The present study was aimed at evaluating chronic stress models in mice with special attention to morphine treatment. We hypothesized that repeated periods of drug withdrawal induce chronic stress. To verify this hypothesis, mice were made dependent on morphine and then subjected to several types of repeated withdrawal. Body weight reduction, thymus involution, adrenal gland enlargement and activation of the hypothalamo-pituitary-adrenal axis were used as signs of chronic stress. The changes were compared to those induced by 'laboratory' models of chronic stress (2 weeks of repeated restraint or rat exposure) and to a disease model of streptozotocin-induced diabetes mellitus (STZ-DM). Mice were made dependent using increasing doses of morphine three times a day for 3 days (10-20-40 mg/kg s.c.). Thereafter, withdrawal was induced either spontaneously (morphine 40 mg/kg injected at 24- or 72-hour time intervals for 2 weeks) or repeatedly precipitated by naloxone (10 mg/kg s.c.) injected daily 3 h after morphine. The results show that repeated periods of spontaneous drug withdrawal (24 or 72 h) in morphine-dependent mice represent a mild stress load. Repeated withdrawal precipitated by naloxone induced clear chronic stress-like changes. Changes observed in the naloxone-precipitated withdrawal model were even more pronounced than those found in laboratory models, namely repeated restraint or exposure to the rat. The most severe chronic stress state developed in mice during untreated STZ-DM. Thus, naloxone-precipitated withdrawal in mice seems to be an appropriate model of chronic stress.