Miscarriage and pregnancy loss statistics: how common it is and why most people underestimate it

She came in at nine weeks for a dating scan. On the screen there was a gestational sac measuring about seven weeks and no cardiac activity. She asked me what she had done. Not what happened, or what comes next. What she had done. She had a list ready: the glass of wine before she knew, the flight to Denver, a heavy box carried up two flights of stairs.

None of those caused the loss. I told her that, and she nodded, and I could see she did not believe me. Almost nobody does at first.

Miscarriage is the most common complication of early pregnancy and one of the least understood by the people it happens to. The pooled global risk is 15.3% of clinically recognized pregnancies, per the 2021 Lancet "Miscarriage matters" series led by Siobhan Quenby. A national survey in Obstetrics & Gynecology, meanwhile, found 55% of Americans believe it occurs in 5% or less of pregnancies. That gap is where most of the guilt lives.

This article covers what the evidence currently shows about how often pregnancy loss happens, how risk shifts with age and with previous losses, what causes it, what it does to mental health, and what changed in stillbirth data and miscarriage care during 2026.

How common is miscarriage

The most defensible global number comes from the 2021 Lancet Miscarriage Matters series, which pooled data across studies and reported a risk of 15.3% of all recognized pregnancies, with a confidence interval of 12.5% to 18.7%. Roughly one in six or seven pregnancies a woman knows about will end this way.

March of Dimes, which is what most American patients actually find when they search, gives a wider range: 10 to 20 in 100 known pregnancies. The two figures are not in conflict. The wider band reflects real variation by population and by how early pregnancies get detected.

That last point is where the numbers get slippery. Every one of these estimates counts recognized pregnancies, so losses that occur before a missed period are not in the denominator. Sensitive early testing has pushed detection earlier, which means women today learn about losses that a generation ago would have registered as a slightly late period. The true biological rate of conception failure is higher than 15.3%, but nobody can put a clean figure on it.

The practical consequence: a woman charting closely with our period calculator and testing early will become aware of more losses than a woman who tests at six weeks. She is not having more miscarriages. She is seeing more of them.

When losses happen

Timing is the single most useful thing to understand about miscarriage risk, because it changes so fast.

According to March of Dimes, 8 out of 10 miscarriages, or 80%, happen in the first trimester before the 12th week. That concentration is why the twelve-week mark became the informal moment for announcing a pregnancy, a convention I have mixed feelings about, since it means many women grieve in the exact window when nobody knows they were pregnant.

Once an ultrasound confirms fetal cardiac activity in a woman with no bleeding or pain, the picture improves quickly. A cohort study by Tong and colleagues in Obstetrics & Gynecology followed asymptomatic women after a normal first-trimester visit and found the subsequent miscarriage risk was 9.4% at six completed weeks of gestation, 4.2% at seven weeks, 1.5% at eight weeks, 0.5% at nine weeks and 0.7% at ten weeks. Across the entire cohort, 11 of 696 pregnancies were lost, a rate of 1.6%.

I quote those numbers often, with a caveat: the six-week figure rests on a smaller subgroup, so the drop from 9.4% to 4.2% is steeper on paper than in reality. The study also enrolled asymptomatic women. Bleeding or a scan measuring behind dates changes the estimate, sometimes substantially.

If your dates are uncertain, the gestational age you are quoted may be off by a week either way, and a week matters here. Women who conceived through IVF have an advantage, since the transfer date is known exactly and our IVF due date calculator reflects that.

Miscarriage risk by age

The clearest data on age comes from a 2019 BMJ study by Magnus and colleagues, which used Norwegian population registers to follow 421,201 pregnancies. Risk traced a J-shaped curve: elevated in the youngest mothers, lowest among women aged 25 to 29 at 9.8%, with the absolute minimum at age 27 at 9.5%, then climbing steadily. By 45 and over, 53.6% of recognized pregnancies ended in miscarriage.

More than half. Read that carefully, though: the denominator is recognized pregnancies rather than attempts to conceive, and the oldest age band rests on fewer pregnancies than the middle of the curve. The steepness of the trend matches what we know about oocyte aneuploidy.

Registry data of this kind captures whole populations rather than clinic patients, which avoids the selection bias that inflates numbers in fertility-center studies. Our article on fertility and age statistics covers the conception side of the same curve.

Recurrent pregnancy loss

One miscarriage is common. Three is not.

The Lancet series put population prevalence at 10.8% of women having had one miscarriage, 1.9% two, and 0.7% three or more. Roughly one in nine women has experienced a single loss; fewer than one in a hundred has had three.

The same BMJ register study showed that risk is not independent from one pregnancy to the next. Compared with women who had no previous loss, the adjusted odds ratio was 1.54 after one miscarriage (95% CI 1.48 to 1.60), 2.21 after two, and 3.97 after three or more. Roughly a quadrupling. That is the statistical basis for evaluating women after repeated losses rather than telling them to keep trying.

Who qualifies for that evaluation changed this year. The American Society for Reproductive Medicine issued its first update to recurrent pregnancy loss guidance in fourteen years (Fertility and Sterility 2026;125:1023-1041), redefining the condition as two or more losses, dropping the requirement that they be consecutive or ultrasound-confirmed, and formally including biochemical losses. It also moves away from broad, reflexive testing panels.

This widens the door. A woman with a biochemical loss in 2023 and a seven-week loss in 2026 now meets criteria for workup, where the old framework told her to wait for a third. Whether insurers follow promptly is another question, and in my experience they rarely do.

Why miscarriages happen

Most early losses are genetic accidents in the embryo, and they are not caused by anything the pregnant person did.

A 2023 Nature Medicine analysis of first-trimester products of conception classified 879 samples, or 50.4%, as karyotypically abnormal, against 866 (49.6%) karyotypically normal. Standard karyotyping misses smaller structural changes, though. With higher-resolution sequencing and extrapolation, the rate of genomic aberrations reached 67.8%. Two thirds.

ASRM's 2026 committee opinion frames it in clinical terms: approximately 50% to 60% of first-trimester miscarriages are due to embryonic aneuploidy, and the share climbs sharply with maternal age, appearing in roughly 50% of miscarriages tested in women under 35 compared with 75% of miscarriages in women over 40. This is the mechanistic explanation underneath the age curve from the BMJ data. Older eggs produce more chromosomally abnormal embryos, and most chromosomally abnormal embryos do not continue.

I say a version of this to every patient who hands me a list of things she thinks she did wrong. The exercise did not do it, and neither did the flight. In most cases the pregnancy was not viable from the moment of fertilization, and no different behavior would have changed the outcome.

The perception gap

Here is the finding that explains much of the private suffering around this topic. In a national survey of 1,084 US adults published in Obstetrics & Gynecology by Bardos and colleagues, 55% of respondents believed miscarriage occurred in 5% or less of all pregnancies. The real figure is roughly three times that.

When you believe something happens to 5% of people and it happens to you, the natural inference is that you are the anomaly and that you caused it. The same survey measured exactly that. Among people who had experienced a miscarriage, 47% felt guilty, 41% felt alone, 28% felt ashamed, and 37% felt they had lost a child.

Grief after pregnancy loss is real grief, and no statistic dissolves it. The isolation is a different matter. That part is manufactured by silence, and silence is the one thing medicine and public communication can actually fix.

Mental health after loss

The psychological aftermath of early pregnancy loss is understudied, and the best prospective data suggests we have been underestimating it badly.

Farren and colleagues, writing in the American Journal of Obstetrics & Gynecology in 2020, followed women after early pregnancy loss and ectopic pregnancy. Criteria for post-traumatic stress were met in 29% of the early pregnancy loss group after one month and in 18% after nine months. Moderate or severe anxiety was reported in 24% at one month and 17% at nine months.

Nine months later, nearly one in five still met criteria for post-traumatic stress. That is a persistent psychiatric outcome following an event the medical system usually treats as routine and self-limited, warranting no follow-up beyond a repeat blood test.

Standard care after a first-trimester loss in the United States runs to a discussion of management options and a visit to confirm resolution. Mental health screening is not part of the default pathway in most practices. Given what universal screening for postpartum depression catches, the absence of any equivalent after pregnancy loss is hard to defend.

States have begun filling the gap with reproductive-loss bereavement leave. California's SB 848 mandates five days of unpaid leave after miscarriage, stillbirth, or failed IVF. Washington's SB 5217 expands paid family and medical leave bereavement to seven paid days starting July 1, 2026. New York has two bills pending in its 2025-26 session. No federal counterpart exists, so whether a woman gets time off after a loss depends on her ZIP code.

Stillbirth in the United States

Later loss is counted differently and tracked separately. In 2023, 20,005 fetal deaths at 20 weeks of gestation or more were reported in the United States, a fetal mortality rate of 5.53 per 1,000, according to the CDC/NCHS National Vital Statistics Report.

The disparities in that dataset are more than twofold. The rate was highest for Native Hawaiian or Other Pacific Islander non-Hispanic women at 10.18 per 1,000 and Black non-Hispanic women at 9.95, lowest for Asian non-Hispanic women at 4.14.

Those patterns track closely with what we see in maternal mortality data, and for overlapping reasons involving access, quality of prenatal care, and untreated chronic conditions.

The national count itself is now contested. A 2025 JAMA analysis by Sullivan and colleagues examined 2.7 million pregnancies and found that more than 1 in 150 births end in stillbirth, higher than the long-cited 1 in 175. In low-income communities it was 1 in every 112 births, and in areas with higher proportions of Black families, 1 in every 95. Nearly 27.7% occurred in pregnancies with no identified risk factor, which undercuts the idea that risk-based screening alone can catch them.

Provisional numbers improved the following year. NCHS Vital Statistics Rapid Release No. 41, published in December 2025, reported a provisional 2024 rate of 5.41 per 1,000 with 19,756 fetal deaths, down 2%, and a late-fetal rate down 4%. No significant change appeared for any race or Hispanic-origin group. The overall number moved; the gap did not.

Part of the problem is measurement. US fetal death certificates are incomplete and inconsistently filed across states, which is what the bipartisan SHINE for Autumn Act (H.R.5469 and S.2858, introduced September 18, 2025) is meant to address, authorizing $5 million a year for FY2026 through FY2030 in state grants for stillbirth data collection. It remains pending. Our overview of healthcare access disparities covers the structural conditions these rates sit inside.

How abortion bans changed miscarriage care

The most consequential recent finding in this field is not about biology. It is about what happens to miscarriage treatment when the law changes around it.

An Oregon Health & Science University study published in JAMA on May 18, 2026 analyzed 123,598 people who miscarried before ten weeks between 2018 and 2024. In states with abortion bans, researchers found a 2.8% increase in expectant management (waiting for the loss to complete on its own), a 2.2% decrease in medication management, and a 13.8% increase in misoprostol-only regimens, which are less effective than mifepristone plus misoprostol.

The mechanism is straightforward, if grim. The drugs used to complete a miscarriage are the same drugs used for medication abortion. When pharmacies and clinicians face legal exposure for dispensing them, the friction lands on miscarriage patients, who wait longer or get a regimen with a higher failure rate. Failed regimens mean more bleeding and more procedures that were avoidable.

Percentage shifts of this size sound small. Applied to a cohort of 123,598, they are not. The study window closed in 2024, before several of the most restrictive state laws took effect.

If you are managing an early loss right now, ask your clinician which options this facility can offer and whether mifepristone is being used alongside misoprostol. Patients should not have to know to ask that. In much of the country in 2026, they do.

Frequently asked questions about miscarriage

How common is miscarriage?

Miscarriage ends 15.3% of clinically recognized pregnancies worldwide (95% CI 12.5-18.7%), according to the 2021 Lancet Miscarriage Matters series. March of Dimes gives the range as 10 to 20 in 100 known pregnancies. Both figures count only pregnancies that were detected, so the true rate of very early loss is higher and difficult to measure precisely.

When do most miscarriages happen?

About 8 out of 10 miscarriages, or 80%, occur in the first trimester before the 12th week, per March of Dimes. Risk drops quickly once cardiac activity is confirmed in a woman without symptoms: a study in Obstetrics & Gynecology found 9.4% at six weeks, 4.2% at seven, 1.5% at eight, and 0.5% at nine weeks.

How does maternal age affect miscarriage risk?

Risk follows a J-shaped curve. In a BMJ register study of 421,201 Norwegian pregnancies, risk was lowest among women aged 25 to 29 at 9.8%, with the minimum at age 27 at 9.5%, then rose to 53.6% at age 45 and older. The driver is embryonic aneuploidy, which increases with egg age.

What causes most miscarriages?

Chromosomal abnormalities in the embryo. A 2023 Nature Medicine analysis found 50.4% of first-trimester losses were karyotypically abnormal, rising to 67.8% with higher-resolution sequencing. ASRM's 2026 guidance attributes 50% to 60% of first-trimester miscarriages to embryonic aneuploidy, found in about 50% of losses in women under 35 and 75% in women over 40.

How many women have recurrent miscarriages?

Per the Lancet series, 10.8% of women have had one miscarriage, 1.9% two, and 0.7% three or more. Risk climbs with each loss: BMJ data shows adjusted odds ratios of 1.54 after one prior miscarriage, 2.21 after two, and 3.97 after three or more. ASRM redefined recurrent pregnancy loss in 2026 as two or more.

Is mental health support recommended after pregnancy loss?

The evidence says it should be. A 2020 American Journal of Obstetrics & Gynecology study found 29% of women met criteria for post-traumatic stress one month after early pregnancy loss and 18% still did at nine months. Moderate or severe anxiety affected 24% at one month. Routine mental health screening after loss is not standard practice in most US clinics.

How common is stillbirth in the United States?

The CDC/NCHS reported 20,005 fetal deaths at 20 weeks or more in 2023, a rate of 5.53 per 1,000. A 2025 JAMA analysis of 2.7 million pregnancies put stillbirth at more than 1 in 150 births, higher than the cited 1 in 175, and 1 in 112 in low-income communities. Nearly 27.7% had no identified risk factor.