Serial Diurnal Salivary Cortisol Profiles in 667 Pregnant Women-Association With Cardiometabolic Complications

Abstract

Context: Maternal obesity, hypertensive pregnancy disorders, and gestational diabetes (GDM) are linked to an increased risk of negative offspring health outcomes. This association may be mediated by maternal hypothalamic-pituitary-adrenal axis (HPA axis) activity, resulting in elevated maternal cortisol levels and fetal exposure, but evidence remains scarce.

Objective: We (1) examined maternal diurnal cortisol profiles longitudinally across gestation, and (2) explored associations with maternal cardiometabolic complications.

Methods: Women in the InTraUterine sampling in early pregnancy (ITU) study (n = 667) provided 7 salivary cortisol samples from awakening to bedtime up to 3 times during pregnancy (median gestational week 19.3, 25.7, and 38.1; n = 9356 samples). Changes in cortisol awakening response (CAR) and diurnal slope (indicative of HPA axis activity) and their associations with maternal body mass index (BMI), hypertensive pregnancy disorders and GDM were examined using linear mixed models.

Results: The CAR declined in 60% to 67% of women, and the diurnal slope attenuated from early to late pregnancy (b = 0.006; P = .001). Higher BMI was associated with less decline in CAR (b = 0.031; P = .0004) and less attenuation in diurnal slope from early to late pregnancy (b = -0.001; P = .006). Hypertensive pregnancy disorders and GDM were not significantly associated with diurnal cortisol profiles.

Conclusion: The attenuation in CAR and diurnal slope support HPA axis hyporesponsivity during pregnancy. Less attenuation of both markers in women with a higher BMI may indicate reduced adaption of the HPA axis to pregnancy, presenting a mechanistic link to offspring health outcomes.

Keywords: body mass index; cardiometabolic disorders; cortisol awakening response; diurnal cortisol slope; hypothalamic-pituitary-adrenal (HPA) axis; pregnancy.

Figure 1.

Flowchart of the exclusion criteria and salivary quality control steps applied to select our final analytic sample. As illustrated, we preprocessed the serial salivary cortisol data at several levels, applying exclusions criteria at the person level (eg, corticosteroid treatment), day level (eg, illness on a specific sampling day), and the individual data point–specific level (eg, sampling delay after awakening). CV%, coefficient of variation percentage; GW, median gestational week. Created with BioRender.com.

Figure 2.

Bar graph of the log-transformed mean cortisol concentration by time of sampling and pregnancy stage. Early pregnancy: median gestational week = 19.3 (IQR = 1.85), midpregnancy: median gestational week = 25.7 (IQR = 1.56), late pregnancy: median gestational week = 38.1 (IQR = 1.28). S1 = awakening, S2 = awakening + 15 min, S3 = awakening + 30 min, S4 = 10:00, S5 = 12:00, S6 = 17:00, S7 = bedtime. Error bars indicate SD.

Figure 3.

Diurnal cortisol profiles in early, mid, and late pregnancy from awakening to bedtime. Mean concentration (in log-transformed μg/L + 1) by mean time of day at sampling and pregnancy stage (early pregnancy = circles [median gestational week = 19.3, IQR = 1.85], midpregnancy = triangles [median gestational week = 25.7, IQR = 1.56], late pregnancy = squares [median gestational week = 38.1, IQR = 1.28]). Average time at awakening and subsequent sampling at +15 min and +30 min varied between pregnancy stages. The cortisol rise from early to mid and late pregnancy in total cortisol output is illustrated in the box plot in the upper right corner. AUCg, area under the curve with respect to ground. Error bars indicate the standard error.

Figure 4.

Body mass index (BMI)-related morning cortisol changes during pregnancy. Mean cortisol concentration (in log-transformed μg/L + 1) at awakening, and +15 minutes and +30 minutes post awakening are shown for women with BMI greater than or equal to 25 (circles) and BMI less than 25 (triangles), illustrating the cortisol awakening response declines less in women with higher BMI in the morning model across all pregnancy stages. Error bars indicate the standard error.

Figure 5.

Body mass index (BMI)-related diurnal cortisol changes during pregnancy. Mean cortisol concentration (in log-transformed μg/L + 1) from the diurnal peak to bedtime in early (circles, median gestational week = 19.3, IQR = 1.85), mid (triangles, median gestational week = 25.7, IQR = 1.56), and late (squares, median gestational week = 25.7, IQR = 1.56) pregnancy is shown for women with BMI less than 25 (left) and BMI greater than or equal to 25 (right), illustrating the association between higher BMI and attenuated diurnal slope from early to late pregnancy in the diurnal model. Error bars indicate the standard error.