Science Wiki

Yellow fever is an acute, mosquito-borne viral hemorrhagic fever that affects humans and non-human primates. It is caused by yellow fever virus (YFV, species Orthoflavivirus flavi), the prototype member of the genus Orthoflavivirus in the family Flaviviridae. The name refers to the jaundice that occurs in severe disease (Latin flavus, "yellow"). Historical names include "yellow jack" (the quarantine flag), "black vomit", and vomito negro. The virus is transmitted by the bite of infected female mosquitoes of the genera Aedes (in Africa), Haemagogus, and Sabethes (in South America), and by the urban mosquito Aedes aegypti. It circulates in three cycles: a sylvatic (jungle) cycle between forest mosquitoes and non-human primates, an intermediate (savannah) cycle in Africa in which semi-domestic mosquitoes bite both primates and people, and an urban cycle in which Aedes aegypti transmits the virus between humans.

Most infections are mild or without symptoms. In a minority of patients, illness progresses after a short remission to a toxic phase with jaundice, kidney failure, bleeding, and shock, and about half of the patients who enter this phase die. Yellow fever was the first human virus to be identified (1901) and the first virus shown to be transmitted by mosquitoes. It is preventable by a single dose of the live attenuated 17D vaccine, developed by Max Theiler in the 1930s, which gives long-lasting protection. There is no licensed antiviral drug, and treatment is supportive. The disease is endemic in 34 countries in Africa and 13 in Central and South America. The World Health Organization (WHO) estimated, from modeling for 2013, between 84,000 and 170,000 severe cases and between 29,000 and 60,000 deaths per year, with most of the burden in Africa. Outbreaks continue in the 2020s despite the availability of the vaccine.

Classification and genotypes

YFV is the type species of the genus Orthoflavivirus and the founder of the "yellow fever group" of mosquito-borne viruses, which includes Wesselsbron, Sepik, Edge Hill, and Banzi viruses. All wild-type YFV belongs to a single serotype, so antibodies induced by the vaccine neutralize wild-type viruses from Africa and South America. Phylogenetic analysis divides YFV into seven genotypes that differ at roughly 10 to 25 percent of nucleotides and a few percent of amino acids:

The origin of YFV is in Africa. Molecular clock estimates differ by method, but place the origin of current lineages within roughly the last 1,500 to 3,000 years, and the introduction to the Americas, which came from West Africa, about 300 to 400 years ago.

Virion

The virion is an enveloped sphere about 50 nm in diameter. In mature virions of orthoflaviviruses, 180 copies of the envelope (E) glycoprotein and 180 copies of the membrane (M) protein lie flat on the lipid envelope. The E proteins form 90 head-to-tail dimers, arranged in 30 "rafts" of three parallel dimers in a herringbone pattern with a quasi-icosahedral lattice (T = 3). The nucleocapsid, a complex of the genome and the capsid (C) protein, is less ordered than the envelope. The E protein has three domains: domain I at the center of the monomer, domain II (the dimerization domain, which carries the conserved fusion loop), and domain III (an immunoglobulin-like domain at the top of the virion that has an important role in receptor binding and carries neutralizing epitopes). In immature virions, the precursor prM covers the fusion loop. Cleavage of prM to M by the host protease furin in the trans-Golgi network produces infectious particles. Unlike dengue and Zika viruses, many yellow fever virus strains lack an N-linked glycan on the E protein, although the pattern varies among strains.

Genome and proteins

The genome is a single molecule of positive-sense, single-stranded RNA of about 10.9 kb (10,862 nucleotides in the reference strains of 17D). It has a type 1 cap at the 5′ end and no poly(A) tail. One open reading frame of 3,411 codons, between a 5′ untranslated region of about 118 nucleotides and a 3′ untranslated region of about 500 nucleotides, is translated as a polyprotein. It is cleaved into three structural proteins (C, prM/M, and E) and seven non-structural proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5):

• NS3 (with NS2B as cofactor) is a serine protease and also an RNA helicase and triphosphatase.
• NS5 is a methyltransferase and the RNA-dependent RNA polymerase, and it antagonizes interferon signaling. YFV NS5 binds the host transcription factor STAT2 after interferon stimulation, which blocks the induction of antiviral genes.
• NS1 is found in replication complexes, on the cell surface, and as a secreted hexamer in blood. Experimental work suggests that secreted NS1 of orthoflaviviruses acts on endothelial cells in a tissue-specific way, and that YFV NS1 can affect liver endothelium. NS1 is also a diagnostic target.
• The 3′ untranslated region carries structured elements, including repeated sequences whose copy number differs among genotypes, and resistance to the host exonuclease XRN1 leads to the accumulation of a subgenomic flavivirus RNA (sfRNA). The sfRNA counteracts host antiviral responses, including RNA interference in mosquitoes.

Replication

The virus attaches to host cells through attachment factors, including heparan sulfate and phosphatidylserine receptors, and enters by receptor-mediated endocytosis. Acidification of the endosome causes a rearrangement of E, which leads to fusion of the viral and endosomal membranes. The genome is translated at the endoplasmic reticulum, where the non-structural proteins remodel membranes into vesicle packets that shelter replication from cytosolic sensors such as RIG-I and MDA5. Virions assemble at the endoplasmic reticulum, mature in the secretory pathway, and are released by exocytosis. The main target cells in the host include dendritic cells and macrophages (early), and hepatocytes, which are the main site of severe disease.

Attenuation of the 17D strain

The 17D vaccine strain was derived from the virulent Asibi strain by serial passage in tissue culture (176 passages in the original series). The comparison of the genomes of Asibi and 17D, first done in the 1980s, showed 68 nucleotide differences and 32 amino acid substitutions, of which 12 are in the E protein. The attenuating changes are distributed in the structural and the non-structural genes, and the genetic basis of attenuation has been explained only in part. Well-documented features of 17D include:

• Substitutions in E, such as the T380R change, which creates an Arg-Gly-Asp (RGD) motif in domain III. Several changes in E alter binding to heparan sulfate and affect how the virus enters cells, and it has been proposed that this limits spread from the skin to the liver. Experimental work has suggested that 17D uses a different endocytic route than Asibi, and that it induces stronger innate immune signaling in dendritic cells.
• Reduced replication and spread in the liver in primates, and loss of the ability to infect mosquitoes by feeding, so that the vaccine virus is not transmitted by mosquitoes in practice.
• Genetic stability, in that reversion to a virulent form has not been seen during the decades of use. Vaccine-associated severe disease results from host factors, as described below, and not from reversion.

Origins and early epidemics

YFV was probably introduced into the Americas from West Africa in the sixteenth and seventeenth centuries by the slave trade, together with the mosquito Aedes aegypti, which breeds in water casks on ships. The first epidemics that are documented with confidence in the Americas were in 1647 to 1648 in Barbados, Guadeloupe, Cuba, and the Yucatán Peninsula. Recurrent epidemics then struck port cities of the Caribbean, North and South America, and Europe, where mosquitoes could survive the summer. Notable examples include Philadelphia in 1793, in which about 5,000 people died (about a tenth of the population), New Orleans in 1853 (about 8,000 deaths), Memphis in 1878, Barcelona in 1821, and Cádiz in 1800. Yellow fever also contributed to the failure of the French attempt to build the Panama Canal in the 1880s, in which tens of thousands of workers died of yellow fever and malaria combined, and it destroyed much of the French army sent to Saint-Domingue in 1802, which influenced the outcome of the Haitian Revolution and the sale of Louisiana to the United States. For centuries the disease was ascribed to miasmas or to contact with the sick.

Discovery of mosquito transmission

The Cuban physician Carlos Finlay proposed in 1881 that yellow fever is transmitted by mosquitoes, and he identified the species, which he called Culex fasciatus, now Aedes aegypti. In 1900, after the Spanish-American War, the United States Army Yellow Fever Commission, led by Walter Reed with James Carroll, Aristides Agramonte, and Jesse William Lazear, tested the hypothesis in volunteers at Camp Lazear near Havana. Lazear died of yellow fever, contracted from a mosquito bite, during the work. The commission showed that:

• A mosquito that feeds on a patient in the first days of illness becomes able to transmit the disease only after an interval of about 12 days (the extrinsic incubation period, EIP).
• Soiled clothing and bedding from patients, the fomites on which public health measures had been focused, did not transmit the disease.
• The agent passed through a bacteria-retaining filter, which made yellow fever the first human disease shown to be caused by a filterable virus (1901).

William C. Gorgas used the findings to control mosquitoes in Havana (from 1901) and in the Panama Canal Zone (from 1904), by eliminating breeding sites and by screening, and Oswaldo Cruz did the same in Rio de Janeiro. Urban yellow fever was eliminated from Havana within months, and from the Canal Zone within a few years, and the work made the completion of the Panama Canal possible. The Rockefeller Foundation later led campaigns against the mosquito in Latin America, and Fred Soper's campaigns to eradicate Aedes aegypti from large parts of the Americas in the 1930s to 1960s succeeded for a time. Mosquitoes recolonized after the campaigns ended.

The Asibi strain and the 17D vaccine

In 1927, Adrian Stokes, Johannes Bauer, and N. Paul Hudson, working in Accra in the Gold Coast (now Ghana) for the Rockefeller Foundation, infected a rhesus macaque, the first susceptible laboratory animal, with blood from a Ghanaian patient named Asibi. The Asibi strain became the parent of the 17D vaccine. A second isolate from Dakar gave the French viscerotropic virus, which was used to make a mouse-brain-passaged neurotropic vaccine in France. That vaccine caused encephalitis in some children and was discontinued in 1982. Max Theiler and Hugh Smith at the Rockefeller Institute passaged Asibi virus for 176 passages through mouse embryo tissue, then whole chick embryo tissue, and then chick embryo tissue with the brain and spinal cord removed. The resulting strain, 17D, lost its ability to cause fatal viscerotropic and neurotropic disease in monkeys while still inducing immunity. It was first used in people in 1937 and 1938, and Theiler received the 1951 Nobel Prize in Physiology or Medicine. All current vaccines are derived from this strain.

Re-emergence

After mass vaccination and vector control reduced the disease in the mid-twentieth century, a decline in vaccination coverage and the reinfestation of cities by Aedes aegypti brought a resurgence from the 1980s. Large outbreaks in Africa and South America since 2015 are described below.

Transmission cycles

Vectors and transmission dynamics

After a mosquito has fed on a viremic host, the virus infects the midgut, escapes into the hemolymph, and infects the salivary glands. The mosquito can then transmit the virus in saliva after the extrinsic incubation period (EIP), which shortens as temperature rises. Experiments have shown EIPs of about 12 days at around 25 °C, much longer at temperatures near 18 °C, and about 4 to 7 days at 30 to 35 °C. Mosquitoes remain infectious for life. For a vector-borne pathogen, the basic reproduction number in the classical Ross-Macdonald model is:

where m is the number of mosquitoes per host, a is the biting rate, b is the probability that an infectious bite infects a host, c is the probability that a mosquito becomes infected from a viremic host, μ is the daily mosquito mortality, n is the EIP in days, and r is the host recovery rate (about 1/3 to 1/5 per day for the period of viremia). The factor e^−_μn_ is the probability that a mosquito survives the EIP. If μ = 0.1 per day, a mosquito survives a 12-day EIP with probability e^−1.2 ≈ 0.30, and a 6-day EIP with probability e^−0.6 ≈ 0.55, so that warming has a large effect on transmission. Estimates of R₀ for urban yellow fever transmitted by Aedes aegypti vary with the setting, and are often placed at about 3 to 6, with higher values in some models. The herd immunity threshold is:

so that R₀ = 3 corresponds to a threshold of 67 percent, and R₀ = 5 to 80 percent. Because the vaccine has an efficacy E of about 0.99 and not 1, the vaccination coverage that is needed is p_c / E, which for R₀ = 5 is about 81 percent. This is the reason why a coverage of about 80 percent is the target in high-risk areas.

Non-vector transmission

Transmission without a mosquito is rare. Transmission by blood transfusion or organ transplantation, and from mother to infant by breastfeeding, has been reported for the vaccine virus and, in a few cases, for wild-type virus. Laboratory infections have occurred. Humans with wild-type infection are viremic for about 3 to 5 days, and during this time they can infect mosquitoes and should be protected from bites.

Non-human primate hosts

Old World monkeys in Africa, such as baboons, vervets, and guenons, are infected with the virus and generally have viremia but little illness, which is consistent with a long history of co-existence. New World monkeys have been exposed to the virus only since it arrived in the Americas, and some are highly susceptible. Howler monkeys (Alouatta) and some marmosets die in large numbers, whereas capuchins and some other primates are less affected. Deaths of howler monkeys in forests are therefore a sentinel for the arrival of the virus, and they come before human cases by days to weeks. During the outbreaks in southeastern Brazil in 2016 to 2019, large numbers of monkeys died, and populations of the brown howler monkey (Alouatta guariba) declined sharply in affected areas. Surveillance of dead monkeys, including a smartphone reporting system in Brazil, is used to trigger vaccination of nearby human populations.

Burden

YFV is endemic in 34 countries in Africa and in 13 countries in Central and South America, and Africa has about 90 percent of the burden. The WHO estimated, for 2013, between 84,000 and 170,000 severe cases and between 29,000 and 60,000 deaths per year, using a model that corrects for under-reporting in surveillance data. The numbers that are reported to the WHO are much smaller, because most cases are not detected or are not confirmed. If a proportion f of infections is symptomatic and a proportion g of the symptomatic cases is severe, then the number of severe cases is about N f g, so that surveillance of severe cases corresponds to a much larger number of infections. The estimates are uncertain, and they depend on vaccination coverage, which is changing.

Africa

Outbreaks occur in the savannah regions of West and Central Africa, and in East Africa. The largest recorded epidemic in Africa was in Ethiopia in 1960 to 1962, with an estimated 100,000 to 200,000 cases and about 30,000 deaths. Repeated outbreaks in Nigeria since the late 1980s, and in Senegal, Ghana, Côte d'Ivoire, Guinea, Cameroon, Uganda, Kenya, and Sudan, have been reported. In 2015 to 2016, an outbreak of the Angola genotype began in the Viana district of Luanda, Angola, in a population that was mostly not vaccinated. It spread across Angola and to Kinshasa in the Democratic Republic of the Congo, with about 7,300 suspected cases, 962 confirmed cases, and about 400 deaths. It was controlled by a vaccination campaign that used about 30 million doses, including fractional doses in Kinshasa. Eleven infected workers who had returned from Luanda to China in 2016 were the first documented importations of yellow fever to Asia.

The Americas

Yellow fever in South America was long limited to the Amazon basin and the central west of Brazil. From 2016, a sylvatic wave of South American genotype I (sublineage 1E) spread south from the Amazon into the Atlantic forest of southeastern Brazil, where large populations lived in states that had not been in the area recommended for vaccination. Between late 2016 and 2019, there were about 2,250 confirmed human cases and about 770 deaths (a case fatality of about 34 percent) in Minas Gerais, Espírito Santo, Rio de Janeiro, and São Paulo, all transmitted by forest mosquitoes without urban transmission. Brazil vaccinated tens of millions of people. The factors that favored the wave include the lack of immunity in the human population, the movement of the virus along forest fragments, and weather. Drought and heat have been suggested as contributors, but the evidence for a causal role of climate in this wave is not conclusive. In 2024 and 2025, human cases and monkey deaths increased in several countries in South America, including Brazil, Colombia, Peru, Bolivia, and Ecuador, with a high case fatality. Current numbers should be checked in the reports of the Pan American Health Organization.

Why not Asia?

Yellow fever has never become established in Asia, although Aedes aegypti and Aedes albopictus are abundant, macaques are susceptible, and the climate suits transmission. The reasons are not settled. Proposed explanations include:

• Cross-protection against yellow fever from earlier infection with other orthoflaviviruses, such as dengue and Japanese encephalitis viruses, which are widespread in Asia. Experiments in animals have given mixed results, so the importance of this effect in people is not clear.
• Differences in the ability of Asian populations of Aedes aegypti to transmit the virus.
• Interference between viruses in co-infected mosquitoes.
• The history of shipping, in which the voyage to Asia was too long for a viremic person to arrive while infectious, a barrier that air travel has removed.

These explanations are not mutually exclusive. Modern aviation and low population immunity to flaviviruses in some cities mean that the risk of introduction cannot be excluded, and the WHO requires proof of vaccination from travelers from endemic countries to some countries under the International Health Regulations.

Climate and environment

Mosquito and viral traits respond nonlinearly to temperature. Models of transmission by Aedes aegypti find a unimodal dependence, with a lower limit at about 18 °C, an optimum of about 29 °C, and an upper limit near 35 °C. Canopy mosquitoes in South America live in shaded, humid conditions with lower temperatures. Climate affects yellow fever through temperature, rainfall, and the water storage practices that support Aedes aegypti, and through the movement of primates and vectors. Deforestation and fragmentation of forests increase contact between people, primates, and mosquitoes. Modeling studies project that climate change could increase the burden in parts of Africa, with the size of the change depending on the scenario and on vaccination coverage. Other factors, such as population growth, urbanization, travel, and vaccination coverage, also change, and they are often more important in the short term.

Stages of disease

The incubation period is 3 to 6 days. A large proportion of infections, from about half to more than 80 percent depending on the setting and the case definition, are asymptomatic or cause only mild illness. In symptomatic disease, the stages are:

1. Period of infection (days 1 to 3 or 4). Fever begins suddenly, with chills, headache, muscle and back pain, nausea, and vomiting. The patient's face is flushed, and the conjunctiva are red. Faget sign, a relatively slow pulse despite high fever, may be seen on the second or third day. Viremia is highest in this phase, and patients can infect mosquitoes. White blood cell counts fall, and transaminases begin to rise.
2. Period of remission (up to 48 hours). Fever falls and symptoms improve. In most patients, this is the start of recovery.
3. Period of intoxication (toxic phase). About 15 percent of patients who develop symptoms go into a toxic phase, in which fever returns and signs of liver, kidney, and coagulation failure appear. They include jaundice, abdominal pain with vomiting, bleeding from the gums, nose, eyes, and stomach (with "black vomit" or coffee-ground hematemesis), reduced urine output, and, in severe cases, delirium, seizures, coma, and shock. Aminotransferases can rise to thousands of units per liter, and the aspartate aminotransferase is usually higher than the alanine aminotransferase. Bilirubin, creatinine, and prothrombin time are raised, platelets fall, and hypoglycemia can occur. Death usually occurs between days 7 and 10. The case fatality among patients in the toxic phase is about 20 to 50 percent, with the higher values in settings with less intensive care. Survivors usually recover completely, and the liver heals without cirrhosis, although fatigue can last weeks to months.

Differential diagnosis

Mild yellow fever resembles many febrile illnesses, such as dengue fever, malaria, and influenza. Severe yellow fever must be distinguished from viral hepatitis (including hepatitis E), leptospirosis, malaria, other viral hemorrhagic fevers (Lassa fever, Crimean-Congo hemorrhagic fever, Ebola virus disease, and Rift Valley fever), rickettsial infections, and drug-induced liver injury. Epidemiological data on residence, vaccination, and exposure are important.

After mosquito inoculation, the virus replicates in local dendritic cells and in lymph nodes and then spreads to the liver, spleen, kidney, heart, and other organs. Hepatocytes, Kupffer cells, and endothelial cells are infected. Severe disease is a result of direct injury by the virus, of an intense host inflammatory response that includes high levels of cytokines such as IL-6, TNF-α, and IL-10, of changes in the endothelium that lead to vascular leakage, and of failure of the liver to produce clotting factors, which leads to bleeding. Early, strong type I interferon responses are protective, and patients with severe disease often have high viral loads and delayed immune responses.

The liver in fatal disease has characteristic changes. Cell death is concentrated in the middle zone of the hepatic lobule (zone 2 of Rappaport), with preservation of cells around the portal tracts and the central veins. Hepatocytes with shrunken, eosinophilic cytoplasm and fragmented nuclei, called Councilman bodies after the pathologist who described them in 1890, represent apoptotic cells. Other features include microvesicular fat, the presence of Torres bodies (eosinophilic intranuclear inclusions), and relatively little inflammation compared with the amount of cell death. The preservation of the reticulin framework of the lobule is one reason that survivors can regenerate the liver. These features are characteristic but are not unique, and diagnosis by histology is now supported by immunohistochemistry and by RT-PCR. Other changes include acute tubular injury in the kidney, myocardial injury, hemorrhage, and edema of the brain in severe cases.

Diagnosis depends on laboratory tests, because the clinical signs are not specific.

• RT-PCR of blood in the first week of illness is the main method for confirming acute infection. Viral RNA can sometimes be detected for longer in whole blood and in urine. In recently vaccinated people, the vaccine virus can be present at low levels in blood for about a week, so assays or sequencing that distinguish wild-type and vaccine virus are used in the investigation of suspected cases after vaccination.
• Serology. IgM antibodies appear from about day 5 to 7 of illness and are measured by an IgM antibody-capture ELISA. They cross-react with antibodies against other orthoflaviviruses, including dengue and Zika viruses, and with vaccine-induced antibodies, so a positive result is confirmed with a plaque reduction neutralization test (PRNT) that compares neutralizing titers against yellow fever virus and related viruses. A fourfold rise in titer in paired samples is evidence of infection.
• Antigen detection. Antigen detection assays for NS1 have been developed. Their use is less standardized than that of RT-PCR and serology.
• Histopathology and immunohistochemistry of liver tissue obtained after death, by autopsy or by minimally invasive viscerotomy, confirm fatal cases. Liver biopsy in living patients with suspected yellow fever is contraindicated because of the risk of bleeding.

No antiviral drug has been shown to be effective, and management is supportive. Patients with suspected severe disease should be treated in hospital, with attention to fluids, glucose, oxygen, and monitoring of kidney and liver function. In the toxic phase, intensive care may include correction of hypoglycemia, treatment of bleeding with blood products (plasma, cryoprecipitate, platelets, and vitamin K), renal replacement therapy for kidney failure, treatment of infections, and prophylaxis against gastric bleeding. Aspirin and other non-steroidal anti-inflammatory drugs are avoided because of the risk of bleeding and kidney injury, and the choice of an analgesic in liver failure (paracetamol is used with care or avoided) depends on local guidelines. Patients should be protected from mosquito bites during the first days of illness so that they do not infect mosquitoes.

Experimental approaches include the following:

• Emergency liver transplantation for fulminant liver failure has been reported in case reports and small series during the outbreaks in Brazil, with some survivors. The evidence is limited, and the approach is feasible only in specialized centers.
• Plasma exchange and other extracorporeal therapies have been used in a few patients, but there is no controlled evidence.
• Nucleoside analogs that inhibit the viral RNA polymerase, notably sofosbuvir (licensed for hepatitis C virus) and galidesivir, have shown activity against YFV in cell culture and in animal models. Their benefit in people has not been established in controlled trials.
• Neutralizing monoclonal antibodies are in preclinical and early clinical development.

Vaccination

All licensed vaccines contain the live attenuated 17D virus grown in embryonated chicken eggs. The three substrains in use are 17D-204 (used by several manufacturers, including Sanofi Pasteur), 17DD (Bio-Manguinhos/Fiocruz in Brazil), and 17D-213 (Chumakov Institute in Russia). The standard dose is 0.5 mL by subcutaneous or intramuscular injection, with a minimum potency of 1,000 international units. The vaccine induces neutralizing antibodies in about 80 to 100 percent of people within 10 days and in nearly all people within 28 days. In 2013, the WHO Strategic Advisory Group of Experts concluded that a single dose gives long-lasting, probably lifelong protection for most people, so a booster dose is not required, and the International Health Regulations were amended on this basis in 2016, so that the certificate is valid for life, beginning 10 days after vaccination. Boosters or repeated doses may still be considered in some groups, such as people who were vaccinated when pregnant, very young, or immunocompromised, and travelers at high risk. Studies of memory T cells that persist for decades after vaccination support long-lasting immunity.

Fractional dosing. During outbreaks, when the supply of vaccine is limited, one-fifth of a standard dose (about 0.1 mL, provided it contains at least 1,000 international units) has been used. Trials found that it gives high rates of seroconversion at 28 days, with antibodies persisting for at least several years in follow-up studies. This approach extends the supply fivefold, so that a stockpile of 6 million doses is equivalent to about 30 million fractional doses. The WHO recommends it as an emergency measure, and not as a replacement for the standard dose in routine immunization, in children under 2 years, in pregnant women, or in people with HIV. A fractional dose does not currently qualify for the international certificate.

Contraindications and precautions. The vaccine is contraindicated in infants younger than 6 months, in people with severe allergy to egg proteins or other components, in people with severe immunodeficiency (such as symptomatic HIV infection with low CD4⁺ counts, transplant recipients, and people receiving strong immunosuppressive drugs), and in people with a history of thymus disease (including thymoma and thymectomy). Vaccination is not recommended at 6 to 8 months of age unless the risk during an epidemic is high. In pregnancy, in breastfeeding, and in older people (age 60 and over), the risks and benefits are considered, and the vaccine is given if exposure is likely.

Adverse events. Most reactions are mild (headache, muscle pain, low-grade fever, and local reactions). Rare serious events are:

Both neurologic and viscerotropic disease occur almost exclusively after the first dose. The vaccine virus in YEL-AVD has generally not mutated to the virulent form. Host factors are responsible in at least some cases, and they include genetic defects in type I interferon signaling and autoantibodies against type I interferons (found in some patients, and often in people with thymoma and in older adults), but the cause in many patients is unknown. Inactivated and other non-replicating vaccine candidates are in development to give protection to people who cannot receive the live vaccine.

The EYE strategy

In 2017, the WHO, UNICEF, and Gavi launched the strategy to Eliminate Yellow Fever Epidemics (EYE), for 2017 to 2026, which covers 40 countries at high risk (27 in Africa and 13 in the Americas), and about a billion people. It has three objectives:

1. Protect at-risk populations, by including the vaccine in routine immunization at 9 to 12 months of age and by preventive mass vaccination campaigns in high-risk areas, with a target coverage of at least 80 percent.
2. Prevent international spread, by the application of the International Health Regulations (the international certificate of vaccination, and vaccination of travelers), and by the maintenance of an emergency stockpile of about 6 million doses managed by the International Coordinating Group on Vaccine Provision.
3. Contain outbreaks rapidly, by early detection (with surveillance of human cases and of monkey deaths, and with regional laboratory networks), reactive vaccination, and vector control.

The strategy has led to preventive campaigns in several countries, including Nigeria, and to an increase in routine coverage. Progress is uneven, and gaps in coverage and the growth of cities and forest-edge populations still leave large groups unprotected. The reports on the strategy should be consulted for current figures.

Vector control and personal protection

Vector control is a complement to vaccination, and it is used mainly to stop urban transmission. It includes removal or covering of water containers and water storage in cities, larvicides (such as Bacillus thuringiensis israelensis and pyriproxyfen), and, during urban outbreaks, adulticides in space spraying or indoor residual spraying. Resistance to insecticides, especially pyrethroids, limits some measures. Release of Aedes aegypti carrying the bacterium Wolbachia is used at a large scale in dengue control. Laboratory studies show that Wolbachia can reduce infection of Aedes aegypti by yellow fever virus, but field evidence for yellow fever is lacking. Personal protection includes repellents, long clothing, and screening, and viremic patients should be kept under bed nets. In the forest, vector control is not practical, and vaccination is the main protection.

Areas of research include the genetic and immunological basis for severe disease and for vaccine adverse events, the correlates of protection and the duration of immunity after fractional dosing and in children, the development of non-replicating vaccines, antivirals and antibody therapy, the ecology of the sylvatic cycle and its spread to new regions, the competence of Aedes albopictus and Asian mosquitoes, and models that combine climate, vector, and human data to anticipate outbreaks and to guide vaccination.

• Dengue fever
• Zika fever
• West Nile fever
• Arbovirus
• Viral hemorrhagic fever
• Aedes aegypti
• Yellow fever vaccine

How is the live-attenuated 17D yellow fever vaccine strain attenuated compared to the virulent Asibi strain?

Derived by Max Theiler through 176 serial passages of the wild-type Asibi strain in murine and chick embryo tissues lacking neural tissue, the 17D vaccine strain carries 68 nucleotide and 32 amino acid substitutions across its genome, including 12 in the envelope (E) glycoprotein. Substitutions such as E-T380R increase binding affinity for cell-surface heparan sulfate, restricting systemic hepatotropic dissemination while triggering robust Toll-like receptor 2, 7, 8, and 9 activation in dendritic cells and eliminating oral infectivity for Aedes aegypti mosquitoes.

Why does yellow fever cause midzonal liver necrosis (Councilman bodies) that heals without cirrhosis in survivors?

Yellow fever virus infects hepatocytes and Kupffer cells, inducing extensive apoptosis concentrated in the intermediate midzone (Rappaport zone 2) of the hepatic lobule, visible histologically as eosinophilic apoptotic Councilman bodies and microvesicular steatosis with minimal inflammatory cell infiltration. Because the underlying extracellular reticulin stromal scaffold and the periportal (zone 1) and centrilobular (zone 3) hepatocyte rims remain structurally intact, surviving patients regenerate normal hepatic architecture without fibrotic scarring or cirrhosis.

What causes rare yellow fever vaccine-associated viscerotropic disease (YEL-AVD)?

Yellow fever vaccine-associated viscerotropic disease (YEL-AVD, occurring in roughly 0.3–0.4 per 100,000 primary doses) is caused by unchecked systemic replication of the genetically stable 17D vaccine virus rather than viral reversion to wild-type virulence. Genomic and immunological investigations demonstrate that many YEL-AVD cases result from host defects in innate antiviral immunity—specifically inherited loss-of-function mutations in IFNAR1 or IFNAR2 or pre-existing neutralizing autoantibodies against type I interferons (IFN-α and IFN-ω), which are enriched in older adults and patients with thymoma.