In a retrospective cohort of 2,056 adults with type 2 diabetes who reached at least 90 days medication-free after sleeve gastrectomy or Roux-en-Y gastric bypass, 882 (42.9%) restarted diabetes medication, a median of 1.04 years later. Longer diabetes duration (HR 1.22 per SD), preoperative insulin use (HR 1.63) and higher baseline HbA1c (HR 1.09 per SD) raised the risk, while gastric bypass lowered it (HR 0.78). Postoperative weight slope added little (C-index 0.62), and HbA1c slope improved fit without meaningfully improving discrimination.
The practical message is that the window for durable remission is largely set before the operation. This fits earlier work, including DiaRem-style scores and long-term Swedish and Cleveland Clinic data, showing that shorter disease duration and no insulin dependence predict lasting remission, and that bypass outperforms sleeve. The finding that weight trajectory adds little is notable: patients cannot count on weight loss alone to protect against relapse if beta-cell reserve is already depleted. The implication for adults with type 2 diabetes is that earlier surgical referral, when pancreatic function is better preserved, likely buys more durable benefit. Medication re-initiation is also not equivalent to failure, since restarted regimens are often lighter than baseline. Limits are real: this is single-health-system observational data, restart reflects prescribing behaviour as well as biology, and unmeasured factors such as C-peptide and diet were absent. Confirmatory rather than paradigm-shifting, but a useful reminder that early intervention matters.