Increased prevalence of antithyroid antibodies identified in women with recurrent pregnancy loss but not in women undergoing assisted reproduction

FERTILITY AND STERILITY VOL. 71, NO. 5, MAY 1999 Copyright 1999 American Society for Reproductive Medicine Published by Elsevier Science Inc. Printed on acid-free paper in U.S.A. Increased prevalence of antithyroid antibodies identified in women with recurrent pregnancy loss but not in women undergoing assisted reproduction William H. Kutteh, M.D., Ph.D., Deborah L. Yetman, B.S., Alexander C. Carr, B.S., Leslie A. Beck, B.S., and Richard T. Scott, Jr., M.D. Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, University of Tennessee, Memphis, and Saint Barnabus Medical Center, Livingston, New Jersey Objective: To assess the prevalence of antibodies to thyroglobulin and thyroid peroxidase (or microsomal) in women with recurrent pregnancy loss and women undergoing assisted reproductive techniques (ART) compared with healthy controls. Design: Retrospective, two-centered study. Setting: University-affiliated private patient centers. Patient(s): Included were 700 women with a history of two or more consecutive pregnancy losses, 688 women with a history of infertility who were undergoing ART, and 200 healthy, reproductive-aged female controls. Intervention(s): Blood was collected before ART cycles, frozen, and assayed. Main Outcome Measure(s): Standardized ELISAs were used to measure antithyroid antibodies and TSH levels. Statistical analysis was performed with use of the two-tailed Fisher s exact test. Result(s): Antithyroid antibodies were identified in 29 of 200 (14.5%) of controls and 158 of 700 (22.5%) of women with recurrent pregnancy loss and 132 of 688 (19.2%) of women undergoing ART. Less than 20% of the women with antithyroid antibodies were clinically hypothyroid. Conclusion(s): Antithyroid antibodies are identified more frequently in women with recurrent pregnancy loss than in controls but not in women undergoing ART. These autoantibodies may be markers of autoimmune activation and have been associated with an increased risk of pregnancy loss and postpartum thyroid disease. (Fertil Steril 1999;71:843 8. 1999 by American Society for Reproductive Medicine.) Key Words: Antithyroid antibodies, recurrent pregnancy loss, in vitro fertilization Received October 21, 1998; revised and accepted December 8, 1998. Presented at the 45th Annual Meeting of the Society for Gynecologic Investigation, Atlanta, Georgia, March 11 14, 1998. Reprint requests: William H. Kutteh, M.D., Ph.D., Department of Obstetrics and Gynecology, University of Tennessee, Memphis, 956 Court Avenue, Room D324, Memphis, Tennessee 38163-2116 (FAX: 901-448-8782). 0015-0282/99/$20.00 PII S0015-0282(99)00091-6 Autoimmune thyroid disorders are characterized by the presence of antithyroid antibodies, specifically antithyroglobulin and antithyroid peroxidase. Thyroglobulin antibodies are present in most patients with autoimmune thyroid disease. Thyroglobulin is a molecule produced by the thyroid cells and stored in thyroid colloid. The primary function of thyroglobulin is the storage and synthesis of thyroid hormones. Thyroid hormones are synthesized on thyroglobulin, which subsequently serves in the synthesis of T3 and T4 (1). Elevations in the serum concentration of this autoantibody may be identified in any condition that results in an abnormal or damaged thyroid follicular structure (2, 3). Antibodies to thyroid peroxidase, previously referred to as microsomal antibody, are less prevalent in some studies. Thyroid peroxidase is an enzyme responsible for iodination of tyrosine residues along with coupling of iodinated residues to form thyroid hormones. The role of TSH, the primary trophic hormone of the thyroid, is thyroid growth and development; therefore, thyroid cell activation stimulates thyroid hormone synthesis (1). Antithyroid antibodies have been reported in apparently healthy populations and are observed more frequently in women during their reproductive years (4). Immunological factors may play an important role in the reproductive processes of fertilization, implantation, and placental development (5). Women have a high 843

degree of immunological responsiveness that is reflected by their increased susceptibility to nonorgan specific and organspecific autoantibodies (6). Such increased susceptibility is supported by the fact that thyroid autoantibodies have been associated with an increased risk for pregnancy loss (7). It also has been reported that 5% 10% of postpartum women demonstrate evidence of thyroid dysfunction (8 10). Some studies have suggested an association between autoimmune factors and reproductive wastage (11). All patients with autoimmune thyroid disease have T cells in their blood and within the thyroid gland that recognize the specific thyroid molecules thyroglobulin, thyroid peroxidase, and the TSH receptor. Some of these T cells are able to kill self thyroid cells and activate B cells to secrete antibodies, which bind to these same thyroid molecules (1). Detection of thyroid reactive T cells, however, is difficult and currently unrealistic as a clinical tool. Alternatively, thyroid autoantibody measurement can be used as a marker of autoimmune thyroid disease related to activated B cells. Such measurements have become increasingly more advanced and have proven to be useful clinical tools. Antithyroid antibodies may serve as peripheral markers for abnormal T-cell function, which may be responsible for pregnancy loss (12). This was further supported by Pratt et al. (13) who demonstrated that detection of thyroid antibodies before conception carried an increased risk of pregnancy loss. Both thyroglobulin and thyroid peroxidase are important factors in hormone synthesis and are major autoantigens in thyroid autoimmune disease. The purpose of our study was to determine the prevalence of thyroglobulin and thyroid peroxidase antibodies in women undergoing assisted reproductive technologies (ART) and women with recurrent pregnancy loss compared with a healthy female population. Moreover, we sought to determine if these women with positive antithyroid antibodies manifested clinical thyroid disease. MATERIALS AND METHODS Patient Population For this study, we included 688 women undergoing ART. All women had an evaluation including history and physical examination, hysterosalpingogram or hysteroscopy, and midluteal progesterone or late luteal phase endometrial biopsy. These women had a history of infertility, which included tubal factors (19%), ovarian dysfunction (26%), endometriosis (17%), and those who suffered from other or unidentified forms of infertility (38%). Women with more than one pregnancy loss or known thyroid disease were excluded. We also evaluated 700 different women with a history of at least two consecutive pregnancy losses fathered by the same individual. The total number ( SD) of prior losses per patient was 3.7 1.6. Most prior losses occurred before 20 weeks (89.3%) at an average ( SD) gestational age of 9.6 5.4 weeks. All women with recurrent pregnancy loss were encouraged to have a complete evaluation that included a history and physical examination, karyotype on both partners, hysterosalpingogram or hysteroscopy, midluteal progesterone or late luteal endometrial biopsy, prolactin, TSH, lupus anticoagulant (partial thromboplastin time, dilute Russell viper venom time, and platelet neutralization), and cervical cultures for Mycoplasma species, Ureaplasma species, and Chlamydia species. Women with known thyroid disease were excluded. The control population consisted of 200 healthy, nonpregnant, reproductive-aged, parous women without a history of reproductive problems. Controls were recruited from a group of healthy blood donors. The use of frozen blood samples from patients and controls conformed to human use guidelines established by the institutional review board. Blood was taken from nonpregnant patients and controls at least 3 months after any pregnancy or miscarriage. Enzyme-Linked Immunosorbent Assays All serum samples were evaluated for the presence of thyroglobulin and thyroid peroxidase antibodies with use of a standard, highly sensitive ELISA. Samples were assayed in duplicate, and the mean values were reported. These commercial ELISA test kits (ImmunoWELL; GenBio, San Diego, CA) consisted of thyroglobulin and thyroid peroxidase antibody-coated plates to which both a set of known negative and positive sera and the patient sample sera were applied and incubated at room temperature. Sample plates were washed several times with wash buffer (0.01 M phosphate-buffered saline and 0.05% Tween- 20) before the addition of a conjugate containing peroxidaseconjugated goat antihuman immunoglobulin G in phosphatebuffered saline and carrier protein. The sample plates then were washed again before the addition of a color-developing substrate followed by addition of the stop solution (0.25 M oxalic acid). The optical density of the wells was determined on a spectrophotometer at a wavelength of 405 nm. Test results from 100 control women were correlated to a second commercially available ELISA kit (Kronus, San Clemente, CA). Correlation coefficients were excellent between both manufacturers ELISA kits (R 0.95 for thyroglobulin and R 0.96 for thyroid peroxidase). To standardize the assay, this optical density value was used to determine the normalized absorbance of the specimen (A n ), calculated through the following equation: A n A s /A c EV. (A s absorbance of specimen; A c mean absorbance of the positive control; EV expected value for the positive control). A n values then were used to determine the respective values in IU/mL. Based on normalized data from the manufacturer, results were reported as follows: 844 Kutteh et al. Antithyroid antibodies and reproduction Vol. 71, No. 5, May 1999

TABLE 1 Prevalence of antibodies to thyroglobulin and thyroid peroxidase in controls, in women with recurrent pregnancy loss, and in women undergoing assisted reproductive techniques. No. (%) of women with antibodies to TG only, TPO only, and TG and TPO Positive finding Controls (n 200) Recurrent pregnancy loss (n 700) P value* Assisted reproductive techniques (n 688) P value TG only 8 (4.0) 54 (7.7).08 43 (6.2).30 TPO only 7 (3.5) 46 (6.6).12 37 (5.4).36 TG and TPO 14 (7.0) 58 (8.3).76 52 (7.6).88 Total 29 (14.5) 158 (22.5).01 132 (19.2).14 Note: TG thyroglobulin; TPO thyroid peroxidase. * Recurrent pregnancy loss versus controls. Assisted reproductive techniques versus controls. negative, 0.17 A n ( 67 IU/mL) for thyroglobulin and 0.26 A n ( 40 IU/mL) for thyroid peroxidase, borderline, 0.18 0.23 A n (68 119 IU/mL) for thyroglobulin and 0.27 0.37 A n (41 64 IU/mL) for thyroid peroxidase; and positive 0.24 A n ( 120 IU/mL) for thyroglobulin and 0.38 A n ( 65 IU/mL) for thyroid peroxidase. For data analysis, borderline values were considered to be negative. Interassay and intraassay variation for antibodies to thyroglobulin and thyroid peroxidase were 5.8% and 6.2%, respectively. Levels of TSH also were determined with use of a commercial ELISA test kit (Genzyme, San Carlos, CA). Sample sera and control standards were added to microwells coated with murine monoclonal anti-tsh. Goat anti-tsh conjugated to horseradish peroxidase then was added to the samples before a 3-hour incubation. The sample plates were rinsed with tap water, and tetramethylbenzidine substrate reagent then was added followed by a stop solution (2 M hydrochloric acid). The absorbance was read at 450 nm, and results then were converted to IU/mL. The normal range for TSH was 0.45 4.5 IU/mL. Interassay and intraassay variations were 8.0%. Statistical Analysis Statistical analysis between groups was performed with the two-tailed Fisher s exact test. This study had a power of 80% to detect a difference of 6% between controls and patients at a significance level of 0.05. RESULTS Antithyroid Antibodies in Control Women Blood samples for 200 controls (healthy, reproductiveaged women) were assayed with use of commercial ELISA kits for the presence of thyroglobulin and thyroid peroxidase antibodies (Table 1). Overall, autoantibodies were detected in 29 of 200 controls (14.5%). Of these 29, 8 (4.0%) had antibody to thyroglobulin alone; 7 (3.5%) had antibody to thyroid peroxidase alone, and 14 (7.0%) control females revealed the presence of both thyroglobulin and thyroid peroxidase antibodies. Antithyroid Antibodies in Women With ART Of the 688 ART patients, 132 (19.2%) tested positive for the presence of thyroid autoantibodies (Table 1). There was no statistically significant difference between the prevalence of antithyroid antibodies in women undergoing ART compared with the control population (P.14). Of the total group, 43 (6.2%) of the women had thyroglobulin antibodies alone, whereas 37 (5.4%) tested positive for thyroid peroxidase antibody alone. Fifty-two (7.6%) of these women had both thyroglobulin and thyroid peroxidase antibodies. None of these three groups, thyroglobulin alone, thyroid peroxidase alone, or thyroglobulin and thyroid peroxidase together, were statistically different compared with the control population. Antithyroid Antibodies in Women With Recurrent Pregnancy Loss Of the 700 patients with recurrent pregnancy loss, antithyroid antibodies were present in 158 (22.5%) (Table 1). This was a significantly higher autoantibody prevalence among women with recurrent pregnancy loss than in the control population (P.01). Equal proportions of these patients tested positive for antibodies to thyroglobulin alone (54 of 700 [7.7%]), for thyroid peroxidase antibodies alone (46 of 700 [6.6%]), and for both thyroglobulin and thyroid peroxidase autoantibodies (58 of 700 [8.3%]). Thyroid-Stimulating Hormone Levels The TSH levels also were assessed in the ART, recurrent pregnancy loss, and control populations (Table 2). From the group of 200 control patients, 11 (5.5%) had abnormally high TSH levels. Of these, 7 (3.5%) were abnormal coupled with positive antibodies to thyroglobulin and/or thyroid peroxidase values. Only 4 (2.0%) from the total control population yielded abnormal TSH values while testing negative for both thyroglobulin and thyroid peroxidase antibodies. Of FERTILITY & STERILITY 845

TABLE 2 Elevated TSH levels in control women, in women with recurrent pregnancy loss, and in women undergoing assisted reproductive techniques. No. (%) of women with elevated TSH levels Positive finding Controls (n 200) Recurrent pregnancy loss (n 700) P value* Assisted reproductive techniques (n 688) P value TSH only 4 (2.0) 30 (4.3).20 34 (4.9).07 TSH and TG or TPO 5 (2.5) 12 (1.7).55 14 (2.0).78 TSH, TG and TPO 2 (1) 11 (1.6).74 7 (1.0) 1.00 Total 11 (5.5) 53 (7.6).35 55 (8.0).28 Note: TG thyroglobulin; TPO thyroid peroxidase. * Recurrent pregnancy loss versus controls. Assisted reproductive techniques versus controls. the 688 women undergoing ART, 55 (8.0%) had abnormally high TSH levels, which was not statistically significantly different from the controls (P.28). Most of these patients (34 [4.9%]) had abnormal TSH levels but tested negative for both thyroglobulin and thyroid peroxidase antibodies, whereas 21 (3.0%) of these patients had abnormal TSH levels along with positive antibodies to thyroglobulin and/or thyroid peroxidase. Neither of these levels reached a statistically significant difference from the control population. Of the 700 test patients with recurrent pregnancy loss, 53 (7.6%) had abnormal TSH levels, which did not differ significantly from the controls (P.35) nor from the ART patients (P.68). Of these patients with abnormal TSH levels, 30 (4.3%) consisted of women with elevated TSH alone, whereas 23 (3.2%) of the abnormal TSH patients simultaneously tested positive for antibodies to thyroglobulin and/or thyroid peroxidase antibodies. Neither of these two groups yielded a statistically significant difference from the controls nor from the patients who underwent ART. Age of Study Subjects Both women with recurrent pregnancy loss and women treated with ART differed significantly in age from the control population (P.01). The mean ( SD) age for the control population was 30.8 6.2 years, for recurrent pregnancy loss patients the mean ( SD) age was 33.3 5.0, and for the ART patients the mean ( SD) age was 34.9 4.0 years. To ascertain that the differences in autoantibody frequency among the three populations was not due to age differences alone, the prevalence of each antibody among various age group ranges also was assessed. Among the control patients, thyroglobulin antibody was most frequent (4.0%) among the 36 40 age group range (Fig. 1). Antibodies to thyroglobulin and thyroid peroxidase were distributed normally with the peak frequency of 7% in the patients with recurrent pregnancy loss and of 5.6% in the women undergoing ART occurring in the 31 35 year age group. Thyroid peroxidase antibody prevalence in the control population peaked at 3.5% in both the 31 35 and 36 40 age groups (Fig. 2). Thyroid peroxidase antibodies in the patients with recurrent pregnancy loss and in women undergoing ART reached their highest prevalence among the 31 35 year age group as well, yielding 5.9% and 6.1%, respectively. Age alone could not account for the differences observed in the prevalence of thyroid antibodies, and women aged 31 35 years with recurrent pregnancy loss or undergoing ART had a higher frequency of both thyroid peroxidase and thyroglobulin autoantibodies compared with control women of the same age. No statistically significant differences or significant trends in the three groups were observed from this analysis. FIGURE 1 Prevalence of antibodies to thyroglobulin in control women (control), in women with recurrent pregnancy loss (RPL), and in women undergoing assisted reproductive techniques (ART). Percentages depict the overall percent of women who had antibodies to thyroglobulin in each indicated age group. 846 Kutteh et al. Antithyroid antibodies and reproduction Vol. 71, No. 5, May 1999

FIGURE 2 Prevalence of antibodies to thyroid peroxidase in control women (control), in women with recurrent pregnancy loss (RPL), and in women undergoing assisted reproductive techniques (ART). Percentages depict the overall percent of women who had antibodies to thyroid peroxidase in each indicated age group. DISCUSSION These data suggest that elevated antithyroid autoantibody titers occur more frequently among women with recurrent pregnancy loss but not in women undergoing ART. A multicentered study that evaluated 544 pregnant women in the first trimester directed attention to a correlation between spontaneous abortion and elevated titers of autoantibodies to thyroglobulin and thyroid peroxidase (7). This increased risk of miscarriage could not be explained by TSH levels, anticardiolipin antibodies, maternal age, or relevant history. Unfortunately, the investigators failed to evaluate these women for other known causes of miscarriages. Similarly, Lejeune et al. (14). reported on 730 consecutive pregnant women attending an antenatal clinic. Elevated antibodies to thyroid peroxidase and thyroglobulin were associated with an increased miscarriage rate. In a subsequent study of 487 ART pregnancies, antithyroid antibodies were found in 106 patients (22%) (15). There was a clinical miscarriage rate of 32% in the antibody-positive group compared with a 16% miscarriage rate in the antibody-negative group; however, no statistically significant difference was noted in the incidence of biochemical pregnancies. Our larger study failed to detect an increased prevalence of antibodies to thyroglobulin or thyroid peroxidase in nonpregnant women undergoing ART compared with nonpregnant control women. Antithyroid testing has been advocated by some as an early marker for implantation failure in the prediction of pregnancy after ART. Some even have suggested that those patients be treated with intravenous gammaglobulin (16). Although antithyroid antibodies do appear to be associated with an increased risk of subsequent miscarriage in pregnancies conceived naturally or through ART (7, 15), current data do not support an association of antithyroid antibodies and implantation failure. Antithyroid antibodies are known to occur in normal, healthy populations, and these autoantibodies are five times more common in women than in men (6). Because of the prevalence of antithyroid antibodies in normal women, interpreting the significance of these antibodies in women with reproductive problems remains difficult. Therefore, the need for aggressive and expensive treatment of women with antithyroid antibodies who are undergoing ART may be unnecessary. Conversely, in women with a history of recurrent pregnancy loss, evaluation of TSH alone may not be enough to determine relevant thyroid abnormalities. In our study of 700 women with recurrent pregnancy loss, 158 women had antithyroid antibodies, but only 23 of those same women were clinically hypothyroid on the basis of elevated TSH levels. These hypothyroid women should have thyroid hormone replaced and thyroid status normalized before attempting a subsequent pregnancy. Thus, if only TSH was assayed, only 14.5% of the total patients with recurrent pregnancy loss with abnormal thyroglobulin or thyroid peroxidase antibodies would have been identified. In a much smaller study, 11 of 28 women with recurrent pregnancy loss (39%) but only 2 of 28 controls (7%) had positive titers of thyroglobulin, thyroid peroxidase, or both antibodies (P.01) (17). Pratt et al. (13) investigated the subsequent pregnancy outcome in 42 euthyroid women with at least recurrent pregnancy loss. Overall, 13 of 42 (31%) women had antithyroid antibodies, and 30 of 42 (71%) women had a successful subsequent pregnancy. It is significant that 8 of 13 (62%) with positive antibodies had a subsequent miscarriage, whereas only 4 of 29 (14%) women without antibodies to thyroid peroxidase or thyroglobulin had a miscarriage. These small studies support the association of antithyroid antibodies and subsequent loss in women with a history of recurrent pregnancy loss. In a recent report, 22 of 75 women with recurrent pregnancy loss (29%) compared with 28 of 74 control women (37%) were found to have antithyroid antibodies (18). Although the authors reported no differences in the populations, their conclusions have been questioned based on the unusually high prevalence of thyroid antibodies in this control group. Moreover, their smaller study only had a power of 80% to detect a 20% difference in positive results between groups. The differences in age between the control and patient test groups did not account for the different prevalences of antithyroid antibodies observed in our study. Women undergoing ART and women with recurrent pregnancy loss were FERTILITY & STERILITY 847

older than the controls and might have been expected to have an elevated prevalence of autoantibodies; however, this was not observed in our study. Women aged 31 35 years with recurrent pregnancy loss or undergoing ART had a higher frequency of positive antithyroid antibodies compared with control women of the same age. Some, but not all studies have revealed an increase in the prevalence of autoantibodies with age. Kontiainen et al. (19), found an increase in the amount of thyroid peroxidase antibodies with age; however, this correlation was not statistically significant. In our study, more of the patients with recurrent pregnancy loss demonstrated an elevated autoantibody titer as age increased, up until the age range of 31 35, after which a decreased frequency was observed. Elevated autoantibody titers in correlation with increasing age were also expected of the ART patients, but our results revealed a similar pattern as the patients with recurrent pregnancy loss, in which the frequency of positive autoantibody titers decreased at the 36 40 age group. Antithyroid antibodies may be associated with an increased risk of pregnancy loss. There are two working hypotheses concerning the pathophysiological role of thyroglobulin and thyroid peroxidase antibodies and pregnancy loss. First is the proposal that the biochemical interaction between hormones and elevated thyroid autoantibodies may in some way directly result in pregnancy loss. Because thyroid function is normal in many patients when thyroid antibodies are detected, many investigators have questioned the possible role of thyroid antibodies in women with pregnancy loss. Alternatively, others consider thyroglobulin and thyroid peroxidase autoantibodies as secondary markers of autoimmune disease rather than the actual cause of pregnancy loss (15). These antithyroid antibodies may be reflective of an abnormal immunological response that results in pregnancy loss. Our data indicate an increased prevalence of antithyroid antibodies in women with recurrent pregnancy loss but not in women undergoing ART compared with controls. We are prospectively observing both populations of patients and plan to report our findings on subsequent pregnancy outcome based on the presence or absence of antithyroid antibodies. Acknowledgments: The authors thank the research committees at GenBio, San Diego, California, and Kronus, San Clemente, California for the generous donation of ELISA kits used in this study to detect antibodies to thyroid peroxidase and thyroglobulin. References 1. O Conor G, Davies TF. Human autoimmune thyroid disease: a mechanistic report. Trends Endocrinol Metab 1990;1:266 72. 2. Bayraktar M, Ergin M, Boyacioglu A, Demir S. A preliminary report of thyroglobulin release after fine needle aspiration biopsy of thyroid nodules. J Int Med Res 1990;18:253 5. 3. Parkes A, Black E, Adams H, John R, Richards C, Hall R, et al. Serum thyroglobulin: an early indicator of autoimmune post-partum thyroiditis. Clin Endocrinol (Oxf) 1994;41:9 14. 4. Geva E, Amit A, Lerner-Geva L, Lessing J. Autoimmunity and reproduction. Fertil Steril 1997;67:599 611. 5. Billingham R, Head J. Current trends in reproductive immunology: an overview. J Reprod Immunol 1981;3:253 65. 6. Chiovato L, Lapi P, Fiore E, Tonacchera M, Pinchera A. Thyroid autoimmunity and female gender. J Endocrinol Invest 1993;16:384 9. 7. Stagnaro-Green A, Roman S, Cobin R, El-Harazy E, Alvarez-Marfany M, Davies T. Detection of at-risk pregnancy by means of highly sensitive assays for thyroid autoantibodies. J Am Med Assoc 1990;264: 1422 5. 8. Amino N, Mori H, Iwatani Y, Tanizawa O, Kawashima M, Tsuge I, et al. High prevalence of transient postpartum thyrotoxicosis and hypothyroidism. N Engl J Med 1982;306:849 52. 9. Hayslip C, Fein H, O Donnell V, Freidman D, Klein T, Smallridge R. The value of serum antimicrosomal antibody testing in screening for symptomatic postpartum thyroid dysfunction. Am J Obstet Gynecol 1988;159:203 9. 10. Jansson R, Dahlberg P, Karlsson F. Postpartum thyroiditis. Baillieres Clin Endocrinol Metab 1988;2:619 35. 11. Gleicher N, El-Roeiy A, Confino E, Friberg J. Reproductive failure because of autoantibodies: unexplained infertility and pregnancy wastage. Am J Obstet Gynecol 1989;160:1378 80. 12. Stagnaro-Green A, Roman SH, Cohen RH, el-harazy E, Wallenstein S, Davies TF. A prospective study of lymphocyte-initiated immunosuppression in normal pregnancy: evidence of a T-cell etiology for post partum thyroid dysfunction. J Clin Endocrinol Metab 1992; 74:645 53. 13. Pratt D, Kaberlein G, Dudkiewicz A, Karande V, Gleicher N. The association of antithyroid antibodies in euthyroid nonpregnant women with recurrent first trimester abortions in the next pregnancy. Fertil Steril 1993;60:1001 4. 14. Lejeune B, Grun JP, de Nayer P, Servais G, Glinoer D. Antithyroid antibodies underlying thyroid abnormalities and miscarriage or pregnancy induced hypertension. Br J Obstet Gynaecol 1993;100:669 72. 15. Singh A, Dantas ZN, Stone SC, Asch RH. Presence of thyroid antibodies in early reproductive failure: biochemical versus clinical pregnancies. Fertil Steril 1995;63:277 81. 16. Sher G, Maassarani G, Zouves C, Feinman M, Sohn S, Matzner W, et al. The use of combined heparin/aspirin and immunoglobulin G therapy in the treatment of in vitro fertilization patients with antithyroid antibodies. Am J Reprod Immunol 1998;39:223 5. 17. Bussem SS, Steck T. Thyroid antibodies and their relation to antithrombin antibodies, anticardiolipin antibodies and lupus anticoagulant in women with recurrent spontaneous abortions (antithyroid, anticardiolipin and antithrombin autoantibodies and lupus anticoagulant in habitual aborters). Eur J Obstet Gynecol Reprod Biol 1997;74:139 43. 18. Esplin MS, Branch DW, Silver R, Stagnaro-Green A. Thyroid autoantibodies are not associated with recurrent pregnancy loss. Am J Obstet Gynecol 1998;179:1583 6. 19. Kontiainen S, Melamies L, Miettinen A, Weber T. Thyroid autoantibodies in serum samples with abnormal TSH levels. APMIS 1994;102: 716 20. 848 Kutteh et al. Antithyroid antibodies and reproduction Vol. 71, No. 5, May 1999