OBJECTIVE: To characterize dosages and frequencies of infliximab in a publicly financed health-care system without administrative restrictions on the use of biologics. METHODS: Data from the Stockholm TNFalpha follow-up registry (STURE) were used. Dosages and frequencies of infliximab were adjusted when clinically indicated. Kaplan-Meier survival function was calculated for maintenance of initial infliximab dosage, initial frequency, and both combined; the influence of baseline characteristics on the adjustments and the effects of these adjustments on clinical outcomes were analysed. All analyses were for patients who continued on therapy only. RESULTS: The 1-year survival on initial infliximab frequency was 96.3+/-1.8%, on initial infliximab dose 70.5+/-4.3%, and on the two combined 68.7+/-4.3% (cumulative survival estimate +/- SE). At 2 years, the corresponding values were 88.1+/-3.4%, 53.4+/-5.1%, and 50.3+/-5.1%, and after 3 years 77.4+/-4.9%, 45.9+/-5.4%, and 41.5+/-5.3%. A higher number of prior disease-modifying anti-rheumatic drugs (DMARDs), use of glucocorticoids, and lower baseline swollen joints counts (SJCs) predicted greater likelihood of later adjustments. Efficacy of dosage increases was modest, as reported previously, while frequency increases appeared more efficacious. CONCLUSION: Infliximab dosages and dosing frequencies are increased frequently in clinical practice: after 3 years, only 42% of patients continued on the original schedule. While frequency increases appear to result in better effect persistence between infusions, the gains from dosage increases are small and may not be better than chance. These data suggest that the total amounts of infliximab needed to obtain satisfactory disease control are greater than the amounts suggested by the original dosing recommendations, and may have bearing on pharmaco-economic issues pertaining to treatment with infliximab and other therapeutic agents.