Hyperemesis gravidarum

In brief

What it is, and how it differs from morning sickness

Most pregnant women feel sick in the first weeks. A 2013 meta-analysis of 59 studies from 13 countries, covering 93,753 women, put the average rate of nausea and vomiting of pregnancy at 69.4% (95% confidence interval 66.5 to 72.3%), with 32.7% of pregnant women having nausea alone and 23.5% still symptomatic in the third trimester; the symptoms were rated mild in 40% of cases, moderate in 46% and severe in 14% ². A confidence interval is the range within which the true value probably lies given the size of the studies. Hyperemesis gravidarum, a Greek-and-Latin name meaning excessive vomiting (Greek hyper, excessive, and emesis, vomiting) of pregnant women (Latin gravidarum), is the severe end of that spectrum, and the same meta-analysis, drawing on 26 studies of 6,155,578 pregnancies, put its prevalence at 1.1% (0.8 to 1.3%), ranging from 0.3% to 3.6% between studies ². The spread in the figures is not only geography. Until 2021 there was no agreed definition, and the 2019 review in Nature Reviews Disease Primers, written when there was none, reported a range of 0.3% to 10.8% depending on what each study chose to count ³³.

The definition that now exists was built by consensus. A Dutch-led group used a Delphi process, a structured series of anonymous survey rounds in which a panel converges on agreement, with 21 candidate criteria; of 277 stakeholders invited, 178 completed the first round and 125 (70%) the second, 20 attended a face-to-face meeting in Windsor in the United Kingdom, and 92% of the 96 who answered a final consultation accepted the result ¹. The panel deliberately included four groups: researchers, women who had lived through the condition and their families, obstetricians and midwives, and the general practitioners, dietitians and nurses who see it outside the hospital ¹. The Windsor definition they produced has four elements: symptoms start in early pregnancy, before 16 weeks of gestational age; nausea and vomiting, at least one of which is severe; inability to eat and/or drink normally; and symptoms that strongly limit the activities of daily living, with signs of dehydration counted as contributory rather than required ¹. What is absent from the definition matters as much as what is in it: there is no weight-loss threshold, no laboratory test and no requirement for hospital admission ¹.

Severity is scored with questionnaires. The PUQE, the Pregnancy-Unique Quantification of Emesis, asks about three symptoms, nausea, vomiting and retching, and in large prospective cohorts its score predicted four things of direct importance: whether women could keep down their multivitamins, their rates of emergency visits and hospitalization, the health cost of their illness, and their own rating of their well-being ⁸. The HELP score, the HyperEmesis Level Prediction tool, was built by the patient-founded Hyperemesis Education and Research (HER) Foundation because the PUQE missed too many women at the severe end: in its validation survey, posted on hyperemesis websites, the HELP score classified 92% of women who reported that nothing goes or stays down as severe, against 58% for the PUQE, and women who had been hospitalized were more likely to be rated severe by it ⁹. The 2024 guideline of the Royal College of Obstetricians and Gynaecologists (RCOG) recommends either tool, at grade C, meaning the recommendation rests on lower-quality observational studies or extrapolated evidence rather than randomized trials ¹⁰. In 2026 a consensus study for researchers, after a Delphi consultation of 155 stakeholders from 10 countries, 51% of them women or families with lived experience, agreed that trials should measure severity with the PUQE-24 ¹¹.

The burden is measured in hospital beds. In England, a study of Hospital Episode Statistics covering 8,215,538 pregnancies in 5,329,101 women between 1997 and 2012 counted 186,800 admissions for hyperemesis in 121,885 pregnancies, a prevalence of 1.5% ³. A commentary in AJOG Global Reports, published online in December 2022, called it the second most common reason for hospitalization in pregnancy ²³, and a 2024 review estimated the annual cost in the United States at more than 1.7 billion dollars ³⁴. Those two figures come from secondary sources rather than primary counts, and the cost estimate should be read as an order of magnitude. A 2025 systematic review of the costs found only seven costing studies, methodologically and clinically heterogeneous, all of which reported a substantial economic burden that rose with severity ³⁵.

How it presents, and what it does to the body

The illness begins as ordinary pregnancy sickness and separates from it by degree. In two Danish early-pregnancy cohorts with repeated blood sampling, symptom scores in normal pregnancies peaked around week 8 and then declined, in step with GDF15 reaching its plateau by week 9 ³¹. In hyperemesis the vomiting does not stop. The clearest picture of how long and how far it can run comes from a systematic review of 146 case reports of Wernicke's encephalopathy, the thiamine-deficiency brain injury discussed later: the 177 women had vomited for a median of 7 weeks and lost 12.1 kg before the injury declared itself, and they had become thiamine-depleted between 10 and 15 weeks of gestation ¹⁷. A 2026 cross-sectional study of 36 women attending a day unit in Dublin measured what they actually took in: a median of 1,050 kcal (interquartile range 481 to 1,461) and 1,196 mL of fluid a day against estimated requirements of 2,057 kcal and 2,300 mL, with lower intakes of energy, sodium, potassium, iron, iodine, vitamin D and thiamine correlating with worse nausea and vomiting scores ⁴⁴.

For decades ketonuria, ketones in the urine, was treated as a marker of the diagnosis ¹². Ketones are the acids the body makes when it burns fat for lack of carbohydrate, and a dipstick turns color in their presence; the logic was that a woman who was truly starving would show them. The evidence does not support the test. A 2014 systematic review of 81 articles on 9 candidate biomarkers found that 65% of all studies had admitted only women with ketonuria, yet in the 5 studies that examined the question there was no association between ketonuria and either the presence or the severity of hyperemesis ¹². In 2020 the Dutch group behind the MOTHER trial (Maternal and Offspring outcomes after Treatment of HyperEmesis by Refeeding) looked at 215 women admitted to 19 hospitals, where ketonuria was not required for entry: the degree of ketonuria at admission predicted neither the severity of nausea and vomiting, nor quality of life, nor weight loss, nor the number of readmissions, but women with more ketones were kept in hospital longer, by 0.27 days (95% confidence interval 0.05 to 0.48) for each grade on the dipstick, which the authors read as clinicians basing length of stay on a number that carried no information ¹³. The 2024 RCOG guideline now states, at grade A, that ketonuria is not an indicator of dehydration and should not be used to assess severity ¹⁰.

The thyroid is often biochemically overactive. In 1992 Goodwin and colleagues in Los Angeles compared 57 women with hyperemesis with 57 controls matched for gestational age: thyroid-stimulating hormone (TSH) was suppressed in 60% of the hyperemesis group against 9% of controls, human chorionic gonadotropin (hCG) correlated with free thyroxine (the active, unbound thyroid hormone in the blood) and inversely with TSH, and the degree of biochemical hyperthyroidism tracked the severity of vomiting ⁴¹. This is transient gestational thyrotoxicosis, driven by hCG's weak action on the thyroid's TSH receptor, and the 2014 systematic review found the association between thyroid hormones and hyperemesis inconsistent across studies ¹². It is a finding to recognize so as not to treat it as Graves' disease (an autoimmune disease in which antibodies drive the thyroid to overwork), not a cause.

Who gets it has been measured at scale. In the English cohort of 8.2 million pregnancies, the strongest independent risk factor was hyperemesis in a previous pregnancy, with an adjusted odds ratio of 4.74 (95% confidence interval 4.46 to 5.05); younger age, less socioeconomic deprivation, Asian or Black ethnicity, a female fetus and a multiple pregnancy all raised the risk, and among comorbidities (other illnesses present at the same time), parathyroid dysfunction (disorder of the small glands that control blood calcium; adjusted odds ratio 3.83), hypercholesterolemia (high blood cholesterol; 2.54), type 1 diabetes (1.95) and thyroid dysfunction (1.85) were most strongly associated, while higher parity (more earlier births) lowered the risk (0.90) ³. An odds ratio compares the odds of an outcome in one group with another; 1.0 means no difference. The excess of female fetuses is consistent: a 2011 meta-analysis found an odds ratio of 1.27 (1.21 to 1.34) for a female child, and in 2026 the Danish cohorts found GDF15 about 19% higher in the mother's blood when the fetus was female, with the rise beginning a day later for male fetuses ^31,45. Catherine Nelson-Piercy, the second author of the 2016 study and a co-author of its 2018 companion, declared personal fees from Sanofi Aventis, Warner Chilcott, Leo Pharma, UCB and Falk outside the submitted work and is a co-developer of the RCOG guideline, of whose 2024 revision she is lead author; the other authors declared none ^3,10,32.

What it does to the mother is set out in the companion English study of 8,211,850 pregnancies ending in live birth or stillbirth: admission for hyperemesis raised the odds of anemia (odds ratio 1.28, 99% confidence interval 1.23 to 1.33), pre-eclampsia (high blood pressure with signs of organ strain in the second half of pregnancy; 1.16), eclampsia (seizures complicating pre-eclampsia; 1.84) and, before birth, deep vein thrombosis (a clot in a deep vein, usually of the leg; 1.94) and pulmonary embolism (a clot that has traveled to the lungs; 2.54), together called venous thromboembolism, while lowering the odds of stillbirth (0.77, 0.66 to 0.89) and of post-term delivery (birth at or after 42 weeks; 0.86) ³². The psychiatric burden is covered in the section on the psychogenic era, because it was for so long mistaken for the cause ^46,47. The rarer physical catastrophes, Wernicke's encephalopathy and refeeding syndrome, belong to the section on thiamine ^17,48.

Pernicious vomiting: the era when it killed

Before the condition's modern name was in common use it was the obstinate, severe or pernicious vomiting of pregnancy, pernicious meaning deadly, and the titles of the nineteenth-century journals record both the deaths and the remedy of last resort. In 1875 E. Copeman wrote in the British Medical Journal on "A Novel Treatment of Obstinate Vomiting in Pregnancy" ⁴⁹, and the same year Æneas Munro presented to the Edinburgh Obstetrical Society a communication on the vomiting of pregnancy, with a case that required the induction of labor ⁵⁰. In 1885 the British Medical Journal printed "Severe Vomiting in Pregnancy: Miscarriage: Death, with Remarks" and a Texas journal "Induced Abortion to Relieve Vomiting in Pregnancy" ^51,52. A paper in the Texas Medical Journal in August 1900 was titled "Pernicious Vomiting of Pregnancy and Its Treatment with Electricity", a reminder that the therapies of the time were as desperate as the diagnosis ⁵³, and in 1909 a death was reported to the Royal Society of Medicine as "A Fatal Case of Toxic Vomiting of Pregnancy with Microscopical Sections of Liver and Kidneys" ⁵⁴. In December 1912 J. Whitridge Williams, professor of obstetrics at Johns Hopkins and the author of the obstetric textbook that still carries his name, published "Further Contributions to Our Knowledge of the Pernicious Vomiting of Pregnancy" in the Glasgow Medical Journal, a 27-page lecture to the Glasgow Obstetrical and Gynaecological Society ⁵⁵. By 1968 a 41-page review in the American Journal of Obstetrics and Gynecology under the plainer title "Nausea and vomiting in pregnancy" marked the shift in name and in the expectation that the condition could be managed ⁵⁶. These papers are cited here for what their titles and their existence show; their texts were not available for this page and their contents are not summarized. No indexed source was found for descriptions of the condition before the nineteenth century, or for when intravenous fluids arrived and what they did to mortality, so neither is claimed here.

The death most often attributed to it is Charlotte Brontë's, in 1855. Her own physician ascribed her death to tuberculosis, and her death certificate records phthisis, the old name for it ^24,57; later observers reading her letters concluded that she died of hyperemesis gravidarum, and in the era when that diagnosis was psychogenic they attributed her vomiting to neurosis and to a rejection of her pregnancy and her femininity ^24,57. How firm is the attribution? Not very. A 1991 analysis in Obstetrics and Gynecology of her last months cast doubt on whether she was ever pregnant at all and argued that tuberculosis with secondary Addison disease, failure of the adrenal glands, could fully explain her terminal signs and symptoms ²⁴. A 2019 article in Clinical Nutrition by two Nottingham clinicians, Simon Allison and the gastrointestinal surgeon Dileep Lobo, accepted hyperemesis as the most likely diagnosis but proposed that what killed her was refeeding syndrome, the metabolic collapse that can follow when a starved person starts eating again, because her nausea and vomiting remitted in the third week of March 1855 and she then declined and died ⁵⁷. Two careful readings of the same letters reach two different diagnoses, neither can be tested, and the honest statement is that she died of a wasting illness in early pregnancy that nobody can now name with confidence.

The condition still kills. In the systematic review of 177 published cases of Wernicke's encephalopathy complicating hyperemesis, 5% of the mothers died and 50% of the pregnancies were lost ¹⁷. The commentary Fejzo and colleagues published online in December 2022 records that maternal deaths from the condition continue to be reported, and that the condition is associated with high rates of fetal loss and termination ²³. In the Wernicke series the deaths followed thiamine depletion ¹⁷; no source on this page counts what else kills women with the condition today.

The psychogenic era, and how it ended

For much of the twentieth century the cause of severe vomiting in pregnancy was sought in the woman. In 1954 Psychosomatic Medicine published "Vomiting in pregnancy; a psychiatric study", and in 1971 Obstetrics and Gynecology published "Female sexuality and life situations. An etiologic psycho-socio-sexual profile of weight gain and nausea and vomiting in pregnancy", a title that is its own summary ^58,70. The 1991 Obstetrics and Gynecology analysis of Charlotte Brontë's death records how this reading was applied in retrospect to a woman dead for a century: her supposed hyperemesis was ascribed to neurosis and a rejection of her pregnancy and her femininity ²⁴. Those two papers are cited for their titles; their methods are not available in the abstracts and are not characterized here.

The era was examined on its own terms in 2002 by Shari Munch, a social-work scholar at Rutgers, in a review of how the condition had been characterized in the medical literature. Her analysis found that biologic and psychologic approaches had existed in parallel tracks throughout the condition's history, and that sociocultural factors rather than scientific evidence had shaped the overarching and predominant paradigm of psychogenesis; she set this in a wider pattern in which women's physical complaints of obscure cause are labeled psychological ²¹. The paper is a reading of the literature, not a trial, and its force comes from what it did not find: an empirical basis, in the literature it examined, for the psychological cause ²¹.

What ended the era was evidence that reversed the arrow. Women with hyperemesis are indeed more depressed and anxious than other pregnant women: a 2017 systematic review of 59 articles, with 12 in meta-analysis, found depression scores higher by a standardized mean difference of 1.22 (95% confidence interval 0.80 to 1.64) and anxiety scores by 0.86 (0.53 to 1.19), a very large effect for depression and a large one for anxiety, though with heterogeneity between studies of 94% and 84% ⁴⁶. A standardized mean difference expresses a difference in units of the spread of the scores, so that 1.22 is more than a full standard deviation. But the Norwegian Mother and Child Cohort, in which 851 of 92,947 women (0.9%) had hyperemesis, measured emotional distress four times: it was raised at 17 and 32 weeks of pregnancy and at 6 months after birth, and had returned to that of other mothers by 18 months, which the authors read as the distress being primarily a consequence of the illness ⁴⁷. Distress that tracks the vomiting and resolves after it is most simply read as an effect rather than a cause ⁴⁷. The same cohort had earlier found that a lifetime history of depression did raise the odds of hyperemesis (adjusted odds ratio 1.49, 95% confidence interval 1.23 to 1.79), but only 1.2% of women with such a history developed it and two thirds of women with hyperemesis had neither past nor current depression, so its authors concluded that depression does not seem to be a main driver ⁷¹. A 2026 meta-analysis of 13 case-control studies in 95,249 women found depression associated with hyperemesis (odds ratio 5.08, 95% confidence interval 2.71 to 9.52; heterogeneity 93%), more strongly when measured at diagnosis (6.04) than before it (2.11), and its authors state that the association does not imply causation and may reflect the illness's psychological impact ⁷².

The residue of the era is in how women say they are treated. In the 808-woman survey of 2007, women who had terminated a pregnancy because of hyperemesis were more than twice as likely to describe their health-care providers as uncaring or as not understanding how sick they were, 52.0% against 24.5% (odds ratio 3.34, 95% confidence interval 2.21 to 5.05) ²⁵. A 2021 British survey of more than 5,000 women reported increased rates of termination and of suicidal ideation, and in a 2024 survey of 1,220 women those who took their medications as prescribed had fewer hospital visits, visits were more likely among women of color, and the authors judged use of medication and nutritional therapy inconsistent and inadequate ^73,74. By 2024 Marlena Fejzo, whose genetic work drove the reversal, could title a review "Hyperemesis gravidarum theories dispelled by recent research", and the 2025 Lancet review states that inaccurate assumptions that the condition is self-limiting result in women being dismissed ^22,26.

Thalidomide, Bendectin and the Daubert case

Two drugs for morning sickness taught medicine opposite lessons. Thalidomide, a sedative, was prescribed to pregnant women for morning sickness from its release in the late 1950s until it was banned in November 1961 and withdrawn from most of the world by 1962, and over 10,000 children were born with severe and debilitating malformations, most famously the shortened limbs of phocomelia, before it was withdrawn; the drug is now used for multiple myeloma, a cancer of antibody-making cells in the bone marrow, and complications of leprosy, and a new generation of damaged children has been identified in Brazil, where it is used for leprosy ²⁷. The mechanism by which it damages the embryo was still being worked out in 2015, with candidates including the protein cereblon, blocked blood-vessel growth and reactive oxygen species, chemically aggressive forms of oxygen that damage cells ²⁷. The lesson taken was that no drug given in early pregnancy could be assumed safe, and every subsequent treatment for pregnancy sickness has been judged in its shadow.

Bendectin was that lesson applied to a safe drug. It combined doxylamine, an antihistamine, with pyridoxine, vitamin B6, and before 1983 between 10% and 25% of pregnancies were exposed to it, over the years as many as 33 million ⁵⁹. In 1983 Merrell Dow Pharmaceuticals voluntarily removed it from the market because of the many product liability suits then pending, not because of any regulatory finding ⁵⁹. In Canada, hospital admissions for excessive vomiting in pregnancy per thousand live births rose by 37% in 1983 over the 1980 to 1982 rate and by 50% in 1984, with similar rises in the United States, and no decrease in the rate of congenital malformations could be shown to offset the cost ⁵⁹. An ecological analysis of the Centers for Disease Control's Birth Defect Monitoring Program from 1970 to 1992 found that the prevalence of specific birth defects did not change as Bendectin use ceased over 1980 to 1984, while the hospitalization rate for nausea and vomiting of pregnancy doubled ⁷⁵. Robert Brent's 1995 review, which went through the epidemiologic studies, secular trends (changes in rates over calendar time), animal studies, dose-response and biologic plausibility, concluded that therapeutic use of Bendectin has no measurable teratogenic effect, called it the most prevalent tortogen-litigen, his coinage for a substance that causes lawsuits rather than defects, and wrote that many of the plaintiffs' experts had failed the scientific standards expected of knowledgeable scientists ²⁸.

The litigation changed American law. In Daubert v. Merrell Dow Pharmaceuticals, decided in 1993, a Bendectin birth-defect suit reached the United States Supreme Court, which used the case to overrule the older Frye standard, under which scientific evidence was admissible only if generally accepted in its field, and to impose a gatekeeper role on trial judges, who must now assess whether expert testimony rests on reliable methods ^60,61. The ruling itself is a primary document that PubMed does not hold; its character is described here as the indexed legal and medical literature summarizes it, and a 2022 law-and-medicine article argues that for malpractice its significance is to change the content of the standard of care, so that doctors testify on whether conduct was reasonable rather than merely customary ⁶¹. The irony is complete: a drug with no measurable teratogenic effect ²⁸ produced the legal standard by which claims of harm are now tested.

The drug came back. The same combination had stayed on the Canadian market since 1979 as Diclectin, and in April 2013 the United States Food and Drug Administration (FDA) approved a 10 mg/10 mg delayed-release tablet of doxylamine and pyridoxine as Diclegis, which a 2014 review described as the only medication then FDA-approved for nausea and vomiting of pregnancy, with studies of more than 200,000 women exposed in the first trimester showing no increased fetal risk ^14,29. Its safety is not in question; its efficacy turned out to be. The pivotal trial, published in 2010 by the Motherisk program in Toronto with colleagues at six United States sites, randomized 131 women to the tablet and 125 to placebo for 14 days and reported a larger fall in PUQE score, 4.8 ± 2.7 against 3.9 ± 2.6 points (P = .006), with 48.9% of treated women against 32.8% on placebo asking to continue the drug afterward ⁶³. P is the probability of seeing a difference at least this large if there were truly none; below .05 is conventionally called statistically significant. In 2018 a Toronto group reanalyzed the individual-level data from the clinical study report, the full trial dossier a company submits to regulators, obtained from Health Canada: 140 women had been randomized to each group, 131 and 125 analyzed, and 101 and 86 provided the final-day outcome; the difference was 0.73 points (95% confidence interval 0.21 to 1.25) when missing data were filled by carrying the last observation forward, and 0.38 points (−0.08 to 0.84), not statistically significant, using complete data only, while the trial's own prespecified minimal clinically important difference had been 3 points ¹⁵. The reanalysts concluded there was a trend toward efficacy but no clinically important benefit ¹⁵. They also set out the trial's commercial and regulatory record: it was sponsored by Duchesnay, the drug's manufacturer; in the published disclosures the lead author had been a paid consultant to the company, which also supported some of his studies, and three co-authors had advisory, speaking or consulting ties to it; the published reports had treated global well-being, prespecified as a secondary outcome, as a primary one; the FDA's summary review had called the effect a "small, but statistically significant improvement" and noted that there were no other FDA-approved treatments for the condition; and a 2016 Health Canada review had judged that "the results of the study are not definitive" ¹⁵. Guidelines on both sides of the Atlantic still place the combination first, on the strength of its safety record and the older evidence for its components, and the 2014 review that announced its return and the 2010 trial both came from the Motherisk program whose director, Gideon Koren, was the trial's lead author ^10,29,63. Koren is an author of four sources on this page ^2,8,29,63; none of them is retracted or under an expression of concern in PubMed as of 19 October 2026, but other papers of his have been retracted, among them a 2020 study of first-trimester dydrogesterone exposure that was first placed under an expression of concern citing "a number of errors and inconsistencies" ⁷⁶.

The hCG hypothesis and its limits

The oldest biological explanation was timing. Human chorionic gonadotropin, hCG, the placental hormone that a pregnancy test detects, rises steeply in the first weeks and peaks at about the time the sickness is worst; the English registry found that a multiple pregnancy raised the risk of admission for hyperemesis ³. The strongest evidence for a direct role was the 1992 Los Angeles study already described: in 57 women with hyperemesis against 57 matched controls, serum hCG was significantly higher, it correlated with free thyroxine (correlation coefficient r = 0.45, on a scale where 0 is no relation and 1 a perfect one) and inversely with TSH (r = −0.48), and both the hormone concentration and the degree of biochemical hyperthyroidism varied directly with the severity of vomiting; unextracted serum from the women stimulated iodide uptake in cultured rat thyroid cells in proportion to its hCG content ⁴¹. That was association in humans and mechanism in a dish, and it explained the thyroid findings well; whether it explained the vomiting was another question.

The question was answered by accumulation. The 2014 systematic review of diagnostic biomarkers found 81 articles on 9 candidates and reported that studies of hCG and thyroid hormones, as well as leptin (a hormone made by fat tissue), estradiol, progesterone and white cell count, showed inconsistent associations with hyperemesis, with only lymphocytes, a kind of white blood cell, tending to be higher ¹². The one biomarker that survived meta-analysis was Helicobacter pylori, the stomach bacterium of its own Atlas page: an odds ratio of 3.2 (95% confidence interval 2.0 to 5.1) against asymptomatic controls, but with a sensitivity of 73% and a specificity of 55%, meaning it was present in too many well women to be useful, and the authors suggested serology, a blood test for antibodies to the bacterium, might help only in selected patients ¹². A 2026 scoping review of 36 sources published from 2021 to May 2026 on the gut and gastric microbiome in hyperemesis found altered microbial diversity and composition but no reproducible signature, judged GDF15-centered endocrine susceptibility to have the strongest mechanistic support, and concluded that dysbiosis, an unbalanced mix of gut bacteria, is an associated modifier, consequence or marker rather than an established independent cause ⁷⁷.

What the hCG hypothesis was probably seeing is now clearer. In the Cambridge Baby Growth Study, serum GDF15 at a median of 15.1 weeks was strongly and positively correlated with hCG, and GDF15, unlike hCG, has a receptor in the brainstem and causes vomiting when injected into animals ^36,38,42. The evolutionary biologist Daniel Stadtmauer's 2025 review notes that placental production of GDF15 rose 100- to 1,000-fold to human-like levels in catarrhine primates, the Old World monkeys and apes, and is low or absent in New World monkeys, rodents and most other mammals, so that pregnancy sickness is phylogenetically restricted, found only in some branches of the animal family tree, but not uniquely human ⁷⁸. hCG remains a plausible contributor through the thyroid and through its correlation with GDF15 ^41,42. The multi-ancestry genome-wide study of 10,974 cases, in its preprint form, found no association with the genes encoding hCG or estrogen or their receptors, and its authors argued that further work on those hormones lacks genetic support ^7,43.

It runs in families, and through the mother

The first modern clue was pedigree. In a 2011 survey of women treated with intravenous fluids for severe nausea and vomiting, with unaffected friends as controls, the sisters of affected women had an odds ratio of 17.3 for the condition themselves (P = .005), and 33% of cases reported a mother with severe nausea and vomiting in pregnancy against 7.7% of controls ⁵. Friend controls and self-report are weak instruments, but the Norwegian registry removed both: linking the Medical Birth Registry of Norway from 1967 to 2006 produced 544,087 mother-and-childbearing-daughter units and 399,777 mother-and-son units, and daughters who had themselves been born after a pregnancy complicated by hyperemesis had a 3% risk in their own pregnancies against 1.1% for other women, an unadjusted odds ratio of 2.9 (95% confidence interval 2.4 to 3.6), while the female partners of sons born after such pregnancies had a risk of 1.2% (odds ratio 1.0, 0.7 to 1.6) ⁶. The condition passed from mother to daughter and not from mother to son's partner, which the authors read as a stronger influence of the maternal than the fetal genotype, while allowing that shared environment along the maternal line could not be excluded ⁶.

The genes were found by a company's customers. In 2018 Fejzo and colleagues reported a genome-wide association study, a scan of common variants across the whole genome for those more frequent in cases than controls, using 23andMe participants for both a yes-or-no hyperemesis phenotype, the trait being studied, and an ordinal severity phenotype, graded in ranked steps; two loci reached genome-wide significance (P < 5 × 10⁻⁸) in both scans and replicated in an independent cohort, chr19p13.11, a band on the short arm of chromosome 19, where the gene is GDF15, and chr4q12, a band on the long arm of chromosome 4, where it is IGFBP7, both already known to be involved in placentation (the building of the placenta), appetite and cachexia ⁶⁵. Several authors were employees of 23andMe, Inc., and one had moved to Amgen, which is a commercial interest to note; the authors themselves wrote that proving causal roles for GDF15 and IGFBP7 would require further study ⁶⁵. The GDF15 locus has since replicated in a Japanese cohort of 23,040 pregnant women, where two variants at 19p13.11 reached significance in the meta-analysis of discovery and replication stages, which answers the objection that the first finding came from a largely European, self-selected sample ⁷⁹.

By 2026 the map had ten loci. A multi-ancestry genome-wide study of 10,974 cases and 461,461 controls across European, Asian, African and Latino ancestries confirmed GDF15, IGFBP7, the progesterone receptor PGR and the GDF15 receptor GFRAL, and added SYN3 (synapsin III), SLITRK1 (SLIT and NTRK-like family member 1) and IGSF11 (immunoglobulin superfamily member 11), with enriched or enhanced expression in the brain; FSHB (follicle-stimulating hormone subunit beta); TCF7L2 (transcription factor 7-like 2, a diabetes-associated gene); and CDH9 (cadherin 9) ^7,43. The downstream analyses are where the biology is: GDF15 and TCF7L2 were expressed primarily in extravillous trophoblast, the placental cells that invade the uterine wall, and a model separating maternal, paternal and fetal effects found opposing effects for GDF15 between the maternal and the fetal genotype, whereas IGFBP7 and PGR were expressed in the mother's spiral arteries and acted only through her genome ⁷. Some of the loci were also associated with abnormal pregnancy weight gain, pregnancy duration, birth weight and pre-eclampsia ⁷; the preprint version of the study also reported the risk loci to be under significant evolutionary selection, a claim its published abstract does not repeat ⁴³. A fetal gene that raises the mother's risk and a maternal gene that lowers it is the genetic signature of the mechanism the next section describes.

GDF15: the 2023 Nature paper

GDF15, growth differentiation factor 15, is a distant member of the transforming growth factor beta (TGF-beta) family that circulates as a 25-kilodalton dimer, a protein of two linked chains ³⁶; it rises in the blood in response to a broad range of cellular stressors and in illnesses marked by cachexia and malaise ³⁷, metformin raises it ²⁰, and so does pregnancy ⁴. GDF15 has its own Atlas page. Its receptor was unknown until 2017, when a group at Janssen reported that GDF15 binds with high affinity to GFRAL, a distant relative of the receptors for a different class of the same superfamily's hormones, that Gfral is expressed in neurons of the area postrema and nucleus of the solitary tract in mice and humans, that deleting the receptor abolished GDF15's effect on food intake and body weight in mice, and that signaling requires the co-receptor RET, a second receptor protein that GFRAL needs in order to signal ³⁶. The 2023 Nature paper and later reviews describe the receptor as restricted to the hindbrain ^4,37,80. The area postrema is the small region at the back of the brainstem that lacks a blood-brain barrier and samples the blood for things that should make an animal vomit; the 2017 paper came from a company interested in GDF15 as an obesity drug, and its authors were Janssen employees ³⁶. In 2020 Borner and colleagues showed what the receptor's location implied: GDF15 caused vomiting in Suncus murinus, the musk shrew, one of the few small laboratory animals that can vomit, and in mice and rats, which cannot, it caused pica (eating non-food such as clay, a stand-in for nausea) and anorexia (here, loss of appetite, not the eating disorder); GDF15 also contributed to chemotherapy-induced malaise in mice ³⁸. That is animal evidence of mechanism, not human evidence of cause.

GDF15 was first found in pregnancy in 2000, under the name macrophage inhibitory cytokine-1: levels in pregnant women's serum were high and rose with gestation, large amounts were present in amniotic fluid and placenta, a trophoblast cell line secreted it, and its discoverers suggested it might promote fetal survival by suppressing the production of maternal inflammatory signals within the uterus ⁶². The human association with vomiting came in 2018 from Cambridge. In 791 women of the Cambridge Baby Growth Study sampled at a median of 15.1 weeks, circulating GDF15 was higher in those reporting vomiting in the second trimester than in those with no nausea or vomiting, 11,581 (95% confidence interval 10,977 to 12,219) against 10,593 (10,066 to 11,147) pg/mL, picograms (trillionths of a gram) per milliliter (P = 0.02), and higher still, at 13,157 pg/mL, in the 11 women taking antiemetic drugs; GDF15 was strongly correlated with hCG and inversely with the mother's pre-pregnancy body mass index, a finding replicated in 231 women of a second cohort ⁴². The differences are modest, the sampling was after the peak of symptoms, and the authors framed the result as support for the concept that GDF15 may be playing a pathogenic role ⁴².

The paper that turned a candidate into a cause was published online in Nature on 13 December 2023, by Fejzo, O'Rahilly and 41 co-authors from Los Angeles, Cambridge, Glasgow, Edinburgh, Sri Lanka, Split in Croatia, the Regeneron Genetics Center in New York and the Hyperemesis Education and Research Foundation in Oregon; the long-acting GDF15 used in the mouse experiments was supplied by Pfizer under a material transfer agreement ⁴. Kimber MacGibbon of that foundation, the patient charity whose website and social media recruited participants for several of the group's surveys, is also a co-author of the 2026 genome-wide study and of the HELP and metformin surveys ^7,9,20; the Nature paper's competing-interests statement is published online and could not be opened for this page. It made five moves. First, it confirmed that higher maternal GDF15 is associated with vomiting and with hyperemesis ⁴. Second, using mass spectrometry, a method that weighs molecules finely enough to tell two versions of a protein apart, to detect a naturally labeled variant of the protein, it showed that the vast majority of GDF15 in the maternal plasma is derived from the feto-placental unit, that is, made by the fetus and placenta rather than by the mother ⁴. Third, by studying carriers of rare and common genetic variants, it found that low levels of GDF15 in the non-pregnant state increase the risk of hyperemesis; the preprint spells out that a rare protein-altering variant, C211G, strongly predisposes mothers, particularly when the fetus is wild-type (carrying the usual version of the gene), markedly impairs secretion of GDF15 from cells and associates with low circulating levels outside pregnancy, and that two common haplotypes (stretches of DNA variants inherited together) that predispose to hyperemesis also associate with lower levels outside pregnancy ^4,39. Fourth, it turned to women with beta-thalassemia, an inherited anemia in which GDF15 is chronically very high (24.2-fold above healthy controls in beta-thalassemia major in a Sri Lankan series of 103 patients ⁴⁰); among women with thalassemia who had carried a pregnancy to a live birth, particularly women with the milder intermedia form, about 5% reported any nausea or vomiting against more than 60% of ethnically and age-matched controls ⁴. Fifth, in mice, the acute food-intake response to a bolus, a single dose given at once, of GDF15 was influenced bi-directionally by prior circulating levels, and a prior long-acting dose markedly reduced the response to an acute one, establishing that the system can be desensitized ^4,39.

The authors' own wording was careful: the findings support a putative causal role for fetally derived GDF15 in the nausea and vomiting of human pregnancy, with maternal sensitivity, at least partly determined by pre-pregnancy exposure, being a major influence on severity ⁴. The claim of cause rests on the convergence of human genetics, which cannot be confounded by the illness itself, with the animal mechanism and the thalassemia observation; what has not yet been reported is the experiment that would close the loop in people, blocking GDF15 in pregnant women and watching the vomiting stop ^81,82. Two findings from 2024 bear on whether that experiment could be safe. From 75,018 whole-exome (all protein-coding DNA read) or genome-sequenced participants in the Pakistan Genomic Resource, a Novartis-led group identified 8 homozygous GDF15 knockouts (two altered copies) and 227 heterozygous carriers (one altered copy) of loss-of-function alleles, including C211G; the knockouts ranged from 31 to 75 years, were fertile, had multiple children and showed no consistent overt phenotype, supporting the hypothesis that GDF15 is not required for fertility, healthy pregnancy or fetal development ⁶⁶. That paper's authors were largely employees of Novartis, AstraZeneca, Astellas and Regeneron, companies with GDF15 programs, which is the commercial interest to record ⁶⁶. And Fejzo's 2024 review in Trends in Molecular Medicine set out the reversal plainly: there is now strong evidence that the condition is associated with poor outcomes, that a common cause has been identified, and that pregnancy sickness is likely a protective evolutionary mechanism no longer necessary for human survival ²².

2024 to 2026: desensitization, metformin, trajectories and antagonists

If sensitivity is set by pre-pregnancy exposure, raising GDF15 before conception should protect. The Nature authors proposed two ways to induce that desensitization: low doses of recombinant GDF15 before pregnancy, whose short-term safety outside pregnancy had been shown in phase 1 trials (first tests of safety in people), or metformin, the diabetes drug, which robustly raises GDF15 in humans ⁴. Fejzo's group tested the second idea the fastest way available: a structured questionnaire posted on the Hyperemesis Education and Research Foundation's social media from January 2023 to September 2024, in which 5,414 participants reported daily use of 32 substances in the month before each pregnancy ²⁰. Metformin before the first pregnancy was associated with a more than 70% lower risk of hyperemesis, adjusted relative risk 0.29 (95% confidence interval 0.12 to 0.71, P = .007), and with 82% lower odds of severe nausea and vomiting or hyperemesis in the second pregnancy (adjusted odds ratio 0.18, 0.06 to 0.59) even after adjusting for the 86% recurrence risk the same survey identified; tobacco use was also associated with lower risk (adjusted relative risk 0.51, 0.30 to 0.86), while selective serotonin reuptake inhibitors, a class of antidepressants (adjusted relative risk 2.41, 1.33 to 4.38), and, before a second pregnancy, cannabis (adjusted odds ratio 3.48, 1.80 to 6.75) were associated with higher risk ²⁰. A relative risk is the risk in the exposed group divided by the risk in the unexposed. This is a retrospective survey of a self-selected online community, run by the same group that proposed the mechanism, with the obvious risk that women who used metformin differ in ways the models did not capture, and the authors themselves called for clinical trials ²⁰.

The independent evidence is mixed. In a Danish registry study of 1,447,968 pregnancies from 1998 to 2018, metformin prescriptions filled before pregnancy did not lower the prevalence of hospital-diagnosed hyperemesis (1.9%) or its recurrence (27%), but use for 6 months or more was associated with a prevalence of 1.1% compared with use for less than 2 months (adjusted odds ratio 0.49, 95% confidence interval 0.31 to 0.75), and use continued up to conception with 1.2% compared with stopping longer before (adjusted odds ratio 0.61, 0.38 to 0.95); the authors, writing from an observational design, call for randomized trials ¹⁸. A smaller North American cohort, in the MotherToBaby pregnancy studies in the United States and Canada, compared 80 women exposed to metformin before conception between 2012 and 2023 with 4,411 unexposed: hyperemesis occurred in 1 of 80 (1.25%) against 97 of 4,411 (2.20%), an adjusted relative risk of 0.50 (95% confidence interval 0.07 to 3.39), while rates of ordinary nausea and vomiting were the same (adjusted relative risk 1.06, 0.94 to 1.19) ⁶⁷. A confidence interval from 0.07 to 3.39 is compatible with a large benefit and with harm; the study was not powered to find a difference and says so ⁶⁷. A 2026 French review of the question states what is true as of this page: the mechanistic rationale is compelling, the clinical signals are encouraging, and no randomized trial has demonstrated preventive efficacy, so no practice recommendation can be made ³⁰.

The hormone's time course was mapped in 2026. In two Danish early-pregnancy cohorts with longitudinal sampling, GDF15 in normal pregnancies rose dramatically until week 7 and plateaued by week 9, fitting a sigmoidal (S-shaped) curve; pregnancies with female fetuses ran about 19% higher, the rise came a day later with male fetuses, symptom scores peaked at week 8 and declined as the plateau arrived, and in women with hyperemesis GDF15 had already plateaued by the time of diagnosis, higher than normal in the early window but with overlapping confidence intervals by week 8 ³¹. Pregnancies that ended in early loss showed lower and slower rises ³¹. The overlap is the clinically important finding: a single GDF15 level cannot diagnose hyperemesis, which is exactly what a mechanism built on sensitivity rather than concentration predicts ^4,31. A separate 2026 research letter from the HER Foundation and Fejzo's group reports glucocorticoid treatment (the class of steroid drugs that includes prednisone and methylprednisolone) associated with lower GDF15 concentrations and improved symptoms, which, if confirmed, would give the oldest rescue drug a mechanism; the letter's abstract is not indexed and the finding should be treated as preliminary ⁸³.

Drugs that act on the pathway exist. Ponsegromab, a humanized monoclonal antibody (engineered so the immune system treats it as human) that neutralizes GDF15, was tested in a Pfizer-funded phase 2 trial (an early test of dose and effect in patients) in 187 patients with cancer cachexia and GDF15 of 1,500 pg/mL or more: at 12 weeks the 100 mg, 200 mg and 400 mg groups gained a median of 1.22, 1.92 and 2.81 kg more than placebo, with improved appetite and activity in the highest-dose group, which the authors took as confirming GDF15 as a driver of cachexia ⁸⁴. A 2025 review of cachexia drugs lists ponsegromab, AV-380 and NGM120 among the agents targeting GDF15, with ponsegromab entering phase 3 trials ⁸⁵. A 2026 review of the medicinal chemistry describes GFRAL agonists (drugs that activate the receptor) and antagonists (drugs that block it), approaches to bind and eliminate circulating GDF15, and patents and ongoing clinical trials targeting the GDF15-GFRAL-RET axis, and lists hyperemesis among the conditions in which GDF15 acts ⁸¹. Whether any of these has reached pregnancy is not settled by the indexed literature: a 2026 case report states, without naming the drug or the trial, that a phase 2 trial of an antibody-based approach to the pathway in hyperemesis has reported encouraging preliminary findings ⁸², but no results of such a trial were found in PubMed on 19 October 2026 and the registry record identified for this page could not be opened, so the report is recorded here and not relied on. The specific problem for pregnancy is laid out in a 2024 review: the drug must reach the mother's hindbrain without reaching the fetus, and proposed answers include non-IgG immunotherapies (IgG is the antibody class actively carried across the placenta to the fetus), small peptide and small-molecule antagonists, and transdermal patches, none of which that review could report as tested in a pregnant woman ³⁴. Whether blocking the hormone is wise at all is argued by the evolutionary biologists: Crespi's 2024 analysis sets out evidence that fetal GDF15 production is typically adaptive for the fetus, through enhanced placental invasion, fewer miscarriages and preterm births and higher birth weight, so that the symptom is one side of a maternal-fetal conflict ⁸⁶, while Stadtmauer's 2025 review re-evaluates the prophylactic, metabolic-rewiring, placental-growth and anti-rejection hypotheses and argues that the question hinges on whether GDF15 also acts locally through non-canonical receptors, receptors other than GFRAL ⁷⁸. The 2026 Nature Genetics finding that the fetal and maternal GDF15 effects point in opposite directions is the genetic form of that conflict ⁷.

Treatment and its evidence

The honest summary of the trial literature is the 2016 Cochrane review: 25 randomized trials involving 2,052 women covered 18 different comparisons, most from single small studies, the quality of evidence for most outcomes was low or very low, and there was little high-quality and consistent evidence supporting any one intervention ⁸⁷. The guidelines are therefore built partly on safety data from large cohorts and partly on consensus. The 2024 RCOG guideline recommends that first-line antiemetics such as H1 antihistamines, phenothiazines and doxylamine-pyridoxine be prescribed initially (grade A); that ondansetron, safe and effective, should not be discouraged as second line if first-line drugs fail (grade B); that metoclopramide is safe and effective (grade B) but second line because of extrapyramidal effects, with intravenous doses given over at least 3 minutes (grade C); that combinations be used when a single drug fails (good practice point); and that intravenous rehydration be normal saline with potassium chloride in each bag, guided by daily electrolytes (grade C) ¹⁰. Where rehydration is given matters too: a 2025 systematic review of costs found outpatient or day-stay units the more cost-effective model of care ³⁵, and the authors of the English registry study argued that earlier management in gynecology day-case units or general practice could help prevent admissions ³. The 2018 practice bulletin of the American College of Obstetricians and Gynecologists (ACOG), number 189, remains the United States document and is discussed in the guidelines section ⁸⁸.

Pyridoxine, vitamin B6, has one good trial: 342 Thai women at 17 weeks or less were randomized to 30 mg a day or placebo, and nausea scores fell significantly more with pyridoxine (P = 0.0008) while the reduction in vomiting episodes did not reach significance (P = 0.0552) ⁸⁹. Ginger has twelve: a 2014 meta-analysis of 12 trials in 1,278 pregnant women found ginger improved nausea compared with placebo by a mean difference of 1.20 points (95% confidence interval 0.56 to 1.84), did not significantly reduce vomiting episodes (0.72, −0.03 to 1.46), seemed to work better below 1,500 mg a day, and did not significantly raise the risk of spontaneous abortion, though the estimate was imprecise (relative risk 3.14, 0.65 to 15.11), with the authors cautioning that the evidence was of low quality ⁹⁰. Both were tested in ordinary nausea and vomiting of pregnancy, not in hyperemesis, and the Cochrane review found that in 92 women with hyperemesis vitamin B6 was associated with a slightly longer hospital stay than placebo (0.80 days) and no clear difference in vomiting ⁸⁷. A 2026 network meta-analysis of 24 placebo-controlled trials in 3,017 women with ordinary nausea and vomiting of pregnancy ranked quince, vitamin B6 combined with pomegranate and mint, P6 acupressure (pressure on a point on the inner wrist), dimenhydrinate and acupuncture combined with doxylamine-pyridoxine highest, with considerable uncertainty, low-to-moderate certainty of evidence and sparse reporting of harms ⁹¹. Pyridoxine is not harmless in excess: a 2024 systematic review counted neurological symptoms such as burning, tingling and unsteadiness in 164 of 1,226 people taking excessive vitamin B6, although its fetal concerns rest on 5 losses among 245 women ⁹². Doxylamine-pyridoxine's evidence, and its reanalysis, are set out in the Bendectin section ^15,63. Pyridoxine and ginger each have their own Atlas entry.

Ondansetron, a blocker of the serotonin 5-HT3 receptor (5-HT, 5-hydroxytryptamine, is serotonin's chemical name) that has its own Atlas page, has the best head-to-head trials and the most argued safety record. In a double-blind trial of 36 women, 4 mg ondansetron beat 25 mg pyridoxine with 12.5 mg doxylamine over 5 days, the median fall in nausea on a 100 mm scale being 51 mm against 20 mm (P = .019), though 13 of 18 completed follow-up on ondansetron against 17 of 18 on the comparator ⁹³. Against metoclopramide in 160 women with hyperemesis given intravenous 4 mg ondansetron or 10 mg metoclopramide every 8 hours for 24 hours, well-being and vomiting episodes were the same (median 1 against 2 episodes, P = .38), but drowsiness (12.5% against 30%), dry mouth (10.0% against 23.8%) and persistent ketonuria were less frequent with ondansetron, and metoclopramide was much cheaper ⁹⁴; a smaller Iranian trial of 83 women over two weeks found less vomiting with ondansetron (P = 0.042) and no difference in nausea ⁹⁵. The placebo-controlled evidence is thinner: in the 2025 Danish VOMIT trial (Validating the effect of Ondansetron and Mirtazapine In Treating hyperemesis gravidarum), the difference between ondansetron and placebo in PUQE-24 change at day 2 was −0.51 points (95% confidence interval −2.32 to 1.30), not significant, and the trial's authors note that ondansetron is recommended as second line despite a lack of placebo-controlled trials ⁶⁹. On safety, a Danish registry study of 608,385 pregnancies found no increase in spontaneous abortion, stillbirth, major birth defects (2.9% against 2.9%), preterm delivery or low birth weight ⁶⁴; a study of 1,816,414 United States Medicaid pregnancies with 88,467 exposed in the first trimester found no association with cardiac malformations (adjusted relative risk 0.99) or malformations overall (1.01) but a small increase in oral clefts, 14.0 against 11.1 per 10,000 births, relative risk 1.24 (1.03 to 1.48), an absolute excess of 2.7 per 10,000 ¹⁶; and a 456,963-pregnancy cohort across Canada, the United States and the United Kingdom that compared ondansetron with other antiemetics, to avoid confounding by the illness itself, found no excess of fetal death (hazard ratio 0.91), spontaneous abortion, stillbirth or major malformations (odds ratio 1.06, 0.91 to 1.22) ⁹⁶. Two 2026 papers add to this: a scoping review of 39 studies found no elevated risk of congenital anomalies overall and conflicting but minimal absolute changes in risk for cardiac and orofacial defects ⁹⁷, and a MotherToBaby cohort of 2,711 pregnancies found no excess of major malformations with antiemetics overall, but ondansetron was associated with small-for-gestational-age birth, a birth weight below the 10th percentile for the length of the pregnancy (adjusted relative risk 1.80, 95% confidence interval 1.23 to 2.60, against untreated nausea and vomiting), which the authors suggest may at least partly reflect the greater severity of illness in women given it ⁹⁸. The RCOG's reading is that women can be reassured regarding a very small increase in the absolute risk of orofacial clefting, to be balanced against the risks of poorly managed hyperemesis ¹⁰. The full malformation debate is on the 5-HT3 receptor page.

Metoclopramide's safety rests on the same Danish registry: among 28,486 women exposed in the first trimester between 1997 and 2011, major malformations occurred in 25.3 per 1,000 against 26.6 per 1,000 unexposed, a prevalence odds ratio of 0.93 (0.86 to 1.02), with no signal in any of 20 malformation categories, no excess of stillbirth and, as with ondansetron, a lower rate of recorded spontaneous abortion (hazard ratio 0.35, 0.33 to 0.38) ⁹⁹. That is not a protective effect; it probably reflects, though the abstract does not say so, that women prescribed an antiemetic have often already passed the weeks when most losses occur. Its problem is tolerability: drowsiness and dry mouth were roughly twice as common as with ondansetron in the single trial analyzed by Cochrane for these outcomes (relative risks 2.40 and 2.38, moderate-quality evidence) ^87,94, and the extrapyramidal movement disorders are why RCOG places it second line ¹⁰.

Corticosteroids are the rescue drug, and the evidence is split by endpoint. In the one placebo-controlled trial of size, 126 women who had failed outpatient treatment were randomized to intravenous methylprednisolone 125 mg followed by a prednisone taper or to identical placebo, all receiving promethazine and metoclopramide; among the 110 with outcomes, 19 in each group were rehospitalized, 34% against 35% (P = .89) ¹⁰⁰. The Cochrane pooling of four studies in 269 women nevertheless found a lower readmission rate with corticosteroids (relative risk 0.69, 0.50 to 0.94), with no difference in days of admission in the one trial that reported it (very low-quality evidence), and the 2026 research letter associating glucocorticoids with lower GDF15 offers a mechanism that has yet to be tested ^83,87. Two newer drugs have small positive trials. Gabapentin is a drug most often used for nerve pain that also substantially reduces chemotherapy-induced and postoperative nausea and vomiting ¹⁰¹. In a double-blind trial that enrolled 31 women with refractory hyperemesis, meaning hyperemesis not responding to standard treatment and needing intravenous fluids, and obtained the primary outcome in 21, it reduced PUQE scores over days 5 to 7 by 52% more than ondansetron or metoclopramide (95% confidence interval 16 to 88%, P = .01), in what the authors themselves called a small trial ¹⁰¹. Mirtazapine, an antidepressant with antiemetic action, was tested in the VOMIT trial in 7 Danish hospitals from 2019 to 2022: of 59 women randomized, mirtazapine (n = 21) lowered the PUQE-24 at day 2 more than placebo (n = 20), a difference of −1.86 (95% confidence interval −3.61 to −0.12, P = .04); against ondansetron the difference was −1.35 (−3.10 to 0.40), not significant; the benefit widened after day 4, and mild to moderate adverse events were more frequent ⁶⁹. The trial had planned 180 women and enrolled 59, so its estimates are imprecise ⁶⁹. Acupuncture with auricular therapy, stimulation of points on the outer ear, added to fluids in a 122-woman Chinese trial, improved quality-of-life scores and shortened the median stay from 6 to 4 days but did not change the PUQE, and there was no sham control, a fake procedure that mimics the real one, so the contextual effect, benefit from the attention and expectation of treatment, cannot be separated ¹⁰².

Thiamine, Wernicke's encephalopathy and feeding

The single most preventable catastrophe of the condition is Wernicke's encephalopathy, the brain injury of thiamine deficiency, which has its own Atlas page under thiamine. Pregnancy raises the demand for thiamine, vomiting empties the body's small store, and the systematic review of 146 case reports describing 177 women gives the natural history: thiamine depletion between 10 and 15 weeks of gestation, a median of 7 weeks of vomiting and 12.1 kg of weight loss before the injury, with nausea and vomiting in 100%, double vision in 37.4% and blurred vision in 27.4% as the warning signs ¹⁷. The failures were of treatment: 63.6% of the women had been given a subtherapeutic dose of thiamine, and in 14.1% the encephalopathy was worsened by intravenous glucose, because glucose metabolism depends on thiamine pyrophosphate, the active form of the vitamin, as an enzyme cofactor, and a glucose load raises the demand on an already empty store ¹⁷. The outcomes were chronic cognitive disorders in 65.4%, pregnancy loss in 50% and maternal death in 5% ¹⁷. The authors recommended 100 mg of intravenous or intramuscular thiamine for every woman with persistent or severe late-onset vomiting, and the 2024 RCOG guideline, at grade D, gives thiamine, either 100 mg three times daily by mouth or intravenously as part of a vitamin B complex, to all women admitted with vomiting or severely reduced intake, especially before dextrose or parenteral nutrition ^10,17. Case reports from 2026 show the injury still happens, and in unexpected forms: acute bilateral visual loss at 16 weeks in a Malaysian woman that reversed within 24 hours of intravenous thiamine, and refeeding syndrome with magnetic resonance changes in the thalami and mammillary bodies (deep brain structures for relaying signals and forming memories, the classic sites of Wernicke's injury) in a British woman after medical management of a miscarriage at 16 weeks, with refractory low potassium and magnesium as the early warning ^48,104.

Feeding by tube does not help when it is routine. The open-label MOTHER trial, in 19 Dutch hospitals between October 2014 and March 2016, randomized 116 women hospitalized for hyperemesis between 5 and 20 weeks to at least 7 days of enteral tube feeding in addition to standard intravenous rehydration and antiemetics, or to standard care alone, with birth weight as the primary outcome ¹⁰⁵. In the intention-to-treat analysis, mean birth weight was 3,160 ± 770 g with tube feeding against 3,200 ± 680 g without, a difference of −40 g (95% confidence interval −230 to 310 g, as published; the interval is not centered on the estimate), and maternal weight gain, length of stay, readmission, symptoms, quality of life, psychological distress, prematurity and small-for-gestational-age birth were all comparable ¹⁰⁵. Only 28 (47%) of the women allocated to tube feeding were treated according to protocol; the rest had the tube removed within 7 days, primarily (34%) because of its adverse effects, which the authors read as evidence that early routine tube feeding is poorly tolerated ¹⁰⁵. The trial does not say that a woman who cannot keep anything down should not be fed; it says that putting a tube into every admitted woman buys nothing. The Dublin intake study, with median energy at half of requirements and every measured nutrient below need, is the argument for assessing nutrition in each woman rather than assuming it ⁴⁴.

Parenteral nutrition, feeding through a vein, is the last resort, and no trial on this page has tested it. The 2024 RCOG guideline's instruction that thiamine be given especially before parenteral nutrition is the practical point ¹⁰, and a 2026 narrative review, which adds no new data, recommends enteral tube feeding in preference to parenteral nutrition when oral intake is inadequate ¹⁰⁶. Refeeding syndrome, the fall in phosphate, potassium and magnesium when a starved body is fed and insulin drives the ions into cells, is the hazard of both routes, and the 2019 reading of Charlotte Brontë's death as refeeding syndrome after her appetite returned, whether or not it is right, is a reminder that the moment of recovery is dangerous ^48,57.

What ACOG and RCOG say now

The American document is Practice Bulletin No. 189 of the American College of Obstetricians and Gynecologists, published in Obstetrics and Gynecology in January 2018 ⁸⁸. Its stated purpose is to review the best available evidence on the diagnosis and management of nausea and vomiting of pregnancy; it notes that the condition is minimized and undertreated because morning sickness is common, that some women avoid treatment from fear of medication, and that early treatment may prevent progression and hospitalization ⁸⁸. A PubMed search on 19 October 2026 found no later ACOG practice bulletin on the subject, and the commentary by Fejzo, MacGibbon and Wisner published online in December 2022 explicitly presented its data as published since the most recent ACOG report of 2018, including the long-term neuropsychiatric findings in offspring and the GDF15 genetics ^23,88. Readers should check whether ACOG has reaffirmed or replaced the bulletin, since reaffirmation notices are not reliably indexed.

The British document is newer and more specific. The Royal College of Obstetricians and Gynaecologists published the revised Green-top Guideline No. 69, The Management of Nausea and Vomiting in Pregnancy and Hyperemesis Gravidarum, in BJOG in February 2024, with Catherine Nelson-Piercy as lead author and Caitlin Dean of the charity Pregnancy Sickness Support among the authors ¹⁰. Its graded recommendations are the spine of this page: PUQE and HELP for severity (grade C); ketonuria not an indicator of dehydration or severity (grade A); first-line H1 antihistamines, phenothiazines and doxylamine-pyridoxine, sold in the United Kingdom as Xonvea (grade A); ondansetron safe and effective as second line, with reassurance about a very small absolute increase in orofacial clefting (grade B); metoclopramide safe and effective but second line for extrapyramidal effects, given intravenously over at least 3 minutes (grade B and C); asking about previous adverse reactions and stopping drugs promptly if they occur (good practice point); normal saline with potassium chloride guided by daily electrolytes (grade C); combinations for non-responders (good practice point); thiamine for every admitted woman, especially before dextrose or parenteral nutrition (grade D); and the instruction that all therapeutic measures should have been tried before termination of pregnancy is considered (grade C) ¹⁰. In a 2016 paper the lead author declared personal fees from several pharmaceutical companies outside the submitted work ³; the 2024 guideline's own declarations of interest were not reviewed for this page.

The most current synthesis is the Lancet review of November 2025 by Nana, Painter, Williamson and Nelson-Piercy, which defines the condition by weight loss, dehydration, electrolyte imbalance and nutritional deficiency, states that women can be rendered so unwell that they are at increased risk of terminating a wanted pregnancy and of suicidal ideation, and sets out the decade's findings on cause, antiemetic safety and consequences for mother and child ²⁶. That definition is not the Windsor one: it names weight loss and electrolyte and nutritional derangement, which the 2021 consensus had left out of its criteria, so the field's two leading statements still describe the condition differently ^1,26. Its authors overlap with the RCOG guideline's, which is a strength for consistency and a reason to read both as one school's view ^10,26.

Where it leads: mother, child, and the decision to end a pregnancy

For the mother, the English cohort gives the population-level picture already set out: higher odds of anemia, pre-eclampsia, eclampsia, venous thromboembolism, induction, cesarean section and preterm birth, with lower odds of stillbirth and post-term delivery, and, although absolute risks were small, more resuscitation and intensive care for the newborn ³². Over the longer term, in 1,413,166 Quebec women followed for up to 32 years, hyperemesis alone was associated with a higher risk of later hospitalization for cardiovascular disease (hazard ratio 1.46, 95% confidence interval 1.38 to 1.54) ¹⁰⁷. For the child at birth, a 2011 meta-analysis found higher odds of low birth weight (odds ratio 1.42, 95% confidence interval 1.27 to 1.58), of being small for gestational age (1.28, 1.02 to 1.60) and of preterm birth (1.32, 1.04 to 1.68), with no association with Apgar scores (a 0-to-10 rating of a newborn's condition in the first minutes), congenital anomalies or perinatal death (stillbirth or death in the first week of life) ⁴⁵. These are associations in observational data; the illness, its treatment and whatever predisposes to it are entangled.

The longer-term findings in children are where the evidence reversed in 2025. The claim that the illness harms the child's brain had built up over a decade. In 469,789 Kaiser Permanente mother-child pairs in Southern California, children exposed to hyperemesis were diagnosed with autism more often (2.87 against 1.71 per 1,000 person-years; adjusted hazard ratio 1.53, 1.37 to 1.70) ¹⁰⁸; an exploratory follow-up of 267 exposed children reported autism in 22 (8%) and in none of 93 unexposed children ¹⁰⁹; and a 2022 systematic review of 19 studies found small increases in neurodevelopmental and mental-health disorders and possibly testicular cancer (odds ratio 1.60, 1.07 to 2.39), on evidence it called few studies of low quality ¹¹⁰. The commentary published online in December 2022 cited the autism and brain-volume findings as reasons for urgency ²³. A Quebec cohort of 1,189,000 children born between 2006 and 2021, of whom 6,904 (0.6%) were exposed to maternal hyperemesis requiring hospitalization, found a 1.21-fold risk of any hospitalization before age 16 (95% confidence interval 1.17 to 1.26), with hazard ratios of 1.50 for neurologic and 1.51 for developmental disorders ¹¹¹. The Singapore GUSTO (Growing Up in Singapore Towards healthy Outcomes) cohort found that children of the 190 mothers (16.2%) with severe nausea and vomiting had more externalizing and internalizing behaviors in early childhood, with effect sizes around a third of a standard deviation ¹¹². Then a New South Wales record-linkage study of 700,082 live births from 2002 to 2010, with 10,282 (1.5%) exposed, found the same associations, children below the national standard in reading (adjusted risk ratio 1.19), spelling (1.24) and numeracy (1.14) and more use of neurodevelopmental disability services (adjusted hazard ratio 1.44, 1.27 to 1.63), and then restricted the analysis to siblings discordant for exposure, where for most outcomes the associations were nullified ⁶⁸. Siblings share the family's genes, income, home and much else, so an association that vanishes between them was probably produced by those shared factors rather than by the illness; the authors called the result reassuring for mothers ⁶⁸. Fejzo replied in a letter titled "Relationship between hyperemesis gravidarum and poor school performance unresolved" ¹¹³, and the Quebec cohort, which had no sibling comparison, stands ¹¹¹. But the direction of the 2025 evidence is away from a direct effect on the child's brain.

It comes back. In the follow-up of the Dutch MOTHER cohort, 73 of the original 215 women answered: 35 (48%) had become pregnant again, 40% of them after deliberately postponing because of hyperemesis, the condition recurred in 89% of those 35 pregnancies, one woman terminated and eight (23%) considered it, and 24 of the 38 who had not become pregnant again said it was because of hyperemesis in the past ¹⁹. The online survey of 5,414 visitors to the HER Foundation's social media found an 86% recurrence risk ²⁰. Registry counts, built on diagnoses recorded in hospital, are far lower: 26.0% in Denmark ¹⁸. In England, a history of the condition raised the odds of admission in the next pregnancy 4.74-fold ³.

Termination for intractable symptoms is the outcome the condition's history did most to hide. In the 2007 survey of 808 women recruited through a hyperemesis website, 123 (15.2%) had ended at least one pregnancy because of it and 49 (6.1%) more than one; the reasons given were inability to care for family and self (66.7%), fear that they or the baby could die (51.2%) and fear that the baby would be abnormal (22.0%), and these women were more than twice as likely to describe their care as uncaring or uncomprehending (52.0% against 24.5%, odds ratio 3.34) ²⁵. A 2021 survey of more than 5,000 British women reported increased rates of termination and of suicidal ideation, and the 2025 Lancet review names termination of a wanted pregnancy and suicidal ideation as risks of the condition ^26,73, while the 2024 RCOG guideline instructs that all therapeutic measures be tried before termination is considered (grade C) ¹⁰. The burden outlasts the pregnancy: in a 2025 systematic review of 18 studies, adjusted odds ratios for postpartum depression ranged from 1.54 for mild to 6.52 for severe nausea and vomiting, study-level estimates rather than a pooled one ¹¹⁴, and a 2026 mixed-methods review of 64 studies found depression and anxiety generally higher in hyperemesis, sometimes persisting beyond pregnancy, with women reporting dismissal and lack of validation ¹¹⁵. Fejzo and colleagues put the policy consequence bluntly in the commentary published online in December 2022: until treatment is universally effective, abortion access must be available for refractory hyperemesis gravidarum ²³.

The reversal, and what is still unknown

The reversal this page exists to record is from it is in her head to it is a hormone from the placenta. The first claim was, on a 2002 reading of the literature, shaped by sociocultural assumption rather than scientific evidence; the second rests on human genetics that the illness cannot confound, on a receptor found in the brainstem's vomiting center and, by later accounts, nowhere else, on direct measurement of where the hormone comes from, on a natural experiment in women with thalassemia, and on an animal mechanism for why the same hormone makes some women desperately ill and others not at all ^4,21,36,40. Fejzo's 2024 review calls the old theories dispelled ²², and the 2025 Lancet review summarizes a decade of new insight into the causes ²⁶. What the reversal has not yet delivered is a treatment built on it: the drugs in use are the antihistamines and dopamine and serotonin blockers of the previous century, the metformin idea rests on a survey, a small cohort and a registry study whose signal appears only with long use, and no result from blocking GDF15 in a pregnant woman has been published ^{10,18,20,67,81,82}.

Three adjacent findings deserve a sentence each. Cannabis, which many women take for nausea, is associated with more of it: in 356,343 pregnancies in a California health system from 2011 to 2022, self-reported daily cannabis use before conception carried an adjusted odds ratio of 2.61 (95% confidence interval 2.40 to 2.84) for severe nausea and vomiting, and daily use in pregnancy 3.80 (3.28 to 4.39), and a 2026 case report in the American Journal of Clinical Nutrition is titled for the problem of telling cannabinoid hyperemesis syndrome from hyperemesis gravidarum, which has its own Atlas page ^116,117. The gut microbiome differs with severity, in a Swedish nested case-control study of 337 women with severe symptoms matched to moderate and mild, with lower richness and diversity in the sickest, but whether that is cause, consequence or marker is unresolved ¹¹⁸. And a 2026 survey of 1,010 women found that those with the most severe symptoms were more susceptible to motion sickness and visually induced motion sickness, with associations the authors call limited in magnitude, which fits a brainstem mechanism without proving one ¹¹⁹.

What is unknown is listed plainly. Why placental GDF15 production rose 100- to 1,000-fold in the catarrhine primates, and what it does for the fetus, are argued, not settled ^78,86. Whether raising GDF15 before conception prevents the disease awaits a randomized trial; the largest observational test so far found a signal only with long use ^18,30. Whether blocking it in pregnancy is safe for the fetus rests so far on eight human knockouts ⁶⁶ and on the argument that its receptor lies only in the hindbrain ⁴; no trial result in pregnancy has been published ⁸². Whether the six new loci of 2026 point to mechanisms beyond GDF15, in appetite, insulin signaling and brain plasticity, is a question the paper raises and does not answer ⁷. And whether the children of these pregnancies are harmed by the illness or by what the illness travels with is, after the 2025 sibling study, an open question that leans toward the latter ^68,111. The condition that was for much of the twentieth century the woman's fault is now a problem in endocrinology with a known hormone, a known receptor and no drug, which is a better place to be stuck.

The tests, one by one

• PUQE-24 score (Pregnancy-Unique Quantification of Emesis)Three questions about the past 24 hours: hours of nausea, episodes of vomiting, episodes of retching, each scored 1 to 5 for a total of 3 to 15 ⁸.The standard severity score in trials and clinics; endorsed, with the HELP score, by the 2024 guideline of the Royal College of Obstetricians and Gynaecologists (RCOG) at grade C ¹⁰, and adopted in 2026 as the core severity measure for hyperemesis research ¹¹In large prospective cohorts it predicted whether women could keep down multivitamins, their emergency visits and admissions, their health costs and their self-rated well-being ⁸.It rated only 58% of women who said that nothing goes or stays down as severe ⁹.
• HELP score (HyperEmesis Level Prediction)A longer questionnaire built by the patient-founded Hyperemesis Education and Research (HER) Foundation to capture what the PUQE misses: inability to eat or drink, weight loss, and the need for intravenous fluids ⁹.Endorsed alongside the PUQE in the 2024 RCOG guideline, at grade C ¹⁰It classified 92% of women reporting that nothing goes or stays down as severe, against 58% with the PUQE, and women who had been hospitalized were more likely to be rated severe by it ⁹.The validation was an online survey posted on hyperemesis websites, so it was tested in women who had already sought out the condition's community ⁹.
• Urine ketone dipstickKetones appear in urine when the body burns fat for lack of food; for decades most studies admitted only women who had them ¹².Still measured on admission in many hospitals, which the evidence argues against ^10,13In 5 studies that looked, ketonuria did not track the presence or severity of hyperemesis, and in 215 admitted Dutch women it predicted neither symptom severity, weight loss nor readmission ^12,13.Women with more ketones were kept in hospital 0.27 days longer per grade of ketonuria, suggesting clinicians were treating the dipstick rather than the patient; the 2024 RCOG guideline says it should not be used to assess severity, at grade A ^10,13.
• Thyroid function testsThyroid-stimulating hormone (TSH) was suppressed in 60% of 57 women with hyperemesis against 9% of matched controls, because hCG weakly stimulates the thyroid ⁴¹.To exclude true thyroid disease as a cause of vomitingThe degree of biochemical hyperthyroidism tracked the severity of vomiting in that 1992 series ⁴¹.Across 81 studies the association of thyroid hormones with hyperemesis was inconsistent ¹², and the 1992 data point to hCG as the thyroid stimulator, so the finding marks the placental hormone rather than thyroid disease ⁴¹.
• Serum GDF15The hormone itself, measured by immunoassay; in 791 Cambridge women at a median of 15.1 weeks it was 11,581 pg/mL (picograms, trillionths of a gram, per milliliter) in those reporting vomiting against 10,593 in those with no symptoms ⁴².Research only; no clinical cut-off existsIn two Danish cohorts GDF15 rose steeply to week 7 and plateaued by week 9, was about 19% higher with a female fetus, and was higher in hyperemesis early but overlapped with normal pregnancy by week 8 ³¹.Because the ranges overlap and the sensitivity of the mother matters as much as the level, a single value cannot diagnose the condition ^4,31.

Where it connects

Sources

119 sources, numbered as they are cited. Every one was checked against PubMed or its publisher before it was cited here; the note under each says what it shows and what it does not.

1. 1
   Jansen LAW, Koot MH, Van't Hooft J, et al. The windsor definition for hyperemesis gravidarum: A multistakeholder international consensus definition.Eur J Obstet Gynecol Reprod Biol · 2021 · 266:15-22doi:10.1016/j.ejogrb.2021.09.004 · PMID 34555550

   The Delphi consensus definition: onset before 16 weeks, severe nausea or vomiting, inability to eat or drink normally, strong limitation of daily life, dehydration contributory. It is a definition for research and diagnosis, not an outcome study.

2. 2
   Einarson TR, Piwko C, Koren G. Quantifying the global rates of nausea and vomiting of pregnancy: a meta analysis.J Popul Ther Clin Pharmacol · 2013 · 20(2):e171-83PMID 23863575

   79 studies: NVP 69.4% (66.5 to 72.3), nausea alone 32.7%, into third trimester 23.5%, mild 40%, moderate 46%, severe 14%; HG 1.1% (0.8 to 1.3), range 0.3 to 3.6%, from 6,155,578 pregnancies. No DOI is registered for this journal article. None of the Koren papers cited on this page is retracted or under an expression of concern in PubMed as of 19 October 2026; other papers of his have been retracted [94].

3. 3
   Fiaschi L, Nelson-Piercy C, Tata LJ. Hospital admission for hyperemesis gravidarum: a nationwide study of occurrence, reoccurrence and risk factors among 8.2 million pregnancies.Hum Reprod · 2016 · 31(8):1675-84doi:10.1093/humrep/dew128 · PMID 27251205

   England 1997 to 2012: 186,800 admissions in 121,885 of 8,215,538 pregnancies (1.5%); prior HG aOR 4.74; parathyroid, cholesterol, type 1 diabetes and thyroid comorbidities. Co-author C. Nelson-Piercy declares pharmaceutical personal fees and RCOG guideline co-development; the other authors declared none.

4. 4
   Fejzo M, Rocha N, Cimino I, et al. GDF15 linked to maternal risk of nausea and vomiting during pregnancy.Nature · 2024 · 625(7996):760-767 (online 13 Dec 2023)doi:10.1038/s41586-023-06921-9 · PMID 38092039

   The central paper: most maternal GDF15 is feto-placental by mass spectrometry; low non-pregnant GDF15 raises HG risk; beta-thalassemia women report little NVP; mice are desensitized by prior exposure. Authors call the causal role putative. Co-author K.W. MacGibbon is affiliated with the Hyperemesis Education and Research Foundation; the long-acting GDF15 for the mouse work was supplied by Pfizer under a material transfer agreement (full text). The competing-interests statement is published online and could not be opened for this page.

5. 5
   Zhang Y, Cantor RM, MacGibbon K, et al. Familial aggregation of hyperemesis gravidarum.Am J Obstet Gynecol · 2011 · 204(3):230.e1-7doi:10.1016/j.ajog.2010.09.018 · PMID 20974461

   Sisters OR 17.3 (P = .005); 33% of cases vs 7.7% of controls had an affected mother. Survey with friend controls; self-reported family history.

6. 6
   Vikanes A, Skjaerven R, Grjibovski AM, Gunnes N, Vangen S, Magnus P. Recurrence of hyperemesis gravidarum across generations: population based cohort study.BMJ · 2010 · 340:c2050doi:10.1136/bmj.c2050 · PMID 21030362

   Norway 1967 to 2006: daughters born of HG pregnancies 3% vs 1.1% (OR 2.9, 2.4 to 3.6); sons' partners 1.2% (OR 1.0). Maternal genotype over fetal, with maternal-line environment not excluded.

7. 7
   Fejzo M, Wang X, Tan Q, et al. Multi-ancestry genome-wide association study of severe pregnancy nausea and vomiting.Nat Genet · 2026 · 58(4):810-820doi:10.1038/s41588-026-02564-4 · PMID 41981316

   10,974 cases, 461,461 controls: ten loci including GDF15, IGFBP7, PGR, GFRAL and six new; GDF15 in extravillous trophoblast with opposing maternal and fetal effects; selected loci associated with pregnancy weight gain, duration, birth weight and pre-eclampsia. A publisher correction was issued in May 2026.

8. 8
   Koren G, Piwko C, Ahn E, et al. Validation studies of the Pregnancy Unique-Quantification of Emesis (PUQE) scores.J Obstet Gynaecol · 2005 · 25(3):241-4doi:10.1080/01443610500060651 · PMID 16147725

   PUQE predicted multivitamin tolerance, emergency visits and admissions, costs and well-being in prospective cohorts. From the Motherisk program. None of the Koren papers cited on this page is retracted or under an expression of concern in PubMed as of 19 October 2026; other papers of his have been retracted [94].

9. 9
   MacGibbon KW, Kim S, Mullin PM, Fejzo MS. HyperEmesis Level Prediction (HELP Score) Identifies Patients with Indicators of Severe Disease: a Validation Study.Geburtshilfe Frauenheilkd · 2021 · 81(1):90-98doi:10.1055/a-1309-1997 · PMID 33487669

   Online survey validation: HELP classified 92% of women reporting nothing stays down as severe against 58% for PUQE. Sample recruited from hyperemesis websites; the developers are the validators.

10. 10
    Nelson-Piercy C, Dean C, Shehmar M, et al. The Management of Nausea and Vomiting in Pregnancy and Hyperemesis Gravidarum (Green-top Guideline No. 69).BJOG · 2024 · 131(7):e1-e30doi:10.1111/1471-0528.17739 · PMID 38311315

    RCOG guideline, February 2024: PUQE and HELP (C); ketonuria not an indicator (A); first-line antihistamines, phenothiazines, doxylamine-pyridoxine (A); ondansetron second line with cleft reassurance (B); metoclopramide second line (B, C); saline with potassium (C); thiamine for all admitted (D); all measures before termination (C).

11. 11
    Jansen LAW, van der Minnen LM, Blaauwgeers A, Dean C, Gauw N, Van't Hooft J, et al. A core outcome measurement set for hyperemesis gravidarum research: a consensus study.BMC Pregnancy Childbirth · 2026 · 26(1)doi:10.1186/s12884-025-07937-x · PMID 42399837

    Systematic review of 362 instruments in 461 studies, then a modified Delphi of 155 stakeholders from 10 countries (51% women or families with lived experience): consensus on 24 instruments, including the PUQE-24 for severity, for all future HG trials.

12. 12
    Niemeijer MN, Grooten IJ, Vos N, et al. Diagnostic markers for hyperemesis gravidarum: a systematic review and metaanalysis.Am J Obstet Gynecol · 2014 · 211(2):150.e1-15doi:10.1016/j.ajog.2014.02.012 · PMID 24530975

    81 articles, 9 biomarkers: no association of ketonuria with presence or severity in 5 studies; H. pylori OR 3.2 (2.0 to 5.1) with sensitivity 73% and specificity 55%; hCG, thyroid hormones and others inconsistent.

13. 13
    Koot MH, Grooten IJ, Post JAMV, et al. Ketonuria is not associated with hyperemesis gravidarum disease severity.Eur J Obstet Gynecol Reprod Biol · 2020 · 254:315-320doi:10.1016/j.ejogrb.2020.08.014 · PMID 33045502

    215 admitted women in 19 Dutch hospitals: ketonuria predicted neither symptoms, quality of life, weight loss nor readmission, but each grade added 0.27 days (0.05 to 0.48) to the stay.

14. 14
    Nuangchamnong N, Niebyl J. Doxylamine succinate-pyridoxine hydrochloride (Diclegis) for the management of nausea and vomiting in pregnancy: an overview.Int J Womens Health · 2014 · 6:401-9doi:10.2147/IJWH.S46653 · PMID 24748822

    States that Diclegis, approved April 2013, is the only FDA-approved medication for NVP and qualified for FDA Pregnancy Category A, under the letter system then in use, on the strength of accumulated safety data. A 2014 statement, not a current regulatory finding.

15. 15
    Persaud N, Meaney C, El-Emam K, Moineddin R, Thorpe K. Doxylamine-pyridoxine for nausea and vomiting of pregnancy randomized placebo controlled trial: Prespecified analyses and reanalysis.PLoS One · 2018 · 13(1):e0189978doi:10.1371/journal.pone.0189978 · PMID 29342163

    Reanalysis from the Health Canada clinical study report: 0.73 points (0.21 to 1.25) with last observation carried forward, 0.38 (−0.08 to 0.84) with complete data, against a prespecified 3-point minimal clinically important difference. The reversal on efficacy. Its full text documents that Duchesnay sponsored the trial, the disclosed Duchesnay ties of the lead author and three co-authors, the promotion of a secondary outcome, the FDA summary review wording and the 2016 Health Canada review.

16. 16
    Huybrechts KF, Hernández-Díaz S, Straub L, et al. Association of Maternal First-Trimester Ondansetron Use With Cardiac Malformations and Oral Clefts in Offspring.JAMA · 2018 · 320(23):2429-2437doi:10.1001/jama.2018.18307 · PMID 30561479

    1,816,414 Medicaid pregnancies, 88,467 exposed: cardiac RR 0.99, any malformation 1.01, oral clefts RR 1.24 (1.03 to 1.48), absolute excess 2.7 per 10,000. The source of the cleft signal.

17. 17
    Oudman E, Wijnia JW, Oey M, van Dam M, Painter RC, Postma A. Wernicke's encephalopathy in hyperemesis gravidarum: A systematic review.Eur J Obstet Gynecol Reprod Biol · 2019 · 236:84-93doi:10.1016/j.ejogrb.2019.03.006 · PMID 30889425

    177 cases: depletion at 10 to 15 weeks, median 7 weeks of vomiting, 12.1 kg lost; 63.6% underdosed, 14.1% worsened by glucose; cognitive disorder 65.4%, pregnancy loss 50%, maternal death 5%. Case-report synthesis, so biased toward severe cases.

18. 18
    van der Minnen LM, Vendelbo JH, Pedersen LH, Painter RC. Preconception metformin and hyperemesis gravidarum in Denmark: clinical implications of growth differentiation factor 15's etiological role.Am J Obstet Gynecol · 2026 · 234(5):1502-1515 (online 30 Dec 2025)doi:10.1016/j.ajog.2025.12.060 · PMID 41478545

    Danish national registry, 1,447,968 pregnancies of 10 weeks or more, 1998 to 2018: hospital-coded HG in 1.9%, recurrence 26.0%. Any preconception metformin did not lower prevalence (1.9%) or recurrence (27%); use for 6 months or more, 1.1% (aOR 0.49, 0.31 to 0.75) against use under 2 months; use continued to conception, 1.2% (aOR 0.61, 0.38 to 0.95). Observational; the authors call for randomized trials.

19. 19
    Nijsten K, Dean C, van der Minnen LM, et al. Recurrence, postponing pregnancy, and termination rates after hyperemesis gravidarum: Follow up of the MOTHER study.Acta Obstet Gynecol Scand · 2021 · 100(9):1636-1643doi:10.1111/aogs.14197 · PMID 34033123

    73 of 215 followed: 48% pregnant again, 40% after postponing; recurrence 89%; one termination, eight considered; 24 of 38 avoided pregnancy because of HG. Small, with a third responding.

20. 20
    Sharma N, MacGibbon KW, Brecht-Doscher A, Cortessis VK, Fejzo MS. Prepregnancy metformin use associated with lower risk of severe nausea and vomiting of pregnancy and hyperemesis gravidarum.Am J Obstet Gynecol · 2025 · 233(6):649.e1-649.e14doi:10.1016/j.ajog.2025.06.055 · PMID 40588059

    Online survey of 5,414 HER Foundation visitors: metformin before first pregnancy aRR 0.29 (0.12 to 0.71); second pregnancy aOR 0.18; tobacco protective; SSRIs and cannabis associated with higher risk; 86% recurrence. Retrospective, self-selected, calls for trials.

21. 21
    Munch S. Chicken or the egg? The biological-psychological controversy surrounding hyperemesis gravidarum.Soc Sci Med · 2002 · 55(7):1267-78doi:10.1016/s0277-9536(01)00239-8 · PMID 12365536

    Literature analysis concluding that biologic and psychologic explanations ran in parallel throughout the condition's history and that sociocultural factors rather than evidence made psychogenesis dominant.

22. 22
    Fejzo MS. Hyperemesis gravidarum theories dispelled by recent research: a paradigm change for better care and outcomes.Trends Mol Med · 2024 · 30(6):530-540doi:10.1016/j.molmed.2024.04.006 · PMID 38782680

    The reversal stated by its principal author: strong evidence of poor outcomes, a common cause identified, NVP a protective evolutionary mechanism no longer necessary.

23. 23
    Fejzo MS, MacGibbon KW, Wisner KL. Pregnant, miserable, and starving in 21st century America.AJOG Glob Rep · 2023 · 3(1):100141 (online 5 Dec 2022)doi:10.1016/j.xagr.2022.100141 · PMID 36536797

    Commentary: second most common reason for hospitalization in pregnancy, maternal deaths still reported, data since the 2018 ACOG report, and the position that abortion access must remain available for refractory disease.

24. 24
    Weiss G. The death of Charlotte Brontë.Obstet Gynecol · 1991 · 78(4):705-8PMID 1923177

    Argues that the hyperemesis attribution, itself framed as neurosis and rejection of femininity, is doubtful, that she may never have been pregnant, and that tuberculosis with Addison disease explains her death. No DOI registered.

25. 25
    Poursharif B, Korst LM, Macgibbon KW, Fejzo MS, Romero R, Goodwin TM. Elective pregnancy termination in a large cohort of women with hyperemesis gravidarum.Contraception · 2007 · 76(6):451-5doi:10.1016/j.contraception.2007.08.009 · PMID 18061703

    808 women via a hyperemesis website: 15.2% had terminated at least once for HG, 6.1% more than once; those women three times as likely to call providers uncaring (OR 3.34). Self-selected sample.

26. 26
    Nana M, Painter R, Williamson C, Nelson-Piercy C. Hyperemesis gravidarum.Lancet · 2026 · 407(10523):78-89 (online 13 Nov 2025)doi:10.1016/S0140-6736(25)01454-0 · PMID 41242332

    The current major review: definition by weight loss, dehydration, electrolyte and nutritional derangement; termination of wanted pregnancies and suicidal ideation; dismissal of women; decade of findings on cause, drug safety and consequences.

27. 27
    Vargesson N. Thalidomide-induced teratogenesis: history and mechanisms.Birth Defects Res C Embryo Today · 2015 · 105(2):140-56doi:10.1002/bdrc.21096 · PMID 26043938

    History: prescribed for morning sickness nearly 60 years earlier, over 10,000 children malformed, current use in myeloma and leprosy, new cases in Brazil; reviews candidate mechanisms.

28. 28
    Brent RL. Bendectin: review of the medical literature of a comprehensively studied human nonteratogen and the most prevalent tortogen-litigen.Reprod Toxicol · 1995 · 9(4):337-49doi:10.1016/0890-6238(95)00020-b · PMID 7496090

    Five-part analysis concluding no measurable teratogenic effect; criticizes plaintiffs' experts. The review evaluates the experts for the defense and the plaintiffs; the author's own role in the litigation is not stated in the indexed record.

29. 29
    Madjunkova S, Maltepe C, Koren G. The delayed-release combination of doxylamine and pyridoxine (Diclegis®/Diclectin ®) for the treatment of nausea and vomiting of pregnancy.Paediatr Drugs · 2014 · 16(3):199-211doi:10.1007/s40272-014-0065-5 · PMID 24574047

    FDA approval April 2013 as Diclegis; on the Canadian market since 1979 as Diclectin; more than 200,000 first-trimester exposures without excess malformations. From the Motherisk program, whose director led the pivotal trial. None of the Koren papers cited on this page is retracted or under an expression of concern in PubMed as of 19 October 2026; other papers of his have been retracted [94].

30. 30
    Deruelle P, Arcieri F, Martineau R. [Metformin for the prevention of hyperemesis gravidarum: From a GDF15-based rationale to early clinical evidence].Gynecol Obstet Fertil Senol · 2026 · in pressdoi:10.1016/j.gofs.2026.06.002 · PMID 42250761

    French review: compelling rationale, encouraging signals, no randomized trial, no practice recommendation possible.

31. 31
    Svenningsen JS, Ostenfeld A, Petersen JF, et al. Early pregnancy GDF15 trajectories and their association with normal gestation, pregnancy loss, and hyperemesis gravidarum.EBioMedicine · 2026 · 131:106428doi:10.1016/j.ebiom.2026.106428 · PMID 42561814

    Longitudinal Danish cohorts: GDF15 rises to week 7, plateaus by week 9; 19% higher with female fetus; HG higher early but overlapping by week 8; early loss lower and slower. Funded by the Independent Research Fund Denmark.

32. 32
    Fiaschi L, Nelson-Piercy C, Gibson J, Szatkowski L, Tata LJ. Adverse Maternal and Birth Outcomes in Women Admitted to Hospital for Hyperemesis Gravidarum: a Population-Based Cohort Study.Paediatr Perinat Epidemiol · 2018 · 32(1):40-51doi:10.1111/ppe.12416 · PMID 28984372

    8,211,850 pregnancies: higher odds of anemia, pre-eclampsia, eclampsia, VTE, preterm and low birth weight; lower odds of stillbirth and post-term birth. Observational; 99% confidence intervals.

33. 33
    Fejzo MS, Trovik J, Grooten IJ, et al. Nausea and vomiting of pregnancy and hyperemesis gravidarum.Nat Rev Dis Primers · 2019 · 5(1):62doi:10.1038/s41572-019-0110-3 · PMID 31515515

    The pre-definition primer: NVP up to 70%, HG 0.3 to 10.8% depending on definition; already names GDF15, IGFBP7, GFRAL and PGR.

34. 34
    Thygerson J, Oyler D, Thomas J, Muse B, Brooks BD, Pullan JE. GDF15 Targeting for Treatment of Hyperemesis Gravidarum.Medicines (Basel) · 2024 · 11(7):17doi:10.3390/medicines11070017 · PMID 39311313

    Review: US annual costs exceeding 1.7 billion dollars; proposes non-IgG immunotherapies, small peptides and molecules, transdermal delivery to keep drug from the fetus. Proposals, not data.

35. 35
    Gazal A, Scarf V, Kay-Smith C, Gidaszewski B, Chua S, Levett K. What are the costs associated with nausea and vomiting in pregnancy and hyperemesis gravidarum? A systematic review.BMC Pregnancy Childbirth · 2025 · 25(1):807doi:10.1186/s12884-025-07921-5 · PMID 40750873

    Seven costing studies from 652 screened, methodologically and clinically heterogeneous: all found a significant economic burden that rose with severity; outpatient or day-stay units were the more cost-effective models of care. One author is from the charity Hyperemesis Australia.

36. 36
    Mullican SE, Lin-Schmidt X, Chin CN, et al. GFRAL is the receptor for GDF15 and the ligand promotes weight loss in mice and nonhuman primates.Nat Med · 2017 · 23(10):1150-1157doi:10.1038/nm.4392 · PMID 28846097

    Identifies GFRAL as the GDF15 receptor, expressed in area postrema and nucleus of the solitary tract in mice and humans, requiring RET; from Janssen, an obesity-drug program.

37. 37
    Borner T, Pataro AM, De Jonghe BC. Central mechanisms of emesis: A role for GDF15.Neurogastroenterol Motil · 2025 · 37(3):e14886doi:10.1111/nmo.14886 · PMID 39108013

    Review of GDF15 as a cause of anorexia through nausea and emesis via hindbrain GFRAL, with hyperemesis, chemotherapy and cyclic vomiting as contexts.

38. 38
    Borner T, Shaulson ED, Ghidewon MY, et al. GDF15 Induces Anorexia through Nausea and Emesis.Cell Metab · 2020 · 31(2):351-362.e5doi:10.1016/j.cmet.2019.12.004 · PMID 31928886

    GDF15 caused vomiting in musk shrews and pica and anorexia in mice and rats; contributes to chemotherapy malaise in mice. Animal mechanism only.

39. 39
    Fejzo M, Rocha N, Cimino I, et al. Fetally-encoded GDF15 and maternal GDF15 sensitivity are major determinants of nausea and vomiting in human pregnancy.bioRxiv · 2023 · preprint, 4 Jun 2023doi:10.1101/2023.06.02.542661 · PMID 37398065

    Preprint of the Nature paper, cited for details absent from the final abstract: the rare C211G variant impairs secretion and predisposes especially with a wild-type fetus; two common haplotypes lower non-pregnant levels; a long-acting dose blunts the acute response in mice. Not peer reviewed in this form.

40. 40
    Karusheva Y, Petry CJ, Yasara N, et al. Association of GDF15 levels with body mass index and endocrine status in β-thalassaemia.Clin Endocrinol (Oxf) · 2023 · 99(2):182-189doi:10.1111/cen.14897 · PMID 36806122

    103 Sri Lankan patients: GDF15 24.2-fold above healthy controls in beta-thalassemia major; the background to the pregnancy observation in the Nature paper.

41. 41
    Goodwin TM, Montoro M, Mestman JH, Pekary AE, Hershman JM. The role of chorionic gonadotropin in transient hyperthyroidism of hyperemesis gravidarum.J Clin Endocrinol Metab · 1992 · 75(5):1333-7doi:10.1210/jcem.75.5.1430095 · PMID 1430095

    57 cases vs 57 controls: TSH suppressed in 60% vs 9%, hCG correlated with free T4 and vomiting severity, serum thyrotropic activity in rat thyroid cells tracked hCG. The strongest form of the hCG hypothesis; association plus in vitro mechanism.

42. 42
    Petry CJ, Ong KK, Burling KA, et al. Associations of vomiting and antiemetic use in pregnancy with levels of circulating GDF15 early in the second trimester: A nested case-control study.Wellcome Open Res · 2018 · 3:123doi:10.12688/wellcomeopenres.14818.1 · PMID 30345390

    791 Cambridge women at median 15.1 weeks: GDF15 11,581 vs 10,593 pg/mL with vs without vomiting (P = 0.02); 13,157 in antiemetic users; strongly correlated with hCG, inversely with BMI. Modest association sampled after the symptom peak.

43. 43
    Fejzo M, Wang X, Zöllner J, et al. Multi-ancestry GWAS of severe pregnancy nausea and vomiting identifies risk loci associated with appetite, insulin signaling, and brain plasticity.Res Sq · 2024 · preprint, 16 Dec 2024doi:10.21203/rs.3.rs-5487737/v1 · PMID 39764105

    Preprint of the 2026 Nature Genetics study [43], not peer reviewed in this form; full text read. Reports the risk loci to be under significant evolutionary selection and constraint, a claim not in the published abstract, and states that the study found no association with the genes encoding hCG or estrogen or their receptors, so that further work on those hormones lacks genetic support.

44. 44
    Sandberg M, Kelly L, Cunningham A, Augustin H, O'Brien EC. Nutritional intakes and their association with symptom severity among women with Hyperemesis Gravidarum.Br J Nutr · 2026 · 1-9 (online 15 Oct 2026)doi:10.1017/S0007114526108496 · PMID 42876064

    36 Dublin women: median 1,050 kcal and 1,196 mL a day against 2,057 kcal and 2,300 mL required; lower intakes correlated with worse PUQE-24. Cross-sectional, small.

45. 45
    Veenendaal MVE, van Abeelen AFM, Painter RC, van der Post JAM, Roseboom TJ. Consequences of hyperemesis gravidarum for offspring: a systematic review and meta-analysis.BJOG · 2011 · 118(11):1302-13doi:10.1111/j.1471-0528.2011.03023.x · PMID 21749625

    Female child OR 1.27; low birth weight 1.42; small for gestational age 1.28; preterm 1.32; no association with Apgar, anomalies or perinatal death.

46. 46
    Mitchell-Jones N, Gallos I, Farren J, Tobias A, Bottomley C, Bourne T. Psychological morbidity associated with hyperemesis gravidarum: a systematic review and meta-analysis.BJOG · 2017 · 124(1):20-30doi:10.1111/1471-0528.14180 · PMID 27418035

    59 articles, 12 pooled: depression SMD 1.22 (0.80 to 1.64), anxiety SMD 0.86 (0.53 to 1.19), with heterogeneity of 94% and 84%. Shows association, not direction.

47. 47
    Kjeldgaard HK, Eberhard-Gran M, Benth JŠ, Vikanes ÅV. Hyperemesis gravidarum and the risk of emotional distress during and after pregnancy.Arch Womens Ment Health · 2017 · 20(6):747-756doi:10.1007/s00737-017-0770-5 · PMID 28842762

    Norwegian Mother and Child Cohort, 851 of 92,947 with HG: distress raised at 17 and 32 weeks and 6 months postpartum, gone by 18 months; the authors read distress as primarily a consequence.

48. 48
    Amoo WA, Sultana M, Slavska M. Refeeding Syndrome Complicated by Wernicke's Encephalopathy Following Severe Hyperemesis Gravidarum and Medical Management of Miscarriage: A Case Report.Cureus · 2026 · 18(6):e110587doi:10.7759/cureus.110587 · PMID 42434661

    Single case: refractory hypokalemia and hypomagnesemia, thalamic and mammillary body MRI changes, recovery with thiamine and electrolytes; illustrates refeeding risk.

49. 49
    Copeman E. A Novel Treatment of Obstinate Vomiting in Pregnancy.Br Med J · 1875 · 1(750):637-8doi:10.1136/bmj.1.750.637-a · PMID 20747863

    Primary historical document; no abstract and no full text indexed; cited for its title, the period name 'obstinate vomiting', only. The treatment it describes is not characterized here.

50. 50
    Munro Æ. Abstract of Communication on Vomiting of Pregnancy; with a Case Requiring the Induction of Labour.Trans Edinb Obstet Soc · 1875 · 3:94-101PMID 29613558

    Primary historical document indexed through PubMed Central with no abstract or DOI; cited for its title only, as a record that labor was induced for the condition.

51. 51
    Horrocks P. Severe Vomiting in Pregnancy: Miscarriage: Death, with Remarks.Br Med J · 1885 · 2(1279):13-4doi:10.1136/bmj.2.1279.13 · PMID 20751312

    Primary historical case report; no abstract indexed; cited for its title only.

52. 52
    Anonymous. Induced Abortion to Relieve Vomiting in Pregnancy.Daniel's Tex Med J · 1885 · 1(2):74-77PMID 36953476

    Primary historical document with no author listed in PubMed and no abstract or DOI; cited for its title only, as a record of abortion as a treatment.

53. 53
    Oliver JP. Pernicious Vomiting of Pregnancy and Its Treatment with Electricity.Tex Med J (Austin) · 1900 · 16(2):60-64PMID 36954737

    Primary historical document from 1900, indexed through PubMed Central with no abstract or DOI; cited for its title as a record of the name and the treatments of the era.

54. 54
    Maxwell RD. A Fatal Case of Toxic Vomiting of Pregnancy with Microscopical Sections of Liver and Kidneys.Proc R Soc Med · 1909 · 2(Obstet Gynaecol Sect):236-9doi:10.1177/003591570900200844 · PMID 19973829

    Primary historical case report; no abstract indexed; cited for its title only.

55. 55
    Williams JW. Further Contributions to Our Knowledge of the Pernicious Vomiting of Pregnancy.Glasgow Med J · 1912 · 78(6):401-427PMID 30434127

    Primary historical document, a lecture by the Johns Hopkins professor of obstetrics; indexed in PubMed through PubMed Central, no DOI and no abstract. Cited only for its title, author and date as evidence of the older name.

56. 56
    Fairweather DV. Nausea and vomiting in pregnancy.Am J Obstet Gynecol · 1968 · 102(1):135-75doi:10.1016/0002-9378(68)90445-6 · PMID 4877794

    A 41-page review often cited as the mid-century synthesis; no abstract is available and its contents are not summarized on this page.

57. 57
    Allison SP, Lobo DN. The death of Charlotte Brontë from hyperemesis gravidarum and refeeding syndrome: A new perspective.Clin Nutr · 2020 · 39(1):304-305doi:10.1016/j.clnu.2019.01.027 · PMID 30777294

    Accepts hyperemesis as most likely and proposes refeeding syndrome after her symptoms remitted in the third week of March 1855 as the cause of death. A reading of letters, not a clinical record.

58. 58
    Harvey WA, Sherfey MJ. Vomiting in pregnancy; a psychiatric study.Psychosom Med · 1954 · 16(1):1-9doi:10.1097/00006842-195401000-00001 · PMID 13134420

    A paper of the psychogenic era, cited for its title and date; no abstract is available and its methods are not characterized here.

59. 59
    Neutel CI, Johansen HL. Measuring drug effectiveness by default: the case of Bendectin.Can J Public Health · 1995 · 86(1):66-70PMID 7728721

    1983 voluntary withdrawal under pending suits; 10 to 25% of pregnancies exposed, 33 million in all; Canadian admissions for vomiting in pregnancy rose 37% in 1983 and 50% in 1984; no fall in malformations. No DOI registered.

60. 60
    Supreme Court of the United States. Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993).United States Reports · 1993 · 509:579

    Primary document, a court opinion that PubMed cannot hold; not indexed in PubMed. No online copy could be opened from the research environment used for this page, so the ruling is characterized only as the indexed sources [19][23] describe it: the Bendectin birth-defect suits reached the Court, which overruled the Frye general-acceptance test and made trial judges gatekeepers of scientific evidence.

61. 61
    Gvozdenovic M. Standard of Care in Medical Malpractice: Deference, Daubert, or Different Direction.J Law Med · 2022 · 29(4):1220-1235PMID 36763027

    Legal analysis stating that in Daubert v Merrell Dow the Supreme Court overruled Frye v United States and imposed a gatekeeper role on trial judges, aiming at more reliable expert evidence. No DOI registered.

62. 62
    Moore AG, Brown DA, Fairlie WD, et al. The transforming growth factor-ss superfamily cytokine macrophage inhibitory cytokine-1 is present in high concentrations in the serum of pregnant women.J Clin Endocrinol Metab · 2000 · 85(12):4781-8doi:10.1210/jcem.85.12.7007 · PMID 11134143

    First report of GDF15, then called MIC-1, in pregnancy: high and rising serum levels, large amounts in amniotic fluid and placenta, constitutive secretion by a trophoblast cell line; the authors suggest it may promote fetal survival by suppressing maternal proinflammatory cytokines.

63. 63
    Koren G, Clark S, Hankins GDV, et al. Effectiveness of delayed-release doxylamine and pyridoxine for nausea and vomiting of pregnancy: a randomized placebo controlled trial.Am J Obstet Gynecol · 2010 · 203(6):571.e1-7doi:10.1016/j.ajog.2010.07.030 · PMID 20843504

    131 vs 125 women, 14 days: PUQE fell 4.8 ± 2.7 vs 3.9 ± 2.6 (P = .006); 48.9% vs 32.8% asked to continue. The pivotal trial for the 2013 approval; see the reanalysis [26]. None of the Koren papers cited on this page is retracted or under an expression of concern in PubMed as of 19 October 2026; other papers of his have been retracted [94].

64. 64
    Pasternak B, Svanström H, Hviid A. Ondansetron in pregnancy and risk of adverse fetal outcomes.N Engl J Med · 2013 · 368(9):814-23doi:10.1056/NEJMoa1211035 · PMID 23445092

    Danish registry, 608,385 pregnancies: no excess spontaneous abortion, stillbirth, major defects (2.9% vs 2.9%), preterm birth or low birth weight; 1,233 exposed for the defect analysis.

65. 65
    Fejzo MS, Sazonova OV, Sathirapongsasuti JF, et al. Placenta and appetite genes GDF15 and IGFBP7 are associated with hyperemesis gravidarum.Nat Commun · 2018 · 9(1):1178doi:10.1038/s41467-018-03258-0 · PMID 29563502

    First GWAS, in 23andMe participants: chr19p13.11 (GDF15) and chr4q12 (IGFBP7) at P < 5 × 10⁻⁸, replicated. Several authors were 23andMe employees; the paper itself says causality needs further study.

66. 66
    Gurtan AM, Khalid S, Koch C, et al. Identification and characterization of human GDF15 knockouts.Nat Metab · 2024 · 6(10):1913-1921doi:10.1038/s42255-024-01135-3 · PMID 39327531

    8 homozygous and 227 heterozygous loss-of-function carriers, including C211G, from 75,018 sequenced Pakistanis; knockouts fertile and without overt phenotype. Authors largely from Novartis, AstraZeneca, Astellas and Regeneron.

67. 67
    Sillis L, Heinonen EW, Ceulemans M, Johnson D, Luo Y, Chambers CD. Metformin for the Prevention of Hyperemesis Gravidarum: An Observational Cohort Study.BJOG · 2025 · 132(12):1772-1778doi:10.1111/1471-0528.18238 · PMID 40443168

    MotherToBaby 2012 to 2023: 1 of 80 metformin-exposed (1.25%) vs 97 of 4,411 (2.20%), aRR 0.50 (0.07 to 3.39); NVP rates equal. Underpowered by design.

68. 68
    Pont S, Bond DM, Shand AW, et al. Long-term health, neurodevelopmental, and educational outcomes of children born to mothers with hyperemesis gravidarum: a population-based sibling-design record linkage cohort study.Am J Obstet Gynecol · 2025 · 233(4):325.e1-325.e20doi:10.1016/j.ajog.2025.02.044 · PMID 40064411

    700,082 NSW births, 10,282 exposed: raised risks of poor reading, spelling, numeracy and neurodevelopmental disability services in the full cohort, nullified for most outcomes among discordant siblings. The 2025 reversal on child outcomes.

69. 69
    Ostenfeld A, Carlsen SE, Jensen AK, et al. Mirtazapine or ondansetron for hyperemesis gravidarum: a randomized placebo-controlled trial.Am J Obstet Gynecol · 2026 · 234(6):1700-1729 (online 30 Dec 2025)doi:10.1016/j.ajog.2025.12.061 · PMID 41478546

    VOMIT trial, 7 Danish hospitals, 59 of a planned 180 women: mirtazapine vs placebo −1.86 PUQE-24 at day 2 (−3.61 to −0.12, P = .04); ondansetron vs placebo −0.51, not significant; more mild adverse events with mirtazapine.

70. 70
    Semmens JP. Female sexuality and life situations. An etiologic psycho-socio-sexual profile of weight gain and nausea and vomiting in pregnancy.Obstet Gynecol · 1971 · 38(4):555-63PMID 4938580

    Cited for its title as a document of the era in which the cause was sought in the woman's sexuality; no abstract or DOI available.

71. 71
    Kjeldgaard HK, Eberhard-Gran M, Benth JŠ, Nordeng H, Vikanes ÅV. History of depression and risk of hyperemesis gravidarum: a population-based cohort study.Arch Womens Ment Health · 2017 · 20(3):397-404doi:10.1007/s00737-016-0713-6 · PMID 28064341

    Norwegian Mother and Child Cohort, 731 HG pregnancies vs 81,055: lifetime depression aOR 1.49 (1.23 to 1.79); only 1.2% of women with prior depression developed HG and two thirds of women with HG had neither past nor current depression; the authors conclude depression does not seem to be a main driver.

72. 72
    Xu J, Qin Z, Wang X, Chen J, Zhuang L. Association Between Depression and the Risk of Hyperemesis Gravidarum: A Systematic Review and Meta-Analysis.Neuropsychiatr Dis Treat · 2026 · 22:567778doi:10.2147/NDT.S567778 · PMID 41908335

    13 case-control studies, 95,249 women: depression OR 5.08 (2.71 to 9.52; I² 93%); stronger when measured at diagnosis (6.04) than before it (2.11). The authors state that the association does not imply causation and may reflect the illness's psychological impact.

73. 73
    Nana M, Tydeman F, Bevan G, et al. Hyperemesis gravidarum is associated with increased rates of termination of pregnancy and suicidal ideation: results from a survey completed by >5000 participants.Am J Obstet Gynecol · 2021 · 224(6):629-631doi:10.1016/j.ajog.2021.03.006 · PMID 33713683

    Research letter; no abstract indexed. Cited for the finding its title states, from a British survey of more than 5,000 women.

74. 74
    de Vera SA, Brecht-Doscher A, Fejzo MS, Brecht ML, Kwon IM, MacGibbon KW. Risk Factors for Infusions, Emergency Room Visits and Hospitalizations for Hyperemesis Gravidarum: New Data and Literature Review.Int J Womens Health · 2024 · 16:1789-1802doi:10.2147/IJWH.S371458 · PMID 39493662

    1,220 survey respondents: visits more likely with extreme weight loss, inability to work, no children and being a person of color; those taking medication as prescribed had fewer visits. Self-selected online sample.

75. 75
    Kutcher JS, Engle A, Firth J, Lamm SH. Bendectin and birth defects. II: Ecological analyses.Birth Defects Res A Clin Mol Teratol · 2003 · 67(2):88-97doi:10.1002/bdra.10034 · PMID 12769504

    CDC birth-defect prevalence 1970 to 1992 did not change as Bendectin use ceased over 1980 to 1984; NVP hospitalization doubled. Ecological, population-level data.

76. 76
    Koren G, Gilboa D, Katz R. RETRACTED ARTICLE: Fetal Safety of Dydrogesterone Exposure in the First Trimester of Pregnancy.Clin Drug Investig · 2020 · 40(7):679doi:10.1007/s40261-019-00862-w · PMID 31583604

    Retracted publication, cited only as a record of retraction. It was first placed under an expression of concern stating that the article 'contains a number of errors and inconsistencies' (Clin Drug Investig 2020;40(7):681, PMID 31916053, DOI 10.1007/s40261-019-00884-4). PubMed also lists as retracted Koren G, Barer Y, Cem Kaplan Y. Fetal safety of medications used in treating infertility. Expert Rev Clin Pharmacol 2020;13(9):991-1000 (PMID 32815747). None of the four Koren papers cited on this page [3][4][24][25] is marked retracted or under an expression of concern in PubMed as of 19 October 2026.

77. 77
    Ratnadewi N, Husin F, Mose JC, Syamsunarno MRAA, Irianti S. Gut and Gastric Dysbiosis in Hyperemesis Gravidarum: A Scoping Review of Microbiome-Targeted Evidence and Translational Pathways.Int J Womens Health · 2026 · 18:632969doi:10.2147/IJWH.S632969 · PMID 42769745

    PRISMA-ScR scoping review of 36 sources published 2021 to May 2026: altered diversity and composition but inconsistent taxa and no reproducible signature; GDF15-centered endocrine susceptibility had the strongest mechanistic support; dysbiosis an associated modifier, consequence or endotype marker rather than an established independent cause; HG-specific probiotic evidence insufficient.

78. 78
    Stadtmauer DJ. Proximate and ultimate causes of pregnancy sickness.Evol Med Public Health · 2025 · 13(1):307-330doi:10.1093/emph/eoaf025 · PMID 41132319

    Evolutionary review: placental GDF15 production rose 100- to 1,000-fold in catarrhine primates and is low or absent in New World monkeys and rodents; re-evaluates adaptive hypotheses.

79. 79
    Yonezawa Y, Takahashi I, Ohseto H, et al. Genome-wide association study of nausea and vomiting during pregnancy in Japan: the TMM BirThree Cohort Study.BMC Pregnancy Childbirth · 2024 · 24(1):209doi:10.1186/s12884-024-06376-4 · PMID 38509478

    23,040 Japanese women: 19p13.11 (GDF15 region) variants significant in meta-analysis of discovery and replication; replicates the European finding in another ancestry.

80. 80
    Breit SN, Tsai VW. Metabolic Messenger: growth differentiation factor 15.Nat Metab · 2025 · 7(9):1732-1744doi:10.1038/s42255-025-01353-3 · PMID 40825850

    Review from the group that first described the hormone as MIC-1: GDF15 is an etiological factor in anorexia-cachexia and in pregnancy nausea and vomiting; GFRAL and its co-receptor RET are highly localized to specific hindbrain regions.

81. 81
    Najjar NR, De Jonghe BC, Hayes MR, Doyle RP. From stress cytokine to drug target: The translational landscape of growth differentiation factor 15.J Pharmacol Exp Ther · 2026 · 393(11):105036doi:10.1016/j.jpet.2026.105036 · PMID 42880196

    Medicinal-chemistry review of GFRAL agonists and antagonists, GDF15-binding approaches, patents and trials; hyperemesis listed among targets, with no pregnancy trial.

82. 82
    Alfaro A, Myxter K, Thomas J, Vaile-Larson L, Pullan JE. Hyperemesis Gravidarum and the Genetics of Growth Differentiation Factor 15 (GDF-15): A Case Report.Cureus · 2026 · 18(8):e115246doi:10.7759/cureus.115246 · PMID 42799287

    Single case with C211G genotyping (mother and child wild-type). Full text read; its discussion states, without naming the drug or the trial and with the references not recoverable, that a phase II clinical trial of an antibody-based approach to GDF15-GFRAL signaling in HG has reported encouraging preliminary findings. A registry entry for NGM Biopharmaceuticals' anti-GFRAL antibody NGM120 in HG (ISRCTN76681798) was identified by a checker's web search but could not be opened from the research environment and is not cited.

83. 83
    Thompson H, Neemuchwala ZH, MacGibbon KW, Housholder A, Fejzo M. Glucocorticoid treatment, lower growth and differentiation factor 15 concentrations, and improved hyperemesis gravidarum symptoms.Am J Obstet Gynecol · 2026 · 235(2):e44-e46doi:10.1016/j.ajog.2026.03.017 · PMID 41903874

    Research letter; no abstract indexed. Cited only for the association its title states, which should be treated as preliminary.

84. 84
    Groarke JD, Crawford J, Collins SM, et al. Ponsegromab for the Treatment of Cancer Cachexia.N Engl J Med · 2024 · 391(24):2291-2303doi:10.1056/NEJMoa2409515 · PMID 39282907

    Phase 2, 187 patients with GDF-15 ≥ 1,500 pg/mL: weight gain over placebo 1.22, 1.92 and 2.81 kg at 100, 200 and 400 mg. Pfizer-funded; shows GDF15 blockade works in cachexia, not in pregnancy.

85. 85
    Muscaritoli M, Molfino A, Orlando S, Tambaro F. Advancements of investigational agents for cancer cachexia: what clinical progress have we seen in the last 5 years?Expert Opin Investig Drugs · 2025 · 34(11):855-867doi:10.1080/13543784.2025.2588640 · PMID 41222020

    Review: GDF-15-targeting agents for cancer cachexia include ponsegromab, AV-380 and NGM120, with ponsegromab entering phase 3 trials. Cancer, not pregnancy.

86. 86
    Crespi BJ. Nausea, vomiting and conflict in pregnancy: The adaptive significance of Growth-Differentiation Factor 15.Evol Med Public Health · 2024 · 12(1):75-81doi:10.1093/emph/eoae008 · PMID 38711789

    Argues fetal GDF15 is typically adaptive for the fetus through placental invasion, fewer losses and higher birth weight, framing hyperemesis as maternal-fetal conflict. Theory with cited evidence, not a trial.

87. 87
    Boelig RC, Barton SJ, Saccone G, Kelly AJ, Edwards SJ, Berghella V. Interventions for treating hyperemesis gravidarum.Cochrane Database Syst Rev · 2016 · (5):CD010607doi:10.1002/14651858.CD010607.pub2 · PMID 27168518

    25 trials, 2,052 women, 18 comparisons, mostly low or very low quality; corticosteroids lower readmission (RR 0.69) but not stay; metoclopramide more drowsiness and dry mouth than ondansetron; B6 slightly longer stay.

88. 88
    American College of Obstetricians and Gynecologists. ACOG Practice Bulletin No. 189: Nausea And Vomiting Of Pregnancy.Obstet Gynecol · 2018 · 131(1):e15-e30doi:10.1097/AOG.0000000000002456 · PMID 29266076

    The US guideline, January 2018; the abstract states the purpose and the problem of undertreatment. No later ACOG bulletin on the subject was found in PubMed on 19 October 2026; reaffirmation status could not be verified.

89. 89
    Vutyavanich T, Wongtra-ngan S, Ruangsri R. Pyridoxine for nausea and vomiting of pregnancy: a randomized, double-blind, placebo-controlled trial.Am J Obstet Gynecol · 1995 · 173(3 Pt 1):881-4doi:10.1016/0002-9378(95)90359-3 · PMID 7573262

    342 Thai women, 30 mg/day: nausea reduced (P = 0.0008); vomiting episodes not significantly (P = 0.0552). Ordinary NVP, not hyperemesis.

90. 90
    Viljoen E, Visser J, Koen N, Musekiwa A. A systematic review and meta-analysis of the effect and safety of ginger in the treatment of pregnancy-associated nausea and vomiting.Nutr J · 2014 · 13:20doi:10.1186/1475-2891-13-20 · PMID 24642205

    12 trials, 1,278 women: nausea MD 1.20 (0.56 to 1.84), vomiting MD 0.72 (−0.03 to 1.46, not significant), no excess miscarriage; low-quality evidence, NVP not hyperemesis.

91. 91
    Frivaldszky L, Obeidat M, Hegyi P, et al. Comparative Effectiveness of Pharmacological and Non-Pharmacological Interventions for Nausea and Vomiting in Pregnancy: A Systematic Review and Network Meta-Analysis.Nutrients · 2026 · 18(8):1293doi:10.3390/nu18081293 · PMID 42075106

    24 placebo-controlled trials, 3,017 women, 16 intervention categories: quince, vitamin B6 with pomegranate and mint, P6 acupressure, dimenhydrinate, and acupuncture with doxylamine-pyridoxine ranked highest, with considerable uncertainty, low-to-moderate certainty of evidence, high heterogeneity and sparse harm reporting. Ordinary NVP, not hyperemesis.

92. 92
    He L, Fan Y, Hu Y, Tian C, Tian Y, Zhang J, Ren Y, Tan J. The potential hazards of high doses of vitamin B6 in treating nausea and vomiting in pregnancy: A systematic review.Int J Gynaecol Obstet · 2025 · 169(1):38-50 (online 21 Nov 2024)doi:10.1002/ijgo.16032 · PMID 39569767

    19 sources, PubMed search only: 164 of 1,226 individuals taking excessive vitamin B6 had neurological symptoms (burning, tingling, paresthesia, ataxia, perioral numbness). Its fetal concerns rest on 4 miscarriages and 1 intrauterine death among 245 women and are weak.

93. 93
    Oliveira LG, Capp SM, You WB, Riffenburgh RH, Carstairs SD. Ondansetron compared with doxylamine and pyridoxine for treatment of nausea in pregnancy: a randomized controlled trial.Obstet Gynecol · 2014 · 124(4):735-742doi:10.1097/AOG.0000000000000479 · PMID 25198265

    36 women: ondansetron 4 mg beat pyridoxine 25 mg plus doxylamine 12.5 mg on nausea (51 vs 20 mm, P = .019) and vomiting; small, with uneven follow-up (72% vs 94%).

94. 94
    Abas MN, Tan PC, Azmi N, Omar SZ. Ondansetron compared with metoclopramide for hyperemesis gravidarum: a randomized controlled trial.Obstet Gynecol · 2014 · 123(6):1272-1279doi:10.1097/AOG.0000000000000242 · PMID 24807340

    160 women, intravenous, 24 hours: equal well-being and vomiting; less drowsiness (12.5% vs 30%), dry mouth and persistent ketonuria with ondansetron; metoclopramide much cheaper.

95. 95
    Kashifard M, Basirat Z, Kashifard M, Golsorkhtabar-Amiri M, Moghaddamnia A. Ondansetrone or metoclopromide? Which is more effective in severe nausea and vomiting of pregnancy? A randomized trial double-blind study.Clin Exp Obstet Gynecol · 2013 · 40(1):127-30PMID 23724526

    83 women, oral, 14 days: less vomiting with ondansetron (P = 0.042), no difference in nausea. Small; no DOI registered.

96. 96
    Dormuth CR, Winquist B, Fisher A, et al. Comparison of Pregnancy Outcomes of Patients Treated With Ondansetron vs Alternative Antiemetic Medications in a Multinational, Population-Based Cohort.JAMA Netw Open · 2021 · 4(4):e215329doi:10.1001/jamanetworkopen.2021.5329 · PMID 33890993

    456,963 pregnancies, Canada, US and UK, active comparator: fetal death HR 0.91, spontaneous abortion 0.82, stillbirth 0.97, major malformations OR 1.06 (0.91 to 1.22).

97. 97
    Morgan S, Quintrell E, Russell DJ, Rhaman A, Gebremedhin A, Wyrwoll CS, Kelty E. A Scoping Review of 5-HT3 Antagonist Antiemetic Medication Safety in the Management of Nausea and Vomiting of Pregnancy.Pharmacoepidemiol Drug Saf · 2026 · 35(9):e70449doi:10.1002/pds.70449 · PMID 42592880

    39 studies, 34 on ondansetron: no elevated risk of congenital anomalies overall; cardiac and orofacial findings conflicting but absolute changes minimal; granisetron data limited.

98. 98
    Sillis L, Heinonen EW, Ceulemans M, Foulon V, Luo Y, Johnson D, Chambers CD. Medications for Nausea and Vomiting of Pregnancy and the Risks for Adverse Birth Outcomes.Pharmacoepidemiol Drug Saf · 2026 · 35(4):e70367doi:10.1002/pds.70367 · PMID 41980593

    MotherToBaby, 2,711 pregnancies 2010 to 2023: no excess major malformations with NVP drugs overall (aRR 1.22, 0.79 to 1.87, vs untreated NVP); ondansetron associated with small for gestational age (aRR 1.80, 1.23 to 2.60, vs untreated NVP), possibly reflecting greater severity. Prospective observational, imprecise.

99. 99
    Pasternak B, Svanström H, Mølgaard-Nielsen D, Melbye M, Hviid A. Metoclopramide in pregnancy and risk of major congenital malformations and fetal death.JAMA · 2013 · 310(15):1601-11doi:10.1001/jama.2013.278343 · PMID 24129464

    28,486 first-trimester exposures: malformations 25.3 vs 26.6 per 1,000 (POR 0.93), none of 20 categories raised, no excess stillbirth.

100. 100
     Yost NP, McIntire DD, Wians FH, Ramin SM, Balko JA, Leveno KJ. A randomized, placebo-controlled trial of corticosteroids for hyperemesis due to pregnancy.Obstet Gynecol · 2003 · 102(6):1250-4doi:10.1016/j.obstetgynecol.2003.08.013 · PMID 14662211

     126 randomized, 110 analyzed: methylprednisolone then prednisone taper vs placebo, all on promethazine and metoclopramide; rehospitalization 34% vs 35% (P = .89).

101. 101
     Guttuso T, Messing S, Tu X, et al. Effect of gabapentin on hyperemesis gravidarum: a double-blind, randomized controlled trial.Am J Obstet Gynecol MFM · 2021 · 3(1):100273doi:10.1016/j.ajogmf.2020.100273 · PMID 33451591

     31 enrolled, 21 with primary outcome: 52% greater PUQE reduction than ondansetron or metoclopramide (16 to 88%, P = .01); registry data in the report give 1.2% major malformations in 659 exposures. Small trial, as its authors say.

102. 102
     Wang X, Zhang L, Chen S, et al. Acupuncture Combined with Auriculotherapy as Adjunctive Therapy for Hyperemesis Gravidarum: A Randomized Controlled Trial Evaluating Hospital Stay, Relapse Rate, and Quality of Life.Complement Ther Med · 2026 · 103448 (online 16 Oct 2026)doi:10.1016/j.ctim.2026.103448 · PMID 42887805

     122 women: better quality-of-life and mood scores, median stay 4 vs 6 days, PUQE unchanged; open label, no sham, so non-specific effects cannot be excluded.

103. 103
     Sanger GJ, Andrews PLR. Histamine and the mechanisms of nausea and vomiting: Translating the uncertain pharmacology of cyclizine.Br J Clin Pharmacol · 2026 · online 14 Sep 2026doi:10.1002/bcp.70807 · PMID 42733333

     Pharmacology review: cyclizine's use for vomiting of known cause, which may include pregnancy, was historically accepted without large, controlled trials; how it inhibits vomiting in pregnancy is not understood.

104. 104
     Ngee Ling L, Abu Kassim SNA, Jalaluddin J, Mohd Zain A. Acute Bilateral Visual Loss as an Atypical Manifestation of Wernicke Encephalopathy Secondary to Hyperemesis Gravidarum: A Case Report and Narrative Literature Review.Cureus · 2026 · 18(8):e115540doi:10.7759/cureus.115540 · PMID 42819704

     Single case at 16 weeks: visual loss, nystagmus, ataxia, recovery within 24 hours of intravenous thiamine; later growth restriction and preterm birth. A case, not a series.

105. 105
     Grooten IJ, Koot MH, van der Post JA, et al. Early enteral tube feeding in optimizing treatment of hyperemesis gravidarum: the Maternal and Offspring outcomes after Treatment of HyperEmesis by Refeeding (MOTHER) randomized controlled trial.Am J Clin Nutr · 2017 · 106(3):812-820doi:10.3945/ajcn.117.158931 · PMID 28793989

     116 women, 19 hospitals: birth weight 3,160 vs 3,200 g (difference −40 g, −230 to 310); all secondary outcomes comparable; only 47% tolerated the tube per protocol.

106. 106
     Alshaikh ABA, Al-Kuraishy HM, Kafy S, Albar M, Sebghatallah A, Shokr MM, Batiha GE. Hyperemesis gravidarum revisited: from GDF15 biology to precision multimodal therapy.Naunyn Schmiedebergs Arch Pharmacol · 2026 · 399(9):12953-12966doi:10.1007/s00210-026-05216-w · PMID 41942591

     Narrative review: stepwise antiemetic therapy, corticosteroids for refractory cases, early correction of fluids, electrolytes and thiamine, and nutritional support preferably by enteral tube feeding rather than parenteral nutrition when oral intake is inadequate. Opinion-level synthesis that adds no new data.

107. 107
     Cécile B, Potter BJ, Lewin A, Healy-Profitós J, Brousseau É, Auger N. Risk of Cardiovascular Disease in Women With a History of Hyperemesis Gravidarum, With and Without Preeclampsia.J Am Heart Assoc · 2023 · 12(11):e029298doi:10.1161/JAHA.122.029298 · PMID 37259983

     Quebec, 1,413,166 women followed up to 32 years: cardiovascular hospitalization HR 1.46 (1.38 to 1.54) with HG alone and 3.54 with HG plus pre-eclampsia. Observational.

108. 108
     Getahun D, Fassett MJ, Jacobsen SJ, et al. Autism Spectrum Disorders in Children Exposed in Utero to Hyperemesis Gravidarum.Am J Perinatol · 2021 · 38(3):265-272 (online 3 Oct 2019)doi:10.1055/s-0039-1696670 · PMID 31581303

     Kaiser Permanente Southern California, 469,789 mother-child pairs 1991 to 2014: autism 2.87 vs 1.71 per 1,000 person-years, aHR 1.53 (1.37 to 1.70); medications did not contribute. Retrospective cohort without sibling control.

109. 109
     Fejzo M, Kam A, Laguna A, MacGibbon K, Mullin P. Analysis of neurodevelopmental delay in children exposed in utero to hyperemesis gravidarum reveals increased reporting of autism spectrum disorder.Reprod Toxicol · 2019 · 84:59-64 (online 27 Dec 2018)doi:10.1016/j.reprotox.2018.12.009 · PMID 30594672

     Exploratory descriptive study: 267 exposed children of 177 mothers vs 93 children of 60 unaffected mothers; autism reported in 22 (8%) exposed and none unexposed. Parent-reported, small control group.

110. 110
     Nijsten K, Jansen LAW, Limpens J, et al. Long-term health outcomes of children born to mothers with hyperemesis gravidarum: a systematic review and meta-analysis.Am J Obstet Gynecol · 2022 · 227(3):414-429.e17doi:10.1016/j.ajog.2022.03.052 · PMID 35367190

     19 studies, 1,814,785 offspring: anxiety OR 1.74, sleep problems 2.94 (2 studies each), testicular cancer OR 1.60 (1.07 to 2.39; 5 studies); all 6 studies of ADHD and autism reported increases, too heterogeneous to pool. The authors call the evidence few studies of low quality.

111. 111
     Auger N, Padda B, Bégin P, Brousseau É, Côté-Corriveau G. Hyperemesis gravidarum and the risk of offspring morbidity: a longitudinal cohort study.Eur J Pediatr · 2024 · 183(9):3843-3851doi:10.1007/s00431-024-05647-8 · PMID 38884821

     1,189,000 Quebec children, 6,904 exposed: any hospitalization HR 1.21; neurologic 1.50; developmental 1.51; digestive 1.40; allergic 1.39. Observational without sibling control.

112. 112
     Syn NL, Chan SY, Chia EWY, et al. Severity of nausea and vomiting in pregnancy and early childhood neurobehavioural outcomes: The Growing Up in Singapore Towards Healthy Outcomes study.Paediatr Perinat Epidemiol · 2021 · 35(1):98-108doi:10.1111/ppe.12703 · PMID 32578237

     1,172 mothers, 190 (16.2%) severe: more externalizing and internalizing behaviors in early childhood, Cohen's d 0.33 to 0.43. Prospective but observational.

113. 113
     Fejzo MS. Relationship between hyperemesis gravidarum and poor school performance unresolved.Am J Obstet Gynecol · 2025 · 234(3):e118 (online 18 Sep 2025)doi:10.1016/j.ajog.2025.09.028 · PMID 40975257

     Letter responding to the New South Wales sibling study [73]; no abstract indexed. Cited for its title and its existence only.

114. 114
     Mignot S, Vergnaud L, Fevre KL, Bornert C, Xavier J, Xavier-David C. Features of nausea and vomiting in pregnancy as predictive factors of post-partum depression: a systematic review.BMC Pregnancy Childbirth · 2025 · 25(1):971doi:10.1186/s12884-025-07976-4 · PMID 41029222

     PubMed indexes the authors with given and family names transposed (as Stéphanie M, Laura V, Linh FK, Célia B, Jean X, Charlotte XD); the first author's affiliation email is stephanie.mignot. 18 studies: adjusted ORs for postpartum depression from 1.54 (mild NVP) to 6.52 (severe NVP) and 2.26 for persistent symptoms; study-level extremes, not pooled.

115. 115
     Mills A, Glaister P, Heys S, Hibberd C, Coxon K. Experiences of psychological burden and care in women with hyperemesis gravidarum and nausea and vomiting in pregnancy: a mixed-methods systematic review.Midwifery · 2026 · 162:104956doi:10.1016/j.midw.2026.104956 · PMID 42660075

     64 studies 2010 to 2025: depression and anxiety generally higher in HG, rising with severity and sometimes persisting beyond pregnancy; women report dismissal and lack of validation. A corrigendum was published (Midwifery 2026:104995, PMID 42805823, DOI 10.1016/j.midw.2026.104995); its content is not indexed.

116. 116
     Young-Wolff KC, Chi FW, Campbell CI, et al. Frequency of Preconception and Prenatal Cannabis Use and Nausea and Vomiting in Pregnancy.Obstet Gynecol · 2025 · 145(5):519-522doi:10.1097/AOG.0000000000005884 · PMID 40080822

     356,343 pregnancies 2011 to 2022: daily preconception cannabis aOR 2.61 for severe NVP, daily prenatal 3.80. Cross-sectional association; direction cannot be fixed.

117. 117
     Scott CH, Prewitt KC, Marshall NE. When the remedy becomes the illness: differentiating hyperemesis gravidarum versus cannabinoid hyperemesis syndrome in pregnancy and nutritional consequences.Am J Clin Nutr · 2026 · 101505 (online 5 Oct 2026)doi:10.1016/j.ajcnut.2026.101505 · PMID 42833911

     Case report; no abstract indexed. Cited for the diagnostic problem its title names.

118. 118
     González-Valdivia C, Tong B, Hjalmarsson S, et al. The gut microbiome in early pregnancy is associated with the severity of nausea and vomiting: a nested case‒control study.Gut Microbes Rep · 2025 · 3(1):2603861doi:10.1080/29933935.2025.2603861 · PMID 41907525

     SweMaMi cohort, 337 severe cases matched to moderate and mild: lower richness and diversity with severity. Association only.

119. 119
     Uğur E, Uludağ B, Ozkaynak A. Vestibular symptom burden in hyperemesis gravidarum: A scale-based comparison with migraine and motion sickness.J Vestib Res · 2026 · online 7 Oct 2026doi:10.1177/09574271261496103 · PMID 42841802

     Online questionnaire of 1,010 women: severe symptoms (8.7%) associated with motion-sickness susceptibility and headache impact; authors call the associations limited in magnitude.