Nearly Half the People in a Trial Put Their Type 2 Diabetes Into Remission. The Follow-Up Is the Interesting Part

On this account, when fat accumulates in the liver beyond what that individual can tolerate, the liver overproduces glucose and exports fat to the pancreas. Fat in the pancreas impairs the beta cells' ability to release insulin normally. The two feed each other.

Crucially, the threshold is personal. Some people can carry a great deal of weight without crossing it, and others cross it at a body mass index that looks unremarkable, which is part of why the usual BMI cut-offs mislead in this region. Remove enough fat from liver and pancreas and beta cell function can recover, which is the mechanism DiRECT appears to have exploited.

It also predicts who will not respond, and that prediction holds up: remission is much less likely the longer someone has had diabetes, because beta cell capacity that has been lost for a decade does not come back.

What happened at five years?

The extension study is where the honest version of this story lives.

Of those in remission at the end of year two, 26 per cent remained in remission at five years. That is a real and durable result for a quarter of them, achieved without medication, and it is far better than the historical assumption of inevitable progression.

It also means around three quarters did not sustain it. The difficulty is not achieving remission, which the trial showed is very achievable. It is maintaining a substantial weight loss for years in a food environment engineered to prevent exactly that.

Reading the five-year data as failure is the wrong conclusion. Long-term weight maintenance is hard for everyone, and the comparison should not be with a hypothetical perfect outcome but with the alternative, which was escalating medication and progressive complications.

Who should consider this?

The best candidates are reasonably clear from the evidence.

Remission is most likely in people diagnosed within the last few years, who have meaningful weight to lose, whose beta cells still have reserve, and who are not on insulin. Someone with twenty years of type 2 diabetes and long-standing insulin therapy is very unlikely to achieve it, and should not be set up for a failure that will be read as personal.

Two warnings belong here. This should not be attempted unsupervised while taking glucose-lowering medication, because a sharp calorie reduction alongside insulin or a sulfonylurea is a direct route to serious hypoglycaemia. Medication has to be adjusted as part of the plan.

And remission does not end surveillance. Eyes, feet and kidneys still need their annual checks, because the years spent above target already happened and their consequences do not go into remission alongside the glucose.