Abstract
BACKGROUND
Treatment of multidrug-resistant or rifampicin-resistant tuberculosis (MDR/RR-TB), although improved in recent years with shorter, more tolerable regimens, remains largely standardised and based on limited drug susceptibility testing (DST). More individualised treatment with expanded DST access is likely to improve patient outcomes.
METHODS
To assess the potential of TB drug resistance prediction based on whole genome sequencing (WGS) to provide more effective treatment regimens, we applied current South African treatment recommendations to a retrospective cohort of MDR/RR-TB patients from Khayelitsha, Cape Town. Routine DST and clinical data were used to retrospectively categorise patients into a recommended regimen, either a standardised short regimen or a longer individualised regimen. Potential regimen changes were then described with the addition of WGS-derived DST.
FINDINGS
WGS data were available for 1274 MDR/RR-TB patient treatment episodes across 2008-2017. Among 834 patients initially eligible for the shorter regimen, 385 (46%) may have benefited from reduced drug dosage or removing ineffective drugs when WGS data were considered. A further 187 (22%) may have benefited from more effective adjusted regimens. Among 440 patients initially eligible for a longer individualised regimen, 153 (35%) could have been switched to the short regimen. Overall, 305 (24%) patients had MDR/RR-TB with second-line TB drug resistance, where the availability of WGS-derived DST would have allowed more effective treatment individualisation.
INTERPRETATION
These data suggest considerable benefits could accrue from routine access to WGS-derived resistance prediction. Advances in culture-free sequencing and expansion of the reference resistance mutation catalogue will increase the utility of WGS resistance prediction.
