Depression. Aaron T. Beck, M.D.

the patient as well as appropriate financial arrangements for his or her care.

      One aspect of the usual depressive episode that is of importance in treatment is the fact that the episode tends to follow a curve, that is, tends to progressively worsen, then bottom out, and then progressively improve until the patient returns to his premorbid state. By determining the time of onset of the depression, the physician can make a rough estimate as to when an upward turn in the cycle may be expected. It is particularly important when assessing the efficacy of specific forms of treatment to take into account the spontaneous start of the upward swing.

      There is some variation in the findings of the numerous studies relevant to duration. Undoubtedly, these variations may be attributed to different methods of observation and to different criteria for making diagnoses and judging improvement. In general, the relatively unrefined clinical studies (which will be discussed presently) indicate a longer duration than do the systematic studies.

      Lundquist⁶ found that the median duration of the attack of depression in patients younger than 30 was 6.3 months, and for those older than 30, 8.7 months. This difference was statistically significant. There was no significant difference between men and women in regard to duration. (As noted previously, he also found acute onset associated with shorter duration.) Paskind⁴ also found in his outpatient group a shorter duration of attacks occurring before age 30 than after age 30. Rennie’s study yielded similar results, the first episode lasting on the average 6.5 months. He found, incidentally, that the average duration of hospitalization was 2.5 months. In Paskind’s series of non-hospitalized depressives, the median duration was 3 months. He found that 14 percent of the episodes lasted one month or less, and that almost 80 percent were completed in six months or less.

      The earlier, less refined studies predominantly reported a period of 6–18 months as the average duration of the first attack: Kraepelin,¹ 6–8 months; Pollack,¹⁷ 1.1 years; Strecker et al.,¹⁰ 1.5 years. The clinical impression of writers of monographs on depression published in the 1950s and 1960s shows similar variation. Kraines¹⁸ stated that the average depressive episode lasts about 18 months. Ayd¹⁴ reported that prior to age 30, the attacks average 6–12 months; between ages of 30 and 50, they average 9–18 months; and after 50, they tend to persist longer, with many patients remaining ill from three to five years.

      In regard to the duration of multiple episodes of depression, there was a prevalent opinion among the earliest clinicians of a trend toward prolongation of the episodes with each recurrence.¹ Lundquist, however, performed a statistical analysis of the duration of multiple episodes and found no significant increase in duration with successive attacks. Paskind’s⁴ study of outpatient cases similarly showed that the attacks do not become longer as the disease recurs. The median duration for first attacks was four months, and for second, third, or subsequent attacks three months.

      The differences in the findings between the rough clinical studies and the statistical studies may reflect a difference in samples and/or different criteria for recovery from the depression. It is probable that certain biases influenced the selection of cases in the less refined studies and, therefore, the samples cannot be considered representative.

      Lundquist found a significant association between prolonged duration and the presence of delusions in younger but not older patients. The presence of confusion, however, favored a shorter duration.

Brief Attacks of Manic-Depressive Psychosis (Bipolar Disorder)

      In 1929, Paskind² described 88 cases of depression of very brief duration, from a few hours to a few days. These patients had essentially the same symptoms as those in his other extramural cases of longer duration and constituted 13.9 percent of his large series of cases diagnosed as manic-depressive disorder. The case histories he presented leave little doubt that they would later be diagnosed as neurotic-depressive (dysthymia) reaction.

TABLE 3-2. Frequency of Single and Multiple Attacks of Depression

      Most of these patients with brief attacks also experienced longer episodes of depression. In 51, the brief attacks came first and were followed from months to decades later by longer attacks lasting from several weeks to several years. In 18, longer attacks occurred first, and were followed by the transient episodes. In nine, there were brief episodes only.

Recurrence

      There is considerable variation in the older literature relevant to the frequency of relapses among depressed patients. Except as indicated, the statistics for manic-depressive psychosis include some manic patients in addition to the depressed patients. In the earlier studies, German authors reported a substantially higher incidence of recurrence than American investigators.⁶ These differences may be attributed to more stringent diagnostic criteria and to longer periods of observation by the German authors.

      Of the more refined studies, Rennie’s reported relapse rate was closer to that of the German writers than to those of the other American investigators. He found that 97 of 123 patients (79 percent) initially admitted to the hospital in a depressed state subsequently had a recurrence of depression. (These figures do not include 14 patients who committed suicide after the first admission or who remained chronically ill.) When the cyclothymic cases (i.e., patients who had at least one manic attack in addition to the depression) are added to this group, the proportion of relapse is 142 patients of 170 (84 percent).

      The Scandinavian investigators Lundquist⁶ and Stenstedt¹² reported, respectively, a 49 percent and a 47 percent incidence of relapse. In comparing their studies with Rennie’s, one can reasonably conclude that the more stringent diagnostic criteria employed by Rennie and the longer period of observation of his sample may account for the higher percentage of relapses in his report.

      The differences in relapse rate are reflected in a striking difference in the rate of multiple recurrences. In Rennie’s series more than half of the depressed patients had three or more recurrences (see Table 3-2). The frequency of multiple recurrences in the cyclothymic cases was particularly high in Rennie’s series. Thirty-seven of the 47 patients in the group had four or more episodes. In Kraepelin’s series, 204 out of 310 cases of this type (67 percent) had one or more recurrences, with more than half having three or more attacks.

      Another important aspect of the recurrent attacks is their duration. The opinion was frequently expressed that the episodes become progressively longer with each recurrence. Rennie, however, in analyzing his data, found that the second episode had the same duration as the initial episode in 20 percent, was longer in 35 percent, and was shorter in 45 percent. Paskind found that the median duration decreased with successive attacks.

      Belsher and Costello¹⁹ reviewed 12 published studies of relapse in unipolar as opposed to bipolar depression. They selected studies that included correlates of relapse, rates of relapse, and a naturalistic follow-up period with no controlled maintenance therapy. They found a number of methodological inadequacies, such as unclear and variable definitions of recovery and relapse, nebulous patient characteristics, and vague inclusion and exclusion criteria. Despite these uncertainties, they were able to conclude that the risk of relapse in unipolar depression goes down the longer a person stays well. Several factors did predict relapse: (1) a history of depressive episodes, (2) recent stress, (3) poor social support, and (4) neuroendocrine dysfunction. Other variables did not predict relapse, including marital status, gender, and socioeconomic status.

Intervals Between Attacks

      In examining the older literature on the intervals between episodes of depression, one is struck by the fact that recurrences may occur after years, or even decades, of apparent good health. The systematic studies offer little encouragement for the notion of a permanent cure analogous to the 5-year cures reported for cancer treatment. Recurrences have been reported as long as 40 years after remission from an initial depression.¹

      The findings presented by Rennie, in particular, are noteworthy in