Abstract Aims Two post-authorisation studies assessed the safety and persistence of patients’ use of nalmefene. Methods The START study (EUPAS5678) was a non-interventional, multi-country, prospective, 18-month (8 follow-up visits) cohort study including outpatients initiating nalmefene for the first time. The multi-database retrospective cohort study (MDRC, EUPAS14083) included baseline and follow-up data from German, Swedish and UK healthcare databases. Both studies permitted ‘all comers’ without explicit exclusion criteria; predefined subgroups of interest included the elderly (≥65 years) as well as patients with significant psychiatric and/or somatic comorbidities. Results START study: Overall, the mean duration of nalmefene treatment was 10.3 ± 7.3 months (N = 1348), with 49.0% of patients treated for ≥1 year; frequent reasons for treatment discontinuation were ‘goal reached’ and ‘drug cost’. The most frequently reported adverse drug reactions (ADRs) were nausea (4.7%), dizziness (3.2%) and insomnia (2.0%). ADR rates appeared higher in the elderly subpopulation (18.6% reported ≥1 ADR vs. 12.0% in the total population) but were not higher in the other predefined subgroups. MDRC study: The database follow-up analysis followed 2892 patients over 18 months for whom the duration of nalmefene treatment was between 2 and 3 months and <5% of patients used nalmefene for ≥1 year. Conclusions Despite the inclusion of a wider patient population (e.g. elderly patients and those with relevant co-morbidities), the safety and tolerability profile of nalmefene given in routine practice was consistent with previous clinical studies. The differing rates of persistence beyond 1 year likely reflect the different methodologies and highlight the relevance of psychosocial support at follow-up visits. Short Summary: Two studies were used to evaluate the safety and persistence of nalmefene in the routine management of alcohol dependence in Europe as part of the overall risk management plan. Although adverse events appeared more common in over 65-year olds, there were no major differences in safety outcomes between the total population, and those with psychiatric and somatic comorbidities or those with a history of seizures. Findings were in line with those from previous clinical studies. The differing rates of persistence beyond 1 year (49% in the prospective study and < 5% in the multi-database retrospective cohort study) likely reflect the different study methodologies and serve to highlight the relevance of follow-up support.