In 67 adults with obesity and type 2 diabetes who underwent standardised Roux-en-Y gastric bypass in the UK or Singapore, remission at 12-15 months was predicted by shorter diabetes duration, lower treatment intensity, lower medication-adjusted HbA1c, and stronger post-prandial insulin secretion. Yet remitters and non-remitters improved adjusted HbA1c almost identically (-38.2 vs -39.5 mmol/mol, p = 0.739), a result replicated in a separate audit of 195 UK patients.
The remission-versus-failure framing has long understated what bypass surgery does for people with longer-standing diabetes. The authors' medication-adjusted HbA1c is a clever fix: it back-calculates glycaemia as if the patient were unmedicated, so a non-remitter who drops from insulin to a single tablet no longer registers as having 'failed'. A roughly 39 mmol/mol (about 3.5 percentage point) fall is large by any standard, exceeding what most added glucose-lowering drugs achieve. The practical message is that remission-prediction scores, such as DiaRem, should not be used to gatekeep surgery, because a low predicted chance of remission does not mean a low chance of benefit. Limitations matter, though. The discovery cohort is small, follow-up is only 12-15 months, and the adjusted HbA1c rests on assumed drug potencies. It also says nothing about microvascular or cardiovascular events, where long-term data are what count. Confirmatory rather than paradigm-shifting, but a useful correction to how surgical candidacy is discussed with patients.