Variation in incidence of type 2 diabetes mellitus: time series of Mexican adolescents.

PURPOSE: To study temporal changes in the cumulative incidence (CI) of type 2 diabetes mellitus during early and late adolescence from 2003 to 2013.

METHODS: This was an ecologic, analytical study of trends over time. Data were weekly reports of new cases (General Directorate of Epidemiology). Specific CI was calculated and standardized by age using the direct method (WHO). Autoregressive Integrated Moving Average models offering a better fit to the observed series were calculated and controlled by intentional screening. Structural break point analysis was performed.

RESULTS: The CI was lower in younger adolescents than in older adolescents. In early adolescence, the incidence was similar in both sexes and stable over time [Autoregressive Integrated Moving Average female: (2,0,2)(0,0,0), male: (1,0,1)(0,0,0); P < .001], whereas in late adolescence, the female incidence was higher than the male incidence and showed a linear increase [female: (1,1,2)(1,0,0), male: (1,0,1)(0,0,0); P < .001)]. The female series showed two structural break points, in 2010 and 2012. The male early adolescent series showed one break point in 2011.

CONCLUSIONS: Although there was an increase in the CI of type 2 diabetes mellitus during the study period, only the female late adolescence series showed an epidemiologically significant linear trend. There was also a brief, limited rise between 2010 and 2012 that affected all adolescents. This suggests that the disease may be triggered by specific events.