Since the two last decades, many authors have broadened the scope of mood disorders to include a larger bipolar spectrum which encompasses the sub-affective conditions, including temperaments. According to this view, the latter conditions represent milder or alternative expressions of the classic bipolar episodes. In successive elaborations, Akiskal et al. hypothesized a complex multicausal approach to bipolar disorder, and studied temperamental dysregulations, which could serve as risk factors for major episodes. Until recently, there have been several studies of patients populations, little is known in control populations. The aim of this report is to compare the rates of three affective temperaments (hyperthymic: TH; depressive: TD; irritable: TI) in non-ill subjects with different risk for mood disorders. (The cyclothymic temperament is studied as part of another report).
Methods: We recruited 185 individuals from: a) staff hospital; b) sibling of patients suffering from bipolar disorder, type I. Twenty subjects were excluded: 7 suffered from personal affective trouble; 12 exhibited cyclothymic traits; and one had familial schizophrenia. In the 165 remaining subjects, the temperamental characteristics were assessed by mean of the Akiskal and Mallya's criteria (1987, semi-structured interviews for affective temperaments, TH, TD, TI). Then, the population of controls was divided in 3 groups as a function of the familial loading for affective disorder and bipolar disorders: the first subgroup (AFN) was free of any antecedent ("super-normal controls", n=99); the second subgroup (AFP) had familial antecedents at the first or second degree (normal controls but at risk for affective disorder, n=33); the third subgroup (FBP) was composed of the siblings of bipolar I patients (subjects at high risk, n=33). Statistical procedures included standard and non-parametric methods: means standard deviation, Fisher's test, Mann-Whitney' and Kuskall-Wallis' tests, Spearman's correlation coefficient. As described by Placidi and collaborators (12), we also used the Z-score (temperamental score strictly higher than the second positive standard deviation: m + 2 sd).
Results: The general demographic characteristics show a higher frequency of women (p=0.02) but a similar mean age (p=0.296, NS) among the groups. The mean scores of the TH and TD are strongly and negatively correlated (Rho coefficient=- 0.397, p=0.01), exhibiting the internal coherence of the responses. The comparison of the temperamental characteristics among the 3 groups exhibits significant differences for the TH and TI (p=0.003). The mean scores are respectively: for the TH, 9.16 4.18 in AFN, 8.33 4.11 in AFP, and 12.16 5.28 in FBP; and for the TI, 8.94 2.25 in AFN, 9.39 2.63 in AFP, and 10.84 2.76 in FBP. Conversely, the TD scores do not significantly differ: 6.01 3.27 in AFN, 6.76 4.34 in AFP, and 7.94 5.28 in FBP. Beyond these first pass results, we also considered the distribution of the subjects as function of the Z-score and the different groups. We found that hyperthymic traits were almost exclusively among the FBP: 15.1% vs 3.0% in the other groups. For the TD, expressed in mean scores, the groups at risk for affective disorders (AFP and FBP) clearly display a percentage of subjects with a more substantial Z-score than the frequencies observed in the AFN: respectively 12.1%, 18.1% and 4.0% for the TD. Concerning traits of all three temperaments, as function of the demographic variables and the Z-score, they are generally predominant in males; however, the TH is more frequent in males only in the AFP and FBP groups (respectively: 8.3% vs none; 21.4% vs 10.5%). The TD is more prevalent among females in AFP and FBP (respectively: 8.3% vs 14.3%; 21.1% vs 14.8%).
Conclusion: Our results clearly show temperamental dysregulations in the subjects at risk for affective disorders: (1) the levels of all three affective temperaments under study are significantly higher in subjects at risk for affective disorder, as compared to individuals free of a family antecedent; (2) the depressive temperament is prevalent in both AFP and FBP, whereas the hyperthymic is specific for FBP. As for Akiskal's model on the multicausal origin of the mood disorders, our data supports temperamental dysregulation as an important familial genetic factor in the vulnerability to manic depressive episodes. We further posit that such temperaments--more specifically, the hyperthymic--could serve as proximal phenotypes for full-blown bipolar disorder.