Madness Explained
Page 11
More importantly, there was considerable evidence that the quadruplets had suffered unfortunate experiences during childhood, which may have contributed to their difficulties as adults. Their father, who often drank excessively, was described as unstable and paranoid. It is likely that he sexually abused some of his daughters. The investigators report that ‘He chose Nora as his favourite, at times fondling her breasts and being intrusive when she was in the bathroom.’ A less than ideal family environment can also be inferred from the following observations:
Iris and Hester engaged in mutual masturbation and the parents, horrified, agreed with an attending physician to have both girls circumcised and their hands tied to their beds for thirty nights. Nora and Myra were not allowed to visit their sisters and ‘couldn’t understand the whole situation’. Three of the girls completed high school; Hester did not. Her parents kept her at home in her senior year and she cried a great deal.29
Rosenthal and his colleagues never seriously considered that these kinds of experiences might have contributed to the women’s problems. Nor did they consider the impact of the study on the women’s lives. Visiting them at home for several days every six months or so, the researchers would take the three co-operating Genains out to dinners and lunches, and talk with them about their current circumstances, activities and problems, creating a powerful incentive for the women to describe their experiences in a way that was consistent with the researchers’ expectations.
In the light of the limitations of much of the genetic research, it is impossible at present to reach definitive conclusions about the contribution of heredity to madness. However, even the hardiest sceptic must concede it unlikely that genes play no role whatsoever. In a review of twin study evidence, American psychiatrist E. F. Torrey estimated the pairwise concordance rate for monozygotic twins to be 28 per cent in the case of schizophrenia, and 6 per cent for fraternal twins. For bipolar disorder it was estimated at 56 per cent for identical twins and 14 per cent for fraternal twins.30 Although these figures indicate that genes play a more substantial role in bipolar disorder than in schizophrenia, they should not lead us to suppose that there is a simple causal arrow pointing from genes to mental illness. Assuming that these figures are correct, about half of those who inherit genes for bipolar disorder do not become ill, and only one in seven of those closely related to an affected person are also affected. On similar reasoning, only about a quarter of those who inherit genes for schizophrenia actually become schizophrenic, and the risk to close relatives of schizophrenia patients is only one in seventeen. These figures therefore provide very strong evidence that non-genetic factors influence the development of mental illness.
The implications of genetic research for Kraepelin’s paradigm have been examined by British psychiatrist Tim Crow31 and by American psychiatrist Michael Taylor.32 Although the findings from family studies have not always been consistent, Crow and Taylor were able to point to studies in which the observed risk of a diagnosis of affective disorder in the relatives of schizophrenia patients was greater than chance,33 or in which there was an increased risk of a diagnosis of schizophrenia in the relatives of bipolar patients.34
Crow has also noted that many of the studies which appear to show that schizophrenia and manic depression breed true have been compromised by what he has termed ‘the fallacy of the excluded middle’ – researchers’ tendency to exclude or ignore individuals whose symptoms do not fit the classic pictures of schizophrenia and bipolar disorder. In studies that have included patients with mixed states, their close relatives have been found to be at increased risk for both schizophrenia and affective disorders. Angst and Scharfetter, for example, have analysed data from the families of patients suffering from unipolar depression, bipolar disorder, schizoaffective illness with predominantly affective symptoms, schizoaffective illness with predominantly schizophrenic symptoms, and ‘pure’ schizophrenia. The proportion of first-degree relatives who were found to suffer from affective disorders or schizophrenia varied systematically across this spectrum. The more ‘schizophrenic’ the patient, the more likely it was that their relatives would suffer from schizophrenia. The more ‘mood disordered’ the patient, the more likely it was that their relatives would suffer from mood disorder (see Table 4.1).35 Crow has interpreted these findings as suggesting that schizophrenia and bipolar disorder lie at two ends of a spectrum of psychosis.
Table 4.1 Data from Angst and Scharfetter (1990) showing the risk that first-degree relatives of psychotic patients will meet criteria for schizophrenia or affective disorder (from T. Crow, ‘The failure of the binary concept and the psychosis gene’, in A. Kerr and H. McClelland (eds.) (1991) Concepts of Mental Disorder: A Continuing Debate. London: Gaskell).
Patients are grouped into five diagnostic categories reflecting the extent to which they suffered from mainly affective symptoms, mainly schizophrenia symptoms, or a mixture of symptoms. The last row shows the ratio of relatives who meet the criteria for schizophrenia vs affective disorder. Ratios less than 1 indicate that most ill relatives meet the criteria for affective disorder whereas ratios greater than 1 indicate that most meet the criteria for schizophrenia.
Diagnosis of patients
Unipolar depression
Bipolar disorder
Schizoaffective (predominantly affective)
Schizoaffective (predominantly schizophrenic)
Schizophrenia
No. of patients
58
31
34
35
105
No. of relatives with schizophrenia
10
4
19
18
31
No. of relatives with affective disorder
24
6
14
4
5
Ratio of schizophrenic to affective relatives (age-corrected)
0.30
0.47
0.92
2.99
5.05
Twin studies have also undermined the Kraepelinian distinction between schizophrenia and manic depression. Several cases of twins have been reported in which one appears to suffer from schizophrenia and the other seems to suffer from an affective illness.36 In a study of schizophrenia conducted by British psychiatrists Ann Farmer and Peter McGuffin in collaboration with US psychologist Irvine Gottesman, seven out of twenty-four identical twin pairs had one member diagnosed as suffering from schizophrenia and the other diagnosed as suffering from a mood disorder according to DSM-III criteria. Of the remaining seventeen pairs, six were concordant for schizophrenia whereas, for the other eleven, only one member of each pair suffered from schizophrenia. Farmer and her colleagues attribute these findings to deficiencies of the DSM-III definition of schizophrenia, but they obviously point to a close relationship between schizophrenia and affective disorder.37
Better than Astrology?
From the point of view of the clinician, one of the most important functions of a diagnosis is prediction. When a doctor interprets your chest pain as evidence of angina, you know that she is predicting a different future for you than if she had interpreted your pain as evidence of indigestion. Similarly, if Kraepelin was correct, a diagnosis of schizophrenia should predict a more unremitting course of illness than a diagnosis of manic depression.
In practice, this claim has proved difficult to test. One reason for this is that the course and outcomes of psychiatric disorders are hard to quantify. American psychiatrists John Strauss and William Carpenter have noted that outcome can be measured across several domains that are only loosely correlated. Whereas clinical outcome is determined by assessing the persistence of symptoms, occupational outcome is defined according to how well the individual maintains a steady job, and social outcome reflects the individual’s ability to maintain an adequate network of social relationships.38
In their studies of schizophrenia patients, Strauss and Carpenter came across patients whose f
unctioning could not be classified as simply good or poor, but varied across these three domains. For example, one woman they interviewed was very delusional but was successfully taking care of her child and holding down a job. Her clinical outcome was therefore poor, but her occupational and social outcomes were excellent. Another, whose symptoms had almost disappeared, had not worked for two years, had lost touch with most of her friends, and spent most of her days sitting in a darkened room.39 Similar dissociations between different kinds of outcomes have been observed in bipolar patients.40
Despite these difficulties, three consistent findings have emerged from outcome studies. The first is that the course of psychosis is very unpredictable. For example, the time between the mood episodes experienced by bipolar patients may be as short as several weeks or as long as many years.41 Some patients experience episodes of depression followed by mania followed by normal functioning, others experience mania followed by depression and eventually remission, whereas the majority follow no particular discernible sequence of episodes.42 However, underneath this variability a pattern can be perceived. Towards the end of his life, Kraepelin’s own data led him to believe that the average time between bipolar episodes decreased as the number of episodes increased. This finding has been replicated by more recent investigators (see Figure 4.1), leading some biological psychiatrists to suppose that each episode leaves the brain more sensitive to future episodes, a process sometimes described as ‘kindling’.43
Similar findings have been obtained in studies of schizophrenia patients, who have sometimes been followed up for decades. For example, in a classic study, Luc Ciompi, a Swiss psychiatrist, observed the fate of a group of patients followed up for more than thirty years. The majority spent less than one tenth of this time as inpatients but a quarter spent more than twenty years in hospital. Some patients suffered a sudden breakdown followed by complete recovery. Others had many episodes of illness that began abruptly. Still others experienced a slow onset of illness and, if periods of remission occurred, recovery would be only partial. Some remained ill continuously.44 A similar study carried out by Manfred Bleuler, the son of Eugen, yielded almost identical results.45 Reflecting on the heterogeneity of the outcomes they observed, both Ciompi and the younger Bleuler were moved to reject the idea that schizophrenia is a simple brain disease. This conclusion is particularly striking in the case of Bleuler, given his relationship to the man who coined the term ‘schizophrenia’.
Figure 4.1 Average cycle length in bipolar (BP) patients from four studies (from F. K. Goodwin and K. R. Jamison (1990) Manic-Depressive Illness. Oxford: Oxford University Press). Note that Kraepelin’s data included unipolar (UP) patients.
A second conclusion that can be drawn from research data on the course of psychosis is that outcome is enormously variable between individuals with the same diagnosis. For example, although Kraepelin held that schizophrenia patients inevitably remain ill for the majority of their lives, the long-term studies of Luc Ciompi and Manfred Bleuler both revealed that this was the case for only a minority of patients (see Figure 4.2). About a third of patients completely recovered over the long term, the remaining patients having intermediate outcomes.
Extreme variability has also been observed in the outcome of patients diagnosed as suffering from bipolar disorder. Kraepelin, it will be recalled, believed that manic depression was a relatively benign illness. Indeed, he estimated that the majority of his manic-depressive patients (many of whom would be diagnosed as suffering from unipolar depression by modern standards) experienced only one episode. Modern studies, by comparison, indicate that a much broader range of outcomes is experienced by bipolar patients.46 For example, in a thirty-five-year follow-up of patients who had been diagnosed as suffering from mania on first admission to hospital, American psychiatrist Michael Tsuang and his colleagues found that 64 per cent had recovered completely, whereas 22 per cent remained seriously ill.47 This outcome was better, on average, than the outcome for a group of patients diagnosed as suffering from schizophrenia, but not strikingly so.
Figure 4.2 Bleuler and Ciompi data on long-term outcome of schizophrenia (redrawn from J. Zubin, J. Magaziner and S. R. Steinhauer (1983) ‘The metamorphosis of Schizophrenia’, Psychological Medicine, 13: 551–71).
Tsuang’s study leads to the third general conclusion that can be drawn from the data on the long-term course of psychiatric disorders. Despite the unpredictable course and extremely variable outcome of psychotic illnesses, patients who receive a diagnosis of manic depression have a better outcome on average than patients who receive a diagnosis of schizophrenia. This finding has been reported in many studies, including a five-year follow-up of the first-admission patients recruited to the World Health Organization’s International Pilot Study of Schizophrenia.48
At first sight, this difference seems to be strong evidence in favour of Kraepelin’s distinction between the two main categories of psychosis. However, as British psychiatrist Tim Crow has pointed out, this conclusion is based on the error of reasoning that he has described as ‘the fallacy of the excluded middle’, namely the tendency to focus on ‘textbook’ patients who appear to be clearly schizophrenic or clearly manic-depressive, while ignoring the large number of patients who have both types of symptoms.49
The effects of controlling for this fallacy were explored in an important study conducted by Robert Kendell and Ian Brockington.50 They argued that, if schizophrenia and manic depression are separate entities, we should see a clear ‘gap’ or discontinuity between the outcomes associated with one disorder and the outcomes associated with the other. On the other hand, if schizophrenia and manic depression are variants of the same disorder, outcome might well be affected by the exact mix of symptoms experienced by the patient. On this latter view, schizoaffective patients should have an average outcome that lies somewhere in between those observed for patients suffering from ‘pure’ schizophrenia and those suffering from bipolar disorder.
Kendell and Brockington studied the progress of a large number of patients who had been classified into ten groups along a continuum of symptomatology ranging from mostly manic-depressive, through mixed, to mostly schizophrenic. Outcome was measured after two years using six different methods, and the findings are shown in Figure 4.3. There is no evidence of a sharp discontinuity of outcomes according to any of the measures. On the contrary, the average outcomes seem to form a gradient that depends on the extent to which patients are suffering from schizophrenia or affective symptoms.
At the beginning of this chapter I suggested that astrological predictions provide a fool’s-gold standard against which to evaluate the predictions achieved by psychiatric diagnoses. We are now in a position to apply this standard. While diagnoses clearly are superior to star signs, this superiority is not striking and is only evident when large groups are studied. When the focus is on the individual, the clinician
Patients are distributed into ten groups along the horizontal axes according to the extent to which they suffered mainly schizophrenia symptoms, a mixture of schizophrenia and affective symptoms, or mainly affective symptoms. There are six different measures of outcome on the vertical axes: overall pattern of illness (ranging from recovery to deterioration); occupational record (based on work records); social outcome (based on the quality of social relations); two measures of symptoms (one grading the symptoms on a spectrum from schizophrenia to affective symptoms, the other indicating the overall severity of schizophrenia symptoms alone) and finally a composite or global measure. Note that worse outcomes are associated with schizophrenia symptoms. However, note also that there is no clear evidence of separate groups of schizophrenia and mood-disorder patients.
Figure 4.3 Relationships between symptoms and outcome in patients with psychosis, from R. E. Kendell and I. F. Brockington (1980) ‘The identification of disease entities and the relationship between schizophrenic and affective psychoses’, British Journal of Psychiatry, 137: 324–31.
wanting to predict what
will happen to her patients in the years ahead would do almost as well by resorting to horoscopes.
The Specificity of Psychiatric Drugs
Diagnoses should allow the clinician to predict not only the course and long-term outcome of a disorder, but also what kind of treatment is likely to be effective. For example, different drugs should be effective for different conditions. Oddly, with the exception of one study, this important clinical property of psychiatric diagnoses has not been systematically investigated. This omission is remarkable, as drugs have been the mainstay of psychiatric treatment for at least fifty years, and were widely used even in Kraepelin’s time.
Today, four main types of medication are available to psychiatrists. The less severe psychiatric disorders are usually treated with anxiolytics (the benzodiazapines such as Valium and Librium, normally used to treat anxiety but sometimes useful in the treatment of mild depression) or anti-depressants (mainly tricyclic drugs such as imipramine and the new selective serotonin re-uptake inhibitors such as Prozac, usually employed in the treatment of depression but also effective in the treatment of anxiety). Although the development of these drugs is certainly of historical interest, I will say no more about them in this book.51Anti-psychotic (also called neuroleptic) medications (for example, phenothiazines such as chlorpromazine – also known as Largactil*) are the most commonly prescribed drugs for schizophrenia patients. Patients with a diagnosis of bipolar disorder, on the other hand, are usually offered various mood-stabilizing medications (for example, lithium carbonate or carbamazapine).