Short-term effect of rosuvastatin versus atorvastatin on the corrected QT interval: a target trial emulation.
To estimate the short-term effect of rosuvastatin versus atorvastatin on the corrected QT interval (QTc) by emulating a published randomized controlled trial (RCT) using electronic health record (EHR) data, and to assess whether target trial emulation (TTE) can replicate RCT findings for a pharmacological safety outcome at substantially greater scale.
Retrospective cohort study emulating a target trial, reported according to the Transparent Reporting of Observational Studies Emulating a Target Trial (TARGET) guideline.
Single tertiary A teaching hospital in China, March 2012 to September 2024.
Of 619,216 cardiology hospitalizations, 165,460 new statin users with suspected coronary artery disease met all eligibility criteria. After 1:1 propensity score matching, 98,860 patients (49,430 per group) constituted the analytic cohort. All standardized mean differences were below 0.013 after matching.
The primary outcome was the change in Fridericia-corrected QT interval (ΔQTcF) from baseline to first follow-up electrocardiogram (24-72 h). Secondary outcomes included newly emerged QTc prolongation, any QTc increase, clinically significant increase (> 30 ms), severe QTc prolongation, and a composite cardiac safety endpoint. Both intention-to-treat and per-protocol effects were estimated.
The mean ΔQTcF in the rosuvastatin group was + 7.71 ms (SD 20.41) versus + 0.31 ms (SD 22.30) in the atorvastatin group, yielding a between-group difference of 7.40 ms (95% CI 7.13 to 7.67; P < 0.001). Newly emerged QTc prolongation occurred in 14.7% versus 10.5% (risk ratio 1.40, 95% CI 1.35 to 1.45). The composite cardiac safety endpoint did not differ (0.4% versus 0.5%; P = 0.24). Results were consistent across eight subgroups and six sensitivity analyses. The TTE estimate was concordant with the published RCT finding of 7.40 ms (heterogeneity P = 1.00), with a 14-fold narrower confidence interval. The negative control outcome analysis showed no residual bias.
Rosuvastatin was associated with a 7.40 ms greater short-term QTcF prolongation than atorvastatin in a cohort 212 times larger than the emulated RCT, without excess clinical cardiac events over a mean follow-up of 48 h. Target trial emulation successfully replicated the RCT finding for a short-term drug safety outcome, demonstrating the framework's value for pharmacovigilance research using routine clinical data.
Retrospective cohort study emulating a target trial, reported according to the Transparent Reporting of Observational Studies Emulating a Target Trial (TARGET) guideline.
Single tertiary A teaching hospital in China, March 2012 to September 2024.
Of 619,216 cardiology hospitalizations, 165,460 new statin users with suspected coronary artery disease met all eligibility criteria. After 1:1 propensity score matching, 98,860 patients (49,430 per group) constituted the analytic cohort. All standardized mean differences were below 0.013 after matching.
The primary outcome was the change in Fridericia-corrected QT interval (ΔQTcF) from baseline to first follow-up electrocardiogram (24-72 h). Secondary outcomes included newly emerged QTc prolongation, any QTc increase, clinically significant increase (> 30 ms), severe QTc prolongation, and a composite cardiac safety endpoint. Both intention-to-treat and per-protocol effects were estimated.
The mean ΔQTcF in the rosuvastatin group was + 7.71 ms (SD 20.41) versus + 0.31 ms (SD 22.30) in the atorvastatin group, yielding a between-group difference of 7.40 ms (95% CI 7.13 to 7.67; P < 0.001). Newly emerged QTc prolongation occurred in 14.7% versus 10.5% (risk ratio 1.40, 95% CI 1.35 to 1.45). The composite cardiac safety endpoint did not differ (0.4% versus 0.5%; P = 0.24). Results were consistent across eight subgroups and six sensitivity analyses. The TTE estimate was concordant with the published RCT finding of 7.40 ms (heterogeneity P = 1.00), with a 14-fold narrower confidence interval. The negative control outcome analysis showed no residual bias.
Rosuvastatin was associated with a 7.40 ms greater short-term QTcF prolongation than atorvastatin in a cohort 212 times larger than the emulated RCT, without excess clinical cardiac events over a mean follow-up of 48 h. Target trial emulation successfully replicated the RCT finding for a short-term drug safety outcome, demonstrating the framework's value for pharmacovigilance research using routine clinical data.