Treatment of isoniazid-resistant tuberculosis with first-line drugs: a systematic review and meta-analysis

Background The results of some reports have suggested that treatment of isoniazid-resistant tuberculosis with the recommended regimens of first-line drugs might be suboptimal. We updated a previous systematic review of treatment outcomes associated with use of first-line drugs in patients with tuberculosis resistant to isoniazid but not rifampicin. Methods In this systematic review, we updated the results of a previous review to include randomised trials and cohort studies published in English, French, or Spanish to March 31, 2015, containing results of standardised treatment of patients with bacteriologically confirmed isoniazid-resistant tuberculosis (but not multidrug-resistant tuberculosis ie, not resistant to rifampicin) in whom failure and relapse were bacteriologically confirmed. Results in patients with drug-sensitive tuberculosis included in the same studies were also analysed. We pooled treatment outcomes with random-effects meta-analysis. Findings We identified 19 cohort studies and 33 trials with 3744 patients with isoniazid-resistant tuberculosis and 19 012 patients with drug-sensitive disease. The pooled rates of failure or relapse, or both, and acquired drug resistance with all drug regimens were 15% (95% CI 12-18) and 3.6% (2-5), respectively, in patients with isoniazid-resistant tuberculosis and 4% (3-5) and 0.6% (0.3-0.9) in those with drug-sensitive tuberculosis. Of patients with initial isoniazid-resistant tuberculosis with acquired drug resistance, 96% (93-99) had acquired multidrug-resistant disease. Treatment of isoniazid-resistant tuberculosis with the WHO standard regimen for new patients resulted in treatment failure, relapse, and acquired multidrug resistance in 11% (6-17), 10% (5-15) and 8% (3-13), respectively; treatment with the standard WHO regimen for previously treated patients resulted in treatment failure in 6% (2-10), relapse in 5% (2-8), and acquisition of multidrug resistance in 3% (0-6). For patients with drug-sensitive disease treated with the standard retreatment regimen the rates were 1% (0-2), 5% (4-7), and 0.3% (0-0.6). Interpretation Treatment of isoniazid-resistant tuberculosis with first-line drugs resulted in suboptimal outcomes, supporting the need for better regimens. Standardised empirical treatment of new cases could be contributing substantially to the multidrug-resistant epidemic, particularly in settings where the prevalence of isoniazid resistance is high.