First trimester β-hCG and estradiol levels in singleton and twin pregnancies after assisted reproduction

Introduction

Assisted conception increased dramatically over recent years worldwide. According to the most recent Center of Disease Control report on Assisted Reproductive Technology (ART) [1] ART contributed to 1.5% of all infants born in the United States. In Europe [2], the percentage of ART neonates varied from 0.2% in Lithuania to 6.1% in Denmark. Among these, the rate of multiple pregnancy is as high as 19.9%–43.6% [1], [2] pointing to the ever increasing prevalence of multiples (especially twins) among ART births.

Twin gestations appear to have an advantage over singletons conceived by ART in terms of early pregnancy loss. Indeed, ART twins have a two to five times lower miscarriage rate of the entire pregnancy compared to singletons [3]. This observation is lower even when the calculation includes a double loss as stochastically occurring two single losses in the same pregnancy (i.e. the vanishing twin syndrome). The explanation for this phenomenon is unclear. ZegersHochschild et al. [4] suggested that women with high reproductive efficacy exposed to ART generate cohorts of good quality embryos with a higher chance of implantation, having multiple gestations. Lambers et al. [5] suggested that multiple implantation at 6 weeks of gestation is predominantly determined by embryo quality but beyond 6 weeks pregnancy is more dependent on the combination of genetic and developmental potential of the embryo(s) and an optimal uterine milieu. La Sala et al. [6] and Matias et al. [3], [7] suggested that higher levels of placental hormones produced by a larger placental mass in twin gestations, might confer advantage to twins over singletons.

Pregnancy-associated hormones like the human chorionic gonadotropin (hCG) and estradiol increase during pregnancy and are essential for successful pregnancy outcome. Functional differentiation of the villous trophoblast is specifically stimulated by estradiol, glucocorticoids and hCG. Because these hormones are temporally secreted in large amounts and present at the feto-maternal interface, they are excellent candidates to be involved in trophoblast differentiation.

The aim of our study was to evaluate βhCG and estradiol levels during the first trimester of ART gestations in twins and singletons that eventually were live born.

Patients and methods

Between November 2008 and April 2012, we studied singleton and twin pregnancies following ART at the Unit of Reproductive Medicine, Hospitalar Center São João. Patients underwent ART (in vitro fertilization, IVF, or intracytoplasmatic sperm injection, ICSI) with fresh double embryo transfers.

Women with positive heartbeat embryos at 7 weeks’ gestation ultrasound were invited to participate in the study and 108 agreed to return between 8 and 9 weeks+6 days for a second ultrasound and for collecting serum samples for βhCG and estradiol. We excluded monochorionic twins, ectopic pregnancies or anembryonic sacs, and pregnancies that ended with missed abortion. We also excluded cases with ovarian hyperstimulation syndrome (OHSS). We followed these pregnancies until delivery. In this study we did not compare those pregnancies with and without adverse outcome but only twins to singletons. All neonates were normal.

We used standard ART stimulation and ovum pick-up protocols. Double fresh embryo transfer was performed between days 2 and 5. Stimulation was individualized. Controlled ovarian hyperstimulation was performed according to standard protocols with recombinant follicle stimulating hormone (r-FSH) or human menopausal gonadotrophin (hMG) along with GnRH antagonist (short protocol) or GnRH agonist (ultra long protocol, in patients with endometriosis). Vaginal micronized progesterone (600 mg/day) was started on the day of oocyte retrieval for luteal support until serum measurement of β-hCG (13 days after embryo transfer). Clinical pregnancy was confirmed by vaginal ultrasound scan at 6 weeks’ gestation.

The β-hCG assay was performed at day 13 after embryo transfer, as routine, and between 8 and 9 weeks+6 days. Serum β-hCG concentrations were measured with a chemiluminescent microparticle immunoassay (CMIA, Abbott Laboratories, Abbott Park, IL, USA). The inter- and intra-assay coefficient of variation was <5% with an assay sensitivity of being <1.2 mIU/L.

The estradiol assay was performed at the same time of the second β-hCG sampling as part of the study. Serum estradiol concentrations were measured with CMIA (Abbott Laboratories, Abbott Park, IL, USA). The inter- and intra-assay coefficient of variation was <7% with an assay sensitivity of being of ≤10 pg/mL. All positive β-hCG cases underwent routine transvaginal ultrasound examinations at 6 weeks. Participants underwent another scan at the time of blood sampling to confirm viability.

Data were collected using the Microsoft Excel^® program (Microsoft Corporation, Redmond, WA, USA). Analysis was performed with the Statistical Package for Social Sciences version 21.0 (SPSS^®, Chicago, IL, USA). Student’s t-test and Mann-U Whitney test were used to compare parametric and non-parametric continuous data, respectively. Fisher’s exact and χ² tests were used to compare categorical data. Significance was set at P<0.05.

All the participants gave written informed consent for participating in the study. The study was approved by the local Ethical Committee of the Hospitalar Center São João, Porto, Portugal.

Results

Of the cohort of 108 pregnancies starting the study (53 singletons, 55 dichorionic twins) we excluded five cases with a vanishing twin, four first trimester abortions, and two cases with adverse outcomes at the second half of gestation. Thus, the study comprised of 50 singleton and 47 twin gestations. A total of 144 neonates were born. All women were Caucasians and non-smoking. The mean maternal age was 33 years (range 23–40 years).

Table 1 shows selected demographic and cycle characteristics. It appears that the groups of singleton and twin pregnancies were similar in terms of age, body mass index (BMI), incidence of primary infertility, duration or causes of infertility and endometrial thickness at the day of embryo transfer. Complete cycle characteristics as well as various other hormone levels were also similar (data not shown).

In contrast, we observed (Table 2) a significant difference between singletons and twins in the levels of β-hCG at day 13 (medians 329 and 544 mIU/mL for singleton and twin pregnancies, respectively), β-hCG and at 8–9 weeks+6 days (medians 105,954 and 224,240 mIU/mL for singleton and twin pregnancies, respectively), as well of the level of estradiol at 8–9 weeks+6 days (median=1460 and 3059 pg/mL for singleton and twin pregnancies, respectively).

Discussion

Our study demonstrates that β-hCG and estradiol levels were more than two-fold increased in twins than in singletons. These findings are in agreement with comparisons of hCG (rather than β-hCG) levels [8], [9] and estradiol levels [10]. Serum hCG and estradiol coordinate and control anatomical modifications associated with pregnancy. Serum hCG is also implicated in promoting tolerance and angiogenesis required for a successful pregnancy [11] whereas estrogens also have powerful effects on immune cells regulating their proliferation, distribution and function [12].

Because no differences existed between of two groups in terms of demographic variables, cycle characteristics, and hormone measurements before and during treatment, we assume that the differences found between groups in the levels of β-hCG and estradiol are related to the number of feto-placental units. Hence, two placentas instead of one in twin pregnancies produce more β-hCG and estradiol and might explain the advantage of twins over singletons in early pregnancy following ART [3], [4], [5], [6], [7] and indirectly support the hypothesis that higher maternal levels of these hormones in twin pregnancies lead to the lower miscarriage rate of the entire pregnancy compared to singletons [3].

In this study, all pregnancies ended as livebirths and we did not compare the hormonal levels to cases with adverse pregnancy outcomes such as vanishing twins and entire pregnancy losses. This limitation results from the fact that miscarriages had already occurred before the time of analysis thus preventing to relate low hormonal levels with future loss.

The strength of the study is the prospective examination of a large number of pregnancies coming from a single ART center with standard treatment protocols. Also, hormone measurements were performed within a narrow gestational age range and analyzed in a single laboratory thus minimizing inter- and intra-assay variability.

Evidently, more studies are required to establish the cause and effect relationship proposed in our working hypothesis.