Monitoring changes in anti-tuberculosis treatment: associated factors determined at the time of diagnosis.

Objectives: To determine predictive factors for changes in standard anti-tuberculosis chemotherapy at the time of diagnosis.

Methods: A prospective study was performed among tuberculosis (TB) patients treated at specialised centres during 2008-2009. Treatment outcome was monitored per standard guidelines. Treatment was considered successful if the patient was cured or completed treatment. Factors associated with treatment modification were analysed at the bivariate and multivariate levels using logistic regression.

Results: A total of 427 patients were included in the study. The initial standard treatment regimen was retained for 249 patients (58.3%), extended to 9 months for 36 (8.4%) and changed for 142 (33.3%). Factors associated with a change of regimen at the multivariate level were female sex, age ≥ 50 years, human immunodeficiency virus infection, comorbidities, alcoholism, hospitalisation and culture-positive sputum. Drug resistance and toxicity were analysed independently. Treatment outcome was successful in 97.2% of cases without a regimen change and in 87.3% of those with a changed regimen (P < 0.001).

Conclusion: Factors associated with changes in the initial anti-tuberculosis regimen should be considered for rigorous follow-up. Results obtained through individualised treatment provided by specialists were good despite the complexity of the cases treated.