Among 14,215 adults with type 2 diabetes and COVID-19 (mean age 60, mean BMI 37), prescription of GLP-1 receptor agonist-based therapies was associated with a 29% reduction in 12-month mortality compared to SGLT2 inhibitors (adjusted risk ratio 0.71; 95% CI 0.53–0.95). However, no meaningful difference emerged for Long COVID diagnosis (aRR 1.01) or probable Long COVID via computational phenotype (aRR 0.94), suggesting the survival benefit does not extend to persistent post-acute sequelae.
This preprint — not yet peer-reviewed — adds meaningful real-world signal to the rapidly expanding case for GLP-1 agonists beyond glycemic control. The active-comparator design against SGLT2 inhibitors (themselves cardio- and reno-protective) is methodologically stronger than placebo comparisons, making the mortality signal harder to dismiss as simple confounding by indication. Still, residual confounding remains plausible: clinicians may preferentially prescribe GLP-1s to patients with better metabolic trajectories. The retrospective EHR design cannot establish causality, and the cohort spans only one pandemic phase (October 2021–April 2023), predating dominant Omicron subvariants.
The null Long COVID finding is notable and somewhat surprising given GLP-1s' anti-inflammatory profile — it may reflect diagnostic heterogeneity in Long COVID phenotyping rather than a true biological null. For clinicians managing obese diabetic patients post-COVID, these results are incrementally encouraging on survival, but do not yet justify GLP-1 prescribing specifically for Long COVID prevention.