Sinnott-Armstrong · Nature communications 2026 · genome-wide association and cohort study · n=~1,232,000

Genetic variants affect diurnal glucose levels throughout the day.

Level 3 - non-randomized controlled study

Large-scale observational genome-wide association and cohort study across multiple biobanks

PubMed 42173832 · doi:10.1038/s41467-026-72432-6 · record verified 2026-08-30

What was done

Genome-wide association study investigating genetic variants that contribute to diurnal (time-of-day) control of blood glucose levels. The discovery cohort comprised ~420,000 individuals from the UK Biobank, with replication across ~100,000 individuals from the Estonian Biobank, ~500,000 from FinnGen, ~160,000 from the VA Million Veteran Program, and ~52,000 from the MGB Biobank. The authors also tested whether sleep modulated glucose levels and if these variants or sleep medications (including melatonin) associated with type 2 diabetes risk.

What was found

Glucose levels were found to be under diurnal genetic control. Significant temporal associations were discovered at the Melatonin receptor 1B locus (MTNR1B, rs10830963, P = 1 × 10⁻²²) and the Cryptochrome 2 circadian pacemaker locus (CRY2, rs12419690, P = 1 × 10⁻¹⁶). Sleep independently modulated glucose levels, and the genetic variants had an independent role in diurnal glucose control. The variants independently modulated type 2 diabetes risk, and sleep medications including melatonin associated with type 2 diabetes (effect sizes not reported in the abstract).

Why it matters

This study provides large-scale human genetic evidence linking circadian pacemaker and melatonin receptor genes directly to time-of-day glucose regulation and type 2 diabetes risk.

Limits

The abstract reports P-values but no effect sizes, beta coefficients, or odds ratios for the genetic loci, sleep effects, or medication associations. Biobank cohorts predominantly feature adult populations of European ancestry, limiting generalizability. Observational links between sleep medications and type 2 diabetes remain susceptible to confounding.

Cited by