Treatment decisions for the use of opioid analgesics in chronic non-malignant pain are based primarily on survey data, as evidence from well-controlled clinical trials has been lacking. Forty-six patients with chronic non-malignant pain were enrolled in a randomized, double-blind, placebo-controlled evaluation of controlled-release (CR) codeine. Following a 3-7-day diary familiarization period, patients were randomly assigned to 7 days of treatment each with CR codeine q12h or placebo. The CR codeine dose was determined from the consumption of acetaminophen+codeine in the 7 days preceding the study. During both phases, breakthrough pain was treated with acetaminophen+codeine every 4 h as required. Pain intensity was assessed at 08:00 h and 20:00 h using a visual analogue scale (VAS) and a 5-point categorical scale, and rescue analgesic consumption was recorded at the time of use. Thirty patients (17 female, 13 male; mean age: 55.1 +/- 13.4 years) completed the study and were treated with a mean daily CR codeine dose of 273 +/- 78 mg (range: 200-400 mg). CR codeine treatment resulted in significantly lower overall VAS pain intensity scores (35 +/- 18 vs. 49 +/- 16, P = 0.0001), categorical pain intensity scores (1.7 +/- 0.6 vs. 2.2 +/- 0.6, P = 0.0001), and in pain scores by day of treatment and by time of day. Daily rescue analgesic consumption was significantly lower on CR codeine, relative to placebo treatment (3.6 +/- 3.5 vs. 6.1 +/- 3.2 tablets/day, P = 0.0001). There was also a significant reduction in the Pain Disability Index (PDI) on CR codeine, compared to placebo (25.0 +/- 7.7 vs. 35.1 +/- 8.2, P = 0.0001). Patients' and investigators' blinded treatment preference was significantly in favor of CR codeine, relative to placebo (73% vs. 10%, P = 0.0160 and 80% vs. 7%, P = 0.0014, respectively). The incidence of nausea was significantly higher on CR codeine than on placebo (32.6% vs. 11.9%, P = 0.013). Ninety-three percent of patients completing the study requested long-term, open-label treatment with CR codeine. Pain intensity scores at the completion of 19 weeks of long-term evaluation were comparable to those during the double-blind CR codeine treatment. We conclude that treatment with CR codeine results in reduced pain and pain-related disability in patients with chronic non-malignant pain.