The clinical importance of routine drug monitoring of serum concentrations of morphine, morphine-6-glucuronide (M6G) and morphine-3-glucuronide (M3G) during chronic morphine therapy is not established. We measured morphine, M6G and M3G serum concentrations in cancer pain patients receiving oral (n = 263, median dose 80 mg/24 hours) or subcutaneous (sc) (n = 35, median dose 110 mg/24 hours) morphine. Regression analyses were performed to investigate if serum concentrations of morphine, M3G and M6G predicted pain intensity (Brief Pain Inventory), health-related quality-of-life variables (EORTC QLQ-C30) and cognitive function (Mini-Mental Score). Serum concentrations were also compared in patients categorized as morphine 'treatment successes' and 'treatment failures'. We observed that serum concentrations of morphine, M6G or M3G did not predict pain intensity, cognitive function, nausea or tiredness. 'Treatment failures' caused by nausea, tiredness, cognitive failure or constipation did not have statistically significant different morphine, M6G and M3G serum concentrations than patients classified as 'treatment successes'. In conclusion, this study did not observe any concentration-effect relationships of morphine, M3G or M6G with pain intensity, nausea, constipation, tiredness or cognitive failure in blood samples obtained during routine clinical drug monitoring in cancer patients. This result suggests that therapeutic drug monitoring as a routine tool during chronic morphine treatment has limited value for clinical decision making.