New insights into mechanisms behind miscarriage

One in five known pregnancies ends in miscarriage. Sit with that number for a moment. It makes miscarriage the most common complication of early pregnancy — not rare, not unusual, just routinely devastating. And yet, for decades, the biological explanation was almost reassuringly simple: most embryos are abnormal, the body recognizes this, and pregnancy fails. That explanation is true, as far as it goes. The problem is how far it goes. Larsen and colleagues, reviewing the latest evidence on miscarriage mechanisms, show that the old story is incomplete — and that what's replacing it changes where we look for treatment entirely. Start with the vocabulary, because the field runs on precise distinctions. A biochemical loss happens before ultrasound confirms anything — a positive pregnancy test, then nothing, usually before six weeks. A clinical miscarriage is confirmed by imaging or histology. Clinical losses split further into early, before twelve weeks, and late, between twelve and twenty-one weeks. Recurrent miscarriage, or RM, is defined by European guidelines as three or more consecutive losses before twenty-two weeks. That threshold matters: RM isn't just sporadic loss that happened to repeat. It's considered a distinct disease entity with its own biology. The scale of sporadic loss is staggering when you map it carefully. Prospective studies using daily urinary hormone tests show only about one-third of all conceptions reach live birth. Roughly thirty percent are lost before implantation, another thirty percent after implantation but before a missed period, and about fifteen percent of conceptions result in a recognized early clinical loss. That rate climbs sharply with age — around ten percent at ages twenty to twenty-four, and fifty-one percent by age forty to forty-four. Recurrent miscarriage is far less common, affecting roughly one percent of women by the strictest clinical definition, or up to three percent if biochemical losses count. For decades, the chromosomal explanation covered most of this. Embryoscopy studies found fetal malformations in eighty-five percent of early clinical miscarriages, and seventy-five percent of those fetuses had an abnormal karyotype. Comparative genomic hybridization — a technique for detecting chromosomal imbalances — found that more than ninety percent of preimplantation embryos had at least one chromosomal abnormality in one or more cells. The picture was clear: human reproduction is inefficient because it filters out defective embryos, and most early losses are that filter working. But the chromosomal story starts to buckle when you look specifically at recurrent miscarriage. Fetal chromosomal aberrations show up in twenty-nine to sixty percent of RM cases — but that incidence actually falls as the number of miscarriages increases. Women who have lost four or five pregnancies are less likely, not more likely, to be losing chromosomally abnormal embryos each time. Something else is going on. Parental chromosomal anomalies — a structural rearrangement in one partner — are found in three to six percent of RM couples, about ten times the background population rate. Yet even carrier couples have roughly the same long-term odds of a healthy child as non-carrier couples, around eighty-three versus eighty-four percent. The chromosomal angle explains some cases; it doesn't explain the pattern. So researchers cast a wider net. Thrombophilic disorders — conditions affecting clotting — are one well-documented category. Acquired antiphospholipid syndrome, in which autoantibodies create a prothrombotic and inflammatory state, has the strongest evidence. Two randomized controlled trials showed that low-dose heparin combined with aspirin increased the chance of live birth in RM patients with antiphospholipid antibodies. Hereditary thrombophilias, like factor V Leiden or prothrombin mutations, have weaker evidence — no anticoagulation trial has shown benefit there, and ongoing studies are still running. Immune dysfunction shows up repeatedly in the data: higher rates of antinuclear, antithyroid, and antiphospholipid antibodies in women with RM, along with specific genetic markers in natural killer cell receptors and HLA class II alleles that correlate with worse prognosis. Mouse experiments showed that injecting human immunoglobulin G from antiphospholipid patients increased fetal loss, an effect blocked by complement inhibition — implicating complement-mediated immune injury in the placenta. On the endocrine side, a meta-analysis of thirty-eight studies found that antibodies against thyroperoxidase nearly doubled the risk of RM, with an odds ratio of 2.3. A first-trimester thyroid-stimulating hormone above 2.5 milli-international units per liter, still within the normal range, nearly doubled miscarriage risk in a large prospective study. The critical caveat, which Larsen and colleagues emphasize, is that none of these factors is specific to recurrent miscarriage, and none is always present. Many women with RM have no identifiable cause at all. That gap is not just a diagnostic frustration — it's a signal that the framework itself is incomplete. Two newer threads have emerged to fill it. The first is paternal. Sperm DNA fragmentation — meaning actual DNA damage in sperm, including single- and double-stranded breaks — has been linked to miscarriage in a meta-analysis of sixteen studies, yielding a pooled risk ratio of 2.16. That's a more than twofold increase in miscarriage risk associated with high paternal sperm DNA damage. In a study comparing fertile donors with couples experiencing unexplained RM, eighty-five percent of RM couples showed a profile with high double-stranded DNA damage, compared with thirty-three percent of fertile donors. This is a meaningful shift in perspective. Miscarriage research has been almost entirely maternal, and sperm DNA integrity data push against that assumption hard. The evidence is currently graded as moderate — no consensus yet on which assay to use or what cut-off value matters — but the association is consistent and large. The second thread is more conceptually radical, and it reframes almost everything else. Larsen and colleagues introduce the idea that the decidualized endometrium — the uterine lining after it has undergone the hormonal transformation that prepares it for pregnancy — functions as a biosensor of embryo quality. The traditional view puts the filter in the embryo: defective embryos fail because they are defective. The biosensor hypothesis puts part of the filter in the mother. A normally selective decidua detects and rejects abnormal embryos before they can establish a clinically apparent pregnancy. Given that more than ninety percent of preimplantation embryos carry chromosomal abnormalities, this maternal checkpoint is doing enormous work constantly, mostly invisibly. Here's what makes this consequential for recurrent miscarriage. If the decidual sensor is disrupted — if its discrimination threshold is lowered — then embryos that would ordinarily be rejected silently before a positive pregnancy test can implant and briefly develop, producing pregnancies that are recognized clinically before they fail. Women with recurrent miscarriage, on this model, may not primarily have defective embryos. They may have an endometrium that has lost the ability to say no. That reframing is consistent with why most women with RM eventually carry a pregnancy to term if they keep trying: the embryos are often fine; it's the selection mechanism that's intermittently failing. The therapeutic implication the paper draws is direct — rather than focusing exclusively on embryo quality or immunosuppression, restoring the decidual selective phenotype becomes a treatment target. Hormonal modulators, particularly progesterone in the early luteal phase, are already in intervention trials aimed at this mechanism. Looking forward, the diagnostic toolkit is expanding. Cell-free fetal DNA can be isolated from maternal blood from seven weeks of gestation, and next-generation sequencing can already detect aneuploidies non-invasively. Larsen and colleagues suggest that sequencing the entire fetal genome from maternal circulation will soon be standard practice — a development that could finally clarify the embryonic contribution to both sporadic and recurrent loss at scale. For couples with known chromosomal rearrangements, in vitro fertilization with preimplantation genetic diagnosis — testing embryos before transfer to select only chromosomally normal ones — is an option some clinicians recommend, though live birth rates appear comparable to spontaneous conception in couples with proven fertility. Endometrial biomarkers, including uterine natural killer cells and soluble human leukocyte antigen G, are identified as candidates for development into clinical tests. For the paternal side, methods to select sperm without DNA damage remain a research priority, though no sperm separation approach has yet demonstrated benefit in trials. What this review ultimately demonstrates is a field in transition. The chromosomal filter, the immunological dysfunction, the endometrial sensor, the damaged sperm — these are not competing explanations so much as layers of a genuinely complex picture coming into focus together. Miscarriage affects one in five known pregnancies, and for the women who experience it repeatedly, the distress is compounded by the absence of answers. The mechanistic clarity that Larsen and colleagues map here is what makes genuine prevention feel, for the first time, like a specific and reachable target rather than a general aspiration. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.