Symptom Severity and the Generalizability of Antidepressant Efficacy Trials

The most commonly used inclusion/exclusion criterion in antidepressant efficacy trials (AETs) is a minimum score on a symptom severity scale. In the present study, we reviewed placebo-controlled AETs published during the past 20 years to determine whether there has been a change in the symptom severity inclusion criterion threshold subsequent to publications that highlighted the unrepresentativeness of the depressed patients studied in AETs. We identified 170 AETs published during the past 20 years and compared the studies published during the past 5 years (2010-2104, n = 56) with the studies published during the previous 15 years (n = 114). The symptom severity threshold for inclusion has increased in the more recent cohort of studies. On the 17-item Hamilton Depression Rating Scale, almost half of the studies of the past 5 years used a cutoff of 22 or greater to select patients versus less than one-fifth of the studies during the previous 15 years (44.0% vs 17.5%; χ(2) = 7.4; P < 0.01). Similarly, the cutoff on the Montgomery-Asberg Depression Rating Scale required for study inclusion has been higher in studies of the past 5 years with approximately three-quarters of the recent studies using a cutoff of at least 25, in contrast to one-quarter of the older studies (76.2% vs 25.0%; χ(2) = 8.2; P < 0.01). A significantly higher percentage of patients in our clinical practice would be excluded on the basis of the severity thresholds of the past 5 years (59.3 ± 13.5 vs 49.0 ± 15.1; t121 = 3.1; P < 0.005). These findings suggest that the results of AETs may not be applicable to less severely depressed patients who make up at least half of the patients treated in routine clinical practice. Questions are raised about the Food and Drug Administration labeling of antidepressants.