NEW ORLEANS — Levothyroxine treatment failed to increase the rate of live births among euthyroid women with thyroid peroxidase antibodies (TPO) and had a history of miscarriage or infertility, according to the TABLET trial.
Among nearly 1,000 women given either 50 μg of levothyroxine per day or placebo, pregnancies were achieved at a slightly higher rate in the levothyroxine group than the placebo group (58.3% vs 56.6%), but the live birth rate was not significantly different between groups (37.4% vs 37.9%, relative risk 0.97, 95% CI 0.83-1.14, P=0.74), reported Rima Dhillon-Smith, MBChB, PhD, and Arri Coomarasamy, MBChB, MD, both of the University of Birmingham in England, and colleagues.
Also, the rates of preterm birth and neonatal outcomes were not significantly different between the groups in the Thyroid Antibodies and Levothyroxine (TABLET) trial, they wrote in the New England Journal of Medicine. The study was presented simultaneously at ENDO 2019, The Endocrine Society annual meeting.
“Thyroid peroxidase antibodies are associated with an increased risk of miscarriage and preterm birth, even when thyroid function is normal,” the authors stated. “Small trials indicate that the use of levothyroxine could reduce the incidence of such adverse outcomes.”
The 2017 American Thyroid Association guidelines state there is “insufficient evidence” to determine whether levothyroxine decreases the risk of pregnancy loss in thyroid peroxidase antibody-positive euthyroid women who just became pregnant. However, because of “its potential benefits in comparison with minimal risk,” they recommend administering levothyroxine to these patients with a history of pregnancy loss, typically at a starting dose from 25 to 50 μg.
Terry Davies, MD, of the Icahn School of Medicine at Mount Sinai in New York City, said the current study is an important attempt to confirm previous data that suggested miscarriages experienced by women with thyroid antibodies could be alleviated through thyroid hormone supplementation.
However, “the results do not support such a conclusion,” Davies, who was not involved in the study, told MedPage Today in an email. He noted that the high miscarriage rate in this study — 28.2% for the study arm versus 29.6% for placebo (RR 0.95, 95% CI 0.73-1.23) for miscarriage at <24 weeks -- is suggestive of a selection bias, and that the effect of the small levothyroxine dose on the women's thyroid function was not individually assessed.
For this study, the authors collected data across 49 U.K. hospitals from women, ages 16-40, who had a history of miscarriage or infertility and were trying to conceive in the following 12 months. Patients were excluded if they were receiving treatment for a thyroid disorder, had a cardiac disease, or were receiving amiodarone or lithium.
Euthyroidism was defined by a thyrotropin level from 0.44 to 3.63 mIU/L and a free thyroxine level from 10 to 21 pmol/L; TPO antibody positivity was determined by individual hospital thresholds. Those in the study with normal thyroid function and TPO antibody positivity as determined by these guidelines were then assigned 1:1 to receive 50 μg of levothyroxine or placebo, beginning in preconception and continuing through the end of their pregnancy.
The conception period lasted 12 months during which any changes in thyroid or adverse events were assessed every 3 months. After becoming pregnant, women attended follow-up appointments at 6 to 8 weeks, 16 to 18 weeks, and 28 weeks.
In total, 476 women were assigned to receive levothyroxine and an equal number received placebo. In general, the levothyroxine and placebo groups were similar in terms of age (mean 32.5 vs 32.7 years) and BMI (mean 26.4 vs 26.5), as well as race or ethnicity, with the majority of women being white (68.9% vs 70.8%), followed by South Asian (23.1% vs 19.7%), black (both 3.8%), and Chinese (both 0.8%), the authors reported.
Both groups previously had a mean two prior miscarriages and a similar percentage were nulliparous in both groups (29.6% vs 27.7%), they noted.
At all time points examined, the levothyroxine group had lower serum thyrotropin concentrations and higher free thyroxine (T4) concentrations than the placebo group, which “indicates a biologic effect of levothyroxine,” according to the authors, but they added that “Although women who receive levothyroxine had significantly lower mean thyrotropin levels and higher free T4 concentrations than those who received placebo, this did not have a clinically relevant effect on the rates of thyroid dysfunction or clinical outcomes.”
The number of serious adverse events was slightly higher in the levothyroxine group compared with placebo (6% vs 4%), but this was not significantly different (P=0.14), they noted.
The authors said a primary study limitation was the dosing of levothyroxine (50 μg once daily), which may “need to be adjusted depending on body weight, thyroid peroxidase antibody level, or thyrotropin concentration.” The data was also limited because TPO antibody testing varied across the hospitals included in this study, they added.
But they noted that the TABLET trial was larger than earlier studies that were “inadequately powered or restricted to a single population. Our inclusion of multiple centers and multiple clinical populations improved the generalizability of the findings.”
“As our study was large and of high quality, we can now be confident that Levothyroxine does not improve pregnancy success for women with thyroid antibodies and normal thyroid function and therefore should not be recommended or used in clinical practice,” Dhillon-Smith said in a press statement. “This will mean no longer providing unnecessary medication to women who do not need it.”
TABLET was supported by the National Institute for Health Research efficacy and mechanism evaluation program.
Generic levothyroxine and placebo were supplied by Sharp Healthcare.
Rima Dhillon-Smith and Coomarasamy disclosed no relevant relationships with industry.