Current diagnostic thresholds, GDF15 axis, stepwise antiemetic treatment and doses in hyperemesis gravidarum, the most severe end of pregnancy nausea and vomiting (ACOG PB 189; RCOG GTG 69, 2024).
⚕️ Target audience: This clinical review is intended for physicians and healthcare professionals; It is a summary of current literature and guidelines for educational purposes. Clinical decisions should be made based on patient-specific evaluation and applicable institutional protocols.
Nausea and vomiting of pregnancy (NVP) is an extremely common condition affecting approximately 70% of pregnancies. Hyperemesis gravidarum (HG) is the most severe end of this spectrum, with reported prevalence ranging widely from 0.3% to 10.8%; This variability is largely due to the lack of a standard diagnostic definition for many years (Fejzo, Nat Rev Dis Primers 2019; PMID 31515515).
Symptoms typically begin between weeks 4–7 of pregnancy and resolve by week 20 in approximately 90% of cases (RCOG GTG 69, 2024). HG is no small clinical burden: It is the most common cause of hospitalization in early pregnancy and the second most common cause of pregnancy hospitalizations overall in the United States. The tendency to relapse is high; The risk of HG recurrence in the next pregnancy was reported to be 86% in a survey-based study (Sharma and Fejzo, AJOG 2025; PMID 40588059).
RCOG Green-top Guideline No. 2024 The second edition of 69 features the international Windsor consensus definition. Accordingly, for the diagnosis of HG, symptoms must begin in the first trimester (before the 16th week of pregnancy, without any other cause), nausea and vomiting must be present together and at least one of them is severe, normal eating and drinking is impaired, and daily life activities are severely restricted; Signs of dehydration are considered contributing criteria. In contrast, ACOG Practice Bulletin No. 189 (2018) states that there is no official consensus definition for HG and in practice uses the criteria of persistent vomiting, fasting ketone, loss of more than 5% of pre-pregnancy weight, and fluid-electrolyte disturbance.
The most common tool for objective monitoring of symptom severity is the PUQE-24 (Pregnancy-Unique Quantification of Emesis) score, which evaluates the past 24 hours. It consists of three items assessing the duration of nausea, the number of vomiting episodes, and the number of retching episodes, with a total score ranging from 3 to 15. The cut-off values given by RCOG in the guideline appendix are clear: mild ≤6, moderate 7–12, severe 13–15 (RCOG GTG 69, 2024).
Limit of PUQE: PUQE-24 is valuable in monitoring the severity and treatment response of mild-moderate NVP; However, it is insufficient to distinguish severe presentation and HG due to the ceiling effect. The RCOG bases further management decisions (e.g. corticosteroid initiation) on clinical and biochemical parameters and positions some steps in its algorithm when the PUQE score is below 13.
NVP and HG are diagnoses of exclusion. Particularly atypical features — first onset of symptoms after 10+6 weeks, abdominal pain, fever, headache, goiter — raise alternative diagnoses, and the RCOG requires that other causes be excluded in this case. Gastrointestinal causes include gastroenteritis, cholecystitis, pancreatitis, appendicitis, hepatitis, peptic ulcer and small bowel obstruction; Genitourinary causes include pyelonephritis, uremia, and kidney stones. In the metabolic-endocrine group, diabetic ketoacidosis, thyrotoxicosis (may be confused with gestational transient hyperthyroidism), Addison's and hypercalcemia; In the neurological group, migraine, increased intracranial pressure and vestibular disease are considered. Pregnancy-specific differentials are molar pregnancy/gestational trophoblastic disease and multiple pregnancy; both increase the risk of HG. Drugs and toxins include opioids, iron and, increasingly, cannabinoid hyperemesis syndrome (Kothari et al., AGA Clinical Practice Update, Gastroenterology 2024).
The understanding of the mechanism of HG was radically changed by a seminal study published in 2023. According to PubMed, Fejzo et al. showed that a TGF-β member hormone called growth/differentiation factor 15 (GDF15), which acts through the GFRAL-RET receptor expressed in the area postrema region of the brainstem and causes nausea, vomiting and food aversion, plays a central role in the risk of HG (Nature 2023; PMID 38092039; DOI ; PMC10808057 full text read, no withdrawal).
Mass spectrometry data from the study revealed that the majority of GDF15 in the maternal circulation was of feto-placental origin; GDF15 is one of the most abundant secreted peptides in the trophoblast. Maternal GDF15 level is significantly increased in HG and vomiting pregnant women, but it is not diagnostic on its own due to overlap. The most revealing aspect of the mechanism is a paradox: low GDF15 level before pregnancy increases the risk of HG. The explanation for this is the sensitization-desensitization balance—people with chronically high basal GDF15 are more tolerant of the pregnancy spike. Three observations that support this model are striking: the prevalence of NVP in β-thalassemia carriers with high lifetime GDF15 levels is significantly lower than in controls; The risk of HG is increased manyfold in carriers of the rare GDF15 C211G variant (this variant reduces non-pregnancy GDF15); and smoking, which increases GDF15, paradoxically reduces the risk of HG.
This mechanism has led to direct therapeutic targets: blocking GDF15 signaling in pregnancy (transplacental transit-limited anti-GDF15/anti-GFRAL antibodies) or providing desensitization by safely elevating GDF15 before pregnancy. The first observational signal of the second approach in humans comes from metformin (below). Classic factors other than GDF15—the thyrotropic effect of hCG (high in multiple and molar pregnancy), estrogen, progesterone-mediated gastrointestinal dysmotility, H. pylori association, and psychosocial factors—complete the context, but the current weight of evidence has shifted to the placental GDF15 axis (Fejzo, Nat Rev Dis Primers 2019; PMID 31515515; DOI ).
Supportive measures apply at all steps: small and frequent, unspiced, high-protein/low-fat meals, avoidance of triggers. It is emphasized that early treatment can reduce the progression of NVP to HG (AGA, Gastroenterology 2024). Ginger may be superior to placebo and shows similar efficacy to B6 and dimenhydrinate; The AGA practical recommendation is 250 mg capsules 4 times daily, but evidence is limited and inconsistent (Cochrane CD007575). The evidence for P6 acupressure is weak.
ACOG PB 189 (2018) — U.S. stepwise approach. First-line treatment is pyridoxine (B6) alone or a doxylamine-pyridoxine combination; both are safe, effective and not teratogenic. B6 monotherapy is typically given at 10–25 mg every 6–8 hours; the U.S. product labels for nausea and vomiting of pregnancy specify different doses for the doxylamine succinate/pyridoxine hydrochloride formulations. For delayed-release DICLEGIS 10 mg/10 mg, the starting dose is 2 tablets at bedtime and the maximum daily dose is 4 tablets in total: 1 in the morning, 1 in the mid-afternoon and 2 at bedtime. For extended-release BONJESTA 20 mg/20 mg, the starting dose is 1 tablet at bedtime and the maximum daily dose is 2 tablets in total: 1 in the morning and 1 at bedtime. These U.S. labels address nausea and vomiting of pregnancy; both state that the products have not been studied in hyperemesis gravidarum. If the response is inadequate, dimenhydrinate, promethazine or prochlorperazine is added. If oral treatment is not tolerated and the patient is dehydrated, IV diphenhydramine, metoclopramide or ondansetron is administered with IV fluids. In refractory cases, methylprednisolone is the last option; corticosteroids are used cautiously before 10 weeks because of the association with cleft lip and palate. (DailyMed, U.S. labels for DICLEGIS and BONJESTA; accessed 1 November 2026).
RCOG GTG 69 (2024) — UK stepwise approach. The dosing table appended to the guideline provides explicit figures. First-line options include cyclizine 50 mg PO/IM/IV every 8 hours; prochlorperazine 5–10 mg PO every 6–8 hours (12.5 mg IM/IV or 25 mg PR); promethazine 12.5–25 mg PO/IM/IV/PR every 4–8 hours; and the doxylamine-pyridoxine combination. Second-line options include metoclopramide, ondansetron, and domperidone. Third-line therapy consists of corticosteroids for refractory HG.
| Step | Medicine | Dose |
|---|---|---|
| First | Cyclizine | 50 mg PO/IM/IV every 8 hours |
| First | Prochlorperazine | 5–10 mg PO every 6–8 hours; 12.5 mg IM/IV; 25mg PR |
| First | Promethazine | 12.5–25 mg PO/IM/IV/PR every 4–8 hours |
| Second | Metoclopramide | 5–10 mg PO/IM/IV every 8 hours (for up to 5 days) |
| Second | Ondansetron | 4–8 mg PO every 6–8 hours; 8 mg IV over 15 minutes, every 12 hours |
| Second | Domperidone | 10 mg PO every 8 hours (caution for QT) |
| Third | Corticosteroid | Hydrocortisone 100 mg IV twice daily → prednisolone 40–50 mg/day PO, gradually decreasing |
Metoclopramide is considered second-line due to the risk of extrapyramidal side effects; Following European Medicines Agency (EMA) review, the RCOG dosing table limits use to a maximum of 5 days (RCOG GTG 69, 2024). The corticosteroid regimen begins with IV hydrocortisone 100 mg twice daily; When clinical improvement is achieved, oral prednisolone is converted to 40–50 mg/day and gradually reduced to the lowest maintenance dose that controls symptoms.
Teratogenicity of Ondansetron is one of the most discussed drug safety topics in pregnancy, and there is a difference of tone among authorities. According to PubMed, in Huybrechts et al.'s study covering 1.8 million pregnancies, no association was found between oral ondansetron and cardiac malformation in the first trimester (adjusted RR 0.99; 95% CI 0.93–1.06), no increase in overall malformation was observed, but a small increase in the risk of oral cleft was found (adjusted RR 1.24; 95% CI 1.03–1.48); absolute risk increases from 11.1/10,000 to 14.0/10,000 births (JAMA 2018; PMID 30561479; DOI). No cardiac, cleft, or global malformation associations were demonstrated in an analysis of IV ondansetron from the same group (JAMA 2020; PMID 31730152; DOI). (DOI) (DOI)
In the UK, the MHRA has incorporated this small risk into regulatory guidance: during the first 12 weeks, ondansetron increases the absolute risk of oral clefts from 11/10,000 to 14/10,000 births (an additional +3/10,000). The MHRA does not ban the drug; it states that when licensed alternatives are inadequate and there is a 'special clinical need,' ondansetron may be used following shared, informed decision-making with the mother. RCOG practice is parallel: ondansetron is not first-line, is deferred until after the first trimester if possible, and if used before 10–12 weeks, the patient should be counseled. The current Cochrane review adds a broader caveat—there is no high-quality evidence for any intervention in NVP/HG; evidence for vitamin B6, doxylamine-pyridoxine, and antiemetics is limited and inconsistent (Cochrane CD007575).
RCOG recommends 0.9% NaCl with potassium chloride added where necessary as the fluid of choice. A critical caution is with dextrose-containing fluids: since glucose loading in a thiamine-depleted patient may trigger Wernicke's encephalopathy, dextrose is not started without thiamine and high doses of parenteral thiamine (e.g. 100 mg) are added when IV dextrose is administered each day (RCOG GTG 69, 2024). Electrolytes (Na, K, corrected Ca) and renal function are monitored; Rapid sodium correction is avoided when correcting hypokalemia and hyponatremia.
Thiamine and Wernicke: All HG patients with prolonged or severe vomiting should be given thiamine—especially before dextrose or parenteral nutrition. RCOG dosage is oral thiamine 100 mg 3 times daily or IV thiamine; IV route is preferred in severe/prolonged vomiting and risk of refeeding. Wernicke's triad (confusion, ophthalmoplegia, ataxia) may begin slowly and episodically in HG; Late diagnosis may leave permanent neurological damage or Korsakoff. The threshold of clinical suspicion should be kept low and refeeding syndrome should be prevented with thiamine and electrolyte replacement.
Thromboprophylaxis with low molecular weight heparin (LMWH) is recommended for women admitted with HG due to the increased risk of venous thromboembolism due to dehydration and immobilization; If there is a specific contraindication such as active bleeding, graduated compression stockings are used. Since the risk only increases significantly when the patient is persistently vomiting, thromboprophylaxis is re-evaluated once dehydration is corrected and mobilization is achieved; The reported odds ratio for venous thromboembolism in HG is 2.5 and for deep vein thrombosis is 4.4 (95% CI 2.4–8.4) (RCOG GTG 69, 2024).
RCOG adopts a three-tier model of care: community-based care, ambulatory day care, and inpatient care. Persistent vomiting and inability to tolerate oral antiemetic/fluid, ketonuria and/or weight loss of more than 5%, electrolyte disturbance, signs of dehydration or accompanying comorbidity (e.g. diabetes) are indications for hospitalization. In patients who are hemodynamically stable and do not have severe metabolic disorders, ambulatory day care (IV fluid + antiemetics) may prevent hospitalization.
Enteral and parenteral nutrition is the last step; It is only considered when all medical treatments have failed. The enteral route (nasogastric, nasoduodenal or nasojejunal tube, percutaneous endoscopic gastrostomy/jejunostomy if necessary) is tried first — enteral nutrition is more effective and safer than total parenteral nutrition. Total parenteral nutrition is preferred only when the enteral route is not tolerated; catheter is a last resort due to the risk of serious complications such as sepsis, thrombosis and metabolic disorder and is initiated with a refeeding syndrome protocol (thiamine first) (RCOG GTG 69, 2024; Fejzo, Nat Rev Dis Primers 2019).
The first human signal for the mechanism-based prevention hypothesis comes from metformin. According to PubMed, in a survey-based study by Sharma and Fejzo, pre-pregnancy metformin use reduced the risk of HG by more than 70% (adjusted RR 0.29; 95% CI 0.12–0.71), based on the hypothesis of desensitization via elevated GDF15; the effect was even more pronounced in a second pregnancy (adjusted OR 0.18). In the same study, SSRIs (adjusted RR 2.41) and cannabis (adjusted OR 3.48) increased the risk of HG (AJOG 2025; PMID 40588059; DOI). These data are observational and survey-based; they do not prove causality, and randomized trials are awaited—therefore, this should currently be viewed as a hypothesis-generating finding rather than a routine prophylactic recommendation. Anti-GDF15/anti-GFRAL antibodies engineered for limited transplacental transfer are still in the clinical trial phase.
Last updated: . Guide numbers and publication years have been verified against current catalogues.