2. Rural Hospital in Tlacolula, Oaxaca, Mexico.
3. Department of Clinical Pharmacology and Epidemiology, Consorzio Mario Negri Sud, Chieti, Italy.
4. Unit of Biostatistics, IRCCS Casa Sollievo della Sofferenza San Giovanni Rotondo, Italy.
Background: Bimodal distribution of glucose concentrations may be a universal phenomenon, but its use in establishing diagnostic cut off values is limited. The purpose of this study was to evaluate the presence of bimodality in the distribution of fasting and two-hour post-load glucose-values, and the cutoff point that best identifies subjects with diabetes in an indigenous population.
Methods: Mexican Indians ≥35 years old were invited to participate in a cross-sectional study, 394 Zapotec and 730 Mixe Indians were included. A normal distribution and a mixture of two normal distributions were fitted to fasting and two-hour post-load glucose-values. To assess the presence of bimodality, the mixture model was compared to the unimodal and the cutoff value for normal glycemia was calculated as the crossing point of the two normal distributions in the mixture model.
Results: Bimodal distribution was observed in the oral glucose tolerance test and the chi-square likelihood ratio statistic showed differences between the unimodal and the normal bimodal models. The cutoff value for diagnosing diabetes with fasting glucose was 6.41 mmol/l, whereas it was 7.96 mmol/l for the tow-hour post-load glucose.
Conclusions: In these Mexican Indians, the fasting and the 2-hour post-load glucose values show bimodal distribution. The cutoff points for diagnosing diabetes are set below the current ones as seen in other populations where the diabetes epidemic is evolving.
Keywords: Bimodality, glucose tolerance test, diabetes mellitus, type 2, prediabetic state, Mexican Indians