Background: Recent World Health Organization guidelines recommend the addition of ethambutol (EMB) throughout standardised treatment of new cases of active tuberculosis (TB) in populations with increased prevalence of isoniazid resistance to reduce the risk of creating multidrug resistance. This could expose patients to the risk of blindness.
Methods: We searched Cochrane, Embase and PubMed electronic databases from 1965 to February 2011 for original studies that prospectively followed all patients treated with EMB for active TB, and routinely ascertained the occurrence of visual toxicity using standard methods. Pooled estimates, overall and stratified by major covariates, were calculated using random effects meta-analysis.
Results: Pooled cumulative incidence of any visual impairment in all patients was 22.5 per 1000 persons treated with EMB (95%CI 10.2-35), and permanent impairment was 4.3/1000 (95%CI 0.3-9.0). After restricting the analyses to arms in which the average dose was 27.5 mg/kg/day or less and treatment was for 2-9 months, the incidence of any visual impairment was 19.2/1000 (95%CI 5.8-33), and permanent impairment was 2.3/1000 persons (95%CI 0-6.1) treated, as the majority of episodes were reversible. In reversible cases, resolution of impairment occurred after an average of 3 months.
Conclusions: In this review, any visual impairment occurred in 22.5/1000 persons receiving EMB at standard doses for up to 9 months, and permanent impairment in 2.3/1000-an important risk. However, these estimates are imprecise, and the studies were of variable quality and the results heterogeneous. Well-designed prospective studies with repeated measurements of multiple visual parameters that clearly describe the degree of permanent impairment are needed.