The heterogeneity of the depressive syndrome: when numbers get serious

Abstract

The current criteria for major depression (1) have been criticized for the heterogeneity of the clinical syndrome they define. This has led to suggestions of alternative classifications, most recently by G. Parker in this journal (2). According to the critics, the polymorphic syndrome is partly to blame for the fact that the treatment of depression still has not moved beyond the trial and error approach and that the genetics and neurobiology underlying the depressive disorder still remain largely unknown and without practical significance (3). Two of the most interesting studies on major depression, conducted within recent years, reflect this problem: In the STAR*D study, which included more than 4000 subjects with major depression, only 30% of the patients remitted on treatment with a current first-line antidepressant drug (Citalopram) and other 30% did not remit after four consecutive treatment trials with antidepressants from different pharmacological classes - despite a formalized and aggressive dosing strategy (4). The hypothesis of increased risk for depression caused by an interaction between the serotonin transporter promoter gene polymorphism (5-HTTLPR) and 'adverse life events', such as childhood maltreatment and medical conditions, has been subjected to intense debate over the past decade. Does the interaction take place or not? This discussion may come to an end after the publication of a recent meta-analysis by Karg et al. (5), which is based on more than 40 000 subjects. The results of the meta-analysis suggest that the 5-HTTLPR - adverse life event - interaction does increase the risk of depression. What can be learned from these two large-scale studies? The fact that only 30% of patients remitted on a second-generation antidepressant is probably not because the drug does not work. Likewise, the fact that it took more than 40 000 subjects to 'prove' the interaction between the serotonin transporter gene and adverse life events is probably not because such interaction does not have an impact on the risk of developing depression. The underlying reason for these rather disappointing results is likely to lie in the heterogeneity of the depressive syndrome. The subjects included in the two studies are probably so dissimilar that any relevant causal effect of risk factors and clinical benefit of treatments are lost in the flood. But exactly how heterogeneous is the depressive syndrome? To give a theoretical answer to this question, we applied simple combinatorics to the nine symptoms considered in the DSM-IV criteria (1): Depressed mood Markedly diminished interest or pleasure Significant weight loss or weight gain Insomnia or hypersomnia Psychomotor agitation or retardation Fatigue or loss of energy Feelings of worthlessness or excessive or inappropriate guilt Diminished ability to think or concentrate or indecisiveness Recurrent thoughts of death or recurrent suicidal ideation The binomial coefficient calculates the number of subsets of k draws from n distinguishable objects without replacement and without regard to order. The result is 227 different combinations with ≥5 symptoms, but this is not the full picture. At least four of the DSM-IV symptoms actually cover two distinct symptoms each: 'weight loss or weight gain', 'insomnia or hypersomnia', 'psychomotor agitation or retardation' and 'feelings of worthlessness or inappropriate guilt'. When these, quite different, individual symptoms are taken into account, the resulting number of combinations is 1497, without considering the psychotic subtype of depression (delusions, hallucinations, or both) or the fact that depressive episodes can occur in both unipolar and bipolar courses of illness. Furthermore, it can be argued that the core symptom 'diminished interest or pleasure' also covers two different symptoms. Thus, the 1497 combinations are actually a conservative estimate. These calculations may be considered as mere academic exercise, but they also prove a point, namely that the current depressive syndrome is at least very, if not too, heterogeneous. Obviously, some of the 1497 combinations have a considerable overlap, but on the other hand, patients fulfilling the diagnostic criteria do not necessarily share any symptoms at all: John, a former successful CEO of a large company, developed depression after the bankruptcy of his company, which led to his firing and subsequent divorce. John has no relatives with depression or other mental disorders. Alice, graduate student in psychology, developed depression without any external stressors. There is a strong family history of depression and bipolar disorder in her family. The clinical pictures of John and Alice are quite different, yet both patients may have been included as subjects in the STAR*D (4) and the gene-environment study by Karg et al. (5). With this degree of heterogeneity among participants in depression-trials, it is not a major surprise that researchers are struggling to tease ou