Citalopram for major depressive disorder in adults: a systematic review and meta-analysis of published placebo-controlled trials

Abstract

Objective To assess the effectiveness of citalopram for major depressive disorder (MDD) in adults, in a systematic review of all published, randomised, double-blind studies comparing it with a placebo. Data sources Cochrane Central Register of Controlled Trials, Medline, PsychINFO and Embase. Study selection Randomised, double-blind, placebo-controlled studies of citalopram in adults with MDD were included. Studies with medically ill or treatment resistant subjects were excluded, as were studies of relapse prevention. Remission of MDD was defined as a primary outcome, and response or change from baseline scores were defined as secondary. Data extraction Remission, response and symptom improvement scores on the Hamilton Depression Scale, Montgomery-Asberg Depression Rating Scale and Clinical Global Impressions-Severity scales were extracted. A random-effects meta-analysis was carried out on the response rates and symptom improvement scores. Included studies were examined for the presence of bias and small study effects. Results Eight studies (n=2025) met the inclusion criteria. Two studies provided data on remission, but only one of these showed a significant difference between citalopram and placebo (RR=1.59, 95% CI 1.10 to 2.31). Meta-analysis of response rates in five studies (n=1010) revealed significant superiority of citalopram (RR=1.42, 95% CI 1.17 to 1.73). Meta-analysis of change from baseline scores in five studies (n=1541) gave a standardised mean difference (Hedges' g) of -0.27 (95% CI -0.38 to to -0.16), showing a reduction in MDD symptoms to be significant for citalopram relative to placebo. There was no evidence of any significant small study effects. The overall quality of reporting was poor, with insufficient information on the methodology or outcomes. Seven studies received industry sponsorship. Conclusions Data concerning remission rates for citalopram, relative to placebo, are inconclusive. Response rates and symptom reduction scores in citalopram-treated patients with MDD are significantly better relative to placebo treatment, according to a meta-analysis of published reports. Evaluation of unpublished data is necessary to assess more definitively the effectiveness of citalopram for MDD.

Figure 1

Summary of the article selection process.

Figure 2

Random effects meta-analysis of symptom responses for citalopram and placebo.

Figure 3

Funnel plot of symptom response odds ratios for citalopram and placebo.

Figure 4

Meta-analysis of standardised mean differences (SMD) in change from baseline HAM-D comparing citalopram and placebo.