BACKGROUND: Pathological gambling (PG) is a relatively common and highly disabling impulse control disorder. A range of psychotherapeutic agents including selective serotonin reuptake inhibitors, antiepileptic drugs, and opioid antagonists are shown to be effective in the short-term treatment of PG. The use of a wide range of pharmacological treatments for PG is consistent with the observation that PG shares features of obsessive-compulsive spectrum disorders, impulse control disorders, and addictive disorders. The aim of the study is to assess the rate of relapse in treatment-responder pathological gamblers after discontinuation of the active treatment. METHODS: Our study sample was composed of 43 male pathological gamblers who had been full responders to 1 of 4 drug treatment regimens (fluvoxamine, topiramate, bupropion SR, or naltrexone) from several previous acute open-label (12-week) comparison studies. Full response was defined as the absence of gambling for a 1-month duration together with improvement on the Clinical Global Improvement scale. The 43 full responders were then followed prospectively for an additional 9 months, which included a 3-month open-label continuation phase and a 6-month medication-free follow-up phase. Follow-up visits were performed on a monthly basis throughout the duration of study. At every follow-up visit, a comprehensive psychiatric diagnostic evaluation was performed on all patients, and patients were assessed for symptoms of gambling using a self-report instrument and collateral family reports. The Clinical Global Impression Improvement scale was also administered at every follow-up visit. Raters were blind to the previous drug treatment. RESULTS: Most patients did not relapse during the 6-month medication-free follow-up phase. Three of 6 patients with fluvoxamine, 3 of 9 with topiramate, 7 of 18 with bupropion SR, and 4 of 10 with naltrexone relapsed. Relapse was strictly defined as gambling behavior at any time during the 6-month medication-free follow-up period. Most of the patients did not gamble during the follow-up period, and the patients that did gamble reported a decrease in gambling losses. CONCLUSIONS: This naturalistic long-term follow-up outcome study demonstrates that among pathological gamblers who respond to a 6-month trial of medication, most patients seem to maintain full-response during a 6-month medication-free follow-up phase. Further studies are needed to confirm our findings.