Acute liver failure, multiorgan failure, cerebral oedema, and activation of proangiogenic and antiangiogenic factors in a case of Marburg haemorrhagic fever.
ABSTRACT A woman developed Marburg haemorrhagic fever in the Netherlands, most likely as a consequence of being exposed to virus-infected bats in the python cave in Maramagambo Forest during a visit to Uganda. The clinical syndrome was dominated by acute liver failure with secondary coagulopathy, followed by a severe systemic inflammatory response, multiorgan failure, and fatal cerebral oedema. A high blood viral load persisted during the course of the disease. The initial systemic inflammatory response coincided with peaks in interferon-γ and tumour necrosis factor-α concentrations in the blood. A terminal rise in interleukin-6, placental growth factor (PlGF), and soluble vascular endothelial growth factor receptor-1 (sVEGF-R1) seemed to suggest an advanced pathophysiological stage of Marburg haemorrhagic fever associated with vascular endothelial dysfunction and fatal cerebral oedema. The excess of circulating sVEGF-R1 and the high sVEGF-R1:PlGF ratio shortly before death resemble pathophysiological changes thought to play a causative part in pre-eclampsia. Aggressive critical-care treatment with renal replacement therapy and use of the molecular absorbent recirculation system appeared able to stabilise--at least temporarily--the patient's condition.
- SourceAvailable from: dtic.mil[show abstract] [hide abstract]
ABSTRACT: Effective countermeasures are urgently needed to prevent and treat infections caused by highly pathogenic and biological threat agents such as Marburg virus (MARV). We aimed to test the efficacy of a replication-competent vaccine based on attenuated recombinant vesicular stomatitis virus (rVSV), as a postexposure treatment for MARV haemorrhagic fever. We used a rhesus macaque model of MARV haemorrhagic fever that produced 100% lethality. We administered rVSV vectors expressing the MARV Musoke strain glycoprotein to five macaques 20-30 min after a high-dose lethal injection of homologous MARV. Three animals were MARV-positive controls and received non-specific rVSV vectors. We tested for viraemia, undertook analyses for haematology and serum biochemistry, and measured humoral and cellular immune responses. All five rhesus monkeys that were treated with the rVSV MARV vectors as a postexposure treatment survived a high-dose lethal challenge of MARV for at least 80 days. None of these five animals developed clinical symptoms consistent with MARV haemorrhagic fever. All the control animals developed fulminant disease and succumbed to the MARV challenge by day 12. MARV disease in the controls was indicated by: high titres of MARV (10(3)-10(5) plaque-forming units per mL); development of leucocytosis with concurrent neutrophilia at end-stage disease; and possible damage to the liver, kidney, and pancreas. Postexposure protection against MARV in non-human primates provides a paradigm for the treatment of MARV haemorrhagic fever. Indeed, these data suggest that rVSV-based filoviral vaccines might not only have potential as preventive vaccines, but also could be equally useful for postexposure treatment of filoviral infections.The Lancet 05/2006; 367(9520):1399-404. · 39.06 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: An outbreak of Marburg hemorrhagic fever was first observed in a gold-mining village in northeastern Democratic Republic of the Congo in October 1998. We investigated the outbreak of Marburg hemorrhagic fever most intensively in May and October 1999. Sporadic cases and short chains of human-to-human transmission continued to occur until September 2000. Suspected cases were identified on the basis of a case definition; cases were confirmed by the detection of virus antigen and nucleic acid in blood, cell culture, antibody responses, and immunohistochemical analysis. A total of 154 cases (48 laboratory-confirmed and 106 suspected) were identified (case fatality rate, 83 percent); 52 percent of cases were in young male miners. Only 27 percent of these men reported having had contact with other affected persons, whereas 67 percent of patients who were not miners reported such contact (P<0.001). Most of the affected miners (94 percent) worked in an underground mine. Cessation of the outbreak coincided with flooding of the mine. Epidemiologic evidence of multiple introductions of infection into the population was substantiated by the detection of at least nine genetically distinct lineages of virus in circulation during the outbreak. Marburg hemorrhagic fever can have a very high case fatality rate. Since multiple genetic variants of virus were identified, ongoing introduction of virus into the population helped perpetuate this outbreak. The findings imply that reservoir hosts of Marburg virus inhabit caves, mines, or similar habitats.New England Journal of Medicine 08/2006; 355(9):909-19. · 51.66 Impact Factor
- New England Journal of Medicine 09/2006; 355(9):866-9. · 51.66 Impact Factor
www.thelancet.com/infection Published online March 5, 2012 DOI:10.1016/S1473-3099(12)70018-X 1
March 5, 2012
Department of Intensive Care
(J van Paassen MD,
M S Arbous MD), Department of
(M P Bauer MD, L G Visser MD,
Prof J T van Dissel MD),
(Prof B van Hoek MD), and
Department of Medical
(N D van Burgel MD,
A C T M Vossen MD), Leiden
University Medical Center,
Leiden, Netherlands; Bernhard
Nocht Institute for Tropical
Medicine, Hamburg, Germany
(J Schmidt-Chanasit MD,
S Ölschläger MSc, T Rieger MSc,
P Emmerich PhD,
C Schmetz MSc,
Prof S Günther MD);
Medizinische Klinik II, Klinikum
der Johann Wolfgang Goethe
Universität, Frankfurt am
Main, Germany (S Schilling MD);
Department of Internal
Medicine, Elkerliek Hospital,
(F van de Berkmortel MD);
Atrium MC, Heerlen,
(F van de Berkmortel); and
Department of Virology,
Erasmus MC, Rotterdam,
(Prof A D Osterhaus PhD)
Prof Jaap T van Dissel,
Department of Infectious
Diseases, C5-P, Leiden University
Medical Center, Leiden, 2300 RC,
Acute liver failure, multiorgan failure, cerebral oedema, and
activation of proangiogenic and antiangiogenic factors in a
case of Marburg haemorrhagic fever
Judith van Paassen, Martijn P Bauer, M Sesmu Arbous, Leo G Visser, Jonas Schmidt-Chanasit, Stefan Schilling, Stephan Ölschläger, Toni Rieger,
Petra Emmerich, Christel Schmetz, Franchette van de Berkmortel, Bart van Hoek, Nathalie D van Burgel, Albert D Osterhaus, Ann C T M Vossen,
Stephan Günther, Jaap T van Dissel
A woman developed Marburg haemorrhagic fever in the Netherlands, most likely as a consequence of being exposed
to virus-infected bats in the python cave in Maramagambo Forest during a visit to Uganda. The clinical syndrome was
dominated by acute liver failure with secondary coagulopathy, followed by a severe systemic infl ammatory response,
multiorgan failure, and fatal cerebral oedema. A high blood viral load persisted during the course of the disease. The
initial systemic infl ammatory response coincided with peaks in interferon-γ and tumour necrosis factor-α concentrations
in the blood. A terminal rise in interleukin-6, placental growth factor (PlGF), and soluble vascular endothelial growth
factor receptor-1 (sVEGF-R1) seemed to suggest an advanced patho physiological stage of Marburg haemorrhagic fever
associated with vascular endothelial dysfunction and fatal cerebral oedema. The excess of circulating sVEGF-R1 and
the high sVEGF-R1:PlGF ratio shortly before death resemble pathophysiological changes thought to play a causative
part in pre-eclampsia. Aggressive critical-care treatment with renal replacement therapy and use of the molecular
absorbent recirculation system appeared able to stabilise—at least temporarily—the patient’s condition.
A 41-year-old woman was admitted on July 5, 2008, to a
Dutch hospital with a 3-day history of fever, rigors,
myalgia, and headache, which developed 4 days after
returning from a 3-week journey through Uganda.
During the journey, the woman observed wild chimp-
anzees and gorillas at close range. She visited a cave not
inhabited by bats and—along with three other
travellers—the python cave in Maramagambo Forest,
which is inhabited by the bat species Rousettus
aegyptiacus, 16 days and 13 days before disease onset,
respectively. She was not bitten or scratched by the bats.
9 days before disease onset, she visited a Twa (pygmy)
village, where she witnessed a severely ill elderly woman
from a distance; 1 day later, she swam in a lake. None of
her eight fellow travellers or two guides had fallen ill.
Before her journey, she was vaccinated against yellow
fever, hepatitis A, and typhoid, and she had been taking
prophylaxis against malaria.
On admission, physical examination did not show
abnormalities except for a relative bradycardia (pulse
80 beats per min and temperature 40°C). Blood
exam ination showed leucopenia (3·3×10⁹/L), thrombo-
cytopenia (72×10⁹/L), and raised aspartate amino-
transferase (139 U/L) and alanine aminotransferase
(141 U/L) con centrations (fi gure 1). A chest radiograph
was normal and malaria was excluded. Despite previous
vaccination, typhoid fever was deemed the most likely
cause for this non-specifi c febrile illness characterised
by relative bradycardia,
thrombocytopenia, and mild hepatitis, and intravenous
ceftriaxone was started. The diff erential diagnosis
included leptospirosis and viral infections such as
dengue fever. On day 2 of her stay in hospital, the
patient developed nausea, vomiting, and watery
mild leucopenia and
diarrhoea. On day 3, blood tests suggested acute liver
failure. Hepatitis E, hepatitis A with previous vaccine
failure, and herpes simplex virus were con sidered, as
were viral haemorrhagic fevers (Marburg, Ebola, and
Crimean-Congo haemorrhagic fevers, Rift Valley fever,
yellow fever with previous vaccine failure, and dengue
The patient, who had so far not been isolated on the
ward, was transferred to Leiden University Medical
Center (LUMC; Leiden, Netherlands) under strict
isolation conditions and admitted directly to the
intensive care unit, where strict isolation was continued
according to the local viral haemorrhagic fever protocol.
This protocol dictates that the patient must remain in a
low-air-pressure isolation room with an anteroom for
changing gowns, gloves, and other items; patients may
leave the room only if strictly necessary for life-saving
procedures that cannot be done at point-of-care (eg,
cardiac catheterisation); and all hospital personnel who
enter the patient’s room must wear disposable gowns,
headwear, plastic removable footwear covers, gloves,
FFP2 masks, and eye protection. Blood sampling is kept
to a minimum and done with a closed system,
transported to the microbiological laboratory by a
dedicated courier, processed and in activated in a
biosafety level 3 laboratory by heat (1 h at 60°C) or
chaotropic salt (guanidine isothiocyanate), and then
distributed to other laboratories.
assessments (eg, coagulation, electrolytes, and blood
gases) were done by a point-of-care apparatus that
remained in the patient’s room. Waste and other
material that left the patient’s room was kept separate
and autoclaved by a biological safety offi cer.
Upon arrival at LUMC at 0100 h on July 8, 2008, the
patient mentioned a non-productive cough, progressive
www.thelancet.com/infection Published online March 5, 2012 DOI:10.1016/S1473-3099(12)70018-X
dys pnoea, abdominal pain located centrally, and
progressive loss of hearing. She seemed alert and had a
temperature of 40°C, blood pressure of 88/48 mm Hg,
and a relative bradycardia of 72 beats per min (fi gure 1).
Examination revealed bilateral conjunctivitis with
conjunctival haemorrhage, a dry dark-red oral mucosa
with palatal petechiae, and a faint papular rash of the
face, abdomen, and back. Serum γ-glutamyltransferase
was 394 U/L, alkaline phosphatase 301 U/L, aspartate
amino transferase 10 050 U/L, alanine aminotransferase
6300 U/L, lactate dehydrogenase 15 500 U/L, amylase
311 U/L, and creatine kinase 919 U/L, with a low
bilirubin concentration of 40 μmol/L, which was
predominantly conjugated. Serum albumin was 32 g/L,
glucose 1·6 mmol/L, ammonia 163 μmol/L, and arterial
lactate 5·2 mmol/L. The activated partial thromboplastin
time had risen to 101 s and prothrombin time to 58 s,
whereas the fi brinogen concentration had dropped to
Figure 1: Haemodynamic, infl ammatory, hepatic, and coagulation parameters from admission to death
Mean values are shown at some timepoints. APPT=activated partial thromboplastin time. MARS=molecular absorbent recirculation system. FFP=fresh frozen plasma.
Maramagambo Forest, Uganda: contact with Rousettus aegyptiacus
Returned from 3-week trip to Uganda
Alkaline phosphatase (U/L)
Blood pressure (mm Hg)
Pulse (beats per min)
Prothrombin time (s)
APPT (s) 33
3 × platelets
20 × FFP
4 × red blood cells
2 × platelets
12 × FFP
5 × red blood cells
1 × platelets
8 × FFP
5·2 25 18 14
Fluid balance +10 L +13 L +9 L
Continuous venovenous haemofiltration
Leucocyte count (×109/L)
Aspartate aminotransferase (U/L)
Alanine aminotransferase (U/L)
www.thelancet.com/infection Published online March 5, 2012 DOI:10.1016/S1473-3099(12)70018-X 3
0·4 g/L (fi gure 1), with a raised D-dimer level of
Abdominal ultrasonography showed normal density
of the liver parenchyma, bile duct diameters, and fl ow
in the hepatic blood vessels. The pancreas appeared
swollen. PCR and serological testing (antigen, IgM, and
IgG) did not show evidence of infection with hepatitis A,
B, C, or E viruses, HIV, or parvovirus B19, or reactivation
of cytomegalovirus, herpes simplex virus 1 or 2, or
varicella zoster virus.
Reverse transcriptase (RT)-PCRs specifi c for the
fi lovirus genus and Marburg virus were positive,1,2
whereas tests for Ebola haemorrhagic fever, Lassa
haemorrhagic fever, Crimean-Congo haemorrhagic fever,
yellow fever, Rift Valley fever, dengue fever, chikungunya
virus, Hanta virus, and West Nile virus were negative.
Antibodies (IgM and IgG) to Marburg virus were not
detected at any time by immunofl uorescence assay.
A PCR for Epstein-Barr virus showed low-grade viraemia
of 4480 copies per mL plasma (Epstein-Barr-virus IgM
negative and IgG positive).
The clinical situation deteriorated rapidly shortly after
arrival at LUMC and a deep shock ensued, necessitating
intense haemodynamic support. A positive fl uid balance
(+10 L on the fi rst day after arrival at LUMC, +13 L on
the second day, and +9 L on the third day) needed
supplementation by high doses of norepinephrine and
epinephrine. In 3 days, 40 units of fresh frozen plasma
(300 mL each), six units of thrombocytes (of fi ve donors
each), and nine units of packed erythrocytes (about
270 mL each) had to be given (fi gure 1). The patient had
respiratory failure due to metabolic acidosis and adult
respiratory distress syndrome and needed mechanical
ventilation half a day after admission to LUMC. The
next day, continuous venovenous haemofi ltration in
combin ation with a molecular absorbent recirculation
system (MARS) was started because the patient
developed anuria and liver failure. Rhabdomyolysis
progressed (creatine kinase >23 000 U/L), resulting in
hyperphosphataemia and hyperkalaemia. Intravenous
administration of glucose and insulin, bicarbonate, and
high-dose calcium in combination with mechanical
hyperventilation were necessary to counter the
Although severe spontaneous bleeding did not occur,
there was prolonged oozing of blood at puncture sites of
venous and arterial catheters. Cerebral oedema was
diagnosed by transcranial doppler testing, and con-
tinuous intravenous administration of mannitol and
hypertonic saline was added. On July 10, the patient’s
condition seemed to stabilise because we were able to
reduce the frequency of blood transfusions and the rates
of electrolyte, mannitol, and hypertonic saline admin-
istration. However, the vasoplegic state ensued (fi gure 1).
Suddenly, dilated, non-responsive pupils developed.
Transcranial doppler testing showed no signs of cerebral
blood fl ow, indicating cerebral herniation. In view of the
unfavourable prognosis, treatment was discontinued,
and the patient died the next morning, on day 9 of her
illness. An autopsy was not done. Afterwards, the room,
including non-disposables and apparatus left inside,
were sterilised by hydrogen peroxide fuming.
Review of the patient’s diagnosis and
experimental fi ndings
Ebola and Marburg haemorrhagic fever viruses (family
Filoviridae) are endemic in sub-Saharan Africa and cause
sporadic cases as well as outbreaks.3 Marburg virus was
discovered after outbreaks of Marburg haemorrhagic
fever in Marburg and Frankfurt, Germany, and Belgrade,
Serbia (formerly Yugoslavia), in 1967, which were caused
by infected macaques imported from Uganda.4,5 The
largest outbreaks of Marburg haemorrhagic fever in
Africa occurred in Durba and Watsa in the north of the
Democratic Republic of the Congo from 1998 to
2000 among miners and in Uige, Angola, in 2005, with
case fatality rates between 80% and 90%.6–8 Additionally, a
few isolated cases of Marburg fever have been reported in
people who had travelled to Africa, some of whom had
visited caves.9–11 Because of the close epidemiological link
between Marburg virus infections and caves,12 bats
infesting the caves have long been suspected of serving as
a reservoir for the virus. This speculation has recently
been confi rmed by detection of Marburg virus RNA and
infectious virus in fruit bats of the species R aegyptiacus
sampled in Uganda,13 Kenya,14 the Democratic Republic of
the Congo,15 and Gabon.16 With the exception of the
original outbreak of Marburg haemorrhagic fever in
1967 and laboratory-acquired cases,17 no symptomatic
fi lovirus infections have been described in Europe. The
rare events of Marburg virus infections that occur outside
of endemic areas off er the possibility to investigate clinical
complications, effi cacy of state-of-the-art treat ment, and
The probable source of infection in our patient was
exposure to Marburg virus-infected bats or bat dung in
the python cave in Uganda. Marburg virus remains
stable in the environment even after drying and can
be transmitted by aerosol,18,19 raising the possibility
of transmission via aerosolised body fl uids or dried
excreta of the bats. Alternatively, bat dung might have
contaminated one of several mosquito bites on our
patient, who wore shorts when visiting the python cave.
Inoculation with the virus during the visit to the cave
implies an incubation period of 13 days, which is longer
than the incubation periods of 5–9 days during the
Marburg haemorrhagic fever outbreak in Germany
1967.4,5 The presumed infection source is in agreement
with the case of an American traveller, who had
recovered from an unexplained febrile illness after
visiting the same cave, and who was retrospectively
tested positive for Marburg haemorrhagic fever once
our case became known.20
www.thelancet.com/infection Published online March 5, 2012 DOI:10.1016/S1473-3099(12)70018-X
When a viral haemorrhagic fever is suspected, measures
should be taken to prevent spread of the infection, which
can be challenging in low-resource settings. As soon as a
viral haemorrhagic fever was suspected, our patient was
isolated according to local viral haemorrhagic fever
protocol. The US Centers for Disease Control and
Prevention recommend the following minimum bio-
safety and biocontainment requirements that should also
be applicable in low-resource settings: the patient should
be isolated, preferably in a single room with adjoining
toilet and separate anteroom as a changing area for
hospital personnel; personnel must wear a scrub suit,
gown, and apron, two pairs of gloves, a mask, headcover,
eyewear, and rubber boots; all non-disposable materials
that have come into contact with the patient should be
cleaned with soap and disinfected with bleach; waste
should be disposed of in a place set aside for this purpose
and burnt at least daily; laboratory workers working with
patient material should wear protective clothing; and
bodies of deceased patients should be sprayed with
bleach, sealed in a body bag that is also sprayed with
bleach and buried with the same precautions as are taken
in the hospital.21 Additionally, all relevant authorities
should be notifi ed when a case of viral haemorrhagic
fever is suspected and contact transmissions should be
monitored. The public health response to the present
case has been described elsewhere.22
Clinical presentation and treatment
Our patient showed typical symptoms of Marburg
haemorrhagic fever, including headache, myalgia,
abdominal pain, nausea and vomiting, diarrhoea,
lethargy, relative bradycardia, and a maculopapular
rash.4–9 Signs of coagulation dysfunction, such as
petechiae, conjunctival injection, and prolonged
bleeding at injection sites, were also observed. The
terminal illness was dominated by acute liver failure
and a severe systemic infl ammatory response resulting
in an ensuing vaso plegic state, followed by multiple
organ failure and brain oedema. Apart from an absence
of hyper bilirubinaemia, the patient had signs of acute
liver failure including hypoglycaemia and hyperammo-
naemia. The blood chemistry values suggested injury of
liver, kidney, and pancreas tissue and are in agreement
with measurements in patients with fatal Marburg or
Ebola haemorrhagic fever.4,5,9,23 There were high
concentrations of lactate dehydrogenase and creatine
kinase, suggesting generalised and muscular tissue
damage. The clinical chemistry was consistent with
histopathological fi ndings in human beings with
Marburg haemorrhagic fever: severe renal parenchymal
damage accompanied by signs of tubular failure, and
widespread necroses that apparently fi rst aff ect the liver
and lymphatic system and later occur in the pancreas,
gonads, adrenal glands, hypophysis, thyroid, kidneys,
and skin.24 Macrophages, pancreatic islet cells,
hepatocytes, adrenal cells, and fi broblast-like cells serve
as important sites of Marburg virus replication in
Increases of D-dimer concentrations suggest that
disseminated intravascular coagulopathy (DIC), a hall-
mark of fi lovirus infection in human beings and non-
human primates,9,23 was present already early in the
disease in our patient. The development of DIC has been
linked to overexpression of tissue factor—an initiator of
the coagulation cascade—by fi lovirus-infected macro-
phages.26 The relevance of the tissue factor–DIC pathway
is also shown by the partial treatment eff ects obtained
with recombinant nematode anticoagulant protein c2
(rNAPc2), an inhibitor of the tissue factor pathway, in the
non-human primate model of fi lovirus infection.27 The
serious coagulopathy in our patient seemed to result also
from liver failure, because thrombocytopenia was initially
not prominent and after six thrombocyte transfusions
the platelet count was maintained above 100×10⁹/L.
However, because of the ongoing coagulopathy, large
amounts of fresh frozen plasma had to be given.
Cerebral oedema as a feature of hepatic encephalopathy
and infl ammation-related impairment of the blood-brain
barrier was probably the cause of death in this patient,28
although intracranial haemorrhage cannot be completely
ruled out; no lateralisation was present on neurological
examination. Terminal CNS manifestations including
coma were also frequent among the patients with a
severe course of Marburg haemorrhagic fever in the
outbreaks in 1967.4,5 The coma in these patients clinically
did not correspond to hepatic encephalopathy; the
condition rather resembled an apallic syndrome.4 As
with our case, the CNS complications were assumed to
be the cause of the lethal outcome, and post-mortem
pathological investigation revealed marked cerebral
oedema with histopathological signs of encephalitis or
haemorrhagic diathesis.24,29 By contrast, terminal coma
was rare (2%) among the patients with Marburg
haemorrhagic fever who died during the outbreak in
Durba (no late-stage data are available for the Uige
outbreak), while haemorrhage was more frequent and
the case fatality rate was higher than in the outbreak in
Europe.7,8 Whether these diff erences in clinical
presentation and case fatality rate are associated with
diff erences in virus strains, host genetics, or the quality
of supportive care remains unknown.
As yet, no causal treatment is available for Marburg
haemorrhagic fever. Advances have been made in the
development of a vaccine, which uses a recombinant
vesicular stomatitis virus that expresses Marburg virus
protein.30 We considered administering this experimental
vaccine to the patient’s husband as post-exposure
prophylaxis, but he refused after weighing the risk and
potential benefi ts.
Virus genomics and diagnostics
Marburg viruses exhibit a high genetic diversity with
up to 22% nucleotide diff erence.31 The sequences do
www.thelancet.com/infection Published online March 5, 2012 DOI:10.1016/S1473-3099(12)70018-X 5
not show a geographical clustering because the bat
population inhabiting a cave may harbour Marburg
viruses of various lineages.13,15 Accordingly, the virus
strains isolated from the miners who presumably
became infected while working in the Durba mine
diff er from case to case.8,12 To characterise the virus
strain that infected our patient (called Leiden strain),
we isolated the virus from blood in Vero cells in the
biosafety level 4 (BSL-4) facility of the Bernhard-Nocht-
Institute (Hamburg, Germany) and completely se-
quenced it (fi gure 2). Additionally, the virus circulating
in blood was completely sequenced without previous
isolation in cell culture. The sequences from blood and
cell culture were 100% identical (GenBank accession
number JN408064). Characteristic genomic features
with respect to coding sequences and regulatory
elements were conserved. Phylogenetic analysis of full-
length genomes showed a close relation of the Leiden
strain to the Popp strain from the outbreak in Germany
1967, which also originated from Uganda (fi gure 2).
Virus load was measured retrospectively in our patient
by real-time RT-PCR and by in-vitro transcribed
Marburg virus RNA as a quantifi cation standard.1 Virus
load was already high on the day of admission, reached
a plateau on July 7 (1·5×10⁸ RNA copies per mL), and
hardly declined thereafter (fi gure 3). The high
concentration of viraemia early in the disease is similar
to that reported in fatal cases of Ebola haemorrhagic
fever and in retrospect could have been considered a
predictor of poor outcome in our patient.32 Because of
the high virus load, we were able to readily identify
characteristic Marburg virus particles by electron
microscopy in serum (fi gure 2). The absence of
detection of IgM and IgG antibodies to Marburg virus
could be interpreted as a consequence of fi lovirus-
induced immunosuppression.33 IgM and IgG are often
absent in patients with Ebola haemorrhagic fever at the
time of death.32,33
Several factors seem to play a part in the pathogenesis
of fi lovirus haemorrhagic fever, including virus-
induced immunosuppression,33 pantropism,25 DIC,26
and endothelial dysfunction.34 Endothelial function can
be impaired during fi lovirus infection via direct and
indirect mechanisms. Marburg virus replicates in
endothelial cells, raising the possibility of direct cell
damage,35 and in experiments with Ebola virus-like
particles, the virus glycoprotein activated endothelial
cells and impeded their barrier function.36 Additionally,
Marburg virus-infected macrophages secrete pro-
infl ammatory cytokines such as tumour necrosis factor
(TNF)α that disturb the barrier function of the
endothelium in vitro.37–39 Raised con centrations of TNFα,
interleukin-6, and interleukin-10 have been reported in
patients with Ebola haemorrhagic fever.40–42 Also, a
marked rise in interleukin-6 precedes death in Marburg
and Ebola virus-infected non-human pri mates,19,43
whereas animals who survive as a result of rNAPc2
treatment show markedly lower interleukin-6 con-
centrations. To obtain further insights into the
expression of proinfl ammatory cytokines and endothelial
dysfunction in Marburg haemorrhagic fever, a range of
cytokines and vascular endothelial growth factors and
their soluble receptors were measured in blood samples
taken from our patient during the course of her illness.
These experiments were done in the BSL-4 facility of the
Bernhard-Nocht-Institute with Quantikine Immuno-
assays (R&D Systems Europe, Abingdon, UK), according
to the manufacturer’s instructions. The results of the
viral load and cytokine measurements were available
after the patient’s death.
Diff erent kinetics of serum cytokines were noted
(fi gure 3). Interleukin-10 concentrations peaked earliest,
on the second day of admission, and then declined
(maximum 88 times higher than threshold), whereas
interferon-γ and TNFα were slightly above threshold on
July 5, and did not peak until July 7–8 (maximum:
interferon-γ 47 times higher, TNFα six times higher),
and thereafter decreased to baseline values. The peak in
these two cytokines coincided with the rapid clinical
and haemodynamic deterioration observed shortly after
transfer to LUMC. By contrast, concentrations of
interleukin-6, placental growth factor (PlGF; a member
of the vascular endothelial growth factor [VEGF] family),
and the soluble form of VEGF receptor-1 (sVEGF-R1 or
soluble fms-like tyrosine kinase-1 [sFlt-1],44 an antagonist
to VEGF and PlGF) were mostly below threshold until
July 8. They sharply increased and peaked on July 9
(maximum: interleukin-6 305 times higher than
threshold, PlGF nine times higher, sVEGF-R1 212 times
higher), and decreased somewhat shortly before death
Figure 2: Electron microscopy and phylogenetic analysis
(A) Typical Marburg virus particle in serum from the patient. Virus particles were pelleted by centrifugation and
subjected to negative stain electron microscopy. (B) Phylogeny of the Marburg virus clade, including the new
Leiden strain, based on full-length genome sequences. The unrooted tree was inferred by the Bayesian Markov
chain Monte Carlo method in MrBayes software. Posterior support values calculated by the BEAST program are
shown on the branches. They show the reliability of the phylogenetic reconstruction and range from 0 to 1; the
maximum of 1 means that the topology of the tree is very reliable.
0·1 nucleotide substitutions per site
www.thelancet.com/infection Published online March 5, 2012 DOI:10.1016/S1473-3099(12)70018-X
of the patient on July 10. These cytokines peaked during
the extreme vasoplegic haemodynamic state, before
fatal cerebral oedema developed. Angiopoietin-2 was
above threshold only on July 9 (three times higher).
Throughout the course of disease, interleukin-12
(75 kDa dimer) remained in the normal range.
The decrease in interferon-γ and TNFα concentrations
and the concomitant increase in interleukin-6,
sVEGF-R1, and PlGF concentrations noted in our case
seem to herald an advanced stage in Marburg virus
patho physiology characterised by activation of both pro-
angiogenic and antiangiogenic factors. This change
in infl ammatory response was followed by cerebral
deterioration and death in this patient. The terminal
rise in interleukin-6, PlGF, and sVEGF-R1 con-
centrations, which coincided with ongoing need for
intense vaso pressor support, probably represents
serious endothelial dysfunction, increased vascular
permeability, and nitric oxide-mediated vasoplegia.
These markers of vascular damage might be directly
involved in development of cerebral oedema and lethal
CNS injury.44,45 In particular, the excess of circulating
sVEGF-R1 and the high sVEGF-R1:PlGF ratio (rising
from 170 [July 9] to 250 and 590 [July 10] shortly before
death) resemble patho physiological changes thought to
play a causative part in pre-eclampsia—a multiorgan
Search strategy and selection criteria
References were identifi ed through searches of PubMed
without limitations in language, with the terms “(marburg
haemorrhagic fever OR marburg virus OR fi lovir*)” in
combination with “case” from 1970 to December, 2011;
without any other term before 1970 for information on
previous case descriptions; and without any other term from
2006 to December, 2011, for information on pathogenesis.
We also consulted the table of known cases and outbreaks of
Marburg haemorrhagic fever on the Centers for Disease
Control and Prevention website.
Figure 3: Cytokines, vascular growth factors, and virus load
(A) Interleukin-10, (B) interferon-γ, (C) tumour necrosis factor-α, (D) interleukin-6, (E) placenta growth factor, (F) soluble vascular endothelial growth factor receptor 1, (G) interleukin-12, and
(H) angiopoietin-2 concentrations, and (I) Marberg virus load were measured.
Tumour necrosis factor-α (pg/mL)
Soluble vascular endothelial
growth factor receptor 1 (pg/mL)
Placenta growth factor (pg/mL)
Interleukin-12 75-kDa dimer (pg/mL)
Marberg virus load (copies per mL)
Normal <7·8 pg/mLNormal <15 pg/mLNormal <15 pg/mL
Normal <15 pg/mLNormal <25 pg/mLNormal <180 pg/mL
Normal <7·8 pg/mL Normal <8000 pg/mL
For the CDC website see http://
www.thelancet.com/infection Published online March 5, 2012 DOI:10.1016/S1473-3099(12)70018-X 7
disease associated with vascular endothelial dysfunction
and cerebral oedema.46,47 Alternatively, a shift in viral
tropism, including infection of the brain, as has been
reported at a late stage in Ebola virus-infected non-
human primates that showed prolonged survival as a
result of treatment, might explain the biphasic clinical
course and the profi le of cytokines and markers of
To our knowledge, this is the fi rst time that sVEGF-R1
and PlGF have been studied in the context of a fi lovirus
infection. Research into the relevance of these factors
and the associated molecular mechanisms in
appropriate animal models might potentially lead to
new intervention concepts for the terminal stage of
fi lovirus infections.
JvP and MPB took daily care of the patient; both contributed equally to
the manuscript. MSA and LGV also contributed to patient management.
BvH overviewed the molecular absorbent recirculation system. JS-C and
SÖ did the fi rst-line Marburg virus diagnostics, sequenced the Marburg
virus, and assessed the phylogeny. PE did the virus isolation together
with SÖ, and assessed viral load and serology. SS and TR did the
cytokine blood assessments. CS did the electron microscopy of the virus
in the blood. FvdB took care of the patient in the referring hospital, until
transfer to the LUMC. NDvB, ADO, and ACTMV were responsible for
diagnostics (eg, diff erential diagnosis) and sample handling at LUMC.
The data analysis and a fi rst draft of the paper was prepared by JTvD and
SG, who also supervised the laboratory experiments in Germany.
JTvD and SG contributed equally to the manuscript. All other authors
commented on and contributed to the fi nal version of the manuscript.
Confl icts of interest
We declare that we have no confl icts of interest.
We thank all the hospital workers who were involved in the care of this
patient, in particular the intensive care unit nursing staff and hospital
hygiene staff of the Elkerliek Hospital and the LUMC. Isolation and
sequencing of Marburg virus strain Leiden was supported by grant
228292 (European Virus Archive) from the European Community. The
funding source was not involved in the design or implementation of
1 Gibb TR, Norwood DA Jr, Woollen N, Henchal EA. Development
and evaluation of a fl uorogenic 5’-nuclease assay to identify
Marburg virus. Mol Cell Probes 2001; 15: 259–66.
2 Panning M, Laue T, Olschlager S, et al. Diagnostic reverse-
transcription polymerase chain reaction kit for fi loviruses based on
the strain collections of all European biosafety level 4 laboratories.
J Infect Dis 2007; 196 (suppl 2): S199–204.
3 Feldmann H. Marburg hemorrhagic fever—the forgotten cousin
strikes. N Engl J Med 2006; 355: 866–69.
4 Martini GA, Knauff HG, Schmidt HA, Mayer G, Baltzer G. [On the
hitherto unknown, in monkeys originating infectious disease:
Marburg virus disease]. Dtsch Med Wochenschr 1968;
93: 559–71 (in German).
5 Stille W, Bohle E, Helm E, van Rey W, Siede W. [On an infectious
disease transmitted by Cercopithecus aethiops. (“Green monkey
disease”)]. Dtsch Med Wochenschr 1968; 93: 572–82 (in German).
6 Roddy P, Thomas SL, Jeff s B, et al. Factors associated with Marburg
hemorrhagic fever: analysis of patient data from Uige, Angola.
J Infect Dis 2010; 201: 1909–18.
7 Colebunders R, Tshomba A, Van Kerkhove MD, et al. Marburg
hemorrhagic fever in Durba and Watsa, Democratic Republic of the
Congo: clinical documentation, features of illness, and treatment.
J Infect Dis 2007; 196 (suppl 2): S148–53.
8 Bausch DG, Nichol ST, Muyembe-Tamfum JJ, et al. Marburg
hemorrhagic fever associated with multiple genetic lineages of
virus. N Engl J Med 2006; 355: 909–19.
9 Gear JS, Cassel GA, Gear AJ, et al. Outbreak of Marburg virus
disease in Johannesburg. BMJ 1975; 4: 489–93.
10 Smith DH, Johnson BK, Isaacson M, et al. Marburg-virus disease in
Kenya. Lancet 1982; 1: 816–20.
11 Johnson ED, Johnson BK, Silverstein D, et al. Characterization of a
new Marburg virus isolated from a 1987 fatal case in Kenya.
Arch Virol Suppl 1996; 11: 101–14.
12 Bausch DG, Borchert M, Grein T, et al. Risk factors for Marburg
hemorrhagic fever, Democratic Republic of the Congo.
Emerg Infect Dis 2003; 9: 1531–37.
13 Towner JS, Amman BR, Sealy TK, et al. Isolation of genetically
diverse Marburg viruses from Egyptian fruit bats. PLoS Pathog
2009; 5: e1000536.
14 Kuzmin IV, Niezgoda M, Franka R, et al. Marburg virus in fruit bat,
Kenya. Emerg Infect Dis 2010; 16: 352–54.
15 Swanepoel R, Smit SB, Rollin PE, et al. Studies of reservoir hosts
for Marburg virus. Emerg Infect Dis 2007; 13: 1847–51.
16 Towner JS, Pourrut X, Albarino CG, et al. Marburg virus infection
detected in a common African bat. PLoS One 2007; 2: e764.
17 Emond RT, Evans B, Bowen ET, Lloyd G. A case of Ebola virus
infection. BMJ 1977; 2: 541–44.
18 Piercy TJ, Smither SJ, Steward JA, Eastaugh L, Lever MS. The
survival of fi loviruses in liquids, on solid substrates and in a
dynamic aerosol. J Appl Microbiol 2010; 109: 1531–39.
19 Alves DA, Glynn AR, Steele KE, et al. Aerosol exposure to the
angola strain of marburg virus causes lethal viral hemorrhagic fever
in cynomolgus macaques. Vet Pathol 2010; 47: 831–51.
20 Anonymous. Imported case of Marburg hemorrhagic fever—
Colorado, 2008. MMWR Morb Mortal Wkly Rep 2009; 58: 1377–81.
21 Centers for Disease Control and Prevention (CDC). Update:
management of patients with suspected viral hemorrhagic fever—
United States. MMWR Morb Mortal Wkly Rep 1995; 44: 475–79.
22 Timen A, Koopmans MP, Vossen AC, et al. Response to imported
case of Marburg hemorrhagic fever, the Netherland. Emerg Infect Dis
2009; 15: 1171–75.
23 Rollin PE, Bausch DG, Sanchez A. Blood chemistry measurements
and D-Dimer levels associated with fatal and nonfatal outcomes in
humans infected with Sudan Ebola virus. J Infect Dis 2007;
196 (suppl 2): S364–71.
24 Gedigk P, Bechtelsheimer H, Korb G. [Pathological anatomy of the
“Marburg virus” disease (the so-called “Marburg monkey disease”)].
Dtsch Med Wochenschr 1968; 93: 590–601 (in German).
25 Geisbert TW, Jaax NK. Marburg hemorrhagic fever: report of a case
studied by immunohistochemistry and electron microscopy.
Ultrastruct Pathol 1998; 22: 3–17.
26 Geisbert TW, Young HA, Jahrling PB, Davis KJ, Kagan E,
Hensley LE. Mechanisms underlying coagulation abnormalities in
ebola hemorrhagic fever: overexpression of tissue factor in primate
monocytes/macrophages is a key event. J Infect Dis 2003;
27 Geisbert TW, Hensley LE, Jahrling PB, et al. Treatment of Ebola
virus infection with a recombinant inhibitor of factor VIIa/tissue
factor: a study in rhesus monkeys. Lancet 2003; 362: 1953–58.
28 Bjerring PN, Eefsen M, Hansen BA, Larsen FS. The brain in acute
liver failure. A tortuous path from hyperammonemia to cerebral
edema. Metab Brain Dis 2009; 24: 5–14.
29 Bechtelsheimer H, Jacob H, Solcher H. [On the neuropathology of the
green monkey (Cercopithecus aethiops) transmitted infectious diseases
in Marburg]. Dtsch Med Wochenschr 1968; 93: 602–04 (in German).
30 Daddario-DiCaprio KM, Geisbert TW, Stroher U, et al. Postexposure
protection against Marburg haemorrhagic fever with recombinant
vesicular stomatitis virus vectors in non-human primates: an
effi cacy assessment. Lancet 2006; 367: 1399–404.
31 Towner JS, Khristova ML, Sealy TK, et al. Marburgvirus genomics
and association with a large hemorrhagic fever outbreak in Angola.
J Virol 2006; 80: 6497–516.
32 Towner JS, Rollin PE, Bausch DG, et al. Rapid diagnosis of Ebola
hemorrhagic fever by reverse transcription-PCR in an outbreak
setting and assessment of patient viral load as a predictor of
outcome. J Virol 2004; 78: 4330–41.
33 Baize S, Leroy EM, Georges-Courbot MC, et al. Defective humoral
responses and extensive intravascular apoptosis are associated with
fatal outcome in Ebola virus-infected patients. Nat Med 1999; 5: 423–26.
www.thelancet.com/infection Published online March 5, 2012 DOI:10.1016/S1473-3099(12)70018-X
34 Aleksandrowicz P, Wolf K, Falzarano D, Feldmann H, Seebach J,
Schnittler H. Viral haemorrhagic fever and vascular alterations.
Hamostaseologie 2008; 28: 77–84.
35 Schnittler HJ, Mahner F, Drenckhahn D, Klenk HD, Feldmann H.
Replication of Marburg virus in human endothelial cells. A possible
mechanism for the development of viral hemorrhagic disease.
J Clin Invest 1993; 91: 1301–09.
36 Wahl-Jensen VM, Afanasieva TA, Seebach J, Stroher U, Feldmann H,
Schnittler HJ. Eff ects of Ebola virus glycoproteins on endothelial cell
activation and barrier function. J Virol 2005; 79: 10442–50.
37 Feldmann H, Bugany H, Mahner F, Klenk HD, Drenckhahn D,
Schnittler HJ. Filovirus-induced endothelial leakage triggered by
infected monocytes/macrophages. J Virol 1996; 70: 2208–14.
38 Bockeler M, Stroher U, Seebach J, et al. Breakdown of
paraendothelial barrier function during Marburg virus infection is
associated with early tyrosine phosphorylation of platelet
endothelial cell adhesion molecule-1. J Infect Dis 2007;
196 (suppl 2): S337–46.
39 Stroher U, West E, Bugany H, Klenk HD, Schnittler HJ,
Feldmann H. Infection and activation of monocytes by Marburg
and Ebola viruses. J Virol 2001; 75: 11025–33.
40 Hutchinson KL, Rollin PE. Cytokine and chemokine expression in
humans infected with Sudan Ebola virus. J Infect Dis 2007;
196 (suppl 2): S357–63.
41 Villinger F, Rollin PE, Brar SS, et al. Markedly elevated levels of
interferon (IFN)-gamma, IFN-alpha, interleukin (IL)-2, IL-10, and
tumor necrosis factor-alpha associated with fatal Ebola virus
infection. J Infect Dis 1999; 179 (suppl 1): S188–91.
42 Baize S, Leroy EM, Georges AJ, et al. Infl ammatory responses in
Ebola virus-infected patients. Clin Exp Immunol 2002; 128: 163–68.
43 Fritz EA, Geisbert JB, Geisbert TW, Hensley LE, Reed DS. Cellular
immune response to Marburg virus infection in cynomolgus
macaques. Viral Immunol 2008; 21: 355–63.
44 Wu FT, Stefanini MO, Mac Gabhann F, Kontos CD, Annex BH,
Popel AS. A systems biology perspective on sVEGFR1: its biological
function, pathogenic role and therapeutic use. J Cell Mol Med 2010;
45 Weis SM, Cheresh DA. Pathophysiological consequences of
VEGF-induced vascular permeability. Nature 2005; 437: 497–504.
46 Maynard SE, Min JY, Merchan J, et al. Excess placental soluble
fms-like tyrosine kinase 1 (sFlt1) may contribute to endothelial
dysfunction, hypertension, and proteinuria in preeclampsia.
J Clin Invest 2003; 111: 649–58.
47 Levine RJ, Maynard SE, Qian C, et al. Circulating angiogenic factors
and the risk of preeclampsia. N Engl J Med 2004; 350: 672–83.
48 Larsen T, Stevens EL, Davis KJ, et al. Pathologic fi ndings associated
with delayed death in nonhuman primates experimentally infected
with Zaire Ebola virus. J Infect Dis 2007; 196 (suppl 2): S323–28.