Miscarriage and celiac disease: what studies report and how strong the evidence is

Short answer

Pooled studies report a higher miscarriage rate in women with celiac disease than in women without it1, 2. One review pooled observational studies5 and the other case-control and cohort studies9; the NIH says observational studies can’t prove cause and effect39. In one Danish cohort the excess appeared before diagnosis, not after14, 15. ESHRE’s 2022 guideline says testing for celiac blood markers is not indicated in recurrent pregnancy loss without celiac symptoms4.

Published by Gluten Verdict · updated · 9 sources, each linked below

Reviewed by Mark Peter, Registered dietitian (RD), on

Armchair at dusk with a cardigan, a mug on the sill and a saffron cushion

Key points

  • Two meta-analyses report a relative risk of miscarriage of 1.391 and 1.352 in women with celiac disease; Tersigni’s rests on case-control and cohort studies9, and Arvanitakis’s on observational studies5.
  • The abstract of a third meta-analysis, of obstetric complications, does not list miscarriage among its outcomes12.
  • In the Danish register study, spontaneous abortion was no more frequent after diagnosis, but 11 extra per 1000 pregnancies before it15, 14.
  • About half of early miscarriages happen when the embryo does not develop properly24; ACOG’s repeated-miscarriage page says about half of miscarriages occur randomly when an embryo receives an abnormal number of chromosomes25.
  • Published sources differ on testing: the 2014 meta-analysis authors write that physicians should investigate33; ESHRE says marker testing is not indicated without symptoms4.

What the pooled studies report

This guide covers four studies. Three are pooled analyses of earlier studies; the fourth is a single nationwide matched cohort study from Denmark13. The table gives what each says about miscarriage, and what it does not.

Four studies on celiac disease and pregnancy loss
StudyDesignWhat it reports on miscarriage
Tersigni and colleagues, Human Reproduction Update, 2014Meta-analysis of case-control and cohort studies found by searching Medline and Embase8, 9In women with celiac disease, a relative risk (RR) of miscarriage of 1.39 (95% confidence interval, CI, 1.15-1.67)1
Arvanitakis and colleagues, Annals of Gastroenterology, 2023Systematic review and meta-analysis of 14 cohort and 4 case-control studies10Spontaneous abortion: RR 1.35 (95% CI 1.10-1.65)2
Saccone and colleagues, American Journal of Obstetrics and Gynecology, 2016Meta-analysis of 10 cohort studies covering 4,844,555 women11Miscarriage is not among its listed outcomes12
Grode and colleagues, Human Reproduction, 2018One nationwide Danish matched cohort: 6319 women with celiac disease and 63166 comparison women13Before diagnosis, 11 extra spontaneous abortions per 1000 pregnancies (adjusted odds ratio 1.12, 95% CI 1.03-1.22)14

The Saccone review’s abstract reports preterm birth, growth restriction, stillbirth, preeclampsia, small-for-gestational-age and low birthweight, not miscarriage12.

Two different questions, two different numbers

The pooled figures answer two separate questions, and they are easy to mix up.

  • How much more often is celiac disease found in women who have had recurrent miscarriage than in the general population? Tersigni’s pooled odds ratio for celiac disease was 5.82 (95% CI 2.30-14.74) in women experiencing recurrent miscarriage16.
  • Is miscarriage more common among women who have celiac disease? The same review reports a relative risk of 1.391.

The same abstract adds a result to read beside both: no increased risk of recurrent miscarriage was found in celiac patients17.

“Recurrent” has a specific meaning. ESHRE defines recurrent pregnancy loss as the loss of two or more pregnancies18, and ACOG as two or more miscarriages19. A single miscarriage falls outside both definitions.

Why these studies cannot show cause

The NIH says observational studies, in which a population’s health or behaviors are observed and analyzed, can’t prove cause and effect but can be useful for finding associations39. The Annals of Gastroenterology review searched for observational studies5 and words its conclusion as an association37; the Danish study is a nationwide matched cohort13. The caveats the Danish authors themselves give are below.

  • The Danish authors adjusted for several confounding factors, and write that residual confounding cannot be ruled out6.
  • They also note that not all spontaneous abortions will come to attention and be registered7, so some spontaneous abortions are not in the count.
  • The biological routes described in the Tersigni abstract come from laboratory work. Tersigni and colleagues write that in vitro studies have provided two main models of placental damage20, and that anti-transglutaminase antibodies may be detrimental to endometrial angiogenesis (blood-vessel growth in the lining of the uterus), as shown in cultured human cells and in a murine (mouse) model21.

A pooled ratio such as 1.39 describes a difference between groups of women. This guide does not read it as the reason for any one loss.

Before and after diagnosis: what the data show

The Danish register study split women’s pregnancies by timing. Comparing women with diagnosed celiac disease with women without it, the chance of pregnancy, live birth and risk of stillbirth, molar and ectopic pregnancy, spontaneous abortion and abortion due to foetal disease was the same15. Before diagnosis, the women with celiac disease had 11 extra spontaneous abortions per 1000 pregnancies14.

The authors’ own summary: prior to the diagnosis an increased risk of adverse pregnancy outcomes was seen, whereas after the diagnosis no influence on reproductive outcomes was found3.

The 2023 pooled analysis also looked at timing, but its abstract states the split for other outcomes: only undiagnosed celiac disease increased the risk for fetal growth restriction, stillbirth, preterm delivery and lower mean birthweight, whereas early diagnosis was not linked to any adverse pregnancy outcomes22. That sentence does not name spontaneous abortion, so the abstract alone does not show how its miscarriage result divides by diagnosis.

These are comparisons between groups of women, and the authors state that residual confounding cannot be ruled out6.

What ACOG says about why miscarriages happen

ACOG’s patient information puts miscarriage in context. Early pregnancy loss happens in about 10 of 100 known pregnancies23. About half of early miscarriages happen when the embryo does not develop properly24. On its repeated-miscarriage page ACOG says about half of miscarriages occur randomly when an embryo receives an abnormal number of chromosomes25. ACOG also says that in almost every case, miscarriage is not a woman’s fault26.

Repeated loss is less common: ACOG estimates that fewer than 5 in 100 women have two miscarriages in a row27. In more than half of women with repeated miscarriages, no cause can be found28. ACOG adds that about 65 in 100 women with unexplained recurrent pregnancy loss have a successful next pregnancy30.

ACOG’s list of medical conditions that may raise the risk of repeated miscarriage does not name celiac disease29. That page is a short patient FAQ, so its silence shows what ACOG chose to list there. It is not a finding that no link exists.

What published guidance says about testing after pregnancy loss

ACOG’s FAQ says that after two miscarriages, a thorough physical exam and testing are recommended19. The page does not name celiac disease29.

ESHRE’s 2022 guideline on recurrent pregnancy loss addresses celiac disease directly. It describes a case-control study of 116 women with unexplained recurrent loss and 116 age-matched controls31. Although the women with recurrent loss had significantly higher serum levels of IgG tTG antibodies than the controls, the proportion of women with antibodies indicative of celiac disease was very low and similar in both groups32. Its conclusion: testing for celiac disease serum markers is not indicated in women with recurrent pregnancy loss in absence of symptoms of celiac disease4.

Authors of the pooled analyses lean the other way, or leave the question open. The 2014 Tersigni review concludes that physicians should investigate women with unexplained infertility, recurrent miscarriage or growth restriction for undiagnosed celiac disease33. The 2016 Saccone authors write that further studies are required to determine whether women with adverse pregnancy outcomes should be screened34.

So sources differ: a 2014 meta-analysis says physicians should investigate33, and ESHRE’s 2022 guideline says marker testing is not indicated without symptoms4. Whether testing applies in a particular case is a conversation for the clinician managing it.

Gluten-free diet and later pregnancies: what is and is not known

What the sources report:

  • Tersigni and colleagues conclude that the risk is significantly reduced by a gluten-free diet35. Their results sentence makes the treated-versus-untreated comparison for growth restriction, low birthweight and preterm delivery, not for miscarriage36.
  • Arvanitakis and colleagues write that early diagnosis and appropriate management with a gluten-free diet may ameliorate the associations37.
  • In the Danish study, spontaneous abortion was the same in diagnosed women as in women without celiac disease15.
  • Coeliac UK states that there is an increased risk of reproductive problems in women with untreated coeliac disease38.

What is not known from these sources: none of the four abstracts we read describes a trial in which women were assigned to a gluten-free diet or to no change in diet5, 9, 11, 13. The studies compare women as they are.

Questions people ask

Does celiac disease cause miscarriage?

Two meta-analyses report a higher rate of miscarriage in women with celiac disease: a relative risk of 1.391 and of 1.352. The reviews pooled case-control and cohort studies9, 10, and Arvanitakis’s searched for observational studies5; the NIH says observational studies can’t prove cause and effect39, and the Danish authors say residual confounding cannot be ruled out6.

Is undiagnosed celiac disease linked to recurrent miscarriage?

One pooled analysis found an odds ratio of 5.82 for celiac disease in women experiencing recurrent miscarriage16, and also reports that no increased risk of recurrent miscarriage was found in celiac patients17. ESHRE’s 2022 guideline says marker testing is not indicated without celiac symptoms4.

Should I be tested for celiac disease after a miscarriage?

That is a decision for a clinician, and published sources differ: one meta-analysis says physicians should investigate women with recurrent miscarriage33, while ESHRE says testing is not indicated without symptoms4.

Could I have caused my miscarriage?

ACOG says that in almost every case, miscarriage is not a woman’s fault26, and that about half of early miscarriages happen when the embryo does not develop properly24.

More in Women, pregnancy and babies

Sources