Spontaneous ketonuria and risk of incident diabetes: a 12 year prospective study

Aims/hypothesis Ketones may be regarded as a thrifty fuel for peripheral tissues, but their clinical prognostic significance remains unclear. We investigated the association between spontaneous fasting ketonuria and incident diabetes in conjunction with changes in metabolic variables in a large population-based observational study. Methods We analysed 8703 individuals free of diabetes at baseline enrolled in the Korean Genome and Epidemiology Study, a community-based 12year prospective study. Individuals with (n=195) or without fasting ketonuria were matched 1:4 by propensity score. Incident diabetes was defined as fasting plasma glucose >= 7.0 mmol/l, post-load 2 h glucose >= 11.1 mmol/l on biennial OGTTs, or current use of glucose-lowering medication. Using Cox regression models, HRs for developing diabetes associated with the presence of ketonuria at baseline were analysed. Results Over 12 years, of the 925 participants in the propensity score-matched cohort, 190 (20.5%) developed diabetes. The incidence rate of diabetes was significantly lower in participants with spontaneous ketonuria compared with those without ketonuria (HR 0.63; 95% CI 0.41, 0.97). Results were virtually identical when participants with fasting ketonuria were compared against all participants without ketonuria (after multivariate adjustment, HR 0.66; 95% CI 0.45, 0.96). During follow-up, participants with baseline ketonuria maintained lower post-load 1 h and 2 h glucose levels and a higher insulinogenic index despite comparable baseline values. Conclusions/interpretation The presence of spontaneous fasting ketonuria was significantly associated with a reduced risk of diabetes, independently of metabolic variables. Our findings suggest that spontaneous fasting ketonuria may have a potential preventive role in the development of diabetes.