Intended for healthcare professionals

Clinical Review

Dengue fever

BMJ 2015; 351 doi: (Published 15 September 2015) Cite this as: BMJ 2015;351:h4661
  1. Senanayake A M Kularatne, senior professor of medicine
  1. 1Department of Medicine, Faculty of Medicine, University of Peradeniya, Peradeniya, Sri Lanka
  1. samkul{at}

This clinical review has been developed for The BMJ in collaboration with BMJ Best Practice, based on a regularly updated web/mobile topic that supports evidence based decision making at the point of care. To view the complete and current version, please refer to the dengue fever ( topic on the BMJ Best Practice website.

The bottom line

  • Dengue fever is a globally important arboviral infection transmitted by the Aedes genus of mosquito (primarily A aegypti, but also A albopictus), found in tropical and subtropical regions

  • The infection is endemic in more than 100 countries, particularly the South East Asia region, western Pacific region, and the Americas

  • The incubation period is 3-14 days (average 7 days)

  • Clinical features include fever, headache, myalgia/arthralgia, and skin flushing/rash, together with leucopenia, thrombocytopenia, and increased liver function

  • Severe thrombocytopenia, haemorrhage, and plasma leakage are the key diagnostic features of the more severe forms of infection

  • Confirmatory tests include detection of viral antigen or nucleic acid and serology

  • Fluid therapy and the identification of the critical phase are the most important aspects of management

Dengue fever is a globally important arboviral infection transmitted by mosquitoes of the Aedes genus (primarily Aedes aegypti, but also A albopictus), an insect found in tropical and subtropical regions.1 Dengue infection causes a range of severe and non-severe clinical manifestations.2 The incubation period is 3-14 days (average 7 days).

Sources and selection criteria

I searched Medline and PubMed from 1980 to date, limited to publications in English. My search strategy used a combination of key words, including “dengue”, “DHF”, “US”, “Africa”, “Europe”, “Asia”, “WHO”, and “Complications”. I supplemented these sources with selected systematic reviews. Additional information cited includes evidence based national guidelines, published consensus statements, and WHO publications.

Who gets dengue fever?

Around two fifths of the world’s population (those in tropical and subtropical countries), or up to 2.5 billion people, are at risk of dengue infection.1 An estimated 50 million infections occur annually worldwide, with 0.5 million of these cases being admitted to hospital for dengue haemorrhagic fever. Approximately 90% of these cases are in children aged less than 5 years.1 The epidemiology is, however, changing both regionally and globally. Classic dengue fever is more common in adults than in children.3

The infection is now endemic in more than 100 countries, particularly the South East Asia region, western Pacific region, and the Americas.1 4 5 6 Severe manifestations such as dengue haemorrhagic fever and dengue shock syndrome, as well as other unusual manifestations, are increasingly being reported in previously unaffected regions.5 6 Multiple epidemics of dengue fever occurred in the United States in the 18th to early 20th centuries. After an absence of 56 years, dengue fever has re-emerged in US states such as Texas and Hawaii, and 796 cases were reported in the US from 2001 to 2007.2 Dengue haemorrhagic fever has been reported in many tropical US territories, and the DENV-2 virus has been implicated in some of these cases.7

All cases of dengue fever in the United Kingdom have been acquired as a result of travel to endemic areas. The Health Protection Agency reported 406 cases of dengue fever in England, Wales, and Northern Ireland in 2010, compared with 166 in 2009.

What causes dengue fever?

Dengue fever is caused by four antigenically distinct dengue virus serotypes: DENV-1, DENV-2, DENV-3, and DENV-4. All four have the capacity to cause severe disease. They are RNA viruses that belong to the Flavivirus genus/Flaviviridae family, which also includes the yellow fever virus, West Nile virus, Japanese encephalitis virus, and the St Louis encephalitis virus.1

The primary vector for spread of infection is A aegypti, a highly domesticated, day biting mosquito, with A albopictus also responsible for transmission. Although the mosquitoes are of Asian origin, they now occur in Africa, Europe, and the US. International travel and transportation of goods has helped the spread of both vector and virus. A considerable genetic variation occurs within each viral serotype, thereby forming phylogenetically distinct genotypes. The virion consists of three structural proteins plus a lipoprotein envelope and seven non-structural proteins, of which non-structural protein 1 (NS1) has diagnostic and pathological importance. Infection with any one serotype confers lifelong immunity to that specific serotype; cross protection to other serotypes, however, lasts only a few months.1 2 8 Some studies have shown that infection with the DENV-1 or DENV-2 serotype may result in more severe infection.9 10

Pathogenesis is linked to the host immune response, which is triggered by infection with the virus.11 Primary infection is usually benign. Secondary infection with a different serotype or multiple infections with different serotypes may, however, cause severe infection that can be classified as either dengue haemorrhagic fever or dengue shock syndrome, depending on the clinical signs.

Antigen-presenting dendritic cells, the humoral immune response, and the cell mediated immune response are involved in the pathogenesis. Proliferation of memory T cells and the production of pro-inflammatory cytokines lead to vascular endothelial cell dysfunction, which results in plasma leakage. The concentration of cytokines such as interferon-γ, tumour necrosis factor-α, and interleukin 10 is higher, and the levels of nitric oxide and some complement factors are reduced. NS1 is a modulator of the complement pathway and plays a role in low levels of complement factors.12 13 After infection, specific cross reactive antibodies, as well as CD4 and CD8 T cells, remain in the body for years.2

Infants can develop severe dengue fever during a primary infection (which is usually benign) owing to transplacental transfer of antibodies from an immune mother. This subsequently amplifies the infant’s immune response to the primary infection.12

Can dengue fever be prevented?

Primary prevention

The World Health Organization recommends strategies for the prevention and control of dengue infection,1 2 and authorities in dengue endemic regions may also produce their own prevention programmes and initiatives. The key to all prevention programmes is disease surveillance to detect epidemics (box 1). Communities in dengue endemic regions should be educated to recognise symptoms and prevent transmission.

Box 1 Measures to prevent the spread of dengue fever1

  • Regular removal of all sources of stagnant water to prevent mosquito breeding grounds

  • Appropriate clothing to cover exposed skin, especially during the day, and the use of insecticides, mosquito repellents, mosquito coils, and mosquito nets

  • Mosquito nets and coils placed around sick patients to prevent transmission

Various worldwide initiatives are in the process of testing genetically modified mosquitoes to help stop the spread of dengue fever.14 A tetravalent vaccine is being developed and may be available in the future.1 6 15


Screening is not applicable as dengue fever is a communicable disease. However, populations may be screened for epidemiological purposes or to check for previous exposure to dengue virus.

Secondary prevention

Recurrence is possible, with different serotypes leading to a secondary infection. The usual primary prevention measures should therefore be followed after recovery from an initial infection.

How is dengue fever diagnosed?

Dengue fever should be considered in any patient presenting with fever, generalised skin flushing, leucopenia, and thrombocytopenia. A correct diagnosis early in the course of infection is important to prevent complications.


Dengue fever should be suspected in any patients residing in countries where the infection is endemic and in those who travelled in such areas within the past two weeks.

The onset of symptoms after the incubation period is usually abrupt. Fever is characteristic of infection and is often abrupt in onset with high spikes of 39.4-40.5°C. It may also be biphasic and have a remittent pattern or be low grade, and generally lasts for five to seven days. In young children fever may cause febrile seizures or delirium. Patients with rapid defervescence may be about to enter the critical phase of infection.2 6

Aches and pains, particularly backache, arthralgia, myalgia, and bone pain, are common. Headache is also typical of infection and is generally constant and towards the front of the head. It usually improves within a few days. Severe retro-orbital pain on eye movement or with a little pressure applied to the eyeball is also usual.

Gastrointestinal symptoms (for example, anorexia, nausea or vomiting, epigastric discomfort or pain), lethargy or restlessness, collapse, or dizziness may also be present. Patients often report a lack of appetite or changes to taste sensation. Gastrointestinal symptoms, weakness, and dizziness may be more noticeable in dengue haemorrhagic fever. Upper respiratory tract symptoms (for example, cough, sore throat) are usually absent, although they may atypically occur in mild infection.

Physical examination

Diffuse skin flushing of the face, neck, and chest develop early with infection. This evolves into a maculopapular or rubelliform rash of the whole body, usually on day 3 or 4 of the fever. Blanching may occur when the skin is pressed.16 The rash fades with time, and during the convalescent phase appears as pallid areas.

Haemorrhagic signs include petechiae, purpura, or a positive tourniquet test (blood pressure cuff inflated to a point midway between systolic and diastolic blood pressures for five minutes; the test is positive if ≥10 petechiae per square inch appear on the forearm). More major haemorrhage can manifest as epistaxis, gingival bleeding, haematemesis, melaena, vaginal bleeding (in women of childbearing age), or bleeding from a venepuncture site. These signs can occur with either dengue fever or dengue haemorrhagic fever.1 2

Hepatomegaly may be present. Plasma leakage is a sign of dengue haemorrhagic fever, and clinical evidence of this includes the presence of ascites, postural dizziness, or pleural effusion.1 2

Circulatory collapse (that is, cold clammy skin, rapid and weak pulse with narrowing of pulse pressure <20 mm Hg with decreased diastolic pressure, postural drop of blood pressure >20 mm Hg, capillary refill time greater than three seconds, reduced urine output) indicates the presence of shock and supports a diagnosis of dengue shock syndrome.1 2

Phases of infection

Dengue infection has three distinct phases2: febrile, critical, and convalescent. The febrile phase is characterised by a sudden high grade fever and dehydration that can last two to seven days.2 The critical phase (box 2) is characterised by plasma leakage, bleeding, shock, and organ impairment and lasts for about 24 to 48 hours. It usually starts around the time of defervescence (this does not always occur), typically days 3 to 7 of the infection.

Box 2 Warning signs of impending critical phase of infection6

  • Abdominal pain or tenderness

  • Persistent vomiting

  • Accumulation of clinical fluid (for example, ascites, pleural effusion)

  • Mucosal bleeding

  • Lethargy or restlessness

  • Liver enlargement >2 cm

  • Increase in haematocrit with rapid decrease in platelet count

Patients with dengue haemorrhagic fever or dengue shock syndrome go through all three stages. The critical phase is bypassed in patients with dengue fever.1 2 17

Laboratory investigations

Initial laboratory investigations

A full blood count should be ordered initially in all patients with symptoms. Typically, leucopenia and thrombocytopenia occur as early as the second day of fever.1 Leucopenia, in combination with a positive tourniquet test, in a dengue endemic area has a positive predictive value of 70-80%.18 19 Leucopenia (with neutropenia) persists throughout the febrile period. In classic dengue fever (box 3), thrombocytopenia is usually mild, although it may also be severe.1

Box 3 Laboratory criteria for diagnosis of dengue haemorrhagic fever or dengue shock syndrome12

  • Rapidly developing, severe thrombocytopenia

  • Decreased total white cell count and neutrophils and changing neutrophil to lymphocyte ratio

  • Increased haematocrit (20% increase from baseline is objective evidence of plasma leakage)

  • Hypoalbuminaemia (serum albumin <35 g/L suggests plasma leakage)

  • Increased liver function test results (aspartate aminotransferase:alanine aminotransferase >2)

The haematocrit may also rise about 10% in patients with dengue fever owing to dehydration.1 The results of liver function tests are usually increased, particularly for alanine and aspartate aminotransferases.1 Clotting studies are not required for diagnosis but may play a useful role in the management of the infection in patients with haemorrhagic signs.

Confirmatory laboratory investigations

Confirmatory tests should be carried out if possible.1 2 This is important because dengue fever can be confused with many non-dengue illnesses. Four types of diagnostic test are available for confirmation of dengue virus infection (table 1).

Table 1

 Confirmatory laboratory tests for dengue fever

View this table:

The choice of test depends on numerous factors, including local availability, cost, time of sample collection, available facilities, and technical expertise. Although direct methods such as viral nucleic acid or viral antigen detection are more specific, they are more costly and labour intensive. Indirect methods (that is, serology) are less specific but are more accessible, faster, and less costly.2 The identification of viral nucleic acid or viral antigen, plus the detection of an antibody response (serology), is preferable to either approach alone.2 Detection of viral nucleic acid or viral antigen is primarily done in the first five days of illness, and serological tests after the fifth day. Some tests differentiate between viral serotypes, although this is not useful clinically.

Virus isolation is possible during the initial viraemic phase. This test is accessible only in some locations and is generally not recommended as results are usually not available in a clinically meaningful time frame.


Imaging studies are required only if dengue haemorrhagic fever or dengue shock syndrome is suspected. A lateral decubitus chest radiograph of the right side of the chest can be ordered to detect clinically undetectable pleural effusion in the early phase of plasma leakage. Ultrasonography of the abdomen is useful to detect the presence of ascites and plasma leak or other disease related changes in abdominal organs, including the liver, gallbladder (oedema may precede plasma leakage), and kidneys.1 2

How is dengue fever managed?

Treatment approach

Treatment is supportive, as no specific antiviral therapy is available for dengue infection, and is based on guidance produced by WHO and other region specific authorities.1 2 6 17 21 22 The only recognised treatment in dengue fever is maintaining adequate hydration, and in dengue haemorrhagic fever and dengue shock syndrome treatment is fluid replacement therapy. In dengue endemic regions, the triage of patients with suspected dengue infection should be done in a specifically designated area of the hospital.

Early diagnosis and optimal clinical management reduce the associated morbidity and mortality. Delays in diagnosis, incorrect diagnosis, use of improper treatments (for example, non-steroidal anti-inflammatory drugs), and surgical interventions are all considered harmful. Educating the public about the signs and symptoms of dengue infection and when to seek medical advice is the key to optimal diagnosis and treatment.

Severity of infection

The most commonly used and practical treatment plan is produced by WHO and is based on the severity of infection.2 This classification separates patients into one of three groups (table 2), depending on the clinical presentation.

Table 2

 World Health Organization guide to severity of infection2

View this table:

Group A

Patients classified as being in group A have the following features and can be managed at home:

  • No warning signs (particularly when fever subsides)

  • Able to tolerate an adequate volume of oral fluids and pass urine at least once every six hours

  • Near normal blood counts and haematocrit.

Group B

Patients classified as being in group B have the following features and require hospital admission:

  • Developing warning signs

  • Co-existing risk factors for serious infection (for example, pregnancy, extremes of age, obesity, diabetes, renal impairment, haemolytic diseases)

  • Poor family or social support (for example, patients live alone or far from medical facilities and without reliable transport)

  • Increasing haematocrit or a rapidly decreasing platelet count.

Group C

Patients classified as being in group C have the following features and require emergency medical intervention:

  • Established warning signs

  • In the critical phase of infection, with severe plasma leakage (with or without shock), severe haemorrhage, or severe organ impairment (for example, hepatic or renal impairment, cardiomyopathy, encephalopathy, or encephalitis).

Management of group C patients

These patients require emergency medical intervention. At presentation, patients may be in compensated or decompensated shock. Access to intensive care facilities and blood transfusion should be available. Intravenous crystalloids and colloids administered rapidly are recommended, according to algorithms produced by WHO.1 2 An attempt should be made to work out how long patients have been in the critical phase and their previous fluid balance.

The total fluid quota for 48 hours should be calculated based on the formula1 17:

maintenance (M)+5% fluid deficit

where M=100 mL/kg for the first 10 kg of body weight, 50 mL/kg for the second 10 kg of body weight, and 20 mL/kg for every kilogram over 20 kg of body weight up to 50 kg; and 5% fluid deficit is calculated as 50 mL/kg of body weight up to 50 kg

  • For example, for an adult who weighs 50 kg, the total fluid quota for 48 hours would be 4600 mL.

The formula may be used for children and adults, although the rate of administering treatment differs between these patient groups, and local protocols should be followed. Ideal body weight should be used in the formula for children.

Other formulas for fluid replacement therapy have been reported, so local protocols should be consulted and followed. The infusion rate should be adjusted according to the usual monitoring variables, and therapy is usually required for only 24-48 hours, with a gradual reduction once the rate of plasma leakage decreases towards the end of the critical phase.2

Giving colloids (for example, dextran 70% or 6% starch) over crystalloids (for example, 0.9% normal saline, Ringer’s lactate) has no clinical advantage.23 24 25 (B evidence.) WHO guidelines clearly indicate when colloids should be used (for example, intractable shock, resistance to crystalloid resuscitation).1 2

Patients should be monitored closely throughout, including vital signs, peripheral perfusion, pulse pressure, capillary refill time, fluid balance, haematocrit, platelet count, urine output, temperature, blood glucose, liver function tests, renal profile, coagulation profile, and other organ function tests as indicated.

Usually the patient’s condition will become stable within a few hours of fluid therapy. If patients remain unstable, other contributory causes such as metabolic acidosis, electrolyte imbalances (for example, hypocalcaemia, hypoglycaemia), myocarditis, or hepatic necrosis should be investigated and managed appropriately. If patients do not improve and the haematocrit falls, internal bleeding should be suspected and a blood transfusion carried out immediately; caution is, however, advised owing to the risk of fluid overload. Consensus is now for early use of colloids and blood transfusion in refractory unstable patients.

Over-enthusiastic treatment and too rapid hydration can lead to fluid overload, causing pulmonary oedema, facial congestion, raised jugular venous pressure, pleural effusion, or ascites. These complications should be treated with restriction of intravenous fluid therapy and bolus doses of intravenous furosemide (frusemide) until patients are stable. Although less common, the disease may take a different course as a result of the complications listed in table 3.

Table 3

 Complications of dengue fever

View this table:

Management of group B patients

These patients require hospital admission. The severity of infection should be assessed. If patients are not in the early critical phase (that is, with plasma leakage) they should be encouraged to take fluids orally (for example, approximately 2500 mL/24 hours for adults, or age appropriate maintenance fluid requirement for children). If this is not possible, or patients enter the critical phase (indicated by rising haematocrit, hypoalbuminaemia, progressive leucopenia, thrombocytopenia, third space fluid loss, and narrowing of pulse pressure with postural drop), intravenous fluid replacement therapy with 0.9% saline (or Ringer’s lactate) should be started using the maintenance (M)+5% fluid deficit formula.1 17 Patients should be monitored closely throughout, including vital signs, peripheral perfusion, fluid balance, haematocrit, platelet count, urine output, temperature, blood glucose, liver function tests, renal profile, and coagulation profile.

Management of group A patients

These patients can be managed at home and should be encouraged to take oral fluids (for example, approximately 2500 mL/24 hours for adults, or age appropriate maintenance fluid requirement for children). Oral rehydration products, fruit juices, and clear soups are better than water. Red or brown fluids should be avoided, as these may lead to confusion about haematemesis.

Patients should be advised to rest. Tepid sponging may be used for fever and paracetamol in usual doses for pain or fever. Non-steroidal anti-inflammatory drugs should be avoided as they increase bleeding tendency.

Patients should be given an instruction leaflet outlining the warning signs and be advised to return to hospital immediately if any develop. Blood counts should be performed daily.


Pregnancy is a risk factor for higher maternal mortality and poor antenatal outcomes. The incidence of caesarean sections, pre-eclampsia, preterm deliveries, reduced birth weight, and vertical transmission of the infection is higher.26 Close observation and meticulous management is therefore important in this patient group. Fluid intake is the same as for non-pregnant women; prepregnancy body weight should be used in the formula.1 27

As pregnancy is associated with various physiological changes such as high pulse rate, low blood pressure, wider pulse pressure, decreased haemoglobin and haematocrit values, and decreased platelet count, baseline variables should be noted on the first day of infection and subsequent results interpreted with caution. It should be remembered that other pregnancy related conditions, such as pre-eclampsia and HELLP (haemolysis, elevated liver enzyme levels, and low platelet levels) syndrome, may also alter laboratory variables.1 27

Detection of plasma leakage (for example, ascites, pleural effusion) is difficult in pregnant women, and so early ultrasonography is recommended.1 27


As the tendency for children to develop dengue haemorrhagic fever or dengue shock syndrome is increased, laboratory variables such as haematocrit, platelet count, and urine output should be monitored regularly.

Assessment of the severity of symptoms in infants aged less than 1 year is difficult compared with that in older children and adults. Infants have less respiratory reserve and are more susceptible to electrolyte imbalances and hepatic impairment. The plasma leakage that occurs in children may be shorter and respond faster to fluid resuscitation.1

Convalescence and discharge

Convalescence can be recognised by the improvement in clinical variables as well as the patient’s appetite and wellbeing. Patients may develop a diuresis, and hypokalaemia may follow. If hypokalaemia occurs, intravenous fluids should be replaced by potassium rich fluid. During recovery patients may also develop a rash or generalised pruritus. Once wellbeing is achieved and patients remain afebrile for 48 hours with an increasing platelet count and stable haematocrit, they can be discharged.2

Adjunctive therapies

Prophylactic platelet transfusions are rarely required (even with very low platelet counts) and are not recommended except in the presence of active bleeding. The clinical value of fresh frozen plasma, corticosteroids, intravenous immunoglobulin, and antibiotics is controversial and more evidence is required before they can be recommended.1 28

Disease notification

In dengue endemic regions, suspected, probable, and confirmed cases of dengue fever should be reported to the relevant authorities as soon as possible so that measures can be instituted to prevent transmission.2

What are the complications of dengue fever?


The mortality rate for severe dengue fever is 0.8-2.5%.1 Children are at an increased risk of severe infection and death, although severe infection in adults is increasingly being reported.1 4 5 6 The risk of children aged 1-5 years dying from dengue fever is fourfold higher than in children aged 11-15 years.20 Even though dengue haemorrhagic fever and dengue shock syndrome are uncommon in adults, a higher morbidity and mortality rate has been reported, especially in older people, which is related to an increased risk of organ impairment.1 6 29 30 31

Long term sequelae

Once patients have recovered, no long term sequelae are associated with dengue fever; some patients may, however, experience postviral fatigue syndrome. Platelet count gradually increases during convalescence, although some patients may develop temporary thrombocytosis during recovery. Complete normalisation of liver function test results may take up to four weeks.


Recurrence is possible with a different dengue virus serotype, leading to secondary dengue infection. Third and fourth recurrences of clinical infection can also occur, but the clinical impact is not clear. Once immunity has developed to all four serotypes, lifelong immunity to each occurs.


Patients do not require monitoring and regular follow-up beyond the acute infectious period. It is advisable to review blood counts and liver function for up to four weeks after discharge.

Patient instructions

Patients should be reassured that there are no long term sequelae associated with dengue fever after recovery, and that they may resume normal activities once physically capable. Based on empirical evidence, patients should be advised to avoid alcohol and strenuous activities during convalescence, as liver function may take up to three weeks to return to normal.

Are there any emerging treatments?


Although some studies report that corticosteroids are effective in dengue fever, more evidence is required before they are recommended.28 32 (C evidence.) In one randomised controlled trial use of oral prednisolone during the early acute phase of dengue fever was not associated with prolongation of viraemia or other adverse effects. The study was not powered to assess efficacy, but it found no reduction in the development of shock or other recognised complications of dengue fever.33

Antiviral drugs

The development of a safe and effective antiviral drug that is active against the dengue virus is a priority. Although this has been attempted, no successful outcomes have been achieved.34 Balapiravir is a prodrug of a nucleoside analogue (R1479) and an inhibitor of hepatitis C virus replication in vivo. It was found to be ineffective as a candidate drug in a randomised, double blind placebo controlled trial of balapiravir in adults with dengue fever.35


A tetravalent vaccine is currently being developed and may be available in the future.1 6 Trial results have shown promise.36 37 38 39 40

Additional educational resources

Resources for healthcare professionals
Resource for patients
  • Centers for Disease Control and Prevention (CDC): dengue—prevention (—discusses ways to reduce the risk of dengue infection


Cite this as: BMJ 2015;351:h4661


  • Competing interests: I have read and understood the BMJ Policy on competing interests and declare the following: none.

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

  • This clinical review series has been developed for The BMJ in collaboration with BMJ Best Practice (, an independent product produced by BMJ Publishing Group Limited. BMJ Best Practice comprises web/mobile topics that support evidence-based decision making at the point of care. Peer review of the content in this clinical review was carried out exclusively according to BMJ Best Practice’s own, independent process ( This adaptation of a BMJ Best Practice topic for a clinical review in The BMJ uses only a portion of content from the latest available web version of BMJ Best Practice. BMJ Best Practice is updated on an ongoing basis, and the content of any BMJ Best Practice topic is expected to change periodically including subsequent to its publication as a clinical review in The BMJ. To view the complete and current versions of all BMJ Best Practice topics, please refer to the BMJ Best Practice website (

  • Content from BMJ Best Practice is intended to support, aid and supplement the expertise, discretion and judgement of licensed medical health professionals who remain solely responsible for decisions regarding diagnosis and treatment of their patients. Content from BMJ Best Practice is not intended to function as a substitute for a licensed medical health professional’s judgement. BMJ Best Practice reflects evidence available to its authors and licensors prior to publication. BMJ relies on its authors to confirm the accuracy of the information presented to reflect generally accepted practices. While BMJ seeks to ensure BMJ Best Practice is up to date and accurate, it does not warrant that is the case. Content from BMJ Best Practice is supplied on an “as is” basis and any statements made to the contrary are void. BMJ Best Practice does not endorse drugs, diagnose patients, or recommend therapy. The full disclaimer applicable to BMJ Best Practice can be found at


View Abstract