Recurrent Pregnancy Loss

The workup, by what the evidence says.

A test-by-test guide for primary care. Three tiers — do these, consider these, skip these — drawn from ESHRE 2022 and RANZCOG C-Gyn 38.

Definition

Two or more pregnancy losses before 24 weeks (ESHRE 2022). Losses need not be consecutive. A biochemical pregnancy (β-hCG-confirmed) counts.

When to investigate

After two losses — the threshold for offering antiphospholipid screening, the one most consistently actionable result. Female age and complete pregnancy history are the strongest prognostic inputs.

Show

The minimum evidence-based panel

Four tests cover the routine workup. Everything else is tailored to history.

Routinely recommended

Offer to every woman with two or more losses. Each result is actionable.

Lupus anticoagulant

Antiphospholipid screen · part 1

Strong

First-line APS test. Treatable cause of recurrent loss with the largest live-birth gain when positive.

Yield: APS in ~10–15% of RPL — the single most actionable finding.
When: After two losses. Repeat positive ≥12 weeks later to confirm.
Treats: If confirmed APS + ≥3 losses → low-dose aspirin + prophylactic LMWH from positive βhCG.
Note: Not for women on warfarin or DOACs — false positives. Withhold heparin 12 h before sample.
Source: ESHRE Rec 15 · RANZCOG R8

Anticardiolipin antibodies

IgG + IgM · part 2 of APS panel

Strong

Pairs with lupus anticoagulant. Together they identify APS — the only treatable serological cause.

Cut-off: Medium/high titre (>40 GPL or MPL units, or >99th centile) is clinically meaningful.
When: After two losses. Confirm ≥12 weeks apart per Sydney criteria.
Pearl: Order LA + aCL together — one swing covers ~95% of APS detectable in primary care.
Source: ESHRE Rec 15 · RANZCOG R8

TSH + TPO antibodies

Thyroid function + autoimmunity

Strong

Subclinical hypothyroidism and TPO positivity are over-represented in RPL. Overt hypothyroidism is treatable.

Yield: Abnormal TSH or TPO+ in ~15–25% of RPL.
Action: If TSH abnormal → add free T4. Treat overt hypothyroidism. Euthyroid TPO+ does not benefit from levothyroxine (ESHRE 2022 update).
Caveat: In a future pregnancy with TPO+ or prior SCH, check TSH at 7–9 weeks gestation.
Source: ESHRE Rec 24, 47, 51 · RANZCOG R15-16

3D transvaginal ultrasound

Uterine cavity + adenomyosis

Conditional

Recommended for all. Distinguishes septate from bicorporeal uterus; 2D US screens for adenomyosis.

Yield: Uterine anomaly in ~10–25% of RPL; adenomyosis newly added in ESHRE 2022.
Order: Request "TVS with 3D reconstruction" — not all clinics offer; sonohysterography (SHG) is acceptable alternative.
Note: MRI only if 3D US unavailable. HSG is inferior to SHG for cavity assessment.
Action: If Müllerian anomaly found → assess renal tract. Septum resection no longer routinely recommended after Rikken RCT 2021.
Source: ESHRE Rec 33–38 · RANZCOG R8

Consider case-by-case

Tailor to family history, age, prior workup or partner factors. Don't reflex-order.

Anti-β2 glycoprotein I

aβ2GPI · APS extended panel

Very low

Optional add-on to the APS workup. One trial suggests treatment may improve live birth when isolated aβ2GPI+ is present.

When: Reasonable to bundle with LA + aCL on the same form. Isolated positivity rare.
Status: ESHRE GPP only; RANZCOG "consider case-by-case".
Source: ESHRE Rec 16

Parental karyotype

Peripheral blood, both partners

Conditional

Balanced translocation carrier rate ~2–4% in RPL. Yield drops sharply with maternal age >39 and <3 losses.

Yield: <2.2% if female age >39, <3 losses, negative family history (Franssen 2005). Higher in younger couples and where there is a family history of pregnancy loss or aneuploidy.
When: If POC testing not available, abnormal POC result, family history of recurrent loss, or consanguinity.
Action: Refer for genetic counselling if abnormal. Live birth rates remain good — IVF/PGT-SR is one option, not the default.
Source: ESHRE Rec 13 · RANZCOG R8

POC chromosome microarray

Array-CGH on miscarriage tissue

Conditional

Explanatory, not predictive. ~55–64% of POC samples show chromosomal anomaly — slightly lower in RPL than sporadic loss.

Yield: Aneuploidy in ~55% of POC in RPL vs ~64% in single losses (Lei 2022).
Why array: Array-CGH preferred over conventional karyotype: fewer failed cultures, less maternal contamination.
Caveat: Result rarely changes management. Counsel pre-test about emotional impact and limitations.
Source: ESHRE Rec 11–12 · RANZCOG R9

Sperm DNA fragmentation

Male factor · diagnostic only

Low

Growing evidence of association with RPL. Most relevant in men with smoking, alcohol, obesity, or older paternal age.

When: Unexplained RPL after standard workup. Particularly informative for unhealthy paternal lifestyle.
Action: Drives lifestyle counselling. No specific reproductive treatment supported by RCT data — PICSI does not have evidence in RPL.
Cost: Private only in Australia; expect ~$200–$300.
Source: ESHRE Rec 39–40 (updated 2022) · RANZCOG R8

Antinuclear antibodies (ANA)

Diagnostic / prognostic only

Low

Modest association with RPL in meta-analysis; ANA+ may worsen prognosis but does not direct treatment.

Use: Consider for explanatory purposes only. Not actionable as a single test.
If positive: Trigger broader connective-tissue review only if clinical features present.
Source: ESHRE Rec 21 · RANZCOG R8

Prolactin

Only with symptoms

Conditional

Only if oligomenorrhoea, amenorrhoea or galactorrhoea. No routine role.

Order: Only when clinical symptoms of hyperprolactinaemia.
Source: ESHRE Rec 27 · RANZCOG R8

Don't routinely order

Common requests with low yield, no actionable result, or evidence against. Useful for pushback.

Inherited thrombophilia panel

FVL, prothrombin G20210A, protein S

Don't · Mod

No evidence anticoagulation improves live birth in inherited thrombophilia with RPL.

Exception: Order only when personal/family VTE history independently indicates testing. Treat for VTE prevention, not for RPL.
Source: ESHRE Rec 14, 44 · RANZCOG R8

Protein C / antithrombin / MTHFR

Extended thrombophilia

Don't · Strong

MTHFR is not a thrombophilic mutation in the clinically relevant sense. Strong evidence against routine testing.

Pearl: MTHFR polymorphisms are common, do not predict pregnancy loss, and do not require treatment beyond standard folate.
Source: RANZCOG R8 (Strong against)

HLA / NK cell / cytokine testing

Immunological

Don't · Strong

Marketed by some private clinics. No therapeutic implication and significant technical challenges.

Caveat: HLA-DRB1*15:01 / DRB1*07 / DQB1*05 can be considered in Scandinavian women with secondary RPL after the birth of a boy — niche.
Source: ESHRE Rec 17–23

Ovarian reserve (AMH / AFC)

For RPL purpose

Don't · Strong

No association between low ovarian reserve and RPL risk. Different clinical question.

When AMH is reasonable: For fertility timeline counselling — but that's a separate conversation from miscarriage risk.
Source: ESHRE Rec 28

Fasting insulin / glucose

For RPL purpose

Don't · Strong

No prognostic value for RPL. Order independently if PCOS or diabetes is clinically suspected.

Note: BMI optimisation and lifestyle remain core advice regardless. Metformin has not been shown to reduce loss in RPL with glucose defects.
Source: ESHRE Rec 26, 54

Androgens / LH

Endocrine extras

Don't · Strong

No consistent association with RPL. No prognostic or therapeutic implication.

Source: ESHRE Rec 30–31

Homocysteine

Metabolic marker

Don't · Strong

Inconsistent evidence of association. Routine high-dose folate not recommended without hyperhomocysteinaemia.

Source: ESHRE Rec 32, 72

Vitamin D (as a test)

For RPL purpose

Don't · Strong

RANZCOG strong against routine testing. Empirical supplementation 1000–2000 IU daily is reasonable and safe.

Approach: Skip the bloods, give the supplement. ESHRE notes high prevalence of deficiency and suggests prophylactic vitamin D may be considered.
Source: RANZCOG R8 · ESHRE Rec 55

Luteal phase insufficiency

Endometrial dating / mid-luteal progesterone

Don't · Strong

Outdated concept in this context. No reliable test, no treatment shown to help RPL.

Source: ESHRE Rec 29