Treating insomnia with talk therapy also nudged blood sugar down in diabetes

Sleep problems are strikingly common in people living with diabetes, and they are tied to a range of physical, psychological and social harms. That raises a practical question: if a non-drug treatment can fix the sleep, does the diabetes itself improve? Cognitive behavioural therapy has a strong record for sleep problems in general, but its effect on blood sugar control in diabetes had not been rigorously synthesised. This review pooled the randomised evidence to find out whether better sleep comes with better glycaemic numbers.

The authors searched ten databases, including Chinese-language sources, from inception to June 2026 for randomised controlled trials comparing cognitive behavioural therapy against usual care or health education in people with diabetes. Ten trials qualified, covering 2827 participants across 4 countries. Researchers independently screened the studies and extracted the data, and the risk of bias in each trial was formally assessed. Results were pooled with random-effects models, with further analyses probing publication bias and the sources of disagreement between trials.

Sleep quality improved significantly with cognitive behavioural therapy compared with usual care or health education. HbA1c, the standard marker of long-term blood sugar control, fell by 0.35% on average, and insomnia severity also improved, though that estimate was less precise. BMI did not change, suggesting the glycaemic benefit was not simply a weight story. The trials disagreed with each other considerably on the size of the benefits, and exploratory analyses found results differed by follow-up time, while disease type, sample size and intervention plan did not explain the variation.

The finding extends the established role of cognitive behavioural therapy in insomnia into diabetes care, suggesting a behavioural sleep treatment can move a core medical marker. A 0.35% fall in HbA1c is a notable shift for a non-drug intervention, though the large disagreement between trials means the true size is uncertain. The authors call for higher-quality trials to settle the best protocol and identify who benefits most. How long the glycaemic improvement lasts after therapy ends also remains an open question.

For anyone managing diabetes alongside poor sleep, this is a reason to treat the sleep problem as part of the condition rather than a side issue. The evidence suggests a structured talking therapy for sleep can come with lower HbA1c, without any change in body weight. It is not a replacement for standard diabetes care, and the variability between trials means individual results will differ. This sits best as a conversation topic for a care team rather than a do-it-yourself fix.