DiRECT: remission had a definition, not a guarantee
A primary-care programme tested weight management and medication-free remission at one year.

Illustrative food image, not a photograph of the study intervention.
Who entered the programme
DiRECT tested an intensive weight-management programme against guideline-based care in 49 general practices in Scotland and the Tyneside region of England. Practices, rather than individual patients, were randomly assigned. Eligible participants were aged 20–65 years, had type 2 diabetes diagnosed within the preceding six years, had a body mass index of 27–45, and were not receiving insulin. The intention-to-treat population included 149 people in each group. These entry criteria matter: this was not a trial of everyone living with diabetes. Its results should not automatically be extended to people with longer-standing disease, insulin treatment or substantially different starting characteristics.
Lean ME, Leslie WS, Barnes AC, Brosnahan N et al.. Lancet (London, England), 2018. PMID 29221645 · DOI
The intervention was a supervised package
The intervention combined a formula total-diet-replacement phase, stepped food reintroduction and structured support for maintaining weight loss. The formula phase supplied 825–853 kcal daily for three to five months; food reintroduction lasted two to eight weeks. Withdrawal of diabetes and blood-pressure medicines was part of the study intervention, not advice for readers to stop treatment themselves. These details distinguish a managed clinical programme from an ordinary collection of dinner recipes. DiRECT did not test a vegan or NANO menu against another menu. Its randomised comparison concerned the full weight-management package delivered through primary care, rather than any individual plant food.
Lean ME, Leslie WS, Barnes AC, Brosnahan N et al.. Lancet (London, England), 2018. PMID 29221645 · DOI
Read the remission endpoint precisely
At 12 months, 68 participants in the intervention group, or 46%, met the remission endpoint, compared with six, or 4%, in the control group. The trial defined remission as HbA1c below 6.5% after at least two months without all antidiabetic medicines. It also assessed weight loss of at least 15 kg, achieved by 36 intervention participants, or 24%, and none of the controls. Remission here is a specified outcome at a specified follow-up time, not a synonym for permanent cure. The one-year result cannot, on its own, establish how many participants would remain in remission over subsequent years.
Lean ME, Leslie WS, Barnes AC, Brosnahan N et al.. Lancet (London, England), 2018. PMID 29221645 · DOI
Weight loss and the limits of the comparison
Across the whole study population, remission ranged from none among 76 people who gained weight to 31 of 36, or 86%, among those who lost at least 15 kg. Mean weight fell by 10.0 kg in the intervention group and 1.0 kg in controls. The weight-loss categories were observed results, not separately randomised assignments, so they do not guarantee an individual's outcome at a chosen weight target. The trial was also open-label: participants, carers and outcome-data collectors knew the allocation, although the statistician did not. The strongest reading is a one-year benefit of this primary-care programme in the population studied.
Lean ME, Leslie WS, Barnes AC, Brosnahan N et al.. Lancet (London, England), 2018. PMID 29221645 · DOI