Hyperemesis gravidarum statistics: prevalence, recurrence, and the GDF15 finding

She was 8 weeks along when she stopped keeping water down. By triage she had lost 11 pounds from her prepregnancy weight, could not stand without retching, and had been told twice that "morning sickness is normal." Her ketones were large, her potassium was low, and she needed IV fluids before she could swallow a tablet. That is hyperemesis gravidarum, the severe end of nausea and vomiting of pregnancy. It still gets minimized until a woman is already dehydrated.

The clinical definition is not subtle once you look for it: persistent vomiting not explained by another disease, signs of starvation, and usually weight loss of at least 5% of prepregnancy weight, per ACOG Practice Bulletin No. 189. What changed the public conversation is a 2023 Nature paper linking fetal production of the hormone GDF15, and a mother's sensitivity to it, to who develops severe disease. For the first time, reporters and patients have a molecular story that matches the lived one.

This article gathers the numbers a clinician or journalist can cite with a source attached. It covers prevalence, hospital and emergency department use, recurrence risk, the GDF15 evidence, work and termination data, and where treatment still falls short.

Source	Population	Measure	Estimate
ACOG Practice Bulletin No. 189 (2018)	Clinical guideline synthesis	HG incidence	0.3-3% of pregnancies
Fiaschi et al., Human Reproduction (2016)	8,215,538 pregnancies, England	Hospital admission for HG	1.5% of pregnancies (121,885 pregnancies; 186,800 admissions)
Gardner et al., Am J Epidemiol (2026)	2,476,492 California single births, 2007-2011	ED or inpatient HG care	2.2% (53,681 mothers)
ACOG Practice Bulletin No. 189 (2018)	Clinical guideline synthesis	Any nausea of pregnancy	50-80%

Sources: ACOG Practice Bulletin No. 189 (2018); Fiaschi L et al. Hum Reprod. 2016; Gardner RM et al. Am J Epidemiol. 2026 (Stanford Medicine summary). Differences reflect definition (clinical HG vs hospital coding vs ED/inpatient encounters) as much as true geography.

Metric	2006	2014	Change
ED visits with primary diagnosis of HG	177,940	227,298	+28%
Mean ED cost per visit (2021 USD)	$2,156	$3,549	+65%
Aggregate ED cost (2021 USD)	$384 million	$807 million	+110%
Share of HG ED visits admitted	7.7%	4.5%	Admissions down ~42%

Source: Geeganage G et al. Emergency department burden of hyperemesis gravidarum in the United States from 2006 to 2014. AJOG Glob Rep. 2023;3(1):100166. Costs inflation-adjusted to 2021 dollars. NEDS/HCUP data.

How common is hyperemesis gravidarum

According to ACOG Practice Bulletin No. 189 (2018), nausea of pregnancy affects 50% to 80% of pregnant people, and vomiting or retching about 50%. Most of that is self-limited "morning sickness." Hyperemesis gravidarum is the severe end. ACOG places incidence at approximately 0.3% to 3% of pregnancies, with the spread driven by different diagnostic criteria and study populations. There is no single universally accepted definition. Criteria most often cited include persistent vomiting not explained by other causes, a measure of acute starvation (commonly large ketonuria), and weight loss, most often at least 5% of prepregnancy weight.

Population studies that require a hospital encounter land in the middle of that range. Fiaschi, Nelson-Piercy and Tata analyzed English Hospital Episode Statistics for 8,215,538 pregnancies among 5,329,101 women (1997-2012), with 186,800 HG admissions during 121,885 pregnancies. Prevalence of HG requiring hospital admission was 1.5%. That is an operational count of people sick enough for secondary care coding, not every woman retching at home.

A 2026 California analysis by Gardner and colleagues at Stanford Medicine, in the American Journal of Epidemiology, covered 2,476,492 single births from 2007 to 2011. Of those mothers, 53,681 (2.2%) received emergency department or inpatient care coded for HG. Including ED care captures people treated acutely without overnight admission, so the rate sits above English inpatient-only estimates. StatPearls' 2025 review puts a global point estimate near 1.1%, still inside the ACOG band. Even 1% of U.S. births is tens of thousands of severely ill pregnant people each year.

Hospitalization and emergency department burden

ACOG states that HG is the most common indication for hospital admission in the first part of pregnancy and second only to preterm labor for hospitalization during pregnancy overall. That ranking still surprises families told for weeks that nausea is expected.

Geeganage and colleagues analyzed the Nationwide Emergency Department Sample for 2006 and 2014. ED visits with a primary diagnosis of HG rose from 177,940 to 227,298 (+28%). Mean ED cost per visit (2021 dollars) climbed from $2,156 to $3,549 (+65%). Aggregate ED costs roughly doubled, from about $384 million to $807 million. The share admitted from those visits fell from 7.7% to 4.5% (~42% relative drop). More presentations, higher unit cost, fewer inpatient conversions: outpatient pathways, observation status, and payer pressure may all play a role. NEDS cannot fully separate them, but the acute-care load clearly grew.

Gardner's 2.2% California rate also marks higher-risk pregnancies. After adjustment, HG care was linked to higher risks of preeclampsia (~18% relative increase), preterm birth (~25%), anemia (~37%), and placental abruption (~14%), plus gestational hypertension and small-for-gestational-age infants. Most HG pregnancies still end well; "just morning sickness" remains the wrong frame once hospital coding is involved. See also our review of pregnancy complications and preeclampsia statistics.

In Fiaschi's English cohort, risk factors included younger age, Asian or Black ethnicity, nulliparity, female fetus, multiple pregnancy, thyroid and parathyroid disease, Type 1 diabetes, and hypercholesterolemia. History of HG was the strongest independent predictor of a new admission (aOR 4.74, 95% CI 4.46-5.05).

Recurrence risk in later pregnancies

Recurrence numbers diverge by definition. A systematic review by Dean and colleagues found five heterogeneous studies reporting rates from 15% to 81%. Hospital-register studies tend to sit lower; self-reported severe-symptom studies sit higher. ACOG notes that recurrence of nausea and vomiting of pregnancy more broadly also ranges from 15% to 81%.

The MOTHER follow-up by Nijsten and colleagues (2021) is a careful prospective look at women already admitted for HG. Of 73 women in follow-up, 35 (48%) became pregnant again; 40% of those who conceived had postponed pregnancy because of HG. Recurrent HG (vomiting with medication use, weight loss, admission, tube feeding, or symptoms severe enough to affect life or work) occurred in 89% of those subsequent pregnancies. One woman terminated because of recurrent HG; eight (23%) considered it. Among 38 who did not conceive again, 24 said HG was the reason.

That 89% is not a population rate for mild nausea. It is the rate among women sick enough for prior admission who risk another pregnancy, the group asking, "Will this happen again?" For them, planning antiemetics, early fluids, and work leave before the positive test is not overkill. Register studies that require a second hospital admission report lower rates, sometimes in the mid-20s. Both can be true: coding undercounts home management; severe-symptom surveys over-represent the worst cases.

Finding	Estimate	Source
Recurrence range across studies	15-81%	Dean et al. systematic review, 2020
Recurrence after prior HG admission (next pregnancy)	89%	Nijsten et al., MOTHER follow-up, 2021
Postponed next pregnancy due to HG	40% of those who conceived again	Nijsten et al., 2021
Did not conceive again because of prior HG	24 of 38 non-conceivers	Nijsten et al., 2021
Prior HG as risk factor for new admission	aOR 4.74	Fiaschi et al., 2016

Sources: Dean CR et al. Eur J Obstet Gynecol Reprod Biol X. 2020; Nijsten K et al. Acta Obstet Gynecol Scand. 2021; Fiaschi L et al. Hum Reprod. 2016. Definitions of recurrence differ by study (hospital coding vs symptom severity).

The GDF15 finding

For decades the leading story was hCG, timed to the peak of first-trimester nausea. It never fully fit the genetics or the extreme outliers. The 2023 Nature paper by Fejzo, Rocha, Cimino, Lockhart, Petry and colleagues, with senior leadership including Stephen O'Rahilly, supplied a mechanism that does.

GDF15 acts on the brainstem and can suppress appetite and provoke nausea. The team confirmed that higher circulating GDF15 associates with vomiting in pregnancy and with HG. Using mass spectrometry in mother-fetus pairs discordant for a common GDF15 protein variant, they showed that the vast majority of GDF15 in maternal plasma during pregnancy comes from the feto-placental unit. In one analysis, maternal contribution averaged under 1% when the mother was heterozygous and the fetus was not.

The counterintuitive half is sensitivity. Rare and common HG-risk variants in GDF15 lower circulating GDF15 in the non-pregnant state. Women heterozygous for the rare C211G mutation, previously linked to at least a ten-fold increase in HG risk, had more than 50% lower circulating GDF15 when not pregnant. Mendelian randomization suggested higher non-pregnant GDF15 is protective (inverse-variance weighted odds ratio 0.70 per standard deviation increase; 95% CI 0.65-0.76). Low prior exposure appears to leave a woman more sensitive when fetal GDF15 surges. Mouse experiments supported desensitization after prior exposure.

Natural history lined up. Women with β-thalassemia, who live with chronically high GDF15, reported nausea or vomiting in only about 5% of pregnancies versus greater than 60% of matched controls. That is a clue that maternal set point matters, not a treatment recommendation. GDF15 levels still overlap between cases and controls, so a blood test is not diagnostic and pathway therapies remain investigational. The finding does reframe the visit: the patient who cannot stand or work is responding to a fetal hormone signal she is primed to feel. A mother or sister with HG also raises pretest probability well above the population 1%.

Work loss and pregnancy termination

Hyperemesis is an employment event as much as an obstetric one. ACOG notes that nausea and vomiting of pregnancy "significantly contributes to health care costs and time lost from work." Piwko and colleagues modeled the U.S. NVP burden for 2012 at about $1.78 billion: roughly $1.06 billion (60%) direct and $716 million (40%) indirect, including lost work and caregiver time. Average cost per woman with NVP was about $1,827, rising with severity. Those are 2012 model dollars, not a claims audit, but the scale explains why disability systems see this diagnosis repeatedly. Geeganage's ED data add hundreds of millions more in emergency visits by 2014 alone.

Termination statistics come largely from self-selected surveys of women who sought HG information or charity support. They describe the severe end, not all pregnancies with any nausea, and remain the best published windows into how often HG ends a wanted pregnancy.

Poursharif, Korst, MacGibbon, Fejzo, Romero and Goodwin surveyed 808 women through an HG website (2003-2005). Of those, 123 (15.2%) reported at least one elective termination because of HG, and 49 (6.1%) reported multiple terminations. Leading reasons included inability to care for family and self (66.7%), fear that they or the baby could die (51.2%), and fear the baby would be abnormal (22.0%). Women who terminated were about three times as likely to say providers were uncaring or did not understand how sick they were (52.0% vs 24.5%; odds ratio 3.34, 95% CI 2.21-5.05).

Nana and colleagues, with Pregnancy Sickness Support in the United Kingdom, surveyed 5,016 participants. Among those with usable data, 4.9% (249/4,994) reported terminating a pregnancy because of HG, and 52.1% (2,601/4,994) had considered it. Occasional suicidal ideation related to severe sickness was reported by 25.5%, and regular ideation by 6.6%. Both outcomes associated with greater severity, poorer function, and poorer perceived care. Those mental-health figures sit near the perinatal risk patterns in our postpartum depression and maternal mental health statistics. The timing differs, but the message is the same: severe pregnancy illness is a psychiatric emergency as well as a medical one.

These survey rates are not the population abortion rate among all people with pregnancy nausea. They show that for a subset with true HG, care failed so completely that ending a wanted pregnancy felt like the only option. That is a quality-of-care failure. When patients later ask about spacing another pregnancy, our ovulation calculator and period calculator can help with timing after recovery, but the medical plan must start before conception: antiemetics on hand, early weight checks, a hospital that will treat rather than dismiss.

Treatment access and gaps in care

ACOG's ladder is clear on paper: diet and ginger for mild disease; vitamin B6 with doxylamine first-line; antihistamines, dopamine antagonists, and serotonin antagonists as disease escalates; IV fluids with thiamine and nutrition support when oral intake fails. Early treatment may reduce progression to hyperemesis and hospitalization. The gap is access, not the guideline text.

In Nana's UK survey, respondents who described primary and secondary care as extremely poor were far less likely to have been offered medication than those who described excellent care (odds ratios 23.99 and 7.39 for primary and secondary care). A related mixed-methods report found 41.2% of women on prescribed medicines had to actively request them. Samples are self-selected, but the direction matches clinic reality: treatment delayed until the patient becomes an excellent historian of her own decline.

U.S. structure adds friction. Medicaid covered 57.6% of HG ED visits by 2014 (up from 47.9% in 2006), per Geeganage, so state rules on antiemetics, home infusion, and observation status shape a large share of acute care. Private plans vary on outpatient IV hydration without an ED stamp. Rural counties short on obstetric beds (the same map as our maternity care deserts and provider shortages) often have no day-unit pathway. The default becomes the ED, which is expensive and poorly designed for multi-week control.

Admission criteria remain inconsistent. StatPearls' 2025 review notes no universal standard for when to admit, so many people cycle through outpatient visits before anyone checks electrolytes. Practical thresholds: inability to tolerate oral intake, failed outpatient antiemetics, significant ketonuria or electrolyte abnormality, sustained weight loss, or inability to function. Thiamine should go in with IV dextrose-containing fluids. Wernicke encephalopathy is rare and preventable, yet it still appears in case reports when that step is skipped. Gardner's California cohort predated 2018 ACOG updates favoring earlier pharmacologic treatment, so a clean national before-after on complications is still missing. Undertreatment's fingerprints are already clear in terminations, suicidal ideation, ED volume, and recurrent disease.

For early pregnancy timing, our implantation calculator can clarify when nausea often starts; it does not replace care if vomiting is continuous. Falling weight or sparse urine warrants same-day evaluation. Families planning after a loss or termination related to HG may also want our miscarriage and pregnancy loss statistics. The woman who lost 11 pounds before anyone believed her stabilized on scheduled antiemetics, thrice-weekly IV fluids, and a work note that treated HG as the disability it is. She delivered at term. She would not risk a second pregnancy without a written plan before conception. Given 89% recurrence after prior admission, that is data literacy.

Frequently asked questions about hyperemesis gravidarum

How common is hyperemesis gravidarum?

ACOG estimates HG affects roughly 0.3% to 3% of pregnancies, while ordinary nausea of pregnancy affects 50% to 80%. English hospital data put admission-level HG near 1.5% of pregnancies; a California study found 2.2% of births involved ED or inpatient HG care. The range mostly reflects how strictly HG is defined.

Is hyperemesis the same as morning sickness?

No. Morning sickness is common nausea and vomiting that usually still allows some oral intake. Hyperemesis gravidarum is the severe end: persistent vomiting, signs of starvation such as ketonuria, and often weight loss of 5% or more of prepregnancy weight. ACOG notes HG is a leading reason for early pregnancy hospital admission.

What is the recurrence risk in a later pregnancy?

Published recurrence rates range from about 15% to 81%, depending on whether studies use hospital codes or self-reported severe symptoms. In the MOTHER follow-up of women previously admitted for HG, recurrence reached 89% among those who conceived again. Prior HG raised odds of a new hospital admission nearly five-fold in a large English cohort.

What did the GDF15 research show?

A 2023 Nature study found that most GDF15 in a pregnant woman's blood comes from the fetus and placenta, that levels are higher with vomiting and HG, and that low GDF15 before pregnancy increases susceptibility. That includes a rare GDF15 variant linked to at least a ten-fold higher HG risk. The work suggests future prevention strategies but is not yet a clinical blood test.

Can hyperemesis lead someone to end a wanted pregnancy?

Yes, in severe cohorts. In one U.S. web survey of 808 women with HG, 15.2% reported at least one termination because of HG. In a UK survey of more than 5,000 respondents, 4.9% reported termination and 52.1% had considered it. Both studies are self-selected; they still document a serious care-failure pathway.

When should someone seek urgent care for pregnancy nausea?

Seek same-day care if you cannot keep liquids down, are losing weight, feel faint, have very dark or infrequent urine, vomit blood, or cannot function at work or home. ACOG supports early treatment to help prevent progression. IV fluids with thiamine and antiemetics are standard when oral therapy fails, not a last resort after weeks of suffering.