RESUMO
Importance: Rifampin-resistant tuberculosis treatment regimens require electrocardiographic (ECG) monitoring due to the use of multiple QTc-prolonging agents. Formal 12-lead ECG devices represent a significant burden in resource-constrained clinics worldwide and a potential barrier to treatment scale-up in some settings. Objective: To evaluate the diagnostic accuracy of a handheld 6-lead ECG device within resource-constrained clinics. Design, Setting, and Participants: This diagnostic study was performed within a multicenter, pragmatic (broad eligibility criteria with no exclusions for randomized participants), phase 3 rifampin-resistant tuberculosis treatment trial (BEAT Tuberculosis [Building Evidence for Advancing New Treatment for Tuberculosis]) in South Africa. A total of 192 consecutive trial participants were assessed, and 191 were recruited for this substudy between January 21, 2021, and March 27, 2023. A low proportion (3 of 432 [0.7%]) of all screened trial participants were excluded due to a QTc interval greater than 450 milliseconds. Triplicate reference standard 12-lead ECG results were human calibrated with readers blinded to 6-lead ECG results. Main Outcomes and Measures: Diagnostic accuracy, repeatability, and feasibility of a 6-lead ECG device. Results: A total of 191 participants (median age, 36 years [IQR, 28-45 years]; 81 female participants [42.4%]; 91 participants [47.6%] living with HIV) with a median of 4 clinic visits (IQR, 3-4 visits) contributed 2070 and 2015 12-lead and 6-lead ECG assessments, respectively. Across 170 participants attending 489 total clinic visits where valid triplicate QTc measurements were available for both devices, the mean 12-lead QTc measurement was 418 milliseconds (range, 321-519 milliseconds), and the mean 6-lead QTc measurement was 422 milliseconds (range, 288-574 milliseconds; proportion of variation explained, R2 = 0.4; P < .001). At a QTc interval threshold of 500 milliseconds, the 6-lead ECG device had a negative predictive value of 99.8% (95% CI, 98.8%-99.9%) and a positive predictive value of 16.7% (95% CI, 0.4%-64.1%). The normal expected range of within-individual variability of the 6-lead ECG device was high (±50.2 milliseconds [coefficient of variation, 6.0%]) relative to the 12-lead ECG device (±22.0 milliseconds [coefficient of variation, 2.7%]). The mean (SD) increase in the 12-lead QTc measurement during treatment was 10.1 (25.8) milliseconds, with 0.8% of clinic visits (4 of 489) having a QTc interval of 500 milliseconds or more. Conclusions and Relevance: This study suggests that simplified, handheld 6-lead ECG devices are effective triage tests that could reduce the need to perform 12-lead ECG monitoring in resource-constrained settings.
