In response to the concerns raised by Rathmann and Bernd (1) about our article (2), we appreciate their valuable feedback and wish to address the issues raised.
First, the exclusion criteria for the baseline cohort are based on ICD-10 codes C00 to D49, which encompass a wide spectrum of neoplasms, both benign and malignant, regardless of functional activity. This approach provides a broader population than focusing solely on malignant neoplasms, as cited in the American Cancer Society's lifetime risk estimate for any type of cancer.
Second, the TriNetX U.S. collaborative network is a federated network of different levels of health care organizations across the U.S. During the pandemic, mortality rates in intensive care units (ICUs) and hospitals (3) exhibited a substantial degree of variability, ranging from 6.6% to 80.8% in ICUs and from 0% to 44.4% in hospitals. Our study spans the years 2021–2022 and includes the reported deaths among vaccinated participants with coronavirus disease 2019 (COVID-19) infection (292 out of 178,203, ∼0.16%) and unvaccinated participants (479 out of 212,093, ∼0.26%). Although the levels we used are higher than the levels of COVID-19–associated deaths reported in Morbidity and Mortality Weekly Report for the years 2021 and 2022 (115.6 and 61.3 per 100,000 standard population, respectively, for the general population), the levels fall within reasonable bounds.
Third, we wish to assure the authors that all HbA1c level cutoffs for diabetes exclusion adhered to the American Diabetes Association guidelines, specifically that for using a threshold of 6.5% or higher.
Fourth, regarding the interpretation of the odds ratio from the study of Kwan et al. (4), we acknowledge that it pertains to the vaccination status before infection as a secondary exposure and is not directly analogous to the primary exposure (new diagnosis of diabetes) and the subsequent outcome (new diagnosis of diabetes following COVID-19 infection). Nevertheless, our study employs a comprehensive cohort design that is distinct from a self-controlled exposure-crossover design, which is known to be susceptible to exposure-related trends (5). Given the evolving nature of the COVID-19 pandemic, cohort studies provide a more comprehensive perspective.
Fifth, for ease of comparison with studies not reporting composite outcomes of type 2 diabetes mellitus (T2DM) and death, we offer the following concise results.
Without accounting for death as a competing risk, the vaccinated group exhibits an 18% lower risk (hazard ratio [HR] 0.82, 95% CI 0.75–0.90) of developing new-onset T2DM by diagnosis, a 16% lower risk (HR 0.836, 95% CI 0.78–0.90) of using antihyperglycemic drugs, and a 19% lower risk (HR 0.81, 95% CI 0.76–0.86) of experiencing composite outcomes comprising T2DM diagnosis and using antihyperglycemic drugs compared with the unvaccinated group. These results align with findings reported by Xiong et al. (6) for recipients of the BNT162b2 vaccine in the Hong Kong cohort.
Last, we thank the authors for pointing out the incorrect wording of “prosperity” rather than “propensity” in the supplementary material. We have requested that an erratum be issued.
In conclusion, we appreciate the author’s feedback and clarification of our study's objectives. We did not claim to be the first to identify the protective effects of COVID-19 vaccination against new-onset T2DM but rather emphasized our unique dose-response analysis. We welcome further discussion and research on this critical topic.
Article Information
Funding. The source of funding for this article was the Kaohsiung Veterans General Hospital (KSVGH112-122).
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
Handling Editors. The journal editor responsible for overseeing the review of the manuscript was Steven E. Kahn.