We thank Diamond et al. for sharing their data from AMIGOS. The authors suggest that the fact that their study included a larger sample size than ours minimizes our findings. On the contrary, their study design, which included only women with unexplained infertility (UI) separated into three groups based on subtle differences in male semen parameters (ranging from normal to mild male factor infertility), significantly limited Diamond et al.’s ability to detect an association between TSH and UI. This is in sharp contrast to our study, which compares two distinct groups of women: one with UI and a second group with a known cause of infertility—male partners with azoospermia/severe oligospermia (sperm count <1 million/mL). Our two distinct groups allowed us to use a small sample size to demonstrate an association between TSH and UI. To detect an association between TSH and UI in a population inclusive only of women with UI would require an extremely large sample size; therefore, it is not surprising that Diamond et al. did not detect an association, even with a sample size of 900.
A second key difference between our study and AMIGOS is our definition of male factor infertility. AMIGOS included couples with a motile sperm count of ≥5 million (1). Although this parameter is predictive of success with intrauterine insemination, this threshold allows for the inclusion of some couples who meet the definition of male factor infertility based on World Health Organization criteria (2). We specifically chose to exclude males with milder forms of male factor infertility from our control group based on data suggesting that in couples with milder forms of male factor infertility, female factors also contribute to the diagnosis of infertility (3, 4). In fact, a study in in vitro fertilization patients demonstrated higher TSH levels in women of couples diagnosed with male factor vs other types of infertility (5). Therefore, a second reason we are not surprised by Diamond et al.’s finding of similar TSH levels in three groups of women with UI and varying degrees of mild male factor infertility is that including couples with mild male factor infertility in the comparator groups likely biased their results toward the null hypothesis.
Third, although our study included multiple TSH assays, we demonstrate that similar percentages of individuals from each group had TSH levels measured using the various assays. Furthermore, when the percentages differed, a sensitivity analysis demonstrated that our findings remained significant. Therefore, we are confident in the fact that the use of multiple assays did not bias our results.
Lastly, unlike AMIGOS, we did not evaluate outcomes. The referenced post hoc analysis includes women with polycystic ovary syndrome and therefore cannot be interpreted in the context of our study (6), but we are intrigued by the data presented in the letter. Although the authors did not find a significant difference in conception between women with TSH <2.5 mIU/L vs ≥2.5 mIU/L, the results of the unadjusted and adjusted analyses are in the direction we would predict based on our findings (lower odds of conception with TSH ≥2.5 mIU/L). Therefore, it is possible that their results would have been significant if AMIGOS had been designed to evaluate for this outcome; for example, with a larger sample size of 3537 patients, the unadjusted 39.6% conception rate in women with TSH <2.5 mIU/L vs 34.1% in TSH ≥2.5 mIU/L would have achieved a P value of <0.05 with 80% power using a χ2 test.
Acknowledgments
Disclosure Summary: The authors have nothing to disclose.
Glossary
Abbreviation:
- UI
unexplained infertility.
References
- 1. Diamond MP, Mitwally M, Casper R, Ager J, Legro RS, Brzyski R, Casson P, Eisenberg E, Zhang H; NICHD Cooperative Reproductive Medicine Network. Estimating rates of multiple gestation pregnancies: sample size calculation from the assessment of multiple intrauterine gestations from ovarian stimulation (AMIGOS) trial. Contemp Clin Trials. 2011;32(6):902–908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Dickey RP, Pyrzak R, Lu PY, Taylor SN, Rye PH. Comparison of the sperm quality necessary for successful intrauterine insemination with World Health Organization threshold values for normal sperm. Fertil Steril. 1999;71(4):684–689. [DOI] [PubMed] [Google Scholar]
- 3. Emperaire JC, Gauzere E, Audebert A. Female fertility and donor insemination. Lancet. 1980;1(8183):1423–1424. [DOI] [PubMed] [Google Scholar]
- 4. Albrecht BH, Cramer D, Schiff I. Factors influencing the success of artificial insemination. Fertil Steril. 1982;37(6):792–797. [DOI] [PubMed] [Google Scholar]
- 5. Cramer DW, Sluss PM, Powers RD, McShane P, Ginsburg ES, Hornstein MD, Vitonis AF, Barbieri RL. Serum prolactin and TSH in an in vitro fertilization population: is there a link between fertilization and thyroid function? J Assist Reprod Genet. 2003;20(6):210–215. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Seungdamrong A, Steiner AZ, Gracia CR, Legro RS, Diamond MP, Coutifaris C, Schlaff WD, Casson P, Christman GM, Robinson RD, Huang H, Alvero R, Hansen KR, Jin S, Eisenberg E, Zhang H, Santoro N; Eunice Kennedy Shriver National Institute of Child Health and Human Development Reproductive Medicine Network. Preconceptional antithyroid peroxidase antibodies, but not thyroid-stimulating hormone, are associated with decreased live birth rates in infertile women. Fertil Steril. 2017;108(5):843–850. [DOI] [PMC free article] [PubMed] [Google Scholar]