Objective
Fetal microchimerism may have a role in development of autoimmune thyroid disorders. Using parity as a surrogate for increasing fetal cell exposure, we analyzed its association with thyroid peroxidase antibody levels.
Study Design
Secondary analysis of serum thyroid analytes determined in 17,298 women from a population-based prospective study between 2001 and 2003. Sera were assayed for thyrotropin, free thyroxine, and antithyroid peroxidase antibodies. We analyzed the relationship between thyroid peroxidase antibodies and increasing parity.
Results
The incidence of abnormally elevated thyroid peroxidase antibody levels (>50 IU/mL) increased with advancing parity, but was not significant after adjustment for maternal characteristics. However, at higher thyroid peroxidase antibody levels (>500 IU/mL), a significant relationship with advancing parity persisted after adjustments ( P = .002).
Conclusion
Advancing parity is associated with an increased risk for high serum concentrations of antithyroid peroxidase antibodies. This suggests fetal microchimerism may play a role in development of autoimmune thyroid disorders.
It is now well-established that there is transfer of fetal cells into the maternal circulation even as early as the first trimester. Fetal microchimerism results when fetal cells, some with stem cell-like properties, persist in maternal blood or other tissues. These so-called pregnancy-associated progenitor cells are thought to play a role in repairing damaged maternal tissues, but they have also been implicated in the genesis of autoimmune disorders through stimulation of a maternal immune response directed against foreign fetal cell antigens. Support for the role of microchimerism in triggering at least some autoimmune diseases come from studies identifying fetal cells in target tissues taken from women who had later developed systemic sclerosis or lupus erythematosus. For example, women with systemic sclerosis are much more likely to be HLA class II compatible with at least 1 of her offspring than is usual. Other circumstantial evidence of such a role for microchimerism is the overwhelming predilection of autoimmune disorders to affect women compared with men. Systemic sclerosis and lupus, for example, have an 8- to 9-fold increased prevalence in females.
Fetal microchimerism has also been implicated in autoimmune thyroid diseases (AITD). Major examples include Hashimoto and Graves’ disease, both characterized by various elevated titers of antithyroid antibodies. In addition, both of these diseases are identified with a 10-fold increased prevalence in women compared with men. There is also direct evidence that supports a role for microchimerism in AITD, as first reported by Srivasta et al. These investigators applied fluorescent in situ hybridization techniques to histopathologic sections of thyroid tissue from women with Hashimoto thyroiditis. Along with lymphocytic infiltration characteristic of this type of autoimmune thyroiditis, they found Y-chromosome-bearing cells in 83% of tissue specimens from women who had delivered male infants. Subsequent investigators confirmed these findings using similar techniques. Taken together, the increased incidence of immunologically mediated thyroid disorders in women as well as the presence of fetal cells in their thyroid glands supports the concept that microchimerism stimulates maternal antithyroid antibodies directed against fetally derived antigens. The most frequent antithyroid antibodies associated with Hashimoto thyroiditis are those directed against thyroid peroxidase (TPO). TPO antibodies directly inhibit this intrathyroidal enzyme and cause thyrocyte destruction by complement-mediated and/or antibody-dependent cell-mediated cytotoxicity.
With this evidence for fetal microchimerism as a stimulus to produce maternal antithyroid antibodies, it would seem reasonable that such antibodies would be detected more often in women with greater exposure to fetal antigens because of an increasing number of pregnancy episodes. Investigations designed to assess the potential link between increasing parity and accrual of TPO antibodies have reported conflicting findings. The current study was designed to evaluate the association of parity with the incidence of serum TPO antibodies in a cohort of more than 17,000 pregnant women.
Materials and Methods
As described previously, after obtaining approval from the University of Texas Southwestern Medical School Institutional Review Board, we stored excess serum at −80°C in aliquots derived from blood samples drawn for rubella antibody screening in all women presenting for prenatal care at Parkland Health and Hospital Systems from Nov. 1, 2000, to April 14, 2003. For this study, we retrieved samples from women screened during the first 20 weeks’ gestation and who were delivered of a singleton infant weighing ≥500 g. Each specimen was thawed and analyzed for concentration of serum TPO antibodies using a chemiluminescent immunoassay (Immulite 2000 Analyzer; Diagnostic Products Corporation, Los Angeles, CA). The analytical sensitivity of this assay was 5.0 IU/mL, and its coefficient of variation was 9.8% within run and 11.3% between runs. TPO levels <10 IU/mL and greater than 1000 IU/mL using the automated assay were not further quantified. TPO antibody levels ≤50 IU/mL were considered to be normal and were classified as TPO antibody negative, whereas levels >50 IU/mL were classified as TPO antibody positive. The TPO threshold of 50 IU/mL was selected considering both the manufacturer recommended threshold (35 IU/mL) and the impact of freezing on antibody levels in samples stored more than 2 years. TPO assay results were linked with our computerized perinatal database at Parkland Hospital. Briefly, this database is created from obstetric data sheets completed by the attending nurse at each delivery. Specially trained research nurses verify the data for consistency and completeness before electronic storage.
TPO antibody status was analyzed according to gravidity, parity, maternal age, and other maternal characteristics that included race, gestational age at screening, and body mass index (BMI kg/m 2 ). Gravidity was defined as the total number of pregnancy episodes regardless of outcome. For gravidity, the current pregnancy was counted along with all past pregnancies, including those with abortive outcomes. Parity was defined as the total number of pregnancy episodes that resulted in delivery of an infant weighing ≥500 g and/or delivery after 23 weeks’ gestation.
Pearson’s χ 2 and Student t test were used for univariate 2-group comparisons. The Mantel-Haenszel χ 2 was incorporated to evaluate trends in frequencies across ordinal categories. Logistic regression was applied to examine the significance for selected pregnancy outcomes adjusted for covariates known to be related to TPO antibody status such as maternal age, race, gravidity, parity, and BMI. Statistical computations were performed using SAS, version 9.2 (SAS Institute, Cary, NC). Two-sided P values less than .05 were judged statistically significant.
Results
During the study period, 17,298 women underwent screening for thyroid-associated analytes. Women with subclinical hypothyroidism and isolated maternal hypothyroxinemia were included, but importantly, all women with overt thyroid disease were excluded. Of those remaining, 6% (n = 1012) had a serum TPO antibody level of >50 IU/mL and thus met the criteria for TPO antibody positivity. The distribution of these levels is shown in Figure 1 .
There was a significant relationship between increasing maternal age and the prevalence of TPO antibody positive status as shown in Figure 2 ( P < .001). This progressively increased from 4% in women aged 20 years or younger to 10% in women >40 years old. This association remained significant after correcting for gravidity, parity, gestational age at screening, BMI, and race ( P < .001).
As shown in Figure 3 , univariate analysis of the risk of TPO antibody positivity according to parity suggested an association with TPO antibody positive status and advancing parity ( P < .01). Nulliparous women served as the referent population. After logistical regression to correct for maternal age, gestational age at screening, BMI, and race, however, this association between parity and TPO antibody positive status was no longer significant ( P = .54). Because fetal cell transfer has been reported to begin as early as the first trimester, we also tested the risk of TPO positivity using primigravid, nulliparous women as the reference group and obtained similar findings ( P = .30).
