Escitalopram was compared to placebo in moderately to severely depressed patients in primary care with citalopram as the active reference. Patients were randomized to receive flexible doses of 10–20 mg/day escitalopram (n=155), 20–40 mg/day citalopram (n=160), or placebo (n=154) over an 8‐week double‐blind period. The primary efficacy parameter was the change from baseline to last assessment in the Montgomery–Asberg Depression Rating Scale total score. Escitalopram produced a statistically significant therapeutic difference of 2.9 points (P=0.002) compared to placebo, and escitalopram was consistently and statistically significantly more efficacious than placebo from week 1 onwards. Analysis of Clinical Global Impression–Severity and Clinical Global Impression–Improvement confirmed the primary efficacy results. By week 8, significantly more patients had responded to treatment with escitalopram than with citalopram (P=0.021) or placebo (P=0.009). Escitalopram was as well tolerated as citalopram and had a similar adverse event profile. Both escitalopram‐ and citalopram‐treated patients had placebo‐level adverse event withdrawal rates (3% and 4%, respectively). This study demonstrates the consistent antidepressant efficacy and excellent tolerability of escitalopram 10–20 mg/day in primary care patients with major depressive disorder.