Viral Haemorrhagic Fevers: an Overview

Viral haemorrhagic fever, abbreviated VHF, names a clinical syndrome rather than a single disease. At least thirty ribonucleic acid (RNA) viruses drawn from four groups can produce it, and they share almost nothing in their genome or replication beyond the illness they cause. Each is an enveloped, single-stranded RNA virus; almost all are maintained in nature by animals, so human infection is usually an accidental spillover from a reservoir species, its excretions, or the bite of an infected tick or mosquito. The one consistent exception is dengue, which is sustained between people by mosquitoes rather than in an animal reservoir. What binds the group together is a shared pathophysiology: high-level viraemia, infection of the body’s first-line immune cells, and a systemic inflammatory response that raises vascular permeability, lowers blood pressure and disturbs coagulation.

The four groups are the Arenaviridae, the Filoviridae, the order Bunyavirales and the Flaviviridae. This four-part scheme is the practical clinical frame; the current taxonomy is finer, and the old single family Bunyaviridae has been split into several families within the order Bunyavirales, of which Nairoviridae (Crimean-Congo haemorrhagic fever), Phenuiviridae (Rift Valley fever) and Hantaviridae (the hantaviruses) carry the haemorrhagic agents. Understanding the syndrome means holding two facts together: the agents are biologically diverse, yet they converge on one recognisable clinical picture.

The viral haemorrhagic fever syndrome

The defining features are fever and malaise progressing, in severe cases, to a fall in blood pressure, coagulation failure and shock. The syndrome sits at the extreme end of cross-species virus transfer: a virus adapted to an animal host replicates so efficiently in human cells, and so effectively blunts human immune defences, that it drives an overwhelming inflammatory illness. VHF does not localise to a single organ in the way that a hepatitis virus targets the liver or an encephalitis virus the brain; it is a systemic disease in which the dominant problem is loss of vascular integrity.

The name is partly misleading. Frank haemorrhage is frequently a minor feature and is rarely the presenting sign, and massive blood loss is uncommon even in the most lethal infections, where it is usually confined to the gastrointestinal tract. The clinically decisive lesion is increased vascular permeability, the same capillary leak that characterises septic shock. A small number of agents from outside this group, including measles and the viruses of fulminant hepatitis, can occasionally cause haemorrhagic manifestations, but they are not part of the syndrome.

The four virus families and their agents

Every VHF agent is an enveloped RNA virus, but the four groups differ in genome structure. The Flaviviridae carry a single positive-sense strand; the Filoviridae a single negative-sense strand; the Arenaviridae two segments; and the members of the Bunyavirales three segments. These structural differences matter for replication and evolution but not for the shared clinical picture. What shapes the human disease pattern more directly is the reservoir and the route of transmission: a tick-borne agent acquired at slaughter produces sporadic single cases, whereas a mosquito-borne agent can drive explosive epidemics.

Family or order	Representative agents	Reservoir and vector	Main regions	Person-to-person	Case fatality (severe)
Arenaviridae (Old World)	Lassa, Lujo, lymphocytic choriomeningitis virus	Rodents (Lassa: Mastomys natalensis); contact with excreta	West and southern Africa	Yes, blood and body fluids	Lassa ~1% overall, higher in hospital; Lujo ~80%
Arenaviridae (New World)	Junin, Machupo, Guanarito, Sabia	Rodents; contact with excreta	South America	Limited	~15 to 60%
Filoviridae	Ebola (including Zaire, Sudan, Bundibugyo), Marburg	Bats; direct contact with body fluids	Sub-Saharan Africa	Yes, high	Ebola ~25 to 90%, Marburg ~24 to 88%
Bunyavirales: Nairoviridae	Crimean-Congo haemorrhagic fever	Hyalomma ticks and livestock	Africa, Europe, Asia	Yes, nosocomial	~30%
Bunyavirales: Phenuiviridae	Rift Valley fever	Mosquitoes; livestock amplifiers	Sub-Saharan and North Africa	No	~1%, higher in the haemorrhagic form
Bunyavirales: Hantaviridae	Hantaan, Seoul, Puumala; Sin Nombre, Andes	Rodents; inhaled excreta	Eurasia (renal), Americas (cardiopulmonary)	Andes only	Cardiopulmonary ~40 to 60%
Flaviviridae	Yellow fever, severe dengue	Mosquitoes	Tropical Africa, the Americas, Asia	No	Yellow fever ~30 to 50% severe; dengue low

Two patterns run through this map. First, rodents and bats supply most VHF agents, because their large, dense populations sustain continuous circulation of the virus. Second, the person-to-person column predicts the risk to health workers: the filoviruses, Lassa, Lujo and Crimean-Congo haemorrhagic fever spread readily through blood and body fluids and have caused devastating hospital outbreaks, whereas Rift Valley fever and dengue pose little direct threat to staff. Crimean-Congo haemorrhagic fever and Rift Valley fever are the two arthropod-borne VHF agents endemic to southern Africa, and they share the vector ecology of the wider group of tick-borne and mosquito-borne viruses.

Shared pathophysiology

The unifying mechanism begins with the cells the viruses infect first. VHF agents preferentially replicate in monocytes, macrophages and dendritic cells, the sentinels that normally raise the alarm against an invading pathogen. Infecting these cells achieves two things at once: it disables early defence, and it turns the infected cell into a vehicle that carries virus to the draining lymph nodes and onward to the liver, spleen and other organs. This helps explain both the high viral loads and the speed with which severe infections overwhelm the host.

Tropism beyond these sentinel cells varies and tracks with severity. Dengue largely infects monocytes and macrophages without killing them, and tissue injury is correspondingly modest. Ebola and Marburg, at the other extreme, infect a very broad range of cells, including endothelium, hepatocytes, fibroblasts and adrenal cortical cells, producing widespread tissue destruction. Most agents lie between, with the liver as a principal target: hepatic infection begins in the fixed macrophages of the sinusoids (Kupffer cells) and extends to the parenchyma, which is why jaundice is prominent in yellow fever and seen in some Rift Valley fever and dengue.

Across the group, the shared immune-evasion strategy is suppression of the type I interferon response, the cytokine system that would otherwise limit viral spread. Viral proteins block it at several points: the Ebola VP35 protein hides viral RNA from the sensor RIG-I, the Ebola VP24 protein blocks interferon signalling through STAT1, and arenavirus and flavivirus proteins interrupt the same pathways. Because double-stranded RNA generated during replication is a strong interferon trigger, every RNA virus must evolve some evasion; the haemorrhagic fever viruses appear to be the subset that suppress it so completely that replication runs unchecked. In the most severe filovirus infections both the antibody and the T-cell responses also fail.

Paradoxically, this immunosuppression coexists with a cytokine storm. The infected monocytes and macrophages release large quantities of pro-inflammatory mediators, including interleukin-6, interleukin-8, tumour necrosis factor alpha, and reactive oxygen and nitrogen species. Sustained high levels are directly toxic: they kill bystander lymphocytes by apoptosis, damage tissue, and, through nitric oxide, lower vascular tone. It is these circulating mediators, rather than direct viral injury to the vessel wall, that are now thought to drive the central lesion. Increased vascular permeability, the capillary leak, follows from cytokine action on the endothelium, with disruption of the tight-junction protein VE-cadherin and, for the hantaviruses, a role for bradykinin. Direct endothelial infection can occur and is documented in fatal Crimean-Congo haemorrhagic fever, but it is no longer seen as the primary cause of the leak.

The bleeding tendency has two roots. Reduced hepatic synthesis of clotting factors combines with a disseminated intravascular coagulation picture, in which the coagulation system is activated throughout the circulation, consuming platelets and factors. Thrombocytopenia is close to universal across the syndrome, though it is usually not severe enough on its own to explain haemorrhage. The multi-organ syndrome resembles septic shock: an early, regulated inflammatory response is associated with recovery, whereas a dysregulated one predicts death. In survivors the difference is measurable as an early, robust CD4 and CD8 T-cell response that non-survivors fail to mount. A further contributor to hypotension is adrenal cortical infection and necrosis, which impairs the steroid synthesis needed to maintain blood pressure and sodium balance and is found in nearly all fatal cases.

The clinical spectrum

The early illness is non-specific: fever, malaise, myalgia and headache, reflecting the release of inflammatory mediators. Onset is characteristically insidious in Lassa fever but so abrupt in yellow fever, Crimean-Congo haemorrhagic fever and filovirus disease that a patient may report the hour it began. Fever is usually high and unremitting, though it can be blunted in pregnant women and the elderly. Bradycardia is a recognised feature of the arenavirus and filovirus infections and of yellow fever. Vomiting and profuse diarrhoea are prominent in filovirus disease and drive early fluid loss, making intravenous rehydration a mainstay of care; abdominal pain can be severe enough to prompt surgery, and hospital outbreaks of both Crimean-Congo haemorrhagic fever and Ebola have begun with a patient taken to theatre.

Signs of vasodilatation and capillary leak give the syndrome its recognisable face: a blanching erythema of the upper trunk and face, conjunctival injection, and an early rash in Marburg, Ebola, Crimean-Congo haemorrhagic fever and dengue that can be hard to see in darker skin. Facial and limb oedema in Lassa is a capillary-leak sign that may appear only once rehydration begins. When bleeding does occur it shows as oozing venepuncture sites, gastrointestinal or urinary blood, and easy bruising; large ecchymoses are characteristic of Crimean-Congo haemorrhagic fever and unusual in the others.

The laboratory picture is suggestive but never diagnostic. Leukopenia is frequent, though the white cell count is typically normal in severe dengue and fatal Lassa may be heralded by a neutrophil leukocytosis. Thrombocytopenia is essentially universal. Rising haemoglobin and haematocrit signal plasma loss from the leaking vasculature and, followed serially, track the adequacy of resuscitation. Aminotransferases are markedly raised in filovirus disease, yellow fever and Rift Valley fever, and prolonged prothrombin and partial thromboplastin times, with fibrin degradation products, mark the disseminated intravascular coagulation of the severe infections. Some presentations carry a specific stamp: third-trimester Lassa and filovirus infection carries maternal and fetal mortality approaching 100%, Rift Valley fever can cause an immune-mediated retinal vasculitis and meningoencephalitis in the second week, and up to about a third of survivors of clinical Lassa fever develop permanent sensorineural hearing loss, making Lassa a leading cause of acquired deafness in its endemic region. Filovirus survivors may develop uveitis, with live virus recovered from the aqueous humour of the eye.

Approach to a suspected case

Because the early illness resembles many common infections, suspicion rests on exposure history far more than on any physical sign. A febrile patient who has recently travelled to or from an area where these viruses circulate must be asked about contact with animals, ticks, or ill people, about attendance at funerals, and about work in healthcare or laboratories. The incubation period across the syndrome runs from about 2 to 21 days, which is why a 21-day window from last possible exposure is used operationally to define who is at risk.

The single most important competing diagnosis is malaria. VHF is conservatively estimated at fewer than one case per million travel episodes to an endemic country, and a febrile returning traveller is at least a thousand times more likely to have malaria than a viral haemorrhagic fever. The corollary is not complacency but discipline: both a treatable common infection and a rare dangerous one must be actively considered, because imported VHF cases have been missed with serious consequences, and malaria and VHF can coexist. Suspicion rightly rises when a patient reports contact with an ill person or a cluster, or presents during a recognised outbreak.

Confirmation requires tests that identify the pathogen directly. Reverse transcription polymerase chain reaction (RT-PCR) on blood is the primary diagnostic tool for most agents, because viraemia is prolonged and high; oral-fluid sampling works for filoviruses and Lassa, and reduces the biosafety risk of a blood draw. Serology by enzyme-linked immunosorbent assay (ELISA) is more useful than RT-PCR for the hantaviruses and sometimes for dengue, where the viraemia may already be falling by the time symptoms bring the patient in; acute-phase serology is otherwise unreliable because the impaired immune response is slow to generate antibody. The practical rule when VHF is a real possibility is to seek advice before sampling: on which tests to send for the likely diagnoses, and on how to collect, package and transport specimens safely to a laboratory equipped to handle them.

The first management priority once VHF is plausible is to interrupt transmission by isolating the patient and applying barrier precautions, before the diagnosis is confirmed, because the risk to others is greatest in the undiagnosed phase. Care is otherwise supportive and can be highly effective: careful attention to circulating volume and electrolytes, blood products, and organ support including dialysis and ventilation where needed. Corticosteroids, non-steroidal anti-inflammatory drugs and aspirin are avoided, the first for lack of benefit and the latter two for bleeding and renal risk. Specific therapies are agent-dependent and are developing: ribavirin has an established role in Lassa fever and haemorrhagic fever with renal syndrome by older evidence, though more recent appraisal regards the benefit in Lassa and Crimean-Congo haemorrhagic fever as unproven, and monoclonal antibodies have transformed the treatment of Zaire ebolavirus disease.

Biosafety levels and containment

Work with these viruses is governed by the biosafety level (BSL) system, a graded set of laboratory containment requirements from BSL-1 for agents that do not cause disease in healthy adults to BSL-4 for the most dangerous. The level combines engineering controls, work practices and personal protective equipment (PPE) to match the hazard.

Biosafety level	Typical containment	Representative agents
BSL-2	Standard practice, limited access, biosafety cabinet for aerosol-generating steps	Dengue, yellow fever vaccine strain
BSL-3	Controlled access, directional inward airflow, respiratory protection, work in a cabinet	Rift Valley fever, hantaviruses, wild-type yellow fever
BSL-4	Maximum containment: positive-pressure suit or class III cabinet, dedicated air and effluent treatment, airlocked entry	Ebola, Marburg, Lassa, Lujo, Crimean-Congo haemorrhagic fever, New World arenaviruses

The filoviruses, Lassa, Lujo, Crimean-Congo haemorrhagic fever and the New World arenaviral agents are handled at BSL-4, the highest level of containment, which is why definitive testing is centralised in a small number of reference laboratories. Diagnostic tests that inactivate the virus first, and molecular assays run under validated conditions, can often be performed at lower containment, which is what allows essential clinical-pathology results to be produced without shipping every sample to a maximum-containment facility. The generic principles of biosafety cabinet use and laboratory practice belong to laboratory health and safety; here the point is which VHF agents demand which level.

Prevention and outbreak response

Sporadic spillover cases cannot be wholly prevented, because the reservoirs are part of the natural landscape. What prevents an outbreak is early detection and a fast response: alert clinicians, accessible diagnostics, and adherence to the International Health Regulations (2005), the legal framework under which countries detect, report and respond to events of international concern. Health workers are often among the first casualties when a case is recognised late, so the containment of a single case and the protection of staff are the same task.

Specific preventive tools vary by agent and are strongest where a vaccine exists. Yellow fever is prevented by a single dose of the live-attenuated 17D vaccine that confers lifelong immunity, and licensed vaccines and monoclonal antibodies now exist against Zaire ebolavirus, though not against the other Ebola species or Marburg. For the tick-borne and mosquito-borne agents, personal protection against bites, safe animal-husbandry and slaughter practices, and vector control carry most of the preventive weight. For the filoviruses, the recognition that virus persists in immune-privileged sites such as semen for months after recovery has added survivor follow-up and safe-sex advice to outbreak control.

South African context

Two viral haemorrhagic fevers are endemic to southern Africa. Crimean-Congo haemorrhagic fever is acquired from Hyalomma tick bites and from contact with the blood of infected livestock at slaughter, so cases are sporadic and concentrate in farming and abattoir communities; it is a notifiable condition and has transmitted to hospital staff. Rift Valley fever behaves first as a disease of livestock: heavy rains drive mosquito-borne epizootics with waves of abortion in sheep and cattle that act as a sentinel, and human infection follows from mosquito bites and from contact with infected animal tissue, occasionally progressing to the haemorrhagic or encephalitic form.

Lujo virus, a southern African arenavirus, is known from a single event: a 2008 nosocomial cluster that began in a patient airlifted from Zambia to Johannesburg and spread to four healthcare workers, killing four of the five people infected, a case fatality of 80%. No further human cases have been recorded since, and the reservoir, presumed to be a rodent, has never been identified.

Imported filovirus disease is a standing preparedness concern rather than an endemic risk, kept prominent by outbreaks in central and east Africa, including the 2026 Bundibugyo virus outbreak. A suspected case sets in motion a defined national response of risk assessment, safe specimen handling, isolation and notification, coordinated through the National Institute for Communicable Diseases (NICD), which houses the country’s viral haemorrhagic fever reference laboratory.

References and recommended reading

• Carson G, Bray M, Roth C. Viral Haemorrhagic Fevers. In: Clinical Virology, 4th edition, Chapter 9. ASM Press; 2016. The comparative account of the VHF syndrome across its four causative families and the foundation for this overview.
• Flórez-Álvarez L, de Souza EE, Botosso VF, et al. Hemorrhagic fever viruses: pathogenesis, therapeutics, and emerging and re-emerging potential. Frontiers in Microbiology. 2022;13:1040093. The current mechanistic account of interferon antagonism, cytokine storm and vascular permeability.
• Hewson R. Understanding Viral Haemorrhagic Fevers: Virus Diversity, Vector Ecology, and Public Health Strategies. Pathogens. 2024;13(10):909. The current reference for VHF taxonomy, reservoir and vector ecology, and control.