Predictors of placebo response in bipolar depression

The aim of this work is to investigate placebo response rates in placebo-controlled randomized clinical trials (RCTs) of pharmacological therapy in bipolar depression (BPD) and to identify predictors of placebo response and clinical trial outcome in BPD. Medline/PubMed publication databases were searched for RCTs of oral drugs used as monotherapy for the treatment of BPD, published between January 1980 and September 2013. Data extracted from 12 manuscripts and one poster, representing a total of 17 clinical trials, were pooled. Pooled response rates for drug and placebo were 55.1 and 39.2%, corresponding to a risk ratio for responding to active treatment versus placebo of 1.29 (P<0.001). The probability of receiving placebo and trial duration correlated with the response rate to placebo. A meta-regression showed that trial duration and baseline severity correlated with the risk ratio of responding to drug versus placebo. There was a trend toward statistical significance for a greater probability of receiving placebo to predict greater drug-placebo 'separation'. In conclusion, several modifiable factors, specifically the probability of receiving placebo, baseline illness severity, and trial duration, correlate with placebo response rates and/or clinical trial outcome in RCTs of pharmacotherapy for BPD, and should be taken into account when designing studies for BPD.