author's personal copy the pathology of dengue hemorrhagic fever

This article was published in an Elsevier journal. The attached copyis furnished to the author for non-commercial research and

education use, including for instruction at the author’s institution,sharing with colleagues and providing to institution administration.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

The pathology of dengue hemorrhagic fever

Anthony S-Y. Leong, MBBS, MD, FRCPA, FRCPath, FCAP, FASCP, FHKAMed (Pathol),Hon FHKCPath, Hon FRCPT,a K. Thong Wong, MBBS, FRCPath,b

Trishe Y-M. Leong, MBBS (hons), FRCPA, FCAP,c

Puay Hoon Tan, MBBS, FAMS, MD, FRCPath, FRCPA,d

Pongsak Wannakrairot, MD, FRCPTe

aFrom the Hunter Area Pathology Service and University of Newcastle, Newcastle, Australia;bDepartment of Pathology, University of Malaya, Selangor, Malaysia;cVictoria Cytology Service, Melbourne, Australia;dDepartment of Pathology, Singapore General Hospital, Singapore; and theeDepartment of Pathology, Chulalongkorn University, Bangkok, Thailand.

An estimated 2.5 billion people are at risk of dengue infection, and of the 100 million cases of dengue feverper year, up to 500,000 develop dengue hemorrhagic fever (DHF) or dengue shock syndrome (DSS), thelife-threatening forms of the infection. The large majority of DHF/DSS occurs as the result of a secondaryinfection with a different serotype of the virus. While not completely understood, there is evidence that thetarget cells include dendritic reticulum cells, monocytes, lymphocytes, hepatocytes, and vascular endothelialcells. Viral replication appears to occur in dendritic cells, monocytes, and possibly circulating lymphoidcells, and damage to these and other target cells occurs through immune-mediated mechanisms related tocross-reacting antibodies and cytokines released by dendritic cells, monocytes, and vascular endothelium.There is evidence of a concomitant cellular activation as well as immune suppression during the infection.The activation of memory T cells results in cascades of inflammatory cytokines, including tumor necrosisfactor-�, interleukins (IL-2, IL-6, and IL-8), and other chemical mediators that increase vascular endothelialpermeability or trigger death of target cells through apoptosis. Pathological studies in humans are uncom-mon, and a suitable animal model of DHF/DSS does not exist. The current treatment of DHF/DSS issymptomatic, and prevention is through vector control. As such, there is a great impetus for the developmentof vaccines and novel therapeutic molecules to impede viral replication in infected cells or counteract theeffects of specific inflammatory mediators on target cells. The role of genetics in relation to resistance toDHF/DSS also requires clarification.© 2007 Elsevier Inc. All rights reserved.

KEYWORDSDengue hemorrhagicfever;Dengue virus;Inflammatorycytokines;Dendritic cells;Pathology;Hepatitis;Apoptosis;Immunology

The global problem

The first record of a disease clinically resembling denguefever (df) can be found in a Chinese medical encyclopedia

dated 992.1 With the expansion of the world shipping tradein the 18th and 19th centuries, both the viruses and theirprincipal mosquito vector, Aedes aegypti, were spread tonew geographic regions, particularly along the tropical trad-ing routes and shipping ports. Because spread was slow andvia sailing ships, long intervals of 10 to 40 years separatedepidemics of the disease. The expansion of dengue wasclosely related to development and economic growth intropical countries and the mode of the viral transmission.

Address reprint requests and correspondence: Anthony S-Y Leong,MD, Hunter Area Pathology Service, Locked Bag 1, HRMC, Newcastle,NSW 2310, Australia.

E-mail: [email protected].

0740-2570/$ -see front matter © 2007 Elsevier Inc. All rights reserved.doi:10.1053/j.semdp.2007.07.002

Seminars in Diagnostic Pathology (2007) 24, 227-236


Economic growth is associated with transmigration of largenumbers of the population into cities and the developmentof new cities; this urbanization resulted in a concomitantincrease in breeding sites for the mosquito. Acceleratedmodern travel, particularly by airplanes, allows the intro-duction of the virus from an endemic area to a dengue-receptive area where both the vector and susceptible popu-lation exist. The virus is usually carried in an infected butasymptomatic person during the incubation period of thedisease, a method of transmission that often accounts forexplosive outbreaks of df.

In the latter part of the 20th century, globalization andrapid urbanization of many developing tropical countriesproduced increased transmission and hyperendemicity ofthe disease. Today, dengue is the most frequent arbovirusinfection with more than 100 million cases annuallythroughout the world, including up to 0.5 million cases ofdengue hemorrhagic fever (DHF) and 24,000 deaths.2,3 Anestimated 2.5 billion people are at risk for the infection inthe subtropical and tropical regions of the world,4 and recentyears have witnessed unprecedented global dengue epidem-ics with large numbers of fatalities5; such statistics make dfone of the most important arbovirus diseases in humanstoday.

The magnitude of the global problem is compoundedby the fact that there is currently no specific treatmentor vaccines for the disease, and control is dependentlargely on public health measures directed against thevectors.

Presentation, pathogenesis, and pathology

The Dengue virus (DV) is a positive-sense, single-stranded RNA surrounded by an icosahedral nucleocap-sid. The virus is composed of three structural proteins[the capsid (C), premembrane (prM), and envelope (E)proteins] and seven nonstructural (NS1-7) proteins. Withinfection, the DV is endocytosed and acidic pH triggers aconformational change of the E protein. This allowsfusion of the virus envelope to the endosomal membraneand release of the viral genome into the cytoplasm, whereit is translated into viral proteins that replicate in the hostendoplasmic reticulum. The newly synthesized viral ge-nomes are packaged as viral core, envelope, and mem-brane proteins into immature virions before being se-creted from the cell.6 The virus belongs to the family ofFlaviviridae and the genus Flavivirus, which includesyellow fever, West Nile, Japanese encephalitis, and St.Louis encephalitis viruses. There are four distinct sero-types of DV (DEN1-4), and infection with any one of theserotypes confers lifelong immunity to that serotype.Whereas DVs exhibit 65% to 70% sequence homology,serotype cross-reactive immunity is only transient andwanes after 6 months so that the host remains susceptibleto the remaining three heterologous dengue serotypes.7

Presentation

DV infection results in protean manifestations rangingfrom a mild nonspecific or subclinical febrile illness to DHFor dengue shock syndrome (DSS). df uncomplicated has anincubation period of 5 to 9 days. The clinical features areage-dependent. Infants and young adults may only havesimple fever with a maculopapular rash. In older childrenand adults, df is characterized by the sudden onset of highfever of about 40°C, chills, severe headache that is mostlyfrontal or retro-ocular, skin rash, general malaise, and severemuscle ache in the lumbar region, legs, and joints. There is lossof appetite with nausea and vomiting, and photophobia withpuffiness of the eyelids and cervical lymphadenopathy maybe present. The fever lasts for 2 to 4 days, and after an afebrileperiod of about 1 day, a second febrile period may follow,producing the typical “saddle back” temperature curve. Atthis stage, an itchy maculopapular rash develops, oftenspreading from the extremities to the trunk to involve theentire body, except for the face. The palms and soles mayalso be red. Convalescence can take several weeks ormonths, but there is virtually no mortality associated.8 Thetriad of fever, rash, and arthralgia requires separation fromother nonarbovirus infections, such as measles, rubella,Epstein–Barr virus, cytomegalovirus, and protozoan infec-tions such as toxoplasmosis. Due to the wide clinical vari-ation, it is not possible to make a definite diagnosis onclinical findings alone. A positive tourniquet test (�10petechiae/inch2 read 1 minute after release of 5 minutesof pressure midway between systolic and diastolic) andleukopenia (WBC �5000 cells/mm3) in a febrile patient ishighly predictive but still nonspecific for df. Real-time poly-merase chain reaction (RT-PCR) to detect the virus remainsthe gold standard, although this test is not available in manydeveloping nations.

The World Health Organization (WHO) distinguishesDHF/DSS from df with hemorrhage, which is considered amild disease.9 DHF/DSS is associated with poorer outcomesand a mortality rate that approaches 5%.10 The classicalfeatures of DHF are high fever, increased vascular perme-ability with hemorrhagic manifestations, thrombocytopenia(platelet count �100,000/mm3), and hemoconcentration orsigns of plasma leakage. DSS is defined as DHF plus eitherhypotension for age or narrow pulse pressure in the pres-ence of clinical signs of shock. The increased capillary per-meability is generally not accompanied by morphologicaldamage to the capillary endothelium, and there are alterednumbers and functions of leukocytes, increased hematocrit,and thrombocytopenia. Raised levels of aspartate-amino-trans-ferase (AST) and alanine amino-transferase (ALT) have beenobserved in 98% and 37% of patients, respectively, indicatingthat liver dysfunction is a very common occurrence in DHF.11

As dengue spreads worldwide, there is evidence that thefour key criteria of severe disease employed in the WHOgrading system (shock, plasma leakage, marked thrombo-cytopenia, and internal hemorrhage) may not be universally

228 Seminars in Diagnostic Pathology, Vol 24, No 4, November 2007


applicable, and many severe cases (including those thatinvolve shock and fatality) may be missed, especially as thecriteria were based on initial observations in children withthe disease in Southeast Asia.12

Other less common manifestations include neurologicalsymptoms from cerebral edema and encephalopathy fromhepatic failure, complications of disseminated intravascularcoagulopathy, diastolic dysfunction, abdominal compart-ment syndrome, adult respiratory distress syndrome,13 acuterenal failure,14 postinfectious fatigue syndrome,15 intracra-nial hemorrhage,16 fulminant hepatitis,17 acute abdomen,18

hemophagocytic syndrome and dyserythropoiesis,19 andmyocarditis.20

Pathogenesis

The pathogenesis of DHF/DSS is poorly understood andhas been the subject of much research interest in recentyears. Although df is a self-limiting disease in the greatmajority of cases, the problem becomes a significant one asthe large numbers of infected persons result in about half amillion cases per year of potentially life-threatening DHF/DSS. Up to 90% of DHF/DSS cases occur in secondaryheterologous DV infection; the remaining are primary in-fections, usually in infants between 6 and 12 months of age.Clearly, secondary infection with a heterologous dengueserotype is a major risk factor. To date, several immuno-logical aspects of DV infection have been studied in detail.These relate to the target cells of the virus and their immu-nological effects and the effects of antibody-mediatedmechanisms in heterologous secondary infections.

Antibody-enhanced viral duplicationAfter introduction of the dengue virus through the bite of

an infected Aedes mosquito, local viral duplication isthought to take place in target epidermal dendritic cells thatare up to 10 times more permissive to dengue infection thanmonocytes or macrophages. A C-type lectin expressed bythe dendritic cells can bind to the dengue envelope (E)protein and probably serves as a coreceptor for viral en-try.21,22 Migration of interdigitating dendritic cells to re-gional lymph nodes also allows transfer of the virus to Tcells. IgG antibodies and enhancement via IgM and com-plement C3 receptors have been implicated in dendritic cellinfection.23,24 The infection of CD14-positive dendriticcells as well as bone marrow dendritic cells leads to theproduction of TNF-�, IFN-�, and IL-10 and inefficientmaturation of infected dendritic cells, which undergo apo-ptosis. The dendritic cells display impaired ability to upregu-late cell surface expression of costimulatory, maturation, andmajor histocompatibility complex molecules, resulting in re-duced T cell stimulatory capacity. There is also an impairedability to stimulate allogenic T cells, which is accompaniedby further enhanced IL-10 production.25,26 These findingssuggest a possible immune evasion strategy of the virus byimpairing antigen-presenting cell function through matura-

tion blockade and induction of apoptosis.27 Natural killercells, which comprise another arm of the innate immunesystem, may also play a role. These cells can clear virus-infected cells either by direct cytotoxicity or via antibody-dependent cell-mediated cytotoxicity. Dengue immune se-rum has been shown to mediate antibody-dependent cell-mediated cytotoxic lysis of DV-infected cell lines.28

Research into the effects of primary DV infection hasbeen hampered by the absence of a true animal model of thedisease. Nonhuman primates display viremia but not DHF.Several promising murine models of DV infections haverecently been developed, including interferon-�/� and �receptor-deficient mice that developed encephalitis andnonobese diabetic/severe combined immunodeficent micereconstituted with human hematopoietic stem cells withskin erythema.29,30

Not surprisingly, because of the severity and mortalityassociated with DHF immunological reactions, secondaryheterologous DV infections have been extensively studied.Antibody-dependent enhancement is one mechanism thathas been proposed to explain the severity of DHF/DSS. Thismechanism evokes the binding of preexisting dengue anti-bodies at nonneutralizing conditions to heterologous DV toenable viral entry into FcRII-bearing target cells.31 Thetarget cells are predominantly cells of the reticuloendothe-lial system of spleen, liver, and bone marrow, includingmonocytes, lymphocytes, Kupffer cells, and alveolarmacrophages.32 The enhanced viral entry into such cellsproduces an increased viral burden in the host. Anti-Eantibodies enhanced infection via FcRII, whereas antiprMantibodies enhanced infection of both FcRII- and non-FcRII-bearing cells.33 Some studies have demonstrated anassociation between viral burden and disease severity,34,35

with the ability of preinfection plasma to enhance infectionof monocytes corresponding to disease severity.36 However,a more recent study contradicts those findings by showingno association between the ability of preillness plasma toenhanced infection in vitro and subsequent dengue viremiaor disease severity in secondary dengue-2 or dengue-3 virusinfections.37

In the small percentage of DHF cases that manifestduring primary DV infection, usually in infants, it has beenpostulated that maternal transmission of nonneutralizingDV antibodies result in the same phenomenon as describedabove in adults with secondary infections.38

Although the viral envelope glycoprotein E is responsi-ble for viral attachment and entry, and is the antigenic targetfor neutralizing antibodies, there are also several viral non-structural proteins that are involved in viral replication andhave other effects in vivo. The dengue nonstructural proteinNS1 has both a secreted and cell-associated form. SecretedNS1 levels have been found to be associated with viremialevels in secondary dengue-2 infection.39 Anti-NS1 antibod-ies appear to have both a protective as well as pathogenicrole in DV infection. In animal models, anti-NS1 antibodiesprotect against lethal flavivirus challenge,40,41 but these

229Leong et al Dengue Hemorrhagic Fever


antibodies to NS1 have also been suggested to have a directrole in the pathogenesis of vascular leakage. Sera from acutedengue-infected individuals are able to bind to humanumbilical cord endothelial cells, and this process can beblocked by the addition of recombinant NS1.42 Further-more, the treatment of such umbilical cord cells with murineantidengue NS1 antibodies, but not Japanese encephalitisNS1 antibodies, induced the production of IL-6, IL-8, andmonocyte chemo-attractant protein type 1, an inflammatoryresponse that was abrogated in the presence of recombinantdengue NS1.43 Anti-NS1 antibodies have also been impli-cated to cause damage to endothelial cells by inducing nitricoxide-mediated apoptosis.42 Antibodies to NS1 persist longafter dengue infection has resolved but do not cause pathol-ogy. Furthermore, the persistence and kinetics of develop-ment of anti-NS1 antibodies during and after secondary DVinfection do not correlate with the timing of plasma leakage,raising questions as to the clinical relevance of this in vitroevidence for molecular mimicry.

On primary infection with DV, antibodies are generatedagainst NS1 and the viral envelope protein E. Serotype-specific and serotype cross-reactive neutralizing antibodiesare directed against the E protein. Enhancing antibodies canaffect the severity of the disease as long as 20 years after theprimary infection, especially during dengue-2 and dengue-3infection following a primary infection with dengue-1 vi-rus.44 However, all serotypes have been shown to be capa-ble of causing severe disease,45 and milder disease is asso-ciated with lower viral loads and high levels of preexistingheterotype neutralizing antibodies during the secondary in-fection. This is not necessarily always true; in the case ofsecondary dengue-3 virus infection, higher levels of thesecross-reactive memory humoral immune responses appearedto be beneficial as demonstrated by reduced viremia levels anddecreased disease severity, but the same did not hold true forsecondary dengue-1 or dengue-2 virus infection.46

Immunological mechanismsThe observation that onset of plasma leakage occurs up

to several days following significant reduction or clearanceof viremia suggests that, instead of the proposed antibody-dependent enhancement due to viral burden, an immune-mediated mechanism may be responsible for the extremecapillary permeability that is characteristic of DHF/DSS.Furthermore, despite high viral loads in secondary infec-tions, progression to more severe forms of the infectionoccurs only in �5% of cases. Serotype-specific and cross-reactive T cells have been detected in the peripheral bloodof individuals with acute DV infections. It is suggested thatthese serotype cross-reacting T cells show low affinity forthe infecting virus and a higher affinity for another virusserotype (presumably from a previous infection), a phenom-enon called the “original antigenic sin.”47 However, definiteidentification of the primary infecting DV serotype is diffi-cult because neutralizing antibodies can be highly serotypecross-reactive after secondary infection, a caveat that questionsthe validity of this hypothesis.27 Importantly, it has been dem-

onstrated that cross-reactive dengue-specific T cells inducehigh levels of cytokine production that may lead to increasedvascular permeability.48

CD4 and CD8 T cell responses after primary DV infec-tion have shown interesting differences. CD4� T cellsproduced greater amounts of IFN-� to homologous DVantigens, but the ratio of TNF-� to IFN-� was higher afterstimulation with heterologous serotype antigens or CD4� Tcell epitopes.49 CD8� T cells have shown partial agonistresponses in vitro in which a heterologous serotype variantcould sensitize target cells for lysis but not cytokine pro-duction or proliferation.

Several studies have suggested a possible role for “im-munodominant” variants of dengue epitopes in which thesequence of DV serotypes may influence the risk of DHF/DSS. Human leukocyte antigen (HLA) A2-restricted CD8�T cell responses in primary dengue-immune individualshave shown both quantitative as well as qualitative differ-ences in their cytokine responses to variant dengue epitopes,suggesting that previous infection as well as the sequence ofheterologous DV infections may affect subsequent clinicaloutcome.50 For some epitopes, a single dengue variant wasable to elicit the highest response in all donors, regardless ofinfecting serotype, suggesting that certain epitopes may beimmunodominant. A recently identified dengue HLA-A11-restricted NS3 T cell epitope has also been found to be pre-sented by HLA A24, an unexpected finding because thesealleles belong to different HLA superfamilies.48 Previous in-fection history to unrelated viruses can also lead to immuno-pathology, as shown in human and animal models.51

The studies performed so far appear to show a concom-itant cellular activation as well as immune suppressionduring acute DV infection. T cell proliferation to mitogensis impaired during the acute infection, and this appears to becaused by a defect in the antigen-presenting cell.52 There isan impairment of the normal plasmacytoid dendritic cellresponse in children who subsequently develop DHF. Thisblunted response is thought to result in inadequate control ofviremia with the subsequent enhanced activation of cross-reactive memory T lymphocytes resulting in DHF.53 Over-all numbers of CD4 and CD8 T cells, natural killer cells,and �� T cells have been found to be decreased in DHFcompared with df, but despite their reduced numbers, CD8and natural killer T cells expressed higher levels of theactivation marker CD69 in patients with DHF comparedwith uncomplicated df. It has been shown that low affinitymemory T cells occur at higher levels than cytotoxic T cells,the latter being lost through apoptosis,47 and DV epitope-specific T cell responses are associated with disease sever-ity.47 Dengue-specific T cells are increased in DHF, butperipheral blood mononuclear cells are unable to produceIFN-� in response to dengue epitope stimulation,48 reflect-ing the reduced ability of monocytes and dendritic cells topresent antigen adequately. Alternatively, antigen-inducedcell death may ensue after the in vitro stimulation of re-



cently activated CD8 T cells,52 a finding that has beendirectly demonstrated.48,54

The short-lived nature of plasma leakage syndrome seenin DHF is another point to support a functional rather thandestructive effect of DV infection on endothelial cells. Invitro studies of human endothelial cell lines infected withDVs can produce pro-inflammatory mediators such as IL-6and IL-8 RANTES (regulated on activation; normal T cellexpressed and secreted), alter intercellular cell adhesionmolecule type 1 surface expression and actin cytoskeletonstructure, and increase permeability to small molecules.55,56

These effects could be partially reversed by neutralizingantibodies directed against IL-8.55 Also, the infection ofhuman umbilical cord endothelial cells with DV in thepresence of antidengue immune serum induced the forma-tion of activated complement via both classical and alter-native pathways with complement activation appearing tobe mediated by dengue NS1 protein.57 Serum from patientswith acute dengue infection induced the activation andapoptosis of cultured endothelial cells that could be partiallyreversed by anti-TNF-� monoclonal antibodies, further sup-porting the role of inflammatory mediators for plasma leak-

age.58 A recent murine model for DV-induced lethal vas-cular permeability has also demonstrated the importance ofTNF-� as a key mediator of DV induced in mice.59

The presence of IgM, �-1-globulin, and fibrinogen hasbeen demonstrated in the cutaneous vessels of patients withDHF with dengue antigen in perivascular mononuclearcells, indicating that the skin rashes associated with DHFare immune-mediated.60

Cytokine cascades, complement, and other mediatorsAlthough incomplete, the current evidence suggests that,

after massive activation of memory T cells, a cytokinecascade that targets vascular endothelial cells is primarilyresponsible for the critical leakage of fluid and protein inDHF/DSS. Such cytokines, including IFN-�, TNF-�, IL-2,IL-6, IL1-�, and IL-8, are released in high concentrationsmostly by T cells, monocytes/macrophages, and endothelialcells and are demonstrable in the serum of patients withDHF/DSS. More recent studies have confirmed the presenceof elevated levels of IFN-�, TNF-�, and IL-10 in patientsfrom Vietnam, India, and Cuba.61-63 Such cytokines havethe potential to induce the release and production of other

A B

C D

Figure 1 Organs from a 2-year-old Thai boy who died from dengue hemorrhagic fever. (A) The eviscerated intestines were markedlyedematous and focally hemorrhagic, and there was a hemoperitoneum of 300 mL. (B) The kidneys were edematous with focal hemorrhageand hemorrhage into the calyces and renal pelvis. (C) The liver was swollen with focal hemorrhage. (D) Both lungs were heavy and beefyin consistency and hemorrhagic.



cytokines so that a complex interactive network results infurther increases in the levels of cytokines and other chem-ical mediators in a cascade, often with synergistic effects onvascular permeability. IFN-� is also able to upgrade theexpression of Fc� receptors on monocytes and macro-phages, further facilitating viral replication. Although thiscascade of cytokine activation and production stimulated bymemory T cells is an attractive proposition in adult DHF/DSS, it fails to explain the occurrence of this severe form ofdengue infection in the primary infection of infants born ofdengue-immune mothers.

Complement has been suggested to have a role in theimmunopathogenesis of DHF/DSS, although its cause re-mains unknown. Large amounts of dengue NS1, comple-ment anaphylatoxin C5a, and the terminal complementcomplex SC5b-9 were found in pleural fluids of patientswith DHF/DSS.57 Other mediators, such as histamine, tissueplasminogen activator, and macrophage inhibitory factor,have also been found in this disease.64,65

Together with the endothelial permeability seen in DHF,there is a marked thrombocytopenia. This has also been seenin df, although the drop in platelet numbers is less marked.It has been suggested that there is transient suppression ofhematopoiesis, although megakaryocytes have not beenshown to be infected with the virus.66 Other suggestions forthe thrombocytopenia include binding of the virus to plate-lets in the presence of virus-specific antibody, antivirusantibodies that cross-react with human platelets, and anti-platelet antibodies, the latter in a mouse model.67

The elevation of serum liver enzymes is a frequent find-ing in DV infection, and infrequently midzonal necrosis isdemonstrable in the liver and fulminant hepatitis may occur.Hepatocytes are therefore a likely target for the virus. Invitro infection of HepG2 cells results in apoptosis mediatedby TNF-related apoptosis-inducing ligand and other chemo-kines, such as IL-8, RANTES, and monocyte chemoattrac-tant proteins, with evidence that these actions may be me-diated by NS5 protein.68

Other mediators, such as histamine, tissue plasminogenactivator, and macrophage migration inhibitory factor, havealso been implicated in the pathogenesis of the severe formsof dengue infection.69

Pathology

There are a limited number of morphological studies inDHF/DSS. In BALB/c mice inoculated with dengue-2 ob-tained from human serum, focal alterations were found inthe liver, kidney, lung, and cerebellum. The presence of thevirus in these organs was confirmed by ultrastructural andimmunolocalization techniques in mosquito cell cultures ofthe infected tissues.70 Hepatocytes were ballooned; portaland centrilobular veins were congested; lungs were focallyhemorrhagic with vascular congestion and focal alveolitis;cerebellar tissue displayed focal neuronal compaction andperivascular edema; and there was increased glomerular

volume with augmented endocapillary and mesangial cel-lularity.

Children who die from DHF have entensive edema ofviscera with leakage of blood and focal hemorrhage (Figure1). In a study of five fatal cases of DHF in Vietnamesechildren, severe hepatitis was observed with midzonal ne-crosis and micro-vesicular steatosis, although in one patientthe liver appeared normal. The necrotic areas showed apo-ptosis by the TUNEL technique, with destruction of bothhepatocytes and Kupffer cells, and there was no recruitmentof polymorphonuclear cells or lymphocytes.71 Anotherstudy of nine fatal cases with fulminant hepatitis confirmedthe presence of DV cDNA by reverse transcriptase in situPCR in more than 80% of the hepatocytes and in manyKupffer cells. Five livers showed massive hepatocyte ne-crosis or apoptosis with no accompanying involvement ofbile ducts (Figure 2). There was rare bile canalicular cho-lestasis, and micro-vesicular steatosis was common. Thepauci-cellular areas of massive hepatic necrosis showedonly rare Kupffer cells to be positive for TNF-� and IL-272

compared with the upregulation of these and many othercytokines seen in the livers of fatal cases of hepatitis C.Thrombotic microangiopathy may be observed in the glo-meruli (Figure 3), most likely the result of disseminatedintravascular coagulopathy from hemoconcentration.14

Localization of DV antigen and RNA by immunohisto-chemistry and in situ hybridization confirmed the presenceof viral antigens in Kupffer cells, lymphoid cells in thesplenic red and white pulp, renal tubular epithelium, vascu-lar endothelium of the liver and lung, monocytes in theliver, spleen, and lung, and peripheral blood monocytes andlymphocytes (Figure 4 A and B), although high levels of

Figure 2 Liver biopsy from a 13-year-old Chinese boy fromSingapore who presented with dengue hemorrhagic fever andfulminant hepatitis. The needle core biopsy showed extensivepauci-cellular necrosis concentrated in the midzone, but in areastending to involve the entire lobule, the portal tract elements werenot involved (arrows). The nuclear debris represented apoptoticbodies.



viral RNA were only found in the antigen-bearing cells ofthe spleen and peripheral blood supporting viral replicationin these latter cells (Figure 4C and D).32 Viral antigen waspresent in the endothelium of the organs studied, but therewas no evidence of viral replication. The presence of in-creased numbers of ultrastructural vacuoles and pinocytic

vesicles in cutaneous vascular endothelium has suggestedthat the presence of viral antigen in these cells representsendocytosis and not infection.73

Genetic predisposition and resistance

The patient’s genetic background appears to be a criticalfactor in determining progression to DHF/DSS. Outbreaksin Cuba have shown a reduced risk of people of Negroidrace for DHF/DSS compared with those of Caucasoidrace,74 an observation coinciding with the low reportedincidence of dengue disease in African and Black Caribbeanpopulations. Although dengue has been documented in 19African nations and the virus repeatedly isolated, only spo-radic cases of dengue have been reported, and mainly innonindigenous populations. Even when outbreaks havebeen reported in Africa and the Seychelles, the clinicalmanifestations have been very mild.75 There is a clearabsence of DHF/DSS despite hyperendemic dengue virustransmission in the Haiti,76 this despite the annual infectionrate in Port-au-Prince being about 30% higher than that ofYangoon, Myanmar, where DHF/DSS rates and fatalitiesare high.

Although polymorphic genes have been suggested aspossible contenders to account for the genetic susceptibility,the HLA region is an obvious candidate as it encodesseveral proteins involved in the immune response, includingcomplement and TNF-�. Certain HLA types have been both

Figure 3 A 46-year-old Chinese woman with dengue hemor-rhagic fever. Her hemoglobin was 3 g/dL, platelets 64,000/dL, andhematocrit 26.4%. Peripheral blood flim showed no evidence offragmented cells. Thrombotic microangiopathy was evident in therenal biopsy (arrows, methanamine silver).

A B

C D

Figure 4 Avidin–biotin peroxidase staining with antidengue polyclonal antibodies for three serotypes (hyperimmune mouse ascitic fluid; kindlyprovided by R.E. Shope. WHO Centre for Tropical Diseases, University of Texas Medical Branch, Galveston, TX). (a) Staining was seen inKupffer cells. (b) The antigen was present in the white pulp of the spleen as well as in scattered larger monocytes in the red pulp. (c) In situhybridization showed viral genome in macrophages and stimulated lymphoid cells of the splenic red pulp and peripheral blood monocytes (d).32



positively and negatively associated with DHF. For exam-ple, variations in HLA-A locus were significantly associatedwith susceptibility to DHF/DSS77; HLA-DR04 has the re-verse association.78 Genetic background may also be relatedto the proliferation of low-affinity T cells, or the persis-tence of cytokine production by these cells, or both factorstogether.

Conclusions

Current information on the immunopathogenesis of DHF/DSS is still fragmentary. The severe forms of the DVinfection occur as secondary infections. The target cellsappear to be dendritic cells, monocytes, hepatocytes, Tlymphocytes, and possibly vascular endothelial cells. Thedisease results from two major immune mechanisms thatinvolve the production of nonneutralizing enhancing anti-bodies that cross-react between the serotypes of DV enhanc-ing viral entry into dendritic cells and monocytes to increasethe viral load and produce inefficient maturation of theinfected cells. The other component of the immune responseinvolves the massive activation of memory T cells sensi-tized in a previous infection. This activation results in theproliferation and release of pro-inflammatory cytokines anda cytokine cascade that targets the susceptible cells causingcell death through apoptosis and is responsible for the fluidand protein leakage and the liver damage characteristic ofDHF/DSS. Research is hampered by the absence of a suit-able animal model for DHF/DSS, and detailed pathologicalstudies in humans are few. Current treatment is largelysymptomatic, and prevention is through vector control,making it imperative that greater understanding of thepathogenesis be achieved with the view of developing novelmolecules to inhibit viral replication and to stem the dam-aging effects of immune mediators. There is also impetusfor vaccine development, and a greater understanding of therelation of resistance to the severe forms of the infection andgenetics is needed.

References

1. Gubler DJ: Dengue/dengue hemorrhagic fever: history and currentstatus. Novartis Found Symp 277:3-16, 2006

2. Gubler DJ: Dengue and dengue hemorrhagic fever. Clin MicrobiolRev 11:480-496, 1998

3. Gibbons RV, Vaughn DW: Dengue: an escalating problem. Br Med J324:1563-1566, 2002

4. Burke DS, Monath TP: Flaviviruses, in Knipe DM, Howley PM (eds):Fields Virology (ed 4, vol 1). Philadelphia, PA, Williams and Wilkins,2001, pp 1043-1126

5. Chaturvedi UC, Nagar R, Shrivastava R: Dengue and dengue hemor-rhagic fever: implications of host genetics. FEMS Immunol MedMicrobiol 47:155-166, 2005

6. Fink J, Gu F, Vasudevan SG: Role of T cells, cytokines and antibodyin dengue fever and dengue hemorrhagic fever. Rev Med Virol 16:263-275, 2006

7. Halstead SB: Neutralization and antibody-dependent enhancement ofdengue viruses. Adv Virus Res 60:421-467, 2003

8. Fernando RL, Fernando SSE, Leong AS-Y: Tropical Infectious Dis-eases. London, Greenwich Medical Media, 2001, pp 279-280

9. World Health Organisation: Dengue Hemorrhagic Fever: Diagnosis,Treatment, Prevention and Control (ed 2). Geneva, World HealthOrganisation, 1997

10. Wilder-Smith A, Schwartz E: Dengue in travellers. N Engl J Med353:924-932, 2005

11. Nugyen TL, Nguyen TH, Tien NT: The impact of dengue hemorrhagicfever on liver function. Res Virol 148:273-277, 1997

12. Balmeseda A, Hammond SN, Perez MA, et al: Short report: assess-ment of the World Health Organisation scheme for classification ofdengue severity in Nicaragua. Am J Trop Med Hyg 73:1059-1062,2006

13. Kamath SR, Ranjit S: Clinical features, complications and atypicalmanifestations of children with severe forms of dengue hamorrhagicfever in South India. Indian J Pediatr 73:889-895, 2006

14. Boonpucknavig V, Soontornniyomkij V: Pathology of renal diseasesin the tropics. Semin Nephrol 23:88-106, 2003

15. Seet RCS, Quek AML, Lim ECH: Post-infectious fatigue syndrome indengue infection. J Clin Virol 38:1-6, 2007

16. Kumar J, Kumar A, Gupta S, et al: Dengue hemorrhagic fever: anunusual cause of intracranial hemorhage. J Neurol Neurosurg Psychi-atry 78:253, 2007

17. Ling LM, Wilder-Smith A, Leo YS: Fulminant hepatitis in denguehemorrhagic fever. J Clin Virol 38:265-268, 2007

18. Khor BS, Liu JW, Lee IK, et al: Dengue hemorrhagic fever patientswith acute abdomen: clinical experience of 14 cases. Am J Trop MedHyg 74:901-904, 2006

19. Lu PL, Hsiao HH, Tsai JJ, et al: Dengue virus-associated hemophago-cytic syndrome and dyserythropoiesis: a case report. Kaohsiung J MedSci 21:34-39, 2005

20. Wiwanitkit V: Dengue myocarditis, rare but not fatal manifestation. IntJ Cardiol 10:122, 2006

21. Pokidysheva E, Zhang Y, Battisti AJ, et al: Cryo-EM reconstruction ofdengue virus in complex with the carbohydrate recognition domain ofDC-SIGN. Cell 124:485-493, 2006

22. Pulendren B, Ahmed R: Translating innate immunity into immuno-logical memory: implications for vaccine development. Cell 124:849-863, 2006

23. Cardosa MJ, Porterfield JS, Gordon S: Complement receptor mediatesenhanced Flavivirus replication in macrophages. J Exp Med 158:258-263, 1983

24. Gollins SW, Porterfield JS: A new mechanism for the neutralisationof enveloped viruses by anti-viral antibody. Nature 321:421-467,1986

25. Palmer DR, Sun P, Celluzzi C, et al: Differential effects of denguevirus on infected and bystander dendritic cells. J Virol 79:2432-2439,2005

26. Chang TH, Liao CL, Lin YL: Flavivirus induces interferon-beta geneexpression through a pathway involving RIG-I-independent IRF-3 andPI3K-dependent NF-kappaB activation. Microbes Infect 8:157-171,2006

27. Green S, Rothman A: Immunopatholgical mechanisms in dengueand dengue hemorrhagic fever. Curr Opin Infect Dis 19:429-436,2006

28. Navarro-Sanchez E, Despres P, Cedillo-Barron L: Innate immuneresponses to dengue virus. Arch Med Res 36:425-435, 2005

29. Shresta S, Kyle JL, Beatty PR, et al: Early activation of natural killerand B cells in response to primary dengue virus infection in A/J mice.Virology 319:262-273, 2004

30. Bente DA, Melkus MW, Garcia JV, et al: Dengue fever in humanisedNOD/SCID mice. J Virol 79:13797-13799, 2005

31. Littaua R, Kurane I, Ennis FA: Human IgG Fc receptor II mediatesantibody-dependent enhancement of dengue virus infection. J Immu-nol 144:3183-3186, 1990



32. Jessie K, Fong MY, Devi S, et al: Localisation of dengue virus innaturally infected human tissues, by immunohistochemistry and in situhybridisation. J Infect Dis 189:1411-1418, 2004

33. Huang KJ, Yang YC, Lin YS, et al: The dual-specific binding ofdengue virus and target cells for the antibody-dependent enhancementof dengue virus infection. J Immunol 176:2825-2832, 2006

34. Vaughn DW, Green S, Kalayanarooj S, et al: Dengue viremia titre,antibody response pattern, and virus serotype correlate with diseaseseverity. J Infect Dis 181:2-9, 2000

35. Libraty DH, Endy TP, Houng HS, et al: Differing influences of virusburden and immune activiation on disease severity in secondary den-gue-3 virus infections. J Infect Dis 185:1213-1221, 2002

36. Kliks SC, Nisalak A, Brandt WE, et al: Antibody-dependent en-hancement of dengue virus growth in human monocytes as a riskfactor for dengue hemorrhagic fever. Am J Trop Med Hyg 40:444-451, 1989

37. Laoprasopwattana K, Libraty DH, Endy TP, et al: Dengue virus (DV)enhancing antibody activity in pre-illness plasma does not predictsubsequent disease severity or viremia in secondary DV infection.J Infect Dis 192:510-519, 2005

38. Kliks SC, Nimmanitya S, Nisalak A, et al: Evidence that maternaldengue antibodies are important in the development of denguehemorrhagic fever in infants. Am J Trop Med Hyg 38:411-419,1988

39. Libraty DH, Young PR, Pickering D, et al: High circulating levels ofthe dengue virus non-structural protein NS1 early in dengue illnesscorrelate with the development of dengue hemorrhagic fever. J InfectDis 186:1165-1168, 2002

40. Chung KM, Nybakken GE, Thompson BS, et al: Antibodies againstWest Nile virus non-structural protein NS1 prevent lethal infectionthrough Fc gamma receptor-dependent and independent mechanisms.J Virol 80:1340-1351, 2006

41. Costa SM, Paes MV, Barreto DF, et al: Protection against dengue type2 virus induced in mice immunized with a DNA plasmid encoding thenon-structural 1 (NS1) gene fused to the tissue plasminogen activatorsignal sequence. Vaccine 24:195-205, 2006

42. Lin CF, Lei HY, Shiau AL, et al: Antibodies from dengue patient seracross-reacts with endothelial cells and induces damage. J Med Virol69:82-90, 2003

43. Lin CF, Chiu SC, Hsiao YL, et al: Expression of cytokine, chemokine,and adhesion molecules during endothelial activation induced by an-tibodies against dengue virus non-structural protein 1. J Immunol174:395-403, 2005

44. Alvarez M, Rodrigue-Roche R, Bernardo I, et al: Dengue hemorrhagicfever caused by sequential dengue 1-3 infections at a long interval:Havana epidemic. Am J Trop Med Hyg 75:1113-1117, 2006

45. Endy TP, Nisalak A, Chunsuttiwat S, et al: Spatial and temporalcirculation of dengue virus serotypes: a prospective study of primaryschool children in Kamphaeng Phet, Thailand. Am J Epidemiol 56:52-59, 2002

46. Endy TP, Nisalak A, Chunsuttitwat S, et al: Relationship of pre-existing dengue virus (DV) neutralizing antibody levels to viremia andseverity of disease in a prospective cohort study of DV infection inThailand. J Infect Dis 189:990-1000, 2004

47. Mongkolsapaya J, Dejnirattisai W, Xu XN, et al: Original antigenic sinand apoptosis in the pathogenesis of dengue hemorrhagic fever. NatMed 921-927, 2003

48. Mongkolsapaya J, Duangchinda T, Dejnirattisai W, et al: T cell re-sponses in dengue hemorrhagic fever: are cross-reactive T cells sub-optimal? J Immunol 176:3821-3829, 2006

49. Mangada MM, Rothman AL: Altered cytokine responses of dengue-specific CD4� T cells to heterologous serotypes. J Immunol 175:2676-2683, 2005

50. Bashyam HS, Green S, Rothman AL: Dengue virus-reactive CD8� Tcell display quantitative and qualitative differences in their response tovariant epitopes of heterologous viral serotypes. J Immunol 176:2817-2824, 2006

51. Page KR, Scott AL, Manabe YC: The expanding realm of heterolo-gous immunity: friend or foe? Cell Microbiol 8:185-196, 2006

52. Mathew A, Kurane I, Green S, et al: Impaired T cell proliferation inacute dengue infection. J Immunol 162:5609-5615, 1999

53. Pichyangkul S, Endy TP, Kalayanarooj S, et al: A blunted bloodplasmacytoid dendritic cell response to an acute systemic viral infec-tion is associated with increased disease severity. J Immunol 171:5571-5578, 2003

54. Simmons CP, Dong T, Chau NV, et al: Early T cell responses todengue virus epitopes ion Vietnamese adults with secondary denguevirus infections. J Virol 79:5665-5675, 2005

55. Talavera D, Castillo AM, Dominguez MD, et al: IL8 release, tightjunction and cytoskeleton dynamic reorganisation conducive to per-meability increase are induced by dengue virus infection of microvas-cular endothelial monolayers. J Gen Virol 85:1801-1813, 2004

56. Dewi BE, Takasaki T, Kurane I: In vitro assessment of human endo-thelial cell permeability: effects of inflammatory cytokines and denguevirus infection. J Virol Methods 121:171-180, 2004

57. Avirutnan P, Punyadee N, Noisakran S, et al: Vascular leakage insevere dengue virus infections: a potential role for the non-struc-tural viral protein NS1 and complement. J Infect Dis 193:1078-1088, 2006

58. Cardier JE, Marinao E, Romano E, et al: Proinflammatory factorspresent in sera from patients with acute dengue infection induceactivation and apoptosis of human microvascular endothelial cells:possible role of TNF-� in endothelial cell damage in dengue. Cytokine30:359-365, 2005

59. Shresta S, Sharar KL, Prigozhin DM, et al: Murine model for denguevirus-induced lethal disease with increased vascular permeability.J Virol 80:10208-10217, 2006

60. Boonpucknavig S, Boonpucknavig V, Bhamarapravati N, et al: Im-munofluorescence study of skin rash in patients with dengue hemor-rhagic fever. Arch Pathol Lab Med 103:463-466, 1979

61. Perez AB, Garcia G, Sierra B, et al: IL-10m levels in dengue patients:some findings from the exceptional epidemiological conditions inCuba. J Med Virol 73:230-234, 2004

62. Nguyen TH, Nguyen TL, Lei HY, et al: Association between sex,nutritional status, severity of dengue hemorrhagic fever, and immunestatus in infants with dengue hemorrhagic fever. Am J Trop Med Hyg72:370-374, 2005

63. Chakravarti A, Kumaria R: Circulating levels of tumor necrosis factor-alpha and interferon-gamma in patients with dengue and dengue hem-orrhagic fever during an outbreak. Indian J Med Res 123:25-30, 2006

64. Huang YH, Lei HY, Liu CC, et al: Tissue plasminogen activatorinduced by dengue virus infection of human endothelial cells. J MedVirol 70:610-616, 2003

65. Chen LC, Lei HY, Liu CC, et al: Correlation of serum levels of macro-phage migration inhibitory factor with disease severity and clinical out-come in dengue patients. Am J Trop Med Hyg 74:142-147, 2006

66. Wang S, He R, Patarapotikul J, et al: Antibody-enhanced binding ofdengue-2 virus to human platelets. Virology 213:254-257, 1995

67. Lin CF, Lei HY, Liu CC, et al: Generation of IgM anti-plateletautoantibody in dengue patients. J Med Virol 63:143-149, 2001

68. Medin CL, Fitzgerald KA, Rothman AL: Dengue virus non-structuralprotein/NS5 induces interleukin-8 transcription and secretion. J Virol79:11053-11061, 2005

69. Pang T, Cardosa MJ, Guzman MG: Of cascades and perfect storms:the immunopathogenesis of dengue hemorrhagic fever-dengue shocksyndrome (DHF/DSS). Immunol Cell Biol 85:43-45, 2007

70. Barreto DF, Takiya CM, Paes MV, et al: Histopathological aspects ofDengue-2 virus infected mice tissues and complementary virus isola-tion. J Submicrosc Cytol Pathol 36:121-130, 2004

71. Huerre MR, Lan NT, Marianneau P, et al: Liver histopathology andbiological correlates in five cases of fatal dengue fever in Vietnamesechildren. Virchows Arch 438:107-115, 2001

72. Carvalho de Macedo F, Nicol AF, Cooper LD, et al: Histologic, viral,and molecular correlates of dengue fever infection of the liver using



highly sensitive immunohistochemistry. Diagn Mol Pathol 15:223-228, 2006

73. Sahaphong S, Riengrojpitak S, Bhamarapravati N, et al: Electron micro-scopic study of the vascular endothelial cell in dengue hemorrhagic fever.Southeast Asian J Trop Med Pub Health 11:194-204, 1980

74. de la C Sierra B, Kouri G, Guzman MG: Race: a risk factor for denguehemorrhagic fever. Arch Virol 152:533-542, 2007

75. Gubler DJ: The global pandemic of dengue/dengue hemorrhagic fever:current status and prospects for the future. Ann Acad Med Singapore27:227-234, 1998

76. Halstead SB, Streit TG, Lfontant JG, et al: Haiti: absence of denguehemorrhagic fever despite hyperendemic dengue virus transmission.Am J Med Hyg 65:180-183, 2001

77. Loke H, Bethell DB, Phung CX, et al: Strong HLA class-I restricted Tcell responses ind engue hemorrhagic fever. A double-edged sword?J Infect Dis 184:1369-1373, 2001

78. LaFluer C, Granados J, Vargas-Alarcon G, et al: HLA-DR antigenfrequencies in Mexican patients with dengue virus infections: HLA-DR4as a possible genetic resistance factor for dengue hemorrhagic fever. HumImmunol 63:1039-1044, 2002


author's personal copy the pathology of dengue hemorrhagic fever

Documents