STUDY QUESTION: Are rare genetic variants in the conceptus associated with chromosomally normal pregnancy loss?
SUMMARY ANSWER: The proportion of probands with at least one rare variant is increased in chromosomally normal loss conceptuses compared with controls (odds ratio 2.0, 95% CI 0.9, 3.0).
WHAT IS ALREADY KNOWN: Among non-consanguineous families, one study of seven chromosomally normal losses to four couples with recurrent pregnancy loss (RPL) and a case report of a family with RPL of which one was known to be chromosomally normal identify compound heterozygote variants in three different genes as possibly causal. Among consanguineous families, RPL of chromosomally normal pregnancies with non-immune hydrops fetalis (NIHF) has been attributed to recessive variants in genes previously implicated for NIHF and new candidate genes.
STUDY DESIGN, SIZE, DURATION: The starting sample was 52 chromosomally normal losses to 50 women, identified in 2003-2005 as part of a cohort study on trisomy and ovarian aging. The analytic sample comprises 19 conceptus-parent trios with DNA from 17 biologic parents (cases). The control group derives from the National Institutes of Mental Health's National Database for Autism Research (NDAR). It comprises 547 trios of unaffected siblings of autism cases and their parents.
PARTICIPANTS/MATERIALS, SETTING, METHODS: We use exome sequencing to identify rare variants in the coding region of the genome. We defined variant rarity in two ways: ultra-rare (absent in gnomAD) and rare (heterozygote <10-3 in gnomAD) and, for autosomal recessive variants, absence of a homozygote in genomAD. We compare the rates of rare damaging variants (loss of function and missense - damaging) and the proportions of probands with at least one such variant in cases versus controls. Secondarily, 1) we repeat the analysis limiting it to variants in genes considered causal in fetal anomalies and 2) we compare the proportions of cases and controls with damaging variants in genes which we classified as possibly embryonic lethal based on a review which was blinded to case-control status.
MAIN RESULTS AND THE ROLE OF CHANCE: The rates of ultra-rare damaging variants (all de novo) are 0.21 and 0.17 in case and control probands, respectively. The corresponding rates for rare potentially pathogenic de novo variants are 0.37 and 0.24, respectively; for autosomal recessive variants they 0.11 and 0.03. The proportions of probands with at least one rare potentially damaging variant were 36.8% among cases and 22.9% among controls (odds ratio (OR) = 2.0, 95% CI 0.9, 3.0). Secondary analyses show no damaging variants in fetal anomaly genes among case probands; the proportion with variants in possible embryonic lethal genes was increased in case probands (OR=12.3, 95% CI 3.4, 61.1). Our data suggest that damaging variants in genes BAZ1A, FBN2 and TIMP2 may be lethal to the embryo.
LIMITATIONS, REASONS FOR CAUTION: The number of case trios (n=19) limits the precision of our point estimates. We observe a moderate association between rare damaging variants and chromosomally normal loss with a confidence interval that includes unity. A larger sample is needed to estimate the magnitude of the association with precision and to identify the relevant biological pathways.
WIDER IMPLICATIONS OF THE FINDINGS: Our data add to a very small literature on this topic. They suggest rare genetic variants in the conceptus may be a cause of chromosomally normal loss.