Stress during pregnancy and its influence on the offspring's development and temperament: a physiological study

Abstract

Generalized Linear Mixed Models (GLMM's) are appropriate to model univariate and multivariate outcome data. These methods are especially suitable to analyze missing no-normal questionnaire data. GLMM's were used to study the associations between antenatal maternal anxiety (measured with The Pregnancy Related Anxiety Questionnaire) and cortisol, as well as antenatal maternal depression (measured with the Edinburgh pregnancy depression scale) and cortisol. In total, there were 192 normal participants in the Prenatal Early Life Stress (PELS) PROJECT TILBURG.

Univariate, bivariate and multivariate GLMM's for longitudinal correlated and missing data showed that the odds of having high levels of fear about the baby's integrity increases 1% per one unit of increase in cortisol in the first tow trimesters (log odds ratio of 1.01 p-value=0.0098) and 7.2% in the last trimester.