Objective: To assess the effect of parity on the occurrence of prediabetes defined as an abnormal fasting plasma glucose (5.6-6.9 mmol/l), an abnormal 2-hour oral glucose tolerance test (7.7-11.1 mmol/l), or both, before 12 weeks gestation or at least 6 weeks after delivery.
Design: Retrospective cohort study.
Setting: Nested on a community trial Delaying the Development of Diabetes Mellitus type 2 (AMAL study) in Oman.
Population: 532 women with a total of 3,196 pregnancies.
Methods: We conducted sets of Cox proportional hazard regression analyses: crude, age-adjusted and full models which adjusted for maternal age, education, family income and year of delivery.
Main outcome measures: Hazard ratio (HR) of the effect of parity on prediabetes.
Results: We enumerated 258 cases of prediabetes over 8,529 person-years of follow up. In the crude model, high parity (> or =5) pregnancies carried a higher risk of prediabetes than low parity (<5) pregnancies (HR = 3.72; 95% CI = 2.80, 4.91), and the prediabetes incidence rate increased in a dose-response fashion over multiple categories of parity. In age-only models, the association attenuated with control of the confounding effect of maternal age (HR = 1.05; 95% CI = 0.76, 1.45). Adjusting for other confounders in the full models yielded similar results to those adjusted for maternal age only.
Conclusions: The apparent effect of parity on the occurrence of prediabetes is attributable to the confounding effect of maternal age rather than to high parity.