What the PROMISE and PRISM trials really tell us

Progesterone is one of the most commonly discussed treatments after miscarriage.

After two or three pregnancy losses, it is understandable for patients to ask:

“Should I take progesterone next time?”

And often the question is even more difficult:

“If progesterone is safe and might help, why wouldn’t I just take it?”

For many years, progesterone has been prescribed in early pregnancy because biologically it makes sense. Progesterone is essential for implantation and for maintaining the endometrium during early pregnancy.

But biological plausibility does not necessarily mean that giving additional progesterone will prevent every miscarriage.

Two major studies — PROMISE and PRISM — have helped us understand something much more nuanced:

Progesterone does not appear to benefit every woman with recurrent miscarriage. The clearest evidence of benefit is in women who have had previous miscarriages and who develop bleeding in a subsequent early pregnancy.

And the more previous miscarriages a woman has had, the more clinically meaningful that benefit may become.

What did the PROMISE trial show?

The PROMISE trial, published in the New England Journal of Medicine in 2015, was one of the landmark studies in recurrent miscarriage.

It included 836 women with unexplained recurrent miscarriage who conceived naturally. Women received either vaginal micronised progesterone 400 mg twice daily or placebo, beginning soon after a positive pregnancy test and continuing until 12 weeks.

The live birth rates were:

• 65.8% with progesterone
• 63.3% with placebo

The difference was not statistically significant.

In other words, giving progesterone routinely to every woman simply because she had a history of recurrent miscarriage did not significantly improve the chance of live birth.

That was an important finding.

It suggested that recurrent miscarriage alone may not be enough to identify the women who will benefit from progesterone.

Then came PRISM — and the story became more interesting

The much larger PRISM trial studied more than 4,000 women presenting with bleeding in early pregnancy.

Again, women received vaginal micronised progesterone 400 mg twice daily or placebo.

Looking at everybody together, the difference was relatively small:

• Live birth: 75% with progesterone versus 72% with placebo.

The primary analysis did not reach conventional statistical significance.

But when researchers examined women according to their previous miscarriage history, an important pattern emerged.

Among women who had experienced one or more previous miscarriages and were now bleeding, live birth was approximately:

• 75% with progesterone versus 70% with placebo.

And among women with three or more previous miscarriages who were bleeding in the current pregnancy, the difference was considerably larger:

• 72% with progesterone versus 57% with placebo.

That is an absolute difference of around 15 percentage points in this subgroup.

This finding has subsequently influenced clinical guidelines.

So when does progesterone actually help?

This is where I think the distinction is important.

There are three very different clinical situations.

1. Previous recurrent miscarriage, but no bleeding in the current pregnancy

The PROMISE trial does not provide convincing evidence that routinely prescribing progesterone improves live birth rates.

That does not mean progesterone can never be considered. It means we should be careful about telling patients that progesterone has been proven to prevent another miscarriage simply because they have had previous losses.

2. Bleeding in early pregnancy, but no previous miscarriage

The evidence also does not demonstrate a meaningful benefit from progesterone in this group.

3. Previous miscarriage AND bleeding in the current pregnancy

This is where the evidence becomes much more persuasive.

The PRISM data suggest a live-birth benefit in women who have experienced previous miscarriage and present with early pregnancy bleeding, with the largest observed effect in women who have experienced three or more previous miscarriages.

The 2023 ESHRE recurrent pregnancy loss guideline therefore states that vaginal progesterone may improve live birth rates in women with three or more pregnancy losses who experience vaginal bleeding in a subsequent pregnancy.

Similarly, current NICE guidance recommends vaginal micronised progesterone 400 mg twice daily for women with a scan-confirmed intrauterine pregnancy who have early pregnancy bleeding and have previously experienced a miscarriage.

Why might bleeding identify a group that responds differently?

We don’t completely know.

One possibility is that early pregnancy bleeding may identify pregnancies in which the interaction between the developing pregnancy, endometrium and hormonal environment is more vulnerable — and where additional progesterone support may therefore have a greater opportunity to make a difference.

But we should be careful not to interpret low progesterone as necessarily being the cause of a miscarriage.

Sometimes progesterone is low because a pregnancy is already failing.

This distinction matters.

Progesterone cannot correct every reason a pregnancy miscarries.

It cannot make a significantly chromosomally abnormal embryo normal, for example, and recurrent miscarriage can have multiple potential contributing factors.

What do I do in clinical practice?

For me, progesterone is not a substitute for properly assessing recurrent pregnancy loss.

When I see someone who has experienced repeated miscarriages, I want to understand the whole clinical story:

• How many pregnancies have been lost?
• At what gestation?
• Was fetal cardiac activity previously seen?
• Were any pregnancy tissues genetically tested?
• What is her age?
• Is there a uterine abnormality?
• Are antiphospholipid antibodies relevant?
• Is there a thyroid or other endocrine issue?
• Was conception spontaneous or through fertility treatment?
• And importantly — is she bleeding in this pregnancy?

Only then does the progesterone question sit in its proper context.

I also think it is important to acknowledge the emotional side of this decision.

After several miscarriages, being told simply to “try again” can be incredibly difficult. Patients understandably want to know whether there is anything they can do differently.

My approach is to explain what progesterone can do, what it has not been shown to do, and where the evidence is strongest, and then make an individual decision together.

What about IVF pregnancies?

This is an important distinction.

Progesterone used as luteal support after IVF or frozen embryo transfer is a different clinical question from progesterone prescribed to prevent miscarriage after spontaneous conception.

For example, after certain frozen embryo transfer protocols the ovaries may not produce the normal corpus luteum progesterone required to support early pregnancy. Progesterone supplementation in that setting is part of the fertility treatment itself.

The PROMISE and PRISM results should therefore not simply be extrapolated to all IVF pregnancies.

Is progesterone safe?

The large trials have been reassuring.

PROMISE did not demonstrate a significant increase in adverse outcomes, and subsequent evidence reviews have not identified an important increase in congenital abnormalities or maternal adverse events with vaginal micronised progesterone.

However, absence of evidence of harm does not mean every medication should automatically be prescribed.

The better question is always:

Is there good evidence that this treatment is likely to help this particular patient?

My take-home message

Progesterone is a good example of why fertility and miscarriage treatment should not become a checklist of medications.

The evidence does not support the simple equation:

Recurrent miscarriage = progesterone.

Instead, PROMISE and PRISM suggest something more useful:

• Previous miscarriage alone: routine progesterone has not convincingly improved live birth.
• Early pregnancy bleeding alone: benefit has not been demonstrated.
• Previous miscarriage + bleeding in the current pregnancy: progesterone may improve the chance of live birth.
• Three or more previous miscarriages + current pregnancy bleeding: this appears to be the group in whom the evidence of benefit is strongest.

For patients who have already experienced pregnancy loss, those distinctions matter.

The goal is not to prescribe everything that might help.

It is to identify the treatment most likely to help this woman, in this pregnancy, at this point in her journey.

Frequently Asked Questions

Should I take progesterone after two miscarriages?

Not automatically. Previous miscarriage by itself does not clearly identify a group in whom progesterone improves live birth. However, if you develop bleeding during a subsequent early pregnancy, the evidence supporting vaginal progesterone becomes stronger.

Does progesterone prevent miscarriage if I start it as soon as I get a positive pregnancy test?

The PROMISE trial investigated this approach in women with recurrent miscarriage and did not demonstrate a statistically significant improvement in live birth.

What progesterone was used in the PRISM trial?

PRISM used vaginal micronised progesterone 400 mg twice daily.

If I have bleeding after previous miscarriages, when should I contact my doctor?

Early pregnancy bleeding should be assessed rather than simply treated at home with progesterone. An ultrasound may be needed to establish the location and viability of the pregnancy and to exclude conditions such as ectopic pregnancy.

Does progesterone help after IVF?

Progesterone has an established role as luteal-phase support in many assisted reproduction protocols. That is a different indication from using progesterone to prevent recurrent miscarriage after spontaneous conception, so the evidence should not be confused.