Background: The literature on predictors of response to treatment of nonchronic major depression has identified shorter duration of illness, acute onset, and less severity of illness as positive predictors. Unfortunately, there are almost no data on predictors of response to treatment for chronic depression. This study examined predictors of response to pharmacotherapy (sertraline or imipramine) in the treatment of outpatients who had DSM-III-R-defined chronic major or double depression.
Method: The acute phase of the Chronic Major Depression and Double Depression Study is a double-blind, randomized, parallel-group 12-week comparison of sertraline and imipramine. Analyses are based on 623 patients who comprised the intent-to-treat sample, of whom 299 were nonresponders and 324 were responders, defined by a priori criteria as either remission or satisfactory therapeutic response. A stepwise logistic multiple regression analysis was performed on candidate clinical, psychosocial, and demographic variables previously identified as statistically significant in an attempt to develop a predictive model of positive antidepressant response.
Results: The sociodemographic variables that were predictive of positive response included living with spouse or partner or being at least a high school graduate. With regard to symptomatology and clinical history, responders had significantly lower baseline depression severity scores. In general, comorbid anxiety, substance abuse, and personality disorders did not influence rates of response. However, the presence of depressive personality traits was associated with a higher nonresponse rate. Among psychosocial variables, longer duration of personal relationships as well as higher baseline quality of life were associated with positive response. A stepwise logistic multiple regression identified 5 variables-living with spouse or partner, higher educational level, passive-aggressive personality, lower introverted-tense personality traits, and higher quality of life--that significantly and independently contributed to the predictive model. This model correctly classified 67% of patients.
Conclusion: A higher baseline quality of life, living with spouse or partner, and having more education were the strongest predictors of response to acute pharmacotherapy among chronically depressed patients. Clinical variables and comorbidity were not identified as independent predictors, although personality traits did appear to influence treatment response. Overall, the predictive value of these baseline measures was modest, and therefore of limited clinical utility.