The relationships between plasma morphine and metabolite (M3G and M6G) concentrations and analgesic efficacy were investigated in an open study of 39 cancer pain patients receiving chronic oral morphine therapy with either morphine sulfate solution or controlled-release morphine tablets. There were no differences in morphine, metabolite kinetics, or analgesic efficacy between equivalent doses of conventional or controlled-release formulations. The increase in morphine plasma concentration after a dose (1 hr for normal release, 2 hr for controlled release) was correlated significantly with the dose of morphine (r = 0.914, P < 0.001). There was a significant reduction in pain intensity (P < 0.05) and increase in pain relief (P < 0.001) after the dose of morphine administration, when compared with the predose score. One-half of the patients had mild and tolerable adverse effects. Patients were classified by mean pain relief between doses as having optimal, moderate, or poor pain control. No simple relationship was found between morphine plasma concentration and pain control. Morphine plus M6G concentrations in the "optimal control" group (751.2 +/- 194 nmol/L), however, were more than twice those found in the "moderate control" group (276.9 +/- 41.9 nmol/L) (P < 0.05), and no patient in the moderate control group had a morphine plus M6G concentration greater than 405 nmol/L. These results support the importance of M6G in morphine analgesia. For these hospitalized patients, there appeared to be a therapeutic range of morphine plus M6G plasma concentrations for optimal pain control with a lower limit of 400 nmol/L predose.