Which Basic Patient Characteristics Are Predictors of Thyroid Dysfunction during Pregnancy?

Presentation Number: SAT-0547
Date of Presentation: June 21st, 2014

Tim IM Korevaar*1, Marco Medici1, Akhgar Ghassabian2, Edward Visser3, Henning Tiemeier1, Theo J Visser4 and Robin P. Peeters5
1Erasmus Medical Center, Rotterdam, 2Erasmus University Medical Center, 3Erasmus Medical Ctr, Rotterdam, Netherlands, 4Erasmus Univ Med Ctr, Rotterdam, Netherlands, 5Erasmus Univ Rotterdam, Rotterdam, Netherlands


Context: Elevated TSH, hypothyroxinemia and TPO-antibody (TPOAb) positivity during early pregnancy are associated with various adverse outcomes such as decreased neurocognitive development of the child. Surprisingly, thyroid dysfunction during pregnancy is not associated with classical symptoms and most women are even asymptomatic. Current guidelines advocate aggressive case finding in high risk cases, but all known risk factors (RFs) lack sensitivity which makes clinical identification of high-risk women very difficult. We therefore investigated whether easily obtainable characteristics are associated with the risk of maternal thyroid dysfunction in pregnancy.
Material and methods: Serum TSH, FT4 and TPOAb levels were determined during early pregnancy in 6278 pregnant women from the Generation R study. Multivariate analyses with backward selection was used to identify RFs for abnormal thyroid function and included maternal age, parity, BMI, smoking, alcohol use, education, child gender and medical/obstetrical history. Analyses of neurocognitive outcomes were additionally adjusted for maternal education, gender, alcohol use, gestational age at birth, birth weight and age at assessment.
Results: RFs for an elevated TSH were BMI >25, non-smoking, ethnicity and parity <2 (odds ratios (OR) range 1.41-2.71). None of the women with ≤1 RF had an elevated TSH (0/105) compared to 6.1% of women that had all RFs (72/1174, p<0.0001)
RFs for hypothyroxinemia were age >31, family history of diabetes, (former) smoking, parity >2, non-Dutch ethnicity, and BMI >25 (OR range 1.54-3.35). Amongst women without RFs, hypothyroxinemia was seen in 1% (6/610), whereas the prevalence increased to 14.4% when ≥4 RFs were present (24/167; OR 24.9, p<0.0001).
RFs for TPOAb positivity were family history of thyroid disease, ethnicity and parity <2 (OR range 1.77-3.99). TPOAb positivity was seen amongst 1.2% (2/174) of women without RF compared to 9.8% amongst women with ≥2 RFs (77/782; OR 8.2, p<0.0001). The number of RFs which were identified for hypothyroxinemia were linearly associated with child IQ scores (β -1.51 ±0.26; p<0.0001) and autistic symptoms (0.005 ±0.002; p=0.006) at 5 years of age, these effects decreased considerably after correction for hypothyroxinemia.
Conclusions: This is the first multivariate identification of RFs for maternal thyroid dysfunction during pregnancy. According to these RFs, women with an increased risk of thyroid dysfunction can be identified in a very simple and sensitive manner. Furthermore, these risk factors were associated with decreased child IQ and increased risk of autistic symptoms.


Nothing to Disclose: TIK, MM, AG, EV, HT, TJV, RPP