Endocrine Abstracts

Objective: Thyroid peroxidase antibody (TPOAb) positivity is the most important risk factor for overt/subclinical hypothyroidism and partly guides levothyroxine treatment indications during pregnancy. TPOAb positivity is usually defined by manufacturer cutoffs which are typically derived in non-pregnant populations with concentrations ≥95th percentile. Besides the general lack of assay uniformity, current TPOAb positivity cutoffs are likely to underestimate gestational thyroid autoimmunity due to immune tolerance of pregnancy. To improve the clinical interpretation of TPOAb measurement and cutoffs during pregnancy, we investigated the association of TPOAb concentrations and commonly used thresholds with maternal thyroid function during pregnancy.

Methods: This study was embedded in the Consortium on Thyroid and Pregnancy. Participants with multiple gestation, pre-existing thyroid diseases, thyroid (interfering) medication usage, or pregnancy by IVF/ICSI were excluded. We assessed the nonlinear association of TPOAb percentiles with SD-scores of thyroid stimulating hormone (TSH) and free thyroxine (FT4), which indicated an effect threshold between the 80th-85th percentile. With ≤80th percentile as the reference group, we assessed the association of each consequent TPOAb percentile with maternal thyroid function, the risk of TSH 2.5-4.0 mU/l, and the risk of TSH >4.0 mU/l.

Results: The final study population comprised individual-participant data of 69,713 pregnant women from 24 cohorts. For TPOAb concentrations from the 89th percentile and upwards there was a higher TSH (effect estimates, range +0.13 SD to +1.05 SD), and from the 91st percentile and upwards there was a lower FT4 (effect estimates, range -0.08 SD to -0.49 SD). Pregnant women with TPOAb concentrations ≥89th percentile had a statistically significant higher risk of TSH 2.5-4.0 mU/l (range ORs: 1.40 for the 89th percentile to 5.96 for the 100th percentile) and a higher risk of TSH >4.0 mU/l (range ORs: 1.92 for the 89th percentile to 38.46 for the 100th percentile). The associations of TPOAb percentiles with TSH and FT4 SD-scores were stronger during early pregnancy than during late pregnancy, fitting with patterns of hCG stimulation and immune tolerance (P for interaction <0.001 for TSH and <0.001 for FT4).

Conclusion: We show that TPOAb concentrations below the manufacturer cutoffs may already reflect clinically relevant thyroid autoimmunity during pregnancy. High normal TPOAb concentrations could be interpreted differently in pregnant women, especially for those in whom the TPOAb status could (partly) define a treatment indication.