Intended for healthcare professionals

Practice Clinical Updates

Management of severe pregnancy sickness and hyperemesis gravidarum

BMJ 2018; 363 doi: (Published 30 November 2018) Cite this as: BMJ 2018;363:k5000
  1. Caitlin R Dean, chairperson, PhD candidate12,
  2. Manjeet Shemar, trustee, consultant obstetrician and gynaecologist13,
  3. Gillian A U Ostrowski, trustee, general practitioner,, associate director145,
  4. Rebecca C Painter, consultant obstetrician2
  1. 1Pregnancy Sickness Support, Dunmore Farm, Par, Cornwall, UK
  2. 2Amsterdam UMC, University of Amsterdam, Obstetrics and Gynaecology, Amsterdam, The Netherlands
  3. 3Birmingham Women’s & Children’s Hospital, Birmingham, UK
  4. 4Bridge Lane Group Practice, Battersea, London, UK
  5. 5NHS England, London, UK
  1. Correspondence to C Dean caitlinrdean{at}

What you need to know

  • Nausea and vomiting in pregnancy (NVP) and hyperemesis gravidarum (HG) can have a profound psychosocial effect on women and their families; some women become suicidal or can consider termination

  • Use an holistic approach to assessing women, including perinatal mental health support, and recognise that no one measure, including ketones, can reliably assess severity of HG

  • Several medications are available to manage HG with a pragmatic stepwise approach

  • Evidence is mounting from biological research that GDF15, the key driver of cancer cachexia, is linked to severe NVP

Nausea and vomiting in pregnancy (NVP) affects around 70% of pregnancies. It is mild for around 40% of women, moderate for 46%, and severe for 14%.1 Mild to moderate NVP usually resolves in the second trimester, with 90% of cases resolving by 20 weeks.2 It does not generally require treatment. By contrast, hyperemesis gravidarum (HG) is a complication of pregnancy rather than a normal part of it and occurs in around 1.5% of pregnancies.3 It is not unusual for HG to persist throughout pregnancy.45 The psychosocial burden of HG can be heavy for women and their families, but effective and holistic treatment may reduce this.67

What causes it?

Historically, the aetiology of HG was poorly understood and it was thought to be caused by endocrine, infectious, psychosocial, and hereditary factors.4 However, recent studies in twins showed that severity and duration of NVP are highly genetic.8 Family history of HG leads to a threefold increased chance of the condition occurring among offspring.9 A genome wide association study found that genetic variants of the growth and differentiation factor 15 protein (GDF15) and insulin like growth factor binding protein 7 (IGFBP7) genes were strongly associated with NVP and HG.10 In another study, women with severe NVP symptoms who were taking anti-emetic medications in the second trimester had higher serum concentrations of GDF15.11 GDF15 is a hormone known to affect appetite, and is highly expressed in placental trophoblasts and decidual stromal cells. High GDF15 is a key driver in cancer cachexia,12 which is characterised by symptoms similar to HG: nausea, weight loss, and muscle wasting. Further research replicating these findings and establishing causality is needed.

How do you diagnose and assess severity of HG?

HG is a syndrome. However, its diagnosis lacks international consensus. This can lead to uncertainty as to the boundaries between HG and NVP and can make diagnosing HG challenging.

The Pregnancy Unique Quantification of Emesis-24 (PUQE) score (fig 1) is a validated tool for assessing the severity of NVP for research and may help to identify women with HG,1314 although the tool has not been validated to assess severity of HG. As yet, no single tool is validated for holistically assessing women’s symptoms, however questioning can seek to understand how symptoms are affecting quality of life, the economic burden,1516 and psychosocial effects.67 The Quality of Life question on the PUQE score does not directly contribute to the score, but helps to provides a holistic assessment of the impact of symptoms on women, particularly for women experiencing profound nausea without vomiting or retching, and it has been validated to correspond inversely with the PUQE score.13

Fig 1
Fig 1

Norwegian PUQE score for assessing symptoms13 including quality of life question, adapted from original PUQE score by Koren et al.14 Reproduced with permission

Typically, clinicians should consider HG in women whose symptoms of nausea and vomiting lead to weight loss of ≥5% of pre-pregnancy weight, dehydration, or electrolyte imbalances.17 Dehydration can be assessed using patient reported fluid intake and output (through vomit, urine, or saliva), reduced or concentrated urine output, skin turgor, dry mucous membranes, reduced blood pressure, and tachycardia.18 Diagnosing HG requires exclusion of other causes of nausea and vomiting (box 1). Additional investigation is required when abdominal or epigastric pain or signs of infection are present, or when symptoms begin after the first trimester.1719

Box 1

Differential diagnoses of nausea and vomiting in pregnancy and suggested investigations

  • Peptic ulcer

  • Appendicitis

  • Gastrointestinal obstruction

  • Urinary tract infection

  • Molar pregnancy

  • Cholecystitis

  • Gastroenteritis

  • Pancreatitis

  • Neurological conditions

  • Drug induced nausea and vomiting

Examinations and investigations to consider

  • Pelvic ultrasound scan

  • Abdominal examination and/or imaging

  • Midstream urine

  • Helicobacter pylori (most commonly serology)

  • Urea and electrolytes

  • Full blood count

  • Thyroid function

  • Liver function

  • Amylase (to exclude pancreatitis)

  • Oesophageal gastroduodenoscopy


A systematic review found that ketonuria does not correlate with markers of severity of HG and is not associated with dehydration outside pregnancy,20 so we advise that it should not be used to identify or assess HG. However, ketonuria persists in many guidelines as a criterion for management decisions and is frequently cited as a marker for dehydration and/or a threshold for admission, discharge, and access to intravenous fluid in both hospital and community guidelines.2122232425 An apparent gap exists between evidence and practice and we recommend this is revisited in future guidelines.

What are the risks?

Mild to moderate NVP during the first trimester is not associated with physical risks for mother or fetus and is thought to be a sign of healthy trophoblast development and ongoing pregnancy, although the precise mechanism for this is not understood.2627 Moderate sickness for a prolonged period of weeks or months may, however, have a negative psychosocial effect and financial implications if a woman is unable to work.7

Women experiencing any degree of HG are at risk of dehydration and malnutrition, and complications include Wernicke’s encephalopathy, metabolic disturbances, and venous thromboembolism.172829 Women with comorbidities that require medication are at increased risk if they are unable to tolerate regular oral medication.30 Although hyperemesis is rarely linked to maternal death, there were six maternal deaths from complications of HG in the UK between 2006 and 20123031 and a recent case report from the US of maternal death caused by HG.32

HG can have profound psychosocial consequences.67 Negative psychological effects of NVP and HG include loss of identity, social isolation, feelings of guilt and suicidal ideation,7 and pressure to terminate the pregnancy.33 Evidence shows that consequences may persist beyond pregnancy, with reports of post-traumatic stress symptoms in up to 20% of women with HG.34 Families may experience financial hardship, strain on relationships, and difficulty caring for children, as well as a sense of grief over the loss of a normal pregnancy experience.73536

The greatest risk to the fetus is termination of pregnancy, estimated (in a UK survey) to occur in around 10% of pregnancies affected by HG, but may be as high as 15%, according to a US cohort study.3337 A meta-analysis of observational studies showed other risks during pregnancy include low birth weight (<2500 kg; odds ratio 1.42; 95% confidence interval 1.27 to 1.58), preterm labour (odds ratio 1.32; 95% confidence interval 1.04 to 1.68), small for gestational weight babies (odds ratio 1.28; 95% confidence interval 1.02-1.60).42938 Case reports have identified vitamin deficiencies.39 Inconsistent evidence suggests that there may be small, long term cardiometabolic and neurodevelopmental effects from fetal exposure to malnutrition in pregnancy or persistent symptoms beyond the first trimester.4041424344

What treatments are available?

There is a lack of robust high quality evidence for the efficacy of treatment and management options for NVP and particularly HG. As two systematic reviews recently showed, the heterogeneity of definition for HG and poor quality of the studies hamper the little research there is.4546

Pharmacological interventions

Non-specialists can initiate medication and refer when medication is failing, or when red flag symptoms develop (box 2). Discuss with patients the available drug treatments, including ondansetron. The aim of treatment is to control symptoms enough to promote quality of life, and to enable sufficient eating and drinking to avoid dehydration and malnutrition. The RCOG has produced guidelines that outline a stepwise approach to treatment (fig 2). Evidence exists for safety of the most frequently used treatments, including metoclopramide and antihistamines.4748495051 Most studies found no association between maternal ondansetron use in early pregnancy and congenital anomalies, but there is inconsistent evidence for a small increase in congenital heart disease associated with ondansetron.49 Nonetheless, the Royal College of Obstetricians and Gynaecologists’ (RCOG) guideline considers ondansetron safe and recommends its use. Furthermore, the potential for harm from the interventions discussed below must be weighed against the harm of uncontrolled symptoms, and women should receive the necessary information to support informed decision making.17

Box 2

Red flag signs

  • Renal function impairment

  • Severe electrolyte disturbance

  • Cognitive impairment or neurological symptoms consistent with Wernicke's encephalopathy or central pontine myelinolysis

  • Indication of suicidal ideation

Fig 2
Fig 2

Treatment options for severe NVP and HG174546

The pragmatic approach advocated in the RCOG guideline makes use of the different pharmacological properties of each drug. As yet, evidence for the effectiveness of combination therapies is lacking for HG, however, this approach is aligned with regimens for nausea and vomiting induced by chemotherapy52 and is commonly practised in the UK.17 Symptoms peak at around 9-11 weeks,19 so effective early interventions may appear to lose efficacy when in fact the condition is worsening. Steroids may be considered when other treatment combinations have been exhausted. The anti-emetic mechanism of steroids is not yet understood, but when used in combination may boost the effects of other anti-emetics.53 Evidence for the efficacy of steroids in HG is weak,454654 and there is some evidence of an increased relative risk of cleft palate in early first trimester exposure. The absolute risk remains small, however,55 and should be balanced with the risks associated with treatment resistant HG. Women commencing steroids early in pregnancy may require a long course of treatment and should therefore be advised of associated risks such as gestational diabetes, increased risk of infections, and adrenal suppression.56 For women with HG that is refractory to optimal pharmacological, supportive, social, and adjunctive therapies, termination of the pregnancy may be the only suitable option.3357

Consider H2 blockers or proton pump inhibitors to treat oesophagitis and reflux, which affect up to 80% of women with severe NVP.5859 Ondansetron can cause problematic constipation, so offer preventive laxatives and advice to manage this. Side effects from metoclopramide are varied, and it is not recommended for use for longer than five days.60 Advise patients who are prescribed metoclopramide about the possibility and signs of tardive dyskinesia and advise them to stop treatment should it occur.

All first and second line treatments are available in oral and alternative forms. European countries provide per rectum administration for anti-emetics, although this is atypical in the UK. Anecdotal evidence from a pregnancy sickness helpline in the UK suggests that per rectal medication would be a welcome option for women, although to date no research exists comparing oral and rectal medication efficiency for HG. In the US, continuous subcutaneous infusion is frequently used for patients managed at home, but evidence for benefit is lacking in both quantity and quality.61

Intravenous fluid therapy

For women who are dehydrated or those with electrolyte imbalance, intravenous rehydration with normal saline and additional potassium chloride is recommended and should be guided by daily electrolyte monitoring.17 Dextrose solutions offer little advantage over saline62 and may precipitate Wernicke’s encephalopathy in patients who are thiamin deficient.1761 Thiamin (vitamin B1) is recommended to reduce the risk of Wernicke’s encephalopathy.2863

Many areas in the UK have outpatient facilities for intravenous rehydration which can help families manage the social and financial impacts of HG. These facilities could help prevent lengthy hospital admissions, lost work hours for partners, and increased childcare costs. However, outpatient treatment may not be appropriate for all women, depending on the severity of HG and personal circumstances.646566 Intravenous therapy at home is also possible in the UK, where community care teams operate, although this is not widely available.67

Nutritional interventions

Despite maternal undernutrition being a key feature of HG,1368 few studies have investigated the utility of nutritional interventions in the management of HG. A single randomised controlled trial69 assessed the effectiveness of nasogastric tube feeding in early HG pregnancies but it was not found to improve outcomes and was poorly tolerated by women. In severe or complex cases enteral or parenteral nutrition may be considered although there is a lack of evidence for safety or efficacy.

Social and psychological support

Appreciating the burden of the condition physically, mentally, socially, and financially, and validating the biological nature of the condition may go a long way in supporting women and reducing the potentially far reaching psychological morbidity associated with it.7646770 The non-specialist’s role is to identify and support, or refer for support, women with HG. Medical treatments may be somewhat limited in effectiveness for HG, quality of life can be enhanced by supportive care and psychological support.6 Evidence suggests that psychological intervention, such as counselling, in conjunction with medical treatment, may improve outcomes.71

Women may find symptoms so profoundly distressing that suicidal ideation has been reported.7 Perinatal mental health teams and liaison psychiatry services may be able to help women manage the distressing physical and mental nature of the condition.70

Various charities can provide support for women and families affected by NVP and HG. Box 3 gives details of such organisations.

Box 3

Organisations that provide free support and information to women affected by NVP and HG

  • Pregnancy Sickness Support (PSS) UK.; tel: 024 7638 2020. Provides website information, helpline, and online support as well as national 1-2-1 peer support for women. PSS also provides education and training for professionals and can assist with service development and are actively involved in research.

  • Hyperemesis Education and Research (HER) Foundation, USA and International; Provides website information and support to women and their families. The HER Foundation has international contacts and can arrange support in almost all countries around the world where charities do not otherwise exist. Actively funds and conducts research.

  • Hyperemesis Ireland, Ireland. (sister organisation to PSS).

  • ZEHG, The Netherlands.

  • Hyperemeesi, Finland.

  • Hyperemesis Gravidarum Norge, Norway.


What self help techniques exist?

Peer support via patient associations may be helpful. Validating the need for time off work and/or household and childcare tasks is often necessary. Eating “little and often” and resting are likely to help women with mild-moderate NVP. Women can keep a diary of symptoms to identify times in a daily pattern when the woman is nausea free, so eating, drinking, and activities can be planned for those times. However, women with HG often present only after trying such measures, when they are unable to eat at all, are predominately resting, and have no periods free of nausea. Such suggestions can lead these women to feel their symptoms have been disbelieved and may have a negative effect on the patient-clinician relationship.72

We were told it was normal—a patient’s perspective on hyperemesis gravidarum

I was excited to be pregnant, and sickness felt like a rite of passage so initially I was genuinely pleased the first morning I was sick.

Except that the vomiting did not stop, and the nausea became crippling. It felt like every cell in my body had been poisoned. Within days I was dehydrated and desperate. I could barely walk unaided; the world was spinning, and the retching was near constant. We sought help but were told it was normal and that we just needed to get on with it. I did not have the strength, tearlessly sobbing over my sick bag, to advocate for myself and insist that, surely, this was not normal.

Throughout the pregnancy every clinician I saw suggested ginger, but I had tried ginger in every form. Their suggestions felt like an accusation of not trying hard enough to cure myself, or that they didn’t think I was that unwell.

Clinicians seemed to be preoccupied with checking if I had ketones in my urine. I never had them—even when I could only produce a tiny dark sample and was so dehydrated. Even when I had barely eaten more than half a slice of toast a day for weeks, was struggling to stand up, and had lost nearly 20% of my body weight, ketones seemed necessary to access intravenous fluids or further medication. The focus on ketones alone made me feel that my experience was dismissed.

I would have liked an evidence based conversation about my situation and the treatments. My situation was in no way unique and many women report similar experiences, particularly with suggestions of ginger causing distress and ketones acting as a barrier to accessing treatment. These two issues are so important to women that patient groups are campaigning to “ditch the ketones.” 83

Can NVP and HG be prevented or prepared for?

The greatest risk factor for HG is a history of the condition in a previous pregnancy.373 The risk identified in the literature varies from 15% to greater than 80%.74 Advise women who have experienced severe NVP or HG in a pregnancy that there is a high chance of HG recurrence. Knowing this provides women with an opportunity to plan their care and treatment in advance with their doctor, which may help to reduce the overall severity and burden of the condition.75Box 4 outlines elements of a pre-pregnancy care plan.78 Some evidence indicates that pre-emptive medication, started before the onset of symptoms, may reduce the duration and severity of symptoms in subsequent pregnancies.7580 Early prescription of anti-emetics found effective in the first pregnancy is encouraged by national guidelines.1775 Suggesting that women plan practically, socially, and financially for the potential return of a prolonged period of illness may help.78 A 2015 report on women’s experiences of termination for HG found the impact the condition had on women’s ability to care for existing children was a key factor in their decision to end the pregnancies.33

Box 4

Elements of a pre-pregnancy care plan for women with a history of severe NVP or HG69

Physical preparation

Achieving a healthy pre-pregnancy weight may have benefit, as low pre-pregnancy weight is associated with increased hospital admission,76 and high pre-pregnancy weight is associated with increased NVP symptoms.19Commencing self-help strategies before becoming pregnant—such as eating little and often—may make it easier to maintain them during pregnancy77

Prophylactic medication

Discussion of what was and was not effective during the previous pregnancy and any side effects experienced would help to plan individual treatment. Some evidence suggests that treatment before onset of symptoms may reduce the overall severity of the condition75

Criteria to assess deterioration

Symptom severity tolerance varies for women depending on available social support, their financial and employment situation, and dependants.78 Therefore, deciding in advance those criteria indicating deterioration may help women to manage the condition within the context of their life and family and prevent unnecessary admissions

Further treatment options if required

Discussion of the further treatment options in advance can ensure that women are able to make informed decisions about their health and gain a sense of control over the situation. A recent survey of women’s experiences of treatment for HG found that many women were not given the opportunity to discuss the risks/benefits of treatment or ask questions, which added to their distress64

Psychosocial management

Psychosocial factors, such as economic strain and inability to care for children, are a major factor in decisions to terminate HG pregnancies3337 Encouraging women to plan in advance with a realistic approach to managing childcare, household duties, and financial requirements may help to reduce these challenges79


Future research

High quality, clinically relevant research on NVP and HG is limited because of underfunding and lack of foundational definition and core outcomes.81 International projects are establishing a definition and core outcomes,82 and research priorities are being set via the James Lind Alliance. For updates, and to take part in research priority setting, visit

Education into practice

  • What hospital and home services are you aware of in your area to help women with hyperemesis?

  • When you last assessed a woman presenting with severe NVP or HG, how did you offer support psychologically? Are there ways you could alter this?

How patients were involved in the creation of this article

Three of the authors of this article have had HG in recent years. Additionally, the lead author (CD) has discussed the article and its key messages with patients who form a patient involvement panel for Pregnancy Sickness Support to ensure they felt the most important points were covered. Notably, they were keen to stress the biological nature of NVP/HG, the inappropriate assessment of ketones, and the heavy psychosocial burden of the condition, so these were included in the need to know box.

Questions for future research

Ongoing research activity

  • Research priority setting (currently under way)

  • An internationally agreed definition and core outcome set for future researchers to use

Questions for future research identified in this article

  • Is combining medications safe and more effective than solo treatment regimens?

  • What is the most effective administration route for anti-emetics, ie, oral, per rectal, or slow subcutaneous infusion?

  • Repeating and expanding on the genetic findings to establish a cause for NVP/HG is also required

Additional educational resources for healthcare professionals and patients

  • Pregnancy Sickness Support is UK based and provides information and resources for healthcare professionals. Most content is free to access; registration enables access to further in depth training:

  • The HER Foundation is US based:

  • Spewing Mummy is a patient focused blog about HG that also has posts for healthcare professionals and downloadable charts, care plans, and assessment tools, including a pre-pregnancy care plan:


  • Competing interests: We have read and understood BMJ policy on declaration of interests and declare the following interests: None. The first author, CRD, works almost exclusively specialising in this condition and is currently undertaking a systematic review and a scoping review on HG as part of a PhD. [repeated below]

  • Much of the cited literature is from the results of database searches conducted during systematic reviews and has been assessed for quality. Two of the authors (CRD and MS) were involved in the RCOG guideline development and are therefore familiar with the quality assessed literature from that as well as the recent Cochrane systematic review literature. RCP has conducted several systematic reviews and original research on HG. Additionally, other experts were consulted for clarification and accuracy checking on specific points such as the genetic research findings. All authors consult with HG patients as active clinicians. All authors were involved in the planning, development, and revision of the article.

  • Provenance and peer review: commissioned; externally peer reviewed.


View Abstract

Log in

Log in through your institution


* For online subscription