- Email: [email protected]
The association of antithyroid antibodies in euthyroid nonpregnant women with recurrent first trimester abortions in the next pregnancy
FERTILITY AND STERILITY Copyright
©
Vol. 60, No.6, December 1993
1993 The American Fertility Society
Printed on acid-free paper in U S. A.
The association of antithyroid antibodies in euthyroid nonpregnant women with recurrent first trimester abortions in the next pregnancy
Donna E. Pratt, M,D,* George Kaberlein, B.S, Alan Dudkiewicz, Ph.D,
Vishvanath Karande, M.D, Norbert Gleicher, M.D.
The Center for Human Reproduction and The Foundation for Reproductive Medicine Inc., Chicago, Illinois
Objective: To evaluate the prognostic value of antithyroid antibodies in euthyroid women with a history of recurrent first trimester abortions on future pregnancy loss. Design: The sera of 42 euthyroid women with a history of three or more consecutive first trimester abortions were evaluated for the presence of antibodies to thyroglobulin and thyroid peroxidase before pregnancy and again as soon as the diagnosis of pregnancy was made. Setting: Medical school-affiliated private infertility center. Patients: Forty-two women with a history of three or more consecutive first trimester abortions who were planning to conceive again. Main Outcome Measure: The presence of antithyroid antibodies in the nonpregnant state and their association with pregnancy loss in the next gestation. Results: Thirteen of 42 women (31 %) were positive for the presence of antithyroid antibodies at the initial screening before pregnancy. All 13 maintained positivity by the time their next pregnancy was diagnosed. Only 12 of those 42 women (29%) experienced a first trimester abortion. Eight of these 12 women (67%) were positive for one or more antithyroid antibody. In contrast, among 30 nonaborting women, only 5 of 30 (17%) exhibited thyroid antibody positivity. The detection of thyroid antibodies before conception carried an increased risk of pregnancy loss in the next pregnancy (8 of 13,62% versus 4 of 29,14%). Conclusion: The presence of antithyroid antibodies in nonpregnant women with a history of recurrent abortion identifies a subgroup of women at significantly increased risk for yet another pregnancy loss in their next gestation. Because organ-specific autoantibodies thus demonstrate similar prognostic significance to nonorgan-specific autoantibodies, it is tempting to conclude that peripheral autoantibody abnormalities seen in habitual aborters only reflect an underlying T -lymphocyte defect, which may be the actual cause of pregnancy loss. Fertil SterilI993;60:1001-5 Key Words: Antithyroid antibodies, recurrent abortion
Pregnancy loss seems to occur with increased frequency in the presence of abnormal autoimmune function. This association was first made in reference to lupus anticoagulant and other phospholipid antibodies (1, 2). Stagnaro-Green et al. (3) and
Received June 8, 1993; revised and accepted August 3, 1993.
* Reprint requests: Donna Pratt, M.D., The Center for Human Reproduction, 750 North Orleans Street, Chicago, Illinois 60610. Vol. 60, No.6, December 1993
Glinoer et al. (4) recently reported a strong correlation between the presence of thyroid antibodies and first trimester pregnancy loss. Those thyroid antibody studies were performed during the first trimester of pregnancy and it is therefore still unknown whether the presence of thyroid antibodies before pregnancy can serve as an independent marker of risk for losing an upcoming gestation. Although the total incidence of recognized and unrecognized pregnancy loss per pregnancy has Pratt et al.
Thyroid antibodies and recurrent abortion
1001
5i
been reported to be as high as 31 % (5), the actual risk in a young population with no prior history of pregnancy loss is probably <15% (6). This is probably best reflected by investigating pregnancy loss rates after infertility treatment where early pregnancy losses are more likely recognized than in other populations. A higher miscarriage rate in gonadotropin-induced cycles (17% to 31%) than in natural cycles (10% to 15%) still reflects the above noted general pregnancy loss rate (7). This rate also was confirmed recently by Mills et al. (8). In a welldesigned multicenter study, they reported spontaneous abortion rates of 11.8% to 20.8% (8). To define a higher risk group of patients for pregnancy loss, a group of so called habitual (or repeated) aborters was chosen for this study. Such patients by definition have had three or more consecutive pregnancy losses and are statistically at greatly increased risk for yet another loss in their next pregnancy (9, 10). They therefore represent a favorable population pool for attempts to determine the potential value of thyroid antibodies as predictors of pregnancy loss in upcoming pregnancies. MATERIALS AND METHODS
The sera of 42 nonpregnant women with a history of three or more consecutive pregnancy losses were investigated for the presence of thyroid antibodies ~ 6 months before pregnancy and again as soon as a pregnancy was diagnosed by rising hCG titers at 4 to 6 weeks from the last menstrual period. Once pregnancy occurred, the patients were followed weekly with ultrasound and hormonal evaluations until 12 weeks gestational age, when the patients were transferred out into their local communities for obstetrical care. Because we were unable to obtain reliable follow up on all patients in the second and third trimester, this study reports only the incidence of first trimester pregnancy loss. A smaller number of additional second trimester losses can be expected and did in fact occur. Patients were selected from a larger pool of repeated aborters after women with obvious other causes for pregnancy loss had been excluded. Phospholipid antibodies to cardiolipin, phosphatidyl serine, phosphatidic acid, phosphatidyl ethanolamine, phosphatidyl inositol, and phosphatidyl glycerol for immunoglobulin (Ig)G, IgM and IgA isotopes were evaluated. Patients with significant uterine anomalies, uterine myomas, parental chromosomal abnormalities detected from peripheral 1002
Pratt et al.
Thyroid antibodies and recurrent abortion
....
Table 1 Demographic Data of Women who Aborted Versus Women who did not Abort'
Age
No. of prior abortions
y
Study group (n = 42) Positive antibodies (n = 13) Negative antibodies (n = 39) Aborted (n = 12) Did not abort (n = 40)
34 33 34 33 34
± 3.1
± 2.9 ± 3.4 ± 2.3 ± 3.7
3.6 4.2 3.4 3.8 3.6
± ± ± ± ±
1.1 1.5 .91 1.0 1.2
'Values are means ± SD.
blood, and autoimmune abnormalities among phospholipid antibodies were excluded. All patients underwent at least a hysterosalpingography, and often a hysteroscopy and laparoscopy as well, as part of their workup. Not excluded were patients with possible luteal phase defect because patients were substituted in the luteal phase with P in oil. The majority of pregnancy losses in the final study group had not undergone prior chromosomal evaluation of products of conception. Chromosomal abnormalities in such earlier losses could therefore not be ruled out. Table 1 summarizes the demographic data of the study population, which ranged in age from 27 to 41 years (mean ± SD, 34 ± 3.1) and reported between three and eight consecutive pregnancy losses (mean ± SD, 3.6 ± 1.1). Only 3 of 42 women reported a live birth before their consecutive losses (secondary aborters) whereas the remainder had no living children (primary aborters). The ages ofthe secondary aborters were 32, 35, and 36 years. All had one live birth that, in all three cases, was the first pregnancy. The number of spontaneous abortions after the live births were 4, 3, and 3, respectively. Only one of the secondary aborters was positive for antithyroid antibodies. This subject was the only secondary aborter to experience another spontaneous abortion. All study patients were determined to be euthyroid before pregnancy by demonstrating normal levels of thyroid-stimulating hormone (TSH) and T 4 • Laboratory Evaluation
Thyroid-stimulating hormone and free T 4 were determined by using the Stratus, automated fluorometric enzyme immunoassay (Baxter Diagnostics Inc., Deerfield, IL). For TSH, plastic carriers containing a solution of mouse monoclonal antihuman TSH IgG complexed onto the surface of Fertility and Sterility
-
glass fiber paper by the addition of goat antisera to mouse Igs, are placed into an analyzer loading station. The carriers are transferred to a circular holding track where they pass through a series of "work stations." At the initial station 90 lIL of the patients' sera is delivered to the carrier. After a 60-second incubation period, labeled antibody conjugate (a solution containing calf intestinal alkaline phosphatase covalently linked to a mixture of two mouse monoclonal anti-human TSH) and substrate wash (a solution containing 1 mmol/L 4methylumbelliferyl phosphate in a diethanolamine buffer) are added at subsequent work stations. At the final station, after a second 60-second incubation period, the quantity of bound enzyme-labeled anti-human TSH is interpreted by fluorometry. The same automated system is used for the determination of free T 4. The carrier that contains rabbit anti -T 4 serum is complexed onto the surface of glass filter paper by the addition of goat antisera to rabbit Igs. The conjugate is calf intestinal alkaline phosphate covalently linked to T 4. The normal range for TSH is defined as 0.35 to 7.0 lIIU/mL and for free T4 as 0.9 to 2.1 ng/dL. The mean sensitivity of the human TSH assay is 0.05 lIIU /mL and 0.2 ng/dL for free thyroxine. The mean interassay coefficient of variation and the intra-assay coefficients of variation in our assay were 3.7 and 3.2, respectively. Thyroid peroxidase antibodies and thyroid globulin antibodies were assayed using Kalibre RIA kits (Kronus, Dana Point, CA). These are both highly sensitive RIAs with a range of detection between 0.1 and 30 U/mL. Assays were standardized against Medical Research Council thyroglobulin and microsomal antigen autoantibody first international reference preparations (coded 66/387 and 65/93, respectively) where 1 U is equivalent to approximately 10 IU (personal communication, Kronus Technical Services). Briefly, the thyroid peroxidase assay is a single antibody RIA in which 50 lIL of prediluted calibrators and 50 lIL of patient sera are placed in appropriately labeled tubes. Fifty microliters of p25 thyroid peroxidase tracer is added to each tube and the tubes are gently mixed and incubated for 60 minutes at room temperature (RT). A protein A reagent (50 lIL) is then added to all tubes, mixed, and incubated for 60 minutes at RT. One milliliter of diluent, containing 0.15 M NaCl 0.02 M Tris-HCl, ph 7.5; bovine serum albumin; and 0.002 M sodium azide is added and centrifuged at 1500 X g for 15 minutes. The supernatants are decanted and the radioactivity in the tubes is counted for 1 Vol. 60, No.6, December 1993
minute in a gamma scintillation counter. The same procedure is used for the thyroid globulin antibodies substituting p25 thyroid globulin tracer (11). Comparisons of this assay to the tanned red cell test indicated an antithyroid peroxidase antibody and antithyroid globulin antibody titer of 1:100 and 1:10, respectively, was equivalent to 10 U/mL by the direct RIA method. Enzyme-linked immunosorbent assay comparisons showed a highly significant correlation (r = 0.99). The lower limits of detection of the ELISAs were equivalent to 5 to 10 U/mL by direct RIA. Assays based on I 125 -labeled protein A showed similar lower limits of detection to the ELISA assay (5 to 10 U/mL). Correlations were also highly significant (r = 0.91) (11). The mean interassay coefficient of variation and the intra-assay coefficient of variation in our assays were 2.6 and 4.2, respectively. Statistical analyses were performed using the Fisher's Exact Test and the unpaired Student's t-test, using GraphPad Instat, GraphPad Software version 1.15 (GraphPad Software Inc., San Diego, CA).
RESULTS Thirteen of 42 (31%) women were positive for either one or both antithyroid antibodies during their initial screening in the non-pregnant state. All 13 were still positive at the time of pregnancy diagnosis. No previously negative woman became positive early in pregnancy. Among 13 thyroid antibody-positive women, 9 were positive for thyroid peroxidase antibody, 11 for thyroid globulin antibody, and 7 for both antibodies. The levels of thyroid peroxidase antibody ranged from 0.30 to 44.67 U/mL (mean ± SD, 15.93 ± 15.387 U/mL), while the levels for thyroid globulin antibody ranged from 0.32 to 40.44 U/mL (10.95 ± 15.38 U/mL) (Table 2). There was no significant difference in mean levels between thyroid peroxidase antibody and thyroid globulin antibody nor was there a significant difference in the antithyroid antibody levels before and during early pregnancy (data not shown). Twelve of 42 women aborted during the first trimester (29%). Eight of the 12 (67%) were positive for one or both antithyroid antibodies. In contrast, only 5 of the 30 women (17%) who did not abort demonstrated positive antithyroid antibodies (P = 0.003). The presence of thyroid antibodies before conception carried a significantly increased risk of pregnancy loss in that 8 of 13 antibody-positive and Pratt et al.
Thyroid antibodies and recurrent abortion
1003
BE
.£1
Table 2
Subject
a
•
Antithyroid Antibody Levels
Age
No. of spontaneous abortions
31 37 32 29 34 35 32 31 39 37 36 34 34
Thyroid globulin
Thyroid peroxidase U/mL
y
1* 2* 3 4* 5 6* 7* 8* 9 10 l1*t 12 13*
s
3 4 4 4 3 5 3 3 8 4 3 6 4
44.96 14.50 Not detected 0.30 9.54 Not detected 4.08 10.11 Not detected 1.50 32.96 Not detected 25.50
4.50 2.54 1.15 3.0 1.50 36.20 40.44 Not detected 1.33 3.24 Not detected 26.20 0.32
* Patients that aborted. t Secondary aborter.
only 4 of 29 antibody-negative women miscarried (62% versus 14%, P = 0.003). Analysis of sensitivity and specificity were 66.6% and 83.3%, respectively. Testing for thyroid antibodies in a population of recurrent aborters gave a positive predictive value of 61.5% and had a negative predictive value of 86.2%. DISCUSSION
This study confirmed the increased risk status of our study population for pregnancy loss. The loss of 30% of pregnancies in the first trimester indicates a rate of pregnancy loss that is almost twice as high as expected in a normal population (5-10). These data, however, also confirm that even women with a large number of first trimester consecutive pregnancy losses still have an excellent chance of carrying a next pregnancy past the first trimester. Among women at increased risk to miscarry, the presence of thyroid antibodies in the nonpregnant state represents a marker ofrisk. Women who demonstrate thyroid antibodies before conception miscarried in 8 of 13 cases (62%) whereas among antibody negative patients only 4 of 29 miscarried (14 %). This represents a statistically significant difference (P = 0.003). The importance of thyroid antibodies is also reflected in the fact that miscarrying patients demonstrated a significantly higher incidence of antibody positivity (67% versus 17%, P = 0.003). Antithyroid antibodies are known to occur in apparently healthy populations and are more fre1004
Pratt et al.
Thyroid antibodies and recurrent abortion
quently observed in women during their reproductive years (12-14). We previously reported a higher incidence of antithyroid antibodies in women with a history of three or more consecutive pregnancy losses as compared with a control population of apparently healthy blood donors (31% versus 19%) (15). Thyroid disease has historically not been associated with pregnancy loss (6, 10, 16). Recently, Stagnaro-Green et al. (3) and shortly thereafter Glinoer et al. (4) reported, however, for the first time a specific association between antithyroid antibodies and pregnancy loss. Nonorgan-specific antibody abnormalities, and especially antiphospholipid antibody abnormalities, have been associated for a much longer time with pregnancy loss (2, 17, 18). To find organ-specific thyroid antibodies in parallel to nonorgan-specific autoantibody abnormalities is, however, not uncommon. For example, thyroid antibodies also have been reported in up to 45% of patients with systemic lupus erythematosus (19); and at least one-third of women with autoimmune-induced premature ovarian failure demonstrate nonorgan-specific autoantibodies as well as antithyroid antibodies (20). Thyroid antibodies, not dissimilar to nonorganspecific autoantibodies, may thus simply serve as peripheral markers of abnormal T-cell function. It is then tempting to speculate that abnormal T -lymphocyte function, and not abnormal autoantibodies per se, is responsible for pregnancy loss. Support for such a concept comes from an increasing recognition that the basic immunologic defect in autoimmune diseases, inclusive of thyroid diseases, probably lies at the T-lymphocyte level (21). Under such a concept, organ-specific autoantibody abnormalities, such as thyroid antibodies and nonorgan-specific autoantibodies, can be expected to have similar prognostic significance for impending pregnancy loss. This is in fact confirmed by this study because patients with classical nonorganspecific autoantibody abnormalities were excluded. One could further conclude that successful treatment of autoantibody-associated pregnancy loss should concentrate on attempts to modulate T-cell function rather than the customary treatment options of antibody suppression and/or modulation of antibody effects (1, 22).
REFERENCES 1. Lubbe WF. Liggins GC. Role of the lupus anticoagulant and autoimmunity in recurrent pregnancy loss. Semin Reprod Endocrinol 1988;6:181-90.
Fertility and Sterility
2. Cowchock S, Smith RJ, Gocial B. Antibodies to phospholipids and nuclear antigens in patients with repeated abortion. Am J Obstet GynecoI1986;155:1002-10. 3. Stagnaro-Green A, Roman SH, Cobin RH, El-Harazy Essam, Alvary-Marfany M, Davies TF. Detection of at-risk pregnancy by means of a highly sensitive assays for thyroid autoantibodies. JAMA 1990;264:1422-5. 4. Glinoer D, Soto MF, Bourdoux P, Lejeune B, Delange F, Lemone M, et al. Pregnancy in patients with mild thyroid abnormalities: maternal and neonatal repercussions. J Clin Endocrinol Metab 1991;73:421-7. 5. Wilcox AJ, Weinberg CL, O'Connor JF, Baird DD, Schlaterer JP, Canfield RE. Incidence of early loss of pregnancy. N Engl J Med 1988;319:189-94. 6. Regan L, Braude PR, Trembath PH. Influence of past reproductive performance on risk of spontaneous abortion. Br Med J 1989;299:541-6. 7. Ranson MX, Bohrer M, Blotner MB, Kemmann E. The difference in miscarriage rates between menotropin-induced and natural cycle pregnancies is not surveillance related. Fertil Steril 1993;59:567-70. 8. Mills JL, Holmes LB, Aarons JH, Simpson JL, Brown ZA, Jovanovic-Peterson LG, et al. Moderate caffeine use and the risk of spontaneous abortion and intrauterine growth retardation. JAMA 1993;269:593-7. 9. Alberman E. The epidemiology of repeated abortion. In: Sharp F, Beard RW, editors. Early pregnancy loss mechanisms and treatment. London: Royal College of Obstetrics and Gynecologists, 1988:9-17. 10. Harper JH, Archer DF, Marchese SG, Muraca-Clemens M, Garner TH. Etiology of recurrent pregnancy losses and outcome of subsequent pregnancies. Obstet Gynecol 1983;62: 574-81. 11. Beever K, Bradbury J, Phillips D, McLachdan SM, Pegg C, Goral A, et al. Highly sensitive assays of autoantibodies to thyroglobulin and to thyroid peroxidase. Clin Chern 1989;35:1949-54. 12. Vargas MT, Brines-Urbina R, Gladman D, Papsin FR,
Vol. 60, No.6, December 1993
13.
14.
15.
16.
17. 18.
19.
20.
21.
22.
Walfish PG. Antithyroid microsomal autoantibodies and HLA -D R5 are associated with post partum thyroid dysfunction: evidence supporting an autoimmune pathogenesis. J Clin Endocrinol Metab 1988;67:327-33. Goodwin MT, Montaro M, Westman J. Transient hyperthyroidism and hyperemesis gravidarum: clinical aspects. Am J Obstet Gynecol 1992;167:648-52. Jansson R. Bernander S, Karlson A, Levin K, Nelsson G. Autoimmune thyroid dysfunction in the post partum period. J Clin Endocrinol Metab 1984;58:681-7. Pratt D, Novotny M, Kaberlein G, Dudkiewicz A, Gleicher N. Antithyroid antibodies and the association with non-organ-specific antibodies in recurrent pregnancy loss. Am J Obstet Gynecol 1993;168:837-41. Montora M, Callea JV, Fraizier SD, Westman JH. Successful outcome of pregnancy in women with hypothyroidism. Ann Intern Med 1981;94:31-4. Maier DB, Parke A. Subclinical autoimmunity and recurrent aborters. Fertil Steril 1989;51:280-5. 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: 1376-80. Magero M, Zoli A, Altomonte L, Mirone L, Barini LSA, Scuder F. The association of silent thyroiditis and active systemic lupus erythematosus. Clin Exp Rheumatol 1992;10:67-70. La Barera A, Miller MM, Ober C, Rebor RS. Autoimmune etiology in premature ovarian failure. Am J Reprod Immunol 1988;16:114-22. Stagnaro-Green A, Roman SH, Cohen RH, El-Harazy E, Wallerstein S, Davies TF. A prospective study of lymphocyte-initiated immuno-suppression in normal pregnancy: evidence of a T-cell etiology for post partum thyroid dysfunction. J Clin Endocrinol Metab 1992;74:645-53. Lubbe WF, Palmer SJ, Butler WS, Liggins SG. Fetal survival after prednisone suppression of maternal lupus anticoagulant. Lancet 1983; 1:1361-3.
Pratt et al.
Thyroid antibodies and recurrent abortion
1005
©
Vol. 60, No.6, December 1993
1993 The American Fertility Society
Printed on acid-free paper in U S. A.
The association of antithyroid antibodies in euthyroid nonpregnant women with recurrent first trimester abortions in the next pregnancy
Donna E. Pratt, M,D,* George Kaberlein, B.S, Alan Dudkiewicz, Ph.D,
Vishvanath Karande, M.D, Norbert Gleicher, M.D.
The Center for Human Reproduction and The Foundation for Reproductive Medicine Inc., Chicago, Illinois
Objective: To evaluate the prognostic value of antithyroid antibodies in euthyroid women with a history of recurrent first trimester abortions on future pregnancy loss. Design: The sera of 42 euthyroid women with a history of three or more consecutive first trimester abortions were evaluated for the presence of antibodies to thyroglobulin and thyroid peroxidase before pregnancy and again as soon as the diagnosis of pregnancy was made. Setting: Medical school-affiliated private infertility center. Patients: Forty-two women with a history of three or more consecutive first trimester abortions who were planning to conceive again. Main Outcome Measure: The presence of antithyroid antibodies in the nonpregnant state and their association with pregnancy loss in the next gestation. Results: Thirteen of 42 women (31 %) were positive for the presence of antithyroid antibodies at the initial screening before pregnancy. All 13 maintained positivity by the time their next pregnancy was diagnosed. Only 12 of those 42 women (29%) experienced a first trimester abortion. Eight of these 12 women (67%) were positive for one or more antithyroid antibody. In contrast, among 30 nonaborting women, only 5 of 30 (17%) exhibited thyroid antibody positivity. The detection of thyroid antibodies before conception carried an increased risk of pregnancy loss in the next pregnancy (8 of 13,62% versus 4 of 29,14%). Conclusion: The presence of antithyroid antibodies in nonpregnant women with a history of recurrent abortion identifies a subgroup of women at significantly increased risk for yet another pregnancy loss in their next gestation. Because organ-specific autoantibodies thus demonstrate similar prognostic significance to nonorgan-specific autoantibodies, it is tempting to conclude that peripheral autoantibody abnormalities seen in habitual aborters only reflect an underlying T -lymphocyte defect, which may be the actual cause of pregnancy loss. Fertil SterilI993;60:1001-5 Key Words: Antithyroid antibodies, recurrent abortion
Pregnancy loss seems to occur with increased frequency in the presence of abnormal autoimmune function. This association was first made in reference to lupus anticoagulant and other phospholipid antibodies (1, 2). Stagnaro-Green et al. (3) and
Received June 8, 1993; revised and accepted August 3, 1993.
* Reprint requests: Donna Pratt, M.D., The Center for Human Reproduction, 750 North Orleans Street, Chicago, Illinois 60610. Vol. 60, No.6, December 1993
Glinoer et al. (4) recently reported a strong correlation between the presence of thyroid antibodies and first trimester pregnancy loss. Those thyroid antibody studies were performed during the first trimester of pregnancy and it is therefore still unknown whether the presence of thyroid antibodies before pregnancy can serve as an independent marker of risk for losing an upcoming gestation. Although the total incidence of recognized and unrecognized pregnancy loss per pregnancy has Pratt et al.
Thyroid antibodies and recurrent abortion
1001
5i
been reported to be as high as 31 % (5), the actual risk in a young population with no prior history of pregnancy loss is probably <15% (6). This is probably best reflected by investigating pregnancy loss rates after infertility treatment where early pregnancy losses are more likely recognized than in other populations. A higher miscarriage rate in gonadotropin-induced cycles (17% to 31%) than in natural cycles (10% to 15%) still reflects the above noted general pregnancy loss rate (7). This rate also was confirmed recently by Mills et al. (8). In a welldesigned multicenter study, they reported spontaneous abortion rates of 11.8% to 20.8% (8). To define a higher risk group of patients for pregnancy loss, a group of so called habitual (or repeated) aborters was chosen for this study. Such patients by definition have had three or more consecutive pregnancy losses and are statistically at greatly increased risk for yet another loss in their next pregnancy (9, 10). They therefore represent a favorable population pool for attempts to determine the potential value of thyroid antibodies as predictors of pregnancy loss in upcoming pregnancies. MATERIALS AND METHODS
The sera of 42 nonpregnant women with a history of three or more consecutive pregnancy losses were investigated for the presence of thyroid antibodies ~ 6 months before pregnancy and again as soon as a pregnancy was diagnosed by rising hCG titers at 4 to 6 weeks from the last menstrual period. Once pregnancy occurred, the patients were followed weekly with ultrasound and hormonal evaluations until 12 weeks gestational age, when the patients were transferred out into their local communities for obstetrical care. Because we were unable to obtain reliable follow up on all patients in the second and third trimester, this study reports only the incidence of first trimester pregnancy loss. A smaller number of additional second trimester losses can be expected and did in fact occur. Patients were selected from a larger pool of repeated aborters after women with obvious other causes for pregnancy loss had been excluded. Phospholipid antibodies to cardiolipin, phosphatidyl serine, phosphatidic acid, phosphatidyl ethanolamine, phosphatidyl inositol, and phosphatidyl glycerol for immunoglobulin (Ig)G, IgM and IgA isotopes were evaluated. Patients with significant uterine anomalies, uterine myomas, parental chromosomal abnormalities detected from peripheral 1002
Pratt et al.
Thyroid antibodies and recurrent abortion
....
Table 1 Demographic Data of Women who Aborted Versus Women who did not Abort'
Age
No. of prior abortions
y
Study group (n = 42) Positive antibodies (n = 13) Negative antibodies (n = 39) Aborted (n = 12) Did not abort (n = 40)
34 33 34 33 34
± 3.1
± 2.9 ± 3.4 ± 2.3 ± 3.7
3.6 4.2 3.4 3.8 3.6
± ± ± ± ±
1.1 1.5 .91 1.0 1.2
'Values are means ± SD.
blood, and autoimmune abnormalities among phospholipid antibodies were excluded. All patients underwent at least a hysterosalpingography, and often a hysteroscopy and laparoscopy as well, as part of their workup. Not excluded were patients with possible luteal phase defect because patients were substituted in the luteal phase with P in oil. The majority of pregnancy losses in the final study group had not undergone prior chromosomal evaluation of products of conception. Chromosomal abnormalities in such earlier losses could therefore not be ruled out. Table 1 summarizes the demographic data of the study population, which ranged in age from 27 to 41 years (mean ± SD, 34 ± 3.1) and reported between three and eight consecutive pregnancy losses (mean ± SD, 3.6 ± 1.1). Only 3 of 42 women reported a live birth before their consecutive losses (secondary aborters) whereas the remainder had no living children (primary aborters). The ages ofthe secondary aborters were 32, 35, and 36 years. All had one live birth that, in all three cases, was the first pregnancy. The number of spontaneous abortions after the live births were 4, 3, and 3, respectively. Only one of the secondary aborters was positive for antithyroid antibodies. This subject was the only secondary aborter to experience another spontaneous abortion. All study patients were determined to be euthyroid before pregnancy by demonstrating normal levels of thyroid-stimulating hormone (TSH) and T 4 • Laboratory Evaluation
Thyroid-stimulating hormone and free T 4 were determined by using the Stratus, automated fluorometric enzyme immunoassay (Baxter Diagnostics Inc., Deerfield, IL). For TSH, plastic carriers containing a solution of mouse monoclonal antihuman TSH IgG complexed onto the surface of Fertility and Sterility
-
glass fiber paper by the addition of goat antisera to mouse Igs, are placed into an analyzer loading station. The carriers are transferred to a circular holding track where they pass through a series of "work stations." At the initial station 90 lIL of the patients' sera is delivered to the carrier. After a 60-second incubation period, labeled antibody conjugate (a solution containing calf intestinal alkaline phosphatase covalently linked to a mixture of two mouse monoclonal anti-human TSH) and substrate wash (a solution containing 1 mmol/L 4methylumbelliferyl phosphate in a diethanolamine buffer) are added at subsequent work stations. At the final station, after a second 60-second incubation period, the quantity of bound enzyme-labeled anti-human TSH is interpreted by fluorometry. The same automated system is used for the determination of free T 4. The carrier that contains rabbit anti -T 4 serum is complexed onto the surface of glass filter paper by the addition of goat antisera to rabbit Igs. The conjugate is calf intestinal alkaline phosphate covalently linked to T 4. The normal range for TSH is defined as 0.35 to 7.0 lIIU/mL and for free T4 as 0.9 to 2.1 ng/dL. The mean sensitivity of the human TSH assay is 0.05 lIIU /mL and 0.2 ng/dL for free thyroxine. The mean interassay coefficient of variation and the intra-assay coefficients of variation in our assay were 3.7 and 3.2, respectively. Thyroid peroxidase antibodies and thyroid globulin antibodies were assayed using Kalibre RIA kits (Kronus, Dana Point, CA). These are both highly sensitive RIAs with a range of detection between 0.1 and 30 U/mL. Assays were standardized against Medical Research Council thyroglobulin and microsomal antigen autoantibody first international reference preparations (coded 66/387 and 65/93, respectively) where 1 U is equivalent to approximately 10 IU (personal communication, Kronus Technical Services). Briefly, the thyroid peroxidase assay is a single antibody RIA in which 50 lIL of prediluted calibrators and 50 lIL of patient sera are placed in appropriately labeled tubes. Fifty microliters of p25 thyroid peroxidase tracer is added to each tube and the tubes are gently mixed and incubated for 60 minutes at room temperature (RT). A protein A reagent (50 lIL) is then added to all tubes, mixed, and incubated for 60 minutes at RT. One milliliter of diluent, containing 0.15 M NaCl 0.02 M Tris-HCl, ph 7.5; bovine serum albumin; and 0.002 M sodium azide is added and centrifuged at 1500 X g for 15 minutes. The supernatants are decanted and the radioactivity in the tubes is counted for 1 Vol. 60, No.6, December 1993
minute in a gamma scintillation counter. The same procedure is used for the thyroid globulin antibodies substituting p25 thyroid globulin tracer (11). Comparisons of this assay to the tanned red cell test indicated an antithyroid peroxidase antibody and antithyroid globulin antibody titer of 1:100 and 1:10, respectively, was equivalent to 10 U/mL by the direct RIA method. Enzyme-linked immunosorbent assay comparisons showed a highly significant correlation (r = 0.99). The lower limits of detection of the ELISAs were equivalent to 5 to 10 U/mL by direct RIA. Assays based on I 125 -labeled protein A showed similar lower limits of detection to the ELISA assay (5 to 10 U/mL). Correlations were also highly significant (r = 0.91) (11). The mean interassay coefficient of variation and the intra-assay coefficient of variation in our assays were 2.6 and 4.2, respectively. Statistical analyses were performed using the Fisher's Exact Test and the unpaired Student's t-test, using GraphPad Instat, GraphPad Software version 1.15 (GraphPad Software Inc., San Diego, CA).
RESULTS Thirteen of 42 (31%) women were positive for either one or both antithyroid antibodies during their initial screening in the non-pregnant state. All 13 were still positive at the time of pregnancy diagnosis. No previously negative woman became positive early in pregnancy. Among 13 thyroid antibody-positive women, 9 were positive for thyroid peroxidase antibody, 11 for thyroid globulin antibody, and 7 for both antibodies. The levels of thyroid peroxidase antibody ranged from 0.30 to 44.67 U/mL (mean ± SD, 15.93 ± 15.387 U/mL), while the levels for thyroid globulin antibody ranged from 0.32 to 40.44 U/mL (10.95 ± 15.38 U/mL) (Table 2). There was no significant difference in mean levels between thyroid peroxidase antibody and thyroid globulin antibody nor was there a significant difference in the antithyroid antibody levels before and during early pregnancy (data not shown). Twelve of 42 women aborted during the first trimester (29%). Eight of the 12 (67%) were positive for one or both antithyroid antibodies. In contrast, only 5 of the 30 women (17%) who did not abort demonstrated positive antithyroid antibodies (P = 0.003). The presence of thyroid antibodies before conception carried a significantly increased risk of pregnancy loss in that 8 of 13 antibody-positive and Pratt et al.
Thyroid antibodies and recurrent abortion
1003
BE
.£1
Table 2
Subject
a
•
Antithyroid Antibody Levels
Age
No. of spontaneous abortions
31 37 32 29 34 35 32 31 39 37 36 34 34
Thyroid globulin
Thyroid peroxidase U/mL
y
1* 2* 3 4* 5 6* 7* 8* 9 10 l1*t 12 13*
s
3 4 4 4 3 5 3 3 8 4 3 6 4
44.96 14.50 Not detected 0.30 9.54 Not detected 4.08 10.11 Not detected 1.50 32.96 Not detected 25.50
4.50 2.54 1.15 3.0 1.50 36.20 40.44 Not detected 1.33 3.24 Not detected 26.20 0.32
* Patients that aborted. t Secondary aborter.
only 4 of 29 antibody-negative women miscarried (62% versus 14%, P = 0.003). Analysis of sensitivity and specificity were 66.6% and 83.3%, respectively. Testing for thyroid antibodies in a population of recurrent aborters gave a positive predictive value of 61.5% and had a negative predictive value of 86.2%. DISCUSSION
This study confirmed the increased risk status of our study population for pregnancy loss. The loss of 30% of pregnancies in the first trimester indicates a rate of pregnancy loss that is almost twice as high as expected in a normal population (5-10). These data, however, also confirm that even women with a large number of first trimester consecutive pregnancy losses still have an excellent chance of carrying a next pregnancy past the first trimester. Among women at increased risk to miscarry, the presence of thyroid antibodies in the nonpregnant state represents a marker ofrisk. Women who demonstrate thyroid antibodies before conception miscarried in 8 of 13 cases (62%) whereas among antibody negative patients only 4 of 29 miscarried (14 %). This represents a statistically significant difference (P = 0.003). The importance of thyroid antibodies is also reflected in the fact that miscarrying patients demonstrated a significantly higher incidence of antibody positivity (67% versus 17%, P = 0.003). Antithyroid antibodies are known to occur in apparently healthy populations and are more fre1004
Pratt et al.
Thyroid antibodies and recurrent abortion
quently observed in women during their reproductive years (12-14). We previously reported a higher incidence of antithyroid antibodies in women with a history of three or more consecutive pregnancy losses as compared with a control population of apparently healthy blood donors (31% versus 19%) (15). Thyroid disease has historically not been associated with pregnancy loss (6, 10, 16). Recently, Stagnaro-Green et al. (3) and shortly thereafter Glinoer et al. (4) reported, however, for the first time a specific association between antithyroid antibodies and pregnancy loss. Nonorgan-specific antibody abnormalities, and especially antiphospholipid antibody abnormalities, have been associated for a much longer time with pregnancy loss (2, 17, 18). To find organ-specific thyroid antibodies in parallel to nonorgan-specific autoantibody abnormalities is, however, not uncommon. For example, thyroid antibodies also have been reported in up to 45% of patients with systemic lupus erythematosus (19); and at least one-third of women with autoimmune-induced premature ovarian failure demonstrate nonorgan-specific autoantibodies as well as antithyroid antibodies (20). Thyroid antibodies, not dissimilar to nonorganspecific autoantibodies, may thus simply serve as peripheral markers of abnormal T-cell function. It is then tempting to speculate that abnormal T -lymphocyte function, and not abnormal autoantibodies per se, is responsible for pregnancy loss. Support for such a concept comes from an increasing recognition that the basic immunologic defect in autoimmune diseases, inclusive of thyroid diseases, probably lies at the T-lymphocyte level (21). Under such a concept, organ-specific autoantibody abnormalities, such as thyroid antibodies and nonorgan-specific autoantibodies, can be expected to have similar prognostic significance for impending pregnancy loss. This is in fact confirmed by this study because patients with classical nonorganspecific autoantibody abnormalities were excluded. One could further conclude that successful treatment of autoantibody-associated pregnancy loss should concentrate on attempts to modulate T-cell function rather than the customary treatment options of antibody suppression and/or modulation of antibody effects (1, 22).
REFERENCES 1. Lubbe WF. Liggins GC. Role of the lupus anticoagulant and autoimmunity in recurrent pregnancy loss. Semin Reprod Endocrinol 1988;6:181-90.
Fertility and Sterility
2. Cowchock S, Smith RJ, Gocial B. Antibodies to phospholipids and nuclear antigens in patients with repeated abortion. Am J Obstet GynecoI1986;155:1002-10. 3. Stagnaro-Green A, Roman SH, Cobin RH, El-Harazy Essam, Alvary-Marfany M, Davies TF. Detection of at-risk pregnancy by means of a highly sensitive assays for thyroid autoantibodies. JAMA 1990;264:1422-5. 4. Glinoer D, Soto MF, Bourdoux P, Lejeune B, Delange F, Lemone M, et al. Pregnancy in patients with mild thyroid abnormalities: maternal and neonatal repercussions. J Clin Endocrinol Metab 1991;73:421-7. 5. Wilcox AJ, Weinberg CL, O'Connor JF, Baird DD, Schlaterer JP, Canfield RE. Incidence of early loss of pregnancy. N Engl J Med 1988;319:189-94. 6. Regan L, Braude PR, Trembath PH. Influence of past reproductive performance on risk of spontaneous abortion. Br Med J 1989;299:541-6. 7. Ranson MX, Bohrer M, Blotner MB, Kemmann E. The difference in miscarriage rates between menotropin-induced and natural cycle pregnancies is not surveillance related. Fertil Steril 1993;59:567-70. 8. Mills JL, Holmes LB, Aarons JH, Simpson JL, Brown ZA, Jovanovic-Peterson LG, et al. Moderate caffeine use and the risk of spontaneous abortion and intrauterine growth retardation. JAMA 1993;269:593-7. 9. Alberman E. The epidemiology of repeated abortion. In: Sharp F, Beard RW, editors. Early pregnancy loss mechanisms and treatment. London: Royal College of Obstetrics and Gynecologists, 1988:9-17. 10. Harper JH, Archer DF, Marchese SG, Muraca-Clemens M, Garner TH. Etiology of recurrent pregnancy losses and outcome of subsequent pregnancies. Obstet Gynecol 1983;62: 574-81. 11. Beever K, Bradbury J, Phillips D, McLachdan SM, Pegg C, Goral A, et al. Highly sensitive assays of autoantibodies to thyroglobulin and to thyroid peroxidase. Clin Chern 1989;35:1949-54. 12. Vargas MT, Brines-Urbina R, Gladman D, Papsin FR,
Vol. 60, No.6, December 1993
13.
14.
15.
16.
17. 18.
19.
20.
21.
22.
Walfish PG. Antithyroid microsomal autoantibodies and HLA -D R5 are associated with post partum thyroid dysfunction: evidence supporting an autoimmune pathogenesis. J Clin Endocrinol Metab 1988;67:327-33. Goodwin MT, Montaro M, Westman J. Transient hyperthyroidism and hyperemesis gravidarum: clinical aspects. Am J Obstet Gynecol 1992;167:648-52. Jansson R. Bernander S, Karlson A, Levin K, Nelsson G. Autoimmune thyroid dysfunction in the post partum period. J Clin Endocrinol Metab 1984;58:681-7. Pratt D, Novotny M, Kaberlein G, Dudkiewicz A, Gleicher N. Antithyroid antibodies and the association with non-organ-specific antibodies in recurrent pregnancy loss. Am J Obstet Gynecol 1993;168:837-41. Montora M, Callea JV, Fraizier SD, Westman JH. Successful outcome of pregnancy in women with hypothyroidism. Ann Intern Med 1981;94:31-4. Maier DB, Parke A. Subclinical autoimmunity and recurrent aborters. Fertil Steril 1989;51:280-5. 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: 1376-80. Magero M, Zoli A, Altomonte L, Mirone L, Barini LSA, Scuder F. The association of silent thyroiditis and active systemic lupus erythematosus. Clin Exp Rheumatol 1992;10:67-70. La Barera A, Miller MM, Ober C, Rebor RS. Autoimmune etiology in premature ovarian failure. Am J Reprod Immunol 1988;16:114-22. Stagnaro-Green A, Roman SH, Cohen RH, El-Harazy E, Wallerstein S, Davies TF. A prospective study of lymphocyte-initiated immuno-suppression in normal pregnancy: evidence of a T-cell etiology for post partum thyroid dysfunction. J Clin Endocrinol Metab 1992;74:645-53. Lubbe WF, Palmer SJ, Butler WS, Liggins SG. Fetal survival after prednisone suppression of maternal lupus anticoagulant. Lancet 1983; 1:1361-3.
Pratt et al.
Thyroid antibodies and recurrent abortion
1005