The New England Journal of Medicine May 09, 2019

A Randomized Trial of Progesterone in Women with Bleeding in Early Pregnancy

Arri Coomarasamy et al.

Bottom Line

In women with bleeding in early pregnancy, first-trimester vaginal progesterone did not significantly improve the overall rate of live births compared to placebo, although prespecified subgroup analyses indicated a significant benefit for women with a history of prior miscarriages.

Key Findings

1. Overall Primary Outcome: The incidence of live births after at least 34 weeks of gestation was 75% (1,513 of 2,025 women) in the progesterone group compared to 72% (1,459 of 2,013 women) in the placebo group, a difference that was not statistically significant (Relative Rate [RR] 1.03; 95% CI, 1.00 to 1.07; P=0.08).

2. Subgroup with ≥3 Prior Miscarriages: In a prespecified subgroup of women with bleeding and a history of three or more prior miscarriages, progesterone significantly increased the live birth rate to 72% (98 of 137 women) compared to 57% (85 of 148 women) with placebo, representing a 15% absolute increase.

3. Subgroup with 1-2 Prior Miscarriages: Among women with bleeding and 1 or 2 previous miscarriages, the live birth rate was 76% in the progesterone group compared to 72% in the placebo group, reflecting a more modest benefit.

4. Subgroup with No Prior Miscarriages: For women with bleeding but no history of miscarriage, the live birth rates were nearly identical (74% vs. 75%), demonstrating no clinical benefit from progesterone.

5. Safety: The incidence of maternal adverse events and congenital anomalies did not differ significantly between the progesterone and placebo groups.

Study Design

Design

RCT

Double-Blind

Sample

4,153

Patients

Duration

Live birth

Median

Setting

Multicenter, UK

Population Women aged 16 to 39 years presenting with vaginal bleeding in early pregnancy (less than 12 weeks of gestation) and a visible intrauterine gestational sac on ultrasound.

Intervention Micronized vaginal progesterone (400 mg) administered twice daily as a suppository from clinical presentation up to 16 weeks of gestation.

Comparator Matched placebo vaginal suppositories administered twice daily from clinical presentation up to 16 weeks of gestation.

Outcome Incidence of live birth after at least 34 weeks of gestation.

Study Limitations

• The primary trial outcome was formally negative across the entire unselected cohort; the clinically actionable findings supporting progesterone use are derived entirely from subgroup analyses (albeit prespecified).

• The intervention was specifically micronized vaginal progesterone; results cannot necessarily be extrapolated to other formulations like oral or intramuscular progesterone, or synthetic progestins like dydrogesterone.

• Therapy could only be initiated after a patient presented with bleeding and had a confirmed intrauterine gestational sac on ultrasound, potentially missing a critical therapeutic window very early in implantation.

• A substantial proportion of eligible women declined participation, which may have introduced selection bias and limited the broader generalizability of the cohort.

Clinical Significance

The PRISM trial fundamentally shifted clinical guidelines regarding threatened miscarriage. While it proved that universally prescribing progesterone to all women with early pregnancy bleeding is unjustified, the strong subgroup data demonstrated clear benefit for women with both early bleeding and a history of miscarriage. As a direct result of this trial, major bodies like the UK National Institute for Health and Care Excellence (NICE) updated their guidelines to explicitly recommend a course of vaginal micronized progesterone (400 mg twice daily) for women experiencing threatened miscarriage who have had at least one prior miscarriage.

Historical Context

The utility of progesterone in sustaining early pregnancy has been vigorously debated for over 60 years. Prior to PRISM, the same research group conducted the PROMISE trial (2015), which evaluated progesterone in women with a history of unexplained recurrent miscarriage but no current bleeding, and found no benefit. However, the scenario of 'threatened miscarriage' (active bleeding) remained controversial, with many clinicians continuing to empirically prescribe progesterone based on underpowered, low-quality legacy studies. PRISM was funded by the UK National Institute for Health Research to definitively settle this debate, serving as the largest and most rigorously conducted trial on the subject to date.

Guided Discussion

High-yield insights from every perspective

Med Student

Medical Student

Why is progesterone biologically necessary to maintain early pregnancy, and what is the physiological source of this hormone before and after the luteal-placental shift?

Key Response

Progesterone maintains the secretory endometrium and suppresses myometrial contractility. Initially, it is produced by the corpus luteum under the stimulation of hCG. Around 8-10 weeks of gestation, the luteal-placental shift occurs, and the developing placenta becomes the primary source. Understanding this is crucial for the rationale behind first-trimester supplementation in threatened miscarriage.

Resident

A patient presents with vaginal bleeding at 7 weeks gestation and reports a history of two prior miscarriages. Based on the PRISM trial, how should you counsel her regarding vaginal progesterone, and what is the specific dosage?

Key Response

Residents must recognize that while the PRISM trial was negative for the overall population, the prespecified subgroup of women with bleeding AND a history of prior miscarriages showed a significant increase in live births. The recommended dosage based on the trial is 400 mg of micronized progesterone vaginally twice daily until 16 weeks of gestation.

Fellow

How do the findings of the PRISM trial regarding threatened miscarriage contrast with those of the PROMISE trial regarding unexplained recurrent miscarriage, and what does this suggest about the underlying pathophysiology of early pregnancy loss?

Key Response

The PROMISE trial showed no benefit for progesterone in women with recurrent miscarriage but NO bleeding. PRISM showed benefit in women with prior miscarriages WHO PRESENT with bleeding. Fellows should recognize that this suggests progesterone may not fix a baseline chronic defect in all recurrent miscarriage patients, but may act as a rescue therapy during an acute threatened miscarriage, possibly by modulating acute inflammatory cascades or overcoming acute corpus luteum insufficiency triggered by the bleeding event.

Attending

Given that the primary outcome of the PRISM trial was statistically non-significant (p=0.08), how do you ethically and accurately frame shared decision-making with a highly anxious patient presenting with threatened miscarriage without overstating the evidence?

Key Response

Attendings must navigate the nuance of a negative primary trial with a positive subgroup. The rationale highlights the importance of explaining that for women without a history of miscarriage, progesterone does not significantly change outcomes. For those with a history, there is a measurable benefit. It emphasizes teaching trainees how to balance statistical rigor with empathetic, pragmatic patient care, discussing the low risk of teratogenicity alongside the realistic chances of success.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD

The PRISM trial investigators emphasized a positive effect in a prespecified subgroup despite a non-significant primary outcome. From a statistical and methodological standpoint, what criteria must be met for a subgroup analysis to be considered credible, and how does the test for interaction play a role here?

Key Response

Methodologically, subgroup analyses in negative trials are highly prone to false positives. To be credible, they must be prespecified, biologically plausible, powered adequately, and most importantly, demonstrate a significant p-value for interaction, showing that the treatment effect genuinely differs between the subgroups (e.g., prior miscarriage vs. no prior miscarriage) rather than just looking at the p-value within the subgroup alone.

Journal Editor

As a peer reviewer, how would you evaluate the authors' decision to highlight a positive subgroup finding in the abstract of a manuscript where the primary outcome failed to reach statistical significance, and what editorial constraints would you impose to prevent 'spin'?

Key Response

Editors must rigorously guard against 'spin,' which occurs when authors emphasize secondary positive findings to obscure a negative primary outcome. A tough reviewer would require the primary negative result to be the focal point of the conclusion, mandate that the subgroup finding be explicitly labeled as 'prespecified but secondary' or 'hypothesis-generating,' and ensure that the abstract accurately reflects the overall statistical failure of the primary endpoint.

Guideline Committee

Following the publication of the PRISM trial, the UK National Institute for Health and Care Excellence (NICE) updated its guideline (NG126) on ectopic pregnancy and miscarriage. How did the committee translate the technically 'negative' overall PRISM trial into a positive practice recommendation, and what specific patient population was targeted?

Key Response

NICE updated NG126 to recommend 400 mg twice daily vaginal progesterone for women with an intrauterine pregnancy confirmed by scan, vaginal bleeding, and a history of previous miscarriage. The committee justified this by determining that the prespecified subgroup evidence was highly robust, biologically plausible, and demonstrated a favorable cost-effectiveness and safety profile, illustrating how guideline committees sometimes prioritize strong, actionable subgroup data over a strict adherence to primary outcome overall significance.

Clinical Landscape

Noteworthy Related Trials

2010

ALIFE Trial

n = 364 · NEJM

Tested

Aspirin combined with nadroparin or aspirin alone

Population

Women with unexplained recurrent miscarriage

Comparator

Placebo

Endpoint

Live birth rate

Key result: Neither aspirin combined with nadroparin nor aspirin alone improved the live-birth rate among women with unexplained recurrent pregnancy loss.

2015

PROMISE Trial

n = 836 · NEJM

Tested

Vaginal micronized progesterone

Population

Women with unexplained recurrent miscarriages

Comparator

Placebo

Endpoint

Live birth after 24 weeks of gestation

Key result: Progesterone therapy did not result in a significantly higher rate of live births among women with unexplained recurrent miscarriages compared to placebo.

2020

MifeMiso Trial

n = 711 · Lancet

Tested

Mifepristone followed by misoprostol

Population

Women with missed miscarriage

Comparator

Misoprostol alone

Endpoint

Resolution of miscarriage within 7 days

Key result: Pretreatment with mifepristone followed by misoprostol resulted in a significantly higher rate of resolution of missed miscarriage than misoprostol alone.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis