Reference ranges and determinants of total hCG levels during pregnancy: the Generation R Study

Abstract

Human chorionic gonadotropin (hCG) is a pregnancy hormone secreted by the placental synctiotrophoblast cell layer that has been linked to fetal growth and various placental, uterine and fetal functions. In order to investigate the effects of hCG on clinical endpoints, knowledge on reference range (RR) methodology and determinants of gestational hCG levels is crucial. Moreover, a better understanding of gestational hCG physiology can improve current screening programs and future clinical management. Serum total hCG levels were determined in 8195 women participating in the Generation R Study. Gestational age specific RRs using 'ultrasound derived gestational age' (US RRs) were calculated and compared with 'last menstrual period derived gestational age' (LMP RRs) and a model-based RR. We also investigated which pregnancy characteristics were associated with hCG levels. Compared to the US RRs, the LMP RRs were lower, most notably for the median and lower limit levels. No considerable differences were found between RRs calculated in the general population or in uncomplicated pregnancies only. Maternal smoking, BMI, parity, ethnicity, fetal gender, placental weight and hyperemesis gravidarum symptoms were associated with total hCG. We provide gestational RRs for total hCG and show that total hCG values and RR cut-offs during pregnancy vary depending on pregnancy dating methodology. This is likely due to the influence of hCG on embryonic growth, suggesting that ultrasound based pregnancy dating might be less reliable in women with high/low hCG levels. Furthermore, we identify different pregnancy characteristics that influence total hCG levels considerably and should therefore be accounted for in clinical studies.

Keywords: Determinants; Pregnancy; Reference range; hCG.

Fig. 1

Gestational age specific reference ranges for total hCG levels during pregnancy. Total hCG reference range values were calculated according to a (semi) parametric 2.5th–97.5th percentiles by GAMLSS modelling in a population-based approach among the entire study population, after exclusion of women with IVF treatment (N = 38), twin pregnancy (N = 90) or TOP pregnancies (N = 2). Colored lines depict the gestational age specific centiles for total hCG levels. Grey area depict areas with higher uncertainty due to small numbers (N per week <40 before week 9 and after week 24)

Fig. 2

The relationship between maternal or fetal characteristics and total hCG MoM levels. Plots show the relationship between pregnancy characteristics and total hCG MoM levels for continuous and categorical variables as predicted mean with 95 % confidence interval. Analyses were performed after exclusion of women with IVF treatment (N = 38), twin pregnancy (N = 90) or TOP pregnancies (N = 2), and were adjusted for maternal age, smoking, BMI, parity, education level, ethnicity and fetal gender

Fig. 3

The relationship between placental weight and total hCG MoM levels. Plots show the relationship between placental weight at birth and total hCG MoM levels as predicted mean with 95 percent confidence interval. Analyses were performed after exclusion of women with IVF treatment (N = 38), twin pregnancy (N = 90) or TOP pregnancies (N = 2; placental weight available in n = 5851) and were adjusted for maternal age, smoking, BMI, parity, education level, ethnicity and fetal gender. For uncomplicated pregnancies we selected women’s first pregnancy registered in our database and excluded pregnancies with a non-live born child, preterm birth, a small for gestational age newborn, hypertensive disorder or pre-existing hypertension resulting in a population of n = 7015 (with placental weight available n = 4999)