Pregnancy length: Secular trends and patterns of variation in Sweden, 1982-2005

Abstract

Anthropological and medical studies rarely investigate the existence of secular trends in the duration of human pregnancy, which is widely assumed to show less variation than traits such as body size, menarche or lifespan. Here we analyze pregnancy duration in the Swedish population between 1982 and 2005, and correlation patterns of four variables: pregnancy length, maternal height, newborn weight and newborn head circumference. Results reveal positive trends of very small magnitude in the four traits. Although bivariate correlations were all significant and positive, multiple linear regression shows a positive independent contribution of newborn size (both weight and head circumference) and a negative independent contribution of maternal height to pregnancy length. We propose that the very weak and negative independent contribution of maternal height to pregnancy duration, in contrast to the stronger and positive contribution of newborn size, explain the absence of significant secular trends in pregnancy duration in Sweden. The results confirm some of the predictions of the maternal investment hypothesis, the 'obstetrical dilemma', and Ellison's metabolic crossover hypothesis. Due to the weak association between pregnancy length and maternal height, we hypothesize that pregnancy length is expected to show limited secular change even in a population undergoing strong secular trends in maternal height.