Is rifampin resistance a reliable predictive marker of multidrug-resistant tuberculosis in China: A meta-analysis of findings

Abstract

Systematic review of multidrug-resistant tuberculosis (MDR-TB) prevalence among rifampicin (RIF)-resistant tuberculosis (RR-TB) patients in 34 provinces of China was conducted to correlate RIF resistance with concurrent isoniazid (INH) resistance. Database searches (PubMed, Embase, China National Knowledge Infrastructure, Chinese Scientific Journal, Wanfang), identified drug resistance surveillance studies conducted between January 1, 2000 and June 30, 2018. Of 1554 records, random-effects meta-analysis of 34 studies of adequate methodological quality yielded 108,366 TB cases for MDR-TB prevalence analysis of RR-TB cases. MDR-TB prevalence among RR-TB cases varied from 57% (Xinjiang; 95% CI 47%, 67%) to 95% (Taiwan; 95% CI 92%, 98%), for a pooled national rate of 77% (95% CI 75%, 80%). Subgroup and meta-regression analyses revealed greater MDR-TB prevalence in previously treated versus new RR-TB cases (P < 0.001), with no significant differences of regional initial drug resistance rates or sampling methods. Regional MDR-TB prevalence among RR-TB cases was lowest (69%) in the Northeast Region (95% CI 65%, 73%) and highest (90%) in Hong Kong, Macao and Taiwan (95% CI 81%, 98%). In China, ∼77% of RR-TB cases are MDR-TB. Thus, RIF resistance cannot effectively predict MDR-TB. Highly variable RR-TB prevalence across China warrants improved TB management.