The argument here is for the workup itself, not for any one drug at the end of it. A handful of specialist visits across a couple of months converts the worst part of recurrent loss — the open-ended dread, the self-blame — into a defensible plan. Sometimes it finds antiphospholipid syndrome, where the treatment is genuinely transformative. Often it finds nothing, which is itself useful: unexplained recurrent loss still carries a strong majority of next pregnancies going to term with supportive monitoring alone. The workup is the move; what it finds dictates the rest.
Most single miscarriages are not anyone's fault and not a sign of anything. The early embryo has to combine two sets of chromosomes correctly, and the process fails at the dice-roll rate of about one in seven pregnancies; the older the woman, the higher the rate (about 30% at age 25, 50% at age 35, 80% at 42) Magnus 2019. Two losses in a row is the moment that random rate stops being a sufficient explanation. It might still be — bad luck stacks — but it might also be something the body is doing every pregnancy that an ordinary OB visit can't see.
That second possibility is what the workup is for. It looks in five places: the blood (for an autoimmune clotting condition called antiphospholipid syndrome), the thyroid (for an under-active gland the reader may not realise she has), the uterus (for a partition, a fibroid, or scar tissue distorting the cavity), the chromosomes (a parental rearrangement that produces non-viable embryos, or the karyotype of a lost pregnancy showing whether it was the dice or something else), and the metabolism (poorly controlled diabetes, obesity, occasionally a pituitary issue). Each maps to a real mechanism: clots on the placenta, a thyroid that isn't supporting implantation, a uterus that can't carry, an embryo built from an unbalanced karyotype, a metabolic environment that is hostile to pregnancy ESHRE 2023.
About half the time the workup finds one of these. About half the time it finds nothing — what specialists call unexplained recurrent loss. The unexplained bucket is not "we have no idea what's happening"; it is "the named mechanisms have been ruled out, and what remains is random aneuploidy stacking and, almost certainly, real but currently unmeasurable factors at the level of implantation and early development." That distinction matters, because the prognosis for the unexplained bucket is genuinely good.
What the evidence actually supports
The five categories above are the consensus list, but they are not equal in evidence and they are very unequal in treatment. The honest map:
Antiphospholipid syndrome is the one transformative finding. Diagnosed by a specific blood-test pattern — lupus anticoagulant, anti-cardiolipin antibodies, or anti-β2-glycoprotein I antibodies, persistent on two tests at least twelve weeks apart Miyakis 2006Barbhaiya 2023. Treatment in the next pregnancy is daily low-dose aspirin plus a daily injection of low-molecular-weight heparin starting at the positive pregnancy test. The Cochrane review pooled the trials and found this combination roughly halves the loss rate — lifting live births from around 40% to around 70–80% Empson 2005. There is not another finding in the workup with a treatment effect that large.
Thyroid testing finds two real things. Overt hypothyroidism — high TSH, low free T4 — is unambiguously linked to loss and is uncontested treat-it-with-levothyroxine territory. The grey zone is subclinical: a TSH between about 2.5 and 10 with normal T4, sometimes with thyroid antibodies. The TABLET trial put this to bed for one common scenario: 952 women with thyroid antibodies and normal TSH, randomised to levothyroxine 50 µg versus placebo before conception, showed live-birth rates of 37.4% and 37.9% respectively — no benefit Dhillon-Smith 2019. So the workup screens, but the treatment threshold is now narrower than older guidelines suggested.
Uterine imaging finds real anomalies that don't always need surgery. A septate uterus — a fibrous wall dividing the cavity — is the strongest-associated congenital shape, and historically it was cut out hysteroscopically on the assumption surgery improved outcomes. The TRUST trial, the first randomised test of that assumption, found live-birth rates of 31% after surgery versus 35% with expectant management — no benefit and a hint of harm Rikken 2021. The trial was small and recruitment slow, so it is not the final word, but the older default of "find a septum, cut a septum" no longer holds Saravelos 2008.
Chromosomal testing has a sequence. The most informative single test is karyotyping the tissue from a loss, when that tissue can be collected — it answers whether the most recent miscarriage was a random aneuploid event (chance restarted) or something inherited. Parental karyotyping comes second, and only when the products-of-conception result shows an unbalanced structural abnormality that suggests one parent carries a balanced version ESHRE 2023. About 3–5% of recurrent-loss couples carry a balanced translocation ASRM 2020.
Metabolic and lifestyle factors carry independent risk. Tight diabetes control before conception (HbA1c below about 6.5%), weight loss to a BMI under 30, and the standard early-pregnancy package — folic acid, no alcohol, caffeine under 200 mg/day, no smoking — are evidence-supported and underused. The obesity link is stronger than people expect: in a series of women with recurrent loss, obese women had a higher rate of euploid miscarriages, meaning losses not explained by random chromosomal accidents — implying the maternal environment itself was contributing Boots 2014.
And there are several things that look like treatments but aren't. Vaginal progesterone for unexplained recurrent loss was tested directly in the PROMISE trial (836 women); live births were 65.8% with progesterone and 63.3% with placebo — no benefit Coomarasamy 2015. There is a narrower honest case: women who have already had three losses and develop bleeding in a new early pregnancy do appear to get a small boost from vaginal progesterone (a few percentage points), but that is a different scenario than empirically taking it from the start Coomarasamy 2019Devall 2021. Aspirin and heparin without an antiphospholipid diagnosis have been tested too — by the ALIFE, SPIN, and ALIFE2 trials — and have not improved outcomes Kaandorp 2010Clark 2010Quenby 2023.
The workup — what to ask for
The European guideline (ESHRE 2023) is the cleanest current list, and most reproductive-endocrinology clinics in the US and UK follow something close to it. The right time to start is after the second clinical pregnancy loss; waiting for a third is the older standard and now considered too conservative ESHRE 2023.
The whole workup runs 2–4 in-person visits across 6–12 weeks. The bottleneck is usually the antiphospholipid retest: confirmation requires a repeat at twelve weeks, so the first positive is suggestive, not actionable. A negative on the core panel narrows the differential to "unexplained" and is, on its own, useful prognostic information — see What changes when this gets done, below.
What to unlearn
"I caused it." The single most common belief after a loss; the single most reliably wrong. Ordinary exercise, ordinary work, ordinary sex, ordinary mild stress, ordinary diet — none of these have been shown to cause sporadic miscarriage. The self-blame is a documented mood injury after loss, not a clue about the cause Quenby 2021.
"My MTHFR gene caused this." Consumer genetics panels often flag two common MTHFR variants as a clotting risk. The current guidance from ESHRE, the American Society for Reproductive Medicine, and the American College of Medical Genetics is the same: don't test for it, don't treat for it. It does not predict loss and there is no proven treatment ESHRE 2023.
"I just need progesterone next time." Empirically taking progesterone from the start in unexplained recurrent loss was tested and didn't work (PROMISE). There is a real but narrow case for vaginal progesterone in women with three prior losses who develop bleeding in a new early pregnancy — a few percentage points of benefit (PRISM), worth discussing with a specialist if that exact situation arises Coomarasamy 2015Devall 2021.
"Aspirin and heparin will help even without antiphospholipid syndrome." Three trials say no Kaandorp 2010Clark 2010Quenby 2023. Anticoagulants are not benign — daily injections, bleeding risk, drug cost. The treatment belongs to the diagnosis.
"Natural killer cells are attacking my pregnancies." A clinic offering a "reproductive immunology" workup with NK cell testing and immunotherapy is selling a story the trials have not supported; ESHRE specifically recommends against ESHRE 2023Tang 2013.
"My fertility is over." The number is much better than the dread makes it feel. Even after three losses with no identifiable cause, about 60–70% of women have a live birth in their next pregnancy with supportive monitoring alone; after a structured workup and any targeted treatment, 70–80% Brigham 1999ESHRE 2023.
What happens if you don't get the workup
The first and worst thing is the one nobody writes a guideline about: not knowing eats people. Every twinge in a future pregnancy gets read as the start of another loss. Every conversation about trying again becomes a conversation about whether you can stand it. The grief work that should be doing its job — moving through, settling somewhere — gets snagged on a question that has no answer because nobody is looking for one. A multi-centre cohort following women through the year after a miscarriage or ectopic pregnancy found clinically significant post-traumatic stress symptoms in about 29% of women at nine months, and meaningful anxiety or depression in nearly a third Farren 2020. Without structured evaluation, that doesn't naturally resolve; it just becomes the background of the next attempt.
The second thing is concrete and clinical. Antiphospholipid syndrome is missed in roughly one in ten women who actually have it, and the untreated next pregnancy has about a 40% chance of going to term — versus 70–80% with aspirin and heparin Empson 2005. That is the difference between trying for a year and getting nowhere and trying for a year and having a baby. Overt hypothyroidism left untreated raises miscarriage and stillbirth risk and, separately, harms early fetal brain development. A deep uterine septum left undiagnosed forecloses the informed conversation about whether surgery or expectant management is the better call for that specific anatomy. A balanced translocation in one partner, left undiscovered, makes every conception a coin flip nobody knew was loaded.
There is also a long-term piece that women aren't usually told. Antiphospholipid syndrome is not only a pregnancy problem — it is a lifelong clotting and cardiovascular condition, and the diagnosis matters for venous thrombosis and stroke risk decades later Miyakis 2006. The workup that gets done because of recurrent loss is, for some women, also the workup that catches the thing that would have shown up at 50 as a clot.
Where this goes wrong in practice
Empirical treatment before the workup. A clinician offering aspirin, heparin, progesterone, or steroids "to be safe" before the diagnostic results are in is treating their own discomfort with uncertainty, not the patient's condition. The cost is months of injections, drug burden, and a confounded next attempt — whatever happens, neither side will know whether the treatment helped, hurt, or was irrelevant.
Stopping at one positive antiphospholipid result. Confirmation requires a repeat at least twelve weeks later. A single positive lupus anticoagulant or anti-cardiolipin can happen for transient reasons (a recent infection, a recent pregnancy) and resolves on retest. The retest is what separates a real diagnosis from a false alarm.
Karyotyping the parents before karyotyping the pregnancy tissue. Parental karyotype is a downstream test — only useful when the products-of-conception result shows an unbalanced rearrangement that points to one parent carrying the balanced version. Running it first wastes money and produces incidental findings nobody asked for.
Falling into the reproductive-immunology rabbit hole. Natural killer cell testing, lymphocyte immunotherapy, IVIG, intralipid infusions — each has a story, none has a trial that supports it, and several are expensive enough to set fertility plans back. ESHRE's recommendation is plain: don't ESHRE 2023.
Stopping the workup at a normal general-OB ultrasound. A standard 2D scan misses uterine septa and subtle cavity abnormalities at a rate that matters; 3D ultrasound (or saline infusion sonography, or hysteroscopy) is the right tool for this question Saravelos 2008.
Conflating recurrent loss with general fertility workup. They overlap but they aren't the same set of tests. A couple who conceives easily and then loses needs the recurrent-loss workup; a couple who has trouble conceiving and also loses needs both. The right specialist (reproductive endocrinology, or a recurrent-loss clinic) keeps the two distinct.
Skipping the mental-health half. The same evaluation that names a clotting condition or rules out a septum should be naming the depression or post-traumatic stress that is already at work. Specialist clinics that handle the medical and psychological sides together produce better next-pregnancy experience Farren 2020.
How age and history change the picture
The workup is the same; the prior probability of what it finds, and what to do with a negative result, shifts.
Under 35. The highest yield per test. Random aneuploidy is a smaller share of the loss rate at this age, so a "real" cause — antiphospholipid syndrome, a uterine septum, a thyroid issue — is proportionally more likely to be sitting in the workup waiting to be found. A clean workup here is also strongly reassuring: the unexplained-but-investigated prognosis (60–70%+ next live birth) lands on a woman with years of biological runway still ahead Brigham 1999.
35 to early 40s. The workup is still worth doing in full, but the conversation alongside it changes. Aneuploidy now accounts for a larger share of the baseline loss rate, and ovarian reserve becomes part of the planning conversation: not because the workup is less useful but because time matters more. Consider asking for AMH and an antral follicle count alongside the standard panel, and have an honest conversation about whether IVF with preimplantation genetic testing is on the table.
Past 40. Much of the loss rate at this age is straightforward age-related aneuploidy — the eggs themselves Magnus 2019. The workup remains worthwhile because antiphospholipid syndrome and thyroid disease don't care about age, and finding either still changes everything; but it should be paired with realistic, specific counselling on cumulative live-birth probability per cycle, on donor-oocyte success rates (which are largely age-independent because the egg is the limiting factor), and on the time-cost of each path.
One important boundary: this is the workup for early miscarriage (before about 10 weeks, sometimes through the first trimester). Losses later than that — second-trimester losses, or losses with a known cause like cervical insufficiency or placental abruption — have a different differential and a different specialist (maternal-fetal medicine rather than reproductive endocrinology). If the losses are happening late, the path described here is the wrong one; ask for a maternal-fetal medicine referral instead.
What changes when this gets done
The first thing that changes is in the weeks the workup is happening, before any treatment has even started. The mornings stop being diagnostic — you are not running silent loops about whether last night's wine or last week's deadline did the thing. The conversation with a partner has a structure now: we are looking, and what is found will tell us what to do. The person across the room from you who didn't quite know what to say has somewhere to look, too.
The second thing changes at the end of the workup. One of three outcomes is now true.
- A named cause, with a real treatment. The clearest version: antiphospholipid syndrome confirmed, daily aspirin and a daily injection of low-molecular-weight heparin starting at the positive pregnancy test, and the next-pregnancy success rate goes from about 40% to 70–80% Empson 2005. The diagnosis also matters past pregnancy — the long-term clotting and cardiovascular risk is now being managed, decades earlier than the average woman with this condition finds out about it.
- A finding that reshapes the plan but isn't a quick fix. A deep uterine septum, a balanced parental translocation, a poorly controlled metabolic situation. None of these come with one-button treatments, but each turns the next attempt from a blind one into a planned one — and into a conversation with a specialist about IVF with preimplantation genetic testing, weight loss before the next cycle, or shared-decision surgery, depending on what was found.
- Nothing on the core panel. The biggest single shift here is psychological — the "unexplained" diagnosis lands very differently when the workup has actually been done than when it is still hanging open. The number to hold onto: about 60–70% of women with three unexplained losses have a live birth in their next pregnancy with supportive monitoring alone; the structured-workup cohort runs to 70–80% Brigham 1999ESHRE 2023. The reframing from I don't know what's wrong with me to we have looked, this is what is known, this is what we do next shows up in measured improvements in anxiety and pregnancy-specific stress in the following attempt Farren 2020.
And there is the smaller, less-promised payoff in the side findings. A subset of women come out of this workup with a treated thyroid, a treated metabolic issue, or a real cardiovascular risk flag they wouldn't have had for another decade. None of those rescue a next pregnancy on their own, but they are the kind of incidental good that makes the workup pay rent past fertility.
What it costs and where to go
The right address is a reproductive-endocrinology clinic, a maternal-fetal-medicine specialist, or a dedicated recurrent-loss clinic — not a general OB/GYN. General-practice OBs vary widely in how complete a workup they run, and the negative experience women describe most often after recurrent loss is a clinician who tells them to "just try again" without examining anything. The referral is the move Toth 2018.
In the US, the panel runs roughly $500 to $2,000 out of pocket when not covered, and is usually mostly covered when ordered by an in-network specialist after a documented second loss. Most state-mandated infertility coverage now extends to recurrent loss; the diagnostic codes specialists use are well-established for claims. In the UK, the NHS covers the workup after the third loss (and increasingly after the second); private RPL clinics run a few hundred pounds. Most European public systems cover after the third loss and many after the second.
Time is the bigger cost. 2–4 visits, blood draws (one of them inconveniently twelve weeks after the first if antiphospholipid antibodies came up positive), and a 3D ultrasound that may need to be timed in the cycle. If antiphospholipid syndrome is confirmed and a new pregnancy starts, the daily heparin injection adds bruising, injection-site soreness, and a real but well-tolerated treatment burden through pregnancy and six weeks after.
One practical thing to do before the first specialist visit: arrange in advance for the next loss's tissue to be sent for chromosomal analysis, if there is one. This is not what people want to think about while they are still hoping the current pregnancy holds, but the karyotype of a miscarriage is the single most informative test in this entire workup, and it is the easiest to lose by default to hospital handling. The clinic can put the paperwork in place so the lab knows what to do if the moment arrives.
Related, but not covered here
A single first miscarriage doesn't get this workup — the background rate is too high and most single losses resolve into a normal next pregnancy without intervention. Later losses (after about ten weeks, and especially in the second trimester) have a different differential and need a maternal-fetal-medicine path, not the one above. Couples who are also having trouble conceiving need a fertility workup alongside this one — the two overlap but they aren't the same set of questions. Stillbirth (after 24 weeks), molar pregnancy, ectopic pregnancy, termination for fetal anomaly, and the general territory of preconception care each warrant their own conversations.
Substance and claimed effects
Recurrent pregnancy loss (RPL) is defined by ESHRE and ASRM as two or more clinical pregnancy losses, consecutive or non-consecutive, before 24 weeks gestation ESHRE 2023ASRM 2020. RCOG retains the older threshold of three consecutive losses before initiating a full workup RCOG 2011. Background rate of clinical miscarriage is ~10–15% per pregnancy; RPL by the 2-loss definition affects ~5% of women trying to conceive, by the 3-loss definition ~1–2% Quenby et al. 2021.
The entry covers the structured workup (clotting, thyroid, uterine, chromosomal, metabolic), the evidence base behind each test and treatment, the prognostic counsel (subsequent live birth probability after evaluation), and the consequences across reproductive planning, fertility, and mental health. The brief names five workup domains plus four consequence dimensions; all are in scope.
Evidence by addressing question
mechanism
Most sporadic first-trimester losses are caused by fetal aneuploidy — random meiotic errors producing chromosomally non-viable embryos Stephenson et al. 2002. The aneuploidy rate in miscarriage products of conception rises steeply with maternal age: ~30% at age 25, ~50% at age 35, ~80% at age 42 Magnus et al. 2019. RPL therefore has two layered mechanisms: (i) random aneuploid losses stacking by chance, more likely the older the woman; and (ii) an underlying maternal or paternal cause that pushes losses above the chance rate. The workup distinguishes the two by hunting for (ii).
Mechanistic categories identified by structured evaluation:
- Antiphospholipid syndrome (APS) — autoantibodies (lupus anticoagulant, anti-cardiolipin, anti-β2-glycoprotein I) damage trophoblast and placental vasculature; inhibit annexin V anticoagulant shield; activate complement and induce intervillous thrombosis Bohlmann et al. 2010. Single most established treatable cause.
- Inherited thrombophilia — Factor V Leiden, prothrombin G20210A, protein C/S deficiency, antithrombin deficiency. Mechanism: hypothesised placental thrombosis. Association exists in case-control series, but causality and treatment response have not held up in RCTs Quenby et al. 2023.
- Thyroid disease — overt hypothyroidism (high TSH, low free T4) causes anovulation, luteal phase defect, and increased loss. Thyroid peroxidase (TPO) antibody positivity correlates with miscarriage even when euthyroid Stagnaro-Green et al. 2017. Mechanism likely a generalised autoimmune diathesis affecting implantation.
- Uterine anatomy — septate uterus (a fibrous septum dividing the cavity) has the strongest association with RPL among congenital anomalies; prevalence ~5% in RPL vs ~3% in general population Saravelos et al. 2008. Submucous fibroids and intrauterine adhesions (Asherman syndrome) distort the cavity and reduce implantation surface.
- Parental chromosomal rearrangement — balanced reciprocal or Robertsonian translocation in one partner produces gametes with unbalanced karyotypes; ~3–5% of RPL couples carry one ASRM 2020. Each conception has an elevated probability of being non-viable.
- Metabolic/endocrine — poorly controlled diabetes (HbA1c > 7%), obesity (BMI ≥ 30), and PCOS-associated insulin resistance increase loss rates Boots et al. 2014. Boots et al. showed obese RPL patients have a higher proportion of euploid miscarriages — i.e. losses not explained by random aneuploidy, pointing at a real maternal contribution.
evidence
Five major causes have meaningful evidence behind detection; treatment evidence is much weaker for several.
- APS — diagnosed per the 2006 Sapporo criteria (now superseded by the 2023 ACR/EULAR criteria) requiring persistent positivity on two occasions ≥12 weeks apart, plus a clinical event including ≥3 consecutive losses <10 weeks, ≥1 loss ≥10 weeks of a morphologically normal fetus, or premature birth <34 weeks due to placental insufficiency Miyakis et al. 2006Barbhaiya et al. 2023. Cochrane meta-analysis: aspirin plus low-dose heparin in confirmed APS lifts live birth from ~40% to ~70–80% (relative risk reduction ~54%) Empson et al. 2005. PREGNANTS cohort confirmed triple-positive APS carries the worst prognosis and benefits most from treatment Saccone et al. 2017.
- Inherited thrombophilia — ALIFE2 (n=326 women with RPL and inherited thrombophilia) randomised to LMWH+standard care vs standard care alone showed no benefit on live birth (72% vs 71%) Quenby et al. 2023. ESHRE 2023 now recommends against routine inherited-thrombophilia screening in unselected RPL outside research ESHRE 2023.
- Thyroid — overt hypothyroidism: treat with levothyroxine to TSH < 2.5 mIU/L preconception; uncontested. Subclinical hypothyroidism (TSH 2.5–10 with normal T4) and euthyroid TPO+: the TABLET trial (n=952 women with TPO antibodies, normal TSH, history of miscarriage or infertility) randomised to levothyroxine 50 µg vs placebo and found no difference in live birth (37.4% vs 37.9%) Dhillon-Smith et al. 2019. ESHRE recommends TSH and TPO testing but treating only overt hypothyroidism and subclinical with TSH > 4 mIU/L; ATA 2017 is slightly more permissive Stagnaro-Green et al. 2017.
- Uterine septum — the TRUST trial (international RCT, n=80) randomised women with a septate uterus and reproductive failure to hysteroscopic septoplasty vs expectant management; live birth at follow-up was 31% with surgery vs 35% without — no benefit from resection Rikken et al. 2021. Genuinely controversial: trial small, recruitment slow, surgeons remain split. ESHRE conditionally recommends resection in RPL with deep, broad septum and lack of alternative explanation.
- Parental karyotype — ESHRE 2023 recommends karyotyping only when prior loss tissue shows an unbalanced structural abnormality (i.e. as a downstream test, not routine screening) ESHRE 2023. ASRM 2020 retains broader recommendation. Detection enables genetic counselling and consideration of PGT-A with IVF ASRM 2018; trial data on PGT-A's effect on cumulative live birth in RPL specifically remains weak Munné et al. 2019.
- Lifestyle/metabolic — observational evidence for smoking cessation, alcohol abstinence, caffeine reduction (< 200 mg/day), and BMI optimisation < 30 lowering miscarriage risk; consistent across cohorts Quenby et al. 2021.
- Empirical therapies that don't work: PROMISE trial — vaginal progesterone vs placebo in 836 women with unexplained RPL, live birth 65.8% vs 63.3%, no significant difference Coomarasamy et al. 2015. PRISM trial and the Devall IPD meta-analysis show a modest benefit of vaginal progesterone in women with bleeding in early pregnancy and prior losses (live birth +5 percentage points if 3+ prior losses) — a different clinical scenario Coomarasamy et al. 2019Devall et al. 2021. SPIN trial — LMWH+aspirin vs intensive surveillance in 294 women with unexplained RPL, no benefit Clark et al. 2010. ALIFE trial (Kaandorp) — aspirin alone or aspirin+LMWH in unexplained RPL, no benefit Kaandorp et al. 2010. Empirical immunotherapy (steroids, IVIG, prednisolone for high uterine NK cells) Tang et al. 2013 remains unproven and harmful when used reflexively.
protocol
The ESHRE 2023 guideline structures the workup into a short list of tests every patient should get, plus conditional tests ESHRE 2023:
- Recommended for all: antiphospholipid antibodies (lupus anticoagulant, anti-cardiolipin IgG/IgM, anti-β2-glycoprotein I IgG/IgM), TSH and TPO antibodies, pelvic ultrasound (preferably 3D) to assess uterine anatomy.
- Conditional: karyotyping of products of conception (POC) after the second and subsequent losses; parental karyotype only when POC shows unbalanced abnormality or strong family history; saline infusion sonography or hysteroscopy when 3D ultrasound suggests anomaly; HbA1c if diabetes risk factors; prolactin and ovarian reserve assessment if cycle irregularity or advanced age.
- Recommended against (routine): inherited thrombophilia screening, natural killer cell testing, HLA typing, TORCH serology, sperm DNA fragmentation (use case-by-case), MTHFR, and most empirical immunotherapy outside trials.
Timing: workup can be initiated after the second loss; results take 6–12 weeks to assemble. APS confirmation requires two positive results ≥12 weeks apart, which sets the minimum interval before treatment can be planned.
Treatment by cause:
- Confirmed APS: low-dose aspirin (75–100 mg daily) starting preconception + prophylactic LMWH (e.g. enoxaparin 40 mg daily) starting at positive pregnancy test, continued through pregnancy and 6 weeks postpartum Empson et al. 2005.
- Overt hypothyroidism: levothyroxine to TSH < 2.5 mIU/L preconception. Subclinical with TSH 4–10: treat per ATA 2017 if TPO+; ESHRE more conservative.
- Diabetes: tight glycaemic control to HbA1c < 6.5% preconception; folic acid 5 mg daily.
- Obesity: weight loss to BMI < 30 before next attempt.
- Septate uterus: shared decision; hysteroscopic resection in deep wide septa per shared decision after TRUST Rikken et al. 2021.
- Parental translocation: genetic counselling; natural conception (most pregnancies still succeed) vs IVF with PGT-A as options.
- Unexplained (≈50% after full workup): close monitoring with early ultrasound (the "tender loving care" effect), lifestyle optimisation, and reassurance — Brigham's series and ESHRE both report subsequent live birth rates of 60–75% with supportive care alone Brigham et al. 1999ESHRE 2023.
contraindications
The workup itself is low-risk: venepuncture and a 3D ultrasound. LMWH carries small bleeding risk and requires injection sites; allergy to heparin is a contraindication (use fondaparinux). Hysteroscopic septoplasty carries small uterine perforation and Asherman risk. The most consequential "contraindication" is empirical treatment without a diagnosis — anticoagulants, steroids, and IVIG given to women with unexplained RPL expose them to drug risks without evidenced benefit Kaandorp et al. 2010Clark et al. 2010.
misconceptions
- "Something I did caused it." Sporadic miscarriage is overwhelmingly random aneuploidy; exercise, work, sex, mild stress, and ordinary diet are not causes. The empiric guilt that follows loss is universal and almost always wrong Quenby et al. 2021.
- "MTHFR caused my losses." Direct-to-consumer genetic panels often flag MTHFR C677T/A1298C as a thrombophilia. ESHRE, ASRM, and the American College of Medical Genetics explicitly advise against MTHFR testing for RPL — no causal relationship; no proven treatment benefit; testing causes anxiety without evidence-based action ESHRE 2023.
- "I need progesterone for next time." Empirical progesterone for unexplained RPL doesn't improve live birth (PROMISE) Coomarasamy et al. 2015. The narrower case — vaginal progesterone for women with first-trimester bleeding and 3+ prior losses — has a small but real signal (PRISM, Devall) Coomarasamy et al. 2019Devall et al. 2021.
- "Aspirin + heparin will help even without APS." Two RCTs (SPIN, ALIFE, ALIFE2) say no Clark et al. 2010Kaandorp et al. 2010Quenby et al. 2023. The treatment carries injection-site burden and bleeding risk.
- "Luteal phase defect needs progesterone." Largely retired diagnosis; endometrial biopsy unreliable; not a recommended workup.
- "My fertility is over." Even after three losses with no identifiable cause, ~60–70% of women have a live birth in their next pregnancy with supportive care; after structured workup and targeted treatment, 70–80% Brigham et al. 1999.
- "NK cell testing will reveal an immune cause." ESHRE recommends against; testing is unstandardised and prednisolone for high uterine NK cells did not improve live birth in a feasibility trial Tang et al. 2013.
failure-modes
- Starting empirical treatment (anticoagulants, progesterone, immunotherapy) without completing the workup — locks in months of drug burden without targeting a real cause.
- Stopping the workup after one negative test. APS confirmation requires two positives ≥12 weeks apart; a single positive lupus anticoagulant or a single positive cardiolipin is suggestive but not diagnostic.
- Karyotyping parents before karyotyping products of conception. The POC result selects who needs further parental testing.
- Treating subclinical hypothyroidism (TSH 2.5–4) reflexively post-TABLET. The trial argues against.
- Doing a hysteroscopic septoplasty on a small partial septum and attributing the next live birth to surgery. Background rate is ~60–70% next time anyway.
- Conflating RPL workup with general subfertility workup. Different protocols. RPL focuses on losses; subfertility focuses on conception. Many couples need both.
- Missing the psychological dimension. Untreated grief, anxiety, and PTSD persist for years and are independent predictors of poor pregnancy experience next time Farren et al. 2020.
audience
Women experiencing recurrent loss are the direct audience, but partners share the diagnostic relevance (parental karyotyping is for both partners). Age stratifies importance:
- Under 35 — workup yields the highest return per test because aneuploidy is less likely to be the dominant mechanism; an identifiable maternal cause is more probable.
- 35–40 — workup still valuable but aneuploidy share rises; counsel on ovarian reserve and time-pressure for further attempts.
- Over 40 — much of the loss may be age-related aneuploidy; workup remains worthwhile but should be paired with realistic conversation about cumulative live-birth probability and gamete options.
Late losses (≥10 weeks, second trimester) have a different differential — cervical insufficiency, abruption, infection — and a different workup path; RCOG and ESHRE separate them.
alternatives
- Watchful waiting after structured reassurance. Genuinely valid for women with no identified cause: ~60–70% next live birth without treatment Brigham et al. 1999.
- IVF with PGT-A for couples with parental translocation or advanced maternal age; reduces miscarriage rate per transfer but cumulative live birth gain is modest and the procedure burden is significant Munné et al. 2019.
- Donor gametes when poor oocyte quality is the dominant mechanism (advanced age, premature ovarian insufficiency).
- Surrogacy when uterine factor is uncorrectable.
- Adoption / no further attempts — the choice to stop trying is a legitimate option, not a failure mode.
practicalities
Full ESHRE-aligned workup costs $500–$2,000 in the US depending on insurance coverage; most components are covered when ordered by a maternal-fetal medicine or reproductive endocrinology specialist. NHS and most European systems cover after the third loss; some after the second. The workup takes 2–4 in-person visits over 6–12 weeks. Specialist referral is the bottleneck — general OB/GYNs vary widely in workup completeness; an RPL clinic or a reproductive endocrinologist is the higher-yield route Toth et al. 2018.
history
The field has cycled through a series of empirical treatments that didn't survive trials: progesterone for "luteal phase defect" (PROMISE, 2015), heparin for unexplained loss (SPIN 2010, ALIFE 2010), heparin for inherited thrombophilia (ALIFE2 2023), levothyroxine for euthyroid TPO+ (TABLET 2019), septum resection (TRUST 2021). The pattern: mechanism-plausible intervention adopted on observational data, RCT later shows no benefit, guidelines pull back. APS treatment is the clear exception — RCT-supported, durable, transformative.
stakes
Without workup: APS is missed in ~10–15% of RPL patients and treatable APS pregnancies that proceed without aspirin+heparin have ~30–40% live birth vs ~70–80% with treatment Empson et al. 2005. Untreated overt hypothyroidism increases miscarriage, stillbirth, and developmental risk. A deep uterine septum left undiagnosed forecloses informed surgical-vs-expectant choice. A balanced translocation in a partner shapes whether to continue natural attempts, pursue PGT-A, or move to donor gametes. Psychologically, the persistence of unexplained loss without structured evaluation deepens self-blame and lengthens grief Farren et al. 2020. APS additionally carries long-term venous thrombosis and cardiovascular risk independent of pregnancy outcome.
payoff
With structured workup completed and any treatable cause addressed: subsequent live birth rate is 70–80% in most series, including the unexplained cohort treated with supportive monitoring alone Brigham et al. 1999ESHRE 2023. The reframing from "I don't know what's wrong with me" to "we have looked, here is what we know, here is what we do next" is itself a mental-health intervention that improves pregnancy experience and reduces anxiety in the next attempt Farren et al. 2020. APS women on aspirin+heparin go from ~40% to ~70–80% live birth; severe parental translocation carriers move toward PGT-A with realistic per-cycle probability of euploid transfer; subclinical thyroid issues are settled by data instead of speculation.
out-of-scope
Single (first) miscarriage workup; ectopic pregnancy; molar pregnancy; second-trimester loss workup (cervical insufficiency, abruption); stillbirth (≥24 weeks); termination for fetal anomaly; preeclampsia; gestational diabetes; preconception care more broadly. Fertility (time-to-conception) workup is adjacent but separate.
The credibility range
Optimist case
RPL has gone from a diagnostic dead-end to a structured workup with one transformative treatment (aspirin+heparin for APS) and a clean list of tests that meaningfully change the prognostic conversation. Even when the workup is "negative," 60–80% of subsequent pregnancies succeed with supportive monitoring. The information itself is the intervention: women who know they don't have APS, hypothyroidism, a septum, or a parental translocation can plan again with realistic odds rather than catastrophising. The mental-health benefit of being investigated is substantial and measurable.
Skeptic case
Outside APS and overt thyroid disease, the workup is mostly ruling things out, not finding causes. Half of RPL remains "unexplained" after every recommended test. Several previously embraced treatments (levothyroxine for euthyroid TPO+, heparin for inherited thrombophilia, septum resection, empirical progesterone, immunotherapies) have failed RCTs. The field is littered with abandoned tests and treatments women paid for and were anxious about. Over-testing — particularly NK cell, MTHFR, sperm DNA fragmentation — drives cost and worry without changing outcomes. The optimist's 70–80% next live birth applies to the unexplained cohort with no intervention beyond supportive monitoring; much of what is sold as "treatment" is regression to the mean.
Author's call
Both are right at different addresses. The ESHRE 2023 short list — APS antibodies, TSH+TPO, 3D pelvic ultrasound, POC karyotype on subsequent losses, lifestyle review — is high-yield and low-burden, and identifying APS is genuinely transformative. Beyond that list, the marginal test rarely pays for itself, and several positively harm by triggering unproven treatments. Centre of gravity for the article: do the short list, treat what is found, decline what is not on the list, plan with the prognostic data that the workup itself provides. Evidence rating: 3 — the workup is well-defined and consensus-backed, but several treatment components have failed RCTs and 50% remains unexplained. Controversy: 3 — subclinical hypothyroidism thresholds, septum management, and empirical immunotherapy are actively debated.
Stakeholder + incentive map
- Reproductive endocrinology and infertility (REI) specialists / RPL clinics — push structured workup; sometimes push PGT-A and IVF beyond evidence base.
- OB/GYNs in general practice — variable workup completeness; some still defer to RCOG's 3-loss threshold.
- Reproductive immunology subcultures — promote NK cell testing, IVIG, lymphocyte immunotherapy; ESHRE and ASRM recommend against; commercial labs sustain the demand.
- Direct-to-consumer genetic testing — flag MTHFR as a thrombophilia; drive anxiety; ESHRE recommends against testing or treating.
- Compounding pharmacies / progesterone manufacturers — push empirical progesterone; PROMISE undermined the general indication but the bleeding-history sub-indication keeps prescribing alive.
- Patients and patient communities — the lived experience of repeated loss creates strong demand for any intervention with a plausible mechanism; this demand is what sustained heparin and progesterone use against RCT evidence.
- Skeptic / evidence-based counterweight — Cochrane, ESHRE 2023 update, RCOG. The pull toward "do less" is structural and intellectually honest.
Population variability
- Maternal age dominates baseline loss rate; an "unexplained" workup in a 41-year-old is differently informative than the same workup in a 31-year-old Magnus et al. 2019.
- Obesity increases loss rate independently and increases the share of euploid miscarriages — i.e. losses not attributable to chance aneuploidy Boots et al. 2014.
- Prior pregnancy history matters: a woman with two losses but a prior live birth has a better next-pregnancy prognosis than one with two losses and no live births (secondary vs primary RPL).
- Number of losses: after 2 losses next-loss risk is ~25%; after 3 it is ~30–40%; risk rises but slower than intuition.
- Ethnicity and geography: literature is heavily Northern European / North American; APS prevalence and thyroid antibody base rates differ across populations.
- Couples with infertility plus RPL differ from couples conceiving readily then losing — the latter has better next-pregnancy odds.
Knowledge gaps
- The 50% of RPL that remains unexplained after structured workup. The "unexplained" bucket likely contains real but unmeasured mechanisms (endometrial receptivity, decidualisation defects, sperm DNA quality, microbiome) and a large component of random aneuploidy stacking.
- Whether any subset of unexplained RPL truly benefits from immunomodulation (TNF-α inhibitors, low-dose prednisolone, intralipid). Current evidence is negative on average but heterogeneous.
- Whether endometrial receptivity testing (ERA, immune profiling) identifies a treatable subgroup.
- Whether PGT-A meaningfully improves cumulative live birth in RPL specifically, vs simply selecting transfers and shifting outcomes earlier.
- Optimal management of subclinical thyroid abnormalities post-TABLET.
- Why some APS women still lose despite full aspirin+heparin treatment ("refractory APS") and what to add (hydroxychloroquine? plasmapheresis?).
- Long-term cardiovascular and autoimmune trajectory of women with antiphospholipid antibodies who don't meet full APS criteria.
- Optimal psychological support model and its measurable effect on pregnancy outcome.
Scope follows the brief. The five workup domains the brief named (clotting, thyroid, uterine, chromosomal, metabolic) and the four consequences (fertility, subsequent pregnancy outcomes, mood, reproductive planning) are all covered end-to-end. Mood is the strongest of the four and earned the only score of 4; the others are addressed in proportion.
Definition threshold. Followed ESHRE 2023's two-loss definition rather than RCOG's three-loss because the ESHRE position is current and consistent with US practice; the entry calls out the older three-loss threshold so a reader running into it from a UK clinician isn't caught off-guard. Couples in the NHS who can only access the workup after the third loss are not misled.
Action and cadence. Chose respond over know or test: the audience is women who have already had two losses, so the action genuinely is the response. as-needed fits the trigger; course was the runner-up because the workup itself is bounded, but the framing is response-to-event, not a course one starts cold.
Score calls worth flagging.
- longevity and energy at 1 each: defensible but could plausibly be 0. The 1 honours the incidental finding rate — undiagnosed APS carries long-term cardiovascular implications; overt hypothyroidism affects daily energy. Going below would be undercounting real downstream good.
- mood at 4: pinned by Farren 2020's PTSD/anxiety numbers and the documented effect of structured evaluation on next-pregnancy psychological experience. Could plausibly be 5 — but the workup itself isn't on the level of an effective psychiatric intervention, so 4 is the honest read.
- evidence at 3: lifted by ESHRE 2023 and the strong APS treatment data, pulled down by the post-RCT collapse of several previously embraced treatments (TABLET, TRUST, ALIFE2, PROMISE, SPIN). Not a 4 because the field is still settling several conditional questions.
- applicability at 2: 2-loss RPL hits ~5% of women trying to conceive. Considered the awareness-audience lift (partners participate; future-pregnancy planners benefit from the framework) but it didn't justify a 3 — this is condition-specific in a way smoking and emergency-recognition aren't.
What was deliberately not covered. Single first miscarriage (different prior probability, different counselling); second-trimester loss workup (different specialist — maternal-fetal medicine — different differential including cervical insufficiency and abruption); fertility workup proper; long-term APS management beyond pregnancy; stillbirth; grief therapy modalities; ART/IVF protocols beyond the PGT-A pointer. Each is named in out-of-scope at high level.
Hard editorial calls.
- Stakes placement: spec recommends stakes before protocol so loss-aversion earns the protocol an honest read. Placed stakes after protocol and misconceptions because the dek already does the loss-aversion job, and the article's centre of gravity is the workup itself — protocol benefits from coming straight after the evidence map. Stakes still appears mid-article and lands strongly.
- Septum management is genuinely contested post-TRUST (n=80, slow recruitment, surgeons split). Article reports TRUST honestly but doesn't tell women to refuse surgery — frames it as shared decision. The editor notes flag this so a reviewer doesn't mistake even-handedness for fence-sitting.
- The narrow PRISM/Devall progesterone case (vaginal progesterone in women with 3+ prior losses and new bleeding) is kept in
misconceptionsrather thanprotocolbecause the empirical-progesterone misconception is more common than the narrow-indication scenario; readers in that specific scenario will read the misconception, see the carve-out, and ask.
Future-link / related-entry candidates.
- Antiphospholipid syndrome as a standalone entry — the long-term clotting and cardiovascular dimension is wider than recurrent loss and would carry its own scoring.
- First (single) miscarriage — distinct counselling, different prior probability, different mood arc.
- Late pregnancy loss / second-trimester loss — maternal-fetal-medicine differential.
- Thyroid in pregnancy — covers preconception TSH targets across pregnancy attempts more broadly.
- Preimplantation genetic testing (PGT-A) — the IVF intervention referenced here.
- Grief and PTSD after pregnancy loss — the psychological half this entry points at but doesn't fully address.
- Preconception care for the broader before-trying package.
Recurrent Pregnancy Loss
The biggest single lift here. Going from "I don't know why this keeps happening" to a real answer — or to a real ruling-out — cuts the worst of the dread and self-blame.
A few hundred to a couple thousand dollars in the US, often covered when a specialist orders it. Free or low-cost in most public systems.
A handful of specialist visits and blood draws over a couple of months. If a clotting condition is found, daily injections during the next pregnancy.
Strong for the core tests and the one transformative treatment (antiphospholipid syndrome). Mixed for several other treatments that sound plausible but haven't held up in trials.
The workup sometimes turns up a thyroid problem, a clotting condition, or unmanaged blood sugar — things worth treating whether or not you're trying to get pregnant.
If the workup finds antiphospholipid syndrome, you also pick up an early warning on lifelong clot and stroke risk — useful well beyond your next pregnancy.
Small chance: if you've been tired and blaming it on stress, an undiagnosed thyroid problem turns up in this workup more often than you'd guess.