# Real-World Assessment of the Xpert MTB/XDR for Detecting Isoniazid and Second-Line Drug Resistance Among TB Patients

## TL;DR

This study evaluates a new test for detecting drug-resistant tuberculosis, finding it effective for some drugs but less reliable for others due to genetic variations.

## Contribution

The study provides real-world validation of the Xpert MTB/XDR assay's performance for detecting drug resistance in TB patients.

## Key findings

- The Xpert MTB/XDR assay showed high sensitivity for isoniazid and fluoroquinolone resistance detection.
- Sensitivity for amikacin, capreomycin, and ethionamide resistance was lower than manufacturer claims.
- Concordance with whole-genome sequencing was high except for ethionamide.

## Abstract

Rapid and accurate detection of drug-resistant tuberculosis (DR-TB) is critical for effective treatment and containment. The Xpert® MTB/XDR (GXXDR) assay is designed to detect Mycobacterium tuberculosis complex (MTBC) and resistance to isoniazid and second-line anti-TB drugs directly from clinical specimens. We evaluated the clinical performance of GXXDR using 61 MTBC-positive specimens with available phenotypic drug susceptibility testing results. GXXDR results were compared to a phenotypic drug susceptibility test (pDST) and whole-genome sequencing (WGS) to assess sensitivity, specificity, and concordance. Resistance to isoniazid, fluoroquinolones, amikacin, capreomycin, and ethionamide was analyzed. Sensitivity comparisons between GXXDR, WGS, pDST, and manufacturer data were performed using Fisher’s exact and Tango tests. GXXDR demonstrated a high specificity for most drugs and a strong sensitivity for isoniazid (93.8%) and fluoroquinolone (92.3%), consistent with manufacturer reports. In contrast, the sensitivity for amikacin (58.3%), capreomycin (35.7%), and ethionamide (27.3%) was significantly lower than stated by the manufacturer (91.9%, 84.0% and 64.7%, respectively), likely due to resistance mutations outside the assay’s target regions. Sensitivity concordance of GXXDR with WGS was high for all drugs, except ethionamide. The GXXDR assay enables rapid and reliable detection of isoniazid and fluoroquinolone resistance in clinical settings, though sensitivity for certain second-line drugs may be affected by regional genetic diversity. These findings underscore the importance of integrating local epidemiological data to optimize molecular diagnostics for DR-TB.