Depression. Aaron T. Beck, M.D.

after the initial episode of depression. His follow-up showed the following relapse rate for his 97 cases of recurrent depressions: less than 10 years after the first attack of depression, 35 percent; 10–20 years, 52 percent; more than 20 years, 13 percent. It should be emphasized that 65 percent had recurrences after remissions of 10–30 years.

      In an earlier study, Kraepelin had tabulated the symptom-free intervals between 703 episodes of depression. Unlike Rennie’s study, Kraepelin’s included intervals after the second and later attacks (as well as intervals between the first and second episodes). He found that with each successive attack the intervals tended to become shorter. Since his series consisted of hospitalized patients, it is interesting to note the same trend among the extramural patients in Paskind’s study. A comparison of the distribution of intervals in ten-year categories is shown in Table 3-3. For the purposes of comparison, Rennie’s results are also included. It should be emphasized that his findings applied only to the first interval. The tendency for his intervals to be longer than Kraepelin’s and Paskind’s may be explained by the fact that the later intervals included in their study were shorter than the first intervals. Kraepelin and Paskind showed a somewhat similar distribution of the intervals, with Paskind’s outpatient cases having longer periods of remission than Kraepelin’s hospitalized cases.

TABLE 3-3. Distribution of Time Intervals Between Manic-Depressive Episodes in Inpatients and Outpatients

      *Includes only first interval (between first and second episode).

TABLE 3-4. Median Intervals for Inpatients and Outpatients (years)

      Another way of expressing the duration of the intervals is in terms of the median duration of the specific intervals. Table 3-4 shows that the median interval is longer in Paskin’s outpatient cases, and also that in both outpatient and hospitalized cases the median intervals tended to be shorter with successive attacks. In Kraepelin’s study, the biphasic cases showed consistently shorter symptom-free intervals than the simple depressions.

      Further support for the observation that after the first recurrence the interval tends to become shorter is found in Lundquist’s study. In the age group older than 30, the mean duration of the first interval was about seven years, and the second interval three years. This difference was statistically significant.

      Lundquist’s data, classified according to three-year intervals, showed that the overwhelming preponderance of relapses occurred in the first nine years. It should be pointed out that his follow-up period was as short as 10 years in some cases, as compared to 25–30 years in Rennie’s series. Hence, it is probable that many of the cases in Lundquist’s series would have shown a relapse if they had been followed for a longer period than 10 years. Lundquist computed the probability of a relapse after a patient has recovered from an initial episode of depression (Table 3-5). These findings were tabulated separately for the young depressives and older depressives, but no significant difference was found between the two groups. It may be noted that the highest probability of recurrence was in the 3–6-year interval.

TABLE 3-5. Probability of Recurrence After Remission from First Attack (%)

      Adapted from Lundquist 1945.

      TABLE 3-6. Outcome According to Clinical Criteria (%)

      Adapted from Kiloh et al. 1988.

Outcome for “Endogenous” Versus “Neurotic” Depression

      Kiloh et al.²⁰ studied the long-term outcome of 145 patients with primary depressive illness who were admitted to a university hospital between 1966 and 1970. Patients were categorized into endogenous and neurotic subtypes. The follow-up period was an average of 15 years later, and data were obtained on 92 percent of the patients. Table 3-6 shows the percentages of those who (1) recovered and remained well, (2) recovered but experienced subsequent depression, or (3) remained incapacitated or committed suicide.

Schizophrenic Outcome

      In Rennie’s 1942 sample of 208 cases of manic-depressive psychosis, four cases changed their character sufficiently to justify the conclusion of an ultimate schizophrenic development. A review of these cases suggested that there was a strong component of schizophrenic symptomatology at the time of the diagnosis of manic-depressive psychosis.

      At about the same time, Hoch and Rachlin²¹ reviewed the records of 5,799 cases of schizophrenia admitted to Manhattan State Hospital, New York City. They found that 7.1 percent of these patients had been diagnosed as manic depressive during previous admissions. Whether there was an alteration in the nature of the disorder, an initial misclassification, or a change in diagnostic criteria was not established by these writers.

      Lewis and Piotrowski²² found that 38 (54 percent) of 70 patients, originally diagnosed as manic depressives, had their diagnoses changed to schizophrenia in a 3–20-year follow-up. In reviewing the original records, the authors demonstrated that the patients whose diagnoses were changed were misclassified initially, that is, they showed clear-cut schizophrenic signs at the time of their first admission. Because of the very loose criteria used in diagnosing manic-depressive disorder in the early decades of the twentieth century, it is difficult to determine what proportion, if any, of the clear-cut manic depressives had a schizophrenic outcome.

      Lundquist reported that about 7 percent of his manic-depressive cases eventually developed a schizophrenic picture.

      Astrup et al.¹¹ isolated 70 cases of “pure” manic-depressive disorder and followed these 7–19 years after the onset of the disorder. They found that none had a schizophrenic outcome. In contrast, 13 (50 percent) of a group of 26 cases diagnosed as schizoaffective psychosis showed schizophrenic symptomatology on follow-up.

Suicide

      At the present time, the only important cause of death in depression is suicide. (The general topic of suicide is broad, and many excellent monographs are available, e.g., Farberow and Schneidman,²³ Meerloo.²⁴) Previously, inanition due to lack of food and secondary infection were occasional causes of death, but with modern hospital treatment such complications are less usual.

      The actual suicide risk among depressed patients is difficult to assess because of the incomplete follow-ups and difficulties in establishing the cause of death. Long term follow-ups by Rennie⁵ and by Lundquist⁶ indicated that approximately 5 percent of the patients initially diagnosed in a hospital as manic depressive (or as having one of the other depressive disorders) subsequently committed suicide.

      In the mid-twentieth century, several studies demonstrated comparatively higher suicide rates among depressed patients. Pokorny²⁶ investigated the suicide rate among former patients in a psychiatric service of a Texas veterans’ hospital over a 15-year period. Using a complex actuarial system, he calculated the suicide rates per 100,000 per year as follows: depression, 566; schizophrenia, 167; neurosis, 119; personality disorder, 130; alcoholism, 133; and organic, 78. He then calculated the age-adjusted suicide rate for male Texas veterans as 22.7 per 100,000. The suicide rate for depressed patients, therefore, was 25 times the expected rate and substantially higher than that of other psychiatric patients.

      Temoche et al.,²⁷ studying the suicide rates among current and former mental institution patients in Massachusetts, found a substantially higher rate of suicide among depressed patients than nondepressed patients. The