Elevated progesterone and its impact on birth weight after fresh embryo transfers

Abstract

Purpose

The purpose of the study was to examine the association between serum progesterone levels on the day of hCG administration and birth weight among singleton live births after fresh embryo transfer.

Methods

This study was conducted as a retrospective cohort database analysis on patients who underwent IVF treatment cycles from January 2004 to April 2012. The study was performed at a University affiliated private infertility practice. All cycles that had achieved a singleton live birth after fresh embryo transfer and for which progesterone was measured on the day of hCG administration were examined. Generalized linear models were used to calculate mean birth weight and z-scores.

Results

We analyzed 817 fresh IVF embryo transfers in which birth weight, gestational age, and progesterone (ng/mL) level on day of hCG administration were documented. While there was a decrease in birth weight as progesterone quartile [≤0.54; >0.54 to ≤0.81; >0.81 to ≤1.17; >1.17 ng/mL] increased, the difference in mean birth weights among the four quartiles was not statistically significant (p = 0.11) after adjusting for maternal age and peak estradiol levels. When dichotomizing based on a serum progesterone considered clinically elevated, cycles with progesterone >2.0 ng/mL had a significantly lower mean singleton birth weight (2860 g (95% CI 2642 g, 3079 g)) compared to cycles with progesterone ≤2.0 ng/mL (3167 g (95% CI 3122 g, 3211 g) p = 0.007)) after adjusting for maternal age and estradiol.

Conclusion

We demonstrated that caution should be exercised when performing fresh embryo transfers with elevated progesterone levels and in particular with levels (>2.0 ng/mL) as this may lead to lower birth weight.

Electronic supplementary material

The online version of this article (doi:10.1007/s10815-017-0920-8) contains supplementary material, which is available to authorized users.

Introduction

National databases and multiple studies have indicated that infants born after assisted reproductive technology (ART) are smaller than those born from natural conception [1–4]. Aside from the impact multiple births has had on birth weight, subfertility itself is believed to be a risk factor for low birth weight (LBW) as it has been demonstrated that spontaneous conception after a diagnosis of infertility increases the risk of adverse birth outcomes [5–7].

The most conclusive evidence linking ART procedures to LBW has arisen from a study by Shih et al. [8] who demonstrated that singletons from frozen embryo transfer procedures have higher mean birth weight and are less likely to be LBW than those from fresh embryo transfers. Several other studies [9, 10] have also demonstrated this finding including our own [11], which has important implications on the potential etiology of LBW in ART cycles.

A key important physiological difference between frozen and fresh embryo cycles is that during frozen cycles, the embryo is not exposed to the supraphysiologic hormonal milieu that is characteristic of controlled ovarian hyperstimulation (COH) during fresh IVF cycles. In addition, Kalra et al. [12] specifically showed that when comparing frozen embryo transfer cycles and fresh transfer cycles in oocyte donor recipients, who do not undergo any ovarian stimulation, no difference in LBW was demonstrated. One of the major factors believed to impact pregnancy in fresh embryo transfers is elevated progesterone. In a meta-analysis of over 60,000 cycles, Venetis et al. [13] concluded that elevated progesterone on the day of hCG administration is associated with a decreased probability of pregnancy in fresh IVF cycles in women undergoing COH using GnRH analogues and gonadotrophins.

To our knowledge, there are no published studies investigating whether there is an association between elevated progesterone levels on the day of hCG administration and live birth weight. During COH, elevation of serum progesterone is generally prevented by suppressing luteinizing hormone secretion with a gonadotropin releasing hormone agonist or antagonist. Despite these medications, subtle increases in serum progesterone are still observed, with elevated progesterone on the day of hCG administration in as many as 35% of cycles [14–16]. This phenomenon has been referred to as premature luteinization, progesterone elevation, and/or premature progesterone rise [17, 18] and, as stated above, has been shown to negatively impact clinical and live birth rates [13, 19, 20].

Although the pathogenesis of LBW in IVF singletons is not known, we hypothesize that elevated progesterone levels on the day of hCG administration may not only negatively impact pregnancy and live birth rates, there may be an impact on placental function resulting in lower birth weight in pregnancies achieved in a high progesterone environment. To address this hypothesis, we investigated the effect of elevated progesterone on the day of hCG administration on the birth weight of live born term singletons after fresh embryo transfer IVF cycles at our center.

Materials and methods

Study population and design

We performed a retrospective study of all fresh IVF cycles from January 2004 to April 2012 that resulted in a singleton live birth and for which progesterone was measured on the day of hCG administration. We included only the first eligible cycle for each woman, and cycles with missing data on birth weight, gestational age at delivery, or progesterone were excluded.

We collected baseline demographic characteristics, cycle characteristics, laboratory values, and cycle outcomes from medical records. The primary outcome was birth weight. Infants who weighed less than 2500 g at birth were considered LBW. Sex-specific birth weight for gestational age (z-scores) was calculated using Fenton growth curves [21]. Infants with z-scores below the 10th percentile were considered small for gestational age (SGA). The institutional review board at Beth Israel Deaconess Medical Center approved this study.

Stimulation protocol

Patients underwent treatment protocols for ovarian stimulation, monitoring, and oocyte retrieval as previously described [22]. Briefly, cycles were monitored with daily serum E2 levels and transvaginal ultrasound examinations beginning on treatment days 6 to 8. When at least three follicles measured 15 to 20 mm, either 250 μg recombinant hCG (Ovidrel; EMD Serono) or 10,000 U of urinary hCG (Novarel; Ferring Pharmaceuticals) was administered subcutaneously. More recently, a GnRH agonist trigger (leuprolide acetate) has been used to mitigate risk of OHSS, however during the period of 2004–2012 less than 3% of cycles used a GnRH agonist trigger. Embryos were transferred into the uterus either 3 or 5 days after oocyte retrieval. Patients received luteal phase support until 8 weeks of gestation. Intracytoplasmic sperm injection and assisted hatching were performed when indicated.

Hormone measurements

Whole blood was collected between 6:30 and 9 a.m. on the day of hCG administration. Progesterone and estradiol levels were measured using the Immulite 2000 Immunoassay System (Siemens Healthcare Global), which has a sensitivity of 0.1 ng/mL for progesterone and 15 pg/mL for estradiol. This assay was used for the duration of the study. Although a consensus is not available on what is a high progesterone value [13, 23], we chose >2.0 ng/mL as a clinically relevant high progesterone level which would impact live birth rates in a fresh IVF cycle.

Patients were also divided into quartiles according to serum progesterone on the day of hCG administration. Patients were assigned to groups as follows: quartile 1, ≤0.54 ng/mL; quartile 2, >0.54 to ≤0.81 ng/mL; quartile 3, >0.81 to ≤1.17 ng/mL; and quartile 4, >1.17 ng/mL.

To investigate the role of estradiol, patients were also divided into quartiles according to their peak estradiol levels on the day of hCG trigger. They were assigned as follows: quartile 1, ≤1396 pg/mL; quartile 2, >1396 to ≤2193 pg/mL; quartile 3, >2193 to ≤3326 pg/mL; and quartile 4, >3326 pg/mL.

Statistical analysis

Descriptive data are presented as median (interquartile range) or count and proportion. Fisher’s exact and chi-square tests were used to compare categorical variables among the progesterone quartiles. Continuous variables were compared using the Kruskal-Wallis test. Generalized linear models were used to calculate mean birth weight and z-scores for each progesterone and estradiol quartile. As we did not observe a linear relationship between progesterone and estradiol quartiles and birth weight, we modeled both progesterone and estradiol as categorical variables. The progesterone models were adjusted for maternal age at cycle start and peak estradiol level, while the peak estradiol models were adjusted for maternal age at cycle start and progesterone level. Generalized linear models were also used to calculate and compare mean birth weight and z-scores between women with progesterone levels >2.0 ng/mL and those with levels ≤2.0 ng/mL. These models were also adjusted for maternal age at cycle start and peak estradiol level. The proportions of LBW (<2500.0 g) and small for gestational age (<10th percentile) infants were compared between cycles with progesterone levels >2.0 and those with levels ≤2.0. p values <0.05 were considered statistically significant.

Results

Participant characteristics

There were 6223 fresh IVF cycles performed during the study period resulting in 1451 live births. After excluding multiple gestations (n = 370) and cycles with missing data on birth weight, gestational age, and progesterone level on day of hCG administration (n = 264), 817 cycles were included in the final analysis. Participant characteristics stratified by progesterone quartile are presented in Table 1. Women in each quartile were similar with regard to maternal age, partner age, gravidity, parity, and body mass index. The numbers of embryos transferred, as well as number of embryos frozen, also were similar. Peak estradiol levels and the numbers of oocytes retrieved increased with increasing progesterone levels across quartiles. The incidence of preterm delivery was 43.7, 46.1, 37.0, and 47.5% with increasing progesterone quartile, respectively.

Table 1.

Baseline characteristics for all cycles and cycles broken down into quartiles of progesterone values on the day of hCG trigger for patients receiving embryo transfers and achieving a singleton live birth

	All	Quartile 1 ≤0.54 ng/mL	Quartile 2 >0.54–≤0.81 ng/mL	Quartile 3 > 0.81– ≤ 1.17 ng/mL	Quartile 4 >1.17 ng/mL
Singleton deliveries	n = 817	n = 206	n = 206	n = 203	n = 202
Age at cycle start (years)	35.0 (32.0–38.0)	34.0 (32.0–37.0)	35.0 (31.0–38.0)	35.0 (31.0–38.0)	35.0 (31.0–37.0)
Partner age (years)	36.4 (32.9–40.5)	36.4 (33.5–40.4)	36.3 (32.5–40.1)	37.2 (34.1–41.3)	36.1 (32.5–39.9)
Body mass index	24.0 (21.5–27.1)	24.6 (21.8–28.9)	24.3 (22.1–27.1)	23.4 (21.0–26.9)	23.5 (21.6–26.6)
Gestational age at delivery	37.1 (35.7–38.6)	37.1 (35.9–38.4)	37.1 (35.6–38.6)	37.6)35.9–38.6)	37.0 (35.6–38.7)
Gravidity
0	364 (44.5)	97 (47.1)	89 (43.2)	88 (43.4)	90 (44.6)
1	230 (28.2)	59 (28.6)	67 (32.5)	52 (25.6)	52 (24.4)
≥2	223 (27.3)	50 (24.3)	50 (24.3)	63 (31.0)	60 (29.7)
Parity
0	583 (71.4)	144 (69.9)	150 (72.8)	140 (69.0)	149 (73.8)
1	192 (23.5)	53 (25.7)	51 (24.8)	48 (23.7)	40 (19.8)
≥2	42 (5.1)	9 (4.4)	5 (2.4)	15 (7.4)	13 (6.4)
Infant sex*
Male	362 (46.1)	98 (49.3)	87 (43.9)	87 (44.4)	90 (46.9)
Female	423 (53.9)	101 (50.8)	111 (56.1)	109 (55.6)	102 (53.1)
Fetal sac number
1	602 (73.7)	157 (76.2)	147 (71.4)	156 (76.9)	142 (70.3)
≥2	215 (26.3)	49 (23.8)	59 (28.6)	47 (23.2)	60 (29.7)
Preterm delivery	356 (43.6)	90 (43.7)	95 (46.1)	75 (37.0)	96 (47.5)
Number of embryos transferred	2.0 (2.0–3.0)	2.0 (2.0–2.0)	2.0 (2.0–3.0)	2.0 (2.0–3.0)	2.0 (2.0–3.0)
Peak estradiol (pg/mL)	2201.0 (1401.0–3330.0)	1613.5 (1050.0–2287.0)	2152.5 (1414.0–3326.0)	2352.0 (1472.0–3401.0)	2955.0 (1997.0–4907.0)
Number of oocytes retrieved	11.0 (7.0–17.0)	8.5 (6.0–12.0)	11.0 (7.0–17.0)	11.0 (8.0–16.0)	15.0 (9.0–20.0)
Number oocytes fertilized	11.0 (7.0–16.0)	8.0 (6.0–12.0)	11.0 (7.0–16.0)	11.0 (8.0–16.0)	14.0 (9.0–19.0)
Number of embryos frozen	0.0 (0.0–2.0)	0.0 (0.0–2.0)	0.0 (0.0–2.0)	0.0 (0.0–1.0)	0.0 (0.0–2.0)
Cycles that used ICSI	366 (44.8)	87 (42.2)	91 (44.2)	97 (47.8)	91 (45.1)

Data presented as median (interquartile range) or n (%)* infant sex not reported in 32 cases

Birth weight

Table 2 shows the association between serum progesterone quartiles on the day of hCG administration and birth weight. While the mean birth weight was 150 g lower in the highest progesterone quartile compared to the lowest quartile, the overall difference in mean birth weight among the four quartiles was not statistically significant (p = 0.11), after adjusting for maternal age at cycle start and peak estradiol levels. There was also no significant difference in mean gestational age adjusted z-score among the four quartiles, after controlling for maternal age and peak estradiol (p = 0.06).

Table 2.

Mean birth weight and z-score stratified by quartiles of progesterone on the day of hCG trigger

	Birth weight						z-score
	Crude			Adjusteda			Crude			Adjusteda
	Mean	95% CI	p	Mean	95% CI	p	Mean	95% CI	p	Mean	95% CI	p
Progesterone quartile (ng/mL)			0.07			0.11			0.08			0.06
≤0.54	3245	3158, 3333		3242	3152, 3332		0.46	0.34, 0.57		0.47	0.35, 0.59
>0.54–≤0.81	3117	3029, 3204		3116	3029, 3204		0.28	0.16, 0.39		0.28	0.16, 0.39
>0.81–≤1.17	3167	3079, 3255		3164	3076, 3253		0.34	0.22, 0.45		0.34	0.22, 0.45
>1.17	3087	2999, 3175		3093	3002, 3183		0.26	0.14, 0.38		0.25	0.13, 0.37

^aProgesterone models are adjusted for maternal age and peak estradiol levels

When stratifying by the clinically relevant progesterone level, we found a significant difference in mean birth weight even after adjusting for maternal age and peak estradiol (Table 3; p = 0.007). There was, however, no significant difference in mean z-score between cycles with serum progesterone >2.0 and ≤2.0 ng/mL (p = 0.07). The median gestational age for participants with a progesterone >2.0 ng/mL was 36.9 weeks (34.0–38.0) while it was 37.3 weeks (35.7–38.6) for those with progesterone ≤2.0 ng/mL. The incidence of LBW was significantly higher in cycles with serum progesterone >2.0 ng/mL (27.3%) compared to those ≤2.0 ng/mL (14.3%; p = 0.04). There was, however, no significant difference in the incidence of small for gestational age infants between cycles with serum progesterone >2.0 (3.0%) compared to those ≤2.0 (3.8%); p = 1.0).

Table 3.

Mean birth weight and z-score stratified by clinically relevant progesterone level on the day of hCG trigger

	Birth weight						z-score
	Crude			Adjusteda			Crude			Adjusteda
	Mean	95% CI	p	Mean	95% CI	p	Mean	95% CI	p	Mean	95% CI	p
Progesterone			0.006			0.007			0.07			0.07
P4 ≤2.0 ng/mL	3167	3122, 3212		3167	3122, 3211		0.34	0.28, 0.40		0.34	0.29, 0.40
P4 >2.0 ng/mL	2854	2636, 3072		2860	2642, 3079		0.07	−0.21, 0.36		0.07	−0.22, 0.36

^aProgesterone models are adjusted for maternal age and peak estradiol levels

There was no difference in mean birth weight or z-score across the estradiol quartiles after adjusting for maternal age and progesterone levels on the day of hCG administration (Supplemental Table 1; p = 0.18 and 0.29, respectively).

Discussion

The clinical recommendation of delaying transfer in a fresh IVF cycle when progesterone levels are elevated has recently been proposed as a way of maintaining higher pregnancy rates [24, 25]. The practice has, however, come under some scrutiny [26–28], even though a large meta-analysis of over 60,000 cycles indicated that higher progesterone levels are detrimental to pregnancy rates [13]. This meta-analysis indicated that even using a threshold ≥0.8 ng/mL, there was a significant association with a decreased probability of pregnancy in fresh IVF cycles. The elevated progesterone threshold levels used by clinics are, however, very dependent on individual clinic thresholds and are not conclusive [13]. We therefore used a cutoff of 2.0 ng/mL as a clinically relevant high level for our analysis. Our study indicates that in addition to the recommendation of delaying transfer when progesterone is elevated, there may be a higher progesterone level where the impact may be on fetal development itself and translate to lower birth weights. In this case of extremely high progesterone on the day of hCG administration, it may be prudent to delay transfer.

The exact mechanism of LBW in IVF pregnancies has not been clearly elucidated. Factors intrinsic to the subfertile couple, as well as the supraphysiological hormone levels, have been suggested as possible mechanisms. We know that estradiol plays a key role in the morphology and functional differentiation of syncytiotrophoblasts, as well as modulation of uteroplacental blood flow critical for optimal fetal growth in primates [29–31]. It also appears that low levels of estradiol as seen in early pregnancies are required to permit the normal placental invasion of uterine spiral arteries [32]. There is evidence that not only is an elevated estradiol level toxic to implantation in mice [33–35], but it also results in aberrant placentation [32, 33]. While our study does not demonstrate an association between elevated estradiol on the day of hCG administration and LBW, several studies have demonstrated this association [12, 36–38].

On the other hand, elevated progesterone on the day of hCG administration is associated with differential vascular endothelial growth factor expression in the endometrium and this is thought to alter endometrial receptivity resulting in decreased implantation [39]. We therefore hypothesized that elevated progesterone may be associated with adverse perinatal outcomes. We grouped our cycles into quartiles according to progesterone levels on the day of hCG administration. We chose those particular quartiles because using similar quartile cutoffs, we previously have shown a negative association with live birth [23].

Our results from 817 cycles demonstrated a decreasing live birth weight with increasing progesterone levels, in particular at higher levels. We also found that cycles with progesterone >2.0 ng/mL had lower mean birth weights than cycles with lower levels of progesterone.

In contrast to fresh embryo transfers, several studies have suggested better outcomes among frozen embryo transfer cycles that utilized an embryo obtained from a cycle with elevated progesterone on the day of hCG administration [20, 25]. As more data emerges regarding the improved perinatal outcomes with frozen as opposed to fresh embryo transfer cycles, including a lower incidence of preterm delivery and small for gestational age neonates, there may be fewer reasons to proceed with a fresh embryo transfer if cryopreservation is an option [10, 24] when progesterone levels are elevated on the day of hCG administration.

One limitation of our study is a lack of access to all obstetric records of the cycles included in our analysis. For example, we cannot comment on the rates of preeclampsia based on progesterone quartiles. In addition, we recognize that the biggest limitation of our study is the small sample size, which prevents us from making definitive conclusions. However, we are unaware of other published reports on this association. In addition, the utilization of z-scores assumes that they are representative of an IVF population. That is not necessarily the case as the distribution of IVF patients would be older than the control population used to establish the z-scores. We therefore conclude that based on these findings and the many studies analyzed by Venetis et al. [40], it seems prudent to exercise caution when performing fresh embryo transfers when progesterone levels are elevated on the day of hCG administration. We await further studies to critically evaluate the association observed in this study.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.