Effect of Anti-thyroid Peroxidase Antibodies on Assisted Reproductive Technologies Outcomes: A Retrospective Cohort Study

Thyroid autoimmunity (TAI) is the most common autoimmune disorder among women of reproductive age and may be associated with adverse outcomes for women undergoing assisted reproductive technologies (ART).[ 1 2 3 4 5 6 7 ] In this study, ‘women’ refers to individuals assigned female sex at birth. TAI is characterised by the presence of anti-thyroid antibodies (ATAs), specifically anti-thyroid peroxidase antibodies (TPOAbs) or anti-thyroglobulin antibodies.[ 1 8 9 ] TPOAbs are considered the more sensitive marker of TAI.[ 2 ] ATAs are found in 8%–14% of women of reproductive age,[ 1 9 ] and more common among women undergoing fertility treatment.[ 1 3 8 10 ] Previous studies have shown conflicting results regarding the impact of TAI on fertilisation. Some reported lower fertilisation rates in women with TAI,[ 3 4 ] whereas a newer systematic review found no significant differences.[ 2 ] The pathophysiological mechanisms behind a potentially detrimental effect of TAI on reproductive outcomes are still unknown.[ 7 11 ] It has been suggested that the presence of ATAs in follicular fluid may lead to antibody-mediated cytotoxicity in the developing ovarian follicle and damage to the maturing oocyte, which may impair its quality and developmental potential.[ 3 ] However, lacking validated biomarkers limits conclusions about TAI’s impact on oocyte quality.[ 2 ] Given the high prevalence of TAI and the importance of identifying potential causes of unexplained infertility, this topic remains highly relevant in reproductive medicine. The primary objective of this study was to assess whether the presence of TPOAbs influences fertilisation rates and blastocyst development in infertile women undergoing in vitro fertilisation (IVF). Secondary analyses compared outcomes between IVF and Intracytoplasmic sperm injection (ICSI) to explore whether ICSI may mitigate potential antibody-related interference with sperm–oocyte interaction.

There were no significant differences in age, BMI, anti-Müllerian hormone (AMH) levels, or the number of oocytes retrieved between TPOAb-positive and TPOAb-negative women in either the IVF or split group [ Tables 1a and 1b ]. Baseline sperm characteristics are presented in Tables 2a and b . In the IVF group, donor sperm was used by 197 patients in the TPOAb-negative subgroup and 16 patients in the TPOAb-positive subgroup. In the split group, donor sperm was used by 20 patients in the TPOAb-negative subgroup and 3 in the TPOAb-positive subgroup. The distribution of female primary diagnosesis presented in Table 3 . Baseline characteristics - in vitro fertilisation ( n =650) Values are expressed as mean±SD and median (IQR). BMI=Body mass index, AMH=Anti-Müllerian hormone, SD=Standard deviation, IQR=Interquartile range, TPOAb=Anti-thyroid peroxidase antibodies Baseline characteristics - split ( n =129) Values are expressed as mean±SD and median (IQR). BMI=Body mass index, AMH=Anti-Müllerian hormone, SD=Standard deviation, IQR=Interquartile range, TPOAb=Anti-thyroid peroxidase antibodies Baseline sperm characteristics - in vitro fertilisation Values are expressed as median (IQR). Baseline sperm parameters are presented descriptively. 197 patients used donor sperm in TPOAb-negative group, and 16 in TPOAb-positive group. These are not included in sperm characteristics. IQR=Interquartile range, TPOAb=Anti-thyroid peroxidase antibodies Baseline sperm characteristics - split Values are expressed as median (IQR). Baseline sperm parameters are presented descriptively. 20 patients used donor sperm in TPOAb-negative group and 3 in TPOAb-positive group. These are not included in sperm concentration and motility. IQR=Interquartile range, TPOAb=Anti-thyroid peroxidase antibodies Female primary diagnoses - in vitro fertilisation and split PCOS=Polycystic ovary syndrome, IVF= In vitro fertilisation, TPOAb=Anti-thyroid peroxidase antibodies In the IVF group [ Table 4 ], fertilisation rates were similar between TPOAb-positive and TPOAb-negative women. Likewise, no significant difference was found in blastocyst formation rates. Adjusted odds ratios for 2PN and blastocyst development by anti-thyroid peroxidase antibodies status - in vitro fertilisation *Sensitivity analyses, partner sperm only. Adjusted for age, BMI, smoking (cigarettes/day), sperm type (donor or partner) and sperm quality. TPOAb-negative was used as a reference. TPOAb=Anti-thyroid peroxidase antibodies, CI=Confidence interval, OR=Odds ratio, BMI=Body mass index No significant differences were observed in the split group regarding fertilisation (OR: 0.82; 95% confidence interval [CI]: 0.56–1.22; P = 0.33) or blastocyst development following IVF (OR: 0.72; 95% CI: 0.46–1.09; P = 0.13) between TPOAb-positive and TPOAb-negative groups. Similarly, for ICSI, no significant differences were found in fertilisation (OR: 0.95; 95% CI: 0.61–1.50; P = 0.814) or blastocyst development (OR: 0.92; 95% CI: 0.58–1.45; P = 0.730) between TPOAb-positive compared to negative controls. These comparisons were not presented in separate tables. Sensitivity analyses restricted to partner sperm cycles yielded comparable results, with no differences in fertilisation outcomes between TPOAb-positive and TPOAb-negative women following IVF (OR 0.82; 95% CI: 0.54–1.26) or ICSI (OR 0.87; 95% CI: 0.54–1.44). In the split group, fertilisation rates were higher with ICSI compared to IVF in both TPOAb-positive and TPOAb-negative women [ Figure 1 ]. Among TPOAb-positive women, fertilisation occurred in 63.6% of oocytes following ICSI compared to 52.0% with IVF, corresponding to a non-significant increase (adjusted OR 1.66; 95% CI: 0.96–2.90; P = 0.071). A similar non-significant increase was observed for blastocyst development (35.4% vs. 26.0%; adjusted OR 1.65; 95% CI: 0.91–3.02; P = 0.101). Fertilisation rates (%) by insemination method and anti-thyroid peroxidase antibodies status in split cycles. Fertilisation rates are presented as percentages of fertilised oocytes per total inseminated oocytes. TPOAB = Anti-thyroid peroxidase antibodies, IVF = In vitro fertilisation, ICSI = Intracytoplasmic sperm injection Among TPOAb-negative women, ICSI was associated with a significantly higher fertilisation rate compared to IVF (67.1% vs. 57.7%; adjusted OR 1.50; 95% CI: 1.20–1.87; P = 0.0004). Blastocyst development was also numerically higher following ICSI (38.7% vs. 33.8%), although this difference did not reach statistical significance (adjusted OR: 1.24; 95% CI: 1.00–1.55; P = 0.053). In sensitivity analyses restricted to cycles using partner sperm only, the association between ICSI and fertilisation was attenuated in TPOAb-positive women (OR 1.54; 95% CI: 0.84–2.86; P = 0.16), whereas the association remained significant in TPOAb-negative women (OR: 1.50; 95% CI: 1.18–1.90; P = 0.001).

This study found no evidence of an association between TPOAbs and impaired fertilisation or blastocyst development in women undergoing IVF. While ICSI was associated with higher fertilisation rates, this effect was observed irrespective of TPOAb status and does not support a specific benefit in this subgroup. These findings suggest that TPOAb positivity alone should not be used to guide the choice between IVF and ICSI. Given the limitations of the study, results should be regarded as hypothesis-generating. Larger, prospective studies using standardised split-cycle designs and researching later clinical endpoints are warranted. All authors made substantial contributions to the conception and design of the study and to the acquisition, analysis and interpretation of data. All authors were involved in drafting the manuscript or revising it critically for important intellectual content, approved the final version to be published and agree to be accountable for all aspects of the work. There are no conflicts of interest. The data underlying this study are not publicly available due to institutional and patient confidentiality restrictions. The authors used ChatGPT 5.2 to enhance language and readability and take full responsibility for the final content.

Normal thyroid function is essential for maintaining reproductive health, as documented in the literature.[ 7 11 13 14 ] Accordingly, thyroid dysfunction has been implicated in menstrual irregularities, reduced fertility and an elevated risk of miscarriage.[ 11 14 15 ] However, TAI implications in euthyroid women are less well understood.[ 9 16 17 ] TPOAbs have been detected in the ovarian follicular fluid, raising concerns about a possible interference with the fertilisation process.[ 3 ] However, the exact mechanism, such as oocyte damage or alterations of the zona pellucida, remains hypothetical.[ 1 ] The role of ICSI as a potentially protective method against the presumed barrier effect of TAI on zona pellucida penetration has not been extensively investigated.[ 8 18 ] In this study, we aimed to investigate whether TPOAbs affect oocyte fertilisation in women undergoing IVF and whether ICSI offers any advantage over conventional IVF in TPOAb-positive women. We found no significant difference in fertilisation or blastocyst development between TPOAb-positive and negative women undergoing IVF, and similar findings were observed in the split group. These results were consistent after exclusion of donor sperm cycles, indicating that sperm type did not confound the observed associations. Although fertilisation rates were numerically higher with ICSI compared to IVF, these differences did not reach statistical significance in TPOAb-positive women. In TPOAb-negative women, ICSI was associated with significantly increased fertilisation, while a non-significant increase in blastocyst development. Taken together, these findings suggest that any potential advantage of ICSI is not specific to TPOAb-positive women but may reflect a more general effect of the technique. The use of split cycles represents a strength of this study, as IVF and ICSI were performed within the same treatment cycle, thereby reducing the influence of patient-level confounding. In addition, no couples were preselected due to the male factor. This reduces the potential influence of male factor infertility; nevertheless, residual confounding cannot be excluded. A recent randomised controlled trial found a lower fertilisation rate with ICSI compared to IVF in non-male factor infertility cases (53.5% vs. 58.1%).[ 19 ] In this study, TPOAb was not reported, and the conclusion suggested that ICSI should not be the default approach in the absence of male factor. Previous studies investigating the role of ICSI in women with TAI have reported conflicting results.[ 5 8 18 ] While some studies have reported higher fertilisation rates with ICSI, such findings are not consistent across the literature and are primarily derived from retrospective analyses, which may be influenced by selection bias and differences in patient characteristics, including prevalence of male factor infertility.[ 5 8 ] In addition, several studies and meta-analyses have evaluated IVF and ICSI outcomes jointly as part of ART, without distinguishing between fertilisation methods, thereby limiting interpretation of any procedure-specific effects.[ 1 2 9 ] A systematic review by Poppe et al . found no difference in fertilisation between TPOAb-positive and negative undergoing ICSI.[ 18 ] Notably, they did not compare to patients undergoing IVF. Interestingly, Huang et al . reported higher fertilisation rates with ICSI in women with TAI but also observed lower clinical pregnancy and live birth rates in this group.[ 8 ] This discrepancy may suggest that TAI does not primarily interfere with the fertilisation process itself but may instead influence later stages of embryo development or implantation. In this context, our findings, derived from a cohort without routine preselection for ICSI, do not support a TPOAb-specific benefit of this approach. The observed higher fertilisation rates with ICSI in both TPOAb-positive and TPOAb-negative women further suggest that any advantage of ICSI is likely related to the technique itself rather than modulation of TAI. Our study has several limitations that should be acknowledged. First, a key limitation of this study is that analyses were performed at the oocyte level, treating oocytes as independent observations. As multiple oocytes originate from the same patient, this approach does not account for within-patient clustering and may lead to underestimation of standard errors and potential inflation of type I error. The retrospective design may lead to potential biases, including unmeasured confounding and lack of control over protocol adherence. In addition, the study population was heterogeneous, which may introduce residual confounding despite adjustment for key variables. While we adjusted for key variables (age, BMI, smoking status and sperm quality), data on thyroid hormone levels, L-thyroxine treatment and male medical history and medication use were not available. Also, the relatively small sample size, particularly in the TPOAb-positive split group, limits our statistical power. This imbalance reflects the natural prevalence of TPOAb positivity in the clinical population and may increase the risk of type II error. Finally, multiple comparisons were performed without formal correction, and results should therefore be interpreted with caution. Importantly, this study focused on early embryological outcomes. While these endpoints may provide insight into whether TPOAbs interfere at the level of fertilisation or early embryo development, they do not capture later clinical outcomes such as pregnancy or live birth. Future studies should address these clinically relevant endpoints.

This cohort study was approved by the Danish Protection Agency, Capital Region (journal no.: 24070552), permitting the use of patient data without the requirement for informed consent. The study included individuals undergoing ART at Herlev Fertility Clinic between December 16, 2019, and November 30, 2024. The study was conducted in accordance with the principles of the Declaration of Helsinki. Data were extracted from Danish Medical Data Centre journals for 886 patients initiating their first cycle of either IVF or split. Seven hundred and seventy-nine patients were included, of whom 650 underwent IVF, and 129 underwent split. Split fertilisation is a previously used technique in which half of the retrieved oocytes are fertilised using conventional IVF and the other half with ICSI within the same treatment cycle. Until 2023, patients eligible for conventional IVF were routinely offered split insemination in their first treatment cycle if they had 8 or more retrieved oocytes. If fewer than 8 oocytes were retrieved, conventional IVF was performed. Patients were excluded if no oocytes were retrieved, TPOAb status was missing or ICSI was performed exclusively. Collected variables included TPOAb status, stimulation protocol, fertilisation outcomes, body mass index (BMI), smoking status, age and sperm quality. Missing data were manually retrieved from patient files. All women included in the study were euthyroid at the initiation of ART, defined as a TSH level below 4.8 mIU/L in accordance with reference intervals at the capital region hospital laboratories. Women receiving L-thyroxine treatment had TSH levels below 2.5 miU/L, as recommended by clinical guidelines. TPOAbs were assessed as part of routine clinical evaluation before ART. Due to the retrospective nature of the study, antibody measurements were performed at different laboratories. Consequently, complete standardised information on laboratory procedures and assay performance parameters was not available for all participants. However, all TPOAb analyses were performed at large, accredited hospital laboratories operating under standardised quality assurance programmes. For patients whose samples were analysed at Herlev Hospital, TPOAb concentrations were measured using a chemiluminescent immunoassay (Siemens Atellica IM Anti-Thyroid Peroxidase II). The assay is CE-marked, accredited, and traceable to WHO international reference materials. TPOAb positivity was defined as concentrations ≥13.8 kIU/L, in accordance with the manufacturer’s recommended clinical decision limit. The patients underwent controlled ovarian stimulation, oocyte retrieval and IVF, either by conventional IVF or split. Patients were stimulated using either a standard agonist protocol or a standard antagonist protocol. Male partners underwent routine clinical evaluation before assisted reproductive treatment. To minimise selection bias related to male factor infertility, couples undergoing ICSI-only treatment were excluded. At our centre, ICSI was indicated in cases of severe male factor, defined as a total progressive motile sperm count below 2 million. Sperm analysis was performed according to routine clinical practice in accordance with the WHO laboratory manual.[ 12 ] The parameters included in the analyses were those considered clinically relevant for fertilisation outcomes: sperm concentration, progressive motility, ejaculate volume and post-processing progressive motile sperm count. Baseline characteristics were compared across the four groups using means and standard deviations or medians and interquartile range. The primary outcome was fertilisation, defined as the presence of two pronuclei (2PN) in mature oocytes. Blastocyst development, assessed on day five or six, was recorded as a secondary outcome. Each retrieved oocyte was treated as an independent observation. Logistic regression was used to compare binary outcomes. All models were adjusted for potential confounders, including age, BMI, sperm quality, sperm type (partner or donor) and smoking status, where appropriate. Sperm quality was evaluated based on the percentage of progressively motile sperm cells, with normal sperm quality defined as >32% according to WHO guidelines.[ 12 ] It was categorised into four groups: Normal, mildly reduced, moderately reduced, and severely reduced. Donor sperm was classified as normal sperm quality. For each logistic regression analysis, odds ratios (ORs) with corresponding P values were reported. Statistical significance was defined as P < 0.05. Analyses were conducted using RStudio (v2024.12.1) and R (v4.3.2) in Quarto Posit PBC, Boston, Massachusetts, USA.