Jennum · Diabetes care 2015 · single-blind randomized crossover trial · n=26

The Impact of Nocturnal Hypoglycemia on Sleep in Subjects With Type 2 Diabetes.

Cited 22 times in the scientific literature.

Level 2 - randomized trial

Randomized crossover trial

PubMed 26407587 · doi:10.2337/dc15-0907 · record verified 2026-08-28

What was done

In a single-blind, randomized crossover trial, 26 participants with type 2 diabetes completed two overnight polysomnography visits: one normoglycemic and one hypoglycemic. Blood glucose was regulated via hyperinsulinemic glucose clamps. On the hypoglycemic night, glucose was lowered after reaching sleep stage N2 to a target of 2.7–2.8 mmol/L for 15 minutes, followed by restoration to normoglycemia for the remainder of the night. On the normoglycemic night, glucose was held at 5.0–7.0 mmol/L throughout. Sleep architecture, EEG-identified arousals/awakenings, and counterregulatory hormones were evaluated.

What was found

During the first 4 hours of sleep (0–4 h after N2 onset), no significant differences were observed in EEG-identified arousals or awakenings between conditions. During hours 4–8, the awakening rate was 27% lower on the hypoglycemic night compared with the normoglycemic night, and 20% lower across the full 0–8 h period (both statistically significant). Total sleep time showed a non-significant trend toward longer duration on the hypoglycemic night (mean 366 vs. 349 min). Counterregulatory hormones (adrenaline, growth hormone, and cortisol) were statistically significantly higher following hypoglycemia than normoglycemia.

Why it matters

This study shows that nocturnal hypoglycemia impairs subsequent awakening responses in type 2 diabetes, potentially compromising a patient's ability to wake up and self-treat with carbohydrates.

Limits

The study had a small sample size (n = 26) and used an artificial, brief clamp-induced hypoglycemia model in a sleep laboratory rather than spontaneous nocturnal hypoglycemia. The trial was single-blinded, and daytime functional or cognitive outcomes were not assessed.

Cited by