Dengue Fever Viruses

Dengue Fever Viruses

Secondary article Article Contents

Duane J Gubler, Centers for Disease Control and Prevention, Fort Collins, Colorado, USA

. Classification

Dengue viruses are the most important arboviruses causing disease in humans. Over half of the world’s population live in areas of risk, and an estimated 50–100 million cases occur each year.

. Structure . Viral Replication . Vectors . Transmission Cycles . Epidemiology . Clinical Features of Infection and Pathogenesis

Classification

. Prevention and Control . Future

Dengue viruses belong to the family Flaviviridae, genus Flavivirus. There are four serotypes: DEN-1, DEN-2, DEN-3 and DEN-4. They belong to a larger, heterogeneous group of viruses called arboviruses. This is an ecological classification, which implies that transmission between vertebrate hosts including humans is dependent upon haematophagous (blood-sucking) arthropod vectors. There are over 70 antigenically related viruses in the genus Flavivirus, including the type species, Yellow fever virus. The genus includes several antigenic complexes, including the dengue complex, the Japanese encephalitis complex and the tick-borne encephalitis complex. The Japanese encephalitis complex includes several wellknown disease pathogens of humans, including Japanese encephalitis, Murray Valley encephalitis, St Louis encephalitis, West Nile, Kunjin, Zika and other viruses, all of which are transmitted by mosquitoes. The tick-borne flaviviruses include Tick-borne encephalitis, Omsk haemorrhagic fever and Kyasanur Forest disease viruses. Some flaviviruses have no arthropod vector, probably having lost the need for this type of transmission during the process of developing evolutionary relationships with their vertebrate hosts. The four dengue viruses make up a unique complex within the genus Flavivirus. Although the four serotypes are antigenically distinct, there is evidence that serologic subcomplexes may exist within the group. For example, a close genetic relationship has been demonstrated between DEN-1 and DEN-3 by using sequence homology and complementary deoxyribonucleic acid (cDNA) hybridization probes. And surprisingly, DEN-2 shows a high sequence homology (71%) with Edge Hill virus, an ecologically distinct flavivirus from Australia.

capsid core of 30–35 nm in diameter, which consists of a capsid protein and single-stranded, positive-sense ribonucleic acid (RNA) genome. The virion envelope is fringed with fine surface projections, the envelope and membrane structural proteins. The dengue (Flavivirus) genome is approximately 11 000 bases in length, and consists of short untranslated regions at the 3’- and 5’- ends with an uninterrupted open reading frame in between (Figure 2). The complete genome sequences are known for several strains of all four dengue virus serotypes. There are three structural proteins (capsid, premembrane and envelope), encoded by sequences near the 5’-end of the genome, and seven nonstructural proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B and NS5) encoded by the remainder of the genome (Figure 2).

E

M

RNA

Nucleocapsid

Structure The dengue viruses have a structure similar to other flaviviruses; they are spherical, about 40–50 nm in diameter (Figure 1), with a lipid envelope, which is apparently derived from the host cell membrane from which the viruses bud. The envelope encloses an isometric nucleo-

Lipid bilayer

Figure 1 The Flavivirus virion structure. Structural proteins: E, envelope; M, membrane.

ENCYCLOPEDIA OF LIFE SCIENCES / & 2001 Nature Publishing Group / www.els.net

1


pr M

C prM

E

Structural

Nonstructural

5’

3’ NS2A NS2B NS1

NS4A NS4B NS3

NS5

Membrane- Protease/polymerase soluble HA

Polymerase

Figure 2 The Flavivirus genome, showing gene order. Structural proteins: C, capsid; E, envelope; prM, premembrane. NS, nonstructural proteins; HA, haemagglutinin.

Viral Replication It is thought that dengue and other flaviviruses infect cells by attaching to cellular receptors through the envelope protein, although specific receptor proteins have not been identified. The viruses are internalized by endocytosis, after which the nucleocapsid is released into the cytoplasm of the cell by membrane fusion, a process that is initiated by a low-pH dependent conformational change in the envelope protein. The dengue gene order (Figure 2) is encoded in open reading frames as follows 5’-C-prM-E-NS1-NS2A-NS2BNS3-NS4A-NS4B-NS5-3’. Translation of the viral messenger RNA (mRNA) is initiated at the 5’-end, and the resulting polyprotein goes through extensive cotranslational and post-translational proteolytic processing and cleavage to form at least 10 mature viral proteins. Both viral and host proteases appear to be involved in processing the dengue polyprotein. A C-terminal hydrophobic membrane anchor domain in the capsid, premembrane and envelope proteins functions both as internal signal sequences involved in the transfer of the polypeptide into the lumen of the rough endoplasmic reticulum, and as membrane anchor domains. Charged amino acid sequences that follow this anchor act to stop the peptide transfer through the endoplasmic reticulum membrane as well as a signalase cleavage recognition site of the host cell signalase in the lumen of the endoplasmic reticulum. The initial processing events that separate the structural protein precursors and the N-terminus of the NS1 are mediated by this signalase. Cell-associated virions are constructed from the structural proteins, and mature virions are released from infected cells through a process involving cleavage of the pre-M protein in the Golgi vesicles.

2

Dengue viruses have three natural hosts: Aedes mosquitoes, lower primates and humans. Dengue viruses are known to cause clinical illness and disease only in humans. The primary site of dengue virus replication after injection into humans by the feeding mosquito is believed to be phagocytic monocytes. Other tissues from which the viruses have been isolated include the liver, lungs, kidneys, lymph nodes, stomach, intestine and brain, but it is not known to what extent virus replicates in these tissues. There is some evidence that the viruses also replicate in endothelial cells and possibly in bone marrow cells. Although encephalopathy has been documented in dengue infection, only limited evidence is available to suggest that dengue viruses cross the blood–brain barrier and replicate in the central nervous system. Viraemia in humans may last 2–12 days (average, 4–5 days) with titres ranging from undetectable to over 108 mosquito infectious doses 50 (MID50) mL. Experimental evidence shows that several species of lower primates (chimpanzees, gibbons and macaques) become infected and develop viraemia titres high enough to infect mosquitoes, but do not develop illness. Viraemia levels in lower primates are more transient, often lasting only 1–2 days if detectable, with titres seldom reaching 106 MID50 mL. Low-passage or unpassaged dengue viruses can only be propagated with consistent results in laboratory-reared mosquitoes and in mosquito cell lines. Mosquito species most commonly used for in vivo propagation include Ae. aegypti, Ae. albopictus and Toxorhynchites spp., all of which can be reared with ease in the laboratory. In the mosquito vectors, dengue viruses infect and replicate to high titre in nearly all tissues, including midgut epithelial cells, ovaries, fat body, brain and central nerve cord ganglia, and salivary glands. No apparent tissue damage is observed in the mosquito, and dengue virus infection does not appear to have any adverse effects on this host. Only three mosquito cell lines show high susceptibility to dengue viruses: C6/36 from Ae. albopictus, AP-61 from Ae. pseudoscutellaris and TRA-284 from Tx. amboinensis. Dengue viruses can also be propagated in baby mice and in several vertebrate cell lines. These all have lower susceptibility to infection than mosquito cells, however, and dengue viruses must be adapted to each system by serial passage before consistent results can be obtained. Baby mice, which are used for the isolation and assay of many other arboviruses, generally show no signs of illness after intracerebral inoculation with most unpassaged strains of dengue viruses. Experimentally, however, some strains can be adapted to produce illness and death in baby mice. Mammalian cell lines commonly used include LLCMK2 and Vero (monkey kidney), BHK-21 (baby hamster kidney), FRhL (fetal rhesus lung) and PDK (primary dog kidney).



Vectors Only species of the genus Aedes appear to be natural mosquito hosts for dengue viruses. Species of the subgenus Stegomyia are the most important vectors in terms of human transmission, and include Ae.(S.) aegypti, the principal urban vector worldwide, Ae.(S.) albopictus (Asia–Americas), Ae.(S.) scutellaris spp. (Pacific), Ae.(S.) africanus and Ae.(S.) luteocephalus (Africa). Species of the subgenus Finlaya (Asia) and Diceromyia (Africa) appear to be important mosquito hosts involved in dengue forest maintenance cycles. Two other species, Ae.(Gymnometopa) mediovittatus (Caribbean) and Ae. (Protomacleaya) triseriatus (North America) have been shown to be excellent experimental hosts for dengue viruses.

Transmission Cycles Dengue viruses exist in nature in three basic maintenance cycles (Figure 3). The primitive forest cycle involves canopydwelling mosquitoes and lower primates. A rural cycle, primarily in Asia and the Pacific, involves peridomestic mosquitoes (Ae. albopictus and Ae. scutellaris spp.) and humans. The urban cycle, which is the most important epidemiologically and in terms of public health impact, involves the highly domesticated Ae. aegypti mosquito and humans. Multiple virus serotypes are maintained in an endemic cycle in most large urban centres of the tropics, with epidemics occurring at periodic intervals. Dengue viruses are only transmitted to humans and lower primates by the bite of an infective mosquito vector. When a competent mosquito vector takes a blood meal from a person during the viraemic phase (see above), virus is ingested with the blood and infects the cells of the mosquito mesenteron. After 8–12 days, depending upon

ambient temperatures, the mosquito vector and the virus strain, the virus will disseminate and infect other tissues, including the mosquito salivary glands. When the mosquito takes a subsequent blood meal, virus is injected into the person with the salivary fluids. Dengue virus infection has no apparent affect on the mosquito, which is infected for life. Ae. aegypti is a highly competent epidemic vector of dengue viruses. It lives in close association with humans because of its preference for laying eggs in artificial waterholding containers in the domestic environment resting inside houses and feeding on humans rather than other animals. It has a nearly undetectable bite and is very restless, in the sense that the slightest movement will make it interrupt feeding and fly away. It is not uncommon, therefore, for a single mosquito to bite several persons in the same room or general vicinity over a short period of time. In addition to transmitting the virus to humans or lower primates, the female mosquito may also transmit the virus vertically through the eggs to her offspring. Although the implications of vertical transmission are not fully understood, it is thought to be an important mechanism in the natural maintenance cycles of dengue viruses, especially in rural and forest settings.

Epidemiology Geographic and seasonal distribution Dengue viruses have a worldwide distribution in the tropics (Figure 4). The viruses are endemic in most urban centres of the tropics, with transmission occurring throughout the year. Epidemic transmission occurs periodically in most virus-endemic areas, usually at 3- to 5-year intervals. Because surveillance in most endemic

Forest/Enzootic

Rural/Epidemic

Urban/Endemic/Epidemic

Aedes mosquitoes

Aedes mosquitoes

Aedes aegypti

? Vertical Primates transmission Primates

Vertical Humans transmission Humans

Vertical Humans transmission Humans

Aedes mosquitoes

Aedes mosquitoes

Aedes aegypti

Ae. aegypti Ae. albopictus

Ae. aegypti

?

Africa = Aedes (Diceromyia) Aedes (Stegomyia) Asia = Aedes (Finlaya) Aedes (Stegomyia)

Ae. polynesiensis Ae. mediovittatus

?Americas = Aedes spp.? Figure 3 Transmission cycles of dengue viruses.


3


Figure 4 Global distribution of epidemic dengue and the principal mosquito vector, 1980–1999. Yellow indicates areas infested with Aedes aegypti; red indicates areas infested with Ae. aegypti and recent epidemic dengue.

countries is poor, cases are not usually reported during interepidemic years, thus misleading tourist and other agencies about the risk of infection. It is well documented, however, that dengue viruses are maintained during interepidemic periods in most tropical areas and, although risk of infection is lower than during epidemic periods, it is still substantial to unsuspecting visitors. Peak transmission of dengue viruses is usually associated with periods of higher rainfall in most dengueendemic countries. Factors influencing seasonal transmission patterns of dengue viruses are not well understood, but obviously include mosquito density, which may increase during the rainy season, especially in those areas where the water level in larval habitats is dependent on rainfall. In areas where water storage containers are not influenced by rainfall, however, other factors such as higher humidity and moderate ambient temperatures associated with the rainy season increase survival of infected mosquitoes, thus increasing the chances of secondary transmission to other persons.

serotypes) and the geographic spread and increased incidence of the severe and fatal form of disease, dengue haemorrhagic fever/dengue shock syndrome (DHF/DSS). Once observed only in Southeast Asia, DHF/DSS has spread in epidemic form to west Asia, the People’s Republic of China, the Pacific Islands and the Americas in the past two decades. The factors responsible for the emergence and spread of epidemic DHF/DSS are not fully understood. The changing disease pattern described above provides support for both principal hypotheses on the pathogenesis of DHF/ DSS, secondary infection and virus virulence (see below). Increased transmission of multiple dengue serotypes (hyperendemicity) increases incidence and thus the probability that severe disease will occur, regardless of whether the underlying cause is due to increased virulence, immune enhancement or, more likely, a combination of both (Figure 5).

Hyperendemicity

Changing epidemiology of dengue A combination of increased and unplanned urbanization, changing life styles and lack of effective mosquito control has made most tropical cities highly permissive for dengue transmission by Ae. aegypti. Increased air travel by humans provides the ideal mechanism for the transport of dengue viruses between population centres. As a result, in the past 20 years there has been a dramatic increase in the movement of dengue viruses within and between regions, resulting in increased epidemic activity, development of hyperendemicity (the cocirculation of multiple virus 4

Increased transmission and movement of viruses

Increased probability of secondary infection

Increased probability of virulent strain selection or introduction

Increased probability of immune enhancement

Increased probability of DHF Figure 5 Hyperendemicity increases the probability of dengue haemorrhagic fever (DHF).



Dengue in humans is primarily an urban disease. Most major epidemics of DHF/DSS occur in tropical urban centres, where large and crowded human populations live in intimate association with the principal mosquito vector, Ae. aegypti.

Clinical Features of Infection and Pathogenesis Dengue infection causes a spectrum of illness in humans, ranging from clinically inapparent to severe and fatal haemorrhagic disease, with the latter representing only the tip of the iceberg. The incubation period may be as short as 3 days and as long as 14 days, but most often is 4–7 days. The majority of patients present with mild, nonspecific febrile illness or with classic dengue fever. The latter is generally observed in older children and adults, and is characterized by the sudden onset of fever, frontal headache, retroocular pain and myalgias. Rash, joint pains, nausea, vomiting and lymphadenopathy are common. The acute illness, which lasts for 3–7 days, is usually benign and self limiting, but it can be very debilitating, and convalescence may be prolonged for several weeks. The haemorrhagic form of disease, DHF/DSS, is most commonly observed in children under the age of 15 years, but it also occurs in adults. It is characterized by acute onset of fever and a variety of nonspecific signs and symptoms that may last 2–7 days. During this acute stage of illness, DHF/DSS is difficult to distinguish from any number of other viral, bacterial and protozoal infections. In children, upper respiratory symptoms caused by concurrent infection with other viruses or bacteria are not uncommon. The differential diagnosis should include other haemorrhagic fevers, hepatitis, leptospirosis, typhoid, malaria, measles, and influenza, among others. The critical stage in DHF/DSS occurs when the fever subsides to normal or below. At that time, the patient’s condition may deteriorate rapidly, with signs of circulatory failure, shock and death if proper management is not implemented. Skin haemorrhages, such as petechiae, easy bruising and purpura/ecchymoses, are the most common haemorrhagic manifestations; gastrointestinal haemorrhage may also occur, usually after, but in some cases before, onset of shock. The World Health Organization (WHO) has defined strict criteria for the diagnosis of DHF/DSS, with four major clinical manifestations: high fever, haemorrhagic manifestations, hepatomegaly and circulatory failure. WHO has classified DHF/DSS into four grades according to severity of illness: grades I and II represent the milder form of DHF, and grades III and IV represent the more severe form, DSS. Thrombocytopenia and haemoconcentration are constant features. There is some disagreement with the WHO case definition, however, because some

patients may present with severe and uncontrollable upper gastrointestinal bleeding with shock and death in the absence of haemoconcentration or other evidence of the vascular leak syndrome. These patients, by the WHO criteria, cannot be categorized as having DHF/DSS. Additionally, hepatomegaly may not be a constant feature in all epidemics of DHF/DSS. Dengue virus infection is associated with a variety of neurological disorders, including headache, dizziness, hysteria and depression. In addition, some patients present with clinical symptoms of viral encephalitis but conclusive evidence that actual central nervous system infection occurs is limited. Treatment for DHF/DSS is symptomatic, and the prognosis of the disease depends upon early recognition, initiation of corrective fluid replacement therapy and management of shock. A definitive diagnosis can only be made in the laboratory by serological and/or virological methods.

Pathology and histopathology The pathology of dengue virus infection is not well understood because systematic postmortem studies have not been done on patients representing all types of clinical expression. The major pathophysiological abnormality in classic DHF/DSS is an increase in vascular permeability, which leads to leakage of plasma. Patients may have serous effusion in the pleural and abdominal cavities and a variable amount of haemorrhaging in most major organs. Studies have not revealed destructive inflammatory vascular lesions, but some swelling and occasional necrosis have been observed in endothelial cells, as well as some perivascular oedema. Limited studies on patients with a fatal outcome have demonstrated focal central necrosis of the hepatic cells, presence of Councilman bodies and hyaline necrosis of Kupffer cells in the liver. Changes in the kidney are suggestive of an immune-complex type of glomerulonephritis. There is depression of bone marrow elements, which improves when the patient becomes afebrile. Biopsy studies of the skin rash have demonstrated perivascular oedema with infiltration of lymphocytes and monocytes.

Pathogenesis There is still considerable controversy about the pathogenesis of DHF/DSS. Evidence suggests that at least two pathogenetic mechanisms are associated with severe dengue infection. Classical DHF/DSS is characterized by a vascular leak syndrome, which, if not corrected, may lead to hypovolaemic shock and death. The underlying pathogenetic mechanism for this syndrome is thought to be an immune enhancement phenomenon, whereby the infecting virus forms a complex with nonneutralizing


5


dengue antibody produced in response to a different serotype of dengue virus, thus enhancing infection of mononuclear phagocytes; these produce vasoactive mediators thought to be responsible for increased vascular permeability. Loss of plasma from the vascular compartment may be mild and transient or severe and prolonged, the latter resulting in severe shock and death. Although classic DHF/DSS is most commonly associated with secondary dengue infections, it has been documented in primary infections as well, which suggests that subneutralizing levels of homologous antibody or other substances may also cause immune enhancement. In vitro studies have shown that not all dengue viruses can be enhanced and that there are qualitative differences in the enhancing ability of dengue antibody. This raises the question of whether dengue virus strains vary in their pathogenicity and, if so, how this influences the immune enhancement hypothesis. Because an animal model is not available, experimental data that demonstrate variation in virulence among dengue viruses do not exist. However, an accumulating mass of both molecular and field data suggest that dengue viruses, like most other animal viruses, do vary in their pathogenicity and their epidemic potential. For example, when DEN-1 and DEN-2 viruses were introduced into the Pacific in the early 1970s after an absence of over 25 years, some islands experienced explosive epidemics with severe and fatal haemorrhagic disease, while others with similar ecology experienced only low-level, sporadic and even silent transmission with mild illness. Virus strain variation was the only logical explanation for these differences. Recent laboratory evidence suggests that a major Cuban epidemic of DHF/DSS in 1981, the first of its kind in the American region, was caused by a DEN-2 virus strain introduced from Southeast Asia that was genetically distinct from the original American DEN-2 virus. The Cuban epidemic is often cited to support both immune enhancement and virus virulence hypotheses, which are not mutually exclusive. The most constant feature associated with the emergence of DHF/DSS in an area is the development of hyperendemicity (the cocirculation of multiple virus serotypes in the same community). This increases the probability of secondary infection, which has been shown to be associated with DHF/DSS, but hyperendemicity is also associated with increased movement of viruses between population centres, which increases the probability of introduction or of genetic selection of virus strains that have greater epidemic potential or virulence (Figure 5). Patients infected with dengue viruses who do not have classic DHF/DSS may experience severe and uncontrolled bleeding, usually from the upper gastrointestinal tract. This severe haemorrhagic disease is often more difficult to manage than classic DHF/DSS. The underlying pathogenetic mechanism for this type of bleeding is clearly different from that of the vascular leak syndrome and involves 6

disseminated intravascular coagulation and thrombocytopenia. A third type of severe and fatal dengue infection, which may or may not involve overt haemorrhagic disease, is encephalopathy. Although many of these infections present clinically as viral encephalitis, evidence that dengue viruses infect the central nervous system has not yet been conclusively documented. Data suggest that neurological symptoms may be secondary to cerebral haemorrhage, oedema or other indirect effects of dengue virus infection, although several recent studies have reported isolation of dengue virus and/or detection of dengue-specific antibody in the cerebrospinal fluid of patients.

Prevention and Control The options available for prevention and control of dengue/DHF/DSS are limited. Although a vaccine for dengue is currently not available, considerable progress has been made in recent years in developing one. Effective vaccination to prevent DHF/DSS will require a tetravalent, live-attenuated vaccine. Promising candidate attenuated vaccine viruses have been developed and have been evaluated in phase I and II trials in Thailand. A commercialization contract has been signed, and the tetravalent vaccine formulation is currently undergoing repeat phase I and II trials. Promising progress has also been made in developing recombinant dengue vaccines using cDNA infectious clone technology. An infectious clone of the DEN-2, PDK-53 live-attenuated vaccine candidate virus has been constructed, and work is currently in progress to construct chimaeric viruses by inserting the capsid, premembrane (PrM) and envelope genes of DEN-1, DEN-3, and DEN-4 into the DEN-2 PDK-53 backbone. The rationale is that these recombinants, through genetic manipulation, may be made to grow better and to be more immunogenic and safe than the original attenuated candidate vaccine viruses. The development of other new technology, such as DNA vaccines, is in its infancy. Despite the promising progress, it is unlikely that an effective, safe and economical dengue vaccine will be commercially available in the near future. Currently, the only way to prevent dengue infection is to control the mosquito vector that transmits the virus. Unfortunately, our ability to control Ae. aegypti is limited. For over 25 years, the recommended method of control was the use of ultralow volume application of insecticides to kill adult mosquitoes. Field trials in Puerto Rico, Jamaica and Venezuela, however, showed that these space sprays were not effective in significantly reducing natural mosquito populations for any length of time. This supports epidemiological observations that ultralow applications have little or no impact on epidemic transmission of dengue viruses.



The only truly effective method of controlling Ae. aegypti is source reduction, that is eliminating or controlling the larval habitats where the mosquitoes lay their eggs. Most important larval habitats are found in the domestic environment, where most transmission occurs. To have sustainability of prevention and control programmes, some responsibility for mosquito control must be transferred from government to citizen homeowners. Mosquito control programmes, therefore, must be community-based and integrated. Persons living in Ae. aegypti-infested communities must be educated to accept responsibility for their own health destiny by helping government agencies control the vector mosquitoes, and thus preventing epidemic dengue/DHF/DSS. Countries with endemic/epidemic dengue should develop active, laboratory-based surveillance systems that can provide some degree of early warning for epidemic transmission. Finally, prevention of excess mortality associated with DHF/DSS can be achieved by educating physicians in endemic areas on clinical diagnosis and management of DHF/DSS. As demonstrated in countries such as Thailand, early recognition and proper management are the keys to keeping DHF/DSS fatality rates low.

Future Continued population growth and urbanization of the tropics, changing life styles, increased air travel and lack of effective mosquito control have been the most important factors responsible for the dramatic increased incidence

and geographic expansion of DHF/DSS in the 1980s and 1990s. Dengue/DHF/DSS has become a global public health problem in the tropics, and it is anticipated that this trend will continue unless something is done to reverse it. More effective integrated prevention and control strategies must be developed and implemented worldwide in the tropics. Ultimately, development of an economical tetravalent vaccine holds the greatest promise for prevention and control.

Further Reading Gubler DJ (1988) Dengue. In: Monath TP (ed.) The Arboviruses: Epidemiology and Ecology, vol. II, pp. 223–260. Boca Raton, FL: CRC Press. Gubler DJ (1989) Aedes aegypti and Aedes aegypti-borne disease control in the 1990s: top down or bottom up. American Journal of Tropical Medicine and Hygiene 40: 571–578. Gubler DJ and Kuno G (eds) (1997) Dengue and Dengue Hemorrhagic Fever. Wallingford, UK: CAB International. Halstead SB (1990) Dengue and dengue hemorrhagic fever. Current Opinion in Infectious Diseases 3: 434–438. Henchal EA and Putnak JR (1990) The dengue viruses. Clinical Microbiology Reviews 3(4): 376–396. Pan American Health Organization (PAHO) (1994) Dengue and Dengue Hemorrhagic Fever in the Americas: Guidelines for Prevention and Control. Scientific Publication No. 548, pp. 1–98. Washington, DC: Pan American Health Organization (PAHO). Thongcharoen P (ed.) (1993) Dengue/Dengue Hemorrhagic Fever. WHO Monograph, Reg. Pub. SEARO No. 22, pp. 1–163, New Delhi: World Health Organization (WHO). World Health Organization (1997) Dengue Haemorrhagic Fever: Diagnosis, Treatment and Control. Geneva: WHO.


7