BACKGROUNDPrevious studies have shown that diabetes mellitus is a common comorbidity of coronavirus disease 2019 (COVID-19), but the effects of diabetes or anti-diabetic medication on the mortality of COVID-19 have not been well described. AIMTo investigate the outcome of different statuses (with or without comorbidity) and anti-diabetic medication use before admission of diabetic after COVID-19.METHODSIn this multicenter and retrospective study, we enrolled 1422 consecutive hospitalized patients from January 21, 2020, to March 25, 2020, at six hospitals in Hubei Province, China. The primary endpoint was in-hospital mortality. Epidemiological material, demographic information, clinical data, laboratory parameters, radiographic characteristics, treatment and outcome were extracted from electronic medical records using a standardized data collection form. Most of the laboratory data except fasting plasma glucose (FPG) were obtained in first hospitalization, and FPG was collected in the next day morning. Major clinical symptoms, vital signs at admission and comorbidities were collected. The treatment data included not only COVID-19 but also diabetes mellitus. The duration from the onset of symptoms to admission, illness severity, intensive care unit (ICU) admission, and length of hospital stay were also recorded. All data were checked by a team of sophisticated physicians.RESULTSPatients with diabetes were 10 years older than non-diabetic patients [(39 - 64) vs (56 - 70), P < 0.001] and had a higher prevalence of comorbidities such as hypertension (55.5% vs 21.4%, P < 0.001), coronary heart disease (CHD) (9.9% vs 3.5%, P < 0.001), cerebrovascular disease (CVD) (3% vs 2.2%, P < 0.001), and chronic kidney disease (CKD) (4.7% vs 1.5%, P = 0.007). Mortality (13.6% vs 7.2%, P = 0.003) was more prevalent among the diabetes group. Further analysis revealed that patients with diabetes who took acarbose had a lower mortality rate (2.2% vs 26.1, P < 0.01). Multivariable Cox regression showed that male sex [hazard ratio (HR) 2.59 (1.68 - 3.99), P < 0.001], hypertension [HR 1.75 (1.18 - 2.60), P = 0.006), CKD [HR 4.55 (2.52-8.20), P < 0.001], CVD [HR 2.35 (1.27 - 4.33), P = 0.006], and age were risk factors for the COVID-19 mortality. Higher HRs were noted in those aged ≥ 65 (HR 11.8 [4.6 - 30.2], P < 0.001) vs 50-64 years (HR 5.86 [2.27 - 15.12], P < 0.001). The survival curve revealed that, compared with the diabetes only group, the mortality was increased in the diabetes with comorbidities group (P = 0.009) but was not significantly different from the non-comorbidity group (P = 0.59).CONCLUSIONPatients with diabetes had worse outcomes when suffering from COVID-19; however, the outcome was not associated with diabetes itself but with comorbidities. Furthermore, acarbose could reduce the mortality in diabetic.