Figure 8 Lab Worker, Winnipeg's Virology Lab

FILOVIRUS HEMORRHAGIC FEVER

(MARBURG AND EBOLA VIRUSES)

James A. Wilkerson M.D.

  The filoviruses are a group of unique organisms that produce a devastating hemorrhagic fever in human and nonhuman primates. The mortality from infection by several of these viruses appears to be exceeded only by rabies, human immunodeficiency virus (HIV), and possibly by prion infections. Although the number of persons infected and the number of fatalities are small compared to other viral infections, the high mortality rates have aroused widespread alarm, and have resulted in at least one movie and two books.

Two genera, Marburg and Ebola viruses, and four Ebola species have been identified:

  Ebola Zaire

  Ebola Sudan

  Ebola Tai (Ivory Coast)

  Ebola Reston.

MARBURG VIRUS, 1967

An epidemic in Athens, Greece, from 430 to 425 BC killed an estimated 300,000 individuals. Some writers have speculated that this disaster resulted from an Ebola viral infection imported from Ethiopia. Green monkeys, presumably the reservoir for the infection, have been found in Athenian murals from this time.

However, the first documented filovirus infection appeared in Europe in 1967. In that year a shipment of African green monkeys (Vevrets) were sent to laboratories in Marburg and Frankfurt, Germany, and Belgrade, Yugoslavia. Twenty-five technologists and investigators working in laboratories in those cities developed hemorrhagic fever: twenty in Marburg, four in Frankfurt, and one in Belgrade. All twenty-five had been directly exposed to African green monkey blood or tissue as the result of performing autopsies, preparing tissue cultures, or similar activities. Seven of these individuals (28 percent) died as a result of their infection. Six secondary infections occurred. Five were in medical staff that had cared for the individuals with primary infections. The sixth was transmitted by a recovering husband to his wife in semen. None of the patients with secondary infections died.

Until 1998 only six Marburg virus infections were subsequently detected. In 1975 a traveler in Zimbabwe developed Marburg hemorrhagic fever. His travelling companion became secondarily infected. A nurse in a Johannesburg hospital where the travelling companion was hospitalized also became infected. Only the original patient died; his travelling companion and the nurse survived.

The next infection occurred in 1980. The victim had been travelling near Mount Egan in western Kenya. Secondary infection developed in his attending physician in Nairobi. The original patient died, but the physician with the secondary infection survived. The third outbreak involved only a single individual. He also had been travelling near Mount Egan in eastern Uganda. No secondary infections followed his lethal infection.

The area around Mount Egan (western Kenya and eastern Uganda) was the origin for the African green monkeys responsible for the original Marburg outbreak. This area, known at the Lake Kyaga region, was the site of extensive efforts to find a source for the virus. Attempts to culture the virus, including trapping animals, birds, and bats, and from a cave on Mount Egan that both victims had visited — and placing trapped animals in the cave — were totally unsuccessful.

Of these thirty-seven infections, twenty-eight were primary. Ten of these were lethal, a mortality rate of 36 percent. Nine individuals developed secondary infections but none died, a feature that has not been observed with other Ebola epidemics. Overall mortality (ten of thirty-seven patients) was 27 percent.

Within the last two years a number of hemorrhagic fever infections has occurred in the Democratic Republic of the Congo (formerly Zaire). Seventy-two patients are thought to have been infected, but the number of patients and the number of deaths are uncertain. Five deaths have been determined to be the result of Marburg virus infection. However, no evidence of filovirus infection could be found in five additional fatal infections, which suggests this outbreak may have multiple etiologies

.

EBOLA ZAIRE, 1976

In 1976 the Ebola virus first appeared in simultaneous outbreaks in Yambuku, Zaire, and in Nzara, Sudan. The Ebola Zaire epidemic started in the Yambuku Catholic Mission Hospital, which was founded in 1935 by Belgian missionaries. This hospital had one hundred twenty beds but a total staff of only seventeen that included a Zairian "medical assistant" and three Belgian nurse/midwives. No full-time physician was on the staff.

The hospital was the area's major medical facility because it had a large supply of drugs. It had a number of rather large outpatient clinics, particularly obstetrics and neonatal clinics, and is reported to have had 6,000 to 12,000 outpatient clinic visits annually (24 to 48 patients daily for a 250 workday year or 20 to 40 patients daily for a 300 workday year.)

The hospital policy was to administer nearly all medications by injection. Presumably this had more of an emotional impact and would constrain patients to follow instructions. However, the hospital could not afford sterile needles and syringes for each injection, and at this time, before the epidemic of HIV, sterile syringes and needles were not considered essential. Each morning five syringes and needles were issued for the outpatient clinics. The needles and syringes were rinsed between injections but were not sterilized.

The index patient for the Yambuku outbreak has never been clearly identified. A thirty-year-old male with "dysentery and epistaxis" was hospitalized for two days on August 28th. Subsequently, investigators who tried to identify this man in the village he had claimed to be his home were unable to locate him or anyone who knew him.

The first clearly identified patient was a forty-four-year-old male teacher who received an outpatient Chloroquine injection for malaria on August 26th. Apparently, this teacher did have malaria at the time of those injections and his infestation was controlled by the Chloroquine injection. However, his fever recurred and he returned to the hospital on September 1st with a fever of 39.2 o C. He came back to the hospital with gastrointestinal bleeding on September 5th and died on September 8th.

At least nine other patients with infections in the first week of September were identified. All had been treated at Yambuku Mission Hospital outpatient clinics. Most returned to their villages after they had received their injections at the hospital.

Approximately 250,000 people live in the area around Yambuku, most in villages of less than 500 persons. During the outbreak, infected individuals were identified in fifty-five villages.

Of 288 individuals whose source of infection could be determined, eighty-five (30 percent) received injections at Yambuku Mission Hospital, and 149 (52 percent) were infected secondarily by contact with patients who did receive injections. Forty-three patients (15 percent) received injections, and also had contact with infected individuals.

The secondary infection rate was high, 27.3 percent in close relatives (spouses, parents and older children). However, secondary infections occurred in only 8.0 percent of other individuals who resided in the house, ate their meals there, but were not directly involved in caring for the infected patients. In one instance, a primary infection was transmitted serially through four succeeding attendants (five generations of infection.) Seventeen percent of male patient contacts became infected in contrast with 36 percent of female patient contacts, which has led to speculation that the infection was sometimes spread by sexual intercourse with female patients.

Isolation ended the epidemic. The major event appears to have been the closure of Yambuku Mission Hospital on the 30th of September. At that time eleven of the seventeen staff had died of infection. The three nurse midwives had also been infected but had been transferred to a hospital in Kinshasa. All three of these individuals also died, and the entire staff of the ward on which these patients were hospitalized was quarantined to that ward for three weeks after the last death. However, no secondary infections appeared among the staff.

Even though the first case appeared on September 1st, and Yambuku Mission Hospital was closed on September 30th, an international commission was not formed until October 18th. Members of that commission were in Yambuku the next day, October 19th, but the last Ebola death occurred on November 5th, only sixteen days after their arrival. As a result, most of the commission's efforts were attempts to learn from survivors what had happened to infected patients.

Essentially no patient laboratory studies were carried out. The only laboratory procedure the Yambuku medical hospital was capable of performing was a urine protein. Individuals with symptoms suggestive of viral hemorrhagic fever who had protein in their urine were considered to be infected. Only five autopsies and three postmortem liver biopsies were performed. Viral isolation studies were carried out on only eight patients, although the virus was successfully recovered from the blood of all these patients.

Two hundred eighty of 318 infected persons died, a mortality rate of 88 percent. All eleven babies liveborn from infected mothers died, but the cause of their deaths remains undetermined. These babies did not present the typical hemorrhagic signs of Ebola hemorrhagic fever and could well have died from many of the other causes of infant mortality prevalent in that area. Of seventy-three pregnant females with lethal infections, eighteen (25 percent) aborted. Of nine pregnant females with nonlethal infections, only one (11 percent) aborted.

The serologic studies performed by the international commission were all indirect fluorescent antibody assays. Members of the commission questioned this procedures reliability at that time, and it is now largely discredited because it is associated with such a high percentage of false positives.

All of the patients infected as a result of Yambuku Mission Hospital injections died. Observers at the site thought that the secondary cases appeared to be less severe, but the statistics did not support that impression.

EBOLA SUDAN, 1976

A simultaneous Ebola outbreak occurred in the village of Nzara in Southern Sudan. The index patient was traced to a specific bat-infested room in a Nzara cotton-weaving factory. This patient was hospitalized in a local hospital, which lead to spread of the infection to other patients and to family members who provided care in the hospital.

The Nzara hospital was a hospital in name only. It consisted of two buildings with two beds each, and a third building with three open wards of eight beds. The roof was thatch and no running water or toilet facilities existed. Masks and gowns were in limited supply. No routine barrier nursing procedures were practiced. Needles and syringes often were washed but not sterilized.

One patient from Nzara was transferred to a large hospital in Maridi, which resulted in the spread of the epidemic to that village and its larger population. Ultimately, ninety-three patients developed infections in the Maridi hospital and 155 infections occurred in the community. Of 284 infected patients in Nzara and Maridi, 151 died (53 percent). A total of forty-one hospital staff deaths occurred in both towns.

Extensive ecologic studies were carried out. The animals investigated included bugs and bats in the Nzara cotton mill. No origin for the virus could be identified.

The causative virus was named Ebola Sudan. In subsequent studies it has been found to be distinctly different from the Ebola Zaire virus.

NZARA, SUDAN, 1979

A second outbreak of the same virus occurred in Nzara in 1979. The index patient was a forty-five-year-old man who was hospitalized on the 2nd day of August with a history of fever of three days duration and vomiting and diarrhea of more recent onset. He developed gastrointestinal bleeding and died in the hospital on the 5th of August. However, this man's illness was not diagnosed as viral hemorrhagic fever. The diagnosis was established only after the deaths at home of three family members who had cared for him.

After realizing the nature of the outbreak, the World Health Organization and other agencies were notified. An international commission with forty-five members was set up to investigate an outbreak that ultimately produced only thirty-four infections and twenty-two deaths.

The outbreak involved essentially only four families. The first family consisted of the index patient and three family members, who cared for him, contracted the infection and also died. The initial patient in the second family occupied the bed next to the index patient in the Nzara hospital. Nineteen of his family members became infected; thirteen of these twenty patients died.

The initial case in the third family was a woman who was caring for her husband, who had been hospitalized following surgery on the same ward as the index patient. That woman survived but two family members caring for her at home developed lethal infections. The initial patient in the fourth family was a nurse caring for members of the second family. She died, but only one of five family members who subsequently became infected died.

The fifth patient also was a nurse caring for members of the second family. None of her family members developed infections.

The infection passed through seven generations. The mortality rate was 89 percent in the first four generations, but only 38 percent in the last three. However, attenuation of the infection is not certain because less severe infections may have been more readily recognized during the later stages of the outbreak.

Subsequently an Ebola Sudan infection occurred in Great Britain in 1977 in a laboratory worker who stuck himself with a needle. He survived his infection, but undoubtedly received more sophisticated medical care than patients in Nzara. No subsequent Ebola Sudan infections have been identified. Of 319 individuals infected by this virus, 173 died, an overall mortality of 54 percent.

EBOLA RESTON, 1989

In 1989 a number of crab-eating Macaques imported from the Philippine Islands died while being held in a primate quarantine facility in Reston, Virginia. (Imported primates must be quarantined for thirty-one days before being released to laboratories. The Reston importer routinely held his animals for forty-five days, although the animals infected during the Ebola outbreak all became ill during the initial thirty days.) The infection was first thought to be the result of simian hemorrhagic fever. That virus was found in some of the animals, but a filovirus also was found.

Discovery of the filovirus was alarming because many investigators and animal handlers had been exposed to the virus. One investigator reportedly suspected the infection was caused by Pseudomonas, opened a culture tube, and smelled it in an effort to detect the unique odor of that organism. One of the animal handlers cut his finger performing an autopsy on one of the animals.

Ultimately, all of the animals were destroyed, but infections subsequently developed in new shipments from the same exporter in the Philippine islands. The infections were thought to have resulted from the persistence of the virus in the Reston facility, in spite of rigorous disinfection, but subsequent findings have indicated that the newly imported animals carried the same virus. Animals from the same exporter in quarantine facilities in Texas and Pennsylvania also developed infections.

Investigation of the Philippine facility disclosed that 52.8 percent of the monkeys dying over a three-month period in that facility had filovirus antibodies. The exporter reportedly has made a conscientious effort to eliminate the infection. Shortly after the Reston outbreak all of the Macaques in his facility were eliminated and the facility was decontaminated. However, the infections recurred when the facility was restocked. Whether the infection is endemic in the monkeys in that area, or a reservoir exists in the immediate vicinity of the facility, has not been determined. In 1992, filovirus infections were found in Macaques shipped to a research facility in Italy by the same exporter. In 1996, infectious were found in animals from this same exporter in a quarantine facility in Alice, Texas.

In the 1996 incident in Alice, Texas, 100 Macaques were imported. The animals were divided into two groups of fifty, which were completely separated. One animal in one group became ill and died; a second animal became ill and was euthanized. Subsequently, the entire group of fifty animals was euthanized. No infections developed in the other fifty animals.

Ebola Reston, as this virus eventually was named because most investigations were performed on isolates from animals at Reston, was the first Asian filovirus to be identified. Contact with African monkeys could not be ruled out for monkeys in the earlier shipments, which had been routed through the Netherlands, but subsequent shipments had no opportunity for contact with monkeys from Africa.

The organism appears to be nonpathogenic for humans although it is quite similar to other Ebola viruses. Four humans, including the individual that cut his finger performing an autopsy, developed antibodies. None had any symptoms of clinical illness. Subsequently, other animal handlers with antibodies have been found, but many of these studies have been performed with assay techniques that yield a high percentage of false positives.

Even though the infection is not pathogenic for humans, it is devastating in Macaques. Few survivors with antibodies to the virus have been found. The infection clearly spread across the room in which animals were caged at Reston. The spread appeared to be by aerosols, but whether the aerosols were pulmonary in origin or were aerosolized urine and feces were not determined. In subsequent experimental studies, artificial aerosols have demonstrated that monkeys are susceptible to being infected by the pulmonary route, and infected animals have viruses in the alveolar spaces of their lungs.

These findings have lead to concern that human infections could be spread as aerosols and to recommendations for vigorous respiratory isolation of hospitalized patients. However, no spread of infection by aerosols in humans has ever been documented.

EBOLA TAI (IVORY COAST), 1994

In 1994, hemorrhagic fever developed in a veterinarian who autopsied a chimpanzee. This person was promptly evacuated to a hospital in Switzerland where she was vigorously treated and subsequently recovered. The virus isolated from this patient was found to be a unique Ebola virus that has subsequently been named Ebola Tai because the outbreak occurred in the Tai National Forest in Ivory Coast.

The outbreak occurred in a unique primate behavioral laboratory. The animals in this facility have been studied for many years and many new observations about primate behavior have resulted. However, twelve of forty chimpanzees died during the outbreak, which greatly impaired the investigations. In retrospect an undiagnosed die-off of eight animals two years earlier was recognized, indicating that the laboratory had lost almost 40 percent of its animals in two years.

The infection again was devastating for the primates. Apparently every infected chimpanzee died; none with antibodies could be found. This infection was the second instance of a filovirus being transmitted from monkeys to humans.

The outbreak occurred at the end of the rainy season, a timing that has been considered indicative that the virus is vector maintained. The vector is presumed to dwell in the canopy of the rain forest. (The Tai National Forest is one of the last primitive rain forests in existence.) On-going efforts to identify a vector and a reservoir by investigative teams from a number of countries have been totally unsuccessful. When insects were injected with the virus to try to identify a vector, the virus failed to grow.

A second Ebola Tai infection in a 25 year old man in 1995 also was nonfatal, which suggests that the Ebola Tai may be less virulent than other Ebola viruses pathogenic to humans.

MORE EBOLA ZAIRE

Mekoukou, Gabon, 1994

In 1994 an outbreak in Mekoukou, Gabon resulted in forty-nine infections and twenty-nine deaths, a 59 percent mortality rate. This outbreak was originally diagnosed as yellow fever and was not carefully investigated. However, the high mortality rate, as well as other aspects of the outbreak, indicate it was a filovirus infection, and filovirus organisms have been isolated from some infected persons.

Kikwit, Zaire, 1995

On April 10th, 1995, a laboratory technologist entered a hospital at Kikwit, Zaire with fever and bloody diarrhea. Ten days later fourteen members of the hospital staff who had participated in his care were hospitalized with similar symptoms.

Although the medical technologist was originally considered the index patient, the outbreak was subsequently traced to a charcoal maker who probably was infected in the rain forest. He became ill on January 6, 1995 and the infection was maintained among caregivers until the medical technologist was hospitalized, resulting in a widespread outbreak.

The city of Kikwit has a population of 400,000 and is near other big cities. In such a densely populated area the potential for a devastating epidemic is great. Surprisingly the medical care available However, more sophisticated medical care was also available.

The cause of the outbreak was not identified for identified until approximately May 8th. When hemorrhagic fever was suspected, blood specimens from fourteen patient were sent to Centers for Disease Control and Prevention (CDC) in Atlanta. Nine hours after the specimens were received the presence of the Ebola antigen or antibodies against the Ebola virus had been confirmed in thirteen. Within forty-eight hours DNA studies had established that 528 base pair segments of the genome from the virus from four patients differed from the original Ebola Zaire by only four base pairs, a variance of less then 1 percent. The segments from the four patients did not differ at all, indicating that all four had been infected with the same virus. When the entire glycoprotein genome was analyzed, the variance from the virus recovered in the 1976 outbreak was only 1.6 percent.

The epidemic was controlled by three measures. Caregivers were trained in the use of protective (barrier) devices, principally gowns, gloves, and masks, which effectively interrupted secondary spread. (In the hospital where most of the infected individuals were treated, rubber gloves were not routinely used for surgery.) An aggressive case identification program was initiated so that all caregivers needing protection could be identified. Finally, an educational program was initiated, which is particularly challenging in a community with essentially no television, few radios, and high illiteracy rates.

Natives of Zaire traditionally care for their dead by carefully washing the body in preparation for the funeral. In order to prevent spread of infection during this procedure, the mayor of Kikwit instituted an ordinance that made such traditional care illegal. (Six months after the outbreak had ended, traditional funereal practices were being resumed even though still illegal, which aroused concern about a recurrence of the outbreak.)

The last patient in the Kikwit outbreak was hospitalized on the 24th of June. A total of 315 infections were identified; 255 were lethal (a mortality rate of 81 percent). As in other outbreaks, medical staff suffered a high morbidity and mortality. Twenty-six percent of the patients were nurses or students. Other caregivers also had a high incidence of infection; 21 percent of the patients were housewives.

In an investigation of the routes of spread of the infection, twenty-seven individuals were considered to be primary infections. These twenty-seven patients were in households that had 173 other members. Twenty-eight of those individuals (16 percent) became secondarily infected. Ten of twenty-two spouses (45 percent) were infected. Only eighteen of the other 151 individuals (14 percent) were infected.

Twenty-four of eighty-one adults (30 percent) were infected. However, only four of ninety-two children (4 percent) were infected. Of the 315 infected patients, only twenty-seven (9 percent) were younger than eighteen, yet 50 percent of the population of Zaire is younger than sixteen. An investigator of this phenomenon concluded that the children were shielded from infection; they did not have greater resistance.

Mayibout, Gabon, 1996

In February 1996 in Mayibout, Gabon a group of villagers killed a chimpanzee, which was subsequently cooked and eaten. (Some newspapers incorrectly reported that the villagers found a dead chimpanzee. The animal was ill, but was not dead when first encountered.) Nineteen persons developed hemorrhagic fever. (Some reports have stated that twenty people were infected.) Thirteen died. Ultimately thirty-one infections developed and twenty-one individuals died (68 percent). The original cases were all in individuals who helped kill, clean and prepare the chimpanzee for cooking. No infections resulted from eating the cooked meat. The later cases were secondary in persons caring for the primary patients. The virus infecting individuals in Gabon has been determined to be similar but not identical to Ebola Zaire virus.

Booué, Gabon, 1996

Also in 1996, the index patient for this outbreak was a hunter who lived in a forest camp. A dead chimpanzee found in the forest at that time was determined to be infected. This outbreak persisted for at least a year. Although it started in Booué, it was spread to Libreville when infected individuals were transported there. Sixty individuals were infected and forty-five died.

Johannesburg, South Africa, 1996

A medical professional that had been treating Ebola virus-infected individuals in Gabon in 1996 came down with the infection after traveling to Johannesburg. He survived, but a nurse who took care of him became infected and died.

Gabon and Republic of the Congo, 2001–2003

From October 2001 to December 2003, several Ebola hemorrhagic fever outbreaks of the Zaïre subtype, were reported in Gabon and the Republic of Congo: Mékambo-Mbomo-Kéllé in 2001-2002, Kéllé-Mbomo in 2003 and Mbandza-Mbomo in 2003. A total of 302 infections and 254 deaths, a mortality rate of 84 percent. CDC participated with the Gabonese and Congolese Ministries of Health, the WHO, the International Center for Medical Research in Franceville, Gabon, and other partners in an international response to these outbreaks. The existence of a state of war in these areas during much of this time increased the difficulties medical care teams encountered.

MORE EBOLA SUDAN

Gulu District, Uganda, 2000

On 8 October, 2000, an outbreak of a severe febrile illness characterized by gastroenteritis. headache, conjunctivitis, and occasional hemorrhagic phenomena was reported in Gulu district, Uganda. On 10 October an isolation ward was established, and the National Institute of virology, Johannesburg, South Africa, confirmed the nature of the infection on 15 October.

Detailed analysis indicated the infecting organism was the same virus responsible for outbreaks in Sudan. The sequences differed by only 3.3 percent of 362 polymerase nucleotides and 4.2 percent of 146 nucleocapsid nucleotides.

An active surveillance system was established during the third week of October to determine the extent of the outbreak, identify foci of disease activity, and detect cases early. The system was designed to be very sensitive and detect all individuals who might have Ebola hemorrhagic fever. Ill persons were encouraged to be assessed at a hospital and to be hospitalized, if indicated, to reduce transmission of the infection.

The most common signs and symptoms on admission during the early weeks of the outbreak were diarrhea (66 percent), asthenia (64 percent), headache (63 percent), nausea and vomiting (60 percent), abdominal pain (55 percent), and chest pain (48 percent). Bleeding was seen in only 20 percent of the individuals and primarily involved the gastrointestinal tract. Fatalities usually developed a rapid progression of shock, increasing coagulopathy, and loss of consciousness. Spontaneous abortions occurred among pregnant women with Ebola infections.

In retrospect, the earliest presumptive infection occurred on 30 August 2000, almost six weeks before the outbreak was recognized. However, the earliest cases remained obscure, which limited tracking of the initial cases and investigation of the possible reservoir for the virus.

Three principal sources for spread of the infection were identified: physical contact with victims of the infection at funerals, contact with infected individuals by multiple caregivers in their homes, and nosocomial spread in hospitals. Even after barrier nursing procedures were instituted, twenty-two health care workers became infected, indicating the need for more intensive education and supervision. Barrier nursing procedures are considered the ultimate manner in which the outbreak was stopped.

The last case was on 9 January 2001. A total of 429 cases were identified, approximately two-thirds in women, who were usually the caregivers. As in previous outbreaks, many infections occurred in health workers. Fatalities numbered 173 for a fatality rate of only 40 percent.

Yambio County, Sudan, 2004

In 2004 the World Health Organization reported twenty infection and five deaths from from Ebola-Sudan in Yambio County in southern Sudan. The infections were laboratory confirmed by both the Centers for Disease Control and Prevention (CDC) and the Kenya Medical Research Institute.

MORE MARBURG

Durba, Democratic Republic of the Congo, 1998–2000

Between 1998 and 2000 154 Marburg virus infections were identified in the Democratic Republic of Congo. Most were in young male workers at a gold mine in Durba, in the northeastern part of the country. Infections were subsequently detected in the neighboring village of Watsa. Of those infected, 128 died, a mortality rate of 83 percent.

Uige Province, Angola, 2004

The most recent filovirus outbreak occurred in Angola. The outbreak is believed to have begun in Uige Province in October 2004. Most infections in other areas have been linked directly to the outbreak in Uige. This outbreak is the only time Marburg virus infections have been identified in Angola

In the early stages of the outbreak 75 percent of the infections were in children. (In other outbreaks children have been infrequently infected, even when living in a household where individuals with filovirus infections were receiving care.) As occurred at the Yambuku Hospital during the first outbreak of Ebola Zaire, injections with needles and syringes that had not been sterilized appears to have played a major role in spreading the infections.

The response to this outbreak by the world health community was immediate. Within twenty-four hours of confirmation of the presence of Marburg virus in nine of twelve patient samples by CDC on 21 March 2005, staff from the Global Outbreak Alert and Response team and the WHO's regional office in Brazzaville, Democratic Republic of Congo, were moving toward Angola. Ultimately more than seventy individuals were in the country.

The few measures needed to contain a filovirus outbreak are straightforward: rapid detection of infections and isolation of infected individuals; infection control in hospitals, mostly by instituting barrier nursing procedures, and preventing funeral and burial practices that include close contact with the bodies. In the Angola outbreak, fear has been one of the greatest impediments to control of spread of the infection. Mobile surveillance and public-awareness teams were forced to suspend operations on three occasions when they were stoned or threatened by hostile residents.

Because so few infected individuals survived, members of their communities associated admission to hospital isolation wards with certain death, and have preferred to provide care at home, often hiding infected individuals and later hiding their bodies. The deaths of sixteen doctors and nurses undermined the morale of the hospital staff. When outside medical staff arrived in Uige, they soon realized that building trust and securing community collaboration was one of their biggest challenges.

However, success was achieved. The last confirmed death occurred on 21 July 2005. A total of 374 infections were identified and 329 deaths occurred, and unusually high case fatality rate of 88 percent.

OVERALL MORTALITY

In the thirty-eight years since the original Marburg outbreak approximately 2,414 filovirus infections have been recognized, and 1,709 deaths have resulted. (WHO and CDC totals for different outbreaks are not always the same.) The overall case fatality rate has been 71 percent. The average number of infections per year is sixty-three and the number of deaths is forty-five.

The number of infections and fatalities clearly indicate that filovirus infection is not a major health problem. Deaths from yellow fever number approximately 30,000 a year. In South America, approximately 500 people die annually from yellow fever, yet yellow fever is considered "controlled" on that continent. Approximately 5,000 people die annually from Lassa hemorrhagic fever.

CFR = Case Fatality Rate

Mortality from infections by different filoviruses varies. The mortality for early Marburg virus infections was only 27 percent and almost no patients with secondary infections have died. Case fatality rates for the 1998 and 2004 outbreaks in Democratic Republic of Congo and Angola have been much higher. Mortality for Ebola Sudan has been 46 percent and many secondarily infected patients have died. Mortality for Ebola Zaire has been 80 percent and no definite difference between primary and secondarily infected patients has been demonstrated.

In further contrast none of the two patients infected by Ebola Tai have died, and Ebola Reston has been non-pathogenic for humans.

CLINICAL FEATURES

The diagnosis of filovirus infection is difficult, and, with only a few exceptions, the infection has only been recognized when patients occur in clusters. (Following the Kikwit outbreak investigators performed serological studies on a number of individuals using more reliable ELISA techniques. Scattered individuals with Ebola antibodies were found, indicating that a few sporadic infections are occurring.) The clinical features of all such infections are quite similar and to a considerable extent are nonspecific. The incubation period ranges from three to ten days. The median age is approximately thirty-seven years. In most outbreaks males have been infected as frequently as females.

The onset is typically abrupt and usually consists of severe headache and fever. The headache commonly radiates to the back of the neck and upper portion of the back. Pharyngitis and gastrointestinal symptoms, both vomiting and diarrhea, appear two to three days later. In an area where malaria and gastrointestinal infections such as Typhoid fever, other salmonella infections, and Shigella infections are rampant, such symptoms are not likely to arouse consideration of hemorrhagic fever.

Typical patients are prostrate. They are very weak, often dehydrated, and may suffer surprising weight loss for such a brief illness. The pharyngitis is often associated with dysphagia and has been described as feeling as if a mass were present in the throat. Exudates are not present and the symptoms may be the result of pharyngeal edema. On day five a rash typically appears. The rash has been described as maculopapular or as morbilliform. It is readily seen in Caucasian individuals, but is much more difficult to discern in darkly pigmented persons. The rash is considered highly significant, however, because similar skin changes are not seen in other hemorrhagic fevers. The rash usually lasts only forty-eight hours.

Another common but nonspecific feature is the presence of severe conjunctivitis.

With the onset of gastrointestinal symptoms many patients have abdominal pain. The pain can be quite severe and some writers have suspected that it is the result of pancreatitis. Too few autopsies have been performed to evaluate this possibility, and laboratory facilities for amylase or lipase determinations have not been available.

On the fifth day, hemorrhagic phenomena typically appear. The bleeding can be from the gastrointestinal tract, including the gums, the respiratory tract, or the uterus. Bleeding from needle puncture sites is typical. The onset of bleeding is considered to be an indicator of a poor outcome. However, the occurrence of hemorrhagic phenomena is inconsistent. In some reports it is stated to occur in as little as one-third of the patients.

Severely ill patients have a sustained high fever and become delirious and combative. Death is due to hypovolemic shock and adult respiratory distress syndrome (ARDS). For patients who survive, convalescence is slow, typically requiring five weeks or longer.

Few clinical laboratory studies have been performed. AST (SGOT) is elevated, usually to a greater degree than ALT (SGPT). Neutrophilia is common. Thrombocytopenia is inevitable. The hemorrhagic phenomena apparently result from platelet dysfunction as well as loss of platelets. Platelet aggregation is diminished one to three days before the appearance of shock. In laboratory studies plasma platelet factor IV have been elevated while platelet factor IV in platelets has been diminished. These alterations have been interpreted as indicating the platelets have been stimulated but cannot aggregate. The viruses proliferate vigorously in aortic endothelium, which reduces the production of prostacyclin required for normal platelet function.

Fibrin split products are increased and prothrombin time and partial thromboplastin time are prolonged. However, whether a true disseminated intravascular coagulation (DIC) is present has been debated. Some investigators have argued that the changes in clotting factors could be the result of extrinsic clotting pathway activation by necrotic tissue.

In the past serologic determinations have been unreliable. False positives have abounded with immunofluorescent assays (IFA) and have resulted in reports of large percentages of some populations with antibodies to the filoviruses. In one report, antibodies were reported in 30 percent of the Central African Republic population. Even backing up the IFA with Western blot determinations still yields false positive results. In one study, forty-two of 550 (7.6 percent) of the workers in animal quarantine facilities were found to have antibodies to filoviruses by both procedures. However, twelve of 449 (2.7 percent) control persons who had no contact with primates, were also positive.

An IgG ELISA antibody detection procedure has been recently developed at CDC and appears to be highly reliable. Individuals with antibody titers as high at 1:1000 by IFA, but no history suggesting they had ever come in contact with the filovirus were found to be negative using the IgG ELISA procedure.

Laboratory diagnosis of filovirus infection requires reverse transcriptase-PCR duplication of viral RNA followed by nucleotide base-pair analysis. Serology is problematic, probably not reliable. Electron microscopy is not reliable. Only Ebola Zaire produces cytopathic changes in tissue culture. (Immunocytochemical diagnosis of filovirus infection in skin biopsies is discussed below.)

In the few autopsies performed on humans and in autopsies on experimentally infected primates, one of the principal findings has been the presence of bleeding, particularly into the gastrointestinal tract and into the body cavities. Typically the liver contains foci of necrosis. Little inflammation is associated with the necrosis, but Councilman-like bodies are present. The necrosis usually is not sufficiently severe to cause organ failure. An unusual type of perifollicular necrosis is seen in lymph nodes and the spleen.

Laboratory studies should be limited to reduce the risks for technologists. Specimens should be transported to the laboratory in plastic bags. The specimens should be handled in biosafety level II cabinets following biosafety level III procedures. Serum should be pretreated with Triton-X 100, and autoanalyzers must be disinfected after carrying out procedures on blood or serum from filovirus infected patients.

Bodies should be covered with a leak proof material, should not be embalmed, and should be cremated or buried in a sealed casket. However, most sealed caskets begin to leak after a few years. The CDC should be consulted about procedures to be followed during surgery, obstetric procedures, or autopsies.

The viruses are inactivated by ultraviolet light and gamma irradiation. Gamma irradiation is the least disruptive for specimens that are to be analyzed. Common disinfectants such as hypochlorite, formalin, ß-propiolactone, phenol and lipid solvents such as ether also inactivate the viruses.

THERAPY

At the present time treatment for filovirus infections is basically supportive. Ribovarin and other antiviral agents have not proven beneficial. Interferon is also ineffective. Supportive care consists of administration of intravenous fluids, treatment of cerebral edema, care for renal failure and care for the complications of the DIC or DIC-like syndrome that these patients develop.

Recent studies have indicated the possibility of developing a more specific therapeutic agent. The glycoprotein on the surface of Ebola and other viruses binds to cells and mediates fusion between the viral envelope and the host cell membrane, enabling the virus to release its contents into the host cell cytoplasm. Conformational changes must take place in the viral glycoprotein spikes to allow it to fuse with the host cell membrane. A low pH and binding to receptor molecules were until recently the only know ways such conformational changes could be induced.

Recently proteolysis by cathepsin B and cathepsin L have been found to produce the conformational change in the surface glycoprotein of Ebola virus. Glycoprotein-mediated infection is substantially reduced in cells lacking these proteases. Treatment with the cathepsin B inhibitor CA074 blocks the conformational change, prevents fusion with the host cell membrane, and thereby prevents infection. CA074 is toxic to cells and is an unlikely candidate for use in clinical situations, but developing nontoxic cathepsin inhibitors that might safely block Ebola replication in humans could offer a means for controlling this deadly infection.

Recently highly specialized vaccines for Marburg and Ebola virus have been developed in an effort to protect care providers and laboratory workers, and possibly to provide protection against bioterrorist attacks. Several prototype vaccines to protect nonhuman primates against Marburg virus have shown promise.

Two vaccines that solidly protect monkeys against Ebola virus have been developed. One has a vesicular stomatitis virus base, and the other uses an adenovirus vector. A single injection of the adenovirus construct protects monkeys, and phase 1 trials of this vaccine have begun.

These potential vaccines require much additional development, and may not be suitable for immunization of large populations, although the economic feasibility and overall need for widespread immunization for such an uncommon infection remains to be determined.

PATHOGENESIS

Filoviruses are known to be spread by body fluids: blood, semen, emesis, and stool. In studies of specimens from the Kikwit epidemic, Zaki and his colleagues at CDC found far more viral organisms in skin than are seen with any other infections. Subsequent studies have suggested direct intact skin-to-intact skin transmission of infection, which may be the route many secondary infections are acquired. (Formalin fixed skin biopsies are safe, reliable specimens for immunocytochemical diagnosis of infections.)

As occurred so dramatically at Yambuku in the first outbreak of Ebola-Zaire, and Uige in Angola, administration of injections with unsterilized syringes and needles plays a major role in spreading the infection. Such practices, which include making multiple aspirations from a single drug container with unsterile needles, are widespread in almost all of Africa and in other developing nations. Hepatitis C is almost epidemic in some of these countries as a result, and a significant component of HIV infection in Africa is suspected to have been transmitted this way.

Once infection has occurred, filoviruses undergo rapid, lytic replication in many cells. Proliferation in macrophages, liver, and endothelium is probably most significant.

Immunosuppression appears to play a major role in progress of the infection. Production of type I interferons is suppressed by the VP35 protein, and interferon action is blocked by the VP24 protein. As a result natural killer cells, macrophages, and dendritic cells are not activated, and effective neutralizing antibodies are not formed.

THE VIRUSES

Filoviruses are single strand, negative sense RNA viruses. Other similar viruses are the paramyxoviruses (measles, respiratory syncitial virus, Newcastle disease among others) and rhabdoviruses, the most significant of which is rabies virus. The filoviruses are among the largest viruses known. They are 80 nm long and range up to 14,000 nm in length. The most infective organisms are 790 (Marburg) to 970 (Ebola) nm long. Individual organisms have a 50 nm helical nucleocapsid with a 20 nm axial space, a 5 nm periodicity, and a derived surface membrane with glycoprotein projections.

The viral genome codes for membrane proteins VP24, VP40, and GP (glycoprotein), NP and L proteins that carry out nucleic acid replication, and structural proteins VP30 and VP35. GP is the primary surface protein, and facilitates cell entry, possibly by binding with membrane receptors. GP is the primary target for humoral antibodies, and contains the twenty-six amino acid immunosuppressive motif.

Ebola and Marburg viruses are distinctly different and do not cross-react antigenically. They are now considered two different genera. The different Ebola viruses do cross-react weakly. The Ebola genome overlaps three times; the Marburg only once. Only Ebola produces immunosuppressive surface glycoprotein. The glycoprotein nucleotide sequence difference among Ebola viruses is 47 percent. Two or three gaps are present. The nucleotide sequence difference between Ebola and Marburg viruses is 72 percent, and six to eight gaps are present. Ebola glycoprotein has abundant terminal sialic acid on O- and N-linked glycans. Marburg has none.

ECOLOGY

Thorough ecologic investigations after every major outbreak have provided no conclusive information about reservoirs of infection or the method of spread from animals to humans. Filoviruses do appear to be zoonotic, but a wide range of trapped animals, birds, and insects have not carried the virus. The organism does not replicate when injected into many different insects. Infected chimpanzees and other primates have been found, and in a few outbreaks have been linked to infections in humans, but such animals are clearly another target species.

The virus can be replicated in some species of bats, and bats native to the areas where the virus is found may prove to be the viruses' carriers. The 1998 to 2000 outbreak in Durba, Democratic Republic of the Congo, involved mostly young male workers in a gold mine. Bats living in the mine may have played a role in transmitting the infection.