Diabetes mellitus is a risk factor for severe dengue in adults, but few studies have examined the association between metformin use and disease severity in dengue. In addition to its effect on glucose control, metformin has been associated with pleiotropic properties in preclinical studies. Using a cohort of laboratory-confirmed adult (≥21 years) dengue patients with diabetes mellitus admitted to Tan Tock Seng Hospital, we conducted a retrospective cohort study involving 131 (58.7%) metformin users and 92 (41.3%) non-users. Dengue severity was categorized as dengue hemorrhagic fever (DHF) or dengue shock syndrome (DSS) in World Health Organization (WHO) 1997 criteria and severe dengue (SD) in WHO 2009 criteria. Multivariable Poisson regression with robust error variance was used to estimate risk ratio (RR). Compared with non-use, metformin use was associated with a decreased risk of developing severe dengue (adjusted risk ratio [aRR] = 0.60, 95% confidence interval [CI]: 0.37-0.98, P = 0.04). Additionally, there was an inverse dose-response relationship (aRR = 0.69, 95% CI: 0.49-0.98, P = 0.04) with dengue severity as classified by WHO 2009 criteria. Use of metformin, however, was not associated with dengue severity based on WHO 1997 criteria; and no dose-response relationship was noted. Our results suggest metformin use could attenuate disease severity in dengue-infected diabetes mellitus individuals.