Abstract
Several interactions between thyroid function and the oocyte physiology have been described. 1,2 On the one hand, thyroid hormone (TH) and thyrotropin (TSH) may act on the oocytes through their respective receptors present on granulosa/cumulus cells and, on the other hand, TH facilitates the effects of the follicle stimulation hormone (FSH) on granulosa cells. 1,2
Therefore, subclinical hypothyroidism and thyroid autoimmunity (TAI) may lead to impaired ovarian stimulation (OS) in vitro outcomes. 3 However, some OS outcomes such as the oocytes retrieval and oocytes cryopreservation (OOR and OCP) rate were never investigated outside an infertility setting. Women planning OCP for anticipated gamete exhaustion (AGE) are a priori fertile 4 and, therefore, we performed this retrospective pilot study to investigate the association between thyroid function (TSH and free thyroxine [fT4]) and the OOR/OCP rate in women consulting for AGE.
Since the start of OCP in our center in 2016 and until the end of 2021, 95 women performed AGE, after having received the agreement of a multidisciplinary concertation team, composed of gynecologists, nurses, embryologists, and clinical psychologists specialized in reproductive medicine. After the exclusion of 21 women treated with levothyroxine (LT4) or unavailable TSH levels, the first cycle of 74 women were considered for analysis (Fig. 1). Thyroid parameters (TSH, fT4, and thyroid peroxidase antibodies [TPOAb]), demographic and OS characteristics and ovarian reserve parameters (FSH, anti-Mullerian hormone [AMH], and transvaginal ultrasound with antral follicle count [AFC]) were determined and noted before the start of AGE (details on the assays and the OS are given in Supplementary Data S1 and Supplementary Data S2).
Participant flow diagram. TSH, thyrotropin.
The study was approved by the institutional review board (CE/22-02-03).
Baseline characteristics, the rate of (high) OOR and (high) OCP, were compared between women with a serum TSH levels <2.5 and ≥2.5 mIU/L. Furthermore, in multivariable linear and logistic regression analyses, the impact of independent variables (serum TSH, fT4, age, body mass index, total OS dose, the use of a pituitary agonist, and AFC) on dependent outcomes ([high] OOR rate and [high] OCP rate) was investigated. The cutoffs for high OS dose (>2500 IU/cycle), low AFC (<7), high OOR rate (OOR ≥7), and high cryopreservation rate (OCP ≥75%) were based on reference articles in the field. 5 –7
Thyroid parameters, serum AMH, and FSH were measured one to two months before the first cycle. The type of OS was mainly based on an ovarian ultrasound with the determination of the AFC before each cycle and women's age.
The results showed that the median (interquartile range) age of the women was 35 (34–37) years, serum TSH 1.71 (1.21–2.30) mIU/L, FSH 7.0 (5.4–8.8) IU/L, AMH 1.7 (1.1–3.1) μg/L, and total OS dose 2080 ± 770 IU (Table 1). The OOR was 8 (5–10) (high OOR 59%) and OCP 5 (3–9) (high OCP 62%). No differences were present between the study (TSH) groups. AFC in all women was 9 (7–12) of whom 14% had an AFC <7; to initiate the OS, a pituitary agonist was used in 11% of women and the mean dose of gonadotrophins to stimulate the ovaries was 2080 ± 770 IU of whom 26% women needed a high dose; all data were comparable between study groups.
Baseline Demographic, Thyroid, Ovarian Reserve Parameters, and Details on First Cycle in All Women and According to Thyrotropin Subgroups in All Women and According to Thyrotropin Subgroups
Continuous data were expressed as median (IQR Q1–Q3) when not normally distributed or as mean ± SD otherwise.
Categorical data were presented as n (%) of cases.
Differences between groups were analyzed by chi-square or Fisher's exact test for categorical data and by a t-test or Mann–Whitney U test for continuous data. High FSH levels are defined as levels >IQR-3, based on the whole study cohort. Age-specific reference ranges for AMH are provided in the user instructions, and we considered low values as those <5th percentile. Data were missing for BMI in 23 women, for smoking habits in 34, for fT4 in 14, for TPOAb in 13, for FSH in 1 and for AMH in 2.
AFC, antral follicle count (n = 70); AMH, anti-Müllerian hormone (n = 72); BMI, body mass index (n = 51); FSH, follicle-stimulating hormone (n = 73); fT4, free thyroxine (n = 60); IQR, interquartile range; OCP, oocytes cryopreserved; OOR, oocytes retrieved; TAI, thyroid autoimmunity; TSH, thyrotropin; TPOAb, thyroid peroxidase antibodies (n = 61).
In the multivariable linear regression analysis, serum TSH was associated with OOR (beta coefficient 1.44 confidence interval [95% CI: 0.40–2.47]; p = 0.008) and in the logistic regression with high OOR (adjusted odds ratio [aOR] 2.38 [95% CI: 1.00–5.63]); p = 0.048 (Table 2). fT4 was not associated with OOR.
Multivariable Linear (Top) and Logistic Regression Analysis (Bottom)
p Values after adjusting for age, BMI, pituitary downsizing type, total ovarian stimulation and AFC. VIF values for confounders were all <5. Statistical tests were considered significant whenever p < 0.05 (bold values).
aOR, adjusted odds ratio; CI, confidence interval.
The reasons for our observation remain speculative. One explanation could be that in women with higher serum TSH levels, the impact on TSH receptors and/or through cross signaling on FSH receptors (FSHR) is more important. As mentioned, TSH receptors are present on oocytes and granulosa cells, and glycoprotein hormones (luteinizing hormone, human chorionic gonadotropin, FSH, and TSH) have a common α- and specific β-subunit with ∼40% sequence similarities, as do their respective receptors. 1,8 Furthermore, polymorphisms in the FSHR gene could increase the ovarian response to endogenous and exogenous TSH/FSH. 9
An argument against that hypothesis is that in one study, TSH supplementation had no short-term effect on the development of follicles and oocytes in human ovarian tissue. 10 Moreover, in a recent study in women with subfertility no association between TSH and OOR was observed, although that might also be explained by the inclusion of women with different degrees of ovarian reserve and subfertility causes. 11
The impact of TSH on the OOR rate persisted also after correction for TPOAb. In studies in assisted pregnancies, the OOR rate is determined by the ovarian reserve, women's age, and the OS dose. 12 In our linear regression, beyond serum TSH levels, only AFC tended to be associated with OOR, but not age (probably since the age range in our cohort of women was very narrow) or the OS dose (women were à priori not infertile).
In the multivariable linear regression analysis, serum TSH was also associated with OCP (beta coefficient 1.07 [95% CI: 0.11–2.02]; p = 0.029). fT4 was not associated with OCP.
TH contributes to oocyte maturation that begins with the germinal vesicle stage, through the vesicle breakdown, metaphase I (MI), to end with metaphase II (MII). 2,13 In a zebrafish and a rat model, propylthiouracil (PTU) induced a concentration-dependent increase of egg production with a concomitant decrease of mature oocyte size. 14,15 TH inhibits the apoptosis signaling pathway of BAX and caspase-3 16 and, therefore, TH deficiency may increase the number of atretic follicles. Since TH transporters and receptors are present on human granulosa cells, the effects of PTU may also be due to a direct action on the ovary by the reduced TH levels and not of elevated TSH levels. However, it remains to be proven whether these animal model data can be projected to the human situation. Furthermore, data on oocyte quality are indispensable to know if TH plays a role mainly on the oocyte quality or quantity.
Our study has some limitations. It is retrospective; some fT4 and TPOAb results are missing, and the diagnosis of TAI that was based on the presence of TPOAb only. The information on tobacco use was not available for all women, and/or was not quantified. Tobacco can decrease serum TSH levels in a dose-dependent way and may affect oocyte quality. 17,18 However, we did not have data on the oocyte quality due to a lack of clear-cut criteria in clinical practice. 6
Depending on the results of future and larger prospective studies, it might become possible that changing thyroid function could improve the OOR and/or OCP rate. But again, good criteria of oocyte quality are necessary to come to more objective conclusions.
Footnotes
Acknowledgments
The authors would like to thank Bénédicte Beckers (biologist PhD) and Carla Willems (MD) for their help on the collection of the data.
Authors' Contributions
K.G.P. and C.A. were involved in the study design, data collection, interpretation, drafting of the article; A.-L.T. was involved in the analysis, drafting of the article; P.K., F.V., and G.S. were involved in the study design, interpretation, critical review of the article; and all authors approved of the final version.
Author Disclosure Statement
K.G.P. received lecture fees from the IBSA Institut Biochimique SA, Berlin-Chemie AG and the Merck company between 2016 and 2022. A.-L.T., P.K., F.V., G.S., and C.A. declare that they have no competing interests.
Funding Information
No funding was received for this article.
Supplementary Material
Supplementary Data S1
Supplementary Data S2