Temporal trends without seasonal effects on gestational diabetes incidence relate to reductions in indices of insulin secretion: the Cambridge Baby Growth Study

Aims The incidence of gestational diabetes has been reported to have risen over the first decade of this century. Some studies have also found it to vary with seasons of the year. We investigated temporal and seasonal trends on gestational diabetes incidence in a single-centre cohort study from Cambridge, UK, and attempted to explain trends using associated risk factors. Methods Using a cosinor model, we tested both temporal and seasonal trends in gestational diabetes incidence in 1074 women recruited to the Cambridge Baby Growth Study in 2001-2009 who underwent oral glucose tolerance tests around week 28 of pregnancy. Results There was a temporal increase in gestational diabetes incidence over the course of recruitment to this study [0.014 (0.005, 0.022) proportional increase per year, p = 2.1 x 10(-3)], but no seasonal effect (p = 0.7). HOMA B [- 0.015 (- 0.025, - 0.005) per year, p = 3.0 x 10(-3)] and the insulin disposition index [- 0.036 (- 0.060, - 0.013) per year, p = 3.0 x 10(-3)], unlike HOMA S, showed negative temporal trends. Risk factor analyses showed a concomitant temporal slight increase in the index of multiple deprivation [0.191 (0.138, 0.257) units per year, p = 4.6 x 10(-10)]. This index was positively associated with HOMA B (p = 6.1 x 10(-5)) but not directly with gestational diabetes (p = 0.6), HOMA S (p = 0.2) or the insulin disposition index (p = 0.4). Conclusions In this cohort, there were temporal, but not seasonal, increases in gestational diabetes incidence between the years 2001 and 2009, which appeared to be related more to reductions in insulin secretion than sensitivity. Possible mediators of this link include confounding factors related to deprivation.