Hsieh et al. (3) reported that after a severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, those with coronavirus disease 2019 (COVID-19) vaccination had a 21% lower risk for diabetes or death than unvaccinated patients. They stated that this was the first study on protective effects of COVID-19 vaccination on diabetes risk. However, important previous studies are discussed only marginally. Furthermore, the current investigation has severe limitations.
Not mentioned are the results of two studies in which the effect estimates suggest a lower risk of diabetes after vaccination, although their authors avoid claiming a causal relationship (1,2). In a representative panel of primary care physicians in Germany, diabetes incidence rates were higher 183 days before than 183 days after vaccination (incidence rate ratio 0.79; 95% CI 0.75–0.83) in people receiving their first COVID-19 vaccines (1). In a population-based cohort study in Hong Kong, results depended on the vaccine used, and diabetes risk was reduced after BNT162b2 vaccination (hazard ratio 0.862; 95% CI 0.828–0.897) only (2). In line with these findings, in a center-based self-controlled exposure-crossover study, the odds of new diabetes diagnoses were lower in vaccinated than in unvaccinated patients (4). However, Hsieh et al. (3) take an odds ratio from this article that they obviously misinterpreted. This odds ratio (0.63; 95% CI 0.47–0.85) for the vaccination status refers to an adjustment variable, and the outcome is incident diabetes after versus before infection, not the development of new-onset diabetes. Taken together, these four studies indicate a protective association of COVID-19 vaccination with diabetes risk.
Potential methodological problems limit the evidence derived from the current study (3). First, the American Cancer Society states that the lifetime risk for any cancer is 40.9% in men and 39.1% in women (https://www.cancer.org/). In the current study, 40% of vaccinated patients and 41% of unvaccinated patients were excluded due to cancers (3). Given a mean age of 48 years for vaccinated and 44 years for unvaccinated participants, this high cancer frequency raises doubt. Second, deaths of 292 vaccinated and 479 unvaccinated participants were reported before index dates, which further raises doubt about data validity. Third, the HbA1c cutoff used for diabetes exclusion was >6.5%, which is not similar to the American Diabetes Association guideline. Fourth, in Table 2 and the online supplementary material, the authors mistakenly use the term “prosperity score” instead of “propensity score.” Fifth, the authors use a composite outcome of diabetes (or antihyperglycemic drugs) and death but never diabetes (or antihyperglycemic drugs) alone, making comparisons with other studies difficult.
The current study makes an important contribution, because the observed dose-response relationship between number of vaccinations and higher diabetes risk is a criterion for a causal association. However, effects of COVID-19 vaccines on pathophysiological processes, particularly insulin sensitivity and secretion and inflammatory markers, should be investigated in experimental studies before causality can be assumed (5). Finally, a problem of observational studies is that vaccinated and unvaccinated people strongly differ in characteristics, e.g., less trust in the medical system, less fear of COVID-19, or less external locus of control (2), which cannot be assessed in routine data.
Article Information
Funding. The German Diabetes Center is funded by the German Federal Ministry of Health, the Ministry of Culture and Science of the State of North Rhine-Westphalia, and grants from the German Federal Ministry of Education and Research to the German Center for Diabetes Research (DZD).
Duality of Interest. W.R. reports the receipt of consulting fees for attending educational sessions by Novo Nordisk outside the topic of the current work. No other 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.