A Novel Immunohistochemical Assay for the Detection of Ebola Virus in Skin: Implications for Diagnosis, Spread, and Surveillance of Ebola Hemorrhagic Fever
- Sherif R. Zaki,
- Wun-Ju Shieh,
- Patricia W. Greer,
- Cynthia S. Goldsmith,
- Tara Ferebee,
- Jacques Katshitshi,
- F. Kweteminga Tshioko,
- Mpia A. Bwaka,
- Robert Swanepoel,
- Philippe Calain,
- Ali S. Khan,
- Ethleen Lloyd,
- Pierre E. Rollin,
- Thomas G. Ksiazek and
- Clarence J. Peters for the Commission de Lutte contre les Epidémies à Kikwit
- Division of Viral and Rickettsial Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia; Red Cross and Hôpital Général de Référence de Kikwit, Kikwit, and World Health Organization, Kinshasa, Democratic Republic of the Congo; Department of Virology, University of the Witwatersrand and National Institute for Virology, Sandringhan, South Africa
- Reprints or correspondence: Dr. Sherif R. Zaki, Infectious Disease Pathology Activity, Centers for Disease Control and Prevention, 1600 Clifton Rd., N.E., Mailstop G-32, Atlanta, GA 30333 (sxz1{at}cdc.gov).
Abstract
Laboratory diagnosis of Ebola hemorrhagic fever (EHF) is currently performed by virus isolation and serology and can be done only in a few high-containment laboratories worldwide. In 1995, during the EHF outbreak in the Democratic Republic of Congo, the possibility of using immunohistochemistry (IHC) testing of formalin-fixed postmortem skin specimens was investigated as an alternative diagnostic method for EHF. Fourteen of 19 cases of suspected EHF met the surveillance definition for EHF and were positive by IHC. IHC, serologic, and virus isolation results were concordant for all EHF and non-EHF cases. IHC and electron microscopic examination showed that endothelial cells, mononuclear phagocytes, and hepatocytes are main targets of infection, and IHC showed an association of cellular damage with viral infection. The finding of abundant viral antigens and particles in the skin of EHF patients suggests an epidemiologic role for contact transmission. IHC testing of formalin-fixed skin specimens is a safe, sensitive, and specific method for laboratory diagnosis of EHF and should be useful for EHF surveillance and prevention.
In April and May 1995, health authorities in the Democratic Republic of the Congo (DRC) reported numerous human deaths associated with a rapidly progressive hemorrhagic fever—like illness. The fatalities occurred in and around Kikwit, a large city located several hundred kilometers east of the capital, Kinshasa [1, 2]. At least 2 large clusters of patients in two different local hospitals were identified, and reports indicated the disease had spread among health care workers involved in patient care and treatment [3]. Initially, the disease was clinically diagnosed as epidemic dysentery; however, because some local medical providers suspected the possibility of a viral hemorrhagic fever (VHF), patient specimens were forwarded to the Centers for Disease Control and Prevention (CDC) for laboratory testing. Immunoassay and polymerase chain reaction (PCR) testing identified Ebola (EBO) virus as the etiologic agent for the outbreak [4–7].
The epidemic of EBO virus hemorrhagic fever (EHF) in Kikwit was the first large outbreak of recognized filovirus disease among humans in almost 20 years [8]. International response teams assisted local officials in controlling the outbreak through the institution of barrier-nursing methods in local hospitals, community education, and active disease surveillance [1, 3, 9]. By the end of the outbreak in July 1995, a total of 316 people had contracted EHF; of these, 250 died [3].
EBO virus is a member of a unique negative-stranded RNA virus family, the Filoviridae, and causes a severe and often fatal hemorrhagic fever [10, 11]. The clinical diagnosis of EHF is often presumptive, and laboratory confirmation is essential [11–13]. Traditionally, the laboratory diagnosis of EHF has been accomplished through virus isolation or serologic assays [14–17]. Because of the biosafety hazards associated with the handling and testing of EBO virus, these assays can be performed in only a few specialized laboratories worldwide. Such testing requires that potentially dangerous biological specimens be transported to these laboratories from remote sites where the disease occurs, using a cold chain and rigorous international packaging and shipping procedures.
This report describes the development of a novel, safe, sensitive, and specific diagnostic immunohistochemical (IHC) test for EBO virus infections, which uses formalin-fixed postmortem skin specimens. Implications for controlling the spread of the virus and surveillance for EHF are also discussed.
Materials and Methods
Patient tissues
Formalin-fixed tissues from 19 patients with suspected EHF were submitted to CDC for pathologic and immunopathologic examination. For 18 patients, various tissues, including liver and skin specimens, were obtained at postmortem examination. Skin-snip biopsy specimens were taken from both the axilla and nape of the neck regions. In 1 patient who survived her illness and developed invasive periorbital mucormycosis, an eyelid biopsy specimen obtained 24 days after disease onset was submitted [18]. In several cases, formalin-fixed tissues were postfixed in glutaraldehyde and processed for electron microscopy (EM) examination. Whole blood or serum were available for virologic and serologic studies for 12 cases. Frozen liver specimens were available from an additional 2 cases (cases no. 5 and 14). Routine hematoxylineosin—stained sections were examined in all cases, and clinical and laboratory reports were reviewed when available.
Cell lines
Controls for light microscopic IHC studies were generated from formalin-fixed and paraffin-embedded pellets of minced normal human tissues mixed with either noninfected Vero E6 (ATCC Vero clone CRL 1586) cells or Vero E6 cells infected with EBO subtypes Sudan, Zaire (EBO-Z), Côte d'Ivoire, or Reston. Vero cells infected with Marburg, yellow fever, Crimean-Congo hemorrhagic fever, and Lassa fever viruses were used as additional controls. The noninfected and various infected Vero E6 cells were grown and maintained in Eagle MEM supplemented with 5% heat-inactivated fetal bovine serum as previously described [19].
Antibodies
The following antibodies were tested for their suitability in IHC assays on formalin-fixed tissues: an EBO-Z—hyperimmune mouse ascitic fluid (HMAF), an EBO-Z—immune rabbit serum, and several monoclonal antibodies directed against the EBO-Z virus nucleoprotein and glycoproteins.
IHC
IHC assays were performed by using a labeled streptavidin-biotin method essentially as described for the detection of Sin Nombre virus antigen [20]. In brief, 4-µm sections of paraffinembedded tissues placed on Fisher Plus slides (Fisher Scientific, Pittsburgh) were deparaffinized and rehydrated by immersion in graded alcohol solutions. The tissue sections were digested in 0.1 mg/mL Proteinase K (Boehringer Mannheim, Indianapolis) in 0.6 MTris (pH 7.5)/0.1% CaCl2. A blocking step was performed using 20% normal swine serum in Tris-saline solution (pH 7.5) with 0.25% Triton ×-100. The optimal dilutions for the primary antibodies were determined in a pilot experiment by using a series of titrations applied to control slides. The primary antibody was applied to the tissue section and incubated at room temperature for 90 min. This step was followed with a 15-min incubation at room temperature with biotinylated swine anti-mouse and anti-rabbit immunoglobulins and a subsequent 15-min incubation at room temperature with streptavidin alkaline phosphatase conjugate (Dako, Carpinteria, CA). The alkaline phosphatase activity was detected by using naphthol/fast red substrate (Dako), and the sections were counterstained in Mayer's hematoxylin (Fisher Scientific) and mounted with aqueous mounting medium (Signet Laboratories, Dedham, MA). The specificity of EBO virus IHC staining was confirmed in all instances by replacing primary antibodies with indifferent antibodies (isotype-identical murine antibodies, HMAF, nonimmune sera, and irrelevant immune sera). Control tissues included formalin-fixed and paraffin-embedded EBO virus— and non-EBO virus—infected Vero E6 cell lines. Additional controls included noninfected cell lines and non-EBO virus autopsy tissues (skin and liver).
Serology
All serum samples were tested by using a combination of three tests: an antigen-detection ELISA, which was developed for identification of EBO viral antigens [17, 19]; an IgM-capture ELISA, using native EBO-Z viral antigens grown in Vero E6 cells; and an IgG ELISA, using detergent-extracted antigens [17, 19].
Virus isolation
Virus isolation was attempted by inoculation of confluent monolayers of Vero E6 cells with the blood and, in some instances, liver biopsy samples. Cells were observed every other day for signs of cytopathic effect. After 2 weeks, spot slides were made with scraped cells, fixed, and tested for the presence of EBO viral antigens by immunofluorescence assay (IFA) [19].
Results
EBO virus IHC assays
Initially, several antibodies were evaluated for use in IHC assays. One rabbit polyclonal antibody and an HMAF were selected for their specific reactivity with EBO virus antigens in formalin-fixed tissues. No staining was observed with noninfected cells or cells infected with viruses other than EBO virus or in non-EBO virus—infected tissues used as controls. The monoclonal antibodies tested either did not react with EBO antigens in formalin-fixed tissues or could not be used because of nonspecific immunostaining. Tissues from all cases were tested with both polyclonal rabbit and mouse antibodies with similar results.
Clinical and pathologic findings
Only 14 of 19 patients with suspected hemorrhagic fever included in this study met the surveillance case definition for EHF [3]; specimens from all of them were positive for EBO antigen as determined by IHC (table 1). Samples were obtained from all but 1 of the patients at death; an eyelid biopsy was submitted from 1 patient (case no. 13) who had previous serologic evidence of EHF and who was hospitalized in Kinshasa with invasive periorbital mucormycosis [18]. The last patient (case no. 14) evaluated in this study died on 16 July 1995. Two of the case-patients were health care workers: One was employed with the local Red Cross and is reported to have become infected by exposure to a cadaver; the other worked as a nurses' aid at Kikwit General Hospital. Of the remaining patients, all except 1, for whom exposure was unknown, had family members listed as the source of infection.
Eight of the patients (57%) were female. The mean age was 36 years (range, 3 months to 65 years). Twelve of the deceased patients had onset between 8 May and 19 June 1995; they were all admitted to Kikwit General Hospital after the arrival of the international response team in May 1995, which allowed some systematic, although still sketchy, collection of clinical details. These 12 patients died within a mean of 10 days (range, 4–14) of disease onset. The patient who survived (case no. 13) had onset in late April, and an eyelid biopsy sample was obtained on day 24 after onset. The most common symptoms reported on the case report form included fever, asthenia, abdominal pain, nausea or vomiting, diarrhea, and headache. Hemorrhagic manifestations were noted in 11 of the 14 cases. Seven had minor bleeding in the form of conjunctival hemorrhage, petechiae, epistaxis, or gum bleeding. More severe bleeding in the form of ecchymosis, bleeding from puncture sites, hematuria, menometrorrhagia, or gastrointestinal bleeding was noted in the remaining 4 patients.
Using light microscopy, we examined available liver and skin tissue samples from each of the 18 fatal cases and the skin biopsy from the nonfatal case. In brief, in all 13 examined liver specimens from EHF patients, we found various degrees of hepatocellular necrosis, microvesicular fatty change, and Kupffer's cell hyperplasia (figure 1A). In all cases, characteristic intracytoplasmic viral inclusions were seen within hepatocytes and were most abundant in periportal zones and surrounding areas of necrosis (figure 1B). The inclusions were usually numerous, eosinophilic, and oval or filamentous.
Figure 1
Histopathologic and ultrastructural features in livers of Ebola (EBO) hemorrhagic fever patients. A, Low-power photomicrograph of hematoxylin-eosin—stained liver showing hepatocellular necrosis, numerous EBO viral inclusions (arrowheads), and sinusoidal dilatation and congestion. Original magnification, ×100. B, Higher-power magnification of several hematoxylin-eosin—stained hepatocytes containing typical intracytoplasmic, filamentous, and eosinophilic EBO viral inclusions. Original magnification, ×100. C, EBO viral inclusions as seen in cytoplasm of infected hepatocytes in thin-section electron micrograph of liver (curved arrows). Abundant and pleomorphic filamentous extracellular EBO virus particles are seen in hepatic sinusoids (arrow). Bar = 1 µm.
Figure 2
Ebola (EBO) antigen distribution in livers of EBO hemorrhagic fever patients, as determined by immunohistochemistry. A, Low-power photomicrograph of liver showing immunostaining of EBO virus antigens throughout tissue section in sinusoids, sinusoidal lining cells, and hepatocytes. Original magnification, ×100. B, Higher-power magnification showing immuno staining of viral inclusions within hepatocytes as well as abundant antigens in sinusoids and sinusoidal lining cells. Original magnification, 1250. A and B, Immunoalkaline phosphatase staining, naphthol fast red substrate with light hematoxylin counterstain.
Figure 3
Ebola (EBO) antigen distribution in skin specimen of EBO hemorrhagic fever patients as determined by immunohistochemistry. A, Extensive amounts of EBO virus antigens are seen primarily within endothelial cells and fibroblasts in dermis. Note rare immunostainig of cell (arrow), perhaps Langerhans', in epidermis. Original magnification, ×100. B, Massive amounts of viral antigens are seen in connective tissue surrounding sweat glands. Original magnification, ×50. C, Higher-power magnification of B, showing immunostaining of endothelial cells and fibroblasts. Note also abundant extracellular antigens and absence of staining within sweat glands. Original magnification, ×100. D, Antigen-positive cell and immunostaining within sweat gland. Original magnification, ×158. A–D, immunoalkaline phosphatase staining, naphthol fast red substrate with light hematoxylin counterstain.
In comparison, the histopathologic changes in the skin were subtle and primarily confined to changes within endothelial cells in the dermis. In all cases, the endothelial cells were uniformly affected, with various degrees of swelling, necrosis, and drop-out. Fibrin thrombi were rare, and vessel wall necrosis was absent. Other histopathologic changes consisted mainly of dermal edema and focal erythrocyte extravasation. Perivascular inflammatory infiltrates were either absent or mild. Microscopic changes in the epidermis were not conspicuous and consisted of mild spongiosis and rare apoptotic cells.
Microscopic examination of liver and skin specimens from 2 of the non-EHF cases was unremarkable. In the remaining 2 non-EHF cases, abundant malarial pigments were observed in the liver tissue, and no specific histopathologic change was seen in the skin.
IHC, serology, antigen detection, and virus isolation test results
IHC, serology, antigen detection, and virus isolation data on 19 people with suspected hemorrhagic fever included in this study are summarized in table 1. Of the 19 patients with suspected hemorrhagic fever whose specimens were submitted for EBO virus testing, 14 were diagnosed as having EHF on the basis of results from IHC analysis alone or in combination with serology, antigen detection, and/or virus isolation (table 1). The results of the different assays were in absolute agreement when more than one assay was performed. Of the 14 patients diagnosed with EHF, all had unequivocal IHC evidence of EBO antigens in the skin. Thirteen cases also showed extensive immunostaining in the liver. Liver tissue was unavailable for case number 13. Of the 14 patients with positive IHC skin tests, 7 were positive by both serology and antigen detection tests (table 2). Sera from 3 patients with a positive IHC skin test were antigen positive and negative for IgG and IgM antibodies; it is noteworthy that all 3 patients had a clinical course of ⩽8 days (table 1). Sera collected on days 27, 39, and 41 of illness from a patient with a positive IHC skin test who survived her illness was positive for IgG and IgM but negative for EBO viral antigens. The skin biopsy was obtained 24 days after disease onset.
Figure 4
Association of Ebola (EBO) virus infection with endothelial cell damage and death in skin specimen of EBO hemorrhagic fever patient. A, Photomicrograph of dermal blood vessel showing prominent karyolysis and karrhyorrhexis. Original magnification, ×158. Hematoxylin-eosin staining. B, Blood vessel immunostained for EBO, showing EBO virus antigens in endothelial cells lining blood vessel. Original magnification, ×250. Immunoalkaline phosphatase staining, naphthol fast red substrate with light hematoxylin counterstain.
Figure 5
Ultrastructural features of Ebola (EBO) virus infection in skin of EBO hemorrhagic fever patient. A, Abundant virus particles (arrowheads) are seen in association with extracellular matrix of dermis. Inclusion-bearing, infected macrophage is also seen within dermal blood vessel. Inclusion consists of lighter-staining granular areas admixed with areas of dense tubular nucleocapsids (arrow). B, Replication of EBO virus within dermal endothelial cells is evidenced by presence of intracytoplasmic inclusion (arrow). Few virus particles, including 1 with characteristic “shepherd's crook” (arrowhead), are seen in proximity to blood vessel. Bars = 500 nm).
Table 1
Nineteen suspected cases of hemorrhagic fever: Available clinical and laboratory tests for Ebola virus infection.
Table 2
Summary of diagnostic tests for Ebola virus infection performed on 19 suspected cases of hemorrhagic fever.
Cellular targets and IHC distribution of EBO virus antigens
IHC evidence for the presence of EBO virus antigens was documented in both the liver and skin of all the patients with EHF. In brief, in all cases, specific immunostaining was observed by using both polyclonal rabbit and mouse sera. The viral antigen—positive cells in the liver were abundant in all cases and seen in both hepatic parenchyma and portal tracts. In the hepatic parenchyma, EBO virus antigens were seen within hepatocytes, sinusoids, and sinusoidal lining cells (figure 2A). Hepatocyte immunostaining was primarily seen in association with areas of hepatocellular necrosis and within cells demonstrating viral inclusions, cytoplasmic eosinophilia, nuclear pyknosis, and cytolysis (figure 2B).
The amount and extent of immunostaining in the skin was also considerable in all EHF cases examined. No difference in intensity and extent of immunostaining was observed between paired skin specimens taken from the axilla and nape of the neck in individual cases. Immunostaining was seen in the dermis and subcutaneous tissue primarily within endothelial cells and fibroblasts (figure 3). Various degrees of endothelial cell swelling, necrosis, and drop-out were associated with EBO virus infection (figure 4). Heavy concentrations of viral antigens were usually seen in areas surrounding the sweat glands (figure 3B—D). In these areas, immunostaining was primarily extracellular in association with areas of cell lysis and karyorrhectic debris. Only rare immunostaining was seen within sweat glands and ducts (figure 3D). Similarly, rare antigen-positive cells were seen within the epidermis. The nature of these cells, epithelial cells compared with Langerhans' cells, could not be discerned by light microscopy.
EM
Virus particles and inclusions were seen in all skin and liver specimens (from EHF patients) that were examined by EM. EM examination of liver tissue revealed a density and distribution of intracytoplasmic viral inclusions similar to that observed by light microscopy and IHC (figure 1C). The inclusions varied in size and consisted of aggregates of viral nucleocapsids. Extracellular EBO virus particles were seen in hepatic sinusoids. Virus particles were usually abundant and occupied most of the hepatic sinusoids in some cases.
EM examination of skin specimens revealed abundant viral inclusions within endothelial cells and fibroblasts (figure 5A–B). Virus particles were also numerous and seen primarily in extracellular areas in association with endothelial cells and fibroblasts. No virus particles or inclusions were seen in surface epithelium or sweat glands and ducts of skin specimens examined. In some sections, virus particles were observed in close proximity (within 2.5 µm) to the dermal epidermal junction.
Discussion
In this study, characterization of histopathology and viral tropism in liver and skin by IHC and EM provided useful insights into the pathogenesis of human EHF. EBO virus antigens were seen primarily within endothelial cells, mononuclear phagocytic cells, and fibroblasts and free within hepatic sinusoids and the interstitium. The distribution of virus particles and inclusions correlated with IHC localization of viral antigens. IHC showed an association of cellular damage with viral infection, suggesting that cell damage may be mediated by a direct cytopathic effect. Significant injury to the vasculature and increased endothelial permeability appear to be central to the pathogenesis of the shock syndrome and bleeding seen in EHF. Infection of mononuclear phagocytes and endothelial cells also appears to play a critical role in the pathogenesis of EHF through the secretion of physiologically active substances, including cytokines and other inflammatory mediators. A recent study suggests that a high degree of immune activation accompanies and potentially contributes to a fatal outcome in EHF patients. Marked elevations of interferon-γ and -α, interleukin- 2 and -10, and tumor necrosis factor-α, were seen in fatal EHF cases [21]. The pathology and pathogenesis of filovirus infections have also been studied in several animal model systems, including monkeys, guinea pigs, hamsters, bats, and suckling mice. The pathologic features in these experimentally infected animals are somewhat similar to those observed in humans and provide an opportunity to study these diseases systematically [22–32].
Traditionally, the diagnosis of EBO virus infections has been performed either through viral isolation or by use of ELISAs for detection of antigens and immunoglobulins. Virus isolation is relatively simple, and the virus multiplies and can be detected by IFA in the cells after a few days. The risk posed by the manipulation and amplification of virus requires that clinical virology be performed in a biosafety level 4 laboratory. The virus is stable in the serum, plasma, or whole blood; in Kikwit, it was possible to isolate the virus from blood specimens retained for a month at ambient temperature (Rollin P, Swanepoel R; unpublished data). For rapid diagnosis, the best and most robust test is an ELISA that allows detection of EBO virus antigen in blood or in tissue suspensions. Survivors of EHF develop IgM and IgG that are detectable by ELISA (IgM capture and indirect IgG) [7,17]. More recently, PCR has been used for rapid diagnosis and virus subtyping, but to be fully exploited and controlled, it can be performed only in specialized laboratories [6, 33].
In the past, pathologic examination of postmortem specimens has been used in the diagnosis and surveillance of yellow fever and other VHFs [34, 35]. However, because the pathologic features in various VHFs are similar and resemble those of other viral, rickettsial, and bacterial infections, a definitive diagnosis cannot be made by histologic examination alone [13]. Several early studies described the pathologic features of human filovirus infections [36–41], but, for the most part, these studies were limited to examinations of specimens from small numbers of patients, largely because of the risk of hemorrhage associated with biopsies and because of biosafety concerns during autopsy [42, 43]. This histopathologic and immunopathologic study of 14 patients represents the largest series of EHF cases evaluated to date. Histopathologic changes in the skin tissue were not pathognomonic and consisted mainly of various degrees of endothelial cell swelling and necrosis. On the other hand, various degrees of hepatocellular necrosis, Councilman's bodies, and Kupffer's cell hyperplasia were seen in the liver tissues of all EHF cases. The most consistent histopathologic finding in this study was the presence of characteristic intracytoplasmic viral inclusions within hepatocytes. These inclusions are especially helpful in strongly suggesting the diagnosis of EHF in cases with appropriate clinical manifestations and history. However, an unequivocal diagnosis can be made only by use of etiologically specific laboratory tests, such as virus isolation, specific antigen or antibody detection, PCR, or as reported here, by IHC.
The effectiveness of IHC as a diagnostic modality was established by absolute concordance with results of other laboratory tests for EBO virus infection (ELISA serology, antigen detection ELISA, and virus isolation) when one or more of the additional assays was used. The IHC test was shown to be extremely sensitive, as evidenced by IHC results demonstrating the presence of EBO virus antigen in skin specimens from all 14 EHF patients and in liver samples from 13 patients (the other EHF patient survived the illness, and therefore no liver biopsy was performed). In fact, 3 patients with a clinical course of <8 days were positive by antigen-detection ELISA and IHC but negative for EBO antibodies by ELISA. The finding of EBO virus antigens in a skin biopsy but not in serum obtained late in the course of the disease in an EHF patient who survived her illness suggests that viral antigens may persist in tissues after being cleared from the circulation or that their detection may be blocked by the formation of immune complexes [7].
The diagnosis of EHF in 1 case was accomplished solely by IHC, illustrating the unique role this method has in identifying EHF cases for which serum samples and frozen tissues are unavailable. Formalin-fixed biopsy specimens are not infectious; they may be shipped without special precautions or refrigeration and obtained in the most basic field conditions. Moreover, formalin remains stable at ambient temperatures until it is needed. This approach is advantageous over the use of viral cultures and fluorescent antibody tests that are commonly used for surveillance purposes and require special handling and transport of infectious material and a cold chain. The EBO skin test allows collection of formalin-fixed samples at minimal risk to those who use the kit or who transport collected samples. Unlike a liver biopsy, a skin biopsy allows for safe and simpler noninvasive sampling and can be obtained without use of needles by using a punch biopsy tool or even a pair of scissors. A skin biopsy taken in the axilla or nape of the neck is less disfiguring to the body and more acceptable for cultural reasons than an autopsy or even a liver biopsy.
EHF is a rare disease, and the number of sporadic cases occurring in tropical Africa is unknown. Investigations in northwestern DRC in 1977–1978 revealed that ∼7% of the residents had immunofluorescent antibodies to the EBO virus [44]. Using IgG ELISA, background prevalence of EBO viral antibodies in residents of the Kikwit area was 2.2% [45]. The lack of specificity of the IFA test may have led to some false-positive results [7], and the lack of proper storage, safe transportation, and an adequate cold chain have limited adequate sample collection for serologic testing and virus isolation. More recent events in Côte d'Ivoire and Gabon also suggest that sporadic human infections do occur [46, 47]. IHC testing of skin biopsy specimens from suspected cases of EHF offers a more practical, cost-effective surveillance mechanism that can be maintained over a long period in a large geographic area without on-site supervision and support. A surveillance system has been set up in the Bandundu region in DRC that combines a program of training physicians to recognize EHF signs and symptoms, the use of a kit to collect appropriate specimens for laboratory confirmation, and implementation of protective isolation procedures in health facilities [42, 48]. Since its development in 1995, the surveillance program using the skin test has continued to expand and has been introduced to other regions in DRC and in neighboring countries. In 1996, the surveillance kits were used to confirm an outbreak of EHF in Gabon that occurred among local residents who had contact with flesh from a chimpanzee found dead in the forest [46]. EBO virus antigens were detected in 8 of 10 formalin-fixed specimens submitted to the CDC (unpublished data). The skin test has also been used to diagnose EBO in monkeys found dead in the jungle and in fact may prove useful in tracking the reservoir [46]. Initial steps are currently being taken to evaluate the use of skin biopsies in the diagnosis of other hemorrhagic fevers. It is not yet known if the skin will provide a sensitive diagnostic specimen for other hemorrhagic fevers, and, therefore, simultaneous liver and skin biopsies are being collected. Preliminary studies of Lassa fever cases showed that about one-third of the skin specimens were positive for Lassa fever virus antigens by IHC (unpublished data).
The finding of EBO virus antigens and virus particles in the skin of EHF patients provides new insights into a possible epidemiologic role for contact transmission and strategies for disease spread and prevention. A study showed that direct physical contact with an infected person during the clinically apparent phase of illness was the most important risk factor for secondary household transmission of EHF during the Kikwit outbreak [49]. Furthermore, even when adjusted for contact with the living patient, touching the cadaver remained an independent risk factor. One possible explanation for the role of direct physical contact in transmission is the presence of abundant virus particles and antigens in the skin in and around sweat glands. In Kikwit, retrospective follow-up identified suspected cases of EHF as early as January 1995 and showed that EHF was occurring in the community for several months prior to amplification of the outbreak via nosocomial transmission among undiagnosed patients at Kikwit General Hospital and Kikwit II Hospital [3]. If EHF had been suspected or confirmed by use of the skin test kit in 1 of the early cases in Kikwit from January to April, the hospital might have taken steps earlier to improve infection control measures, such as the use of sterile needles, gloves, hand washing, and isolation of suspected patients [42]. These measures may not have prevented all transmission cases of EHF in the community, but they may have helped protect many of the 80 health workers who contracted the disease [3, 50].
The case history of the last recognized patient during the Kikwit outbreak emphasizes the important role of skin biopsies in long-term surveillance and disease prevention. The woman who died of EHF was a 27-year-old housewife from Nzinda, Kikwit. She died at home after discharge from Kikwit III Hospital, where she had been hospitalized for management of a septic abortion. She was brought in dead to Kikwit II Hospital on 16 July 1995. Her clinical features included fever, headache, vomiting, anorexia, diarrhea, asthenia, abdominal pain, dyspnea, and melena. Her death was attributed to infection complicating the abortion, and the post-mortem samples were not identified in any way to indicate that it was obtained from a patient with a highly suspicious clinical illness. Skin and liver specimens were collected in the morgue on 17 July 1995 and transported to CDC, where EBO antigens were visualized by IHC. Follow-up surveillance indicated that she had had contact with a prior case. Thirteen family members of this individual were followed for 21 days without subsequent disease in the household [48]. The retrospective identification of this case from Kikwit at the end of an EHF outbreak highlights the difficulties in establishing a clinical diagnosis, even in an area where such an outbreak has just occurred and clinical suspicion is high, and it highlights the utility of postmortem skin snips as a surveillance mechanism. The long-term viability of the skin surveillance system will be proven over time. However, the IHC testing of skin for EBO provides a simple, easy-to-use, and safe assay that empowers health personnel in Africa to confront a devastating disease that lurks in the environment and reemerges periodically but unpredictably.
Acknowledgment
The authors thank John O'Connor for helpful discussions and editorial comments.
- © 1999 by the Infectious Diseases Society of America
References
This Article
- AbstractFree
- » Full Text (HTML)Free
- Full Text (PDF)Free
Classifications
Services
Citing Articles
Google Scholar
PubMed
Related Content
Share
- Email this article
-
CiteULike
Connotea
Delicious
Facebook
Google+
Mendeley
Twitter
Published on behalf of
IDSA Members: For your free access to this journal, log in via the IDSA members area.
The Journal
Impact Factor: 6.288
5-Yr impact factor: 6.103
For The Media
For Authors
- Services for authors
- Self-archiving policy
- IDSA Conflict of Interest policy
- Instructions for Authors
- ICMJE Form
- Submit Now!
Open access options for authors visit Oxford Open
This journal enables compliance with the NIH Public Access Policy
Corporate Services
Alerting Services
Related IDSA Publications
