Perinatal Outcome of Children Born to Mothers with Thyroid Dysfunction or Antibodies: A Prospective Population-Based Cohort Study

ORIGINAL Endocrine ARTICLE Care Perinatal Outcome of Children Born to Mothers with Thyroid Dysfunction or Antibodies: A Prospective Population-Based Cohort Study Tuija Männistö, Marja Vääräsmäki, Anneli Pouta, Anna-Liisa Hartikainen, Aimo Ruokonen, Heljä-Marja Surcel, Aini Bloigu, Marjo-Riitta Järvelin, and Eila Suvanto-Luukkonen Departments of Obstetrics and Gynecology (T.M., M.V., A.P., A.-L.H., E.S.-L.) and Clinical Chemistry (A.R.), and Institute of Health Sciences (T.M., M.-R.J.), University of Oulu, 90014 Oulu, Finland; Department of Child and Adolescent Health (A.P., H.-M.S., A.B., M.-R.J.), National Public Health Institute, 90101 Oulu, Finland; and Department of Epidemiology and Public Health (M.-R.J.), Imperial College London, London SW7 2AZ, United Kingdom Context: There are only a few large prospective studies involving evaluation of the effect of maternal thyroid dysfunction on offspring and observations are inconsistent. Objective: The objective of the study was to investigate the effects of thyroid dysfunction or antibody positivity on perinatal outcome. Setting and Participants: The study included prospective population-based Northern Finland Birth Cohort 1986 including 9247 singleton pregnancies. First-trimester maternal serum samples were analyzed for thyroid hormones TSH, free T 4 (ft4) and antibodies thyroid-peroxidase antibody (TPO-Ab) and thyroglobulin antibody (TG-Ab). Mothers were classified by their hormone and antibody status into percentile categories based on laboratory data and compared accordingly. Main Outcomes: Outcomes were perinatal mortality, preterm delivery, absolute and gestational age-adjusted birth weight, and absolute and relative placental weight. Results: The offspring of TPO-Ab- and TG-Ab-positive mothers had higher perinatal mortality, which was not affected by thyroid hormone status. Unadjusted and adjusted (for maternal age and parity) risk for increased perinatal mortality was an odds ratio of 3.1 (95% confidence interval 1.4 7.1) and 3.2 (1.4 7.1) in TPO-Ab- and 2.6 (1.1 6.2) and 2.5 (1.1 5.9) in TG-Ab-positive mothers. TPO-Ab-positive mothers had more large-for-gestational age infants (2.4 vs. 0.8%, P 0.017), as did mothers with low TSH and high ft4 concentrations vs. reference group (6.6 vs. 2.5%, P 0.045). Significantly higher placental weights were observed among mothers with low TSH and high ft4 or high TSH and low ft4 levels as well as among TPO-Ab-positive mothers. Conclusions: First-trimester antibody positivity is a risk factor for perinatal death but not thyroid hormone status as such. Thyroid dysfunction early in pregnancy seems to affect fetal and placental growth. (J Clin Endocrinol Metab 94: 772 779, 2009) A considerable number of pregnant women have undiagnosed or insufficiently treated thyroid dysfunction during pregnancy (1 4), and its suspected adverse impact on the outcome of the offspring is currently a subject of much discussion (5 7). Even subclinical hypothyroidism in the presence or absence of thyroid antibodies is regarded as a threat to a woman s ISSN Print 0021-972X ISSN Online 1945-7197 Printed in U.S.A. Copyright 2009 by The Endocrine Society doi: 10.1210/jc.2008-1520 Received July 15, 2008. Accepted December 15, 2008. First Published Online December 23, 2008 fertility, pregnancy outcome, and the development of the child. Several pregnancy and perinatal complications have been considered to be associated with maternal thyroid dysfunction: spontaneous abortions (1, 8), fetal death (2), preterm delivery (1, 2, 4, 9), small head circumference and low birth weight (10), and impaired neuropsychological development (11). In addition, Abbreviations: BMI, Body mass index; CI, confidence interval; ft3, free T 3 ; ft4, free T 4 ; LGA, large for gestational age; NFBC, Northern Finland Birth Cohort; SGA, small for gestational age; TG-Ab, thyroglobulin antibody; TPO-Ab, thyroid-peroxidase antibody. 772 jcem.endojournals.org J Clin Endocrinol Metab. March 2009, 94(3):772 779

J Clin Endocrinol Metab, March 2009, 94(3):772 779 jcem.endojournals.org 773 Target population n=9575 Group A (Clinical hypothyroidism) n=54 Group B (Subclinical hypothyroidism) n=224 Group C (Subclinical hyperthyroidism) n=204 Denied use of data n=251 (2.5%) Insufficient or missing samples n=3014 (31.8%) TSH and ft4 analyses Reference group n=4719 Group D (Clinical hyperthyroidism) n=77 NFBC 1986 n=9479 (99%) FMC samples available n=5805 (61.2%) TPO-Ab analyses Multiple pregnancy n=222 (2.3%) Gestational week at FMC sampling >20 n=187 (2.0%) TPO-Ab-positive n=288 TPO-Ab-negative n=5475 TG-Ab analyses TG-Ab-positive n=285 TG-Ab-negative n=5420 Subjects and Methods Study population and data collection The prospective Northern Finland Birth Cohort (NFBC) 1986 covered 99% of all births with calculated term between July 1, 1985, and June 30, 1986, drawn from the two northernmost provinces of Finland (9362 mothers, 9479 children) (16, 17). We included only singleton pregnancies in the present study (n 9247). The cohort has been followed up since the 12th gestational week of pregnancy. The first questionnaire on demographic, biological, health behavioral, and socioeconomic characteristics of the mothers/families covered the period up to the 24th gestational week, when the mothers were enrolled in the study if still pregnant. The second questionnaire covered maternal health and health behavior during pregnancy and the perinatal period. The mothers were assisted in completion of the questionnaire by the nurses, who ensured that all questions were answered. The third questionnaire contained items about pregnancy complications and diseases, delivery, and neonatal outcome and was completed in the maternity hospitals by the attending midwives. All women gave birth at hospital. The ethics committees of the University of Oulu and the National Public Health Institute approved this study. Informed consent was obtained from all subjects. Reference group: TSH between 5th and 95th percentiles, ft4 between 5th and 95th percentiles Group A: TSH over 95th percentile, ft4 under 5th percentile Group B: TSH over 95th percentile, ft4 between 5th and 95th percentiles Group C: TSH under 5th percentile, ft4 between 5th and 95th percentiles Group D: TSH under 5th percentile, ft4 over 95th percentile NFBC 1986=Northern Finland Birth Cohort 1986 FMC=Finnish Maternity Cohort TSH=thyroid-stimulating hormone, ft4=free thyroxine, TPO-Ab=thyroid-peroxidase antibody, TG-Ab =thyroglobulin antibody FIG. 1. Flow chart of the study population. normal but low concentrations of thyroid hormone in the mother have been associated with increased prevalence of breech presentation (12). Thyroid autoantibodies have also been hypothesized to be independently related to increased rates of spontaneous abortion and preterm delivery (1, 9, 13). The results of individual studies on these perinatal outcomes are, however, controversial, probably because of differences in study design, populations, and laboratory data. The precise prevalence of thyroid dysfunction during pregnancy and its suspected association with adverse perinatal outcome can be detected only by means of well-designed populationbased prospective cohort studies with proper documentation. To our knowledge there are only a few large prospective cohort studies (2, 4, 14, 15) involving evaluation of the effect of maternal thyroid dysfunction on offspring and none of these studies provide large amounts of biochemical data. In this prospective, population-based cohort study with comprehensive biochemical measurements from the early pregnancy, we evaluated the effect of maternal thyroid dysfunction and autoimmunity on perinatal outcomes including preterm delivery, birth measures, presentation at birth, and perinatal mortality. Serum samples and laboratory assays The biochemical data were obtained by assay of serum samples from the Finnish Maternity Cohort. This is a biobank, consisting of serum samples collected from all pregnant women in Finland and approved under Finnish law. The law provides the usage of the samples in studies promoting public health. The samples were stored at 25 C. The effect of freezing, thawing, and frozen storage on thyroid laboratory parameters has been reported previously (18). Quantitative analyses of thyroid hormones TSH, free T 3 (ft3), and free T 4 (ft4) and autoantibodies thyroid-peroxidase antibody (TPO-Ab) and thyroglobulin antibody (TG-Ab) were performed by way of chemiluminescent microparticle immunoassays, using an Architect i2000 automatic analyzer (Abbott Diagnostics, Abbott Park, IL). The lower limits of detection and intraand interassay coefficients of variation were 0.0025 mlu/liter, 1.7 and 5.3% for TSH; 5.1 pmol/liter, 3.6 and 7.8% for ft4; 1.53 pmol/liter, 2.3 and 5.0% for ft3; 1.0 IU/ml, 2.5 and 9.8% for TPO-Ab; and 1.0 IU/ml, 2.7 and 8.2% for TG-Ab. The number of serum samples analyzed was 5805 (61.2% of the whole cohort); only samples of a sufficient size ( 1 ml) were included in this study (Fig. 1). The mean gestational age at sampling was 11.0 wk (SD 3.6), and only samples drawn before or at the 20th gestational week were accepted (98% of the samples). Categorization of the study population and outcomes For association analyses the data were categorized using percentiles of laboratory values because the reference values given by the manufacturer of the analyzer (Abbott) apply to a nonpregnant population, and these values may differ from those in a pregnant population (19). Our study is population based and large enough to create our own reference values and to take into account the effect of freezing and storage (18). Concerning thyroid hormones, the mothers with serum concentrations of both TSH and ft4 between the fifth and 95th percentiles were considered to have normal thyroid function (reference group, n 4719). The subjects with thyroid dysfunction were divided into four groups with respect to thyroid hormone levels (Fig. 1): 1) group A, with TSH

774 Männistö et al. Thyroid Dysfunction, Antibodies, and Perinatal Outcome J Clin Endocrinol Metab, March 2009, 94(3):772 779 TABLE 1. Biochemical characteristics of the samples analyzed n Geometric mean 95% CI Fifth percentile 95th percentile TSH (mu/liter) 5779 1.06 1.03 1.08 0.19 3.6 ft4 (pmol/liter) 5726 15.3 15.3 15.4 11.96 20.5 ft3 (pmol/liter) 5737 5.13 5.10 5.15 3.85 6.63 TPO-Ab (IU/ml) 5763 6.23 6.03 6.44 2.0 167.7 TG-Ab (IU/ml) 5705 12.5 12.3 12.7 6.4 47.7 over the 95th percentile and ft4 under the fifth percentile (n 54; clinical hypothyroidism); 2) group B, with TSH over the 95th percentile and ft4 between the fifth and 95th percentiles (n 224; subclinical hypothyroidism); 3) group C, with TSH under the fifth percentile and ft4 between the fifth and 95th percentile (n 204; subclinical hyperthyroidism); and 4) group D, with TSH under the fifth percentile and ft4 over the 95th percentile (n 77; clinical hyperthyroidism). Groups A D were compared with the reference group in the analyses. With respect to thyroid antibodies, we considered mothers to be TPO-Ab or TG-Ab positive if the concentration of the antibody was over the 95th percentile (Fig. 1). The high cutoff level was used instead of manufacturer s reference values because the samples in this study were from pregnant women and after frozen storage (18). The fifth and 95th percentiles as well as geometric means and 95% confidence intervals (CIs) of all laboratory data are presented in Table 1. The mothers positive for thyroid antibodies were compared with antibody-negative mothers. All data concerning the mothers and their obstetric histories were obtained from the questionnaires. The data on perinatal outcomes included gestational age; preterm delivery (birth 37th gestational week); birth measurements birth weight, small for gestational age (SGA) and large for gestational age (LGA) infants, birth length, ponderal index (birth weight/birth length 3 ), head circumference ; Apgar scores; perinatal mortality stillborns and early neonatal deaths ( 7 d after birth) ; neonatal deaths; malformations; presentation at birth; mode of delivery; absolute and relative placental weight; and umbilical cord length. Statistics Categorical variables were compared using Pearson s 2 test or Fisher s exact test, as appropriate. Student s t test for independent samples was used to compare continuous variables between two groups. When four thyroid dysfunction groups were compared with reference group, the comparisons were made by ANOVA followed by pairwise comparisons with Student s t test. The concentrations of thyroid parameters were logarithmically transformed for the calculation of geometric means. Univariate logistic regression analysis was applied to estimate susceptibility to LGA infants, low birth weight ( 2500 g), noncephalic presentation, low Apgar scores ( 7), and perinatal mortality in relation to thyroid parameters. The results were further adjusted for maternal age and parity by multiple logistic regression. P 0.05 was deemed significant. Statistical analysis was performed using SPSS 14.0 software (SPSS Inc., Chicago, IL). Results The demographic data of the mothers in each group according to their thyroid hormone/antibody status are presented in Table 2. These data did not differ substantially from that in the whole cohort. There were clinically minor but statistically significant differences in the background factors of the mothers whose samples were analyzed compared with those with inadequate samples for analysis with respect to maternal age, body mass index (BMI), smoking habit, screening time, and prevalence of diabetes. However, the differences were minor considering clinical significance. There were significant differences between the reference group and groups A D in maternal age, BMI, parity, smoking habits, and prevalence of previous thyroid diseases (Table 2). The mothers of group C were older; those of group B were heavier; groups B, C, and D had higher parity; the groups A, B, and C included fewer smokers, and the groups A, B and D had more often history of thyroid disease than mothers of the reference group. When the mothers were stratified according to antibody status, TPO-Ab- and TG-Ab-positive mothers had higher BMIs and had more often previous thyroid diseases than antibody-negative mothers. In addition, TG-Ab-positive women were older, had higher parity, and were less frequently smokers than TG-Abnegative mothers (Table 2). Therefore, the results were adjusted for maternal age and parity. The infants in group A (clinical hypothyroidism: high TSH and low ft4 levels) had a higher mean ponderal index than infants of the reference group. Absolute and relative placental weights were higher in group A (Table 3). No significant differences were seen in any of the perinatal outcomes when group B (subclinical hypothyroidism: high TSH and normal ft4 levels) was compared with the reference group (Table 3). The infants of group C (subclinical hyperthyroidism: low TSH and normal ft4 levels) less often had Apgar scores of 7 or less at 5 min than infants of the reference group (Table 3). They had a risk of 0.4 of having low Apgar scores at 5 min when compared with the reference group (Table 4). In group D (clinical hyperthyroidism: low TSH and high ft4 levels) both absolute and relative placental weights as well as absolute birth weight and number of LGA infants were higher than in the reference group (Table 3). The infants of group D were at a 2.7-fold greater risk of being LGA compared with infants of the reference group (Table 4), but the risk was no longer significant after adjusting for maternal age and parity. Mothers who were TPO-Ab positive more often had both low birth weight infants and LGA infants than TPO-Ab-negative mothers (Table 5). However, no differences were observed in the frequencies of SGA infants between the groups. The infants of TPO-Ab-positive mothers were at a 2-fold increased risk of being LGA and a 1.7-fold risk of low birth weight (Table 4). In addition, TPO-Ab-positive mothers had a higher relative placental weight than TPO-Ab-negative mothers (Table 5). The offspring of TPO-Ab- and TG-Ab-positive mothers showed 2- to 3-fold greater perinatal mortality than those of antibody-negative mothers (Tables 4 and 5). In TPO-Ab-positive group, four of seven perinatally deceased infants were born very preterm (before gestational wk 28), and in the TG-Ab-positive

J Clin Endocrinol Metab, March 2009, 94(3):772 779 jcem.endojournals.org 775 TABLE 2. Demographic characteristics of the pregnant women Reference group (n 4719) Group A (n 54) Group B (n 224) Group C (n 204) Group D (n 77) TPO-Ab positive (n 288) TPO-Ab negative (n 5475) TG-Ab positive (n 285) TG-Ab negative (n 5420) Age (yr) 28.0 (5.3) 28.7 (5.1) 28.6 (5.8) 29.5 (5.5) a 28.9 (5.8) 28.6 (5.1) 28.2 (5.4) 29.1 (5.4) a 28.1 (5.4) Gestational week at 10.7 (2.8) 10.9 (3.3) 10.9 (2.8) 10.5 (2.2) 10.4 (2.2) 10.3 (2.4) 10.8 (2.8) 10.4 (2.7) 10.7 (2.8) screening BMI (kg/m 2 ) 22.1 (3.4) 22.5 (3.8) 22.6 (3.7) a 22.5 (3.5) 22.0 (2.9) 22.7 (4.0) a 22.2 (3.4) 22.7 (4.0) a 22.2 (3.4) Nulliparous 1628 (34.5%) 12 (22.2%) 69 (30.8%) 45 (22.1%) a 17 (22.1%) a 81 (28.1%) 1842 (33.6%) 73 (25.6%) 1828 (33.7%) Parity 1.3 (0 15) 1.5 (0 7) 1.6 (0 13) a 1.9 (0 12) a 2.1 (0 12) a 1.3 (0 13) 1.4 (0 15) 1.7 (0 13) a 1.4 (0 15) Smokers 1307 (27.7%) 3 (5.6%) a 40 (17.9%) a 31 (15.2%) a 17 (21.1%) 74 (25.7%) 1486 (27.1%) 45 (15.8%) a 1491 (27.5%) Alcohol use during pregnancy 271 (5.7%) 3 (5.6%) 11 (4.9%) 10 (4.9%) 3 (3.9%) 16 (5.6%) 310 (5.7%) 15 (5.3%) 310 (5.7%) Obstetric history Spontaneus abortions 919 (19.5%) 12 (22.2%) 36 (16.1%) 45 (22.1%) 12 (15.6%) 59 (20.5%) 1063 (19.4%) 67 (23.5%) a 1045 (19.3%) Low birth weight infants 212 (4.5%) 1 (1.9%) 8 (3.6%) 9 (4.4%) 5 (6.5%) 13 (4.5%) 246 (4.5%) 15 (5.3%) 242 (4.5%) Preterm births 262 (5.5%) 0 11 (5.3%) 9 (4.4%) 7 (9.1%) 19 (6.6%) 293 (5.3%) 16 (5.3%) 294 (5.4%) Fetal deaths 65 (1.4%) 0 1 (0.4%) 1 (0.5%) 1 (1.3%) 4 (1.3%) 67 (1.3%) 6 (2.2%) 64 (1.2%) Neonatal deaths 56 (1.2%) 0 3 (1.3%) 1 (0.5%) 2 (2.6%) 3 (1.0%) 63 (1.1%) 5 (1.8%) 60 (1.1%) Infertility treatments 253 (5.4%) 3 (5.6%) 10 (4.5%) 14 (6.9%) 4 (5.2%) 18 (6.3%) 290 (5.3%) 17 (6.0%) 287 (5.3%) Previous diseases Thyroid diseases 34 (0.7%) 1 (1.9%) a 13 (5.8%) a 4 (1.9%) 2 (2.6%) a 12 (4.2%) a 33 (0.6%) 6 (2.2%) a 39 (0.7%) Diabetes 10 (0.2%) 0 2 (0.9%) 0 0 1 (0.3%) 11 (0.2%) 0 12 (0.2%) Autoimmune diseases 6 (0.1%) 0 0 0 0 1 (0.3%) 5 (0.1%) 1 (0.4%) 5 (0.1%) Values are mean (SD) or n (%). Reference group had TSH between the 5th and 95th percentile and ft4 between the 5th and 95th percentile; group A (clinical hypothyroidism) had TSH over the 95th percentile, ft4 under 5th percentile; group B (subclinical hypothyroidism) had TSH over 95th percentile, ft4 between 5th and 95th percentile; group C (subclinical hyperthyroidism) had TSH under the 5th, ft4 between the 5th and 95th percentile; group D (clinical hyperthyroidism) had TSH under the 5th percentile, ft4 over the 95th percentile. a P 0.05 when comparing groups A D together or separately with the reference group or comparing the antibody-positive group with the antibody-negative group.

776 Männistö et al. Thyroid Dysfunction, Antibodies, and Perinatal Outcome J Clin Endocrinol Metab, March 2009, 94(3):772 779 TABLE 3. Perinatal outcome of children grouped according to maternal thyroid hormone status Reference group (n 4719) Group A (n 54) Group B (n 224) Group C (n 204) Group D (n 77) Birth weight (g) 3568 (548) 3676 (657) 3514 (542) 3564 (524) 3703 (568) a Low birth weight 149 (3.2%) 2 (3.7%) 8 (3.6%) 7 (3.4%) 2 (2.6%) SGA 93 (2.0%) 0 5 (2.3%) 2 (1.0%) 1 (1.4%) LGA 116 (2.5%) 3 (5.6%) 4 (1.8%) 3 (1.5%) 5 (6.6%) a Birth length (cm) 50.5 (2.4) 50.6 (2.7) 50.2 (2.5) 50.5 (2.2) 50.9 (2.3) Ponderal index (kg/m 3 ) 27.6 (2.4) 28.4 (2.8) a 27.5 (2.3) 27.5 (2.1) 27.9 (2.2) Head circumference (cm) 35.2 (1.5) 35.4 (2.4) 35.0 (1.6) 35.1 (1.6) 35.2 (1.2) Gestational weeks at birth 39.4 (1.7) 39.4 (2.3) 39.3 (1.8) 39.4 (1.5) 39.4 (1.5) Preterm births Less than 37 gestational 204 (4.3%) 4 (7.4%) 7 (3.1%) 9 (4.4%) 2 (2.6%) weeks Less than 34 gestational 64 (1.4%) 1 (1.9%) 5 (2.2%) 3 (1.5%) 1 (1.3%) weeks Noncephalic presentation 175 (3.7%) 1 (1.9%) 7 (3.2%) 6 (3.0%) 4 (5.2%) Apgar score 7 or less At 5 min 317 (6.7%) 6 (11.1%) 14 (6.3%) 6 (2.9%) a 4 (5.2%) At 10 min 102 (2.7%) 1 (2.6%) 5 (2.9%) 4 (2.5%) 2 (3.4%) At 15 min 42 (1.3%) 1 (2.3%) 3 (2.0%) 2 (1.3%) 0 Perinatal mortality 39 (0.8%) 1 (1.9%) 4 (1.8%) 0 0 Fetal death 24 (0.5%) 1 (1.9%) 1 (0.4%) 0 0 Neonatal death 17 (0.4%) 0 3 (1.4%) 0 0 Malformations 141 (3.0%) 1 (1.9%) 8 (3.6%) 3 (1.5%) 2 (2.6%) Placental weight (g) 642 (132) 688 (142) a 636 (133) 644 (139) 696 (140) a Relative weight of 18.1 (3.3) 19.1 (4.3) a 18.2 (3.2) 18.1 (3.3) 18.9 (2.9) a placenta (%) Umbilical cord length (cm) 56.1 (12.0) 56.0 (9.9) 56.6 (12.0) 56.5 (13.6) 55.6 (11.5) Cesarean sections 620 (14.1%) 8 (15.7%) 29 (14.0%) 27 (13.8%) 7 (9.5%) Values are mean (SD) or n (%). Reference group had TSH between the 5th and 95th percentile, ft4 between the 5th and 95th percentile; group A (clinical hypothyroidism) had TSH over the 95th percentile, ft4 under the 5th percentile; group B (subclinical hypothyroidism) had TSH over the 95th percentile, ft4 between the 5th and 95th percentile; group C (subclinical hyperthyroidism) had TSH under the 5th, ft4 between the 5th and 95th percentile; group D (clinical hyperthyroidism) had TSH under the 5th percentile, ft4 over the 95th percentile. a P 0.05 when comparing groups A D together or separately with the reference group. group, three of six were born very preterm. Mothers who were TG-Ab positive had an almost 2-fold risk of more often having children showing noncephalic presentation at birth (Tables 4 and 5). In addition, we did an analysis in which we evaluated the perinatal outcomes of thyroid dysfunction groups after including and excluding the antibody-positive mothers from the reference group. The results did not differ from those reported. We evaluated independently the effect of maternal underweight (BMI 20 kg/m 2 ) and overweight (BMI 25 kg/m 2 )on the outcomes presented in Table 4. Maternal underweight was not associated with any of the adverse perinatal outcomes. However, maternal overweight was a significant risk factor odds ratio 1.7 1.2 2.5 for LGA. No association was found between maternal overweight and perinatal mortality and low birth weight in our study population. Discussion Maternal thyroid autoantibody positivity but not thyroid hormone status at the end of first trimester was associated with elevated perinatal mortality rate in our study. Preterm delivery was not associated with either thyroid autoantibodies or thyroid hormone status. Greater placental weight and an increased number of LGA infants were associated with both low TSH and high ft4 levels and presence of autoantibodies. We are aware of only four prospective cohort studies (2, 4, 14, 15) in which the effect of thyroid hormone status on perinatal outcome has been investigated. In the study by Allan et al. (2), all thyroid function measurements were made among a small subcohort (n 627), and the women were classified only according to their TSH levels. The large study by Matalon et al. (15) involved evaluation of the perinatal outcome of levothyroxinetreated women, but there were no laboratory data. The two studies by Casey et al. (4, 14) did not include thyroid antibody testing. Thus, the present work is the only large populationbased prospective cohort study with extensive laboratory data concerning the effect of maternal thyroid dysfunction and autoimmunity on perinatal outcome. Although our cohort is large, the rarity of some outcomes such as perinatal mortality and the fact that serum samples could be obtained only once can be regarded as a limitation of the study. However, the most changes in thyroidal function can be observed in the first trimester and are due to placental hormone stimulation (20). In addition, screening programs in routine maternity care usually take place in the first trimester. Our study population is a cohort of normal population, in which thyroid dysfunction and antibody positivity are presumably mostly asymptomatic. The mothers with overt thyroid dysfunction (hypo- or hyperthyroidism) have an increased

J Clin Endocrinol Metab, March 2009, 94(3):772 779 jcem.endojournals.org 777 TABLE 4. Estimated risks of perinatal outcomes in association with maternal thyroid dysfunction (presented as odds ratios) Perinatal outcome Thyroid status n (% of total) Univariate OR (95% CIs) Adjusted OR (95% CIs) LGA TSH and ft4 between fifth-95th 116 (2.5%) 1.0 percentiles TSH fifth percentile, ft4 95th 5 (6.6%) 2.7 (1.1 6.9) 2.5 (0.96 6.3) percentile TPO-Ab negative 130 (2.4%) 1.0 TPO-Ab positive 14 (5.0%) 2.1 (1.2 3.7) 2.2 (1.2 3.8) Birth weight 2500 g TPO-Ab negative 168 (3.1%) 1.0 TPO-Ab positive 15 (5.2%) 1.7 (1.01 3.0) 1.7 (1.01 3.0) Noncephalic presentation TG-Ab negative 186 (3.5%) 1.0 at birth TG-Ab positive 18 (6.3%) 1.9 (1.1 3.1) 1.9 (1.1 3.1) Apgar score at 5 min 7 or TSH and ft4 between fifth-95th 317 (6.7%) 1.0 less percentiles TSH fifth percentile, ft4 between 6 (2.9%) 0.4 (0.2 0.95) 0.5 (0.2 1.1) fifth and 95th percentile Perinatal mortality TPO-Ab negative 43 (0.8%) 1.0 TPO-Ab positive 7 (2.4%) 3.1 (1.4 7.1) 3.2 (1.4 7.1) TG-Ab negative 44 (0.8%) 1.0 TG-Ab positive 6 (2.1%) 2.6 (1.1 6.2) 2.5 (1.1 5.9) ORs adjusted for maternal age and parity. OR, Odds ratio. risk for adverse perinatal outcomes, and hence, they require a more extend follow-up during pregnancy and possibly alterations in their treatment (3). Finland is one of the few countries in the Western world in which iodine deficiency practically does not exist because iodine has been provided as a supplement in the population since the 1940s. The greatest iodine sources in the Finnish diet are dairy products and table salt. Studies from the early 1980s (21, 22) have shown that the daily iodine intake was sufficient, and the iodine intake exceeds the recommended amount, even in the pregnant population (23). Therefore, we presume that the iodine intake was sufficient during the pregnancies of NFBC 1986 mothers, and it can be supposed that cases of thyroid dysfunction arise from thyroid diseases and not from environmental factors. In addition, because practically all pregnant Finnish women have maternity care that is well organized and free of charge, provide TABLE 5. Perinatal outcome of children grouped according to maternal thyroid antibody status TPO-Ab positive (n 288) TPO-Ab negative (n 5475) TG-Ab positive (n 285) TG-Ab negative (n 5420) Birth weight (g) 3551 (658) 3570 (543) 3606 (596) 3567 (548) Low birth weight 15 (5.2%) a 168 (3.1%) 7 (2.5%) 174 (3.2%) SGA 8 (2.9%) 106 (2.0%) 7 (2.6%) 107 (2.0%) LGA 14 (5.0%) a 130 (2.4%) 11 (4.0%) 132 (2.5%) Birth length (cm) 50.3 (2.8) 50.5 (2.3) 50.5 (3.0) 50.5 (2.3) Ponderal index (kg/m 3 ) 27.6 (2.5) 27.6 (2.4) 27.8 (2.6) 27.6 (2.4) Head circumference (cm) 35.1 (2.1) 35.2 (1.5) 35.2 (1.9) 35.2 (1.5) Gestational weeks at birth 39.2 (2.3) 39.4 (1.7) 39.3 (1.9) 39.4 (1.7) Preterm births Less than 37 gestational weeks 17 (5.9%) 237 (4.3%) 12 (4.2%) 238 (4.4%) Less than 34 gestational weeks 8 (2.8%) 70 (1.3%) 5 (1.8%) 73 (1.3%) Noncephalic presentation 11 (3.9%) 197 (3.6%) 18 (6.3%) a 186 (3.5%) Apgar score 7 or less At 5 min 20 (7.0%) 354 (6.5%) 17 (6.0%) 352 (6.5%) At 10 min 8 (3.6%) 117 (2.7%) 8 (3.6%) 116 (2.7%) At 15 min 5 (2.3%) 49 (1.3%) 6 (2.7%) 48 (1.3%) Perinatal mortality 7 (2.4%) a 43 (0.8%) 6 (2.1%) a 44 (0.8%) Fetal death 4 (1.4%) 26 (0.5%) 4 (1.4%) 26 (0.5%) Neonatal death 3 (1.1%) 19 (0.3%) 2 (0.7%) 20 (0.4%) Malformations 8 (2.8%) 165 (3.0%) 7 (2.5%) 163 (3.0%) Placental weight (g) 654 (146) 642 (131) 641 (128) 643 (133) Relative weight of placenta (%) 18.7 (3.9) a 18.1 (3.3) 18.1 (4.0) 18.1 (3.3) Umbilical cord length (cm) 56.9 (11.1) 56.1 (12.0) 56.4 (11.0) 56.1 (12.0) Cesarean sections 40 (15.1%) 708 (13.8%) 36 (13.5%) 703 (13.9%) Values are mean (SD) or n (%). Perinatal mortality was death during pregnancy or during the first week of life. a P 0.05 when comparing the antibody-positive group with the antibody-negative group.

778 Männistö et al. Thyroid Dysfunction, Antibodies, and Perinatal Outcome J Clin Endocrinol Metab, March 2009, 94(3):772 779 serum samples, and give birth in public hospitals, studies on perinatal outcome are relatively reliable in our country. Perinatal mortality was 2- to 3-fold greater in the infants of thyroid autoantibody-positive mothers compared with those in the antibody-negative groups. This association remained, even after adjusting for maternal age and parity. There were no differences between the groups A-D (hormonal thyroid dysfunction) and the reference group with respect to perinatal mortality. Therefore, thyroid hormone status did not affect the perinatal mortality rate in our study. It is likely that autoimmunity is associated with increased perinatal mortality, not thyroid hormone levels as such. However, there were three mothers with both subclinical hypothyroidism (in group B) and TPO-Ab positivity in the perinatal mortality group. Previously, increased rates of fetal death have been considered to be associated with high TSH levels (2), but 80% of the women with high TSH levels were thyroid antibody positive in that study. It seems that subclinical hypothyroidism and TPO-Ab positivity together may be a risk factor for perinatal mortality, but TPO-antibodies in itself may also have an independent effect. In addition, all the TG- Ab-positive mothers having perinatally deceased newborns were euthyroid. We conclude that the presence of thyroid autoantibodies during pregnancy indicated increased risk for perinatal mortality. This risk may, however, be explained by preterm births. The perinatally deceased infants in both TPO-Ab- and TG-Ab-positive groups were more often born preterm. Regarding preterm deliveries we found no significant differences between the groups. Just a slight tendency toward an increase in preterm deliveries ( 34th gestational week) was seen among TPO-Ab-positive mothers. Stagnaro-Green et al. (9) previously reported a similar finding, but they could not confirm their results statistically due to small sample size. In a study by Negro et al. (24), TPO-Ab positivity was significantly related to preterm delivery, the rate of which decreased with levothyroxine treatment. These findings, as well as those reported by Casey et al. (4), who found an increased rate of preterm delivery among women with subclinical hypothyroidism, suggest that an increase rate of preterm delivery among TPO-Ab-positive women may be related to impaired thyroid function. The mothers with high TSH combined with low ft4 levels also showed a nonsignificant tendency to have an increased rate of preterm delivery, but this was not the case when ft4 levels were normal. Antibodies could possibly explain this because 50% of the mothers with high TSH and low ft4 levels were antibody positive. Because preterm delivery is a relatively rare event and the reasons for it are multifactorial, an association with thyroid dysfunction cannot be confirmed in our study. The role of autoimmunity and thyroid autoantibodies in preterm deliveries needs further research. The TPO-Ab-positive mothers had an increased rate of low birth weight infants ( 2500 g), but after adjusting birth weight by gestational age, no increase in the rate of SGA infants was seen. Therefore, the increased rate of low birth weight infants in the TPO-Ab-positive mothers might be explained by preterm delivery because almost 65% of low birth weight infants in this group were also born preterm. Three of the fifteen low birth weight infants in the TPO-Ab-positive group were born to mothers with subclinical hypothyroidism (group B). Thus, it seems that TPO-Ab positivity in itself associated with increased rates of low birth weight infants, probably through preterm births. Recently in a large study in which normal-weight, overweight, and obese pregnant women were compared (25), it was shown that an overweight condition and obesity are significant risk factors with regard to fetal death, premature delivery, low Apgar scores, and low birth weight. In the studies by Casey et al. (4, 14) and Allan et al. (2), maternal weight was substantially greater than in our study. The mean BMI of 22.3 kg/m 2 in our study reflected the normal range. Thyroid diseases may have an effect on the BMI of the mother, and BMI is possibly an confounding or intervening factor when evaluating the effect of thyroid dysfunction. Therefore, no adjustment for maternal BMI was done. We found the growth of the fetus and the placenta to be affected by thyroid dysfunction and autoimmunity. To our knowledge, the association of thyroid dysfunction or autoimmunity with gestational age-adjusted birth measurements has not been reported earlier. Mothers with low TSH and high thyroid hormone levels (group D, clinical hyperthyroidism) had an almost 3-fold increased risk, and TPO-Ab-positive mothers had a 2-fold increased risk of having LGA infants. Multiparity is a known risk factor for LGA (26), and our results are in accordance with this because the risk for LGA infants in mothers with low TSH and high ft4 disappeared after adjusting the results for parity. Maternal gestational diabetes, which is associated with LGA infants as well, was found in only one case in group D and none in the TPO-Ab-positive group. The effect of clinical hyperthyroidism or TPO antibodies on the rates of LGA infants is unknown. The relationship is probably multifactorial and maternal BMI and placental function may have an effect on the rates of LGA infants. Placental weight was associated with thyroid dysfunction in an U-shaped manner. Greater placental weight was observed in mothers with high TSH and low ft4 levels (group A, clinical hypothyroidism), mothers with low TSH and high ft4 levels (group D, clinical hyperthyroidism), and TPO-Ab-positive mothers. To our knowledge, this is the first study in which the effect of thyroid dysfunction or autoimmunity on placental weight has been reported. Explanatory factors for higher placental weights could be smoking, maternal weight, and birth weight. These factors may have had an effect on our results because TPO-Ab-positive mothers were slightly heavier, mothers with clinical hypothyroidism smoked less often and their infants had higher ponderal indexes, and mothers with clinical hyperthyroidism had a higher prevalence of LGA infants. The role of thyroid hormone or antibody concentrations on placental growth is unknown, but placenta may have an active role on thyroid hormone metabolism through enzyme function and transportation (27). Noncephalic presentation at birth was more common among TG-Ab-positive mothers, although breech presentation alone was not statistically significantly more frequent. The difference in the rates of noncephalic presentations persisted between TG- Ab-positive and -negative women, even after adjusting for maternal age and parity. Mothers who are TG-Ab positive show an

J Clin Endocrinol Metab, March 2009, 94(3):772 779 jcem.endojournals.org 779 increased incidence of subclinical hypothyroidism during pregnancy (1), which may be related to increased rates of breech presentation (12), but no differences in the rates of noncephalic presentation were observed in our study in relation to thyroid hormones. In Finland there is an active policy with regard to external manipulation of the fetus to cephalic presentation, which may have affected the rate of noncephalic presentation in our study. Subclinical hyperthyroidism during pregnancy is probably transient and due to placental hormone stimulation (20). In our study fewer children with low Apgar scores were born to mothers with low TSH levels together with normal ft4 concentrations (group C, subclinical hyperthyroidism) compared with the reference group. The relationship between pregnancy complications and Apgar scores are multifactorial and the effect of subclinical hyperthyroidism on Apgar scores is unclear. Casey et al. (14) also reported that women with subclinical hyperthyroidism did not have any adverse outcomes. In conclusion, thyroid autoimmunity detected during the first trimester of pregnancy seems to be independently associated with increased perinatal mortality, probably through preterm births. Thyroid autoantibodies are also associated with an increased rate of LGA infants. Therefore, it seems that rather than thyroid hormones themselves, thyroid autoantibodies more often have an adverse effect on perinatal outcome. Testing for them might be advisable, at least in risk pregnancies. Acknowledgments We thank Ms. Sarianna Vaara and Ms. Tuula Ylitalo for their valuable work regarding NFBC 1986. We also thank Mr. Jouni Sallinen and Mr. Frank Quinn (Abbott Laboratories) for providing laboratory reagents. Address all correspondence and requests for reprints to: Eila Suvanto-Luukkonen, Department of Obstetrics and Gynecology, University of Oulu, P.O. Box 5000, 90014 Oulu, Finland. E-mail: eila.suvantoluukkonen@ppshp.fi. This work was supported in part by grants from the Alma and K. A. Snellman Foundation (Oulu, Finland); the Jalmari and Rauha Ahokas Foundation (Finland); the Lilly Foundation (Finland); Oulu University Scholarship Foundation (Oulu, Finland); the Finnish Medical Association of Clinical Chemistry; the Foundation of Northern Ostrobothnia Hospital District (Finland); and the Academy of Finland. Disclosure Statement: The authors have nothing to declare. References 1. Glinoer D, Riahi M, Grun JP, Kinthaert J 1994 Risk of subclinical hypothyroidism in pregnant women with asymptomatic autoimmune thyroid disorders. J Clin Endocrinol Metab 79:197 204 2. Allan WC, Haddow JE, Palomaki GE, Williams JR, Mitchell ML, Hermos RJ, Faix JD, Klein RZ 2000 Maternal thyroid deficiency and pregnancy complications: Implications for population screening. J Med Screen 7:127 130 3. Alexander EK, Marqusee E, Lawrence J, Jarolim P, Fischer GA, Larsen PR 2004 Timing and magnitude of increases in levothyroxine requirements during pregnancy in women with hypothyroidism. N Engl J Med 351: 241 249 4. Casey BM, Dashe JS, Wells CE, McIntire DD, Byrd W, Leveno KJ, Cunningham FG 2005 Subclinical hypothyroidism and pregnancy outcomes. Obstet Gynecol 105:239 245 5. Toft A 2004 Increased levothyroxine requirements in pregnancy why, when, and how much? N Engl J Med 351:292 294 6. Spong CY 2005 Maternal thyroid disorders and preterm birth: another piece of the puzzle? Thyroid 15:349 350 7. Hollowell JG, LaFranchi S, Smallridge RC, Spong CY, Haddow JE, Boyle CA 2005 2004 where do we go from here? summary of working group discussions on thyroid function and gestational outcomes. Thyroid 15:72 76 8. Abalovich M, Gutierrez S, Alcaraz G, Maccallini G, Garcia A, Levalle O 2002 Overt and subclinical hypothyroidism complicating pregnancy. Thyroid 12: 63 68 9. Stagnaro-Green A, Chen X, Bogden JD, Davies TF, Scholl TO 2005 The thyroid and pregnancy: a novel risk factor for very preterm delivery. Thyroid 15:351 357 10. Blazer S, Moreh-Waterman Y, Miller-Lotan R, Tamir A, Hochberg Z 2003 Maternal hypothyroidism may affect fetal growth and neonatal thyroid function. Obstet Gynecol 102:232 241 11. Haddow JE, Palomaki GE, Allan WC, Williams JR, Knight GJ, Gagnon J, O Heir CE, Mitchell ML, Hermos RJ, Waisbren SE, Faix JD, Klein RZ 1999 Maternal thyroid deficiency during pregnancy and subsequent neuropsychological development of the child. N Engl J Med 341:549 555 12. Pop VJ, Brouwers EP, Wijnen H, Oei G, Essed GG, Vader HL 2004 Low concentrations of maternal thyroxin during early gestation: a risk factor of breech presentation? BJOG 111:925 930 13. Stagnaro-Green A, Roman SH, Cobin RH, el-harazy E, Alvarez-Marfany M, Davies TF 1990 Sep 19 Detection of at-risk pregnancy by means of highly sensitive assays for thyroid autoantibodies. JAMA 264:1422 1425 14. Casey BM, Dashe JS, Wells CE, McIntire DD, Leveno KJ, Cunningham FG 2006 Subclinical hyperthyroidism and pregnancy outcomes. Obstet Gynecol 107:337 341 15. Matalon S, Sheiner E, Levy A, Mazor M, Wiznitzer A 2006 Relationship of treated maternal hypothyroidism and perinatal outcome. J Reprod Med 51: 59 63 16. Järvelin MR, Hartikainen-Sorri AL, Rantakallio P 1993 Labour induction policy in hospitals of different levels of specialisation. Br J Obstet Gynaecol 100:310 315 17. Järvelin MR, Elliott P, Kleinschmidt I, Martuzzi M, Grundy C, Hartikainen AL, Rantakallio P 1997 Ecological and individual predictors of birthweight in a northern Finland birth cohort 1986. Paediatr Perinat Epidemiol 11:298 312 18. Männistö T, Surcel HM, Bloigu A, Ruokonen A, Hartikainen AL, Järvelin MR, Pouta A, Vääräsmäki M, Suvanto-Luukkonen E 2007 The effect of freezing, thawing, and short- and long-term storage on serum thyrotropin, thyroid hormones, and thyroid autoantibodies: implications for analyzing samples stored in serum banks. Clin Chem 53:1986 1987 19. Stricker R, Echenard M, Eberhart R, Chevailler MC, Perez V, Quinn FA, Stricker R 2007 Evaluation of maternal thyroid function during pregnancy: the importance of using gestational age-specific reference intervals. Eur J Endocrinol 157:509 514 20. Pekonen F, Alfthan H, Stenman UH, Ylikorkala O 1988 Human chorionic gonadotropin (hcg) and thyroid function in early human pregnancy: circadian variation and evidence for intrinsic thyrotropic activity of hcg. J Clin Endocrinol Metab 66:853 856 21. Lamberg BA, Haikonen M, Makela M, Jukkara A, Axelson E, Welin MG 1981 Further decrease in thyroidal uptake and disappearance of endemic goitre in children after 30 years of iodine prophylaxis in the east of Finland. Acta Endocrinol 98:205 209 22. Varo P, Saari E, Paaso A, Koivistoinen P 1982 Iodine in Finnish foods. Int J Vitam Nutr Res 52:80 89 23. Erkkola M, Karppinen M, Jarvinen A, Knip M, Virtanen SM 1998 Folate, vitamin D, and iron intakes are low among pregnant Finnish women. Eur J Clin Nutr 52:742 748 24. Negro R, Formoso G, Mangieri T, Pezzarossa A, Dazzi D, Hassan H 2006 Levothyroxine treatment in euthyroid pregnant women with autoimmune thyroid disease: effects on obstetrical complications. J Clin Endocrinol Metab 91:2587 2591 25. Raatikainen K, Heiskanen N, Heinonen S 2006 Transition from overweight to obesity worsens pregnancy outcome in a BMI-dependent manner. Obesity 14:165 171 26. Juntunen K, Kirkinen P, Kauppila A 1997 The clinical outcome in pregnancies of grand grand multiparous women. Acta Obstet Gynecol Scand 76:755 759 27. James SR, Franklyn JA, Kilby MD 2007 Placental transport of thyroid hormone. Best Pract Res Clin Endocrinol Metab 21:253 264