Abstract
In the June 2012 issue of the Thyroid, Scoccia et al. (1) aimed to investigate whether pregnancy rates after in vitro fertilization (IVF) procedure differed in women with hypothyroidism treated with levothyroxine (LT4) as compared to euthyroid women. The authors found that women with hypothyroidism, despite treatment with LT4, had significantly decreased pregnancy (14.3 vs. 46.6%, p=0.004) and live birth rates (14.3 vs. 37.3%, p=0.035) following IVF as compared to euthyroid women.
These results seem surprising, as there is grade B level recommendation by the United States Preventive Service Task Force (USPSTF) that, in women with subclinical hypothyroidism, treatment with LT4 improves the obstetrical outcome (2). Nevertheless, in our opinion, the conclusions of the current study (1) have to be interpreted with caution due to several reasons.
The sample sizes are relatively small. The pregnancy and live birth rates of 14.3% in the group of women with hypothyroidism mean than only three women achieved clinical pregnancy and delivered a healthy baby. This is certainly a small number to allow deduction of safe conclusions.
Concerning the treatment of hypothyroidism, the authors claim “Patients included in our study undergoing IVF treatment had a mean baseline thyrotropin (TSH) level of <2.5 μU/mL.” Nevertheless, the mean±standard deviation of the serum TSH concentration in women with hypothyroidism treated with LT4 was 2.5±1.3 μU/mL with a range of 0.53–4.0 μU/mL, suggesting that many women had TSH >2.5 μU/mL. According to the Endocrine Society guidelines (2), this is not adequate. If hypothyroidism is diagnosed before pregnancy, adjustment of the preconception LT4 dose is necessary to reach a TSH concentration not higher than 2.5 μU/mL (USPSTF recommendation level, I; evidence, poor; grade, 1). Of course, the threshold of 2.5 μU/mL had not been set in 2003, the year the first women were recruited for the study. However, by applying today's criteria, the group of women with hypothyroidism, as a whole, was undertreated.
Despite the attempt to treat the women with hypothyroidism by administering LT4, these women entered the IVF procedure having higher serum TSH concentrations as compared to euthyroid women (2.5±1.3 vs. 1.8±0.9, p=0.021). Thus, the women with hypothyroidism were undertreated, not only by applying the threshold of 2.5 μU/mL (see above), but compared to a group of euthyroid women. In any case, the observed differences in reproductive outcomes could be attributed to the higher TSH concentrations.
There were no available data on thyroid autoimmunity (TAI), as it is clearly stated in the discussion. Nevertheless, the authors claim “Similar pregnancy rates in IVF have been reported for euthyroid women with or without thyroid antibodies, while HYPO-Rx women with thyroid antibodies were observed to have a lower clinical pregnancy rate than euthyroid women with thyroid antibodies (1).” Our team, in a meta-analysis, has shown that presence of TAI is associated with an increased risk for spontaneous miscarriage in subfertile women achieving a pregnancy through an IVF procedure (fixed effects, relative risk 1.99 [95% confidence interval (CI) 1.42–2.79], p<0.001) (3). In addition, in a systematic review, we have shown that serum TSH concentrations may be increased during or within one month after IVF, exceeding the threshold of 2.5 mU/mL (4). In a similar way, a large meta-analysis (5) demonstrated that the presence of TAI was associated with an increased risk of unexplained subfertility (odds ratio [OR] 1.5, [CI 1.1–2.0]), miscarriage (OR 3.73 [CI 1.8–7.6]), recurrent miscarriage (OR 2.3 [CI 1.5–3.5]), preterm birth (OR 1.9 [CI 1.1–3.5]), and maternal postpartum thyroiditis (OR 11.5 [CI 5.6–24]) when compared with the absence of TAI. Therefore, presence of TAI seems to be a crucial parameter to interpret the reproductive outcome.
In this retrospective study, it was not possible to include a control group of women with hypothyroidism not being on LT4 treatment. In a similar study though (6), significant differences in reproductive outcomes, after an IVF procedure, were demonstrated when subfertile women with subclinical hypothyroidism on LT4 (n=35) were compared to subfertile women with subclinical hypothyroidism on placebo. LT4 therapy of subclinical hypothyroidism did not affect the miscarriage rate (p=0.44), but did affect the pregnancy rate (p=0.0001).
Additional methodological characteristics of the study that have to be taken under consideration are (i) the retrospective nature of the study, which adds selection, information, and recall biases, (ii) the age of the study, as the women were recruited between January 2003 and December 2007, (iii) the use of different protocols for controlled ovarian hyperstimulation (COH) (e.g., use of gonadotropin releasing hormone agonist or antagonist protocols, (iv) use or not of human menopausal gonadotropin), which adds additional confounding parameters that may have an impact on reproductive outcomes (pregnancy rate, live birth rate), and (v) the threshold of maternal age less than 37 years as an inclusion criterion, the impact of which on the main outcome (i.e., live birth rate) is difficult to estimate.
We believe that the study of Scoccia et al. (1) is a valuable addition to the literature, as it suggests that even small increases in maternal TSH concentration just before a pregnancy may have an unfavorable impact on reproductive outcome. We fully agree with the authors that larger prospective studies are necessary to assess confounding variables and determine optimal TSH concentrations before and during COH for IVF.