Recurrent Miscarriage Workup: What the Evidence Shows Actually Makes a Difference

Uterine cavity assessment is recommended as part of the standard RPL workup by all major guidelines, with the SOGC 2025 guideline rating screening for anatomic factors as a strong recommendation with high-quality evidence. The initial assessment typically involves pelvic ultrasound, with saline infusion sonography, hysteroscopy, or MRI used for further characterisation when anomalies are suspected.

The uterine septum is the most common congenital anomaly associated with RPL, and hysteroscopic septum resection is considered reasonable in women with RPL (SOGC: conditional recommendation, low-quality evidence). However, no RCTs definitively demonstrate that septum resection improves live birth rates — the evidence is based primarily on observational comparisons. Intrauterine adhesions (Asherman syndrome) have a clearer treatment pathway: hysteroscopic adhesiolysis is recommended to restore normal cavity anatomy and improve the likelihood of conception (SOGC: strong recommendation, high-quality evidence).

Submucous fibroids distorting the cavity are generally accepted as contributing to RPL, and hysteroscopic myomectomy is considered appropriate. Intramural fibroids without cavity involvement and uterine anomalies such as unicornuate, bicornuate, or didelphys uteri have weaker associations with RPL, and surgical intervention for these variants lacks robust supporting evidence. The key clinical distinction is between anomalies that distort the endometrial cavity — which have the strongest association with loss — and those that alter uterine shape without affecting the implantation surface.

Thyroid dysfunction — both overt hypothyroidism and hyperthyroidism — is associated with increased miscarriage risk, and thyroid function testing (TSH, with free T4 if abnormal) is recommended as part of the standard RPL workup. Treatment of overt thyroid dysfunction to achieve euthyroidism before and during pregnancy is uncontroversial and supported by consistent observational evidence.

The more clinically nuanced question is whether thyroid peroxidase (TPO) antibody positivity in euthyroid women warrants levothyroxine treatment to prevent miscarriage. The T4LIFE trial — a multicentre, randomised, double-blind, placebo-controlled trial — specifically addressed this question in euthyroid TPO-antibody-positive women with RPL and found no reduction in miscarriage rate with levothyroxine compared to placebo. This result significantly weakens the rationale for treating TPO-antibody positivity alone in the absence of abnormal TSH.

Current guidelines recommend screening for thyroid dysfunction but do not recommend levothyroxine treatment for euthyroid women with TPO antibodies and RPL. Subclinical hypothyroidism (TSH above the upper reference limit but below 10 mIU/L with normal free T4) remains a grey area, with some guidelines supporting treatment to a TSH below 2.5 mIU/L in the first trimester — though the evidence base for this threshold is limited and evolving.

Progesterone supplementation for RPL has been debated for decades, and the evidence has crystallised around a specific clinical scenario. A Cochrane network meta-analysis (Coomarasamy et al. 2021) found that vaginal progesterone may reduce miscarriage in women who present with vaginal bleeding in early pregnancy and have a history of RPL. The SOGC 2025 guideline notes that progesterone therapy may be helpful, especially when started in the luteal phase.

The PRISM trial (progesterone in spontaneous miscarriage) — a large RCT of 4,153 women — found a small, non-significant overall benefit of vaginal progesterone for women with early pregnancy bleeding, but a pre-specified subgroup analysis showed a more substantial effect in women with a history of three or more losses (live birth rate 72% vs. 57%, RR 1.28, 95% CI 1.08–1.51). While subgroup analyses must be interpreted cautiously, this finding has influenced guideline recommendations supporting progesterone use in this specific population.

The form of progesterone matters: vaginal micronised progesterone (400 mg twice daily) is the formulation studied in the PRISM trial. Oral dydrogesterone has also shown benefit in some trials, though the evidence base is less extensive. Intramuscular progesterone and synthetic progestins do not have equivalent evidence. Importantly, progesterone is not recommended for universal miscarriage prevention in women without a history of RPL or current bleeding symptoms.

Immunological theories of RPL have generated substantial patient interest and commercial testing activity, but the evidence base remains insufficient to support routine immunological workup. The SOGC 2025 guideline is unequivocal: immune testing should not be a routine part of the RPL workup (strong recommendation, high-quality evidence). Immunomodulation treatments — including IVIG, intralipid infusions, TNF-alpha inhibitors, and lymphocyte immunotherapy — are not recommended (strong recommendation, high-quality evidence).

Specifically, peripheral blood natural killer (NK) cell testing — one of the most commonly ordered tests in the commercial RPL space — has no validated relationship to endometrial or decidual NK cell function and does not predict pregnancy outcome. Uterine NK cells, which differ phenotypically and functionally from peripheral NK cells, play a role in spiral artery remodelling and placentation, but there is no standardised method for their quantification or clinical interpretation. The ESHRE 2023 update similarly concludes that NK cell testing should not be offered as part of RPL investigation.

Emerging research on chronic endometritis (diagnosed by hysteroscopy and CD138 immunostaining of endometrial biopsies) represents a potentially actionable area, with the SOGC 2025 guideline noting that chronic endometritis may benefit from antibiotic treatment. However, the evidence is preliminary, diagnostic standardisation is lacking, and no large RCTs have confirmed that treatment of chronic endometritis improves live birth rates in RPL specifically. This remains an area of active investigation rather than established practice.

After a thorough, evidence-based workup, more than 50% of RPL cases will have no identifiable cause. This is, paradoxically, often the best prognostic category. Longitudinal studies consistently show that women with unexplained RPL have a 60–75% chance of a successful subsequent pregnancy with supportive care alone — defined as early pregnancy monitoring, access to a dedicated early pregnancy clinic, and emotional support from informed clinicians.

The therapeutic implications of the "unexplained" category are important: in the ALIFE trial (NEJM 2010), neither aspirin plus LMWH nor aspirin alone improved live birth rates compared to placebo in women with unexplained RPL. This means that the most commonly prescribed empirical treatments for unexplained RPL — anticoagulants — do not work. The psychological impact of receiving a negative workup can be distressing for couples, and clinicians should frame the result explicitly: the absence of an identifiable cause is associated with a favourable prognosis, not an unresolvable problem.

Lifestyle modifications with some evidence of benefit include reducing caffeine intake, smoking cessation, avoiding alcohol, and maintaining a healthy BMI. There is evidence supporting a Mediterranean dietary pattern (from the Nurses' Health Study II) for improved pregnancy outcomes. The SOGC 2025 guideline emphasises the importance of trauma-informed, compassionate care and access to early pregnancy assessment clinics as components of management for all couples with RPL, regardless of aetiology.