Autoimmune Thyroid Disease and Female Fertility: Clinical Evidence on Ovarian Reserve, Infertility Treatment Outcomes, and Pathophysiology

This work was supported by a Grant‐in‐Aid for Scientific Research (24 K02583) awarded to Akira Iwase by the Japan Society for the Promotion of Science, Japan, and by a Child and Family Health Research Grant (25DA0701) from the Child and Family Agency, Japan.

We conducted a systematic literature review using PubMed to identify relevant studies published from inception until August 2025. The following search terms were used: “thyroid diseases OR hyperthyroidism OR hypothyroidism,” “thyroid autoimmunity,” “ovarian failure OR ovarian insufficiency” ( Data S1 ). The literature search was primarily performed by one author (Akira Iwase) and subsequently reviewed by the other authors. For clinical studies, we selected those published in 2018 or later to reflect the most recent evidence. We also searched for articles on etiology and pathophysiology using the same keywords; however, the literature was selectively extracted and organized as a narrative review. TSH and gonadotropins share structural similarities, and their cross‐reactivity may be clinically relevant under specific conditions. A well‐known example is thyrotoxicosis in early pregnancy caused by elevated hCG levels acting on the TSH receptor, which is enhanced in the presence of activating mutations of the latter [ 5 ]. In addition, Van Wyk–Grumbach syndrome is a rare condition in which markedly elevated TSH levels due to severe hypothyroidism stimulate the FSH receptor, potentially leading to precocious puberty [ 6 ]. However, these represent highly specific and uncommon conditions, and it is unlikely that elevated TSH levels in typical hypothyroidism directly affect the HPO axis. TSH receptors are expressed in oocytes and granulosa cells at specific developmental stages [ 7 , 8 ]. TSH stimulation increases the intracellular cAMP levels in granulosa cells obtained during oocyte retrieval for in vitro fertilization [ 7 ]. Furthermore, experimental studies using mouse granulosa cells have demonstrated that TSH stimulation enhances aromatase activity and estradiol production [ 9 ]. Therefore, TSH may indirectly influence the HPO axis by affecting ovarian sex steroid production. In recent years, increasing attention has been directed toward kisspeptin neurons, which regulate the HPO axis. Kisspeptin neurons receive signals related to body weight and energy metabolism, suggesting that they influence the HPO axis [ 10 , 11 ]. TSH also influences the production of leptin, a chemokine derived from fat, suggesting that it may affect the HPO axis via leptin‐mediated pathways [ 12 , 13 ]. Although these effects may lead to temporary ovulation disorders, it remains uncertain whether they have a decisive impact on ovarian reserve or assisted reproductive technology (ART) outcomes. Thyroid hormones target a wide range of organs, and thyroid hormone receptors (THRs) are present in many cells and tissues. THRs are expressed in human oocytes, cumulus cells, and granulosa cells, with thyroid hormones detectable in follicular fluid [ 14 ]. Thyroid hormones stimulate granulosa cells [ 15 ], promote preantral follicle growth, and reduce apoptosis [ 16 ]. Moreover, the effects of T3 may be mediated through the PI3K/Akt pathway, a signaling pathway significantly involved in the activation of primordial follicles [ 15 , 17 , 18 ]. With respect to follicular development, maintaining a balance between inhibitory and promotive control is essential for supporting ovulation and sustaining a continuous ovulatory cycle [ 19 , 20 ]. Therefore, disruption of follicular regulatory mechanisms by thyroid hormones may contribute to changes in TAI‐associated ovarian reserve, although further detailed studies are required. Autoimmune diseases result from abnormal immune responses directed against tissues mediated by autoantibodies. Hashimoto's disease appears to have a greater impact on fertility and ovarian function than hypothyroidism without autoimmune involvement, suggesting that autoimmune mechanisms contribute to reproductive phenotypes. Consistent with this, approximately 40% of women with premature ovarian failure have been reported to be positive for at least one autoantibody, of which anti‐thyroid antibodies are the most prevalent [ 21 ]. The presence of TgAbs and TPOAbs in the follicular fluid of women with TAI has been confirmed, with the antibody levels positively correlated with serum autoantibodies [ 22 ]. However, the specific ovarian antigens that may be recognized by thyroid autoantibodies remain unclear. Although thyroid peroxidase is expressed in granulosa cells, its direct recognition by thyroid autoantibodies has not yet been established [ 23 ]. On the other hand, autoantibodies against the zona pellucida in the serum of women with premature ovarian insufficiency (POI) were reported to cross‐react with thyroid tissue [ 24 ]. If autoantibodies other than thyroid autoantibodies are directed against follicular cellular components, follicular development may be impaired. Anti‐ovarian antibodies have been proposed as a potential mechanism underlying POI. Candidate target antigens for these autoantibodies include not only the zona pellucida, but also gonadotropin receptors and steroidogenic enzymes [ 25 ]. Addison's disease is an autoimmune disorder affecting the adrenal glands, and autoantibodies against steroidogenesis‐related enzymes can react with ovarian tissues [ 26 , 27 , 28 ]. Autoimmune polyglandular syndrome (APS) is classified into three types: APS1 is involved in adrenal autoimmunity, whereas APS2 is involved in adrenal and thyroid autoimmunity, both of which are associated with a relatively high prevalence of POI [ 29 , 30 ]. In contrast, POI also occurs in APS3, which does not involve adrenal autoimmunity, suggesting the presence of autoimmune mechanisms that affect the ovaries independent of adrenal autoimmunity [ 31 ]. Adamyan et al. reported that antibodies against CYP11A1, CYP19A1, and CYP21A2 have high diagnostic value for POI in the absence of Addison's disease [ 32 ]. Kasahara et al. identified autoantibodies directed against ovarian tissue in the blood of women with TAI and POI and identified prostate, ovary, testis‐expressed protein (POTE) as the antigen recognized by these antibodies. The POTE is involved in the regulation of follicular development and may represent one of the mechanisms underlying POI associated with TAI [ 33 ]. The presence of ovarian autoantibodies may lead to follicular damage and impairment of ovarian reserve because POI can occur simultaneously with other autoimmune conditions. However, no anti‐ovarian autoantibodies associated with TAI have been established for antigen identification and follicular damage. Cytokines play an important role in the immunopathology of TAI [ 34 ]. TPOAb titers have been shown to be associated with Th1 cytokine production in patients with TAI. Interferon‐gamma and tumor necrosis factor alpha are Th1‐derived cytokines present in TAI [ 35 ]. Both these cytokines may contribute to apoptotic changes in ovarian cells, leading to follicular atresia [ 35 , 36 , 37 , 38 ]. In addition, pathological processes related to TAI may result in excessive production of reactive oxygen species (ROS) [ 39 , 40 ]. Increased expression of superoxide dismutase 1 in the ovaries of hypothyroid model mice may reflect upregulation of ROS, which can adversely affect follicular development [ 41 , 42 ]. Cytokines and ROS have also been reported to be associated with age‐related decline in ovarian function and endometriosis, and may represent one of the mechanisms underlying follicular dysfunction in TAI. Nonetheless, detailed studies evaluating cytokine and ROS levels in the ovaries of patients with TAI compared with controls have not yet been conducted. Therefore, further investigations are required to clarify their contributions to the development of POI. The effects of thyroid dysfunction and autoimmunity on ovarian function are summarized in Figure  1 . Potential mechanisms underlying ovarian dysfunction associated with thyroid autoimmunity and hypothyroidism. POI results from follicular depletion and has a significant effect on female fertility, with premature estrogen deficiency contributing to increased morbidity and mortality. The term POI, which has replaced the term premature ovarian failure, emphasizes that the pathological condition originates in the ovaries and the associated functional disorder is progressive [ 43 ]. POI typically occurs in women younger than 40 years and is diagnosed based on amenorrhea lasting longer than 4–6 months, in combination with elevated FSH levels and decreased estradiol concentrations [ 44 , 45 ]. Recently, the European Society of Human Reproduction and Embryology published guidelines defining POI as a condition characterized by loss of ovarian activity before the age of 40 years, with amenorrhea or irregular menstrual cycles for at least 4 months, elevated gonadotropins (FSH > 25 IU/L), and low estradiol levels [ 46 ]. TAI is one of the most common autoimmune diseases in women of reproductive age [ 47 ] and is sometimes diagnosed in association with POI [ 48 ]. However, for both conditions, there is often a time lag between disease onset and the appearance of symptoms leading to a diagnosis. Moreover, as mentioned above, the pathophysiological mechanisms by which TAI affects ovarian function remain unclear; therefore, no definitive conclusions can be drawn regarding a causal relationship. Table  1 summarizes observational studies that have examined the association between POI and TAI. A large‐scale cohort study using health insurance databases reported that compared to women without Hashimoto's disease, those with the disease had a 1.8‐fold increased risk of amenorrhea and a 2.4‐fold increased risk of infertility [ 50 ]. This study demonstrated a significant increase in risk during a follow‐up period of less than 5 years. In contrast, observational studies that considered POI as a risk factor for autoimmune diseases have reported increased rates of Hashimoto's disease and other autoimmune diseases among women with POI [ 51 , 52 , 53 , 54 , 55 ]. The population‐based study by Savukoski did not evaluate TAI alone, but instead examined autoimmune diseases present prior to the diagnosis of POI as an outcome, which may suggest a causal relationship [ 52 ]. Furthermore, logistic regression analysis reported increased positivity rates for TPOAbs and TgAbs in women with POI [ 56 ]. Thyroid autoimmunity and ovarian failure. Abbreviations: AITD, autoimmune thyroid disease; HR, hazard ratio; NS, not specified; OR, odds ratio; POI, premature ovarian insufficiency; RR, risk ratio; TAI, thyroid autoimmunity. Recently, Mendelian randomization analysis, which enables randomization in observational studies, has been developed. Its use in examining celiac disease and POI identified hypothyroidism (13.46%) and systemic lupus erythematosus (26.36%) as significant mediating factors [ 57 ]. In addition, bidirectional Mendelian randomization using a two‐sample approach allows for the assessment of the directionality of causal relationships. Using this method, Luo et al. demonstrated a significant causal relationship between POI and the risk of TAI, rather than TAI and the risk of POI [ 58 ]. In this study, autoimmune diseases identified as significant risk factors for POI included Addison's disease and systemic lupus erythematosus. Although several studies have applied this method, definitive conclusions regarding the causal relationships between TAI and POI remain difficult to establish. With the clinical application of serum anti‐Müllerian hormone (AMH) levels, the concept of ovarian reserve has become widely accepted. Ovarian reserve is a quantitative indicator of the number of eggs remaining in the ovaries at a given time point, and AMH is currently regarded as the most reliable indicator. POI can be considered an extreme condition in which the ovarian reserve is significantly reduced; however, the use of AMH as an indicator allows for a more nuanced evaluation of the relationship between declining ovarian reserve and TAI. Table  2 summarizes the results of recent publications evaluating TAI and AMH values. The definition of TAI varied slightly across studies. A downward trend in AMH levels has been observed in adult women with TAI [ 59 , 60 , 61 , 62 , 63 ], whereas several reports have found no significant differences in AMH levels [ 49 , 64 , 65 , 66 , 67 , 68 ]. When differences in TSH levels are present between TAI and non‐TAI groups, there appears to be a tendency toward corresponding differences in AMH levels. Osuka et al. demonstrated a negative correlation between TSH and AMH (r = −0.398, p  = 0.0399), despite observing no significant difference in AMH levels between those with and without thyroid autoantibodies [ 49 ]. AMH in TAI and non‐TAI. Note: Values are mean ± SD or median (IQR). Abbreviations: NS, not specified; TAI, thyroid autoimmunity. Regarding the correlation between antibody titer and AMH, Notaro et al. demonstrated that women in the low ovarian reserve subgroup had a higher prevalence of positive TgAbs than those in the normal ovarian reserve subgroup (60.7% vs. 39.3%, p  = 0.038), despite no significant difference in AMH levels between TAI and non‐TAI groups [ 65 ]. Similarly, TgAbs and TPOAbs have been reported to be negatively associated with AMH levels (r = −0.114, p  = 0.013; r = −0.294, p  = 0.047, respectively) [ 60 , 61 ]. It remains unclear whether autoantibodies exert a direct effect on ovarian reserve or whether their presence is associated with lower AMH levels as an indicator of TAI severity. In addition, these associations may be age‐dependent. In a comparison between TPOAb‐positive and TPOAb‐negative women, a significant tendency toward decreased AMH in the former was not observed in a subgroup analysis of women younger than 35 years [ 63 ]. Through a meta‐analysis, Hasegawa et al. demonstrated an increase in AMH levels in adolescents with TAI and a decrease in AMH levels in individuals who developed TAI after 20 years of age [ 69 ]. It has been hypothesized that autoantibodies and other factors interfere with the regulation of follicular development, resulting in abnormal follicular development in young people, in turn leading to an increase in AMH levels. This phase is followed by follicular depletion and a subsequent decrease in AMH levels. Long‐term longitudinal studies beginning early in life are required to test this hypothesis. Recently, numerous studies have evaluated the efficacy of levothyroxine supplementation in the treatment of subclinical hypothyroidism (SCH). The decision to initiate levothyroxine therapy to prevent miscarriage and preterm birth remains complex and continues to be debated, particularly in relation to the TSH cutoff values used to define SCH [ 70 , 71 ]. In contrast, relatively few studies have assessed the efficacy of levothyroxine supplementation in TAI from the perspective of ovarian reserve. Popa et al. reported no significant difference in AMH levels between patients with Hashimoto's disease who received levothyroxine and those who did not. Notably, there was no significant difference in TSH levels between the two groups [ 67 ]. Kuroda et al. reported that levothyroxine supplementation improved AMH levels in patients with Hashimoto's disease; however, their cohort included individuals without thyroid autoantibodies, and reduced TSH levels were observed [ 72 ]. TAI has been discussed as a factor affecting fertility and for its potential relationship with infertility treatment outcomes. In recent years, the use of levothyroxine to prevent miscarriages and premature births in women with SCH has been widely debated. Levothyroxine supplementation is recommended based on the idea that latent hypothyroidism before pregnancy should be corrected in response to an increase in thyroid hormone requirements in early pregnancy [ 73 ]. Conversely, concerns have been expressed regarding the appropriateness of using a TSH cutoff value of 2.5 IU/mL as an indication for intervention [ 70 , 71 ]. The ongoing controversy surrounding levothyroxine supplementation for SCH is related to the unclear effects of TAI on fertility. To date, no firm conclusions have been reached regarding the impact of TAI on ART outcomes. Table  3 summarizes recent reports that have examined ART treatment outcomes of TAI. Two studies reported significant differences in clinical pregnancy and/or live birth/ongoing pregnancy rates [ 74 , 75 ]. One study focused on euthyroid women after levothyroxine supplementation, including those with overt hypothyroidism and SCH before treatment [ 74 ]. Another focused on euthyroid women; however, TSH levels differed significantly between the non‐TAI and TAI groups (1.81 ± 1.2 vs. 2.35 ± 1.70 IU/mL, p  < 0.05) [ 75 ]. ART outcomes in TAI and non‐TAI. Note: Values are mean ± SD or median (IQR). Abbreviations: CPR, clinical pregnancy rate; LBR, live birth rate; NS, not specified. TAI, thyroid autoimmunity. The other studies (Table  3 ) did not demonstrate a significant difference in ART outcomes between women with and without TAI [ 59 , 66 , 67 , 76 , 77 , 78 , 79 , 80 ]. Moon et al. demonstrated that patients with SCH exhibited significantly lower clinical pregnancy rates than euthyroid women, regardless of thyroid autoimmunity status [ 79 ]. Additionally, eight studies (Table  3 ) evaluated the number of retrieved oocytes, two of which reported a statistically significant reduction in oocyte yield in women with TAI [ 75 , 77 ]. Although definitive conclusions are difficult to draw, euthyroid TAI in the absence of elevated TSH levels did not appear to have a significant impact on ART outcomes.

There is a long history of research on POI and Hashimoto's disease. However, neither a clear causal relationship nor definitive pathophysiological mechanisms have been established. Although the hypothesis that follicular depletion is caused by autoimmune inflammation remains supported, by the time follicular depletion has occurred and POI is diagnosed, the inflammatory activity has often subsided. This temporal limitation makes it difficult to assess the natural disease course prior to this point and evaluate local ovarian inflammation at different stages. Further investigations are required to verify this hypothesis. In contrast, when AMH is used as an indicator, multiple studies have reported correlations between thyroid autoimmunity, Hashimoto's disease, and reduced AMH levels in adults. However, the degree of AMH reduction is not significant enough to indicate a state leading to POI. Thus, these alterations may be more appropriately interpreted as reflecting endocrine changes affecting ovarian reserve, rather than follicular loss secondary to ovarian inflammation. Further studies are required to elucidate the complex effects of endocrine abnormalities and autoimmune disorders on follicular development. Furthermore, conflicting results have been reported regarding the impact of infertility on treatment outcomes. It remains unclear whether these issues stem from quantitative or qualitative abnormalities in oocytes and/or additional effects on implantation and embryo development. In conditions classified as autoimmune thyroid disease, the effects of thyroid autoantibodies and function are interrelated; however, independent mechanisms are conceivable. Therefore, it is challenging to clearly distinguish their respective effects on ovarian reserve and ART outcomes. Studies with more narrowly defined populations, as well as longitudinal analyses evaluating changes in thyroid autoantibody titers and AMH levels before and after levothyroxine supplementation for thyroid dysfunction, may provide important insights into this issue.

Infertility affects approximately 8%–12% of reproductive‐aged couples and is recognized as a major global health issue [ 1 ]. Fertility declines with advancing age, and the increasing trend toward delayed childbearing may contribute to the rising infertility rates. However, the pathological and clinical aspects of declining fertility beyond aging warrant further investigation. Gonadotropins and female hormones involved in human reproduction are essential for gamete production and pregnancy maintenance; therefore, endocrine disorders can lead to reproductive dysfunction and reduced fertility [ 2 ]. Given its high prevalence among women of reproductive age, thyroid autoimmunity (TAI) is one of the most significant autoimmune disorders in this population [ 3 ]. Although thyroid dysfunction has long been associated with menstrual disorders, its effects on ovarian function and infertility have been increasingly investigated in recent years. Thyroid‐stimulating hormone (TSH) is a heterodimeric protein similar to the pituitary gonadotropins follicle‐stimulating hormone (FSH), luteinizing hormone, and human chorionic gonadotropin (hCG), all having a common α chain. It has been speculated that fluctuations in TSH may directly or indirectly affect the hypothalamic–pituitary–ovarian (HPO) axis. In addition, reproductive hormones are affected by energy metabolism, and because thyroid hormones regulate metabolic processes, the reproductive system may undergo various modifications following thyroid dysfunction. In addition, the relationship between diagnostic autoantibodies and disease presence and severity in autoimmune thyroiditis (Hashimoto's disease) has been well studied. TAI is characterized by the presence of thyroid peroxidase antibodies (TPOAbs) and/or thyroglobulin antibodies (TgAbs) that play a role in hypothyroidism. Although thyroid autoantibodies are not directed against ovarian components, autoimmune mechanisms similar to those observed in Addison's disease are thought to affect ovarian function [ 4 ]. This review provides an update on the pathophysiology and clinical findings of TAI and fertility, primarily from the perspective of ovarian reserve, and discusses future research directions and clinical challenges.

The authors declare no conflicts of interest. Akira Iwase is the Editor‐in‐Chief of Reproductive Medicine and Biology and is a co‐author of this article. Seung Chik Jwa and Satoko Osuka are Editors of Reproductive Medicine and Biology and are coauthors of this article. To minimize bias, they were excluded from all editorial decision‐making related to the acceptance of this article for publication.

Data S1: rmb270052‐sup‐0001‐DataS1.xlsx.