Among 19,237 adults with type 2 diabetes who had isolated bypass surgery, one in four filled no glucose-lowering prescription after discharge. Compared with those patients, mortality was lower with metformin (HR 0.60), GLP-1RA/SGLT2i (HR 0.71) and insulin (HR 0.75). Metformin also beat insulin on mortality (HR 0.61), and was linked to fewer heart failure hospitalizations than no therapy or insulin, and fewer strokes than insulin.
The headline number deserves skepticism. Metformin's 40% mortality reduction exceeds anything seen in randomized trials of diabetes drugs, which is a classic signature of confounding. Patients who fill prescriptions after major surgery are likely healthier, better followed, and more engaged with care, and inverse probability weighting cannot fully erase that healthy-adherer effect. Insulin's apparent inferiority to metformin likely reflects sicker patients being given insulin in the first place. The GLP-1RA/SGLT2i group, only 9.4% of the cohort, produced the least precise comparisons, which is unfortunate because trials show cardiovascular and kidney benefits for these classes. The finding is more useful as a care-gap signal than as a ranking of drugs. A quarter of post-CABG diabetic patients appear untreated, and that gap merits attention. For patients, the practical takeaway is to ask the surgical or primary care team who owns glucose management after discharge. This is confirmatory and hypothesis-generating rather than practice-changing; a randomized comparison of post-CABG regimens is still needed.