Schizophrenia and Suicide
Humphry Osmond, M.R.C.P., F.R.C.Psych.,
 and Abram Hoffer, M.D., Ph.D.2
Most experienced clinicians know that suicide is a danger in schizophrenia. About 30 years ago, one of the wisest, Professor Nolan D. C. Lewis (1927),.wrote, "among the frank mental disorder groups, apparently suicide occurs more often among dementia praecox patients than in any other types. The reaction usually happening during the earlier stages of the conflict before regression has proceeded far enough to attenuate the reality principle to any extent." Some years later he noted (1933) that while the danger of suicide in depressive states now seemed to be widely understood, far less attention had been paid to its occurrence in dementia  praecox. A study of textbooks supports this contention for they have remarkably little to say on this topic. Indeed, generally speaking, from the late 19th century on, they have dealt briefly with the whole question of suicide which does not even appear in the index of one widely read  book purporting to deal with the day-to-day work of psychiatry. Lewis (1889), Kraft Ebing (1904), Kraepelin (1904), Jelliffe and White (1919), Bleuler (1924, 1950),
Bryce State Hospital, Station 3, Tuscaloosa, Alabama 33401.
 -3A-2727 Quadra Street, Victoria, B.C. V8T 4E5. This paper is reprinted from the Journal of Schizophrenia, Vol. 1, No. 1, 1967, with some corrections and additions. Dr. Osmond at the time of writing was at the N.J. Neuro-Psychiatric Institute, Princeton, New Jersey, and Or. Hoffer was at University Hospital, Saskatoon, Saskatchewan. G.R.S.G. Supported in part by funds from Grant No. 1-S01 -05558-01 granted by United States Public Health Service.
Muncie (1939), OIkon (1945), Hall (1949), Henderson and Gillespie (1951), Noyes (1954), Strecker et al. (1951), and even Jaspers (1963), although paying attention to suicide in the affective psychoses, say little or nothing about it in schizophrenia. Skot-towe (1964) who gives an extremely clear, sensible, detailed, and highly  perceptive account of handling suicidal tendencies in depressions, does not mention this as a danger in schizophrenia, nor does he suggest that it is a likely or even probable outcome of this illness. Menninger (1951) has a long and gruesome section on murder in schizophrenia. He also describes at least two suicides which sound as if their victims were schizophrenic, but states: "Not all suicides are melancholic, although most of them are. There are suicides with inferiority complexes, with sexual abnormalities and  psychopathies, with various types of brain disease, such as paresis, and most important of all, some are apparently normal people." Even the compendious Bellak (1948, 1958), usually so informative on all matters concerning schizophrenia, does not mention suicide in his earlier survey. There are only three references to it in his later work, none suggesting that it is a frequent cause of death. The huge American Handbook of Psychiatry (1959) has 49 references to suicide, but does not discuss the frequency of its occurrence in this or any other psychiatric illness. Mayer Gross et al. (1955) 57
do indeed state that schizophrenia often leads to suicide, but they produce no conclusive or even suggestive evidence to support this opinion. Schneidman and Farberow (1957) have many suggestions for preventing suicide, but give no figures indicating that there are greater risks in some illnesses than in others. Ayd (1962) in a characteristically practical and useful pamphlet does not mention schizophrenia as a cause of suicide. Stengel (1963), writing from a very different viewpoint, discusses alcoholism, the effects of physical illness, preservation of the family, the need for early diagnosis in depression, and measurements against social isolation, but again says nothing about schizophrenia. It is safe to say that none of these authorities has progressed beyond that first reference from a paper written more than 35 years ago. According to the official figures, at least 20,000 suicides occur annually in the United States, about 5,000 in Britain, and 2,000 in Canada. We do not know what proportion of all those who kill themselves are represented by these statistics. Many coroners and juries prefer some less definite verdict if this can be given without obviously violating the evidence. Ayd (1962) states that more people die from suicide than from tuberculosis, diabetes, nephritis, nephrosis, and poliomyelitis combined. Coe (1963) observes that in Minnesota suicide is the third highest cause of death among 15- to 19-year-olds, exceeded only by accident and cancer. It would clearly be very useful to know which illnesses are prone to end in successful suicide and at what stage of a particular illness this is likely to occur.
Suicide Among Schizophrenics
 During the last 10 years we have followed the fortunes of two groups of recently diagnosed schizophrenic patients, one of which had been treated with massive nicotinic acid (Hoffer, 1963) and the other which did not receive this vitamin. This latter group of 450 patients were observed for seven years on average, and during that time, nine of them committed suicide. These  patients had been diagnosed by competent  psychiatrists who used the rather conservative criteria of Bleuler (1950). A rough calculation shows that the annual suicide rate for these  patients was about 280 per 100,000. The general suicide rate in Saskatchewan at this time was about nine per 100,000. Automobile accidents killed 17 per 100,000, cancer 122 per 100,000, and heart disease about 250 per 100,000. If the suicide rate had been as frequent among the general population as among these schizophrenic  patients and ex-patients, about 2,600 people in Saskatchewan would have taken their lives annually, but in fact, only about 70 died in this manner. If Saskatchewan has the usual  proportion of people suffering from schizophrenia, that is, about 1 percent, or  probably slightly more, and if they commit suicide at the same rate as our patients, then sufferers from schizophrenia would account for 25 to 30 suicides yearly, about
 of the total. For the United States, this would suggest that about 6,000 schizophrenics kill themselves a year, many of them young people on the threshold of adult life. The figures for Britain and Canada would be about 1,800 and 700 respectively. Is there any other evidence to support or refute such a grim conclusion?
Suicide in Mental Hospitals
 Levy and Southcombe (1953) found that in their hospital 38 suicides occurred per 10,000 admissions. Almost one-half of these were during the first three months in hospital. Exactly half of all these deaths were schizophrenic, two-thirds of whom were under 44 years old. Manic-depressive illness accounted for only one-fifth of suicides, and these were nearly all in patients of the age 50 or over. Of those schizophrenics who committed suicide, five-sevenths were diagnosed as being paranoid and only one-seventh as catatonic. This suggests that better organized  patients who are more likely to be socially viable were also more likely to have the skill, energy, and determination to end their lives. While this study tells us nothing about patients who are out of hospital, it indicates that those who have a  better chance of leaving are also more likely to kill themselves. Banen (1954) reported on 23 suicides of patients, either in or on leave 58
from a V.A hospital. Of these, 18 were diagnosed as schizophrenic and five as suffering from manic-depressive psychoses. The manic-depressive patients were aged 30 to 55. Nine, exactly half of the schizophrenic patients, were under 30, four being under 25, and only two over 40.  Norris (1959) discussed schizophrenic patients with a mean of 3
 years in hospital. In 714 males, there were five suicides, and in 766 women, only one suicide, and she calculated that the male suicides were 17.4 times as frequent as in the population of London as a whole and the females 5.4 times as frequent. These three papers strongly support the view that suicide is a grave danger in schizophrenic patients while in hospital, but what about those who are not in hospital? These deaths might conceivably be due to bad conditions in the hospitals themselves.
Follow-up Studies
 Romano and Ebaugh (1938) followed up 600 newly admitted patients of the Denver  psychopathic ward from January, 1933, to December, 1936. All of these had been diagnosed as schizophrenic. They lost well over one-quarter of their sample, but still found that eight, four men and four women, had committed suicide. Rennie (1939) discussed 500 schizophrenic  patients who were first admissions to the Phipps Clinic. It is not easy to be sure how long these  patients were followed up, for his reference of "from one to 26 years" is obscure and imprecise; 170 patients were lost to follow up, 150 patients never left hospital (they did not, of course, remain in the Phipps), 100 died—27 from tuberculosis, seven men and four women committed suicide. Rupp and Fletcher (1940) followed 641 newly admitted schizophrenics for from
to 10 years. At the end of their study, 14  percent of these patients were dead; pulmonary tuberculosis came first, with 43 deaths, suicide coming second with 10, five males and five females. Clark and Mallett (1963) made an admirably detailed study of 76 schizophrenic  patients, whose average age was 22, and compared them with 74 slightly older depressed  patients, carefully selected to avoid schizophrenic features. During the three years after their first admission three of the schizophrenics committed suicide, and one drowned in peculiar circumstances. Of these three, two were men and one a woman. None of the depressives killed themselves. It is curious that these authors did not find this discrepancy  between the number of suicides in schizophrenic and affective illnesses of sufficient importance to mention it in the text. Gurel (1963) was kind enough to put at our disposal a study-which he has made for the Veterans Administration of newly admitted and newly readmitted male schizophrenics; 1,254 of these were followed in the community for about four years; during that time, 21 committed suicide. Of 65 other functional psychoses, two committed suicide. By combining these figures, including those from Saskatchewan (see Table 1), we have 3,518 schizophrenics whose mean follow-up time was at the very most eight years; 62 of these patients committed suicide. In other words, one in 56 killed themselves during an eight-year period or less, or in every year one out of 450 died in this way. This is very close to the Saskatchewan figures, being about 220 per 100,000, or something in excess of 20 times the normal suicide rate of the countries concerned. Although a variety of actuarial corrections for age, sex, etc., should be made, we can safely say that this is far higher than the usual suicide rate.
Suicide in Schizophrenia Compared With the Affective Psychoses
 Suicide in schizophrenia seems to be at least as frequent as in the affective psychoses, especially in younger people. Because there are few long-term follow-up studies of patients with affective  psychoses whose age is the same as those with schizophrenia, our findings are suggestive, but not conclusive. But they do, however, indicate that the present emphasis upon the affective  psychoses as the main psychiatric illness associated with suicidal risk is misplaced and ought to be changed. 59
Follow-Up Studies of Schizophrenic Suicides
Number of 
Length of 
Number of 
Number of 
Romano and Ebaugh, 1938
4-yrs. Max.
Rennie, 1939
1-26 yrs.
Ruppand Fletcher, 1940
4-14-10 yrs.
Osmond and Hoffer, 1962
9-10 yrs.
Clark and Mallett, 1963
3 yrs.
Gurel, 1963
4 yrs.
Theory of Suicide in Schizophrenia
 Schizophrenia strikes hardest in late adolescence and early adulthood, and it seems likely that as Mayer Cross et al. (1964) noted, some of the most distressing and seemingly inexplicable suicides in young people are  probably due to early, unrecognized, and neglected effects of this formidable disease, which ought to make its victims particularly liable to attempt to destroy themselves. Long ago Durkheim (1951) suggested that there were three very different kinds of suicide which he called altruistic, egoistic, and anomic. He con-sidered that these were all exaggerations of social virtues; social solidarity—altruism, in-dividuality—egoism, and flexibility—anomie. Durkheim's schema is far more sophisticated and inclusive than most of those currently used in  psychiatry today, as any reader of psychiatric texts soon discovers. He held that altruistic suicide occurred when social solidarity is very high so that the life of an individual is perceived as being relatively unimportant compared with that of the group. Egoistic suicide is very different for here a progressive emphasis on the value of individuality, characteristic of some  phases of civilization, results in some people  becoming so detached from major social in-stitutions such as God, society, country, and all collective sentiments that they can feel and recognize no authority beyond themselves. In times of dislocations, stress, and anxiety these highly individualized people find that they lack the group support which they now need and are liable to end their lives in despair. Anomic suicide occurs when the "conscience collective," that system of social norms which reflects a commonality of beliefs and feelings, is disrupted. This leads to a disastrous "freedom" from social restraints. Durkheim states, "When our desires are freed from all moderating influence, when nothing limits them, they become themselves tyrannical and their first slave the very subject who experiences them." Durkheim was using sociological terms and did not concern himself with the psychology of the individual. It would seem that altruistic suicide is unlikely to occur often among schizophrenics, except perhaps very early in their illness for, as it has been shown elsewhere (Stengel, 1963), they are usually lacking in social cohesion. Egoistic and anomic suicide, however, could very easily be precipitated by this illness, not because society has overvalued individuality or because social norms have broken down, but  because the schizophrenic illness produces exactly these effects in those afflicted by it. Many sufferers from the paranoid varieties of schizophrenia grossly overvalue the individual as opposed to the social collective, because during their illness their perceptions of themselves and other people have become distorted. These enriched and enlarged perceptions (Kaplan, 1964; Landis, 1964) can themselves cause the sick  person to lose touch with social norms by giving him an altered and often inflated sense of the  possible. Morality with its easily understood rules is replaced by Durkheim's "tyranny of freedom" and the terror of what has been incorrectly  perceived as being enormously increased freedom of choice. For moral, that is acculturated, people who have internalized the values and attitudes of 60
of 11