Abstract
Hamayal et al found a significant correlation between metformin exposure and poor sleep quality in patients with diabetes, highlighting a significant clinical issue. However, the study's validity, reproducibility, and generalizability are limited by methodological limitations. Causal inference is not possible due to the cross-sectional and single-center design, convenience sampling, and selection bias. The use of a self-report instrument, such as the Pittsburgh Sleep Quality Index, increases measurement and recollection bias and risks false classification of sleep disturbance. The relationship between metformin exposure and sleep quality is prone to residual confounding by diabetes duration, glycemic control, comorbidity burden, medication use, depressive symptoms, and socioeconomic factors. Reverse causality is also possible, as disrupted sleep may aggravate glycemic indices. Future research should adopt a prospective, multicentric design, objective measures of sleep, standardized definitions of case and outcome, and examine putative mechanisms involving antidiabetic interventions and sleep physiology.