetics, but many are of a nature which would be unremarkable but for the fact that they complicate therapy.

Page 536 CHAPTER 11

Thus some show an unusual need to eat and find great difficulty in adhering to dietary regimes. This may be intensified during periods of loneliness, depression or tension. Explosive rebellion may be seen in adolescents, with wilful neglect of treatment.

In children the disorder lends itself to incorporation in disturbed parent-child relationships. The anxiety of the parents may be transferred to the young child, or a perfectionistic mother may gain control over the illness at the cost of behaviour difficulties. The child on his part may utilise the diabetes to manipulate the home environment, using food as a weapon or form of retaliation. Surveys of diabetic children have been undertaken to explore possible psychological effects in detail, though with varying results (Sterky 1963; Swift etal. 1967; Gath et al. 1980). In general adjustment has seemed remarkably good, though the sample investigated by Swift et al. showed minor abnormalities on measures of dependence/independence, self perception, manifest and latent anxiety and sexual identification.

In adult life employment or marriage prospects may stand to be affected. Pruritis and decreased sexual interest may contribute to emotional complications, and impotence and amenorrhoea can be early complaints even in undiagnosed diabetics. Surridge etal. (1984) found indications of delayed psychosexual development when diabetes sets in at an early age. Earlier reports of the frequency of impotence in men and anorgasmia in women (Kolodny 1971; Kolodny et al. 1974) may have been an overestimate, but there can be little doubt that such problems often occur. Disturbances of ejaculation are also probably common (Fairburn et al. 1982). The physical handicaps later imposed by ocular and other complications bring further problems of their own.

A major fear among many insulin takers is the occurrence ’of a hypoglycaemic attack. They particularly dread attacks which lack the adrenergic warning, and in which loss of self control or bizarre behaviour may occur. It now seems clear, moreover, that ’subclinical’ hypoglycaemia is commoner than previously suspected in diabetics, as a result of overtreatment, and that this may account for considerable chronic disability.

Thus Gale and Tattersall (1979) found, from overnight metabolic studies, that nocturnal hypoglycaemia occurred in 22 of 39 poorly controlled insulin-treated patients. This was often sustained for periods of several hours. Overt hypoglycaemic symptoms by day had been very mild or absent. The features occurring in such a situation included lassitude and depression in 15 patients, undue difficulty in waking in 11, early morning headache in seven and nocturnal fits in two. By day the worst affected appeared pale, apathetic, torpid and demoralised.

Others complained of lethargy, depression and difficulty in concentration. Reduction of insulin dosage relieved or abolished all such problems without loss of overall diabetic control. Schwandt et al. (1979) have described similar examples. Nine out of 45 unstable diabetics turned out to have been chronically overtreated with insulin; their symptoms included excessive appetite, polydipsia, vertigo, mood swings, irritability and chronic fatigue.

Descriptive studies have shown that all forms of mental illness may occur in association with diabetes but there is little evidence about prevalence and forms. Apart from disturbances associated with overt or covert hypoglycaemia there is probably little that is specific. Increased fatigue and diminished energy were the most prominent mental symptoms in Surridge et al.’s (1984) survey, also depression and irritability. However, carefully controlled observations do not appear to have been carried out. Surveys of psychiatric hospital populations have sometimes indicated a higher prevalence than expected of diabetes mellitus (Waitzkin 1966a, 1966b; Clayer & Dumbrill

1967), but chiefly of the late-onset non-insulindependent form. This may merely reflect the age and tendency to obesity among long-stay patients; chlorpromazine, moreover, may lead to hyperglycaemic responses on the glucose tolerance test.

Finally, in the management of manic-depressive illness in diabetics it may sometimes be necessary to consider the possibility that insulin requirements increase during markedly depressive phases:

Crammer and Gillies (1981) reported a woman with long-standing manic-depressive disorder which for some years had been cyclical, occurring approximately every 20 weeks. For several years she had also had late-onset diabetes, well controlled with oral hypoglycaemic drugs. During a particularly severe depressive episode the diabetes became out of control, requiring soluble insulin injections. Electroconvulsive therapy relieved the depression, but coincidentally the insulin requirement declined dramatically. With recurrence of the depression the same changes were seen, again resolving on recovery. During her manic phases, by contrast, no changes in diabetic treatment were required.

Kronfol et al. (1981) reported a similar example of a patient who required increased insulin during recurrent phases of a depressive illness.

Brain damage in diabetes

When a diabetic patient develops organic psychiatric disorder the question arises how far this may be attributable to some aspect of the diabetic process. Episodes of hypoglycaemia or diabetic coma may have contributed to brain damage, or associated atherosclerosis may be responsible.


In very young children Ack et al. (1961) produced some evidence that intelligence may be impaired, possibly as a result of damage to the immature brain by episodes of hypoglycaemia or acidosis. Thirty-eight diabetic children were compared with their siblings on the Stanford Binet test, and those with onset below the age of 5 years were found to score an average of 10 points lower than the controls. By contrast those with onset over the age of 5 were unaffected. The result may have reflected some degree of brain damage, or merely the psychological impact of a chronic disease at such a young age.

Ives (1963) surveyed 380 adult diabetics in a general hospital and found that 45 were mentally abnormal. Eighteen showed ’organic brain syndromes’, 14 were mentally deteriorated, nine had personality disorders and four were psychotic. Hypertension was commoner in the mentally impaired patients than in the group as a whole, suggesting that cerebral atherosclerosis was probably the responsible factor. In others the mental abnormalities may have been coincidental, but episodes of diabetic coma and hypoglycaemia had been more frequent in the total group of 45 than in the remainder.

Bale (1973) compared 100 patients from a diabetic clinic with age- and sex-matched controls drawn from visitors to the hospital. All patients were under 65 years of age and had had diabetes for 15 years or more. On the Walton Black New Word Learning Test 17 diabetics, but no controls, scored in the brain-damaged range. A significant relationship was observed between low scores on the test and the apparent severity of past hypoglycaemic episodes. The frequency of cerebrovascular accidents was higher in the diabetic group than the controls, but only one patient with a cerebrovascular accident scored within the brain-damaged range. It thus appeared that mild cognitive difficulty was not uncommon in long-standing diabetics, and that hypoglycaemic episodes rather than cerebrovascular disorder might be the principal factor responsible.

Two more recent studies have addressed the issue. Wredling et al. (1990) selected 17 insulin-dependent diabetics with histories of recurrent severe hypoglycaemic episodes and compared them to 17 diabetics without such histories. The two groups were comparable in terms of sex, age, duration of diabetes, injection frequency, dose of insulin and socioeconomic parameters. On neuropsychological evaluation the group who had had hypoglycaemic episodes performed significantly less well on tests of finger tapping, digit span, Necker cube reversals and maze learning.

Langan et al. (1991) tested 100 insulin-dependent patients on an extensive psychometric battery, after excluding those who had evidence of other causes of

brain damage including cerebrovascular disease. A questionnaire was used to assess the number, frequency and severity of hypoglycaemic episodes experienced. Significant correlations were found between the frequency of severe hypoglycaemias and an index of intellectual decline (discrepancy between NART and WAIS IQ), all of the WAIS performance subtests and a choice reaction time test. Speed of information processing as measured by the PASAT (p. 198) correlated with the estimated number of hypoglycaemic episodes sustained during the patient’s lifetime. No such relationships could be discerned for measures of verbal memory or verbal IQ. Contrasting groups were then identified with and without a history of severe hypoglycaemias: though balanced in terms of premorbid IQ and duration of diabetes the former were significantly slower on choice reaction time tests and showed a greater IQ decrement.

All such findings, while strongly suggestive of a pathogenic role for hypoglycaemia, are not of course conclusive. It is possible that the association between impaired cognitive capacity and hypoglycaemic episodes reflects the poorer management of diabetes by patients who are less well-endowed intellectually. Deary arid Frier (1996) review these and other studies and conclude that the verdict remains at present ’unproven’. Prospective studies over a considerable period of time will be needed to clarify the situation decisively.

A patient reported by Mace (1987) raised the possibility of memory impairment secondary to self-induced hypoglycaemias. A 29-year-old computer manager had abused many drugs and drunk excessively from his teens, and diabetes was diagnosed at 25. He developed severe hypoglycaemic episodes, some with convulsions and automatisms and prolonged bouts of disinhibited behaviour. In hospital it was noted that such episodes were frequent when he was left to measure and administer his own insulin, but stopped when it was given under nursing supervision. He denied deliberate abuse of insulin, but phials and syringes were found concealed among his possessions along with other drugs. His memory was clearly faulty for events of the previous year, and testing confirmed the presence of severe verbal memory deficits. It was difficult to know how far the insulin hypoglycaemias had contributed to this rather than the abuse of alcohol and other drugs.

The role of arterial disease is clear in patients who develop cerebrovascular accidents and will often be incriminated in those who develop dementia. Diabetes is recognised as a key risk factor for atherosclerotic brain infarction, along with hypertension and heart disease (Dyken etal. 1984). Its impact was clearly demonstrated in the Framingham prospective study of middle-aged and elderly persons, in relation to stroke, cardiovascular disease and claudication (Kannel & McGee 1979).

Careful autopsy studies have also confirmed the association

Page 538 CHAPTER 11

with cerebral atherosclerosis. Alex et al. (1962) showed that cerebral infarctions were one and a half times as common in diabetics as controls, with a greatly increased frequency of proliferative lesions in the small cerebral vessels. Grunnet (1963) graded the severity of atheroma in the circle of Willis in 107 cases, and showed an increased frequency and severity in all age groups when compared to controls. Between the ages of 30 and 70 years the incidence was almost doubled. The duration of diabetes and the presence or absence of hypertension were not closely correlated with the severity of such changes, but they seemed to be related to high blood cholesterol levels and frequent acetonuria.

Cerebral blood flow studies have indicated another mechanism whereby the risk of cerebrovascular damage may be increased in diabetics (Dandona etal. 1978). Even when resting levels are normal, the reactivity of the vessels to carbon dioxide inhalation is unusual in a high percentage of patients. The expected increase in cerebral blood flow may fail to occur, or paradoxically there may be a substantial fall. This failure to compensate appropriately in response to increased cerebral metabolic demands may in the long term make its own contribution to brain damage.

Reske-Nielsen and Lundbaek (1963) found both diffuse and focal changes in the brains of long-standing diabetics, partly attributable to vascular disease and hypertension but possibly also deriving in part from the effects of metabolic disturbances on the neurones. Diffuse degeneration was observed in neurones and nerve fibres of the cerebrum, cerebellum and brain stem, and gliosis was often considerable. Such changes may account for Perlmuter et al.’s (1984) finding that older non-insulin-dependent diabetics performed worse than controls in terms of memory and learning, even after excluding those with overt dementia or cerebrovascular disease.

Diabetic coma

The development of ketotic coma is a serious medical emergency, still responsible for some 10% of deaths due to diabetes each year (VVindebank & McEvoy 1989). Very occasionally patients may present in this way without being known diabetics, but usually the ketosis develops in a patient with the established disorder which has got out of control. Insulin may have been omitted, or there may have been precipitants by way of infection, physical trauma, gastrointestinal disturbance or alcohol excess. The role that has been claimed for emotional precipitants has already been discussed.

Prodromal symptoms consist of weakness, thirst, dull headache, abdominal pain, nausea, vomiting and drowsiness. The onset may be abrupt or insidious. It is sometimes very gradual over several hours, so that a patient with a dangerous level of ketosis may still be fully ambulant. Air hunger and heavy laboured breathing develop. The patient becomes increasingly listless and in about 10% of cases sinks into coma, sometimes after a period of restlessness, irritability and confusion.

The pulse is rapid and feeble and the blood pressure low. Dehydration is marked, the face flushed, and acetone may be smelled on the breath. Investigations reveal large amounts of sugar and acetone in the urine, and elevated sugar and ketone bodies in the blood. Acidosis is marked and the blood urea is raised. Cerebral oedema and disseminated intravascular coagulation are grave developments. Management is discussed by Hammond and Wallis (1992).

On initial examination the picture may be hard to distinguish from hypoglycaemic coma or from advanced renal failure. Overdosage with salicylates can also give a closely similar picture. In older subjects the differential diagnosis must sometimes include a cerebrovascular accident, since glycosuria may also occur in such a situation.

Hyperosmolar (non-ketotic) coma

Neurological complications

Peripheral neuropathy can be a severe and distressing complication. In the middle aged and elderly sensory changes usually predominate, with paraesthesiae, pain and cramps in the calves, absent knee and ankle reflexes and diminished vibration sense. Loss of postural sense may lead to ataxia. In younger patients the affection may be more severe, with both motor and sensory disturbances involving all four limbs. Both forms tend to improve when the diabetes is properly controlled. Microvascular changes in the vasa nervorum may be

is principally encountered in older type II diabetics. There is no associated ketoacidosis, but extremely high levels of serum osmolality and glucose are seen, sometimes developing over several days or weeks. It often arises as a complication of infection or other medical problems and mortality is again high. The presenting picture may be of gradually increasing lethargy and impairment of consciousness, but unlike ketotic coma seizures and focal neurological signs are common. The patient may at first be thought to be suffering from an acute stroke, presenting with hemiparesis, aphasia or simple or complex hallucinations (Guisado & Arieff 1975). The condition is usually reversible with correction of the metabolic abnormalities. Autonomic changes can include hyperpnoea and hypertension.Page 530 ENDOCRINE DISEASES AND METABOLIC DISORDERS

responsible, or some as yet incompletely understood metabolic complication.

Isolated 3rd or 6th cranial nerve palsies are not uncommon, and are probably attributable to focal infarctions in the cranial nerve trunks. These usually clear spontaneously after a few weeks or months. In elderly diabetics pupillary changes may include miosis, irregularity and a sluggish reaction to light, sometimes amounting to the classic Argyll Robertson pupil (pseudotabes). Lesions in the midbrain are presumably the cause.

Other complications include a variety of forms of radiculopathy, amyotrophy, myelopathy and autonomic disturbances.

Insulinomas and other forms of hypoglycaemia

Though rare, insulin-secreting tumours of the pancreas are an important cause of hypoglycaemia, especially to the psychiatrist. The disorders which they produce are extremely varied and usually intermittent with normal health between attacks. Problems of differential diagnosis are therefore considerable and will be dealt with in some detail. The discovery of the condition is often long delayed, resulting hi prolonged ill health and very occasionally in irreversible brain damage.

Pathology and pathophysiology of insulinomas

Insulin-secreting tumours gained recognition in the late

1920s. By far the most common are benign adenomas of the islet cells. Those found during life are usually 0.5-5 cm in diameter, whereas those seen at autopsy are often microscopic in size. A tumour too small to be palpated at operation may therefore be responsible for symptoms. Two-thirds are found in the body and tail of the pancreas rather than the head, and multiple tumours are common. Rarely ectopic insulinomas occur in the vicinity of the duodenum or porta hepatis.

The great majority are benign, but approximately 10% are malignant and can metastasise to other parts of the body. Microadenomatosis throughout the pancreatic tissue is another rare possibility, or diffuse insular hyperplasia without tumour formation. These latter conditions sometimes occur familially along with adenomas of the parathyroid, pituitary or other endocrine glands as part of the ’pluriglandular syndrome’ (MEA type 1) (p. 528).

The type of tumour which produces insulin is cornposed of beta cells. Other forms of islet-cell tumour may be non-functioning or ’ulcerogenic’, the latter being associated with gastric hypersecretion and peptic ulceration (Zollinger-Ellison syndrome).

The primary defect in insulin-secreting tumours appears to be inability to control the storage and release of insulin, rather than simply excessive production. The result is a constant slow excessive release coupled, for unknown reasons, with sudden excessive discharges. In consequence the plasma insulin level shows large and abrupt fluctuations.

Syndromes of hypoglycaemia

Marks (1981a), from whom much of the present account is taken, demarcates four categories of disturbance all of which may be seen with insulinomas. He prefers the term ’neuroglycopaenia’ when referring to clinical syndromes, and reserves the term ’hypoglycaemia’ to describe the level of sugar in the blood. Correlations between the two are not exact, and occasionally profound lowering of blood sugar may be found without apparent effect on brain function.

Acute neuroglycopaenia

This is the common syndrome that follows overdosage with insulin or oral hypoglycaemic agents, though similar attacks may occur with insulinomas. The patient experiences vague malaise with anxiety and panic, or an unnatural detached feeling akin to depersonalisation. This is accompanied by feelings of hunger, palpitations and restlessness, and is shown objectively by tachycardia, tremor, flushing, sweating and ataxic gait. Angina may be precipitated if coronary artery disease exists. Brief episodes of unconsciousness may occur and epileptic attacks can be provoked. Occasionally focal neurological disturbances such as diplopia, hemiparesis or dysphasia may be seen without obvious diffuse cerebral disturbance. In severe examples progression occurs to coma.

Subacute neuroglycopaenia

This also occurs episodically and is more characteristic of insulinomas. Subjective symptoms are slight, and all of the above features may be only minimally developed. Instead there is clumsy performance at habitual tasks, and behaviour that is out of character for the person affected. He may behave in a disinhibited, foolhardy or aggressive manner which closely resembles alcoholic intoxication, y Others become apathetic and withdrawn, with slurred speech and somnolence. Disorientation and mental confusion are usually in evidence but consciousness is retained until late.

The degree of functional impairment is out of proportion to subjective discomfort, and the person typically

Page 540 CHAPTER 11

lacks realisation of the changes within himself. Moreover negativism is a common feature, so that he may fail to seek help or take appropriate action even if partially aware of the disturbance. Often it is only after recovery that the person realises he has been unwell at all.

Chronic neuroglycopaenia

This, though rare, is virtually confined to patients with insulinomas. There are no dramatic symptoms or signs but a change of personality tends to develop insidiously. Defective memory and intellectual deterioration ultimately lead to severe dementia. Emotional changes may be prominent, with irritability, apathy or emotional lability, and psychotic features with paranoid delusions may develop. The course may be punctuated with episodes of acute or subacute neuroglycopaenia, or may be uniformly smooth so that the possibility of hypoglycaemia goes unsuspected. The symptoms and signs are unaltered by food, and little further deterioration is observed with fasting unless this is very prolonged.

Hyperinsulin neuronopathy

Peripheral neuropathy may develop, with paraesthesiae in the hands and feet and wasting of muscles. The wasting is usually in the distal musculature but can sometimes be proximal. The cause of the hypoglycaemia is almost always an insulinoma. It may occasionally be the presenting feature of the illness, but is more commonly overshadowed by preceding coma or other acute disturbances.

Clinical manifestations of insulinomas

Insulinomas occur equally in males and females. They present usually between the ages of 20 and 50, though cases are reported at all ages including childhood. Familial examples are described, perhaps as part of the pluriglandular syndrome. Episodes of odd behaviour and disturbances of consciousness are the main reasons for referral. Symptoms have commonly been present for months or years by the time the diagnosis is made, sometimes for as long as 30 years. Very occasionally long remissions may be detected in the histories.

Almost any psychiatric syndrome may be simulated, depending partly no doubt on the make-up of the individual. The detailed content of attacks may differ from one occasion to another and diagnostic confusion is common. The essential clue usually lies in the episodic and recurrent nature of the attacks. An added difficulty is that organic features are not always in evidence, the change of consciousness sometimes being so slight that it passes unnoticed except to those familiar with the patient.

Typically attacks of subacute, or less often acute, neuroglycopaenia occur as described above. These gradually increase in frequency, initially occurring at intervals of several weeks or months but often occurring several times per week by the time of presentation. Attacks may cornmence abruptly, or with a slow build-up of weakness, ataxia and increasing confusion. During witnessed attacks one may see sweating, nystagmus, incoordination or focal neurological signs such as hemiparesis or positive Babinski responses. Actual coma is rare except in the most severe episodes.

At first attacks are commonest at noon or late afternoon, and only later do they occur more characteristically before breakfast and during the night. Only a quarter of patients give a clear history of a relationship to fasting and only 10% to exercise, while relief by eating is even more rarely noticed by the patient (Marks 1981b).

Typically the patient has complete amnesia for the content of the attacks, and occasionally for additional , periods during which behaviour was seemingly normal. Between attacks he usually feels quite well.

In a small proportion there is progressive mental or physical disability with little or no history of episodic disturbance. This has already been described under the heading of chronic neuroglycopaenia.

The diagnostic problems are well illustrated by cases in the literature:

A man of 44 suffered from attacks of confusion with bizarre behaviour over a 5-year period, each lasting from a few minutes to several hours, and increasing in frequency until they were occurring four or five times a week. He had a partial or complete amnesia for most of what happened during attacks. Initially they were interpreted as hysterical fugue states.

On one occasion, when motoring home from a funeral, he abruptly began to drive at breakneck speed and seemed quite unaware of his passenger. On another he wandered in a semiconfused condition about the corridors at work for 2 hours, carrying on coherent conversations with certain people he met but ultimately removing his shin and staring foolishly at people who tried to talk to him. Once when chopping wood he suddenly clutched his axe in a menacing manner and wandered about the neighbourhood with a dazed and glassy-eyed expression; after this he was taken to hospital where his speech was confused and mumbling and he showed constant grotesque and purposeless movements. He then abruptly recovered with amnesia for the entire episode.

In the attack which led to the correct diagnosis he had frightened a fellow employee by brandishing a knife. This informant was able to give a dear description of many episodes, describing marked pallor, sweating, limpness, unsteady drunken behaviour and double vision, and emphasising their liability to occur after excessive exertion or towards the end of the morning.


Before removal of the insulinoma, detailed studies were made of the interrelationships between the patient’s behaviour and blood sugar levels. During prolonged starvation experiments it was possible to demonstrate intermittent disturbances of awareness which became progressively more marked even though the blood sugar values were essentially unchanged. Periods of acute motor excitement with confusion and aggressive behaviour could similarly begin and end without significant changes in the blood sugar. (Romano & Coon 1942)

A 27-year-old man was referred as a psychiatric emergency on account of bouts of aggressive and destructive behaviour. He showed disorientation and inappropriate behaviour during attacks, with sweating and violent tremor, and had no subsequent recollection of what occurred. He had always been backward and had had a head injury at 18. His father had shown similar bouts of violent behaviour.

The initial differential diagnosis had included aggressive psychopathy, mental subnormality with behaviour disturbance, and post-traumatic epilepsy. Prolonged fasting provoked a typical attack at 16 hours, associated with hypoglycaemia and relieved by glucose. At laparotomy multiple islet-cell tumours were found. A diagnosis of multiple endocrine adenomatosis (pluriglandular syndrome) was made since hyperparathyroidism was also present. His father had died with ’islet-cell-secreting tumour of the pancreas and calcification of the kidneys’.

(Carney etal. 1971)

A physicist of 32 had attacks consisting of inappropriate aggressive behaviour, sometimes involving attacks on people or destruction of his own belongings. Attacks might last for 30 minutes to 1 hour and were succeeded by dense amnesia for what had occurred. If they did not end in an epileptic fit or simple unconsciousness, they would lead on to a period of emotional disturbance and confusion for which memory would subsequently be incomplete and fragmentary. The EEG showed some inconstant theta activity arising mainly in the temporal leads. A 12-hour fasting blood sugar was normal at

84 mg/100 ml.

He was regarded as suffering from temporal lobe epilepsy but anticonvulsant treatment was ineffective. Subsequent and more thorough blood sugar studies indicated an islet-cell tumour and this was confirmed at operation.

(Whitty, Stores & Lishman 1977)

Investigation of insulinomas

The finding of a low blood glucose is essential to the diagnosis. The exact level at which symptoms of cerebral dysfunction may be expected is somewhat variable, but they usually do not appear until the level is below 2.2 mmol/1 (Marks 1981c). The range is wide, however, depending on the rate of fall and other concurrent metabolic factors. If an actual attack is witnessed blood should be taken immediately for glucose estimation, also where possible for the measurement of plasma insulin and C-peptide levels (see below). The ability of intravenous glucose to bring prompt relief should then be assessed. If spontaneous attacks do not occur while the patient is under observation, an attempt may be made to provoke an attack by fasting, coupled if necessary with exercise. Routine overnight fasts may need to be repeated on a number of occasions, but will fail to reveal significant changes in perhaps 10% of cases; the fast may then need to be prolonged for up to 72 hours. Close supervision is of course necessary, and with any attack that occurs the efficacy of intravenous glucose to terminate it should be tried immediately blood has been withdrawn. The hypoglycaemia seen with insulinomas is characterised by inappropriately high levels of plasma insulin. At a fasting glucose level below 2.5 mmol/1, a plasma insulin concentration greater than 5 mU/1 is diagnostic of insulinoma (Turner 1996). Cpeptide levels are also inappropriately raised; C-peptide is secreted molecule for molecule along with insulin by the pancreas, providing a very useful measure of pancreatic beta cell function.

In equivocal cases a ’suppression test’ may circumvent the need for prolonged fasting and brings important evidence of its own. Patients with insulinomas show increased sensitivity to a slow infusion of insulin, and more importantly fail to suppress the production of Cpeptide in the normal manner.

The availability of refined plasma insulin measurements and C-peptide assays has rendered most other tests superfluous, for example tolbutamide or glucagon stimulation tests. These also carry a certain hazard. It should only rarely be necessary nowadays to do an exploratory laparotomy to search for an undisclosed tumour.


The earliest change to be detected is the appearance of delta rhythms on overbreathing. Then spontaneous theta and delta waves occur, and later asynchronous irregular waves and a flattened tracing. The abnormalities are often most marked over the temporal lobes. In hypoglycaemic coma the EEG consists of high voltage delta waves as in other comas of metabolic origin. After severe hypoglycaemic episodes, abnormalities may persist in the EEG for several days and then revert to normal.

These sequential changes were studied in the course of therapeutic insulin comas, formerly used for the treatment of schizophrenia. The electroencephalographic response to hypoglycaemia appears to vary considerably y in different subjects, though the same individual may show closely similar responses in each episode of a series. In general, the correlation between the symptoms provoked and the electroencephalographic changes is closer than that between symptoms and blood sugar levels.

The value of the EEG in helping towards a diagnosis of

Page 542 CHAPTER 11

insulinoma is controversial. Rose (1981) considers it useful to take serial records during fasting, and if symptoms occur with a low blood sugar and an abnormal tracing intravenous glucose can then be shown to restore the record to normal within minutes.

After removal of the insulinoma the EEG soon returns to normal, including the response to overbreathing.

Differential diagnosis of insulinomas

The following psychiatric and neurological conditions may often be suspected before the true diagnosis is revealed:

Neurotic disorder

Personality disorder

Manic-depressive or schizophrenic psychoses

Presenile or senile dementia

A space-occupying lesion

Carotid artery stenosis


Intoxication with alcohol

Other metabolic disorders

Peripheral neuropathy has already been mentioned as a rare form of presentation. Narcolepsy may sometimes be considered, or attacks of somnolence may suggest the Kleine-Levin syndrome

Among disorders of the cardiovascular system, angina, vasovagal attacks, orthostatic hypotension and StokesAdams attacks may all come under suspicion.

The differential diagnosis from other forms of hypoglycaemia is considered below. Of particular importance in psychiatric practice are essential reactive, alcohol-induced and factitious hypoglycaemia.

An important step when hypoglycaemia is confirmed is to differentiate between the various conditions with which it may be associated. The old classification into organic and functional hypoglycaemia has outgrown its usefulness, since the former rests entirely on finding some identifiable cause. When an insulinoma was missed at operation the case would have been erroneously labelled as functional; and the cause remains elusive in some of the organic forms which are important in general medicine.

The most useful classification in practice is into hypoglycaemias which are provoked by fasting and those which are not. The latter are labelled the ’stimulative hypoglycaemias’ since they develop only in response to some identifiable stimulus. The distinction is helpful since attempted provocation by fasting always forms the first step in investigation.

such as uraemia or liver failure may be closely simulated, or endocrine disorders such as thyrotoxicosis, hypoparathyroidism or often suspected when attacks are witnessed by on-lookers and the picture may be very similar indeed. When the patient is known to be alcoholic, episodes of amnesia may be ascribed to ’alcoholic blackouts’. Intoxication with other drugs may similarly be a relatively rare form of presentation of hypoglycaemia except in childhood. But any episodic and recurrent neurological disorder is likely to raise the question of epilepsy, and the content of attacks may closely resemble the automatisms of complex partial seizures.may be suggested by attacks of dysphasia or hemiplegia, or vertebrobasilar insufficiency when vertigo and diplopia are prominent features.may be suspected when attacks provoke epilepsy, or when headache and focal neurological symptoms accompany clouding of consciousness. Subdural haematoma may be considered in view of the negative or equivocal neurological findings between attacks.may be closely simulated in cases with chronic deterioration of intellect and personality. Superimposed episodes of acute and transient disturbance may resemble the step-like course of multi-infarct dementia. These are serious diagnostic mistakes since the brain damage that occurs may ultimately become irreversible.may very occasionally be diagnosed. Rare cases of insulinoma have been reported to present with acute depressive psychoses, no doubt as a result of individual predisposition. In other cases the bizarre nature of the behaviour during attacks may suggest an ingravescent schizophrenia, and in chronic neuroglycopaenia thought disorder and paranoid delusions may be prominent.may be suggested by a history of episodes of aggressive or antisocial conduct, for which the patient claims amnesia or only a hazy recollection. Aggressive psychopathy may be diagnosed in more extreme suggested by the occurrence of transient symptoms which are described vaguely and diffusely, often variable from time to time, and unbacked by physical signs when the patient is examined between attacks. Episodes of anxiety, panic and depersonalisation may suggest anxiety neurosis; transient disturbances of consciousness with periods of amnesia may suggest hysteria. The latter may also be diagnosed when transient neurological features such as dysphasia or hemiparesis follow attacks. The differentiation between neurosis and essential reactive hypoglycaemia is discussed on p. 543.Other forms of hypoglycaemia

Fasting hypoglycaemias

The fasting hypoglycaemias include those due to insulinomas and those associated with liver disease and endocrine disorder. Fasting hypoglycaemia may occur with all varieties of liver disease, both trivial and serious, and shows


little correlation with the severity of the hepatic disorder. Among endocrine disorders Addison’s disease and hypopituitarism are the commonest causes.

Other causes, unlikely to come before the psychiatrist, include glycogen storage disease, neonatal hypoglycaemia and the so-called idiopathic hypoglycaemia of childhood. The differentiation of these conditions from insulinoma requires careful medical investigation; the absence of raised plasma insulin levels during fasting and the presence of normal C-peptide suppression after insulin administration ultimately provide the distinction.

Stimulative hypoglycaemias

Essential reactive hypoglycaemia

This is the commonest form of stimulative hypoglycaemia and must be carefully distinguished from insulinoma. The most important clinical difference is that symptoms resembling acute neuroglycopaenia occur after ingesting food but are not provoked by fasting. They do not occur before breakfast, whereas this is common with insulinoma. In effect the disorder represents an exaggeration of the normal physiological response to the ingestion of carbohydrate.

Females are affected more frequently than males, and most cases occur between the ages of 30 and 40 years. It appears to occur particularly in asthenic and emotionally labile persons, and may be associated with minor psychiatric instability in a manner which initiates a vicious circle.

Common symptoms are episodes of weakness, faintness, palpitation and irritability, often in association with feelings of hunger, nausea, headache and vertigo. The patient may complain of ’blackouts’, but objective evidence of impairment of consciousness is rare. An interesting medicolegal example, in which a patient committed a serious offence but had no knowledge of it afterwards, is described by Bovill (1973). The symptoms are commonest in mid morning, usually 2-5 hours following food, though the patient rarely comments spontaneously on such an association. Exercise may provoke or aggravate the symptoms, but food or glucose do not bring decisive relief. Between attacks the patient often reports that he feels run down and is functioning below his optimum. The episodes may occur intermittently over several years, but do not show the progression in frequency and severity characteristic of insulinomas.

Fasting blood glucose levels are normal, but a 6-hour glucose tolerance test will show an excessive fall 1.5-4 hours after glucose ingestion. The diagnosis should be made only when both hypoglycaemia and symptoms of neuroglycopaenia are reproduced during the procedure.

A pronounced rise of plasma cortisol in response to the rebound hypoglycaemia will add confidence to the diagnosis.

It is often uncertain whether the degree of reactive hypoglycaemia demonstrated in such patients should be considered abnormal. Many authorities suggest that only a minority have a true defect of glucose homeostasis as a cause of symptoms. A number of examples probably represent neurotic disorder, and since reactive hypoglycaemia in some degree is just as likely to occur in neurotic as non-neurotic individuals one must be cautious in attributing the symptoms to the blood sugar changes which are demonstrated.

Ford et al. (1976), for example, found that among 30 patients referred as possible examples of the syndrome, only 18 showed reactive hypoglycaemia on testing. Moreover, half of the sample were markedly psychiatrically unwell, mostly with depression or anxiety neurosis. Many showed hysterical or obsessional personality patterns. While the patients attributed their emotional distress to hypoglycaemia the overall findings ran counter to this. Thus the number and severity of symptoms experiencedemotional lability, depression, headache, tremor, tachycardia, weakness, dizziness-were unrelated to the degree of hypoglycaemia shown on glucose tolerance tests. Moreover, whether or not they had demonstrated reactive hypoglycaemia their scores for emotional disturbance on the MMPI were similar. Careful histories often showed a close association between the onset of the complaints and some precipitating stress.

Ford et al. concluded that while essential reactive hypoglycaemia remained a definite clinical entity, causing genuine somatic discomfort, it was nevertheless considerably overdiagnosed in the face of multiple non-specific complaints. Patients with unstable or obsessional personalities may come forward with symptoms that would be tolerated with less anxiety by other persons; when a glucose tolerance test is then performed the symptoms’may all too easily come to be attributed to any hypoglycaemia revealed, incidental though this may be.

Rennie and Howard (1942), moreover, described a group of patients suffering from tension states and depression in whom the reactive hypoglycaemia appeared, in fact, to be secondary to the psychiatric disorder. After appropriate psychiatric treatment the patients were found to have lost their presenting symptoms, but in addition repetition of glucose tolerance tests often showed that these too had returned to normal.

Clearly therefore the interrelationships between psychiatric disorder, physical symptoms and low blood sugar levels are complex. It is nevertheless worth trying the effect of a high-protein low-carbohydrate diet in the more marked examples of the disorder, and especially when there is no overt neurosis towards which treatment can be directed. ’

Alcohol-induced hypoglycaemia

Hypoglycaemia may occur in response to various drugs

Page 544 CHAPTER 11

and poisons of which alcohol is the most important. Inhibition of gluconeogenesis by alcohol appears to be the predominant factor, though impaired response of the pituitary-adrenal axis, and enhanced insulin secretion following an oral glucose load, can also play a part.

The true importance of alcohol-induced hypoglycaemia has been recognised relatively recently. It occurs mostly in chronic alcoholics, usually 6-36 hours after a large intake, and quite often presents with hypoglycaemic coma. It may sometimes develop sooner, then running a considerable risk of being overlooked in the face of obvious intoxication. Recovery is usually prompt after the administration of glucose though hydrocortisone may be required as “well. It is possible that the condition accounts for a considerable number of deaths in alcoholic subjects.

Recurrent attacks are rare but have been reported (Fredericks & Lazor 1963; de Moura et al. 1967). It is uncertain whether some special sensitivity to alcohol has been acquired in such cases, since the alcoholic history is often of long duration before the complication makes its appearance.

An especially severe form of alcohol-induced hypoglycaemia is liable to occur in insulin-dependent diabetics. Arky et al. (1968) first drew attention to this, reporting patients who were repeatedly admitted to hospital in hypoglycaemic states after an alcoholic debauch. Their five patients illustrated the severity of the neurological damage that could result from hypoglycaemia of such combined origin-two died without recovery from coma and three were left with permanent memory impairments or dementia.

Factitious hypoglycaemia

The clandestine use of insulin or sulphonylureas to induce hypoglycaemic symptoms can pose a difficult diagnostic problem. It is seen predominantly among nurses and other paramedical personnel, or in close relatives of diabetics who have access to the agents. It has also been noted among teenage female diabetics who misuse their insulin (Lancet 1978). Some patients addicted to heroin have described obtaining insulin from the same sources of supply.

The hypoglycaemias may be induced for the benefit of the symptoms experienced or as pan of attention-seeking behaviour. Sometimes the practice appears to be allied to Munchausen’s syndrome, the patients submitting to laparotomy and even subtotal pancreatectomy.

The presentation is with symptoms suggestive of insulinoma. Many patients are admitted in coma. Scarlett et al. (1977) found that symptoms had been present for 2 months to 6 years among their seven insulin abusers, two of whom had had subtotal pancreatectomies before the true cause was discovered. One had been diagnosed as diabetic for 17 years and this had probably always been erroneous. Another described the exquisite pleasure of going to sleep after injecting insulin, not knowing whether she would regain consciousness. A sinister facet of Scarlett et al.’s series was that two patients had also given insulin to their children, one with a fatal result.

Jordan et al. (1977) described patients abusing sulphonylurea and presenting in a similar manner. One of their patients began taking her husband’s chlorpropamide while severely depressed, another was an alcoholic whose hypoglycaemia was at first ascribed to alcohol. Jordan et al. estimated that the problem had become as common as genuine insulinoma, requiring therefore the utmost care to screen for factitious hypoglycaemia before such a diagnosis was made.

The risk of error is increased by the finding of inappropriately raised plasma insulin during the hypoglycaemic episodes, both with insulin and sulphonylurea abuse. However, C-peptide levels (p. 541) will be low when exogenous insulin has been taken, and this will immediately raise suspicion of the practice. After oral hypoglycaemic agents the picture can be virtually indistinguishable from insulinoma since C-peptide levels are likely to be raised as a result of beta-cell stimulation; the diagnosis must then be made by demonstrating the drugs in the blood or urine.

Under the heading of pseudoneuroglycopaenia Marks (198la) describes a quite separate group of patients with disturbed personalities, also paramedical workers or relatives of diabetics, who present with attacks simulating acute neuroglycopaenia, sometimes gradually proceeding to coma. Here the blood glucose and EEG remain normal throughout the episodes. Recovery occurs spontaneously or may be provoked by saline injection. The situation is considered to represent an hysterical conversion reaction.

Other stimulative hypoglycaemias

After gastrectomy the ’dumping syndrome’ occurs during or immediately after a meal; other similar forms of disturbance may be some hours delayed. The symptoms resemble those of essential reactive hypoglycaemia. Glucose tolerance tests again often produce conflicting results, and the true relationship of symptoms to reactive lowering of blood sugar remains uncertain.

A number of drugs other than alcohol may provoke hypoglycaemia. The effect of salicylates can be profound, especially in children, leading occasionally to fatal results. They potentiate the effect of oral hypoglycaemic agents. Propanolol and other non-selective beta blockers can


have a similar effect. Among poisons those present in certain toadstools are markedly hypoglycaemic.

Finally under the heading of stimulative hypoglycaemia, the reactions liable to occur in insulin-treated diabetics must be remembered. Some diabetics are particularly prone to insulin-induced hypoglycaemias and others are incautious of diet.

In the stimulative hypoglycaemias fasting fails to produce abnormal blood sugar levels and the other confirmatory tests for insulinoma will be negative, the sole exception being sulphonylurea abuse as described above.

Treatment of hypoglycaemia

Sugar or glucose tablets should be carried by patients awaiting investigation when hypoglycaemic attacks are known to occur, and relatives should be instructed about the urgent need to give glucose in the earliest stages of an attack. In severe reactions, or if the patient is comatose, glucose should be given intravenously, 10-20g as a

33.3% or 50% solution.

Treatment of hypoglycaemic coma is always a grave emergency. Whenever its presence is considered a possibility intravenous glucose should be given immediately after withdrawing blood for laboratory investigations. The response is usually rapid, though it may sometimes be slow if the coma has been of long duration or the patient is markedly hypothermic. The possibility of coincident hypoadrenalism or hypopituitarism must then be borne in mind, and parenteral hydrocortisone should be tried. The response may be dramatic if irreversible brain damage has not occurred.

The essential treatment for insulinoma is surgical removal at the earliest opportunity after the diagnosis is established. Sometimes more than one adenoma is present and re-operation may occasionally be necessary. When the diagnosis of hyperinsulinism has been fully confirmed by tests, yet no tumour is palpable, subtotal pancreatectomy is indicated. If operation is refused, or if the tumour proves to be malignant, palliative treatment may help for a while in the form of large doses of steroids or diazoxide, together with constant food intake at 3hourly intervals during the day.

Patients with essential reactive hypoglycaemia should not be given repeated doses of glucose, but must be encouraged to take frequent meals which are high in protein and fat.

Outcome of hypoglycaemia

The immediate outcome for most attacks is excellent. The symptoms usually resolve spontaneously even without specific treatment. Sometimes, however, neurological and psychiatric manifestations may outlast the actual period of hypoglycaemia. Dysphasia, hemiparesis and stupor have been reported to persist for hours or even weeks after restoring the blood sugar to normal, also negativism, restlessness, apathy and prolonged behaviour disorder (Markowitz et al. 1961). Such disturbances are thought to rest on a vascular basis which ultimately resolves. Recovery from hypoglycaemic coma may similarly be delayed for hours or even weeks in occasional cases. After prolonged coma recovery may be incomplete, with evidence of permanent brain damage and dementia (Arky et al. 1968).

The outcome after surgery for insulinoma is excellent in the great majority of cases. The typical patient who has presented with intermittent episodes of neuroglycopaenia ceases to have such attacks, whatever form they have taken, and normal health is restored.

While this is the experience of most observers, occasional reports have indicated a less favourable picture on prolonged follow-up. Markowitz et al. (1961) followed six early cases after a lapse of 2 5 years, and found that five had shown either persisting or newly acquired mental disturbances in the interim. These were sometimes severe, including manic-depressive psychosis, irrational or erratic behaviour and other aberrations of personality, and were thought possibly to represent the aftermath of brain damage sustained during the hypoglycaemic episodes. A surprisingly high incidence of peptic ulceration and haemorrhage had also developed, even though ulcerogenic pancreatic tumours had not appeared to be present.

The rarer presentations In the form of chronic psychosis or intellectual deterioration may be expected to improve to some degree after operation, but here recovery is seldom complete. In these cases, as after prolonged episodes of hypoglycaemic coma, irreversible brain damage often proves to have occurred. Thus among 100 patients with psychosis or ’long-term insanity’ in the detailed review by Laurent et al. (1971), 37 showed little or no improvement after removal of the hypoglycaemic tumour. The longer the duration of symptoms before the diagnosis, the greater was the risk of permanent mental disability.

Cerebral pathology of hypoglycaemia

The pathological changes in the brain after acute episodes of hypoglycaemia have mostly been studied in cases of insulin shock. The brain may show oedema and vascular congestion. Survival after profound hypoglycaemic coma may be associated with ventricular dilatation, cortical atrophy and shrinkage of the hippocampi.

Page 546 CHAPTER 11

Microscopically the neurones show ischaemic cell changes and associated gliosis. The cell damage may occur in scattered foci, but is more typically laminar with emphasis on the third and fifth layers (Brierley 1981). Relative sparing of the visual cortex is usual, but the corpus striatum and hippocampus appear to be especially vulnerable, much as in anoxia (p. 548). However, with hypoglycaemia the Purkinje cells of the cerebellum are said to be less prone to be affected (Richardson et al 1959).

The pathology that may result from more minor degrees of intermittent hypoglycaemia has not been documented, nor the changes that result from longcontinued chronic neuroglycopaenia. The pathological basis for the rare examples of persistent neuropsychiatric disability short of gross dementia is therefore incompletely understood.

Cerebral anoxia

Barcroft (1920) distinguished four main varieties of anoxia – anoxic, anaemic, stagnant and metabolic which serve as a framework for reviewing the clinical conditions in which cerebral anoxia can occur. These are shown in Table 18 along with a fifth category of ’overutilisation anoxia’.

Anoxic anaemia

Anoxic anoxia is due to deficient oxygenation of the arterial blood. It may result from respiratory insufficiency in pulmonary diseases such as chronic bronchitis, emphysema or pneumonia, from the administration of general anaesthetics, from asphyxia or drowning, or from lack of oxygen in the inspired air at high altitudes.

Table 18 Causes of cerebral anoxia.

Anoxic anoxia

Chronic bronchitis, emphysema, pneumonia, pulmonary embolism, general anaesthesia, asphyxia, drowning, high altitudes

Anaemic anoxia

Carbon monoxide poisoning, gastrointestinal bleeding, severe anaemia

Stagnant anoxia

Cerebral arteriosclerosis, peripheral circulatory failure (shock), cardiac surgery, congestive cardiac failure, cardiac arrest, paroxysmal dysrhythmias, myocardial infarction

Metabolic (toxic) anoxia

Hypoglycaemia, cyanide poisoning, carbon disulphide poisoning

Overutilisation anoxia

Epileptic seizures

(end of table).


A respiratory cause, and particularly a respiratory infection, must be constantly borne in mind when unexplained mental confusion appears in the elderly, or when a fluctuating acute organic reaction is superimposed upon a known dementing process. The anoxia which results from respiratory disease is often further complicated by the accumulation of carbon dioxide (hypercapnia) as described on p. 562.

General anaesthesia inevitably carries a hazard of anoxia. In particular, minor degrees which would be tolerated by a fit person may prove disastrous when the blood volume is low (Dinnick 1964). Thus anaesthesia contributes to some of the acute delirious episodes which appear postoperatively, especially those which emerge immediately upon recovery. A mild degree of persistent dementia may very occasionally make its first appearance after the administration of an anaesthetic for some relatively minor surgical procedure. This is particularly likely in the elderly or arteriosclerotic in whom the reserves of cerebral function have already been precarious.

Bedford (1955) reviewed a population of over 1000 geriatric subjects who had had operations in the preceding 15 years, and found that in more than a third the relatives claimed that the patients had ’never been the same again’. The statement proved to be unjustified in the majority, but in approximately one case in 10 there was reason to think that the anaesthetic may have contributed to memory difficulties or intellectual impairment. In 18 cases seen personally, Bedford found a severe dementia which appeared to have followed anaesthesia directly. Other more minor forms of intellectual disability were found, as in the case of a surgeon who had been obliged to cease work on account of memory difficulties.

In contrast to Bedford’s findings, Simpson et al. (1961) were able to identify only one out of 681 elderly patients who suffered permanent intellectual impairment following anaesthesia. This was a careful prospective investigation involving assessments both before and after operation, and comparisons between those who had had local and general anaesthetics. Simpson et al. concluded that when patients deteriorate after operation the great majority do so by virtue of reasons unrelated to brain damage or anaesthetic agents. Progression of the original disease or the development of new disease can usually be identified as the responsible factor.

Anaemic anoxia

Anaemic anoxia results from deficient oxygen-carrying power of the blood as in carbon monoxide poisoning.


Abrupt blood loss may be the cause as in haematemesis or unsuspected gastrointestinal haemorrhage. Severe anaemia is an important cause which runs an especial risk of being overlooked:

A man of 70 was admitted to hospital with a 12-month history of jerking movements of the head and arms, and loss of weight and appetite for 2 years. He was noisy and restless by day and night, and emotionally labile with frequent laughing and crying. At night he was agitated and confused and moaned continually. There were incessant coarse movements of the limbs and head but no definite parkinsonian features. The EEG was markedly abnormal with a diffuse excess of theta and paroxysmal delta activity. He was thought to be suffering from presenile dementia secondary to cerebral arteriosclerosis.

His haemoglobin was found to be only 6.1 g/100 ml, probably as a result of chronically bleeding haemorrhoids. Three pints of blood were transfused and he improved remarkably, with resolution of the involuntary movements and of the mental abnormalities.

Four months later the appearances of dementia and the motor abnormalities returned in the course of a transient depressive illness, but on this occasion responded well to antidepressant medication. The EEG now showed only mild abnormalities in the form of generalised theta activity on overbreathing. He remained well when followed up during the next 2 years.

Carbon monoxide produces much of its effect by inactivating the oxygen-carrying power of the haemoglobin molecule. Since carboxyhaemoglobin is relatively stable the effects of exposure to carbon monoxide may be longer-lasting and more profound than a transient reduction in oxygen supply per se. But carbon monoxide also acts directly on tissue cells by inhibiting the oxydoreduction enzyme system, contributing an element of metabolic anoxia as well.

Stagnant anoxia

Stagnant anoxia is due to arrest or reduction of blood flow. It occurs in cerebral arteriosclerosis, or more acutely in peripheral circulatory failure due to shock. Important cardiac causes are heart surgery, congestive cardiac failure, temporary cardiac arrest, episodes of severe dysrhythmia as in paroxysmal ventricular tachycardia, and myocardial infarction. A silent coronary infarction can be particularly misleading. Abrupt loss of consciousness at the onset may be misdiagnosed as a cerebrovascular accident, especially if epileptiform jerking has occurred. An episode of severe collapse followed by confusion, but leaving behind no residual neurological signs, should therefore be viewed with suspicion (Slater & Roth 1969). The myocardial infarction is especially prone to be overlooked in hypertensive patients whose blood pressure may drop to normal levels after the episode.

With stagnant anoxias there are additional complicating factors. In addition to oxygen lack there is a deficiency of glucose supply to the brain, and accumulation of waste products leading to acidosis and other local metabolic derangements.

Metabolic or ’toxic’ anoxia

Metabolic or ’toxic’ anoxia occurs when there are faaors which interfere with the utilisation of oxygen by the tissues even though there is no lack of oxygen supply. This is seen with hypoglycaemia and with poisoning by substances such as cyanide or carbon disulphide. As mentioned above carbon monoxide has a histotoxic effect in addition to reducing the oxygen-carrying capacity of the blood.

Over-utilisation anoxia

Overutilisation anoxia exists when there is an increased demand in relation to supply. This occurs locally in the brain during epileptic seizures. ,

Clinical features

The clinical piaure varies with the individual situation, depending on the rate at which anoxia develops, its duration and the degree of concomitant physical exertion. Individual tolerance also varies widely, with more severe effects and a greater likelihood of sequelae in persons who are already existing near the threshold of neuronal reserve. Coexisting hypotension is often an important, and sometimes a crucial, factor in determining the severity of brain damage, as after temporary cardiac arrest (Brierley 1970).

Most forms of cerebral anoxia are transient events, presenting with impairment of consciousness of varying severity, confusion, disorientation or delirium. There may be muscular twitching or tremor, and epileptic fits may occur. In sustained hypoxia, as at high altitudes, mental obfuscation may develop more gradually and personality change may be the obtrusive feature as described below. In either event the diffuse nature of the cerebral affeaion is apparent.

Unless the anoxia is sufficiently prolonged or severe to cause death there are rarely enduring sequelae. A dense amnesic gap for the acute condition is usually all that remains. In a minority, however, permanent memory difficulties may result, or more global impairment of intellea extending even to severe dementia. Neurological sequelae may consist of extrapyramidal disturbances with parkinsonism or athetosis. Hemiplegia, blindness or

Page 548 CHAPTER 11

pseudobulbar palsy may very occasionally remain. After the anoxia which follows abrupt systemic hypotension there may be a monoplegia and sensory loss affecting the arm predominantly, due to the selective involvement of neurones at the borderland territories of supply of the anterior and middle cerebral arteries (see below). Parietal lobe deficits may also be severe.

The psychological changes produced at high altitudes and the clinical complications of carbon monoxide poisoning are considered in more detail below. The sequelae of cardiac surgery are also briefly reviewed.

Cerebral pathology

The pathological changes in the central nervous system in fatal cases have been described by Hoff et al. (1945) and Richardson et al. (1959). They are usually remarkably similar regardless of the cause, and resemble those seen after hypoglycaemia (p. 545). Brierley (1976) presents a comprehensive review.

If death occurs within a few minutes there is little to be detected other than scattered petechial haemorrhages. Cerebral oedema occurs early, and changes in cerebral blood flow raise the intracranial pressure further in a manner which aggravates the situation (Brock 1971). If the patient survives long enough there may be widespread degeneration and necrosis of nerve cells with corresponding glial proliferation. The cells of the third, fourth and fifth cortical layers are particularly susceptible (laminar cortical necrosis), also the Purkinje cells of the cerebellum and the cells of the corpus striatum. Bilateral necrosis of the globus pallidus is often a marked feature. The hippocampus and the parastriate cortex are also characteristically affected severely. Areas of subcortical demyelination eventually become apparent if the patient survives for more than a few weeks. The subcortical U fibres are, however, characteristically spared.

The biochemical changes set in train by hypoxic brain damage, and the particular role of glutamate excitotoxicity in leading to cell damage and death, are gradually being clarified by experimental studies. Some of the principal findings to date are described on pp. 164-5. Closely similar biochemical changes appear to ensue whether the brain damage is hypoxic or hypoglycaemic in origin, or the result of direct brain trauma.

A different pathological picture results when blood flow to the brain has been abruptly curtailed despite a sustained normal arterial oxygen tension. Thus the sudden systemic hypotension occasioned by temporary cardiac arrest, or severe myocardial infarction, may lead to ischaemic neuronal destruaion in restricted areas of the brain. Brief periods of ischaemia cause selective cell necrosis in the hippocampus and cerebellum, the CA1 zone of the hippocampus being particularly affected. Longer and more severe episodes of hypotension produce infarcts largely confined to the boundary zones between major arterial territories in the cerebrum and cerebellum. Brierley (1970, 1976) reviews the experimental and clinical evidence on the matter. Changes can be minimal or absent in the hippocampi and diffuse laminar cortical necrosis does not occur. Boundary zone necrosis is often most severe in the parieto-occipital regions where the territories of the anterior, middle and posterior cerebral arteries meet. Involvement of the subcortical white matter is roughly proportional to the severity of the cortical lesion, sometimes extending inwards as far as the wall of the lateral ventricle. Accompanying lesions in the basal ganglia are variable but often circumscribed and severe. In the cerebellum the lesion is at the junction of the superior and posterior inferior cerebellar arteries, forming a wedge with its base at the cortical surface and its apex in the central white matter.

Such classic boundary zone lesions may sometimes be combined with diffuse ischaemic changes, depending on the speed of development and severity of the hypotension. If the fall in cerebral perfusion pressure is abrupt and considerable, ischaemic necrosis will tend to be confined to the boundary zones; if it is slow and relatively moderate the changes will often be more generalised, since the accompanying vasodilatation will allow a more uniform distribution of the available but inadequate blood flow. Coincident respiratory failure will often also have occurred in examples encountered clinically, and a cornbination of the two factors – primary hypoxia and systemic hypotension – will often have been operative together.

The electroencephalographic changes in anoxia are identical with those of hypoglycaemia as described on p.


Anoxia at high altitudes

The psychological changes produced at high altitudes have received a good deal of attention, with renewed interest in view of the current vogue for climbing without supplementary oxygen. Important early observations were made on expeditions such as the International High Altitude Expedition to Chile (McFarland 1932,1937) and in experiments conducted in low pressure chambers where high altitudes could be simulated (Haldane et al. 1919; Barcroft 1925). Aviation experiences, particularly during the First World War, underlined the importance of this area of study. The earlier literature is comprehensively reviewed by Van Liere and Stickney (1963).


The slow ascent in mountaineers allows a process of adaptation. Individual variability is marked, some subjects being affected at 2500 m and some only at 4500 m Hypoxia due to the decline in partial pressure of oxygen in the inspired air is accepted as the principal cause, though fatigue, extreme cold, mountain glare and isolation will also contribute to the picture.

Ryn (1988), reporting on a group of Polish alpinists, has described the typical progression from states of dysphoria or sometimes elation at 3000-4000 m, to features of an acute organic reaction at over 7000 m. The early dysphoria shows as apathy, depression, mental exhaustion and slowing of thinking processes. Some subjects, by contrast, show elevation of mood coupled with irritability, argumentativeness and impulsivity. At rather higher levels there may be cyclic alternations between the two pictures. The symptoms of the acute organic reaction at 7000 m include a marked decrease in motor and mental activity, clumsy behaviour and a blunted awareness of dangers. Periods of dense amnesia lasting from minutes to days may occur, also excessive somnolence. Brief lapses of consciousness are fpllowed by confusion, purposeless activity and ataxia. Occasional subjects become severely disorientated, especially towards night-time, with tearfulness and visual and auditory hallucinations. West (1986) quotes examples of bizarre subjective experiences, such as seeing pulsating cloud-like objects in the sky, or strong feelings that a second person is present. Disturbances of judgement and faulty reasoning can impose serious risks to the climber and others. Headache, nausea, anorexia and tachycardia are common additional features.

A prolonged stay at high altitudes may be accompanied by an entire change of personality. Along with episodes of mental confusion there is progressive indifference and apathy, mild depression and persistent inability to concentrate. Refreshing sleep is hard to obtain, and insomnia worsens the emotional and intellectual abnormalities. Dreams are frequently of an unpleasant and frightening nature, though at still higher altitudes dreaming apparently ceases to occur.

On return to normal altitudes there are rarely enduring deficits of a conspicuous nature. An early report by Pugh and Ward (1956), however, described residual memory difficulties in a mountaineer who had climbed to 8500 m without additional oxygen. Townes et al. (1984) made careful observations on members of the 1981 American Medical Research Expedition to Everest, testing them before and after the ascent, and found evidence of mild transient deficits in memory, learning and verbal expression. These had disappeared at the 1-year follow-up. However, impairments in finger tapping speed persisted at 1 year suggesting that motor deficits might be longerlasting.

Townes et al’s subjects had used supplementary oxygen, unlike those reported by Regard et al (1989). Eight world-class climbers who had reached summits of over 8500 m were tested an average of 7 months later. In comparison to controls several were still mildly impaired on tests of concentration, verbal and figural memory and cognitive flexibility (Stroop test and concept matching tasks). By contrast perceptual ability, reaction times, language and visuospatial abilities were within normal limits. They regarded their results as indicative of longlasting frontotemporal dysfunction; this was supported by EEG abnormalities in the temporal regions in the two most severely affected subjects.

Among aviators flying without oxygen the onset of symptoms is more dramatic, appearing often at 3000 m though depending again on the rate of ascent and individual susceptibility. Difficulties are universal above 6000 m without additional oxygen, and complete loss of consciousness is likely to occur at 7500m. The first manifestations are variable – fatigue and mental slowing, euphoria, or surly and pugnacious behaviour. Headache, nausea and vomiting are again usually in evidence. Motor control soon becomes impaired and reaction times are slowed.

The subject is typically unaware of the change in himself and becomes oblivious of danger. He may feel his mind to be unusually keen and develop a dangerous fixity of purpose from which he cannot be dissuaded. Thus foolhardy and reckless behaviour is common. Birley (1920) reports some interesting examples of altered judgement in pilots during the 1914-18 war. One returned from a high photographic reconnaissance flight well pleased with his effort until it was found that he had taken 18 exposures on the same plate; others had attacked enemy formations without any plan of campaign, making tactical errors of which they would not ordinarily be guilty. One officer had been known to wave his hand in friendly greeting to the enemy.

Experimental observations in the decompression chamber, or with re-breathing apparatus, have amply confirmed these changes and shown their essential dependence on cerebral hypoxia. Severe exposures can lead to loss of consciousness without warning. Less severe degrees produce a picture closely resembling alcoh’blic intoxication, with headache, drowsiness, confusion, muscular weakness and incoordination. An initial stage of euphoria is again often seen with feelings of self satisfaction and unusual power. The lifting of inhibitions sometimes leads to hilarious silly behaviour, anger or outbursts of uncontrolled laughter. Both the quantity and quality of

Page 550 CHAPTER 11

mental work are affected; reasoning is slow, calculations faulty and writing shows perseverative errors. Distractibility becomes marked and attention to detail is lost. Choice reaction times are affected earlier than simple reaction times, but these too are slowed as neuromuscular control deteriorates. Visual and kinaesthetic perception are impaired, and later auditory perception also. Memory and learning ability can be shown to decline from an early stage. The lack of subjective awareness of the changes and the dangerous loss of initiative which accompanies them have also been fully confirmed.

Carbon monoxide poisoning

In addition to attempted suicide with domestic gas supplies, carbon monoxide poisoning may occur accidentally in connection with slow combustion stoves, bathroom geysers and car exhaust fumes in enclosed spaces. Domestic risks have been greatly diminished in the UK since the introduction of North Sea gas which is not in itself poisonous, but carbon monoxide is still produced if faulty equipment leads to incomplete combustion. In industry poisoning is seen in connection with blast furnaces and when explosions occur in mines.

The clinical picture was fully documented by Bour etal. (1967). Lowered efficiency and self control lead imperceptibly to loss of consciousness without any intervening delirium. Complete unconsciousness is usually rapidly attained, resulting in coma of variable duration. Diffuse hypertonicity is common, with trismus and up-going plantar responses. Paroxysms of decerebrate rigidity may occur. Hypotonic forms are rare and carry a graver prognosis. Sphincter and swallowing difficulties are often present, and the corneal and pupillary reflexes may be abolished. The complexion is plethoric, and hyperpnoea is often extreme with a forceful expiratory phase. The classic ’cherry red’ colour is in fact rare. The cerebrospinal fluid is normal.

In Shillito et al.’s (1936) series of 21 000 cases twothirds were successfully resuscitated. If death is to follow this is usually within a few days, but sometimes survival extends for several weeks or months. Very occasionally the patient recovers for 10-15 days before a fatal relapse. Long persistence of coma after removal from the contaminated atmosphere carries an unfavourable prognosis, likewise prolonged circulatory collapse, fluctuating pyrexia, hyperglycaemia, uraemia or acidosis.


As coma regresses the patient shows a period of disorientation and confusion. Sometimes there is a brief phase of irritability and restlessness, but usually the predominant features are apathy and general inertia. Approximately one-fifth of patients show prolonged delirium, ranging from several hours to several weeks (Smith &- Brandon 1970). Hallucinations and delusions are conspicuous by their absence (Shillito et al. 1936). Speech is slowed and difficult for a time, and thought processes reduced. Amnesic difficulties are usually much in evidence and often the last to clear:

A woman of 70 was admitted in coma and only several weeks later had recovered sufficiently to respond to simple commands and reply to questions. For the next 2 weeks she was grossly disorientated in time, giving the year variously as 1888 or 1918, and being equally wide of the mark for her age. Another week passed before she realised she was in hospital, yet insisted it was in Armagh, the place of her childhood, though in fact she was in Belfast. At this stage she was able to repeat the months of the year correctly, and could repeat four digits, but could retain nothing of a name and address after a 3-minute interval. She could remember the names of her brothers and sisters and recall some episodes from her childhood. Two weeks later she had improved remarkably. She was then consistently orientated in time and place, had good general insight and was able to recall the circumstances leading right up to her accidental coal gas poisoning. . (Allison 1961)

Sometimes a classic Korsakoff syndrome emerges, or alternatively the patient’s profound inertia, indifference and slowing of cerebration may cause the memory to appear to be more defective than is really the case (’une sorte d’indifference amnestique’ – Ajuriaguerra and Rouault de la Vigne 1946). Agnosia, constructional apraxia or dysphasia may also be seen later in recovery.

Extrapyramidal signs which are absent during the stage of coma may emerge on recovery, with cogwheel rigidity of the musculature, athetosis or immobility of facial expression. Transient hemiplegia is not uncommon, and 8th nerve affections may produce deafness and vertigo. In most cases the neurological abnormalities gradually clear along with the psychological disturbances, though sometimes this may take several weeks or months. The degree of recovery from early severe disability can sometimes be remarkable:

A 24-year-old woman was admitted to hospital in deep coma. There were repeated tonic seizures, breathing was laboured and she required a tracheostomy. After several day^ in coma she passed into a muttering resistive state for some weeks, then became mute and unresponsive for 3 weeks more. At this stage she showed a quadriparesis and bilateral extensor plantar responses. Six weeks later she began to show some recovery of consciousness and ability to move, but remained confused and severely aphasic. Four months after the initial exposure, however, there had been a good deal of recovery; she was then talking well and was discharged from hospital. She continued to


improve in a remarkable fashion, and by 8 months there was only a moderate hemiparesis and slight mental impairment.

(Richardson et al. 1959)

Latent interval

A strange feature, repeatedly noted in large series of cases, is the occurrence of a latent period between recovery from coma and the onset of profound neurological or mental disorder (Bour et al. 1967; Choi 1983; Lee & Marsden 1994). This may also follow other forms of cerebral anoxia, as after anaesthesia or cardiac arrest (Plum et al. 1962). Apparently normal health is regained, but 2-10 days later, or even after several weeks, there is an abrupt relapse with the appearance of extrapyramidal disturbance and personality and cognitive changes. Delirium or coma may recur. Sometimes the patients have been discharged or even returned to their jobs in the interim. Complete recovery is again often attained:

A man of 37 was accidentally gassed while intoxicated. He was unconscious for 2 days but for the rest of the week seemed entirely recovered. On the seventh day he began to act peculiarly, became unresponsive to questions and the right plantar response was found to be up-going. Several weeks later he was still dull, apathetic and disorientated, but thereafter improved and was ultimately discharged fully recovered.

A woman of 51 was unconscious for 2 days, but was discharged at the end of a week fully recovered apart from a general feeling of weakness. After 2 weeks, however, her gait became unsteady, and she became confused and bewildered. Her face was blank and expressionless and she developed a coarse tre-mor of the hands. She gradually improved over the next few months, with disappearance of the mental and neurological abnormalities.

(Shillito et al. 1936)

A latent interval of this nature has been reported in 3% of patients and in 12% of those who are hospitalised (Choi 1983). The apathy, mutism and indifference which develop may then be falsely interpreted as reflecting psychiatric disorder, especially if the patient has attempted suicide (Zagami et al. 1993). Delayed sequelae tend to be commoner when the initial period of unconsciousness has been prolonged, but Choi observed several examples in patients who had remained mentally clear during the acute episode. The ultimate outcome is often reasonably good, 75% of the patients reviewed by Choi recovering within a year, while the remainder showed no improvement or died. Parkinsonism or mild memory impairment persisted in a fifth of those who recovered.

The underlying pathology appears to be demyelination within the cerebral hemispheres though the mechanism responsible is unclear. In 15 patients reported by Chang et al. (1992) MRI invariably showed confluent regions of high signal intensity in the centrum semiovale and periventricular white matter. Ten also showed alterations in the thalamus and putamen suggestive of iron deposition, and nine showed bilateral ischaemia or necrosis of the globus pallidus. Repeat imaging in three patients whose clinical symptoms improved showed decreases in the extent and intensity of the white matter changes. In Hayashi et al.’s (1993) patient an area of demyelination apparent on CT in the centrum semiovale was observed to enlarge during the latent period, suggesting that the white matter damage was progressive throughout.

Enduring sequelae

The majority of patients are left only with an amnesic gap for the period of coma, together with a variable amount of retrograde amnesia and patchy memory for part of the period of recovery. The commonest neurological sequelae are extrapyramidal disturbances, usually parkinsonian in nature or sometimes athetotic. In the most severe examples the patient is left in an akinetic mute state. This carries a grave prognosis and a high mortality. Global impairment of intellect is uncommon but permanent defects of memorising may be seen. Neurological disabilities are almost invariably evident when gross mental impairments remain, though occasional patients are described with purely cognitive and behavioural deficits (Gordon & Mercandetti 1989; Mendez & Doss 1995).

It was formerly claimed that persistent sequelae were rare. Thus over a 10-year period Shillito et al. (1936) found that 21 000 cases of carbon monoxide poisoning had occurred in New York City with 14 500 recoveries, but only 39 patients (0.27% of survivors) could be identified who had been hospitalised with enduring defects. It seemed therefore that the great majority of patients either succumbed or recovered completely. Lee and Marsden (1994) point out, however, that carboxyhaemoglobin levels were not measured in Shillito’s survey, so that patients with intoxication by drugs or alcohol and minimal carbon monoxide exposure may have been included. Most of those with permanent sequelae had been deeply comatose for hours or days, and had often required long periods of assisted respiration when first discovered. Van Amberg (1942) found that recovery within an hour of removal from the contaminated atmosphere carried a uniformly good prognosis.

The investigations of Smith and Brandon (1970,1973), however, revealed a more serious picture. Two hundred and six cases of carbon monoxide poisoning were traced from a defined geographical area (Newcastle-upon-Tyne),

42% being accidental and 58% due to suicide attempts. Of the 135 survivors, three patients (2.2%) were severely

Page 552 CHAPTER 11

affected at the time of discharge, all showing dementia and one with parkinsonism in addition. Two others had spastic hemiplegia, though it was possible that this had developed immediately before exposure.

Follow-up was then undertaken among those who had been domiciled within the city at the time of poisoning. Seventy-four patients were traced at an average of 3 years after the event. Eleven had died (15%), three from suicide and eight from other causes. Eight (11%) had evidence of gross neuropsychiatric damage which was directly attributable to the poisoning, in the form of cognitive disability, personality change or severe neurological abnormalities. In some cases deficits such as parkinsonism, dementia or personality change had become evident only after discharge from hospital, emphasising the importance of following every case for sufficient time to allow for the emergence of delayed sequelae. Five of the eight showed a severe global deterioration of intellect.

Of the 63 alive at follow-up, 27 patients (43%) cornplained of impaired memory and this correlated highly with objective deficits on the Wechsler Memory Scale. Twenty-one (33%) showed a deterioration of personality subsequent to the poisoning, with a marked association between personality and memory deficits. The commonest personality change was towards increased irritability, verbal aggressiveness, violence, impulsiveness and moodiness, a constellation described by the authors as ’affective incontinence’. Thus, over and above the 11 % with gross sequelae, a large number appeared to have suffered milder brain damage, resulting in personality and affective change with associated mild cognitive impairment. This was a great deal more common than in a matched control group of patients who had survived barbiturate overdosage. The level of consciousness on admission to hospital correlated significantly with the development of gross neuropsychiatric sequelae and with complaints of memory impairment.

An example from Smith and Brandon’s (1973) series with severe personality change was as follows:

A 33-year-old miner of good premorbid stability was accidentally exposed to carbon monoxide in a coal mine explosion. He was comatose when found some minutes later, and subsequently was delirious, irritable and aggressive for 4 hours. Oxygen was given and he was discharged from hospital the same day. Subsequently he was forgetful, had difficulty in coping with his previous level of work, was increasingly irritable, restless and argumentative, and on occasion was violent towards his wife. He was more impulsive and outspoken, more anxious and more prone to paranoid misinterpretations. His physical energy was markedly diminished. In the month before interview he had been charged with indecent assault on five young girls. His wife verified his previous mental stability and the deterioration of memory and personality subsequent to exposure. An independent witness, unaware of the patient’s fate after the accident, attested to his previously stable personality. Financial compensation had never been considered.

Lee and Marsden (1994) attempted to determine features that might predict the level of disability 1 year later among 31 patients who developed neurological sequelae. Eight had shown a progressive course to a persistent vegetative state, whereas 23 had relapsed after a latent interval. The former were significantly younger than the latter, and the mean duration of coma had been significantly longer (10 days and 2 days, respectively). Levels of carboxyhaemoglobin, however, did not differ between the groups. None of the eight vegetative patients improved.

Among the 23 with delayed sequelae two patterns could be seen – relapse to a parkinsonian state with behavioural and cognitive impairment (nine cases) or relapse to a bed-bound state with mutism, apathy and severe rigidity (14 cases). Such relapses occurred rapidly over the course of several days. The majority of the former, but less than half of the latter, improved over subsequent months. Occasional patients in either group showed a second phase of deterioration. A third of the CT scans were normal at the start, and contrary to previous reports the presence of features such as white matter low density or globus pallidus lesions were not good predictors of outcome. Follow-up scans could show new lesions or progressive changes even in patients who improved. Bruno et al. (1993) similarly found that MRI changes did not correlate well with clinical outcome.

Cerebral pathology

Lapresle and Fardeau (1967) describe the characteristic changes in the brain in cases of carbon monoxide poisoning coming to autopsy. The most frequent lesion is necrosis of the globus pallidus, sometimes limited to a small and circumscribed portion bilaterally. Similar lesions are common in Ammon’s horn, again variable in extent and sometimes affecting Sommer’s sector only. The cerebral cortex shows necrotic foci with intense capillary proliferation and degeneration or disappearance of nerve cells, but the typical laminar cortical damage seen with other forms of anoxia is rare. The cerebellum also shows necrosis but the Purkinje cells are relatively spared, again in contrast to their special vulnerability in other forms of anoxia.

Lesions are common in the white matter of the centrum ovale. Here foci of necrosis can be extensive and coalescent with areas of demyelination. The latter appears to be particularly characteristic of cases which have shown a latent interval during their clinical course. Axis cylinders then tend to be relatively preserved even when demyelination


is extensive in the deep white matter and periventricular areas.

Vascular lesions are widespread, with endothelial swelling leading to thrombosis and miliary diapedetic haemorrhages. Sometimes there is marked general vasodilatation. Cerebral oedema occurs regularly in the early stages and may contribute to the ischaemic changes. The lesions of the globus pallidus and Ammon’s horn are thought to rest on a common vascular basis.

Brucher (1967) discusses the aetiology of the lesions. The cerebral pathology differs from that of anoxia due to other causes in certain important respects- laminar cortical damage is rare, the Purkinje cells are relatively preserved and lesions are prominent in the white matter. This suggests that carbon monoxide may have a specific histotoxic action in addition to its effect in reducing the oxygen-carrying capacity of the blood. Alternatively both the white matter changes and the lesions in the globus pallidus could depend critically on circulatory failure, i.e. hypoxia compounded by systemic hypotension and acidosis (Ginsberg 1985). Hence the importance in treatment of prompt attention to hypotension and the acid-base balance in addition to the administration of oxygen. A haemorrhagic origin for the pallidal lesions is supported by MRI findings indicative of the presence of methaemoglobin within them (Silverman et al. 1993).

Chronic carbon monoxide poisoning

From time to time it has been claimed that intermittent exposures to low concentrations of carbon monoxide can have a cumulative effect, resulting in chronic disability. In Scandinavian countries during the 1939-45 war this became a focus of concern among persons handling ’producer gas’ or driving vehicles in which this was the source of power (Grut 1949). Other industrial occupations at possible risk are foundrymen, gas workers and blastfurnace workers.

Proponents of the disorder point out that breathing as little as 0.05% of carbon monoxide in the inspired air leads to symptoms after several hours, and that experiments with animals show pathological changes in the central nervous system after prolonged sublethal exposures (Beck 1936). Beck reported 97 patients who had suffered many weeks or months of exposure to domestic gas, vehicle exhaust or blast furnace fumes, and concluded that a definite clinical syndrome was produced. The chief complaints were of dull frontal headache, vertigo, nervousness, generalised weakness, neuromuscular pains, anorexia, digestive disturbances, dyspnoea, palpitations, tremulousness, weakness of the legs with ataxia, and paraesthesiae of the extremities. Almost all showed prominent mental manifestations and many were ’confirmed neurotics’. Psychological symptoms included depression, restlessness, anxiety, mental retardation, memory deficits and periods of confusion. Drowsiness and insomnia were common. Some showed speech defects, tinnitus, visual disturbances or impairment of the sense of smell. Four patients showed parkinsonian features.

The syndrome, however, has come heavily under suspicion (Lindgren 1961), and evidence for an objective physical basis has not been forthcoming (Slater &- Roth

1969). Hunter (1978) pointed out that carbon monoxide is not a cumulative poison and that small amounts are readily ventilated out of the blood; he therefore concluded that there was no such clinical condition as chronic carbon monoxide poisoning. However, the subject remains under review in view of risks to health from urban atmospheric contamination (e.g. Goldsmith & Landaw 1968). It is also important to beware of minor exposures resulting from chronically malfunctioning domestic equipment. The occurrence of fatigue, headache, diarrhoea and flu-like symptoms amongst several members of a household, or recurring at a particular time and place, should raise suspicion of possible minor intermittent exposure (Drug & Therapeutics Bulletin 1988c). Evidence incriminating intermittent exposure to carbon monoxide as the cause of increased liability to atherosclerosis in smokers is also relevant in this regard (Astrup 1972).

Sequelae of cardiac surgery

During the early era of open heart surgery complications by way of anoxic brain damage and stroke were frequent, occurring in over 20% of patients (Hotson 1989). Improvements in surgery and in the techniques for extracorporeal circulation have substantially reduced the incidence of disabling sequelae, but more sensitive measures show that after-effects of a subtle nature are still not uncommon. Moreover, as confidence in coronary artery bypass surgery has grown the operation is offered to patients at greater risk by virtue of age, advanced ischaemic heart disease or arteriosclerosis elsewhere.

The principal risks are from cerebral hypoperfusion during operation, and embolism caused by material dislodged from the aorta and cardiac chambers. Major) embolism leading to stroke is commoner when the heart needs to be opened as for valve replacement, but may also occur after coronary bypass surgery. Multiple microembolism is a great deal commoner, resulting from the passage of air, platelet aggregates, fibrin or fat into the cerebral arterial tree.

Page 554 CHAPTER 11

It can often be difficult to apportion the blame between microembolism and hypoperfusion in leading to some of the common diffuse cerebral sequelae described below. Early neuropathological studies showed evidence of both, with scattered embolic infarcts and anoxic changes most prominent in the hippocampus. Major advances in reducing such complications have come from the incorporation of microfilters in the extracorporeal circulation, from adequate heparinisation and haemodilution, and from maintenance of the mean arterial blood pressure during operation above 50 mmHg.

Clinically obvious sequelae

The more obvious sequelae include stroke, ophthalmic abnormalities and postoperative confusion or delirium. Shaw et al. (1985) examined 312 patients immediately after coronary bypass surgery and found neurological changes in 61%. Serious morbidity was, however, rare. Definite stroke occurred in 5% though this was rarely severely disabling. Ophthalmic abnormalities were found in 25% by way of retinal infarcts, visual field abnormalities, or reductions in visual acuity for which no definite cause was discovered. Three per cent showed prolonged depression of consciousness not attributable to sedative drugs, and 1 % a delirious state with hallucinations and paranoia. One patient sustained fatal cerebral hypoxic damage. Primitive reflexes were detected in 39%, chiefly palmomental and pout reflexes, and 12% had sustained brachial plexus injuries from torsion or compression during surgery.

Postoperative delirium is usually transient, resolving within a few days. Sometimes it follows a lucid interval of 2 or 3 days and the cause may then be uncertain. It has been reported in over a quarter of subjects in some series (Mayou 1986). Sleep disturbance, medication and the relative sensory deprivation induced by the stay in an intensive care unit are probably often contributory factors.

’Soft’ neurological signs and neuropsychological deficits

The high prevalence of primitive reflexes noted by Shaw et al. may reflect hypoxic-ischaemic damage. Other minor neurological signs may be detected immediately postoperatively, such as transient nystagmus, impairment of coordination or depressed reflexes. These almost invariably recover within a few days. Neuropsychological assessment has likewise shown that early deficits are common, with falls from preoperative levels in 80% of patients (Shaw et al. 1986). Such deficits chiefly involve problems with attention and concentration, psychomotor speed and verbal learning. Almost half of those who had deteriorated in Shaw et al.’s sample made no complaints of impairment, and only 10% were overtly disabled in everyday ward activities. However, dysfunction on testing correlated significantly with the presence of new palmomental reflexes, suggesting that it was likely to be due to diffuse brain damage of a subtle nature.

Sotaniemi et al.’s (1986) long-term follow-up study raised the strong possibility that such deficits were not always evanescent. Forty-four patients who had had valve replacements were examined 10 days postoperatively for minor neurological signs and at 8 weeks for neuropsychological change. Almost three-quarters showed improvement on neuropsychological testing, while a quarter were impaired in relation to preoperative performance. At follow-up 5 years later the group with early neurological signs or psychometric impairment were more liable to report some degree of memory difficulty, and showed significantly poorer performance on testing than the remainder. When divided into those who had had short or long perfusion times during operation, the latter performed significantly less well at 5 years, irrespective of the presence or absence of early neurological signs.

The immediate postoperative outcome thus appears to have implications for long-term psychological performance, with predictive value attaching even to asymptomatic early deficits. A good deal of effort has accordingly been devoted to exploring their possible origins.

Studies from the cardiothoracic unit at the Middlesex Hospital, London, have compared patients undergoing coronary bypass surgery with controls having major vascular or other thoracic surgery (Smith et al. 1986; Newman et al. 1987; Treasure et al. 1989). Transient neurological dysfunction as reflected in ’soft’ neurological signs was found in some 60% of bypass patients at 24 hours compared with 20% of controls. This, however, may have been largely related to the dosage of analgesics and sedatives employed. Deterioration on a battery of neuropsychological tests was detectable in almost threequarters of bypass patients at 8 days, and was significantly more likely to occur with older age, longer time on extracorporeal circulation and lower perfusion pressures. By 8 weeks significant recovery had taken place but deficits persisted in over a third. Similar neuropsychological changes were observed in half of the controls but these patients were somewhat older. Cerebral blood flow measurements using intravenous i”Xe showed significant reductions in the coronary bypass group at 8 days, returning to preoperative levels at 8 weeks. This mirrors Henriksen’s (1984) finding with single photon emission


tomography. There were no significant changes in blood flow in the comparison group.

Follow-up of a substantial proportion of these bypass patients showed that the deficits present at 8 weeks tended to remain relatively stable, being still detectable 12 months later (Newman etal. 1990). Patients with enduring deficits were again significantly older than those without, and had had significantly longer periods on extracorporeal circulation. So far these results could reflect either microembolism or haemodynamic factors related to perfusion, since the influence of both will stand to increase with the duration of extracorporeal perfusion.

Investigating possible influences further, the same group has examined the effects of different forms of oxygenation during extracorporeal bypass, and the effects of introducing arterial line filters. A sheet membrane oxygenator, in comparison with a bubble oxygenator, produced a marked decrease in retinal microembolism as assessed by retinal fluorescein angiograms, also a trend towards reduction in neuropsychological deficits on the eighth postoperative day (Blauth et al. 1989). Comparisons between patients randomised to the presence or absence of a 40 um micropore filter in the extracorporeal circulation showed that it significantly reduced the incidence of neuropsychological deficits at 8 days (Pugsley et al. 1994). Doppler ultrasound, used to monitor the presence of microemboli in the middle cerebral artery, confirmed that they were reduced when the filter was used, and the number deteaed correlated with the likelihood of the patient having neuropsychological deficits later. Clearly, therefore, microemboli, whether gaseous or particulate, make a contribution to the subtle deficits deteaed postoperatively. The extent to which other aspeas of the perfusion process may contribute through hypoxic brain damage remains uncertain.

In a small preliminary study Toner et al. (1994) have performed MRI imaging before and after bypass surgery. One week postoperatively four of 15 patients showed additional MRI changes by way of punaate hyperintense areas which could have been due either to microembolism or perfusion-related ischaemia. Aberg et al. (1982,

1984) have produced biochemical data which point towards the influence of hypoxia-ischaemia; adenylate kinase, a marker of cell membrane disruption, was shown to be elevated in the cerebrospinal fluid in patients who had had extracorporeal bypass 24 hours earlier, and to correlate with the extent of the ensuing neuropsychological deficits.

Cardiac transplantation

carries all of the above risks, with additional problems due to postoperative immunosuppression and the attendant risk of infeaions. Hallucinations, delusions and disorganised behaviour can occur during the first 2 weeks, and multiple faaors will often contribute to such pictures including the administration of steroids and sedatives. When developing later than this, psychotic disturbances will often be indicative of intracranial infeaion (Hotson & Pedley 1976).Uraemia

Uraemia may result from primary disease of the kidneys or from extrarenal causes. Any process which results in a prolonged and severe reduaion of blood flow through the kidneys can produce potentially reversible renal failure due to tubular damage, for example shock or dehydration following operations, burns or crush injuries. The uraemia is then often aggravated by the increased protein catabolism that ensues.

Disordered mental funaioning forms a prominent part of the syndrome of uraemia, both in chronic renal failure and in these more acute disturbances. In Stenback and Haapanen’s (1967) large consecutive series of patients seen in a renal unit, mental manifestations occurred in 60%, rising to 75% when the blood urea exceeded

250 mg/100 ml. Mental changes were as common in the acute as in the chronic uraemic patients.

Presentation with psychiatric features

Uraemic patients are usually obviously unwell, but occasionally the mental changes can be the most prominent manifestation and lead direaly to psychiatric consultation. This is more likely when the uraemia has developed slowly. The piaure may simulate neurasthenia with symptoms of lethargy, anorexia and depression. Or early dementia may be suspeaed in view of sluggish comprehension and difficulty with memory. Elderly demented patients may suffer additional mental impairment from unsuspeaed uraemia, sometimes as a result of prostatic hypertrophy and an associated hydronephrosis. More rarely an acute organic reaaion may be the first indication, and uraemia must be constantly borne in mind in the differential diagnosis of delirium of uncertain aetiology.

Common psychological accompaniments

By far the commonest mental disturbance is progressive torpor and drowsiness with the insidious development of intelleaual impairment. Stenback and Haapanen (1967) describe the following sequence of changes: The first complaint is usually of feeling generally

Page 556 CHAPTER 11

unwell, with fatigue and incapacity for physical or mental effort. The difficulty with concentration is characteristically episodic, so that the patient performs well for short periods of time but cannot sustain mental activity. There may be no other definite features on examination at this stage, though headache and anorexia are usually prominent complaints.

With further progression memory becomes obviously impaired, and episodes of disorientation and confusion appear. Listlessness and apathy prevail but an anxious restlessness may sometimes be seen. Depression and emotional withdrawal are usually marked. There may be petulant demanding behaviour, with negative attitudes towards treatment and hostility towards attendants, which later clears with clinical improvement. Both the impairment of consciousness and the changes of mood typically fluctuate markedly, with lucid periods during which behaviour returns to normal. Neurological accompaniments include fascicular twitching, tremor, multifocal myoclonus and muscle cramps. Flapping tremor of the outstretched hands may be observed, similar to that of hepatic encephalopathy (p. 564), On recovery there is patchy or complete amnesia for the periods of disorientation and confusion.

Episodes of acute delirium appear in a third of cases, with apprehension, bewilderment and fleeting hallucinations which may be terrifying in content. As with delirium due to other causes the picture changes rapidly from time to time and paranoid developments are common. Eventually more profound impairment of consciousness develops with increasingly sluggish comprehension and reaction, slurring of speech, incontinence and ultimately coma.

Epileptic fits develop in about a third of cases, more frequently in acute than chronic uraemia. They are usually a late feature unless hypertension is present at an early stage. The combination of clinical signs of cerebral depression together with excitability, as manifest in seizures, is rare in other metabolic encephalopathies (Raskin 1989). When the blood pressure is very high there may be episodes of hypertensive encephalopathy, with severe headache and transient attacks of blindness, dysphasia or monoplegia.

Peripheral neuropathy is common in chronic renal failure. Sensation is impaired distally and symmetrically, reflexes are diminished and intense painful paraesthesiae sometimes occur (Mawdsley 1972). Even in the absence of clinically overt neuropathy impairment of peripheral nerve conduction can often be shown. The ’restless legs syndrome’ (p. 640) develops in a considerable proportion of patients, with or without overt neuropathy, presenting with unpleasant sensations in the lower limbs, worse

in the evenings and relieved by movements (Raskin 1989).

Psychoses with uraemia

Occasionally, and presumably depending on the premorbid constitution, the picture may closely simulate a schizophrenic or affective psychosis. Baker and Knutson (1946) reported a patient initially diagnosed as suffering from catatonic schizophrenia, but with rapid fluctuations between psychotic and normal periods. Menninger’s (1924) patient presented with mounting paranoia over many months, culminating in a florid psychotic illness with clear organic features. Psychotic depression may likewise occur, and occasional manic reactions have been observed. In all of these, close examination will usually betray organic mental symptoms. The obvious impairment of physical health, with weakness, weight loss and anorexia, should alert one to the true situation.

Depression may usually be treated safely with tricyclic antidepressants, though patients receiving dialysis will warrant careful monitoring of blood levels (Rosser 1976). Phenothiazines should where possible be avoided on account of their enhanced liability to induce dystonia and other movement disorders in the presence of uraemia.

’Acute azotaemic psychotic encephalitis’

At one time it was believed that certain conditions existed in which an acute psychotic illness was accompanied by uraemia but was not directly caused by it. A primary cerebral pathology was thought to lead on the one hand to the psychosis and on the other to the uraemia. In the French literature an ’acute azotaemic psychotic encephalitis’ was described, presenting with acute delirium and fever and progressing to coma and usually death, with evidence of degenerative and inflammatory lesions in the central nervous system. The case was vigorously argued that the uraemia was secondary to lesions in vegetative centres of the nervous system rather than their cause, just as uraemia might occasionally complicate epidemic encephalitis or delirium tremens (Marchand 1953). The ’fatal catatonia’ of German workers was viewed similarly (Stauder 1934; Arnold 1949).

More recent knowledge allows most of these cases to be seen as examples of acute renal failure secondary to extrarenal causes, with the psychotic manifestations being attributable to the uraemia in the usual way. Alternatively, acute renal failure may have complicated a psychosis already brought about by other factors. It is noteworthy that the illnesses often followed operations, deliveries or acute infections. Links have been drawn


with the ’neuroleptic malignant syndrome’ of presentday practice (p. 626).


The blood urea is raised and electrolyte disturbances are common. Sodium depletion can be marked, the serum phosphate elevated, and bicarbonate and calcium reduced. In the later stages the serum potassium may rise. Anaemia can be considerable in patients with chronic uraemia.

Electroencephalographic changes develop roughly in proportion to the severity of the clinical condition, and are rarely absent when the blood urea exceeds

60 mg/100 ml (Tyler 1968). They resemble the changes seen in hepatic decompensation, namely lowering of voltage with loss of well-developed alpha activity, then progressive slowing and disorganisation with runs of waves occurring at 5-7 per second which ultimately replace all other activity. Secondary hyperparathyroidism may play a part in contributing to the slowing (Raskin 1989). Epileptiform changes frequently appear.

The pressure of the cerebrospinal fluid is usually slightly raised, sometimes with moderate elevation of protein but without increase of cells.

Aetiology of mental disturbances

Mental and neurological changes become commoner with increasing elevation of the blood urea, but there is by no means a simple linear relationship. Nor can urea be regarded as responsible for all or even the majority of the symptoms that result; considerable improvement can follow dialysis when urea is present in the dialysis fluid, and experimentally urea has not been found to have a strong neurotoxic effect. Other abnormalities must therefore play a direct aetiological role, with the level of blood urea serving mainly as an indicator of the severity of overall metabolic disturbance. Nevertheless estimation of the blood urea remains an important clinical indicator for monitoring the progress of the patient’s condition.

Other derivatives of protein such as uric acid may play a part, as may other toxins so far unidentified. Electrolyte disturbances are certainly important in many cases. Changes in sodium, potassium, calcium, chloride, phosphate, acid-base balance and osmolality can all be blamed in individual instances. The rapidity of the shifts appears to be the essential factor, whether this is in the direction of normality or abnormality (Tyler 1968).

Drugs have been strongly incriminated as a cause of mental disturbance in patients with chronic renal failure, accounting for over a third of the episodes in some series

(Richet & Vachon 1966; Richet et al 1970). Sedatives and antibiotics appear to be mainly responsible, either by virtue of accumulation when the drug is excreted by the kidneys, or as a result of increased susceptibility of the central nervous system in the uraemic patient.

Arterial hypertension brings its own contribution by way of transient neurological disturbances, fits and headache. In many cases a part is played by raised intracranial pressure, changes in cerebral blood flow and cerebral oxygen utilisation, and altered permeability of small blood vessels.

Neary (1976) has drawn attention to other causes of neuropsychiatric disturbance in patients with chronic renal failure. The risk of intracranial infection is increased in uraemia, especially when immunosuppressive drugs are used after renal transplantation; encephalitis due to herpes simplex or cytomegalovirus may be hard to diagnose in the prodromal stages, leading to behavioural disturbance and change of personality. Low-grade meningitis can lack the typical physical signs and present as depression with chronic headache. The use of anticoagulants in maintenance haemodialysis may result in a subdural haematoma.

The psychological stresses associated with haemodialysis or transplantation bring a range of problems of their own. Salmons (1980) has reviewed the problems inherent in such management of chronic renal failure, with disruptions in work, daily life and family relationships. Not surprisingly there is a high incidence of depression, anxiety and disturbed sexual functioning among such patients. Short-lived psychotic episodes may be observed, usually in clear consciousness but often marked by ’organic’ features such as visual hallucinations or loosely held delusions. Steroids administered after transplantation may make an important contribution to such developments (p. 628). The risk of brain lymphoma is very considerably increased after transplantation, largely but not entirely attributable to immunosuppressant treatment (Raskin 1989).

Disturbances occasioned by dialysis

Gradual improvement usually follows dialysis, with return of mental clarity a short while after chemical normality has been achieved. Occasionally, however, the time lag may be several days in duration.

If the patient is lucid beforehand he usually remains so, provided time is taken over the procedure. However, too rapid dialysis carries a hazard of worsening the clinical situation (dialysis disequilibrium syndrome). Headache, muscle twitching, fits and confusion may develop, sometimes progressing to coma with signs of brain stem compression

Page 558 CHAPTER 11

(Peterson & Swanson 1964; Mawdsley 1972). This is commoner when the initial level of blood urea has been very high, and usually occurs towards the end of the dialysis.

The exact cause is unknown, but is probably related to rapid changes in blood biochemistry and the acid-base balance. In some cases the cause appears to lie in the more rapid clearance of urea and other osmotically active substances from the blood than from the central nervous system, resulting in an osmotic gradient which draws water into the brain and cerebrospinal fluid sufficiently to raise the intracranial pressure. Electroencephalography carried out during dialysis shows changes consistent with such an hypothesis and reversible with hypertonic fructose (Kennedy et al. 1963). Similar disturbance may result from quite another mechanism, when glucose enters the blood from the dialysing solution and provokes reactive hypoglycaemia (Rigg & Bercu 1967).

Another serious complication of dialysis has been traced to the aluminium content of the water used in preparing the dialysate. This, known as ’dialysis encephalopathy’ or ’dialysis dementia’, was formerly one of the commonest causes of death in certain units (Burks et at. 1976). The earliest signs were difficulty with speech or episodes of confusion appearing during the dialysis procedure. Over time the mental difficulties became permanent, with dysarthria, dysphasia, apraxia and slowly progressive generalised dementia. Paranoia, bizarre behaviour and episodes of delirium were sometimes prominent. Neurological accompaniments were flapping tremor, facial grimacing and myoclonus. Focal seizures and various pareses often occurred.

Slow worsening of the disorder usually led to death. Treatment was ineffective and progression could continue despite restoration of normal renal function by transplantation. At autopsy the changes in the brain were slight and non-specific.

Most patients developing the disorder had been on haemodialysis for several years. It was found that the aluminium content of the brain, and especially the grey matter, was increased fourfold in such patients compared with uraemics dying from other causes (Alfrey et al. 1976; McDermott et al. 1978b). This was at first ascribed to the aluminium in the phosphate-binding gels employed, but it eventually emerged that the regional incidence of encephalopathy in various centres correlated with the aluminium content of the water used as dialysate (Parkinson et al. 1979). Since taking steps to circumvent the problem the disease has now virtually disappeared.

In areas where the incidence was high an association could be shown between duration of time on dialysis and certain measures of cognitive impairment, even before the severe disorder was declared (English et al. 1978). The EEG was similarly often markedly abnormal for many months before signs appeared.

Cerebral pathology

Gross lesions in the central nervous system are rare in the absence of marked hypertension. The most constant change is scattered neuronal degeneration, with chromatolysis and vacuolisation of nerve cells. In chronic cases this progresses to pyknosis and areas of cell loss. Olsen (1961) systematically examined the brains of 104 patients dying of uraemia, and found such changes most frequently in the brain stem nuclei, reticular formation and cortex, though variable in location from case to case. He was unable to confirm other features which have traditionally been emphasised, such as focal glial proliferation, areas of demyelination and cerebral oedema.

Changes in the cerebral vessels with thrombosis and infarction are common, but correspond to hypertension and arteriosclerosis rather than to uraemia itself. A general haemorrhagic tendency may be seen when other factors are present to explain it.

In general, therefore, the cerebral pathology shows little that is specific for uraemia, and the picture is dominated to a high degree by the disease processes responsible for the uraemia and the secondary complications that develop.

Electrolyte disturbances

A delicate balance must be maintained in the chemical environment, both intracellularly and extracellularly, to maintain the proper functioning of the central nervous system. In this certain relatively simple components have been identified: the correct acid-base balance must be preserved, the proper gradient of sodium and potassium across the cell membrane, and the correct concentration of calcium and perhaps magnesium ions. These factors may be disturbed separately or together in many disease processes and mental symptoms may follow.

The metabolic dynamics involved in the production of mental symptoms are often complex, since disturbance of one aspect of electrolyte balance can have repercussions upon others. Alterations in cerebral blood flow may follow and complicate the situation further. Nevertheless the correct appreciation of the primary disturbance is of the utmost importance if appropriate treatment is to follow.

Electrolyte disturbance plays a prominent part in certain endocrine disorders and in uraemia as already


described. It complicates respiratory disorders, and can assume great importance postoperatively when the patient is maintained on intravenous fluids for long periods of time. A variety of other causes will be mentioned below where appropriate.

Water depletion

Water depletion can arise from simple unavailability as in shipwreck, or may be seen clinically in the presence of severe weakness from any physical illness, in severe dysphagia and in coma. Intense thirst and dryness of the mouth is coupled with a greyish ill appearance and loss of weight. Signs of dehydration are less obvious than in sodium depletion since the greatest loss is from the intracellular compartment. The plasma sodium, chloride and urea tend to rise. Increasing mental confusion gives way to delirium and coma. The administration of water relieves the situation, but if intravenous fluids are required a 5% glucose solution should be given. Seizures may occur with high sodium levels, or due to brain oedema during too rapid rehydration. ,

The elderly are especially at risk in view of their narrow limits of physiological balance, diminished capacity for renal tubular absorption, and liability to chronic debilitating disease. Jana and Romano-Jana (1973) have described four cases of ’hypernatraemic psychosis’ in which elderly patients were admitted to a psychiatric hospital in a confused and disorientated state. Discovery of a raised serum sodium led to the intravenous administration of hypotonic fluids with rapid restoration of normal mental function. All four had been ambulatory on admission and did not complain of excessive thirst. The cause of the dehydration remained uncertain, but diminished fluid intake may have been occasioned by primary emotional disturbance and cognitive changes.

Water intoxication

Overhydration may result from excessive infusion of 5% glucose solution postoperatively, from the administration of excessive quantities of vasopressin, or if too much water is drunk in the presence of renal failure. An important cause in psychiatric patients can be compulsive water drinking, as described on p. 525. The serum osmolality and sodium are low.

Anorexia, nausea and vomiting are early signs, with marked lassitude and changes of mood. Headache and blurring of vision may occur. Later there is impairment of consciousness, delirium and coma. Muscle cramps and twitches are sometimes seen, and epileptic fits are common.

Sodium depletion

Sodium depletion occurs in tropical climates when common salt is omitted during the process of acclimatisation. Clinically it is seen after severe vomiting and diarrhoea from any cause, in Addison’s disease, in salt-losing nephritis and pyelonephritis, and postoperatively when patients are maintained on intravenous glucose alone. Severe hyponatraemia may follow extensive burns. The syndrome of ’inappropriate antidiuretic hormone secretion’ (SIADH) has many causes, chief among which are carcinoma of the lung, duodenum or pancreas, the sequelae of encephalitis, stroke, brain tumour or trauma, and a number of drugs including hypoglycaemic agents, thiazide diuretics and psychotropic agents as described on p. 526 (Illowsky & Kirch 1988). Hyponatraemia and hypotonicity of the plasma are accompanied by continuing excretion of sodium in the urine.

The classic symptoms and signs of heat exhaustion include weakness, dizziness, pallor, profuse sweating, diminution of urine, rapid pulse and respiration, low blood pressure and cramping pains hi the abdomen and limbs. The onset is usually sudden, and the response to sodium chloride by mouth is dramatic. However, the picture can be misleading when it sets in very gradually. Saphir (1945) described 10 cases which were missed until the serum chloride was estimated. The resemblance to neurosis was striking, and in some patients malingering had been suspected. Prominent mental complaints included irritability, depression without cause and intense anxiety.

When associated with medical diseases the presenting features are usually of lassitude, apathy and weakness. Giddiness and hypotensive faints are common, with muscle cramps and weakness. Anorexia develops with nausea and vomiting. The patient appears dehydrated but thirst is rarely a prominent complaint. Mental confusion ultimately appears with disorientation, delusions and hallucinations. If untreated the condition progresses to coma. The occurrence of fits is an ominous sign with a mortality of over 50% (Riggs 1989).

Treatment consists of giving salt by mouth, or when necessary intravenous saline. The administration of water alone or glucose in water can be dangerous, as the hypotonicity is aggravated further. The danger of over-rapid correction of severe hyponatraemia hi causing central^ pontine myelinosis is described on p. 586.

Potassium depletion

The normal renal mechanisms for the conservation of potassium are relatively inefficient, and accordingly

Page 560 CHAPTER 11

potassium deficiency is liable to develop in any disease associated with chronic starvation or anorexia. The cornmonest cause is again inadequate intake when patients are maintained postoperatively on intravenous fluids. Gastrointestinal causes include hyperemesis gravidarum, chronic diarrhoea due to ulcerative colitis, malabsorption syndromes such as steatorrhoea, or the long-continued use of purgatives. Other causes include Cushing’s disease, the massive diuresis which may arise in diabetes mellitus, the diuretic phase of acute renal insufficiency, or the rarer potassium-losing nephritis and renal tubular acidosis. It may also result from the administration of diuretics, ACTH or adrenal steroids. Excessive transfer of potassium into the cells may occur after myocardial infarction, in familial periodic paralysis, and when diabetic ketosis is treated vigorously with glucose and insulin.

Judge (1968) has reported that potassium deficiency may arise in elderly persons simply because their diet contains inadequate amounts. Depression, apathy, weakness, paranoid ideation and disturbance of sleep rhythm follow. The serum potassium level may sometimes be normal, yet improvement in mood, alertness and activity follows supplementation. Others may have just sufficient potassium in the diet, but a secondary deficiency is produced by an attack of diarrhoea and vomiting, an acute infection or a small cerebrovascular accident. In addition such persons may take a purgative or be given a diuretic which aggravates the deficiency.

The usual presentation in all cases is with lethargy, apathy and depression which can be profound in degree. Anorexia, constipation and abdominal distension are common accompaniments and paralytic ileus may develop. Severe muscle weakness may extend to flaccid paralysis with rhabdomyolysis and myoglobinuria.

Hysteria may be diagnosed, especially since the tendon reflexes are sometimes preserved until very low potassium levels are reached. Moreover, in familial periodic paralysis (p. 721) emotional stress may precipitate an attack. Mitchell and Feldman (1968) report an example of a patient with renal tubular acidosis who was first thought to show conversion hysteria:

A woman of 30 complained of marked weakness of the arms and legs after a fall some hours earlier, and was unable to walk. There was a long previous history of anorexia, occasional vomiting, constipation and muscle spasms, and there had been several similar falls in preceding months. Marital difficulties were prominent and a hysterical conversion reaction was diagnosed. She slowly regained strength and was discharged after 4 days.

One month later she was readmitted with extreme weakness of several hours’ duration, and complained of sleeping excessively and frequent headaches. The serum potassium was low at 2.4 mEq/1 but again spontaneous recovery occurred. Two days later there were similar complaints, the serum potassium was 1.8 mEq/1 and the electrocardiogram showed typical changes. There had been no evidence of impairment of consciousness at any stage.

Apprehension and irritability are sometimes marked and an anxiety state may be simulated. Emotionalh induced hyperventilation may be thought to be responsible for paraesthesiae, vague muscle discomfort and transient visual disturbances. Very occasionally a typical acute organic reaction is seen with disorientation, confusion, impairment of memory and delirium.

The electrocardiogram shows characteristic changes which may alert one to the situation – a small T wave, prolongation of the Q-T interval and depression of the ST segment. A low serum potassium confirms the diagnosis but is not always found in the early stages.

Potassium chloride or bicarbonate should be given orally. Nasogastric or intravenous infusion of potassium chloride may be needed if oral intake is impossible, but this must be undertaken with great care.

Potassium excess

Hyperkalaemia occurs with acute renal failure, in severe crises of Addison’s disease and in diabetic coma. It can be induced by potassium-sparing diuretics and angiotensinconverting enzyme (ACE) inhibitors (Drug & Therapeutics Bulletin 1991b). Cardiac toxicity precludes the appearance of nervous system manifestations in most cases, with bradycardia due to heart block, ventricular dysrhythmia or fibrillation and ultimately ventricular asystole (Riggs 1989). The most frequent neurological manifestation is flaccid muscle weakness akin to that seen with potassium depletion. The dullness and lethargy of patients with Addison’s disease is probably mainly related to sodium depletion.


The clinical manifestations of hypercalcaemia have already been described in the section on hyperparathyroidism (p. 529). Malignancy is another important cause

- carcinoma with secondary bone deposits, multiple myeloma and Hodgkin’s disease. The hypercalcaemia is then often due to a combination of increased bone destruction caused by metastases, and bone resorption occasioned by a variety of humoral factors secreted or induced by the tumour. Of these, parathyroid hormonerelated peptide is particularly important and has been demonstrated with squamous cell carcinoma of the lung and carcinomas of the breast and kidney (Meader & Vella 1992). Other causes include sarcoidosis, hyperthyroidism,


the excessive administration of vitamin D, and the prolonged ingestion of calcium especially when taken as milk with an antacid (’milk-alkali syndrome’). Long-term treatment with lithium may very occasionally produce hypercalcaemia by elevating the level of parathyroid hormone (Christiansen et al. 1976).

Petersen (1968) reported six examples, mostly due to vitamin D intoxication, with clinical pictures similar to those seen with hyperparathyroidism. Thirst, asthenia, depression and tension states were the main manifestations, and three patients showed acute organic psychoses. Weizmanefa/. (1979) found that seven of 12 patients with hypercalcaemia due to malignant disease had prominent psychiatric symptoms. Three showed depression or anxiety, sometimes severe, three developed acute organic reactions and one an acute paranoid psychosis. In all ca’ses the mental symptoms disappeared within 2-6 days of the serum calcium being restored to normal. There is obvious risk of viewing the emotional disorder as a reaction to the basic disease, or of ascribing organic mental symptoms to intracerebral complications.


Hypocalcaemia may result from a deficiency of calcium or vitamin D in the diet, producing rickets in children and osteomalacia in adults. Hypoparathyroidism, chronic steatorrhoea and chronic nephritis are other possible causes. Anticonvulsant therapy leads to hypocalcaemia in a substantial proportion of epileptic patients. Acute severe forms are usually the sequelae of thyroid or parathyroid surgery, or arise as a complication of acute pancreatitis (Riggs 1989).

In children there is a characteristic triad of convulsions, laryngeal stridor and carpopedal spasm (p. 532). In adults the usual complaint is merely of painful cramps or tingling paraesthesiae in the limbs. The characteristic signs of latent tetany and the common mental manifestations are described in the section on hypoparathyroidism (pp. 531-3).

Tetany due to alkalosis is described below.


The clinical pictures attributable to magnesium depletion are ill denned and controversial, largely because magnesium deficiency is usually associated with other concurrent metabolic derangements. It may arise after prolonged parenteral feeding, in chronic alcoholism, in delirium tremens, and in cases of severe malnutrition or malabsorption associated with diarrhoea and vomiting. Renal loss of magnesium occurs in diabetic acidosis and renal tubular necrosis, also with cisplatin use in children (Hayes etal. 1979).

Vallee et al. (1960) found that a picture identical with hypocalcaemic tetany could result, with convulsions, carpopedal spasm, Chvostek’s and Trousseau’s signs, and athetoid movements of the limbs. They described such cases in which the serum calcium was normal and parenteral magnesium sulphate promptly abolished the symptoms. Hanna etal. (1960) presented three cases of pure magnesium deficiency with a rather different picture. Convulsions were present, but tetany in the sense of spontaneous muscle cramps was absent. The most marked manifestations were depression, irritability, vertigo, ataxia and muscle weakness. Fishman (1965) described myoclonic jerks and bizarre multifocal seizures. Other reports have included tremors, fasciculation, choreiform movements, mild confusion and disorientation, delirium of sudden onset and stupor, all of which may be reversed by intramuscular magnesium sulphate (Flink

1956; Hammarsten & Smith 1957; Randall etal. 1959; Hall 1973). Shils (1969) kept volunteers on magnesiumdeficient diets for periods of many months, resulting in lethargy, tremors, fasciculations and spontaneous carpopedal spasms, all responding to the administration of magnesium. However, all symptomatic subjects had developed secondary calcium depletion despite adequate intake and absorption, as well hypokalaemia. The contribution of magnesium deficiency alone could therefore not be determined. More recent evidence on the subject is reviewed by Tucci (1981) and Riggs (1989).


A number of workers have investigated serum magnesium levels in patients with schizophrenia. Elevated magnesium has often been reported, but also lowered levels or no difference from controls. Alexander and Jackson (1981) review the inconsistent findings in the literature.

is rare, occurring when there is excessive intake in a setting of impaired renal function. Depression of the nervous system results in lethargy, confusion and loss of the tendon reflexes, proceeding in severe cases to flaccid paralysis due to blockade of neuromuscular transmission.Low serum zinc

Zinc deficiency is rare, but occurs in certain malnourishe’d populations and in countries where bread with a high phytate content is consumed. It can be found with regional enteritis and with malabsorption syndromes. The excretion of zinc is increased in liver disease, diabetes, some renal diseases and with certain drugs (Lancet 1973; Drug & Therapeutics Bulletin 1988b).

Page 562 CHAPTER 11

The syndrome most closely tied to low serum zinc is diminished acuity of taste and smell (hyposmia, hypogeusia), first described by Henkin et al. (1971) and shown to be responsive to the administration of oral zinc sulphate. Many such patients had developed the hypogeusia soon after a respiratory illness, while in others it appeared spontaneously. The strong perversions of taste and smell that were sometimes present could precipitate emotional disturbance including profound depression.

Henkin et al. (1975) monitored the low serum zinc produced by histidine in the treatment of six patients with progressive systemic sclerosis. They were able to follow the sequential appearance of anorexia, dysfunction of taste and smell, and ultimately the development of neurological and psychiatric features. The patients became dizzy and unsteady, with cerebellar symptoms in the form of ataxic gait and intention tremor. Several were irritable and easily upset, with depression and periods of weeping. Others showed memory impairments, lethargy, auditory and visual hallucinations and pronounced emotional lability. The disturbances correlated with the degree of lowering of serum zinc and with indicators of total body zinc loss. All were quickly reversed following the administration of zinc sulphate.

Staton etal. (1976) described a young male patient presenting with a picture resembling catatonic schizophrenia who responded rapidly to zinc administration after failing to respond to other treatments:

He had presented with auditory and visual hallucinations, loose associations, blunted affect and disorientation in time and place. Drooling, negativism and catatonic postures became established. Phenothiazines were ineffective and led to severe extrapyramidal disturbance. After some months electroconvulsive therapy was tried and produced only transient improvement. A low serum zinc and high serum copper were discovered and oral zinc sulphate and pyridoxine were commenced. Thenceforward he made excellent progress and remained well 1 year later.


Metabolic alkalosis results from repeated vomiting as in pyloric stenosis, or the ingestion of large quantities of sodium bicarbonate given, for example, in the treatment of peptic ulcer. An important cause from the psychiatric point of view is the respiratory alkalosis which results from overbreathing. Attacks of hyperventilation are common in states of anxiety and in histrionic patients under stress. Hyperventilation also follows the ingestion of large quantities of salicylates.

In the cases associated with vomiting there is often apathy, delirium and stupor, but concomitant potassium and water depletion may often play a part. Tetany is liable

to occur, because the proportion of ionised calcium is reduced even though the total serum calcium is normal.

The sequence of changes seen with hyperventilation are reviewed by Wyke (1963). Experimental studies show that hyperventilation increases suggestibility and facilitates the induction of hypnosis. At an early stage perception is increased, but later dulled. As consciousness becomes impaired and awareness of the environment diminishes the EEG begins to slow, and high voltage delta waves ultimately appear. Impairment of memory and calculation develop when the dominant frequency reaches

5 Hz. Psychological studies show impaired performance on tests of reaction time, manual coordination and word association, and there is often subsequent amnesia for events of the period.

When hyperventilation accompanies anxiety the emotional instability is increased, creating a vicious circle. Vertigo and paraesthesiae further reinforce the patient’s concern. Mental confusion may become marked, and myoclonic jerks or epileptic phenomena may be precipitated. In severe cases the condition may progress to loss of consciousness.

It is uncertain whether these clinical phenomena depend directly on the lowering of carbon dioxide tension in the blood, or on the rising pH and other metabolic changes in the environment of the neurones. Hypoxia resulting from vasoconstriction of the cerebral arterioles may make a further contribution. The decrease in the ionised calcium of the blood is almost certainly responsible for the tetanic phenomena.


Metabolic acidosis may result from renal failure or from diabetes mellitus with ketosis. In chronic diarrhoea there may be loss of sodium bicarbonate in the stools. Respiratory disorders such as emphysema or status asthmaticus which lead to underventilation similarly cause acidosis due to carbon dioxide retention.

In metabolic acidosis the most prominent result is stimulation of the respiratory centre with deep and rapid respiration. Consciousness is progressively impaired and mental confusion or delirium are seen in varying degree. The precise clinical picture in the individual case is largely determined by the underlying condition ^nd other associated metabolic derangements.

Respiratory acidosis (hypercapnia) provides a more distinctive clinical picture. Inhalation of 6-7% carbon dioxide can be shown to impair psychological functioning and lead to perseverative responses. In chronic respiratory disease mental dulling and drowsiness are common, and it has long been known that if oxygen is given alone


this can sometimes impair consciousness further and precipitate mental confusion and irrational behaviour. This was at first ascribed to overoxygenation, but is now recognised to result from the increased carbon dioxide retention which occurs as dyspnoea is relieved.

Westlake et al. (1955) reviewed the clinical findings in carbon dioxide retention due to emphysema. Mental disturbances were usually present when the blood pH was below 7.2 or the arterial tension of carbon dioxide above 100 mmHg, ranging from mild impairment of consciousness with irritability, disorientation and confusion, to delirium with auditory and visual hallucinations. Headache, muscle twitching and sweating were common accompaniments. When the pH fell below 7.1 or the tension of carbon dioxide rose above 120 mmHg, there was increasing lethargy and drowsiness leading ultimately to coma. The intracranial pressure was often raised, and papilloedema was sometimes seen. The disturbances were usually transient, because the pH is ultimately restored by renal activity, but in a minority of cases the outcome could be fatal.

The mental changes are thought to be due to the direct action of acidaemia or hypercapnia on the metabolism of cortical neurones. The rise of intracranial pressure is ascribed to the accompanying cerebral vasodilatation. The EEG shows delta waves, sometimes paroxysmal or episodic, at high arterial levels of carbon dioxide.

Hepatic disorder

Striking neurological and psychiatric changes may be seen in patients with liver disease. The range and extent of these features were clearly defined in the classic paper by Summerskill et al. in 1956. An important point was the demonstration that neuropsychiatric disorder could sometimes dominate the picture, even when unequivocal evidence of liver disease was not immediately obvious. The correct diagnosis might therefore easily be overlooked.

Seventeen patients were reported, mostly with portal cirrhosis. Neuropsychiatric symptoms were the presenting feature in eight patients in whom liver disease had not previously been suspected. Four had been admitted to psychiatric hospitals initially, and in three others a psychiatric opinion had been sought. Only three were jaundiced and only seven had large livers. Five showed little biochemical evidence of liver dysfunction. The signs of most value in supporting the diagnosis were palmar erythema, spider naevi, finger clubbing and loss of body hair. Foetor hepaticus was prominent in every patient and the spleen was constantly enlarged.

Since then the topic appears to have received less attention than it deserves, but new interest attaches to it in view of improved methods of treatment of liver disease which allow much longer survival. Collis and Lloyd (1992) provide a valuable recent review.

Psychiatric features of hepatic encephalopathy

The manifestations are extremely variable, making it hard to demarcate a single entity of ’hepatic encephalopathy’. Essentially the picture is of a chronic organic reaction, characterised by acute exacerbations and remissions and accompanied by neurological abnormalities which also wax and wane. Neurotic or psychotic features may be prominent, depending on the premorbid personality, and change of personality may feature in large degree. The disturbances may persist for many years, sometimes with complete remissions for long periods of time, or the condition may become chronic and constant (’acquired chronic hepatocerebral degeneration’) (Victor etal. 1965). The picture then resembles dementia, often with persistent motor abnormalities by way of pyramidal and extrapyramidal disturbance.

Impairment of consciousness is always present during exacerbations of the disorder. Warning signs are a fixed staring appearance and reduction of spontaneous movements. Hypersomnia is an early feature, sometimes with overpowering attacks of sleepiness by day and inversion of sleep rhythm. Later this progresses to periods of marked confusion, semicoma or coma, though deterioration can be arrested at any level. Coma at first resembles normal sleep, but later progresses to total unresponsiveness.

Rapid changes in the level of consciousness are accompanied by delirium with hallucinations, mainly in the visual modality. Episodic ’twilight’ states with sudden onset and ending may also occur. One patient reported by Summerskill et al. (1956) experienced panoramic scenes in bright colours of frightening bears and wolves, and synaesthesia so that a dripping tap in the right peripheral field of vision was experienced as a cold drip on the right cheek. Recent memory is impaired in proportion to the impairment of consciousness, and confabulation is sometimes much in evidence. Dense amnesias with retrograde loss follow periods of coma.

The mood often shows abrupt swings, sometimes with depression, sometimes euphoria. Personality changes may be marked and are sometimes the presenting feature. During exacerbations uninhibited behaviour may be released, or previous traits of irritability or joviality are exaggerated. Between exacerbations an enduring disturbance of personality is characteristically seen, reminiscent of frontal lobe disorder with blunted affect, loss of drive and initiative, incongruous jocularity, tactlessness, defective

Page 564 CHAPTER 11

insight and loss of finer aspects of social judgement. Murphy et al. (1948) noted similar changes and found that ’the jaundiced, pot-bellied cirrhotic patients are usually the jovial clowns of the ward’. Such patients would often insist on going home or claim perfect health despite all evidence to the contrary.

Other features include short-lived depressive episodes, hypomanic reactions, anxiety attacks or obsessivecompulsive behaviour. Occasionally paranoid reactions which have set in during episodes of confusion persist on recovery.

Episodes typical of acute schizophrenia or hypomania have also been reported, sometimes in the absence of impairment of consciousness, but usually accompanied by other signs of hepatic encephalopathy (Read et al.1967). These may develop as little as weeks or months after portal-systemic anastomotic operations. They are rare, however, and their causal relationship to liver dysfunction remains in doubt.

Neurological features

The neurological abnormalities are more specific, and tend to worsen or remit in parallel with the psychiatric symptoms.

Motor disorders are the outstanding feature. There may be little to detect in remission other than mild exaggeration of tendon reflexes, unobtrusive tremor and a characteristic blank or grimacing facial expression. But at some stage almost every patient shows the characteristic flapping tremor (asterixis) which is brought into evidence when the arms are held outstretched. This consists of rapid bursts of flexion-extension movements at the metacarpophalangeal and wrist joints, superimposed upon a fine 6-9-per-second tremor. It is aggravated by fatigue, anxiety and excitement but is absent at rest. The disorder is characteristic but not entirely specific for hepatic encephalopathy, and is occasionally seen in uraemia, respiratory failure or severe heart failure (Sherlock & Dooley 1993).

Deterioration is accompanied by a combination of extrapyramidal and pyramidal features – dysarthria, ataxia, gross tremor, muscular rigidity, hyperreflexia and clonus. The plantar reflexes usually remain flexor until the stage of coma is reached, so that the unusual cornbination of rigidity, clonus and down-going plantar responses is frequently seen. In deep coma the muscles become flaccid and reflexes may be abolished.

These gross neurological disorders can remit abruptly and dramatically from one day to the next along with the psychiatric features. Serial tests of handwriting may serve as an indicator of progress. Other neurological abnormalities include dysphasia with perseverative speech disturbances, blurring of vision, diplopia and nystagmus. Constructional apraxia has been stressed as an early and sometimes persistent feature.

Read et al. (1967) have described some less common syndromes of neurological defect. One small group of patients developed progressive paraplegia, and another cerebellar dysfunction along with parkinsonian features. Occasional patients developed myoclonus or epilepsy, or were at first suspected of cerebral tumours on account of dementia coupled with focal neurological signs.


Biochemical evidence of liver dysfunction will usually be found but may sometimes be relatively slight.

The EEG can be of help in diagnosis and prognosis and shows a close correlation with the grades of neuropsychiatric disturbance (Parsons-Smith et al. 1957; Kennedy et al. 1973). Changes occur early in the progression towards coma, and even before psychological abnormalities have appeared. The earliest change is slowing of the alpha rhythm and the appearance of 5-7-per-second theta waves, most marked in the frontal and temporal regions. Theta comes to replace the alpha activity increasingly as consciousness is progressively impaired. Later, characteristic triphasic waves are seen, the appearance of which suggests a poor prognosis. Further deterioration is shown by decrease in amplitude, blunting of the triphasic waves and periods of flattening.

These changes are not specific for hepatic failure, and occur with many other metabolic derangements uraemia, hypokalaemia, anoxia, carbon dioxide retention, hypoglycaemia, B12 deficiency and in the early phases of raised intracranial pressure – but in the conscious patient with liver disease they are virtually diagnostic of impending coma.

Lumbar puncture reveals cerebrospinal fluid under normal pressure. The protein is sometimes raised during coma but there is no pleocytosis (Sherlock & Dooley

1993). Glutamic acid and also glutamine may be increased.

Brain imaging may show minor degrees of cerebral atrophy, chiefly sulcal widening over anterior brain regions. This is more evident in patients, with chronic persisting encephalopathy than in those with relapsing-remitting episodes (Zeneroli et al. 1987), but it is sometimes seen even in seemingly well-compensated patients (Tarter et al. 1986a; Moore etal. 1989). Such findings have been observed in non-alcoholic cirrhotics, confirming that they are unlikely to be due to alcoholism alone (Tarter et al. 1986a; Bernthal et al. 1987). In some


investigations the degree of atrophy has shown significant relationships to biochemical indices of liver dysfunction and to certain measures of psychological deficit, but such correlations have not been striking.

Psychometric studies have confirmed the presence of deficits in encephalopathic patients with an accent on attentional processes (Elsass et al. 1978). Among cirrhotic patients generally, cognitive performance has been shown to be poor, especially on measures of memory and visuospatial ability (Tarter et al. 1986b, 1988). By contrast verbal abilities are generally well preserved. Again such impairments are found in both alcoholic and nonalcoholic patients, and sometimes with significant relationships to biochemical measures of liver dysfunction. Arria et al. (1990) have shown that psychometric impairments are more prominent in cirrhotic patients with low vitamin E levels, suggesting that impaired absorption of vitamin E and perhaps of other nutrients may make its own contribution to cognitive deficits.

Subclinical hepatic encephalopathy

Efforts have been made to detect patients with ’subdinical’ or ’latent’ hepatic encephalopathy so that appropriate measures may be undertaken to avert its full development. Thus even among patients without EEG abnormalities, and with normal psychiatric or neurological examinations, it has been observed that certain changes may herald the development of encephalopathy later. The sensitivity of visual evoked potentials in this regard was shown by Zeneroli et al. (1984), and the value of the WAIS Block Design Test and the Trail Making Test by Gitlin et al (1986) and Sood et al. (1989). Davies et al (1991) recommend the use of a simple screening battery including the drawing of a five-pointed star, and a ’number connection test’ which is a modified version of the Trail Making Test. Weissenbom et al. (1990) compared the relative sensitivities of EEG patterns, visual evoked responses, P300 latency and the number connection test. The last two appeared to be optimal in the detection of impending encephalopathy.

Differential diagnosis

Summerskill et al.’s (1956) original patients had received various diagnostic labels including neurotic anxiety, hysterical ataxia, depression, frontal lobe tumour, cerebral arteriosclerosis, narcolepsy, psychomotor epilepsy, multiple sclerosis, Wilson’s disease and parkinsonism.

The fluctuations in severity differentiate the condition from dementia, but may closely simulate drug intoxication. The distinction from Wilson’s disease (p. 661) can be important; the neurological disturbance is then nonfluctuating, the motor abnormalities usually consist of choreoathetoid movements rather than a flapping tremor, and a Kayser-Fleischer corneal ring is virtually always present.

Delirious episodes may at first suggest delirium tremens, though hallucinations are less vivid and the patient is rarely fearful. The differentiation from alcoholic intoxication can be difficult because the two may coexist.

Depression may be misdiagnosed in view of the fixed facial expression, psychomotor retardation and depressive swings of mood. Catatonia may be suggested by episodes of mutism and stupor. Moreover psychoses of an affective, schizophrenic or paranoid type may be precipitated at an early stage.

Finally, in the presence of known liver disorder which is under treatment, the disturbances must be distinguished from those due to hyponatraemia caused by a low sodium diet and the administration of diuretics.

Aetiology of mental disturbances

It is now established that the neuropsychiatric disturbances are similar whatever the underlying liver pathology. Hepatocellular failure, portal hypertension or surgically induced portal-systemic anastomosis all lead to essentially similar pictures. The feature they share is a circulatory pathway by which portal blood may enter systemic veins and reach the brain without being metabolised by the liver (Sherlock & Dooley 1993). In primary hepatocellular failure the shunt is through the liver itself, because the cells cannot metabolise the contents of the portal blood completely. In cirrhosis the shunt is via the collateral vessels which become established. The mental disturbance is therefore essentially due to cerebral intoxication by intestinal contents which have failed to be metabolised by the liver.

The precise toxin responsible is uncertain, but it is clearly nitrogenous in nature. Thus the picture can be decisively influenced by the level of protein in the diet. A large intake of protein produces exacerbations and may even precipitate coma, while rigid protein restriction ameliorates the condition. The ammonium ion itself was at one time thought to be directly responsible, but this now appears unlikely.

Other metabolic disturbances may aggravate the situation by making the brain more susceptible to toxic in-V fluences. Acid-base imbalance, electrolyte disturbance, hypotension and anoxia may all contribute in this way. Not uncommonly acute episodes are found to have been precipitated by haemorrhage into the gastrointestinal tract, infection, injudicious sedation or the administration of a potent diuretic.

Page 566 CHAPTER 11


Treatment is a matter for specialist medical supervision and is described by Sherlock and Dooley (1993). The fundamental requirement is rigid restriction of dietary protein, often coupled with oral antibiotics to decrease bacterial ammonia formation in the gut. Enemas and purgation may help. Lactulose and lactitol have proved to be beneficial, likewise the benzodiazepine antagonist flumazenil (Ferenci et al 1989). Branched-chain amino acid supplements are possibly useful in patients who cannot tolerate sufficient protein in the diet to prevent a negative nitrogen balance (Horst et al. 1984), Bromocriptine is employed in patients with long-lasting encephalopathy who prove resistant to other measures.

Liver transplantation now offers hope to patients with end-stage liver disease. Powell etal. (1990) demonstrated its effectiveness in improving cognitive and neurological function in a patient with chronic progressive encephalopathy after many years of portal-systemic shunting:

A 44-year-old woman had developed splenomegaly at 19 years of age and severe bleeding from oesophageal varices 2 yean later. Liver biopsy showed ’post-necrotic scarring’ and portacaval anastomosis was performed. During her thirties she developed intermittent confusion and abnormal behaviour, and had several admissions with ascites, oedema and recurrent chest infections.

When seen at 44 she was unable to perform simple household duties or shop or manage her finances. Her speech, memory and gait had deteriorated progressively over the preceding 2 years. Though orientated, memory and arithmetic were poor, likewise abstracting and visuospatial ability. She was unable to write her name and address without making spelling errors. Neurological examination showed dysarthria, coarse tremor and clasp-knife rigidity in the legs. Her gait was ataxic.

After liver transplantation her cognitive and neurological status was considerably improved at 3 months, with further gains after a year. By then she was fully independent in all aspects of daily living and managed her household and finances normally. Detailed testing showed some residual deficits and dysarthria persisted, but the tremor had resolved and her gait was normal.

Nevertheless transplantation may be attended by psychosocial problems as discussed by Heyink et al. (1990). Of 29 patients seen at an average of 3.5 years after operation,

10 had had serious problems by way of depression and three had made suicide attempts. There had often been difficulties in coming to terms with altered status and function within the family after many years of invalidism, in adjusting to altered appearance occasioned by steroids, and in coping with worries about rejection of the organ. Six patients had difficulties in adjusting to the presence of the new liver within their bodies, feeling it to be strange or having fantasies about the donor. Such problems were generally found to be more intrusive than the physical consequences, though they mostly ameliorated with time.


In the early stages, at least, the mental and neurological features are usually found to be reversible, including quite severe degrees of cognitive impairment. Frequently the therapeutic result is excellent. The prognosis depends principally on the extent of liver cell failure. Acute hepatitis has the worst prognosis, whereas a favourable outcome can be expected in chronic disorders with relatively good liver cell function but extensive collateral circulation. The prognosis is always better if a precipitant such as Infection or haemorrhage can be identified and treated.

Cerebral pathology

In cases coming to autopsy the brain is usually normal macroscopically though it may be oedematous. The most striking histological change is diffuse proliferation and enlargement of astrocytes in the cerebrum, cerebellum, putamen and globus pallidus. The neurones themselves show relatively minor alterations.

Such a picture appears to be virtually specific for liver disease. It can develop within a few days of the onset of encephalopathy and bears an approximate relationship to the duration and severity of neuropsychiatric symptoms. In chronic cases there may be cerebral softening with cortical thinning, and neuronal degeneration is most marked in the deeper layers of the cortex (Victor et al. 1965).

Other psychiatric disorders with liver disease

Surveys have shown that in addition to hepatic encephalopathy, patients with liver disease are vulnerable to a range of other psychiatric disorders. Ewusi-Mensah et al. (1983) examined 71 patients with alcoholic liver disease and 71 with non-alcoholic liver disease, using the life-time version of the Schedule for Affective Disorders and Schizophrenia. Two-thirds of the former and onethird of the latter had, or had a history of, psychiatric illness. Depression, anxiety disorder and antisocial personality were all commoner in the alcoholics. Sarin et al. (1988) in a similar comparison reported psychiatric morbidity in three-quarters of alcoholics with liver disease and a quarter of non-alcoholic cirrhotics. The commonest diagnoses according to Research Diagnostic Criteria were


anxiety neurosis (34%) and depression (27%). Patients with a past or present history of hepatic encephalopathy were excluded from the latter survey. Among the alcoholics no differences could be found in relation to the severity of the liver disorder, those with hepatitis scoring as highly as those with cirrhosis. No schizophrenic psychoses were seen in either series.

Trzepacz et al. (1989) made a prospective study of 247 consecutive candidates for liver transplantation, all with end-stage liver disease. By DSM-HI criteria 17% had delirium, 20% an adjustment disorder (mainly depression), 5% major depression and 1% dementia. Nine per cent showed alcohol abuse or dependence and 2% other drug misuse. All subjects and their families scored highly in terms of stress, this centring largely on uncertainties regarding the impending transplant. Such figures underline the important role for psychiatric services in units which care for patients with liver disorder, and the need for counselling when transplantation is being considered.

Acute porphyria

The classification of the several varieties of porphyria is discussed by Kappas et al. (1989) and McColl et al. (1996). The commonest form in the UK is the Swedish or ’acute intermittent’ type. This is inherited through an autosomal dominant gene on chromosome 11, resulting in a 50% reduction of porphobilinogen deaminase activity. The other major form is the South African type, ’porphyria variegata’, with a similar though genetically distinct basis. The deficiency is in protoporphyrinogen oxidase activity. This can declare itself either by light-sensitive skin lesions of a bullous or erythematous nature, or in a manner similar to the acute intermittent type (see below). When in the latter form this too is occasionally seen in European countries. The third acute form is ’hereditary coproporphyria’ with not dissimilar modes of presentation, due to deficiency in coproporphyrinogen oxidase activity. Other rare varieties of little relevance to psychiatry include the ’symptomatic porphyria’ secondary to severe liver disease, and ’congenital erythropoietic porphyria’ which is associated with skin lesions and haemolytic anaemia. All represent inherited or acquired disorders of the many different enzymes involved in haem biosynthesis.

The historical researches of Macalpine and Hunter (1966, 1969) and Macalpine et al. (1968) have brought special interest to the disorder. Evidence has been presented which suggests that George Hi’s prolonged and puzzling mental illness may have been associated with porphyria, and the disease has been traced in the Royal Houses of Stuart, Hanover and Prussia back to Mary Queen of Scots. Doubt has, however, been cast on these conclusions (Levy 1970). Porphyria has also been proposed as the basis for van Gogh’s illnesses (Loftus & Arnold 1991).

Clinical features of acute intermittent porphyria

The clinical features of the acute intermittent type declare themselves at any age from puberty onwards, but mostly in the third decade. Since the disease often exists in latent form a family history may not be forthcoming. Indeed with the advent of porphobilinogen deaminase determinations it is now apparent that some 90% of persons who inherit the biochemical defect remain clinically unaffected throughout their lives (Kappas et al. 1989). Symptoms are more common in females than males.

Attacks may take many forms which renders diagnosis difficult. Valuable reviews of the clinical features are provided by Stein and Tschudy (1970), Kappas et al. (1989) and McColl et al. (1996). Typical symptoms consist of acute abdominal pain, or pain in the chest, limbs or back, often associated with nausea, vomiting, headache and severe constipation. Epileptic fits occur in some 20% of cases, and status epilepticus may develop. Hypertension is a common accompaniment and may persist between attacks. A rapidly developing and predominantly motor peripheral neuropathy may ensue or can be the presenting feature, with weakness, diminution or loss of tendon reflexes, numbness, paraesthesiae or pain in the limbs. This can progress to severe paralysis with embarrassment of respiration. After an attack weakness or wasting may occasionally persist as an enduring disability.

Mental disorder accompanies attacks in 25-75% of cases, and psychiatric symptoms can dominate the picture. The patient becomes emotionally disturbed, sometimes acutely depressed or anxious and sometimes restless and violent. Marked emotional lability is common with histrionic, demonstrative behaviour. Clouding of consciousness and confusion may progress to delirium, with hallucinations, delusions and noisy disturbed behaviour. Coma sometimes develops abruptly. Psychotic developments may resemble schizophrenia and paranoid reactions are not uncommon.

Precipitation of attacks

The majority of attacks are precipitated by identifiable factors though some appear to arise spontaneously. The main precipitants include inadequate nutrition, endocrine factors, alcohol excess, intercurrent illnesses and infections, and a wide range of drugs. The importance of nutrition is shown by attacks which

Page 568 CHAPTER 11

occur while on reducing diets, especially when these lead to precipitous loss of weight. Some unusually sensitive persons may experience attacks after missing several meals (Kappas et al 1989). Endocrine influences are shown in the increased frequency of attacks at the time of the menses and their decline after the menopause. Synthetic oestrogens and progesterone are known to induce attacks. Among drugs barbiturates are notorious for the grave risk of aggravating the disorder if given in an attack to sedate the patient. Thiopentone given in the course of laparotomy may endanger life by precipitating paralysis and respiratory failure. Other drugs which may precipitate or worsen attacks include anticonvulsants (carbamazepine, phenytoin, primidone, sodium valproate and succinimides), nortriptyline, tranylcypromine, diazepam, nitrazepam, meprobamate, sulphonamides, erythromycin, ergot preparations, phenylbutazone, metoclopramide, tolbutamide, clonidine, methyldopa, chloroquine and the contraceptive pill. Those judged safe include aspirin, narcotic analgesics, penicillin, streptomycin, tetracycline, propranolol, chlorpromazine, paraldehyde, and probably chlormethiazole (Moore & Disler 1988; Kappas et al. 1989; McColl et al. 1996).

There has been much discussion of the role played by emotional disturbance in precipitating attacks. This was reviewed by Ackner et al. (1962) who found the evidence to be inconclusive. In their group of 12 patients, only three of the many episodes occurred at times of acute or chronic emotional stress. The patients and their relatives did not consider stress to be causally related, and detailed enquiry revealed many stresses in the past which had failed to provoke attacks.

Premorbid stability

Similar controversy surrounds the background of emotional instability which has been said to characterise porphyric patients. It has been claimed, for instance, that porphyria occurs specifically if not exclusively in patients with severe neurosis or personality disorder, and that a family history of psychiatric disorder is very common (Roth 1945).

Ackner et al (1962) failed to confirm these features. The previous mental health and stability of their patients was unremarkable between attacks, and only one had a family history of mental illness. There was no strong evidence of abnormal personalities among the relatives. Ackner et al. suggest that the evidence for personality disorder is often confused with evidence of episodic disturbed behaviour during attacks, and that the latter is liable to add to a general impression of emotional instability. In fact it was

surprising how little effect the disabling and capricious disease had had upon most of the patients in their series, despite frequent hospitalisations with mistaken labels of hysteria and personality disorder.


The diagnosis is confirmed by the detection of excess porphobilinogen and rf-aminolaevulinic acid in the urine and possibly also in the serum, and the demonstration of reduced porphobilinogen deaminase activity in erythrocytes. Quantitative tests are preferable to qualitative screening tests on urine samples, which should be collected without acid and refrigerated to protect them from light. Porphobilinogen is itself colourless, but may change to a red uroporphyrin after standing, acidification or heating. Freshly voided urine may therefore be colourless, or pale pink varying to deep mahogany depending on conditions at the time.

The absence of raised porphobilinogen and daminolaevulinic acid in the urine does not exclude the diagnosis, since some patients fail to hyperexcrete; conversely low porphobilinogen deaminase activity does not entirely confirm that the attack was a manifestation of the disease because many gene carriers are asymptomatic. The correct diagnosis thus depends on both sets of estimations coupled with knowledge of the clinical history. Caution is also required on account of the false positive results that can occur in the urine with certain febrile illnesses, in lead poisoning and in patients receiving phenothiazine drugs (Reio & Wetterberg 1969).

In the South African variety and in coproporphyria, examination of the stools is more reliable than the urine, excess porphyrins producing a brilliant pink fluorescence in extracts under ultraviolet light.

In the acute intermittent variety there is a tendency for the porphyrin metabolites to reach a peak during exacerbations of the disorder and to decrease to low or negligible levels between attacks. This, however, is by no means invariable and persistently elevated levels may occur. Ackner et al. (1961, 1962) failed to find a clear relationship between the levels of porphobilinogen and daminolaevulinic acid in the urine and the presence or absence of symptoms. The excretion of the two cornpounds fluctuated considerably over time, sometimes with quite high levels in symptomless cases and sometimes with no marked increase in output during mild attacks.

The widespread belief that there is a necessary association between increased excretion and acute attacks therefore appears to be erroneous. It may have arisen because


the chance production of coloured urine is quite often the first pointer to the diagnosis in patients who have suffered from unexplained symptoms for a considerable time. In any case a coloured urine should not necessarily be expected in acute attacks, since the colour depends on physical characteristics of the urine and does not reflect the levels of excretion.

The EEG often shows abnormalities during attacks, with slowing of dominant frequencies and an excess of intermediate slow activity. Sometimes, however, it remains entirely normal. Isolated records are therefore of little help in the diagnosis, but serial recordings can occasionally be useful in confirming the organic origin of symptoms in attacks of uncertain nature.

Differential diagnosis

Porphyria is notorious for leading to mistakes in diagnosis, and patients are sometimes admitted repeatedly to psychiatric units before the condition is discovered.

An impression of psychiatric illness is reinforced by the patient’s emotional instability during attacks, and the long history of intermittent physical complaints for which no cause has previously emerged. Diagnoses of personality disorder or severe neurosis are commonly made. Hysteria may be suspected when the patient complains of weakness of the limbs and varied aches and pains unbacked by physical signs. Psychotic developments may likewise obscure other aspects of the disorder and lead to a primary diagnosis of depressive illness or acute schizophrenia.

Other patients are admitted to general medical wards with suspected appendicitis on account of abdominal pain and vomiting. Intestinal obstruction may be diagnosed when there is severe constipation. Laparotomy therefore features in the history of many porphyric patients. Other cases may be mistaken for the acute ascending polyneuropathy of Guillain-Barre, or the combination of fits and hypertension may suggest hypertensive encephalopathy.

Treatment and outcome

The chief aim must be prevention of attacks by educating the patient about adequate nutritional intake and the drugs which must be avoided. Abstention from alcohol should be advised. Any infection which arises must be treated immediately. On occasion the onset of an attack can be aborted by increasing carbohydrate intake.

The symptomatic treatment of the acute attack is described by Kappas et al. (1989) and McColl et al. (1996). Vomiting may be controlled with phenothiazines, and pain with dihydrocodeine or if necessary pethidine or morphine. The drugs of choice for emotional disturbance are chlorpromazine, promazine or trifluoperazine. Epileptic seizures are perhaps best treated with chlormethiazole, though most will settle spontaneously as the attack resolves. Barbiturates must be avoided under all circumstances.

Intravenous haematin is now used increasingly and appears to be effective in curtailing attacks. Adequate intake of carbohydrate should always be ensured, if necessary by giving dextrose intravenously. Electrolyte imbalance may need attention, especially lowering of the serum sodium in the face of persistent vomiting. It is known that inappropriate secretion of antidiuretic hormone is liable to accompany attacks. A careful watch must also be kept for threatened respiratory embarrassment. Tracheotomy and assisted respiration may occasionally be needed in the more severe attacks.

The duration of attacks varies widely from a few days to several months. The majority subside completely without enduring defects. Some patients, however, are left severely crippled with weakness or muscular wasting, and some remain psychotic for long periods of time. Ultimately there is usually full physical and mental recovery.

About kraeplinpsychiatry

I am a practising psychiatrist. I have strong liking for both biological and psychodynamics aspects of psychiatry. This blog is made to collect my thoughts , sort of self diary where i can use input of others to come to better conclusions. Thanks.
This entry was posted in Lishman. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s