Risk factors for extensive drug resistance in multidrug-resistant tuberculosis cases: a case-case study

Settings Identification of extensively drug-resistant tuberculosis (XDR-TB) may be delayed because of the lack of availability of molecular testing for second-line drugs (SLDs). Early suspicion of XDR-TB is therefore necessary to avoid developing further drug resistance. Objective To identify the characteristics associated with XDR-TB among multidrug-resistant TB (MDR-TB) cases before the availability of second-line drug susceptibility testing (DST) results. Methods All MDR-TB cases with available second-line DST results recorded in France from 1998 to 2013 were classified as simple MDR-TB (no resistance to fluoroquinolones [FQs] or second-line injectable drugs [SLIDs]), pre-XDR-TB (resistance to FQs or SLIDs) and XDR-TB cases (resistance to both). Results A total of 833 MDR-TB cases were analysed, including 168 (20%) pre-XDR and 62 (7%) XDR-TB cases. A previous history of treatment was acknowledged among 41% of the cases; 12% were human immunodeficiency virus-positive. Characteristics independently associated with XDR-TB were foreign birth (OR 9.5), previous anti-tuberculosis treatment (OR 2.6), smear positivity (OR 4.5) and ethambutol (EMB) resistance (OR 9.1). Characteristics independently associated with pre-XDR-TB compared to simple MDR-TB cases were male sex (OR 1.6), birth in Europe (OR 2.6) and EMB resistance (OR 1.9). Conclusion The presence of clinical or bacteriological characteristics associated with XDR-TB should lead to rapid molecular testing for resistance to SLDs before starting tailored treatment.