BACKGROUND: Duloxetine has demonstrated analgesic effect in chronic pain states. This study assesses the efficacy of duloxetine in chronic low back pain (CLBP). METHODS: Adult patients with non-radicular CLBP entered this 13-week, double-blind, randomized study comparing duloxetine 20, 60 or 120 mg once daily with placebo. The primary measure was comparison of duloxetine 60 mg with placebo on weekly mean 24-h average pain. Secondary measures included Roland-Morris Disability Questionnaire (RMDQ-24), Patient's Global Impressions of Improvement (PGI-I), Brief Pain Inventory (BPI), safety and tolerability. RESULTS: Four hundred four patients were enrolled, 267 completed. No significant differences existed between any dose of duloxetine and placebo on reduction in weekly mean 24-h average pain at end-point. Duloxetine 60 mg was superior to placebo from weeks 3-11 in relieving pain, but not at weeks 12-13. Duloxetine 60 mg demonstrated significant improvement on PGI-I, RMDQ-24, BPI-average pain and BPI-average interference. Significantly more patients taking duloxetine 120 mg (24.1%) discontinued because of adverse events, versus placebo (8.5%). CONCLUSIONS: Duloxetine was superior to placebo on the primary objective from weeks 3-11, but superiority was not maintained at end-point. Duloxetine was superior to placebo on many secondary measures, and was well-tolerated.