Agreement: I Agree Body: Dear Editor We read with interest the cohort study by Krüger et al estimating cardiovascular effects of tirzepatide versus sitagliptin using a target trial emulation framework benchmarked against randomized trials.1 While the study's pre-registration, transparency, and overlap weighting are commendable, several methodological features warrant caution when interpreting the reported hazard ratios. First, the analysis adopted an on-treatment (per-protocol) causal contrast with a median follow-up of only 5–6 months, yet estimated 1-year risks.1 Supplemental Table S6 reveals that treatment discontinuation or switching accounted for 37.5% of censoring in the tirzepatide group versus 41.6% in the sitagliptin group, with differential rates of nursing home admission or switching to other glucagon-like peptide-1 receptor agonists (5.9% vs 10.3%).1 These imbalances raise the specter of informative censoring—if healthier patients persist on tirzepatide while frailer patients discontinue sitagliptin, the on-treatment analysis may artifactually favor tirzepatide. The authors applied overlap weighting to address baseline confounding, but inverse probability of censoring weighting (IPCW) was not reported to address differential loss to follow-up. Second, although the baseline characteristic tables are comprehensive (Supplemental Table S3),1 multimorbidity and polypharmacy were modeled as independent binary variables. In real-world practice, patients with type 2 diabetes and atherosclerotic cardiovascular disease typically present with clustered comorbidities (e.g., chronic kidney disease plus heart failure plus hypertension) and complex glucose-lowering regimens. Modeling each drug and disease as a separate covariate assumes additive, non-interacting effects that may not reflect biological reality. Third, the early separation of cumulative incidence curves within 3 months and the magnitude of the all-cause mortality reduction (hazard ratio 0.55)1 are striking for a drug class whose atherosclerotic benefits typically accrue over longer intervals. Landmark analyses at 3 and 6 months, with re-assessment of baseline characteristics at each time point, would help disentangle true biological effects from channeling bias and immortal time phenomena. We encourage the authors to consider sensitivity analyses using IPCW for on-treatment censoring and landmark approaches to validate these important findings. References [1] Krüger N, Schneeweiss S, Wang SV. Tirzepatide and the risk of atherosclerotic cardiovascular events. BMJ. 2026;394:e100011. [2] Hernán MA, Robins JM. Causal Inference: What If. Boca Raton, FL: Chapman & Hall/CRC; 2020. [3] Robins JM, Finkelstein DM. Correcting for noncompliance and dependent censoring in an AIDS clinical trial with inverse probability of censoring weighted (IPCW) log-rank tests. Biometrics. 2000;56(3):779-788. [4] Thomas LE, Li F, Pencina MJ. Overlap weighting: a propensity score method that mimics attributes of a randomized clinical trial. JAMA. 2020;323(24):2417-2418. [5] Anderson JR, Cain KC, Gelber RD. Analysis of survival by tumor response. J Clin Oncol. 1983;1(11):710-719. Conflict of Interest Disclosures: None reported. Funding/Support: None reported. Use of AI, LLM, or Chatbots: The authors confirm that no artificial intelligence, large language models, or chatbots were used in the drafting of this letter. No competing Interests: Yes The following competing Interests: Electronic Publication Date: Friday, August 21, 2026 - 04:45 AI use: No, I have not used AI Highwire Comment Subject: Tirzepatide and the risk of atherosclerotic cardiovascular events: population based cohort study Workflow State: Released Full Title: Methodological Considerations in Trial-Emulated Observational Studies: On-Treatment Censoring, Polypharmacy, and the Limits of Short-Term Follow-Up Highwire Comment Response to: Tirzepatide and the risk of atherosclerotic cardiovascular events: population based cohort study Check this box if you would like your letter to appear anonymously:: Last Name: Tuersun First name and middle initial: Adili Email: 25111030024@m.fudan.edu.cn Address: Shanghai 201203, PR China Occupation: PhD Affiliation: School of Pharmaceutical Sciences, Fudan University BMJ: Additional Article Info: Rapid response