Response rates to fluoxetine in subjects who initially show no improvement.

Objective: This study sought to investigate the likelihood that subjects will respond to continued antidepressant therapy when little or no benefit has yet been observed.

Method: Six hundred twenty-seven subjects diagnosed with DSM-IV major depressive disorder were recruited in a 12-week open-label trial with fluoxetine, which was designed as a preliminary phase to a subsequent 52-week continuation trial, which was conducted in 1997-2003. For each week of the study, a calculation was made for all subjects who had heretofore demonstrated little or no improvement as to the likelihood of converting to a positive response in subsequent weeks as measured by the Clinical Global Impressions scale, the primary outcome measure. In order to compare our findings with prior research, we focused primarily on outcomes at weeks 6, 8, and 12.

Results: The likelihood of converting to a positive response decreased the longer subjects remained unimproved. When week 6 was used as the end point, the likelihood of converting to a positive response for unimproved subjects at week 1 was 36% (n = 302); the respective conversion rates for weeks 2-5 were 29% (n = 208) at week 2, 18% (n = 151) at week 3, 17% (n = 120) at week 4, and 9% (n = 91) at week 5. When week 8 was used as the end point, the likelihood of converting to a positive response for unimproved subjects at week 4 was 23% (n = 118) and, at week 6, was 10% (n = 61). Finally, when week 12 was used as the end point, the likelihood of unimproved subjects at weeks 4, 6, and 8 converting to a positive response at week 12 was 50% (n = 117), 33% (n = 60), and 30% (n = 46), respectively.

Conclusions: The study adds to a small, but growing literature that gives clinicians some guidelines to help decide whether to continue an antidepressant trial when little or no benefit has yet been observed.

Trial Registration: clinicaltrials.gov Identifier: NCT00427128.