Celiac disease is not more prevalent in patients undergoing IVF and does not impact reproductive outcomes with or without treatment: A large, prospective cohort study

The prevalence of seropositive celiac disease is consistent with that of the general population, and the patients had IVF outcomes equivalent to their age- controlled peers.

Volume 110, Issue 3, Pages 437–442


Caroline R. Juneau, M.D., Jason M. Franasiak, M.D., H.C.L.D.'Correspondence information about the author M.D., H.C.L.D. Jason M. FranasiakEmail the author M.D., H.C.L.D. Jason M. Franasiak, Linnea R. Goodman, M.D., Diego Marin, M.S., Katherine Scott, B.S., Scott J. Morin, M.D., Shelby A. Neal, M.D., Jeffrey E. Juneau, M.D., Richard T. Scott, M.D., H.C.L.D.



To study the prevalence of celiac disease in the infertile population undergoing in vitro fertilization (IVF) and assess outcomes.


Prospective cohort study.


A single infertility center from January 2016 to March 2017.


Women 18–45 years of age participating in IVF.


Patients had serum tissue transglutaminase (tTG) and endomysial (EMA) IgA testing to screen for celiac disease and completed a 10-question “yes or no” survey to assess their medical history, previous testing, dietary habits, and pertinent symptoms.

Main Outcome Measure(s)

IVF cycle outcomes were compared between seronegative and seropositive patients.


Of 1,000 patients enrolled, 995 completed serologic screening and 968 underwent oocyte retrieval. Eighteen patients screened positive for both tTG and EMA (1.8%) and 10 additional patients (1.0%) screened positive for one of the two antibodies. The number of mature oocytes retrieved, fertilization rates, and blastulation rates were equivalent between seronegative and seropositive patients. There were 987 patients who completed the questionnaire (98.7%), and 84 reported being gluten free (8.5%). Those who reported being gluten free were no more likely to be antibody positive than the general population. Furthermore, a low-gluten diet was not associated with markers of ovarian reserve, oocytes retrieved, fertilization, blastulation, sustained implantation and pregnancy loss rates.


The prevalence of seropositive celiac disease was consistent with that of the general population (2.8%). Patients who were seropositive for celiac disease–related antibodies had outcomes equivalent to seronegative patients, and patients with a gluten-free diet did not have improved outcomes.

Read the full text here.


Go to the profile of Concepción Núñez
over 3 years ago

Juneau et al studied an interesting topic such as the relationship between celiac disease (CD) and infertility and tried to solve several questions still under debate. It is worth noting the high number of patients recruited and the prospective design of the study. Recent popularity of the gluten-free diet (GFD) gives high relevance to evidence that instauration of a GFD does not improve reproductive outcomes in women undergoing in vitro fertilization. However, we consider that the interpretation of their data in relation to CD needs to be taken cautiously.

Their study involves the highest sample size related to CD prevalence and infertility. They observed 2.8% of seroprevalence considering anti-TG2 and/or EMA and claimed that this value was consistent with the observed in the general population. The authors used for comparison the prevalence of 2.3% reported by Singh et al in a recent meta-analysis (Celiac disease in women with infertility: a meta-analysis. J Clin Gastroenterol 2016;50:33–9). However, 2.3% was the frequency of CD that Singh et al found in women with infertility due to any cause, which indeed resulted in 3.5 increased odds of having CD compared to control population. In a very recent meta-analysis of the same authors (Singh et al. Global Prevalence of Celiac Disease: Systematic Review and Meta-analysis. Clin Gastroenterol Hepatol. 2018 Jun;16(6):823-836), they found 1.4% of pooled global seroprevalence considering anti-TG2 or EMA. Age at onset was not considered for seroprevalence, but when analyzing biopsy-proven CD they observed almost half of prevalence in adults (0.5% vs 0.9%).

A second important conclusion was that CD did not affect reproductive outcomes with or without treatment. The authors compared reproductive outcomes between seropositive and seronegative women to conclude that those parameters were not affected by the disease. However, it must be considered that the poor reproductive outcomes of the seropositive women could be in fact caused by CD. To our knowledge, the debate exists about whether reproductive failures in women can be due to undiagnosed CD, but we are not sure that among women with infertility the worst outcomes should be expected in those with CD. Finally, the effect of the GFD on fertility in CD women was not evaluated, since only 3 out of the 84 patients following a GFD, and therefore used to evaluate the effect of the diet, were seropositive. We consider that the title does not reflect the content of the paper.

It is also remarkable that the authors indicate that intestinal biopsy is the standard criterion for diagnosis, but invasive and extremely impractical for routine clinical use in assisted reproduction patients. Biopsy is essential for the diagnosis of CD in adults and we consider that clinicians should encourage patients to undergo an intestinal biopsy in order to obtain a certain diagnosis. In addition, following that criterion, the paper only refers to seroprevalence. However, as can be seen in Table 2, which shows the results of the questionnaire completed by the 98.7% of the participants, 20 patients underwent a biopsy to confirm diagnosis. It would be interesting to know the results of those biopsies to have a value of biopsy-proven CD.

In addition to specific serology and biopsy, CD diagnosis can be supported by other tools: genetic test and response to the GFD. Most of the seropositive women have not started a GFD, therefore the authors cannot test this response. However, the genetic test could be performed. Although genetics is not valid to confirm CD, it shows a high negative predictive value and could identify false positive serological results.

Therefore, we are not sure that the presented data shed light on such a controversial subject, but they are not enough to claim that CD screening is not indicated for the infertile population.

Concepción Núñez, PhD. Instituto de Investigación Sanitaria del Hospital Clínico San Carlos (IdISSC), Madrid, Spain

Natalia López Palacios, MD, PhD. Hospital Clínico San Carlos, Instituto de Investigación Sanitaria del Hospital Clínico San Carlos (IdISSC), Madrid, Spain