The tolerability of and adherence to intermittent short-term rifabutin-isoniazid preventive treatment was assessed in subjects dually infected with Mycobacterium tuberculosis and the human immunodeficiency virus (HIV). In a randomised, open-label, phase II pilot study, 44 subjects received either rifabutin 300 mg and isoniazid 750 mg twice weekly for 3 months (group A, n = 16) or the same regimen with rifabutin at 600 mg (group B, n = 14), or isoniazid 300 mg/day for 6 months (group C, n = 14). Three, two and four subjects in groups A, B, and C, respectively, did not complete their treatment (one case of flu-like syndrome in group B; one methadone withdrawal syndrome in group A; and patient decision in two cases in group A and four in group C). Overall, adverse events were reported by four, nine, and seven subjects in groups A, B and C, respectively. Intermittent combined rifabutin + isoniazid for 3 months had lower default rates than daily standard isoniazid for 6 months. The regimen with rifabutin at 300 mg dose compared favourably to standard isoniazid, and warrants larger efficacy studies to assess its role for the prevention of latent tuberculosis in HIV-infected subjects.