Monitoring heart risk in rifampicin-resistant TB treatments

Treating rifampicin-resistant tuberculosis (RR-TB) often requires medicines that can lengthen the heart’s electrical recovery time, measured as the QT interval. Excessive QT prolongation can raise the risk of dangerous heart rhythms, but there has been limited evidence about which patients are most likely to develop this problem and how closely they need to be monitored. To address that gap, researchers working on the STREAM Stage 2 trial carried out a post-hoc analysis of trial data to look for risk factors for marked QT or QTcF interval prolongation and to test a practical monitoring approach. The corresponding author on the study was Gareth Hughes. Using data already collected in the trial, the team examined who developed a QT or QTcF interval of 500 ms or more, and evaluated whether a specific monitoring strategy could accurately flag people at higher risk, with the goal of informing safer treatment plans for people on regimens for RR-TB.

The analysis included participants allocated to three treatment regimens: 9-month control (n=202), 9-month oral (n=211) and 6-month (n=143). The investigators assessed risk factors for development of a QT or QTcF interval ≥500ms and tested the diagnostic accuracy of a monitoring strategy for QT/QTcF prolongation. QT/QTcF ≥500ms occurred on all regimens: 9-month control (n=14 (6.9%)), 9-month oral (n=8 (3.8%)) and 6-month (n=6 (4.2%)). The participating country with the highest number of QT/QTcF interval ≥500ms events was Mongolia (18 events [64%]). A higher baseline QTcF was significantly associated with development of QT/QTcF ≥500ms (OR 1.05; 95% CI 1.03 to 1.07, p<0.001). There was a suggested association between moxifloxacin (compared to levofloxacin, OR 2.49; 95% CI 0.92 to 6.71, p = 0.07) and a higher baseline TSH (OR 3.52; 95% CI 0.84 to 14.73, p=0.08) and increased odds of QT/QTcF ≥500ms. The monitoring strategy performed well in the control (sensitivity 100%; specificity 62%; positive predictive value (PPV) 13% and negative predictive value (NPV) 100%) and oral (sensitivity 100%, specificity 59%, PPV 6% and NPV 100%) regimen groups.

These findings point to two practical insights. First, baseline QTcF measurement was a clear predictor of who later developed marked prolongation, suggesting that a simple ECG before or early in treatment can identify people at higher risk. Second, the country-level result, with Mongolia contributing a large share of events, indicates that local factors may influence risk and that monitoring plans might need to be tailored by setting. The suggested link between moxifloxacin (versus levofloxacin) and higher baseline TSH with QT/QTcF ≥500ms are signals that warrant attention, though they were not definitive in this analysis. The tested monitoring strategy showed excellent sensitivity and NPV, meaning it was good at catching those who would develop significant QT prolongation and at reassuring clinicians when risk was low; its low PPV means many people flagged will not go on to reach ≥500ms, so follow-up testing or clinical judgment remains important. Overall, the results support targeted ECG monitoring based on baseline QTcF and local context to improve safety in RR-TB treatment.