Acne vulgaris usually improves with long-term oral antibiotic therapy; however, some patients fail or only partially respond to the commonly prescribed antibiotics. Only a proportion of these patients warrants treatment with isotretinoin suggesting a need for additional therapeutic options. We have therefore performed an open retrospective study of the efficacy of trimethoprim as a third-line antibiotic in the treatment of patients whose acne vulgaris failed to respond to at least 2 courses of antibiotics. A total of 56 patients were reviewed, who had all failed to respond adequately to a minimum of 2 courses of antibiotics. All patients received trimethoprim in a dosage of 300 mg twice daily for at least 4 months, unless it had been withdrawn due to side effects. Topically they were given 1% clindamycin lotion twice daily to the affected areas. The severity of the patient's acne was graded on the face, back and chest. The changes in acne grades were evaluated using non-parametric statistics (Wilcoxon matched pairs). At 4 months there were significant improvements from the grades at initiation in all three sites (p = 0.005 or less). Twenty-one patients remained on the treatment for 8 months, and a significant improvement in the changes of the acne grades remained (p = 0.02 or less). Two patients had the trimethoprim stopped due to side effects. We can therefore recommend this regime as a third-line treatment in the management of acne.