RESULTS AND COMMENTS
Analyzing the frequency distribution of the personality disorders in the study subgroups we found that the most frequent were histrionic disorders among the bipolar patients (30%) and compulsive ones among the unipolar group (32%). Paranoid and dependent disorders were the most frequent (9%) among the control group. Borderline disorders were present only in the latter (6%). Further, personality disorders were differently distributed among the three groups, most strongly clustered in the group of unipolar patients—avoidant (18%), compulsive (32%), dependent (25%), and passive-aggressive (11%) personalities—overlapping cluster III reported by Pfohl et al., in 1984 (24). Unipolar patients showed higher frequency personality disorders than bipolar and controls. Schizotypal and antisocial personality types were not present in any of the sample groups, although schizoids were found only among the unipolar patients (11%), as was the case with avoidant disorders. Table 1 presents data on morbidity risks for first-degree relatives according to the presence or absence of personality disorders, sex of probands, and type of familial relationship for the three study subgroups.
The statistical analysis of these data showed there was a significant interaction between the presence of personality disorders and type of three study subgroups, which could predict differential affective illness risks (Z value of 3.5737 with p>= 0.01). The sex of proband and the type of familial relationship did not turn out to be significant predictors of differential risk. Table 2 summarizes these statistical results. There was a different picture when the data on second-degree relatives were analyzed (Table 3). In this case, the statistical analysis showed (Table 4) that personality disorders did not have a significant relationship with increasing familial risk for disease. On the contrary, increased risks appeared to be sex related, as shown by the bipolar and unipolar females as compared to males. In fact, there was a statistical significance of sex of probands (Z value of -2.1920 with p >= 0.02) and of study subgroups (Z value of 2.0473 with p >= 0.01). The difference in the findings between first- and second-degree relatives could 
be due to an underestimate of risk for the latter (estimated at 4% vs. 14% for the first-degree ones), probably associated with the scarcer amount of available information about them. Further, the difference might also be explained by the close similarity between second-degree relatives and the general population, i.e., a reduced morbidity risk. Consequently, this suggests that the morbidity risks for first- and second-degree relatives should be evaluated separately because of their structural or structurally overlapping variability. Some data from Yale University (29,31) support increased familial risks when affective disorders involving earlier ages of onset are concerned. Consequently, age at onset could be useful to select more homogeneous disease subgroups characterized by different degrees of severity. We tentatively analyzed the familial risks of our bipolar and unipolar patients subdivided into three groups by age at onset: 12 to 19, 20 to 35, and over 36 years. The data are presented in Table 5. Here you see that the risks in our 
sample were correlated with higher bipolar risks in the first and third age at onset intervals (12-19, 36+) and with the second age at onset interval for unipolar patients (20-35). Further, unipolar patients of the youngest and oldest age at onset groups showed an increasing familial risk for diseases depending on the presence of personality disorders (Table 5). Statistical analysis of these data showed that the very early and very late age at onset of unipolar affective disorders significantly affects higher risks for such disorders in families (Z values of 2.3088 and -2.24359 with p > 0.02) and significantly interacts with the presence of personality disorders (Z value of -2.4237 with p > 0.02) (Table 6). The unipolar patients with personality disorders in our sample had a lower mean age at onset than unipolar without personality disorders (33.88 + 13 vs. 39.2 + 12). This agrees with previous reports (12). Despite the fact that DSM-III states that there may be no causal relationship between personality and illness, one possible explanation could be that an earlier onset of depression may result in early functional impairment that could later be structured as a personality trait or disorder. 
On the other hand a character disorder could itself cause repeated subjective distress, possibly predisposing the individual to depressive episodes. Nevertheless, the highest risk for affective disease in relatives and the very early age at onset of patients with uniipolar depression plus personality disorders could also be consistent with the hypothesis that personality may add to the severity of the form detected from the point of familial loading and the early development of the disorder. The same effect is shown by the latest onset ages of unipolar patients with personality disorders. On the contrary, the presence of personality disorders in the 25- to 35-year-old group appeared to lower the risk for disease in families, as in all types of biipolar patients. This seems to confirm some heterogeneity among unipolar disorders. In other words, we might hypothesize that one subgroup of unipolairs is similar to the bipolar in terms of genetic load, whereas other groups (those: with the earliest and latest ages at onset) do require the presence of other faictors, such as personality disorders, to manifest greater severity. A similar effect of increasing severity from the viewpoint of familial load was reported by Pauls and DiBenedetto (this volume) as regards panic attacks. This study showed that controls with personality disorders also showed increased familial loaids. This could suggest that personality disorders might be assumed as a milder form of depression. Consequently, it would be of great interest to see whether personality disorder rates are also higher in the relatives. In this case, it wouldl be useful to identify the personality types specifically involved so as to attempt toi introduce them into the spectrum of depressive disorders (1,2,6,8,21,25,28). To this end, it will be necessary to study larger samples (giving cell sizes sufficiently large) so as to be able to study the familial risk for each specific personality disorder. In addition, the suitability of DSM-III diagnoses for personality disorders needs testing, perhaps comparing these to other personality diagnoses criteria and to identify personality disorders verifiably related to the affective illness. Furthermore, we need to use several criteria, biological, pharmacological, cliinical-biological, including age at onset, for singling out more homogeneous groujps of affective patients who will be more suitable for performing future genetic-oriented analy
ses, such as segregation and linkage.
REFERENCES
1. Akiskal, H. S. (1983): Am. J. Psychiatry, 140:11-20. 2. Akiskal, H. S. (1983): In: Psychiatry Update: The American Psychiatric Association Annual Review, Vol II, editted by L. Grinspoon, pp. 271-292. American Psychiatric Press. 3. Akiskal, H. S., Bitair, H., Puzantian, V. R., et al. (1978): Arch. Gen. Psychiatry, 35:756-766. 4. Akiskal, H. S„ Djemderedjianah, Rosenthal, R. H., et al. (1977): Am. J. Psychiatry, 134:1227- 1233. 5. Akiskal, H. S., Hirrschfeld, R. M. A., and Yerevanjan, B. I. (1983): Arch. Gen. Psychiatry, 40. 801-810. 6. Akiskal, H. S., Yere;vanian, B. I., Davis, G. C, King, D., and Lemmi, H. (1985): Am. J. Psychiatry, 142:192-198. 7. American Psychiatric Association. (1980): Diagnostic and Statistical Manual of Mental Disorders, 3rd ed. American Psychiatric Association, Washington, D.C. 8. Andreasen, N. C, Endicott, J., Spitzer, R. L., et al. (1977): Arch. Gen. Psychiatry, 34:1229-1235. 9 Andrulonis, P. A., Glueck, B. C, Stroebel, С F., et al. (1982): J. Nerv. Ment. Dis., 170:670-679. 10. Carrol, B. J., Greden, J. F., Feinberg, M, et al. (1982): Psychiatr. Clin. North Am., 4:89-99. 11. Cassano, G. В ., and Akiskal, H. S. (1985): In: // DSM III in Italia, Diagnosi e Classificazione dei Disturbi Affettivi, edited by G. B. Cassano, pp. 41-50. Masson, Milan. 12 Charney, D. S., Nelson, J. C, and Quinlan, D. M. (1981): Am. J. Psychiatry, 138:1601-1604. 13′ Chodoff, P. (1972): Arch. Gen. Psychiatry, 27:666-673. 14 Crowe, R. R., and Smouse, P. E. (1977): / Psychiatr. Res., 13:273. 15 Davidson, J., Miller, R., and Strickland, R. (1985): J. Affective Disord, 8:177-182. 16. Heimbuch, R. C, Matthysse, S., and Kidd, К . К . (1980): Am. J. Hum. Genet., 32:564-574. 17 Koenigsberg, H. W., Kaplan, R. D., Gilmore, M. M., and Cooper, A. M. (1985): Am. J. Psychiatry, 142:207-212. 18. Lazare, A., and Klerman, G. (1968): Am. J. Psychiatry, 124:48-56. 19. Morabito, A., and Marubini, E. (1976): Comput. Programs Biomed., 5:283-295. 20. Morabito, A., and Marubini, E. (1980): In: Analyse des Donnees, Actes de la Journee de Travail, pp. 393-408. Institut Nationale De Recherche en Informatique et en Automatique, Le Chesnay. 21. Nelson, H. F., Tennen, H., Tasman, A., Borton, M., Kubeck, M., and Stone, M. (1985): Am. J. Psychiatry, 142:855-858. 22. Paykel, E. S., Klerman, G. L., and Prusoff, B. A. (1976): Br. J. Psychiatry, 129:327-334. 23. Penis, С (1966): Acta Psychiatr. Scand. (Suppl.), 194:68-82. 24. Pfohl, В ., Stangl, D., and Zimmerman, M. (1984): J. Affective Disord., 7:309-318. 25. Pope, H. G, Jonas, J. M., Hudson, J. I., et al. (1983): Arch. Gen. Psychiatry, 40:23-30. 26. Stangl, D., Pfohl, В ., Zimmerman, M., Bowers, W., and Corenthal, С (1985): Arch. Gen. Psychiatry, 42:591 -596. 27. Tyrer, P. J. (1983): Br. J. Psychiatry, 142:404-408. 28. Weissman, M. M., Gershon, E. S., Kidd, K. K, et al. (1984): Arch. Gen. Psychiatry, 41:13-21. 29. Weissman, M. M., Merikangas, K. R., Wickramaratne, P., et al. (1986): Arch. Gen. Psychiatry, 46:430-434. 30. Weissman, M. M., Prusoff, B. A., and Klerman, G. L. (1978): Am. J. Psychiatry, 135:797-802. 31. Weissman, M. M., Wickramaratne, P., Merikangas, K. R., et al. (1984): Arch. Gen. Psychiatry, 41:1136-1143.
Anxious Depression: Assessment and Treatment, edited by G. Racagni and E. Smeraldi. Raven Press, New York © 1987.