Effects of daily administration of melatonin before bedtime on fasting insulin, glucose and insulin sensitivity in healthy adults and patients with metabolic diseases. A systematic review and meta-analysis.

BACKGROUND: Melatonin is increasingly used as a pharmacological sleep aid but it is also emerging as a regulator of glucose homoeostasis. Yet, previous research has been ambiguous with reports of both positive and negative effects of melatonin on glucose metabolism. OBJECTIVES: To assess the effect of daily treatment with melatonin on fasting glucose, insulin, insulin sensitivity and haemoglobin A1c (HbA1c) levels. DATA SOURCES: MEDLINE, EMBASE, CENTRAL, clinicaltrials.gov and clinicaltrialsregister.eu were systematically searched. ELIGIBILITY CRITERIA, PARTICIPANTS AND INTERVENTIONS: All randomized, placebo-controlled studies with melatonin treatment were assessed. We included studies with daily melatonin treatment (≥2 weeks) of healthy adults or patients with metabolic diseases. METHODS: Hedges' g differences were calculated for the metabolic parameters of the included studies, heterogeneity was assessed with χ2 and I2 tests and meta-analyses were performed with the random-effects model. RESULTS: Long-term treatment with melatonin did not change fasting glucose significantly compared with placebo (g: -0.07 [-0.22 to 0.08], n = 603) but it reduced fasting insulin levels slightly (g: -0.27 [-0.50 to -0.04], n = 278) and trended towards reduced insulin resistance (HOMA-IR) (g: -0.20 [-0.44 to 0.03], n = 278). HbA1c levels were largely unaffected by melatonin treatment compared with placebo (g: 0.14 [-0.19 to 0.46], n = 142). CONCLUSIONS: With the available literature, melatonin seems to be a glucose-metabolic safe sleep aid in patients with metabolic diseases and in healthy adults. It may even have beneficial glucose-metabolic effects as fasting insulin levels were reduced in this meta-analysis, but the confidence intervals of the meta-analyses are wide, underscoring the need for further research within this field.