Genome sequencing predicts isoniazid resistance in New York tuberculosis strains

Tuberculosis remains a major global health challenge, and Isoniazid (INH) is one of the most important antibiotics used both to treat active disease and to prevent it. Resistance to INH is the single most common type of drug resistance in Mycobacterium tuberculosis complex (MTBC), and that resistance can come from changes in a number of different genes and regions. To understand how much of INH resistance can be explained by genetic changes and to see whether a genomic approach could replace or reduce laborious lab testing, a team including corresponding author Joseph Shea carried out a six-year, two-phase study. The researchers examined a total of 3,696 MTBC strains collected in New York State. In phase 1 they compared genetic testing and phenotypic testing side-by-side, and in phase 2 they used what they learned to try a new molecular testing algorithm intended to reduce the need for conventional phenotypic drug susceptibility testing (DST). The goal was to see if whole genome sequencing (WGS) could predict which strains were INH resistant quickly, accurately and at lower cost.

The study used paired testing and clear benchmarks to measure how well whole genome sequencing (WGS) worked. In phase 1 the team analyzed 1,767 strains with both phenotypic drug susceptibility testing (DST) and genotypic DST, comparing the results directly. WGS proved highly accurate: sensitivity for detecting INH resistance was 90.3%, specificity was 99.8%, and the negative predictive value for INH susceptibility was 98.8%. Based on those results the investigators developed a molecular testing algorithm that reduced the use of phenotypic DST and then applied that algorithm in phase 2 to another 1,929 MTBC strains. Across all 3,696 isolates, the prevalence of INH resistance in New York was 10.2%. WGS predicted 337 isolates to be INH resistant. Of 41 additional strains that showed phenotypic INH resistance, 38 were found to have mutations in genes known to be associated with INH resistance, including katG, inhA, mabA, mabA-inhA and the oxyR-ahpC intergenic region, showing that known loci explain most resistance detected.

These findings show that WGS can serve as a reliable molecular tool to predict Isoniazid (INH) drug resistance in MTBC and that most INH resistance has a molecular basis in known resistance loci. For public health laboratories and clinicians, this means WGS can be used to streamline testing workflows, reduce the number of slower phenotypic DSTs required, cut costs, and shorten turnaround time (TAT) for reporting resistance. Faster, accurate detection of INH resistance can inform treatment and prophylaxis decisions sooner, which is important because delayed recognition of resistance can lead to ineffective therapy. The study’s results—high sensitivity, very high specificity, and strong negative predictive value—support broader adoption of WGS-based approaches in settings where sequencing capacity exists, while still recognizing that a small fraction of phenotypic resistance may arise from mutations not yet fully characterized.