Thyroid Disorders and In Vitro Outcomes of Assisted Reproductive Technology: An Unfortunate Combination?

Abstract

Background:

The impact of thyroid disorders on in vitro outcomes of assisted reproductive technology (ART) remains controversial. Therefore, the aim of our study was to investigate whether thyroid peroxidase antibodies (TPO-Abs)/thyroid autoimmunity (TAI) or thyroid function (serum thyrotropin [TSH])/subclinical hypothyroidism are associated with an altered number of oocyte retrieval (NOR), fertilization rate (FR), and embryo quality (EQ).

Methods:

Cross-sectional study in 279 women in a single center, comprising 297 cycles and 1168 embryos. In vitro data (NOR, FR, and EQ) were documented in two groups; one according to thyroid function in women without TAI (TSH ≤2.5 and >2.5 mIU/L) and one according to the presence/absence of TAI (determined by TPO-Abs). EQ was evaluated according to international criteria and classified as excellent/good and poor. Women treated with levothyroxine (LT4) were excluded. Furthermore, the impact of thyroid parameters on outcomes, normal NOR (>6 or 8) and high FR (>60%), was verified in a multivariable logistic regression model.

Results:

In women without TAI, 27% had TSH levels >2.5 mIU/L, the prevalence of TAI was 8%, and overall, 6% of women had TSH levels >4.2 mIU/L. NOR, FR, and EQ were comparable between study groups. In the regression analysis, women aged ≥30 years and receiving a high ovarian stimulation dosage (>2300 IU/cycle) had lower rates of normal NOR (odds ratio [OR] 0.18 [95% confidence interval, CI 0.04–0.72]; p = 0.016 and OR 0.17 [CI 0.06–0.48]; p < 0.001, respectively).

Conclusions:

Our results do not suggest an impact of thyroid antibodies/autoimmunity and (dys)function on ART in vitro outcomes.

Introduction

Thyroid hormones (TH) may interact directly with oocyte physiology via the presence of TH receptors in granulosa/cumulus cells and indirectly by facilitating the actions of the follicle stimulation hormone (FSH) on granulosa cells in terms of the steroidogenic activity, cellular morphology, the expression of the LH/CG receptor, and by protecting them against apoptosis via the PI3-K pathway (1 –3). Thyrotropin (TSH) receptors are also present on human ovarian tissue and granulosa cells, but their function needs to be investigated further (4).TH and thyroid antibodies are present in follicular fluid (FF) in concentrations that correlate with those in the serum, and recently, it has been suggested that thyroid peroxidase (TPO) might be present in granulosa cells (5 –8).

Therefore, thyroid antibodies can be cytotoxic for the zona pellucida and hamper the interaction between the oocyte and the sperm cell (9,10).

Subclinical hypothyroidism (SCH) and thyroid autoimmunity (TAI) have been associated with an impaired ovarian reserve, lower oocyte quality, decreased fertilization rates (FRs) and a lower rate of high-quality embryos (6,11 –16). However, this was not the case in other studies (7,17 –19). Heterogeneity in study results may be explained by differences in the anti-Müllerian hormone (AMH) assays, as well as the timing of the measurement and reference ranges of TSH, and in the definition of ovarian reserve and embryo quality (EQ).

Therefore, the aim of our study was to investigate whether thyroid peroxidase antibodies (TPO-Abs)/TAI or thyroid function (serum TSH)/SCH were associated with altered in vitro outcomes of an assisted reproductive technology (ART) procedure, and more in particular with the number of oocyte retrieval (NOR), the FR, and EQ.

Materials and Methods

Overall study design

All data were retrieved from the CHU St-Pierre clinic of medically assisted procreation (PMA), a public university hospital in Brussels, Belgium. The enrollment consultation is systematically completed with a biological analysis, including the measurement of TSH, free thyroxine (fT4), and TPO-Abs. Furthermore, demographic data, including the women's background and fertility (obstetrical) history, are also noted in the records.

We report here on data of a cross-sectional analysis of infertile women (period 01-01-2017/31-05-2018), nested within an ongoing prospective collection of women's infertility history and biological data. The determination of the background of the women was based on a history taken by the social workers that includes systematically the nationality at birth and the origin of the women.

From 631 women visiting the PMA clinic during the study period, 279 were included of whom 1168 embryos were scored for their quality. Women who consulted for social/oncological reasons (n = 35), in whom thyroid function or antibody tests were not complete, who performed their analyses more than three months before the pickup day (n = 287), and those treated with levothyroxine (LT4) before screening (n = 30) were excluded.

Figure 1 illustrates the study selection process in a flowchart.

FIG. 1.

Flowchart of the study selection process. ART, assisted reproductive technology; ATD, antithyroid drugs; BMI, body mass index; fT4, free thyroxine; LT4, levothyroxine; TPO-Abs, thyroid peroxidase antibodies; TSH, thyrotropin; TT, thyroid tests.

In both study groups (one according to the presence/absence of TAI and one in women without TAI with TSH levels ≤2.5 and >2.5 mIU/L), parameters were expressed as continuous data for TSH, fT4, TPO-Abs, women's and partners age, BMI, parity, number of cycles, units of ovarian stimulation (OS), oocyte retrieved, FRs, and embryos obtained, transferred, and frozen. Older age (≥30 years), obesity (≥30 kg/m²), smoking (yes/no), history of miscarriage (≥2), primary infertility, and EQ (excellent/good and poor) were expressed as categorical data.

EQ was assessed using criteria established in the Vienna Consensus workshop (20). Embryos were considered of good quality if they reached four-cell stage on day 2 after fertilization and eight-cell stage with Grades 4 or 5 on day 3. They were considered of poor quality if they arrested, degenerated, or only reached Grade 2; and they were considered fair if they fell in-between the first two other groups. We pooled good and fair embryos into one group.

The study was approved by the institutional review board (CE/19-01-16).

ART procedure

OS was performed with hMG (Menopur^®; Ferring), recombinant FSH (Puregon^®; MSD or Gonal-F^®; Merck-Serono or Bemfola^®; Gedeon Richter), or corifollitropin alfa (Elonva^®; MSD). The dose of gonadotropins was determined on an individual basis according to the woman's age, day three serum FSH value, and antral follicle count. Pituitary inhibition was obtained by a gonadotropin-releasing hormone (GnRH) analog (long or short protocol) (Suprefact^®; Sanofi-Aventis or Decapeptyl^®; Ipsen) or GnRH antagonist (Orgalutran^®; MSD or Cetrotide; Merck-Serono). Cycle monitoring was carried out using transvaginal sonography and laboratory assays. When three or more leading follicles reached 17 to 18 mm, 5000–10,000 IU of hCG (Pregnyl^®; MSD) was administered. Oocyte retrieval was performed transvaginally and ultrasound-guided 34 to 36 hours after hCG injection.

The collected cumulus–oocyte complexes (COCs) were counted and routinely incubated in an organ dish; then oocytes were stripped of cumulus cells with hyaluronidase and mechanical denudation. Semen was prepared by the gradient technique. Oocytes were subsequently inseminated using either conventional in vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI). After having excluded severe male factor infertility, the indications for IVF technique were tubal factor or unexplained infertility, and ovulation failure with all possible conventional methods. ICSI was preferred in case of previous fertilization failure with conventional IVF, retrieval of poor-quality oocytes or of a small number of oocytes, high titers of antisperm antibodies, and fertilization with frozen/thawed oocytes.

In idiopathic infertility during the first ART attempt, half IVF or half ICSI was preferred to identify a possible fertilization problem that would explain infertility.

Serum assay

All laboratory analyses were provided by the laboratory of hormonology at our institution.

Serum TSH was measured using a sandwich immunoassay with electrochemiluminescence detection on a Roche analytical platform (TSH; Roche Diagnostics, Mannheim, Germany). The coefficient of variation (intermediate precision) is 3.5% at 1 mIU/L, and reference values are 0.27–4.2 mIU/L (P2.5-P97.5).

Serum fT4 was measured using a competitive immunoassay with electrochemiluminescence detection on a Roche analytical platform (fT4; Roche Diagnostics). The coefficient of variation (intermediate precision) is 3.5% at 12.8 pmol/L. Reference values are 12–22 pmol/L (P2.5–P97.5).

Serum TPO-ab levels were measured using the Elecsys electrochemiluminescence immunoassays on a Cobas 6000 immunoanalyzer (Roche Diagnostics). The inter- and intra-assay CV for TPO-Abs is <5% and ≤7%, respectively, and the reference values are <34 kIU/L.

Statistical analysis

Data were stored in a Microsoft Excel database. Continuous data are expressed as median (Q1–Q3) when not normally distributed, and as mean ± SD for normally distributed data. Categorical data are presented as numbers (percentage) of cases. Differences between groups (TAI− vs. TAI+ and TSH ≤2.5 vs. TSH >2.5 mIU/L) were analyzed by chi-square or Fisher's exact test for categorical data and by the Mann–Whitney U test for continuous data. Correlations between continuous variables were quantified using Spearman's ρ correlation coefficient when data were not normally distributed, and a Pearson's correlation for normally distributed data.

For the logistic regression analysis, the independent variables were fT4 levels in the whole range, TSH 2.5–4.2 mIU/L, TSH >4.2 mIU/L, TAI (TPO-Abs ≥34 kIU/L), age ≥30 years, BMI ≥30 kg/m², tobacco use, male infertility (corrected for donor sperm), age of the male ≥40 years, type of pituitary downregulation, high OS/cycle (>2300 IU), ICSI as type of ART (without taking split cycles into account), and >1 previous ART cycle. Two dependent outcomes were taken into account: age-adapted normal oocyte retrieval rate (for women <36 years >8 oocytes and ≥36 years >6) and high FR (>60%) (20).

All statistical tests were considered significant whenever p < 0.05.

Results

Table 1 shows baseline demographic and thyroid parameters in the two study groups.

Table 1.

Baseline Demographic and Thyroid Parameters in the Two Study Groups

Thyroid function and autoimmune status	All women	TAI−	TAI+	p-Level	TAI–TSH ≤2.5 mIU/L	TAI–TSH >2.5 mIU/L	p-Level
Continuous data^a	n = 297	n = 272 (92%)	n = 25 (8.0%)		n = 199 (73%)	n = 73 (27%)
Categorical data (n (%))	n = 297	n = 272 (92%)	n = 25 (8.0%)		n = 199 (73%)	n = 73 (27%)
TSH (mIU/L)	1.93 (1.30–2.61)	1.88 (1.28–2.56)	2.27 (1.73–3.86)	<0.001	1.53 (1.17–2.03)	3.04 (2.72–3.73)	<0.001
2.51–4.20	68 (23%)	61 (22%)	7 (28%)	0.526	N/A	N/A
>4.2	17 (6%)	12 (4%)	5 (20%)	0.001	N/A	N/A
fT4 (pmol/L)	15.5 ± 2.0	15.5 ± 2.0	15.5 ± 2.2	0.900	15.5 ± 2.0	15.5 ± 1.9	0.945
TPO-Abs (kIU/L)	15 (15 –17)	15 (15–15)	150 (69–213)	<0.001	15 (15–16)	15 (15–15)	0.091
Age (years)	36.4 ± 5.0	36.4 ± 5.0	36.5 ± 4.3	0.949	36.4 ± 4.9	36.5 ± 5.3	0.861
≥30	264 (89%)	241 (89%)	23 (92%)	0.605	179 (90%)	62 (85%)	0.249
BMI (kg/m²)	25.0 ± 4.6	25.0 ± 4.7	24.6 ± 40	0.670	24.8 ± 4.8	25.6 ± 4.5	0.250
Obesity (BMI ≥30 kg/m²)^b	37 (13%)	35 (13%)	2 (9%)	0.481	24 (13%)	11 (15%)	0.512
Smoking	38 (13%)	36 (13%)	2 (8%)	0.453	28 (14%)	8 (11%)	0.502
Ethnic background women
Caucasian	137 (46%)	123 (45%)	14 (56%)	0.301	91 (46%)	32 (44%)	0.781
North African	55 (19%)	49 (18%)	6 (24%)	0.461	34 (17%)	15 (21%)	0.510
Sub-Saharan	75 (25%)	72 (27%)	3 (12%)	0.111	60 (30%)	12 (16%)	0.023
Others	30 (10%)	28 (10%)	2 (8%)	0.716	14 (7.0%)	14 (19%)	0.004
Age, male	40.9 ± 8.1	41.1 ± 8.1	39.0 ± 7.7	0.218	41.0 ± 8.2	41.3 ± 7.9	0.772
Smoking in male	85 (29%)	76 (28%)	9 (36%)	0.394	57 (29%)	19 (26%)	0.772

Significant values are indicated in bold.

Continuous data are expressed as mean ± SD when normally distributed, or as median (Q1–Q3) otherwise.

Data available in 260 women.

fT4, free thyroxine; Q1–Q3, quartile 1 and 3; N/A, not applicable; TAI, thyroid autoimmunity; TPO-Abs, thyroid peroxidase antibodies; TSH, thyrotropin.

Serum TSH and fT4 levels of all women included were 1.93 (1.30–2.61) mIU/L and 15.5 ± 2.0 pmol/L, respectively. Median (Q1–Q3) TPO-ab levels for all women were 15 (15 –17) kIU/L.

Eight percent of all women had TAI (n = 25). In women with TAI, 28% (n = 7) had TSH levels 2.5–4.2 mIU/L and 20% (n = 5) had TSH levels >4.2 mIU/L. The prevalence of TSH levels >4.2 mIU/L was significantly lower in women without TAI, 4% (n = 12); p = 0.001.

In women without TAI (n = 272), the majority had low-normal TSH levels ≤2.5 mIU/L (n = 199; 73%). By definition, serum TSH levels in women with TAI were significantly higher compared with those in women without TAI (2.27 [1.73–3.86] vs. 1.88 [1.28–2.56] mIU/L; p < 0.001).

Overall, the mean age of the women was 36.4 ± 5.0 of whom 89% were ≥30 years; 13% were obese and 13% smoked.

In women with a sub-Saharan background and without TAI, the prevalence of serum TSH levels ≤2.5 mIU/L was higher compared with TSH levels >2.5 mIU/L (30% vs. 16%; p = 0.023). The opposite was true in a pooled group of women with different backgrounds (Hispanic, Turkish, and Asian), in whom the prevalence of TSH levels >2.5 mIU/L was higher compared with TSH levels ≤2.5 mIU/L (19% vs. 7.0%; p = 0.004).

Table 2 shows details on infertility/obstetric history and in vitro outcomes in the two study groups.

Table 2.

Details on Infertility/Obstetric History and In Vitro Outcomes in the Two Study Groups

Thyroid function and autoimmune status	All women	TAI−	TAI+	p-Level	TAI–TSH ≤2.5 mIU/L	TAI–TSH >2.5 mIU/L	p-Level
Continuous data^a	n = 297	n = 272 (92%)	n = 25 (8.0%)		n = 199 (73%)	n = 73 (27%)
Categorical data (n (%))	n = 297	n = 272 (92%)	n = 25 (8.0%)		n = 199 (73%)	n = 73 (27%)
Parity	0 (0–1)	0 (0–0.5)	0 (0–1)	0.205	0 (0–1)	0 (0–1)	0.943
Miscarriage (≥2)	17 (5.7%)	16 (5.9%)	1 (4.0%)	0.698	14 (7.0%)	2 (2.7%)	0.182
Primary infertility^b	143 (52%)	129 (51%)	14 (56%)	0.646	88 (48%)	41 (60%)	0.079
Number of cycles	1 (1–2)	1 (1 –3)	1 (1–2)	0.370	1 (1 –3)	1 (1–2.5)	0.955
Cause of infertility
Female	86 (29%)	83 (31%)	3 (12%)	0.051	59 (30%)	24 (33%)	0.608
Male	75 (25%)	63 (23%)	12 (48%)	0.006	47 (24%)	16 (22%)	0.768
Mixed	98 (33%)	94 (34%)	4 (16%)	0.059	67 (33%)	27 (37%)	0.610
Idiopathic	38 (13%)	32 (12%)	6 (24%)	0.080	26 (13%)	6 (8%)	0.272
GnRH antagonist/agonist (%)	81/19%	78/22%	100/0%	0.072	78/22%	77/23%	0.878
Ovarian stimulation (IU)	1961 ± 966	2014 ± 982	1492 ± 669	0.056	2049 ± 1008	1906 ± 908	0.484
ART type used
ICSI	164 (55%)	147 (54%)	17 (68%)	0.179	109 (54%)	38 (52%)	0.690
IVF	62 (21%)	59 (22%)	3 (12%)	0.254	45 (23%)	14 (19%)	0.543
Split cycles	71 (24%)	66 (24%)	5 (20%)	0.632	45 (23%)	21 (29%)	0.294
Oocytes retrieved	6.5 ± 4.4	6.5 ± 4.3	7.1 ± 4.9	0.514	6.4 ± 4.2	6.7 ± 4.6	0.681
FR (%)	61.2 ± 29.8	61.4 ± 29.9	58.7 ± 29.8	0.664	60.4 ± 30.5	64.0 ± 28.1	0.386
Embryos obtained	3.9 ± 3.1	4.0 ± 3.1	3.8 ± 2.9	0.786	3.9 ± 3.1	4.1 ± 3.2	0.682
Embryos transferred	0.9 ± 0.8	0.8 ± 0.8	0.9 ± 0.6	0.692	0.8 ± 0.8	0.9 ± 0.8	0.929
Embryos frozen	0 (0–2)	0 (0–2)	0 (0–1.5)	0.742	0 (0–2)	0 (0–2)	1.000
Number of embryos for quality evaluation	1168	1074	94		776	298
Excellent/good	513 (44%)	476 (44%)	40 (43%)		355 (46%)	121 (41%)
Poor	651 (56%)	598 (56%)	54 (57%)	0.741	421 (54%)	177 (59%)	0.129

Significant values are indicated in bold.

Continuous data are expressed as mean ± SD when normally distributed, or as median (Q1–Q3) otherwise.

Data available in 277 women.

ART, assisted reproductive technology; FR, fertilization rate; GnRH, gonadotropin-releasing hormone; ICSI, intracytoplasmic sperm injection; IVF, in vitro fertilization.

The prevalence of women with primary infertility was 52% and comparable between study groups. The median (Q1–Q3) number of cycles women had before the one taken into account for this study was 1 (1–2) and comparable between groups.

In women with TAI, the prevalence in whom the cause of infertility was due to a male factor, was higher compared with that in women without TAI (48% vs. 23%; p = 0.006). ICSI was applied mostly as type of ART (55%), with a slightly (nonsignificant) higher percentage applied in women with TAI compared with that in women without TAI (68% vs. 54%; p = 0.179). GnRH antagonists were used mostly to initiate the ART procedure (81%) and the mean dosage of gonadotropins used to stimulate the ovaries was 1961 ± 966 IU (comparable between groups).

The mean (SD) number of oocytes retrieved was 6.5 ± 4.4, the FR was 61.2% ± 29.8%, the number of embryos obtained was 3.9 ± 3.1, that of transferred embryos 0.9 ± 0.8, and the median (Q1–Q3) of embryos frozen 0 (0–2). All results were comparable between groups.

The number of embryos scored for quality evaluation was 1168. Overall, 44% were scored as good to excellent and 56% as poor; results were comparable between groups.

The FRs in women with TAI only treated with ICSI (n = 16) versus IVF (n = 3) were comparable (59% vs. 63%; p = 0.806).

Table 3 shows the correlations between the in vitro outcome data and thyroid parameters expressed as continuous variables.

Table 3.

Correlations Between the In Vitro Outcome Data and Thyroid Parameters Expressed as Continuous Data

Thyroid parameter	TSH (mIU/L)		fT4 (pmol/L)		TPO-Abs (mIU/L)
In vitro outcomes	r	p	r	p	rho	P
Oocytes picked up	0.030	0.611	0.035	0.557	0.056	0.310
FRs (%)	0.032	0.603	−0.065	0.292	−0.053	0.380
Embryos obtained	0.032	0.578	−0.011	0.860	0.022	0.702
Embryos transferred	0.023	0.706	0.087	0.160	0.004	0.954
Embryos frozen	0.026	0.674	−0.115	0.064	0.031	0.614
Embryo quality
Excellent/good	−0.028	0.646	−0.062	0.315	−0.020	0.740
Poor	0.039	0.522	0.011	0.857	0.024	0.700

For the correlation between the outcomes and TSH/fT4, a Pearson's test was used, and for that with TPO-Abs, a Spearman's test.

None of the in vitro outcomes correlated with serum TSH, fT4, or TPO-ab levels.

For details and p-values see (Table 3).

Table 4 shows the results of the multivariable logistic regression analysis with thyroid, demographic, and ART characteristics as independent variables, and normal NOR and high FRs as dependent outcomes.

Table 4.

Multivariable Logistic Regression Analysis with Thyroid, Demographic, and Assisted Reproductive Technology Characteristics as Independent Variables, and Normal Number of Oocyte Retrieval and High Fertilization Rates as Dependent Outcomes

Dependent outcomes	Normal oocyte retrieval rate^a	High FR (>60%)
Independent variables	aOR [CI]	aOR [CI]
Independent variables	p	P
TSH >2.5–4.2 mIU/L^b	0.79 [0.31–2.05]	0.71 [0.11–4.73]
TSH >2.5–4.2 mIU/L^b	p = 0.632	p = 0.721
TSH >4.2 mIU/L	1.29 [0.17–9.52]	1.65 [0.04–68.65]
TSH >4.2 mIU/L	p = 0.804	p = 0.791
fT4 all levels	0.99 [0.81–1.24]	0.85 [0.62–1.17]
fT4 all levels	p = 0.995	p = 0.317
TAI (TPO-Abs ≥34 kIU/L)	0.44 [0.11–1.74]	0.42 [0.03–5.36]
TAI (TPO-Abs ≥34 kIU/L)	p = 0.243	p = 0.501
Age (≥30 years)	0.18 [0.05–0.75]	2.35 [0.29–19.05]
Age (≥30 years)	p = 0.018	p = 0.422
BMI (≥30 kg/m²)	0.89 [0.32–2.49]	0.99 [0.08–12.58]
BMI (≥30 kg/m²)	p = 0.823	p = 0.995
Tobacco use	1.29 [0.41–4.06]	0.62 [0.08–4.55]
Tobacco use	p = 0.667	p = 0.638
Caucasian background	0.99 [0.43–2.29]	0.49 [0.10–2.34]
Caucasian background	p = 0.989	p = 0.372
Male infertility^c	N/A	0.86 [0.08–8.77]
Male infertility^c	N/A	p = 0.899
Age male (≥40 years)	N/A	1.76 [0.30–10.44]
Age male (≥40 years)	N/A	p = 0.535
Pituitary suppression method	2.13 [0.65–6.96]	0.33 [0.01–18.02]
Pituitary suppression method	p = 0.211	p = 0.589
High ovarian stimulation (>2300 IU)	0.17 [0.06–0.48]	2.63 [0.34–20.40]
High ovarian stimulation (>2300 IU)	p < 0.001	p = 0.355
ICSI as type of ART^d	N/A	1.23 [0.13–11.76]
ICSI as type of ART^d	N/A	p = 0.858
ART cycle >1	0.93 [0.41–2.11]	0.27 [0.06–1.27]
ART cycle >1	p = 0.857	p = 0.098

Significant values are indicated in bold.

Normal oocyte retrieval rate is age dependent: for women <36 years this is >8 oocytes and ≥36 years >6.

Controls corrected with serum TSH levels >4.2 mIU/L.

Corrected for donor sperm.

Corrected for split ART cycles.

aOR, adjusted odds ratio; CI, 95% confidence interval; TAI, thyroid autoimmunity.

Normal NOR (>6 or >8 oocytes according to age) and high FRs (>60%) were not associated with any of the thyroid parameters. Variables associated with a normal NOR rate were age ≥30 years (odds ratio [OR]: 0.18 [95% confidence interval, CI 0.04–0.72]; p = 0.016) and a high OS per cycle (>2300 IU) (OR: 0.17 [CI 0.06–0.48]; p < 0.001).

Discussion

Women pregnant after an ART procedure with thyroid disorders have impaired live-birth rates compared with women without thyroid disorders. The precise underlying pathophysiological mechanism for that association remains unclear, but an impact of thyroid disorders on the earliest stages of the ART procedure—that is, on the number of oocytes retrieved, the FR, and the EQ—has been suggested (7,11 –15,21,22).

We did not observe an association between thyroid function (TSH, fT4), thyroid antibodies (TPO-Abs), and the NOR, FR, or EQ. Moreover, our results did not change when SCH (TSH >4.2 mIU/L) or TAI (TPO-Abs ≥34 kIU/L) was taken into account.

Concerning the absence of an association between TAI and NOR, our data are in line with those in several other studies (7,12,15,21,23 –26). The NOR is determined by a number of parameters, such as the ovarian reserve and stimulation, the underlying causes of infertility, and the women's age (27). In our study, age and other baseline characteristics were comparable between women with and without TAI, but in the logistic regression analysis, age ≥30 years was a negative predictor of a normal NOR. It has been hypothesized that TAI might have a negative impact on the NOR, by decreasing the ovarian reserve (14,16), although in a large number of other studies, this was not the case (7,15,18,19,24 –26).

Thyroid antibodies are present in FF in concentrations that correlate with those in the serum, and therefore may have a direct cytotoxic impact on the oocytes (5 –10). Opposite study results might have been due to differences in study populations (idiopathic low ovarian reserve and genetic causes, among others) and in the methodology used to determine ovarian reserve. It is noteworthy that the detrimental impact of TPO-Abs in the study by Korevaar et al. was independent of the FSH levels, and that there was no impact of Tg-Abs (16).

Since the presence of TAI is the most important cause of SCH, the impact of TAI on the NOR and ovarian should be adjusted for thyroid function (28,29). In our study, the NOR was not associated with thyroid (dys)function. To the best of our knowledge, the only other study that investigated the impact of thyroid function (all TSH levels and as a group with TSH >2.5 mIU/L) on the NOR was also negative (12).

In some studies, ovarian reserve showed a correlation with serum TSH (13,16,18). In the study by Polyzos et al., no correlation was present between TSH/fT4 and low ovarian reserve of all causes (18). However, when the authors investigated genetic causes of low ovarian reserve only, a significantly higher prevalence of (subclinical) hypothyroidism was present compared with women with unexplained low ovarian reserve (and adjusted for age) (18). In a study by Weghofer et al., AMH levels were significantly lower in women with serum TSH levels >3.0 mIU/L after adjustment for TAI and age (13). Finally, in a study by Korevaar et al., lower serum free triiodothyronine (fT3) levels were associated with a lower antral follicle count (16). T3 is the bioactive form of TH and mainly produced intracellularly via deiodination of T4 in peripheral tissues. Animal studies have shown that the ovary can concentrate iodine, produce local T3 through deiodination, and that T3 has an effect on estrogen availability, and maybe on other hormones in the FF (1,8,30,31).

FF is easily available during oocyte pickup and could represent an optimal source of noninvasive biochemical predictors of oocyte quality (32). However, the impact of TAI on oocyte quality was investigated only in one study, without revealing a significant association (7). It is noteworthy that oocyte quality is ill-defined due to a lack of consensus of the criteria, and therefore, in most studies (including ours) oocyte quality was not investigated as outcome.

FRs were comparable between women with and without TAI, in analogy with most data in the literature (7,15,22,25,33,34). However, in two studies, lower FRs were present in women with TAI (6,35). The authors speculated that TPO-Abs may bind to the surface of the egg and/or embryo and interfere with fertilization and subsequent embryo development. Furthermore, we did not observe an association between thyroid (dys)function (TSH/fT4 and SCH) and FRs. Our results were comparable with those in the study by Michalakis et al. (TSH values 2.5–4 mIU/L) (36). However, in the study by Cramer et al., TSH levels were inversely proportional to the FR, and in the study by Scoccia et al., lower FRs were present in hypothyroid women not treated optimally with LT4; in the meta-analysis by Velkeniers et al., LT4 treatment improved the FR in women with SCH (11,37,38). Differences in study results on the FR may be due to different definitions of thyroid dysfunction and/or the absence of information on TAI.

In some studies, ICSI has been associated with better FRs compared with that in IVF cycles (39). ICSI has been used for over 20 years, and high FRs have been reported since its implementation (40). The concept that in women with TAI, ICSI might be the better ART technique is based on the fact that, in contrast to IVF, it may overcome the impact of thyroid antibodies in the FF (hampering the interaction between the oocyte and the sperm cell) (6,39). However, in our logistic regression analysis, ICSI was not a predictor for a high FR. In the study by Andrisani et al., in which no difference in the FR was present in relation to TAI, IVF was used more frequently in women with TAI compared with that in women without TAI (15). In the two studies with a lower FR in women with TAI, Zhong et al. did not report on the type of ART used (the selection of fertilization program was based on the semen condition on the day the oocytes were collected), and in the study by Monteleone et al., the application of ICSI was comparable between women with and without TAI (6,35).

In the future, head to head studies should be performed comparing the FR in women with TAI treated either with IVF or with ICSI to clarify this issue.

A successful fertilization depends on a number of covariates besides the ART technique. The most important are the woman's age, the cause of infertility, and thyroid function. In a recent meta-analysis by Poppe et al., TSH levels were <3.0 mIU/L in all studies included (34). In the study by Zhong et al. (with a lower FR in women with TAI), 13.8% of women with TAI had an increased TSH level, which was only observed in 2.4% of women with TAI, but the authors mention that TSH levels were not determined for all patients, an important study limitation (35).

Finally, we assessed EQ in 1168 embryos and classified them as of good/moderate or poor quality according to expansion, inner cell mass, and trophectoderm criteria (20). No difference was observed in relation to thyroid parameters nor in the continuous or categorical analysis. Also in three other studies, EQ was not associated with TAI (7,17,23). However, in as many other studies, a negative impact of TAI (corrected for age and thyroid function) on the number of top embryos compared with that in women without TAI has been reported (6,12,15). Monteleone et al. suggested using ICSI in infertile women with TAI since it requires no interaction between the sperm cell and the zona pellucida (6). However, in the study by Weghofer et al., an impaired EQ in women with TAI was independent from the use of ICSI (12). In that study, serum TSH levels in the whole range (0.45–4.5 mIU/L) were not correlated with EQ, but in women with TSH levels >2.5 mIU/L, a higher rate of poor EQ was present (12). Scoccia et al. showed that hypothyroid women who were optimally treated with LT4 had a similar EQ with that in euthyroid women, and in a study by Kim et al., including women with SCH (TSH levels >4.0 mIU/L), LT4 treatment increased the number of top-quality embryos compared with that in the placebo group (33,37).

Differences in results on the EQ may have been due to heterogeneity in the definition of poor EQ. In our study and that by Weghofer et al., poor EQ included arrested and degenerated embryos, while in the study by Andrisani et al., arrested and degenerated embryos were excluded (12,15). Finally, it is also possible that the impact of thyroid antibodies/hormones on the EQ is erased due to the replacement of the FF after the pickup by a standard growth medium.

Overt hypothyroidism or severe SCH (TSH >4.0 mIU/L) can lead to infertility via different pathways such as hyperprolactinemia, menstrual irregularity, and others (22,28). Therefore, the American Society for Reproductive Medicine recommends screening infertile women for serum TSH and for TPO-Abs if the TSH is >2.5 mIU/L (41). The American Thyroid Association guidelines on thyroid disorders and pregnancy state that there is insufficient evidence to recommend for or against universal screening for altered TSH concentrations before conception, except in women planning ART (42). However, no position was taken on the testing of thyroid antibodies as such. Based on our in vitro study results, we do not add evidence that strengthens the systematic screening of TSH/TPO-Abs in infertile women (42 –44). However, screening may be important for the later phase of the ART procedure, since TPO-Abs are associated with the development of (subclinical) hypothyroidism after OS, and indicate which women treated with LT4 should adapt their dosage before OS (42 –44).

Indeed, Busnelli et al. showed that 80% of hypothyroid women pregnant after OS and with TAI had to increase their dosage of LT4 within the first seven weeks of pregnancy, compared with 40% without TAI (44). Women with TAI are also more prone to develop SCH during pregnancy, even when they were euthyroid in the first trimester of pregnancy (28,42). Moreover, it is known that the prevalence of women with the polycystic ovarian syndrome and idiopathic infertility have a higher prevalence of TAI compared with fertile women (22,28).

One minor study result was the lower and higher prevalence of TSH levels >2.5 mIU/L in women with a sub-Saharan background and in a group composed of women with different backgrounds (Hispanic, Turkish, and Asian).

In a previous study in pregnant women, we have shown that women with a sub-Saharan background have a lower prevalence of TAI, what might explain their lower prevalence of TSH levels >2.5 mIU/L (45). It is difficult to give a precise explanation why the prevalence of TSH levels >2.5 mIU/L was higher compared with that of TSH levels ≤2.5 mIU/L in the group of women including different backgrounds (Hispanic, Turkish, and Asian). Differences in the thyroid/pituitary set point may play a role, as it was shown in blood spots performed for birth screening that mean TSH levels differed according to the ethnicity of the babies (46).

Strengths of our study are the real-world setting, including women with different causes of infertility, treated with different ART types and with different ethnicities. Inclusion criteria were strict, women with alternative indications of ART, or treated with LT4, were excluded. The time frame in which thyroid function and autoimmunity was determined in relation to the day of pickup was limited to three months before OS; in other studies this time frame was not always mentioned, and when thyroid function is determined after OS, it might have changed (43,44). Our analysis of thyroid parameters in relation to the study outcomes was done in a continuous and categorical way, and data were collected in a prospective way.

Our study has also some limitations. The numbers of patients with TAI and SCH (TSH levels >4.2 mIU/L) is rather low, in part, due to the exclusion of women treated with LT4. Semen parameters were expressed in a semiquantitative way as male infertility or not.

The diagnosis of TAI was based on the presence of TPO-Abs only, and not testing for Tg-Abs may have led to misclassification, since isolated positivity for Tg-Abs may occur in 5% of infertile women (47). We did not measure TH, TPO-Abs, or other hormones such as steroids, estrogens, and progesterone in the FF, and we did not characterize ovarian reserve and quality.

In our study, the absence of any impact of thyroid parameters on the in vitro parameters might have been due to methodological issues. However, we believe that the effectiveness of the ART procedure in its totality, starting with the pituitary downregulation and ovarian hyperstimulation, the oocyte recruitment and cultivation in a special growth medium, and the ART techniques (IVF and ICSI) for the fertilization are all very efficient and as such may overcome the potential impact of thyroid disorders on NOR, FR, and EQ.

Finally, the importance of thyroid disorders might appear only once women are pregnant, and therefore, clinical pregnancy outcome data of this cohort of women need to be investigated further.

In conclusion, no impact of TAI or (dys)function was observed on the number of oocytes retrieved, FRs, and EQ in women undergoing ART.

Footnotes

Acknowledgments

The authors thank Bénédicte Beckers (biologist PhD) and Siham Benallal (MD) for their help on the collection of the data.

Author Disclosure Statement

K.P. received lecture fees from the IBSA Institut Biochimique SA (satellite meeting of the European Thyroid Association) in 2016 and the Berlin-Chemie AG company (ETA educational thyroid meeting) in 2017, 2018, and 2019.

Funding Information

No specific grant/fellowship from any funding agency in the public, commercial, or not-for-profit sector.

References

Colicchia

, Campagnolo

, Baldini

, Ulisse

, Valensise

, Moretti

. 2014. Molecular basis of thyrotropin and thyroid hormone action during implantation and early development. Hum Reprod Update, 20:884–904.

Canipari

, Mangialardo

, Di Paolo

, Alfei

, Ucci

, Russi

, Santaguida

, Virili

, Segni

, Misiti

, Centanni

, Verga Falzacappa

. 2019. Thyroid hormones act as mitogenic and pro survival factors in rat ovarian follicles. J Endocrinol Invest, 42:271–282.

Maruo

, Hayashi

, Matsuo

, Yamamoto

, Okada

, Mochizuki

. 1987. The role of thyroid hormone as a biological amplifier of the actions of follicle-stimulating hormone in the functional differentiation of cultured porcine granulosa cells. Endocrinology, 121:1233–1241.

Aghajanova

, Lindeberg

, Carlsson

, Stavreus-Evers

, Zhang

, Scott

, Hovatta

, Skjoldebrand-Sparre

. 2009. Receptors for thyroid-stimulating hormone and thyroid hormones in human ovarian tissue. Reprod Biomed Online, 18:337–347.

Wakim

, Polizotto

, Buffo

, Marrero

, Burholt

. 1993. Thyroid hormones in human follicular fluid and thyroid hormone receptors in human granulosa cells. Fertil Steril, 59:1187–1190.

Monteleone

, Parrini

, Faviana

, Carletti

, Casarosa

, Uccelli

, Cela

, Genazzani

, Artini

. 2011. Female infertility related to thyroid autoimmunity: the ovarian follicle hypothesis. Am J Reprod Immunol, 66:108–114.

Medenica

, Garalejic

, Arsic

, Medjo

, Bojovic Jovic

, Abazovic

, Vukovic

, Zarkovic

. 2018. Follicular fluid thyroid autoantibodies, thyrotropin, free thyroxine levels and assisted reproductive technology outcome. PLoS One, 13:e0206652.

Monteleone

, Faviana

, Artini

. 2017. Thyroid peroxidase identified in human granulosa cells: another piece to the thyroid-ovary puzzle?. Gynecol Endocrinol, 33:574–576.

Kelkar

, Meherji

, Kadam

, Gupta

, Nandedkar

. 2005. Circulating auto-antibodies against the zona pellucida and thyroid microsomal antigen in women with premature ovarian failure. J Reprod Immunol, 66:53–67.

10.

Lee

, Ng

, Lau

, Liu WM

O W

S, Yeung

, Kung

. 2009. Increased fetal abortion rate in autoimmune thyroid disease is related to circulating TPO autoantibodies in an autoimmune thyroiditis animal model. Fertil Steril, 91:2104–2109.

11.

Cramer

, Sluss

, Powers

, McShane

, Ginsburgs

, Hornstein

, Vitonis

, Barbieri

. 2003. Serum prolactin and TSH in an in vitro fertilization population: is there a link between fertilization and thyroid function?. J Assist Reprod Genet, 20:210–215.

12.

Weghofer

, Himaya

, Kushnir

, Barad

, Gleicher

. 2015. The impact of thyroid function and thyroid autoimmunity on embryo quality in women with low functional ovarian reserve: a case-control study. Reprod Biol Endocrinol, 13:43.

13.

Weghofer

, Barad

, Darmon

, Kushnir

, Gleicher

. 2016. What affects functional ovarian reserve, thyroid function or thyroid autoimmunity?. Reprod Biol Endocrinol, 14:26.

14.

Chen

, Huang

, Tzeng

, Huang

, Chen

. 2017. Idiopathic low ovarian reserve is associated with more frequent positive thyroid peroxidase antibodies. Thyroid, 27:1194–1200.

15.

Andrisani

, Sabbadin

, Marin

, Ragazzi

, Dessole

, Armanini

, Dona

, Bordin

, Ambrosini

. 2018. The influence of thyroid autoimmunity on embryo quality in women undergoing assisted reproductive technology. Gynecol Endocrinol, 34:752–755.

16.

Korevaar

TIM

, Minguez-Alarcon

, Messerlian

, de Poortere

, Williams

, Broeren

, Hauser

, Souter

. 2018. Association of thyroid function and autoimmunity with ovarian reserve in women seeking infertility care. Thyroid, 28:1349–1358.

17.

Fumarola

, Grani

, Romanzi

, Del Sordo

, Bianchini

, Aragona

, Tranquilli

, Aragona

. 2013. Thyroid function in infertile patients undergoing assisted reproduction. Am J Reprod Immunol, 70:336–341.

18.

Polyzos

, Sakkas

, Vaiarelli

, Poppe

, Camus

, Tournaye

. 2015. Thyroid autoimmunity, hypothyroidism and ovarian reserve: a cross-sectional study of 5000 women based on age-specific AMH values. Hum Reprod, 30:1690–1696.

19.

Osuka

, Iwase

, Goto

, Takikawa

, Nakamura

, Murase

, Kato

, Bayasula, Kotani

, Kikkawa

. 2018. Thyroid autoantibodies do not impair the ovarian reserve in euthyroid infertile women: a cross-sectional study. Horm Metab Res, 50:537–542.

20.

Embryology

ESIGo

, Alpha Scientists in Reproductive

Medicine

. Electronic address cbgi 2017. The Vienna consensus: report of an expert meeting on the development of ART laboratory performance indicators. Reprod Biomed Online, 35:494–510.

21.

Busnelli

, Paffoni

, Fedele

, Somigliana

. 2016. The impact of thyroid autoimmunity on IVF/ICSI outcome: a systematic review and meta-analysis. Hum Reprod Update, 22:793–794.

22.

Poppe

, Velkeniers

, Glinoer

. 2008. The role of thyroid autoimmunity in fertility and pregnancy. Nat Clin Pract Endocrinol Metab, 4:394–405.

23.

Kilic

, Tasdemir

, Yilmaz

, Yuksel

, Gul

, Batioglu

. 2008. The effect of anti-thyroid antibodies on endometrial volume, embryo grade and IVF outcome. Gynecol Endocrinol, 24:649–655.

24.

Karacan

, Alwaeely

, Cebi

, Berberoglugil

, Batukan

, Ulug

, Arvas

, Camlibel

. 2013. Effect of antithyroid antibodies on ICSI outcome in antiphospholipid antibody-negative euthyroid women. Reprod Biomed Online, 27:376–380.

25.

Lukaszuk

, Kunicki

, Kulwikowska

, Liss

, Pastuszek

, Jaszczolt

, Meczekalski

, Skowronski

. 2015. The impact of the presence of antithyroid antibodies on pregnancy outcome following intracytoplasmatic sperm injection-ICSI and embryo transfer in women with normal thyreotropine levels. J Endocrinol Invest, 38:1335–1343.

26.

Sakar

, Unal

, Atay

, Zebitay

, Verit

, Demir

, Turfan

, Omer

. 2016. Is there an effect of thyroid autoimmunity on the outcomes of assisted reproduction?. J Obstet Gynaecol, 36:213–217.

27.

Moon

, Kim

, Lee

, Jee

, Suh

, Kim

, Lee

, Lim

, Kim

. 2016. Nomogram to predict the number of oocytes retrieved in controlled ovarian stimulation. Clin Exp Reprod Med, 43:112–118.

28.

Krassas

, Poppe

, Glinoer

. 2010. Thyroid function and human reproductive health. Endocr Rev, 31:702–755.

29.

Veltri

, Decaillet

, Kleynen

, Grabczan

, Belhomme

, Rozenberg

, Pepersack

, Poppe

. 2016. Prevalence of thyroid autoimmunity and dysfunction in women with iron deficiency during early pregnancy: is it altered?. Eur J Endocrinol, 175:191–199.

30.

Cecconi

, Rucci

, Scaldaferri

, Masciulli

, Rossi

, Moretti

, D'Armiento

, Ulisse

. 1999. Thyroid hormone effects on mouse oocyte maturation and granulosa cell aromatase activity. Endocrinology, 140:1783–1788.

31.

Revelli

, Delle Piane

, Casano

, Molinari

, Massobrio

, Rinaudo

. 2009. Follicular fluid content and oocyte quality: from single biochemical markers to metabolomics. Reprod Biol Endocrinol, 7:40.

32.

Uyar

, Torrealday

, Seli

. 2013. Cumulus and granulosa cell markers of oocyte and embryo quality. Fertil Steril, 99:979–997.

33.

Kim

, Chae

, Kang

, Chang

. 1998. Influence of antithyroid antibodies in euthyroid women on in vitro fertilization-embryo transfer outcome. Am J Reprod Immunol, 40:2–8.

34.

Poppe

, Autin

, Veltri

, Kleynen

, Grabczan

, Rozenberg

, Ameye

. 2018. Thyroid autoimmunity and intracytoplasmic sperm injection outcome: a systematic review and meta-analysis. J Clin Endocrinol Metab [Epub ahead of print]; DOI: 10.1210/jc.2017-02633.

35.

Zhong

, Ying

, Wu

, Zhou

, Xu

, Wang

, Li

, Shen

, Li

. 2012. Relationship between antithyroid antibody and pregnancy outcome following in vitro fertilization and embryo transfer. Int J Med Sci, 9:121–125.

36.

Michalakis

, Mesen

, Brayboy

, Yu

, Richter

, Levy

, Widra

, Segars

. 2011. Subclinical elevations of thyroid-stimulating hormone and assisted reproductive technology outcomes. Fertil Steril, 95:2634–2637.

37.

Scoccia

, Demir

, Kang

, Fierro

, Winston

. 2012. In vitro fertilization pregnancy rates in levothyroxine-treated women with hypothyroidism compared to women without thyroid dysfunction disorders. Thyroid, 22:631–636.

38.

Velkeniers

, Van Meerhaeghe

, Poppe

, Unuane

, Tournaye

, Haentjens

. 2013. Levothyroxine treatment and pregnancy outcome in women with subclinical hypothyroidism undergoing assisted reproduction technologies: systematic review and meta-analysis of RCTs. Hum Reprod Update, 19:251–258.

39.

Lee

, Lee

, Park

, Yang

, Lim

. 2017. Comparison of clinical outcomes between in vitro fertilization (IVF) and intracytoplasmic sperm injection (ICSI) in IVF-ICSI split insemination cycles. Clin Exp Reprod Med, 44:96–104.

40.

Van Steirteghem

, Nagy

, Joris

, Liu

, Staessen

, Smitz

, Wisanto

, Devroey

. 1993. High fertilization and implantation rates after intracytoplasmic sperm injection. Hum Reprod, 8:1061–1066.

41.

Practice Committee of the American Society for Reproductive

. 2015. Subclinical hypothyroidism in the infertile female population: a guideline. Fertil Steril, 104:545–553.

42.

Alexander

, Pearce

, Brent

, Brown

, Chen

, Dosiou

, Grobman

, Laurberg

, Lazarus

, Mandel

, Peeters

, Sullivan

. 2017. 2017 Guidelines of the American Thyroid Association for the diagnosis and management of thyroid disease during pregnancy and the postpartum. Thyroid, 27:315–389.

43.

Poppe

, Glinoer

, Tournaye

, Schiettecatte

, Devroey

, van Steirteghem

, Haentjens

, Velkeniers

. 2004. Impact of ovarian hyperstimulation on thyroid function in women with and without thyroid autoimmunity. J Clin Endocrinol Metab, 89:3808–3812.

44.

Busnelli

, Somigliana

, Ferrari

, Filippi

, Vannucchi

, Fugazzola

, Fedele

. 2016. The long-term impact of controlled ovarian hyperstimulation on thyroid function. Endocr Pract, 22:389–395.

45.

Veltri

, Belhomme

, Kleynen

, Grabczan

, Rozenberg

, Pepersack

, Poppe

. 2017. Maternal thyroid parameters in pregnant women with different ethnic backgrounds: do ethnicity-specific reference ranges improve the diagnosis of subclinical hypothyroidism?. Clin Endocrinol (Oxf), 86:830–836.

46.

Peters

, Brooke

, Heales

, Ifederu

, Langham

, Hindmarsh

, Cole

. 2016. Defining the newborn blood spot screening reference interval for TSH: impact of ethnicity. J Clin Endocrinol Metab, 101:3445–3449.

47.

Unuane

, Velkeniers

, Anckaert

, Schiettecatte

, Tournaye

, Haentjens

, Poppe

. 2013. Thyroglobulin autoantibodies: is there any added value in the detection of thyroid autoimmunity in women consulting for fertility treatment?. Thyroid, 23:1022–1028.