Previous PageIndexNext Page


Center for Disease Control, Atlanta, Georgia 30333, U.S.A.

So few specimens of tissue from fatal cases of Ebola virus disease have been available for pathologic study that no description can be considered representative, and any analysis of pathogenetic mechanisms should be recognized as speculative. No report of gross pathologic findings is available and specimens for histopathologic study have consisted only of liver tissue from three cases in Zaire (sent to CDC by the WHO Field Team), and liver, spleen, and kidney tissues from two cases in the Sudan (sent to CDC by Dr. D.S. Ridley, Hospital for Tropical Diseases, London, and Dr. D.I.H. Simpson, London School of Hygiene and Tropical Medicine). Clinical pathology information will be presented by other contributors to this Colloquium.

Because of this paucity of tissue and lack of information of Ebola virus pathology it will be necessary to draw upon findings from Marburg virus studies carried out after the 1967 and 1975 episodes. Findings from 1967 primarily derive from the papers by P. Gedigk, H. Bechtelsheimer, G. Korb, and H. Jacob 1,2,3 ; pathologic findings from the single fatal case in South Africa in 1975 are not formally available but some observations are included here based upon a set of histologic slides sent to CDC by Professor J.H.S. Gear and his 4 colleagues of the South African Institute of Medical Research, Johannesburg (4).

Ebola Virus Infection

The three liver specimens from confirmed Ebola virus disease cases in Zaire were remarkably similar. There was fatty change and necrosis of hepatocytes and Kupffer cells (Figures 1,2). This necrosis was focally distributed throughout lobules, in some cases involving single cells and in other cases extending from central veins to lobular peripheries (Figures 3,4). The sequence of hepatocyte necrosis involved an initial cytoplasmic eosinophilia, then a shrinking, darkening and dissolution of nuclei (nucleoclasia). Intact cells with hyalinized cytoplasm and ghostlike nuclei apparently remained in place for some time; but finally rarifaction, swelling and cytolysis occurred, leaving large amounts of karyorrhectic debris in situ. Considering the extent of this necrosis, there was remarkably little inflammatory infiltration into sinusoids. There were large numbers of hepatocyte mitoses, but often binucleate cells underwent the same necrotic changes as they were entrapped by expanding foci of infection.

Extraordinary large and vivid intracytoplasmic eosinophilic inclusion bodies were present in hepatocytes in two of the three Zaire cases; further study showed that these represented an extreme and that large numbers of smaller and less distinct inclusions were present in all three cases (Figures 5,6). Because smaller inclusions had indistinct margins and an eosinophilic color matching the early hyaline change of the cytoplasm of infected hepatocytes, they could only be identified with certainty in rather normal cells. Inclusion body stains helped in this identification, but as in most acute hepatocellular infections, the presence of large numbers of Councilman-like bodies in areas of necrosis was a complicating factor. This matter has practical importance since histologic identification of Marburg/Ebola virus inclusions, in the presence of Councilman-like bodies, could have presumptive diagnostic value in laboratories where virus isolation or immunofluorescent methods are not available. The identity of the virus inclusions in the three Ebola cases was confirmed by thin-section electron microscopy of formalin-fixed liver tissue (Figure 7). Although preservation was poor, the inclusions found within the cytoplasm of hepatocytes were clearly made up of massed tubules which were identical to the internal constituent (? the nucleocapsid) of virus particles (Figure 8). These structures were indistinguishable from those found in Vero cell cultures infected with Ebola or Marburg viruses; similarly they were indistinguishable from hepatocyte inclusions in human, monkey and guinea pig(5,6,7) infections caused by Marburg virus . In the same three Zaire liver speci mens, large number of virus particles were found in all of the extracellular spaces (sinusoids, spaces of Disse and areas of necrosis); these were indistinguishable from Ebola virus particles in cell culture and Marburg virus particles in cell culture and in vivo (Figure 9). No Aiver necrosis was evident in the very small tissue specimens collected from fatal cases of hemorrhagic fever in the Sudan. There was some necrosis and calcification in tubules and glomerular tufts of kidney and there were necrotic cells in the spleen specimens. However, because no information was received about the course of disease nor laboratory confirmation of the diagnosis, it is prudent to put off interpretation of this exceptional sparing of the liver.

Marburg Virus Infection

Comparison of the liver lesions of the Zaire Ebola cases with those of Marburg infections in 1967 and 1975 indicates precise similarities. At this point, further comparisons are impossible, but a brief review of Marburg disease pathology may be of value. In the fatal Marburg virus infection in South Africa in 1975, hepatocellular necrosis was the most pronounced pathologic finding (Figures 10, 11). The focal necrosis was similar to that in the Zaire Ebola cases, but the damage seemed almost synchronous, that is, large numbers of hepatocytes seemed to have been caught at the time of death of the patient in a similar state of early cytoplasmic hyalinization and eosinophilia without frank dissolution (Figures 12, 13). Small inclusions and Councilman-like bodies were identified in this liver tissue (Figures 14, 15). There were necrotic cells in other organs, but not in numbers like those in the liver. In the kidney there was tubular necrosis and in glomerular capillaries there were multifocal fibrin thrombi characteristic of disseminated intravascular coagulopathy (DIC) 8 (Figure 16). There was also pulmonary edema and an effusion of macrophages into alveoli (Figure 17). It is hoped that a detailed pathologic description of this case will be published soon by South African pathologists.

Because study of the Marburg virus disease of 1967 was so comprehensive, the findings have special comparative value(1,2,3 ). Gross pathologic findings included evidences of hemorrhagic diatheses into skin, mucous membranes, soft tissues, visceral organs and into the stomach and intestines. There was swelling of spleen, lymph nodes, kidney and especially brain. Microscopically, focal necroses were found in many organs, most conspicuously in the liver, lymphatic system, testes, and ovaries. Liver necrosis, identical to that described in Ebola virus infection, was especially prominent, and as in the latter there was no favoring of any particular zone of lobules. Necrotic foci grew by expansion. Inclusion bodies were prominent, as were Councilmanlike bodies and basophilic karyorrhectic debris. Liver biopsies from convalescent patients indicated rapid regeneration coinciding with the decline of serum transaminase levels. Lymphoreticular organ changes included necrosis of 1) lymphoid follicles, 2) the red pulp of spleen, and 3) the medulla of lymph nodes. An eosinophilic "thrombic" debris was left in situ as a result of this necrosis. There were histologic evidences of hemorrhagic diathesis in many organs. In the brain there was a diffuse panencephalitis with glial nodule formation, perivascular lymphocytic cuffing, and evidence of interstitial edema.


The pathophysiologic alterations which make Marburg and Ebola virus infections so devastating have not been studied systematically. The cause of the hemorrhagic diatheses was searched for in tissues from the fatal Marburg cases in 1967 but no vascular lesions were identified. The increase in vascular permeability, associated reduced effective circulating blood volume, interstitial edema in visceral organs and brain, and DIC may all stem from the liver necrosis and renal tubular necrosis. It is not clear how the shock syndrome in this disease relates to activities of pharmacologic mediators of capillary permeability and complement split products. Whatever the underlying mechanism, it was concluded after the 1967 Marburg episode that DIC and cerebral edema played an essential role in the fatal outcome of infection. The magnitude and rapidity of the destructive events of Marburg or Ebola infection would predict that the terminal DIC/shock syndrome would be most difficult to deal with and this is the case in fact. The success in South Africa in saving two Marburg patients with heparin and supportive treatment and the success in the United Kingdom in saving one Ebola patient with convalescent plasma does not tell us too much about the pathogenetic mechanisms of these hemorrhagic fevers, but they do remind us of the regenerative capacity of the liver and kidney tubules.

Differential Pathologic Diagnosis

Although no pathognomonic lesion has been found which would permit certain histopathologic diagnosis of Marburg/Ebola infections, some pathologists considered that the overall pattern of lesions is unique and distinguishable from 1,2 yellow fever, infectious hepatitis, and other well-known infections. Others deem the differential diagnosis extremely difficult, especially in the context of examination of tissues from a single case and in a setting where Lassa Fever must also be considered. In any event, it is clear that pathologic examination is not a substitute for etiologic diagnosis(8), but in areas without full laboratory facilities, histologic diagnosis may be as important as it has been for many years in yellow fever diagnosis. The recent production by Drs. Y. Robin and J. Renaudet (with the support of the WHO) of an excellent 35mm transparency set with accompanying text, entitled 'La Fièvre Jaune -- histopathologique positif et differentiel', provides an important new resource for the differential pathologic diagnosis of hemorrhagic fevers in the African setting where Lassa fever, yellow fever, Crimean hemorrhagic fever, typhoid fever, infectious hepatitis, leptospirosis, and other infectious diseases occur. This slide set is complemented by the recent description of the pathology of human Lassa fever by Winn and his colleagues(9,10), and recent consideration of the differential diagnosis of hemorrhagic fevers in 11 Africa (from the viewpoint of Lassa fever) by Monath and Casals (11).

Pathology in Experimental Animals

After the 1967 Marburg virus episode, several attempts were made to develop animal models for further study of the pathology and pathogenesis of the infection. Monkeys (Cercopithecus aethiops), guinea pigs, and hamsters were found to be most valuable(6,7,12,13). After serial passage of the virus in guinea pigs, infection was invariably lethal; the disease was marked by a swollen and friable liver and spleen. Microscopically, these changes coincided with focal liver necrosis and congestion, hemorrhage and destruction of lymphoid elements in the spleen. In addition to similar liver and spleen lesions, meningitis and hemorrhagic vascular lesions of the brain parenchyma were most characteristic of the hamster disease. Severe hepatocellular necrosis and lymphoreticular necrosis marked the fatal disease in monkeys. In each of these three species, hemorrhagic diatheses were found, and although the pathogenetic characteristics of the human hemorrhagic fever were not reproduced precisely, similarities were such that the models would likely have been of value in testing prophylactic or therapeutic regimens (e.g. passive or active immunization, interferon, or chemotherapeutic agents). However, little was done with these Marburg infection models except for the descriptive studies.

Investigation of Ebola virus infection in experimental animals has been extremely limited so far. In guinea pigs, pathologic studies carried out at the Microbiological Research Establishment at Porton Down by Bowen and his colleagues 14 and at the CDC 15 on virus from Zaire and from the Sudan indicated that focal liver necrosis was the most prominent lesion and that splenic white pulp necrosis and some -lymph node necrosis also occurred. As had been true of Marburg virus in studies done after the 1967 episode, the guinea pig liver disease varied in severity from a progressively destructive lethal course involving most hepatocytes when serially passaged virus was used, to an apparently self-limiting infection with calcification of necrotic hepatocytic foci when unpassaged virus was used. In a similar way, Ebola virus from Zaire seemed more capable of producing progressive hepatitis, and virus from the Sudan more often caused arrested, calcified lesions. Correspondingly, more guinea pigs survived infection with the Sudan virus. In these preliminary studies virus inoculum dose was not carefully controlled so no inference may be drawn as to virulence differences of the various isolates, but such studies will be done.

Ebola virus has not yet been studied in hamsters, and results of pathologic study of monkeys study of monkeys inoculated at Porton Down (for appraisal of interferon sensitivity of the virus) are not yet available.

In conclusion, it would seem that the pathologic alterations found in Ebola virus infection of man and experimental animals are similar to those in Marburg virus infection. Even in the absence of comprehensive comparative studies, it seems safe to draw upon our experiences with Marburg virus and to immediately direct our concern to developing the means for Ebola virus prophylaxis and treatment. The opportunity for a comprehensive study of the pathologic alterations in fatal human Ebola virus infection has been lost in the extremely difficult and hazardous circumstances in the field, but our discussion of the means of medical intervention can proceed, nevertheless, based upon presumed pathophysiologic characteristics of the disease.


This is not a comprehensive listing of the Marburg virus pathology literature. Extended bibliographies are found in references 1, 2, 3, 6, 7, and 8 and many other papers containing pathologic observations are included in the book Marburg Virus Disease (edited by Martini, G.A., and Siegert, R.) Springer-Verlag, New York, 1971.

1. Gedigk, P., Bechtelsheimer, H., Korb, G. (1971) in Marburg Virus Disease (edited by Martini, G.A., and Siegert, R.) Springer-Verlag, New York, pp. 50-53.
2. Bechtelsheimer, H., Korb, G., Gedigk, P. (1971) in Marburg Virus Disease (edited by Martini, G.A., and Siegert, R.) Springer-Verlag, New York, pp. 62-67.
3. Jacob, H. (1971) in Marburg Virus Disease (edited by Martini, G.A., and Siegert, R.) Springer-Verlag, New York, pp. 54-61.
4. Gear, J.S.S. et al. (1975) British Medical Journal, 4: 489-493.
5. Kissling, R.E., Robinson, R.Q., Murphy, F.A., Whitfield, S.G. (1968) Science 160: 888-890.
6. Kissling, R.E., Murphy, F.A., Henderson, B.E. (1970) Annals of New York Academy of Sciences, 174: 932-945.
7. Murphy, F.A., Simpson, D.I.H., Whitfield, S.C., Zlotnik, I., Carter, G.B. (1971) Laboratory Investigation, 24: 279-291.
8. Wulff, H., Conrad, L. (1977) in Comparative Diagnosis of Viral Diseases, Vol. 2 (edited by Kurstak, E., and Kurstak, C.) Academic Press, New York, pp. 3-33.
9. Winn, W.C., Jr., Monath, T.P., Murphy, F.A., Whitfield, S.G. (1975) Archives of Pathology, 99: 599-604.
10. Winn, W.C., Jr., Walker, D.H. (1975) Bulletin of World Health Organization, 52: 535-545. 50
11. Monath, T.P., Casals, J. (1975) Bulletin of World Health Organization, 52: 707-715.
12. Zlotnik, I., Simpson, D.I.H. (1969) British Journal of Experimental Pathology, 50: 393-406.
1. 13. Simpson, D.I.H. (1969) Transactions of Royal Society of Tropical Medicine, 63: 303-314.
13. Bowen, E.T.W., Lloyd, G., Harris, W.J., Platt, G.S., Baskerville, A.,Vella, E.E. (1977) The Lancet, 1: 571-573.
14. Johnson, K.M., Lange, J.V., Webb, P.A., Murphy, F.A. (1977) The Lancet, 1: 569-571.

Fig. 1. Ebola virus infection, Zaire. Liver. Focal hepatocellular necrosis with karyorrhexis, but little interstitial inflammation. In this case the severe damage to hepatocytes is not marked by any architectural disarray. Hematoxylin and eosin; X 110.

Fig. 2. bola virus infection, Zaire. Liver. Focal necrosis with fatty change and modest inflammatory response. In this case massive lobular structural changes were associated with the infection. Hematoxylin and eosin; X 110.

Fig. 3. Ebola virus infection, Zaire. Liver. Hepatocellular necrosis expanding from foci, marked in this case by zones of cells which are intact but undergoing eosinophilic cytoplasmic change, pyknosis, and dissolution of nuclei. Hematoxylin and eosin; X 250.

Fig. 4. Ebola virus infection, Zaire. Liver; same case as in Figure 3. A discrete focus of necrosis illustrating the paucity of inflammatory infiltration. Hematoxylin and eosin; X 375.

Fig. 5. Ebola virus infection, Zaire. Liver. Intracytoplasmic inclusion bodies (arrows) can only be discerned from Councilman-like bodies when found within intact hepatocytes. Hematoxylin and eosin; X 1400.

Fig. 6. Ebola virus infection, Zaire. Liver. Inclusion bodies (arrows), which are magenta in color when stained by hematoxylin and eosin, were not rendered more discernable from Councilman-like bodies by any of the several inclusion stains tried. Hematoxylin and eosin; X 1600.

Fig. 7. Ebola virus infection, Zaire. Liver; formalin-fixed tissue processed for electron microscopy. Identity of inclusion bodies (arrows) was confirmed as being identical to structures found in Ebola virus infected cell cultures. Thin section; X 11,000.

Fig. 8. Ebola virus infection, Zaire. Liver; formalin-fixed tissue processed for electron microscopy. At higher magnification inclusions such as that illustrated in Figure 7 were found to consist of cylindrical structures in an amorphous matrix. Thin section; X 46,000.

Fig. 9. Ebola virus infection, Zaire. Liver; formalin-fixed tissue processed for electron microscopy. Large numbers of virus particles within a distended sinusoid in an area of severe hepatocellular necrosis. Thin section; X 40,000.

Fig. 10. Marburg virus infection, South Africa, 1975. Liver. Focal hepatocellular necrosis marked by cytoplasmic eosinophilia, nuclear pyknosis and nuclear dissolution. Hematoxylin and eosin; X 250.

Fig. 11. Marburg virus infection, South Africa, 1975. Liver. A focus of infection in which total dissolution of liver cells has left only acellular debris. Hematoxylin and eosin; X 250.

Fig. 12. Marburg virus infection, South Africa, 1975. Liver. Hepatocellular necrosis in this area is marked by nearly synchronous reduction of cells to intact forms with only pyknotic or ghostly remnants of nuclear profiles. Hematoxylin and eosin; X 350.

Fig. 13. Marburg virus infection, South Africa, 1975. Liver. Higher magnification showing pyknosis (center) and nuclear dissolution (lower right) in intact hepatocytes. Hematoxylin and eosin; X 1400.

Fig. 14. Marburg virus infection, South Africa, 1975. Liver. Inclusion body (arrow) within the cytoplasm of an intact hepatocyte. Hematoxylin and eosin; X 1400.

Fig. 15. Marburg virus infection, South Africa, 1975. Liver. The identification of inclusion bodies (arrow) is extremely difficult with a background of extensive necrosis in which Councilman-like bodies are often produced. Hematoxylin and eosin; X 350.

Fig. 16. Marburg virus infection, South Africa, 1975. Kidney. Multifocal fibrin thrombi (arrows) in glomerular capillaries, characteristic of disseminated intravascular coagulopathy. Hematoxylin and eosin; X 450.

Fig. 17. Marburg virus infection, South Africa, 1975. Lung. Effusion of macrophages into alveolar space was associated with pulmonary edema (complicated by Candida infection). Hematoxylin and eosin; X 450.

Previous PageTop Of PageNext Page