Objective: The severity of self-reported withdrawal symptoms varies during detoxification of opioid-dependent patients. The aim of this study is to identify subgroups of withdrawal symptoms within the detoxification trajectory and to predict the severity of withdrawal symptoms on the basis of drug-related and sociodemographic characteristics. Design and setting: A prospective study carried out in an in-patient setting in four addiction treatment centres in the Netherlands. Participants: Two hundred opioid-dependent patients who participated in a randomized controlled trial and completed more than 75 percent of the administrations of the subjective opioid withdrawal scales during rapid detoxification. Intervention and main outcome measure: Main outcome measure was the severity of opioid withdrawal as measured by the subjective opioid withdrawal scale during detoxification (18 measurements). Predictor baseline data were obtained on sociodemographic background, severity of addiction, psychopathology, personality disorder, and craving. Statistics: Those variables found to be statistically significant in univariate analyses were entered into multivariate regression models to predict the severity of subjective withdrawal. Results: No distinct subgroups could be identified despite substantial individual variability throughout the detoxification trajectory. The multiple regression results showed only four variables to predict the severity of withdrawal symptoms: baseline withdrawal symptoms, intravenous heroin use in the last 30 days, anxiety, and cluster C personality disorder. The variance explained by these sociodemographic variables was low while the largest amount of variance was explained by baseline withdrawal symptoms (27 percent). Conclusions: The results of the present study provide evidence that the severity of withdrawal symptoms during detoxification treatment is moderately predicted by the baseline severity of their withdrawal symptoms and not by drug- and patient-related characteristics.