The study shows incidences on the scale of 15-25 cases per 100,000. The relative incidence ratio may have increased significantly but the absolute increase was small enough that the handful of people with higher A1C wouldn't have budged the average.