Weekly changes in serum β -human chorionic gonadotropin,estradiol,and progesterone levels for pregnancy assessment in women with unexplained recurrent miscarriage

Abstract

Objective

To investigate the efficacy of weekly measurement of β-human chorionic gonadotropin, estradiol, and progesterone levels for pregnancy assessment in patients with unexplained recurrent miscarriage.

Methods

We retrospectively included patients with unexplained recurrent miscarriage whose serum β-human chorionic gonadotropin, estradiol, and progesterone levels were measured three or more times during early pregnancy. The patients were divided into two groups according to pregnancy outcomes. The Δβ-human chorionic gonadotropin, Δestradiol, and Δprogesterone levels of the two groups were compared, and the corresponding cutoff values were determined using receiver operating characteristic analyses for predicting pregnancy outcome.

Results

A total of 140 patients were analyzed. From the 6th week to 10th week, the serum β-human chorionic gonadotropin and estradiol levels of the ongoing pregnancy group were significantly higher than those of the miscarriage group. The serum Δβ-human chorionic gonadotropin levels identified ongoing pregnancy or miscarriage with an area under the receiver operating characteristic curve of 0.841 (95% confidence interval: 0.769–0.914) and diagnostic cutoff value of 30,633 mIU/mL. The Δestradiol levels identified ongoing pregnancy or miscarriage with an area under the receiver operating characteristic curve of 0.839 (95% confidence interval: 0.752–0.926) and a cutoff value of 45 pg/mL.

Conclusions

Weekly remeasurement of serum β-human chorionic gonadotropin and estradiol may be useful for pregnancy risk assessment in patients with unexplained recurrent miscarriage in clinical practice.

Keywords

β-Human chorionic gonadotropin estradiol recurrent spontaneous abortion dynamic measurement pregnancy outcome

Introduction

Recurrent miscarriage refers to three or more consecutive losses before 20 weeks of gestation according to the World Health Organization¹; it affects 1%–2% of all fertile women who become pregnant.² In clinical practice, it has been accepted that two consecutive losses, rather than three losses, is sufficient for the diagnosis of recurrent miscarriage on the basis of similar recurrence rates.³ However, predisposing factors could not be identified in more than half of the women with recurrent pregnancy loss even after a thorough examination.⁴ Therefore, women with a history of recurrent miscarriage are often distressed, and additional hospital appointments are considered crucial to provide reassurance during a subsequent pregnancy.⁵

Subsequent pregnancies in women with recurrent miscarriage are usually monitored initially by measuring the serum level of β-human chorionic gonadotropin (β-hCG) and later by serial measurements in a more frequent manner considering the fear of miscarriage again. To date, there have been a number of previous studies investigating biomarkers such as leptin,⁶ inhibin A, and activin A⁷ for predicting pregnancy outcomes in patients with a history of recurrent miscarriage. However, no generally accepted markers have been identified.

Serum β-hCG has been widely used in early pregnancy assessment. Studies have also suggested that serum progesterone (Prog)^8,9 and estradiol (E2)¹⁰ can identify women with miscarriage. Nevertheless, whether serum β-hCG, E2, and Prog can usefully predict the outcome of a pregnancy remains debatable. In the clinic, the evaluation of serum hormone levels, including β-hCG, E2, and Prog, is generally conducted at only a single point during a normal pregnancy. However, their levels are dynamic during pregnancy, and more frequent visits are needed to determine the serial levels at multiple time points in patients with recurrent miscarriage. Regarding the pregnancy outcome of this specific population, the utility of dynamic changes in the serum levels of β-hCG, E2, and Prog as predictive markers has rarely been reported thus far. In this study, we aimed to determine whether weekly changes in β-hCG, E2, and Prog levels, whose measurement method was newly developed in the present study, measured during pregnancy can predict pregnancy outcomes in patients with unexplained recurrent miscarriage.

Methods

Patients

This retrospective study was approved by the Ethics Committee of Xiamen Maternity and Child Care Hospital (No. KY-2024-085-K01). Patients with recurrent miscarriage who visited the family planning clinic of Xiamen Maternity and Child Care Hospital (Xiamen, China) between 1 January 2020 and 31 October 2022 were consecutively enrolled. The inclusion criteria for this study were as follows: (a) intrauterine pregnancy confirmed by a clinician; (b) more than two consecutive miscarriages before 20 weeks of gestation; (c) conception with the same partner; (d) no live births subsequent to the consecutive miscarriages; and (d) three or more consecutive measurements of serum levels of β-hCG, E2, and Prog during early pregnancy. The exclusion criteria were as follows: (a) abnormal karyotypes of both partners; (b) the existence of any anatomical abnormality likely to be responsible for miscarriage; (c) antiphospholipid syndrome; and (d) the existence of factor V Leiden, prothrombin G20210A mutations, protein S and C, and antithrombin deficiencies.¹¹ Additionally, patients with incomplete clinical data and who were already diagnosed with ectopic pregnancy or biochemical pregnancy were excluded from the study.

Study design

The patients were classified into two groups according to pregnancy outcomes (ongoing pregnancy or miscarriage) at the 20th week of gestation. Patient data, such as age; reproductive history; last menstrual period; the first three consecutive measurements of serum levels of β-hCG, E2, and Prog since the initial visit; findings of early pregnancy ultrasound examination; and clinical pregnancy outcome during the follow-up period until the 20th week of gestation, were obtained from medical records. All patient details have been de-identified. The reporting of this study conforms to STROBE guidelines.¹² Gestational age was calculated based on the regular last normal menstrual period and then confirmed or corrected according to ultrasound. The weekly change in serum hormone level (Δβ-hCG, ΔE2, and ΔProg) was calculated according to the following formula: ((serum level at third visit) – (serum level at initial visit))/interval days × 7.

Statistical analysis

Statistical analysis of the clinical data was performed using SPSS Statistics version 24.0 (SPSS Inc., Chicago, IL, USA). Continuous data were compared using the Mann–Whitney U test. Categorical data were compared using the chi-square test or Fisher’s exact test, where appropriate. A receiver operating characteristic (ROC) curve was constructed to determine an appropriate cutoff value for weekly changes in serum hormone levels to predict pregnancy outcomes during the follow-up period until the 20th week of gestation. P-values of <0.05 were considered to indicate statistical significance.

Results

Clinical characteristics

A total of 140 women with unexplained recurrent miscarriage were included in this study. The age of the patients ranged from 19 to 43 years. Among them, 106 (75.71%) were assigned to the ongoing pregnancy group. Additionally, 34 (24.29%) women suffered from miscarriage and were included in the miscarriage group. There were no significant differences between the two groups in terms of age (P = 0.498), gravidity (P = 0.316), parity (P = 0.604), miscarriage (P = 0.395), or cesarean section (P = 1.000). The characteristics of the two groups are summarized in Table 1.

Table 1.

Clinical data of the study population.

Characteristics	Ongoing pregnancy(n = 106)	Miscarriage(n = 34)	P
Age, years	31.42 ± 3.91	31.76 ± 4.45	0.498
Gravidity
2	51 (48.11)	14 (41.18)	0.316
3	34 (32.08)	9 (26.47)
≥4	21 (19.81)	11 (32.35)
Parity
0	66 (62.26)	21 (61.76)	0.604
1	35 (33.02)	10 (29.41)
≥2	5 (4.72)	3 (8.82)
Number of miscarriages
2	78 (73.59)	22 (64.71)	0.395
3	18 (16.98)	6 (17.65)
≥4	10 (9.43)	6 (17.65)
Previous cesarean section
Yes	9 (8.49)	2 (5.88)	1.000
No	97 (91.51)	32 (94.12)

Continuous data were presented as the mean ± SD, and categorical data were presented as count and percentage.

Serum β-hCG, E2, and Prog levels at different gestational weeks and changes in their absolute values over time per patient

We compared serum β-hCG, E2, and Prog levels weekly between the two groups. As shown in Table 2, at the 3rd–5th weeks of gestation, the serum β-hCG and E2 levels of the miscarriage group and ongoing pregnancy group had no remarkable statistical differences, whereas at 6th–10th weeks of gestation, the serum β-hCG and E2 levels of the ongoing pregnancy group were significantly higher than those of the miscarriage group. The changes in serum β-hCG, E2, and Prog levels over the indicated weeks are presented in Figure 1. The changes in the absolute values of serum β-hCG, E2, and Prog levels over time per patient are shown in Figure 2. Serum β-hCG and E2 levels appeared to increase with gestational age, whereas no such association was observed for Prog.

Table 2.

Comparison of serum β-hCG, estradiol, and progesterone levels between the ongoing pregnancy group and miscarriage group at different gestational weeks.

Gestational week	Case number		β-HCG (mIU/mL)					Estradiol (pg/mL)					Progesterone (ng/mL)
	Ongoing pregnancy	Miscarriage	Ongoing pregnancy		Miscarriage		P	Ongoing pregnancy		Miscarriage		P	Ongoing pregnancy		Miscarriage		P
	Ongoing pregnancy	Miscarriage	Median	IQR	Median	IQR	P	Median	IQR	Median	IQR	P	Median	IQR	Median	IQR	P
3rd	11	7	112	106	165	1047	0.375	237	183	221	129	0.328	24.41	8.87	27.48	12.90	0.179
4th	60	22	1719	3047	1886.50	5256	0.457	348	247	337	132	0.241	23.58	10.03	19.41	10.11	0.003
5th	81	26	10339	18605	8246.50	9257	0.212	448	301	427	272	0.839	22.75	13.97	18.02	7.39	0.058
6th	80	25	46368	62029	17623	21174	0.001	594	474	408	242	0.001	20.1	10.17	17.08	10.18	0.068
7th	50	16	98606	103261	24717.50	54138	<0.001	722	526	361	294	<0.001	16.00	8.54	12.26	10.59	0.006
8th–10th	36	6	140419	69980	51042	80467	<0.001	1106	1105	639	453	0.012	17.44	5.32	18.92	10.55	0.875

β-hCG: β-human chorionic gonadotropin; IQR: interquartile range.

Figure 1.

Scatter diagram and trend line of β-hCG, E2, and Prog levels in the ongoing pregnancy group and miscarriage group. (a) β-hCG; (b) E2 and (c) Prog. β-hCG: β-human chorionic gonadotropin; E2: estradiol; Prog: progesterone.

Figure 2.

Change in absolute values of serum β-hCG, E2, and Prog levels between each measurement during early pregnancy. Levels at each point were plotted. The X-axis shows the days between measurements. Values of the ongoing pregnancy group are presented in the upper row, whereas those of the miscarriage group are presented in the bottom row. (a) β-hCG; (b) E2; (c) Prog. β-hCG: β-human chorionic gonadotropin; E2: estradiol; Prog: progesterone.

ROC analyses to determine the cutoff value of Δβ-hCG, ΔE2, and ΔProg in predicting pregnancy outcomes

We calculated the weekly changes in serum hormone levels (Δβ-hCG, ΔE2, and ΔProg) and found that the levels of Δβ-hCG (ongoing pregnancy: median, 37,990 mIU/mL; interquartile range (IQR), 35,109 mIU/mL; miscarriage: median, 8613 mIU/mL; IQR: 15,822 mIU/mL; P < 0.001) and ΔE2 (ongoing pregnancy: median, 217 pg/mL; IQR, 259 pg/mL; miscarriage: median, 24 pg/mL; IQR, 122 pg/mL; P < 0.001) were significantly different between the two groups and may be correlated with pregnancy outcomes during the follow-up period. However, ΔProg was not significantly different (ongoing pregnancy: median, −0.47 ng/mL; IQR, 4.75 ng/mL; miscarriage: median, −0.50 ng/mL; IQR, 5.41 ng/mL; P = 0.825). The ROC curves of Δβ-hCG, ΔE2, and ΔProg for achieving ongoing pregnancy at the 20th gestational week are shown in Figure 3. For Δβ-hCG, the cutoff value was 30,633 mIU/mL, with the corresponding Youden index being 0.612. The area under the curve was 0.841, which was notably accurate for predicting ongoing pregnancy. For ΔE2, the cutoff value was 45 pg/mL, with the corresponding Youden index being 0.621. The area under the curve was 0.839, which was notably accurate for predicting ongoing pregnancy. For ΔProg, the cutoff value was −2.22 ng/mL, with the corresponding Youden index being 0.119. The area under the curve was 0, which was not accurate for predicting ongoing pregnancy (Table 3).

Figure 3.

ROC curve of Δβ-hCG, ΔE2, and ΔProg for achieving ongoing pregnancy at the 20th gestational week. ROC: receiver operating characteristic; β-hCG: β-human chorionic gonadotropin; E2: estradiol; Prog: progesterone.

Table 3.

Cutoff points of Δβ-hCG, ΔE2, and ΔProg and performance in predicting pregnancy outcomes.

Performance	Δβ-hCG	ΔE2	ΔProg
Cutoff	30,633 mIU/mL	45 pg/mL	−2.22 ng/mL
Accuracy (95% CI)	0.721 (0.638–0.792)	0.864 (0.794–0.914)	0.636 (0.550–0.714)
Sensitivity (95% CI)	0.971 (0.829–0.998)	0.706 (0.523–0.843)	0.412 (0.251–0.592)
Specificity (95% CI)	0.642 (0.542–0.731)	0.915 (0.841–0.958)	0.708 (0.610–0.790)
PPV (95% CI)	0.465 (0.347–0.586)	0.727 (0.542–0.861)	0.311 (0.186–0.468)
NPV (95% CI)	0.986 (0.911–0.999)	0.907 (0.831–0.952)	0.789 (0.691–0.863)
AUC (95% CI)	0.841 (0.769–0.914)	0.839 (0.752–0.926)	0.513 (0.402–0.623)
Youden index	0.612	0.621	0.119

β-hCG: β-human chorionic gonadotropin; E2: estradiol; Prog: progesterone; PPV: positive predictive value; NPV: negative predictive value; AUC: area under curve; CI: confidence interval.

Discussion

This observational study retrospectively investigated the association between the serum Δβ-hCG, ΔE2, and ΔProg levels and pregnancy outcome in patients with recurrent miscarriage. The results showed that Δβ-hCG and ΔE2 levels were higher in the ongoing pregnancy group than in the miscarriage group. Normal pregnancy and miscarriage can be distinguished using Δβ-hCG (cutoff value: 30,633 mIU/mL) and ΔE2 (cutoff value: 45 pg/mL) levels; however, ΔProg does not provide useful differentiation. Furthermore, Δβ-hCG and ΔE2 may serve as predictors of pregnancy outcomes in patients with recurrent miscarriage.

Previous studies have also reported that the levels of these hormone indicators indicate pregnancy outcomes, which is consistent with the present study results to some extent. For β-hCG, it has been broadly accepted that a lower β-hCG level is associated with a poor pregnancy outcome and that the dynamic level of β-hCG may serve as a better indicator, with a growth rate of 66% per 48 h, in predicting pregnancy outcome.¹³ For E2, Deng et al.¹⁴ reported that serum levels of E2 at 5–6 weeks and 7–9 weeks of gestation could detect miscarriage with an area under the ROC curve of 0.866 (95% confidence interval (CI): 0.793–0.938, P = 0.000) and 0.709 (95% CI: 0.616–0.801, P < 0.001) and diagnostic cutoff values of 320 and 576 pg/mL, respectively. They also combined the dual markers and three markers with Prog and β-hCG, demonstrating that the combination can be used to better predict miscarriage. Another study conducted by Li et al. concluded that low values and low growth rates of E2 indicate poor pregnancy outcomes, with cutoff values of 489.5 pg/mL in the 5th and 6th weeks, 590.5 pg/mL in the 7th week, and 614.5 pg/mL in the 8th week.¹⁰ For Prog, its clinical value in miscarriage risk assessment remains controversial due to large individual differences and insufficient evidence regarding appropriate cutoff levels for risk stratification.^10,14,15

Nonetheless, it has rarely been reported whether the dynamic changes in these three hormone levels within individual patients can predict pregnancy outcomes. Patients with recurrent miscarriage are willing to visit the clinic more frequently during pregnancy. Women and doctors commonly and easily test serum E2, Prog, and β-hCG to evaluate the pregnancy status. Thus, we chose to assess the dynamic change in these three hormone levels, rather than a single test at one time point, in this study. Additionally, the change was adjusted for the time interval within individual patients in the study population. The use of dynamic monitoring of serum hCG and estrogen led to early prediction of pregnancy outcome and thus had many benefits, including eliminating the differences in β-hCG and E2 levels in the same gestational week between individuals^16,17 and saving the costs of examinations through fewer, unplanned sonography tests. However, this study had several limitations. First, the results were limited by the inherent limitations of a retrospective study design via selection bias. Patients who were lost to follow-up or who sought treatment too late may have not been included in the study. Second, it was a single-institutional experience that included a small number of patients. This reduced the robustness of the statistical analysis. Third, the time interval between measurements of serum hormone levels during pregnancy varied among individuals, which might have influenced the analysis. To address these concerns, our results require external validation in a larger prospective multicenter cohort in the future.

In conclusion, we retrospectively demonstrated the value of weekly changes in serum β-hCG and E2 levels for pregnancy outcome assessment in women with unexplained recurrent miscarriage. An increase in the serum level of β-hCG above 30,633 mIU/mL per week and that of E2 higher than 45 pg/mL per week probably indicate favorable pregnancy outcomes. Therefore, dynamic monitoring of serum β-hCG and E2 may be recommended for pregnancy risk assessment in clinical practice.

Footnotes

Acknowledgements

Not applicable.

Author contributions

Jian An conceptualized the study. Jian An, Yan Wang, and Ruide Su obtained the data. Jian An and Ye Lu analyzed the data. Ruide Su, Yan Wang, and Ye Lu wrote the original draft. Bizhu Lin and Jian An reviewed and edited the manuscript. Bizhu Lin supervised the study. All authors have reviewed the manuscript. All authors have read and approved the final manuscript.

Data availability statement

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declaration of conflicting interests

The authors declare no conflicts of interest.

Ethics approval and consent to participate

This study was reviewed and approved by the Ethics Committee of Xiamen Maternity and Child Care Hospital. All methods were conducted in accordance with the relevant guidelines and regulations. This study was performed in compliance with the “Declaration of Helsinki.” The confidentiality and anonymity of the data as well as the voluntariness of participation were assured.

Funding

Not applicable.

ORCID iD

Jian An

References

Stirrat

GM.

Recurrent miscarriage. Lancet 1990; 336: 673–675.

Rai

Regan

Recurrent miscarriage. Lancet 2006; 368: 601–611.

Branch

Gibson

Silver

RM.

Clinical practice. Recurrent miscarriage. N Engl J Med 2010; 363: 1740–1747.

Speroff

Fritz

MA.

Recurrent early pregnancy loss. In: Clinical gynecologic endocrinology & infertility. 7th ed. Philadelphia, PA: Lippincott Williams & Wilkins, 2005; ch 28 P1069–1103.

Meaney

Everard

Gallagher

, et al. Parents’ concerns about future pregnancy after stillbirth: a qualitative study. Health Expect 2017; 20: 555–562.

Laird

Quinton

Anstie

, et al. Leptin and leptin binding activity in women with recurrent miscarriage: correlation with pregnancy outcome. Hum Reprod 2001; 16: 2008–2013.

Prakash

Laird

Tuckerman

, et al. Inhibin A and activin A may be used to predict pregnancy outcome in women with recurrent miscarriage. Fertil Steril 2005; 83: 1758–1763.

Allen

Jr Lek

, et al. Serum progesterone distribution in normal pregnancies compared to pregnancies complicated by threatened miscarriage from 5 to 13 weeks gestation: a prospective cohort study. BMC Pregnancy Childbirth 2018; 18: 360.

Lek

Allen

Jr , et al. Validation of serum progesterone <35 nmol/L as a predictor of miscarriage among women with threatened miscarriage. BMC Pregnancy Childbirth 2017; 17: 78.

10.

Zhang

, et al. Significance of dynamically monitoring serum estrogen and β-human chorionic gonadotropin in early pregnancy assessment. J Clin Lab Anal 2021; 35: e23559.

11.

Pasquier

De Saint Martin

Bohec

, et al. Enoxaparin for prevention of unexplained recurrent miscarriage: a multicenter randomized double-blind placebo-controlled trial. Blood 2015; 125: 2200–2205.

12.

Von Elm

Altman

Egger

; STROBE Initiativeet al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Ann Intern Med 2007; 147: 573–577.

13.

Chu

Hardy

Beeson

, et al. What is the best method for managing early miscarriage? BMJ 2020; 368: l6438.

14.

Deng

Sun

, et al. Prediction of miscarriage in first trimester by serum estradiol, progesterone and β-human chorionic gonadotropin within 9 weeks of gestation. BMC Pregnancy Childbirth 2022; 22: 112.

15.

Puget

Joueidi

Bauville

, et al. Serial hCG and progesterone levels to predict early pregnancy outcomes in pregnancies of uncertain viability: a prospective study. Eur J Obstet Gynecol Reprod Biol 2018; 220: 100–105.

16.

Tanbo

Zucknick

Eskild

Maternal concentrations of human chorionic gonadotrophin in very early IVF pregnancies and duration of pregnancy: a follow-up study. Reprod Biomed Online 2018; 37: 208–215.

17.

Tulchinsky

Hobel

Yeager

, et al. Plasma estrone, estradiol, estriol, progesterone, and 17-hydroxyprogesterone in human pregnancy. I. Normal pregnancy. Am J Obstet Gynecol 1972; 112: 1095–1100.