BRITISH MEDICAL JOURNAL

27 AUGUST 1977

54A1

PAPERS

case

AND

ORIGINALS

of Ebola virus infection

R T D EMOND, BRANDON EVANS, E T W BOWEN, G LLOYD

British Medical Journal, 1977, 2, 541-544

Summary
In November 1976 an investigator at the Microbiological
Research Establishment accidentally inoculated himself
while processing material from patients in Africa who
had been suffering from a haemorrhagic fever of unknown
cause. He developed an illness closely resembling
Marburg disease, and a virus was isolated from his
blood that resembled Marburg virus but was distinct
serologically. The course of the illness was mild and may
have been modified by treatment with human interferon
and convalescent serum. Convalescence was protracted;
there was evidence of bone-marrow depression and virus
was excreted in low titre for some weeks. Recovery was
complete. Infection was contained by barrier-nursing
techniques using a negative-pressure plastic isolator and
infection did not spread to attendant staff or to the

community.
Introduction
In the late summer of 1967 a serious outbreak of an unknown
infectious disease occurred in Germany and Yugoslavia. It
affected 31 people, seven of whom died. A strange new RNA
virus was isolated from the patients, and the source of the
outbreak was traced to vervet monkeys (Cercopithecus aethiops)
imported from Uganda. Since many of the cases were centred on
the West German town of Marburg, the disease was designated
Marburg disease.' The original outbreak subsided and no further
cases were recognised until 1975, when a young man was

Department of Infectious Diseases, Royal Free Hospital, London

R T D EMOND, FRCP, DTM&H, consultant physician
BRANDON EVANS, MB, MRCP, senior registrar
Microbiological Research Establishment, Porton, Salisbury, Wiltshire
E T W BOWEN, FIMLT, head of special pathogens unit
G LLOYD, MSc, member of special pathogens unit

admitted to hospital in South Africa having recently travelled

extensively in Rhodesia. This patient was found to have Marburg
disease and infection spread to his travelling companion and to a
nurse. The original patient died but the other two survived. The
source of the infection was not determined.2
Just over a year later, in July to November 1976, a serious
outbreak of haemorrhagic fever occurred in the Western
Equatoria province of the Sudan and the adjacent Equateur
Region of Zaire.3 Infection spread rapidly among the local
people, particularly within the hospitals. There was an
appallingly high death rate-30-800, in the Sudan4 and 890,,,
in Zaire. In view of the severity of this outbreak specimens were
sent to high-security laboratories in England, Belgium, and the
United States of America for identification of the agent responsible. All three laboratories isolated a virus that resembled
Marburg virus morphologically but was serologically distinct.567
The name Ebola was given to the prototype strain.

Case report
On the 5 November 1976 one of the investigators at the Microbiological Research Establishment, Porton Down, accidentally
pricked his thumb through a protective rubber glove while transferring
homogenised liver from a guinea-pig infected with this new virus.
According to standard safety protocol he immediately removed the
glove and immersed his thumb in hypochlorite solution then squeezed
it vigorously. There was no bleeding and careful examination with a
hand lens failed to reveal a puncture wound. He was kept under
surveillance, and on the sixth day became ill.

CLINICAL COURSE

Shortly after midnight on 11 November his temperature rose to

37 4C. During the early morning he complained of central abdominal
pain and nausea. He did not vomit or have the headache or myalgia
that had been a feature in other cases. Later that day he was seen at the
Microbiological Research Establishment, where a blood sample was
taken before he was transferred to the high-security infectious
diseases unit at Coppetts Wood Hospital and placed in a Trexler
negative-pressure plastic isolator.8
When he was admitted he felt physically exhausted and complained
of anorexia, nausea, and constant central abdominal pain. There
were no other symptoms. His temperature was 38C with a relative
bradycardia. He was alert and did not seem to be particularly ill.

542

BRITISH MEDICAL JOURNAL

Apart from slight abdominal tenderness there were no other abnormal

findings. In view of the hazards to laboratory staff it was considered
unwise to undertake haematological or biochemical studies until the
results of the virological tests were known. Since it appeared highly
probable that the illness was due to this virulent Marburg-like virus,
treatment was started that same evening with human interferon, which
had been prepared by stimulating peripheral lymphocytes with
Sendai virus in vitro.9 Interferon was given by intramuscular injection
in a dose of 3 million units every 12 hours for 14 days.
The next morning his temperature was normal and he was free
from symptoms, but later in the evening his temperature rose again
to 39C. Apart from loss of appetite there were no other symptoms.
By this time direct electron microscopy had shown Marburg-like
virus particles in the patient's blood. In view of this finding it was
thought advisable to give the patient convalescent serum. Since the
new virus was serologically distinct from the original Marburg virus
it was necessary to obtain the serum from people convalescing after
the recent African outbreak. 450 ml serum obtained from Zaire was
heated at 60C for one hour to inactivate virus and tested for hepatitis
B surface antigen and antibody (HBsAg and HBsAb). The serum
was given by slow intravenous infusion over a period of four hours
from 1.30 am on 13 November. Blood samples were taken at frequent
intervals to ascertain virus and antibody levels.
On 13 November the patient had no appetite, but was otherwise
free from symptoms. Examination showed an inflamed throat, but
exudate was not present. Some small lymph nodes were palpable in
the neck and axillae, though these were not tender. A few erythematous
maculopapular lesions were noted on his back over the shoulders. The
muscles were not tender. The cardiovascular system, respiratory
system, and abdomen were normal. Urine was free from protein and
output was

satisfactory.

During the early morning of the fourth day of illness, 14 November,

his temperature fell to normal after a profuse bout of sweating. At this
stage he still felt relatively well and the only change was an extension
of the rash over the chest wall. About midday he had a sudden violent
bout of shivering followed by a sharp rise in temperature to 40C.
This was accompanied by nausea, retching, and a single episode of
vomiting. Since admission he had been constipated, but at this point
he had a loose bowel action. His mental state began to change and over
the next 24 hours there was striking deterioration in concentration and
memory. Protein was detected in his urine for the first time and
persisted thereafter until the fever subsided. Over the next 72 hours,
when the illness was at its height, there was severe malaise and extreme
weakness. Profuse watery diarrhoea developed and continued for two
days accompanied by persistent vomiting. The rash spread to all parts
of his body and ultimately became confluent. There was no bleeding
into the skin or mucous membranes. The throat remained inflamed
and a few small patches of thrush were detected. The abdomen was
slightly distended, but there was no tenderness or guarding. He was
mildly dehydrated and the urinary output was falling. Metoclopramide
was prescribed for the vomiting and Lomotil for the diarrhoea.
On the sixth day of illness, 16 November, a further 330 ml of
convalescent serum from the Sudan, pretreated in the same manner,
was infused and followed by Hartmann's solution to correct the
dehydration. Next day his urinary output fell to its lowest volume of
830 ml despite adequate fluid replacement and a satisfactory blood
pressure. At this stage his appetite began to return, but swallowing
induced pain in the throat and behind the sternum. Examination
showed extensive candidiasis. Diarrhoea and vomiting had become
less frequent and ceased on the 18 November. The thrush responded
to treatment with amphotericin B lozenges and the dysphagia settled
within a few days. The erythematous stage of the rash began to fade
on the 19 November, disclosing a petechial element over the limbs.
On the same day he complained of stiffness of the small joints of his
hands and to a lesser extent of the wrists and knees.
After 20 November his general condition improved. His fever
subsided-to a low level, his energy began to return, and there was
dramatic improvement in his interest and ability to concentrate,
though he could barely recollect the acute phase of his illness. The
joint symptoms did not persist. The temperature returned to normal
on 22 November but there was a further slight flicker of fever on the
next two days, after which the temperature remained normal (see
figure).

Output of urine was normal by 23 November.

At this stage it

decided to take specimens for clearance tests at weekly intervals

and it was arbitrarily agreed that three negative sets of cultures from
throat swab, blood, urine, and faeces would be an acceptable standard
for discharging the patient from isolation. The discovery of virus in
semen was not thought to justify further isolation, especially as the
patient fully appreciated the implications. Subsequently he made an
uneventful but slow recovery over 10 weeks.
was

27 AUGUST 1977

Vomiting
Diar rhoea
Rash

pain

Sore throat

Abdominal

Proteinuria Date

Clinical

course

of disease.

At the end of the acute stage of the illness he had lost a considerable
amount of weight, which he regained slowly during convalescence.
The rate of growth of hair slowed during the acute illness and during
convalescence there was considerable loss of hair from his scalp.
There were no other clinical complications. Electrocardiograms taken
during the acute stage between days 5 and 9 were normal, though the
amplitudes of the T-waves were lower than in a recording made on
27 January during convalescence. Blood urea, and sugar concentrations
and liver function were normal during convalescence. The HBsAg and
HBsAb tests on blood were negative. The result of a chest radiograph
was normal. During the early period of convalescence the haemoglobin
level and the white blood cell counts were depressed and did not
recover fully until 8 February 1977, three months after the onset of
illness (table I).

TABLE I-Results of haematological and biochemical

Haemoglobin (g/dl)
Packed cell volume (O)
Mean cell haemoglobin concentration
(g/dl)
White blood count ( x 10 '/1)
Platelets ( x 10-9/1)
Serum aspartate aminotransferase
(IU/1)
Serum alanine aminotransferase (IU/1)
Alkaline phosphatase (IU/l)
Total proteins (g/l)
Urea (mmol/l)
Sugar (mnuol/l)
z

investigations

11 Jan

8 Feb

15 Feb

11 1
36
31

13 2
40
33

13-2
38
34

3-6
203
<10

4-525
190

<10
35
75
4-37
5-58

72 7

4 275
<10

<10
72-9

Conversion: SI to traditional units-Urea: 1 mmol/l 6 mg/lOO ml. Sugar: 1

18 mg/100 ml.

mmol/1

The Trexler negative-pressure plastic isolator and the techniques

used for disposing of waste proved to be effective in preventing spread
of Ebola virus from the patient to attendant staff and to the general
community. Of the 24 nurses who were directly concerned in the care
of the patient, six became ill with acute respiratory infections, which
lasted on average two days. Four of the five doctors looking after the
patient developed a 'flu-like illness with some gastrointestinal
symptoms. At onset these illnesses caused concern but the problems

BRITISH MEDICAL JOURNAL

543

27 AUGUST 1977

invariably resolved within two or three days and antibody studies later
showed no evidence of Ebola virus infection among either medical or
nursing staff.

VIROLOGICAL INVESTIGATIONS

The first specimen of blood was collected about 14 hours after the
patient became feverish; this was six days after the accident. This blood
specimen was examined by electron microscopy and virus particles
were seen which were similar to those of Ebola virus. Guinea-pigs
inoculated with this blood specimen developed a febrile illness and
electronmicroscope examination of their blood and tissues showed
particles which were again similar to those of Ebola virus. These
observations are consistent with an infection due to Ebola virus.
Virus isolations and serological studies were also made on specimens
of blood collected daily during the acute phase of the illness and on
blood, urine, faeces, throat swabs, and seminal fluid collected during
the convalescent phase. The highest levels of virus in the blood
(104-5 guinea-pig infective units/ml) were recorded on the first and
second day of the illness. After the start of interferon treatment and
serotherapy, the level dropped dramatically to 3-10 guinea-pig
infective units/ml and remained at this level until the viraemia
disappeared on the ninth day of illness (table II).
No virus was isolated from faeces, urine, and throat swabs collected
between days 14 and 27. Ebola virus was, however, isolated from
specimens of seminal fluid collected on days 39 and 61 but not on days
76, 92, and 110.
After the infusion of 450 ml of convalescent serum (fluorescent
antibody titre of 1/128-1/256) on day 2 circulating antibody levels of
1]16 were recorded in the patient's blood from days 3 to 9. This
-increased to 1/32 on day 10 and gradually increased to a fluorescent
antibody titre of 1/128 by day 34. The patient was then subjected to
plasmapheresis between 16 and 25 February 1977. A total of seven
units of plasma was taken, which resulted in the fluorescent antibody
level dropping from 1/128 to 1/32, and a specimen of blood collected
on 5 May 1977 had a fluorescent antibody titre of 1/16.

Discussion

The

of the accident and the absence-of a visible

emphasise the invasiveness of Ebola virus and
the high susceptibility of man. Although the new Ebola virus is
serologically distinct from the original Marburg virus, the pattern
of illness in our patient closely followed the course of Marburg
disease as described in Germany and South Africa. The course
and duration of the illness were similar and a characteristic
clinical syndrome was produced by the exanthem, excessive
fatigue, and considerable gastrointestinal disturbance. There
were, however, some minor differences, notably the absence of
headache and myalgia, which were prominent in Marburg and
Johannesburg. The rash emerged after the standard prodromal
period and had the morbilliform appearance described in the
previous outbreaks of Marburg disease. The evolution of the
rash differed from measles in that the lesions appeared first over
nature

puncture mark

the back and not on the head and neck. A painful throat was not
a feature of the early stage but-developed later when there was
frank evidence of candidiasis.
While the course of the illness was milder than expected from
reports elsew'here, the pattern and duration of symptoms were
not modified. The relatively mild course of the illness and the
absence of haemorrhage might have been determined by
treatment with interferon and convalescent serum, but the
value of these preparations could not be accurately assessed from
experience- with one patient. Treatment was started with interferon 20 hours after the onset of illness and convalescent serum
was first.given 47 hours after onset. There was no obvious
clinical improvement after treatment, but there was a striking
fall in the level of circulating virus. On the first day of illness a
blood sample was found -to contain 104-5 guinea-pig infective
units/ml; on the day after starting treatment with interferon
there was no change in the amount of virus, but on the next day,
after infusion of serum, the level in the- blood dropped to 10 5
guinea-pig infective units/ml. Since there is known to be a time
lag before interferon produces an effect on vi-rus levels it is not
possible to assess the relative effectiveness of the two preparations
in clearing the blood. Subsequently virus was detected in low
titre in the bloodstream throughout the acute stage of the
illness but disappeared on the 9th day of illness, before the
temperature had returned to normal (see figure). The second
infusion of serum had no effect on the amount of virus. The
antibody levels achieved in the patient's blood after infusion
were consistent with the dilution of the convalescent serum
(table II). The oliguria and proteinuria present at the height of
the illness could have been attributed to deposition of immune
complexes in the kidney, especially in view of the transient
arthralgia at the end of the acute stage, but these features were
recorded in severe cases during the original Marburg outbreak,
when no serum was given.
Treatment of the convalescent serum to ensure safety presented serious problems. Marburg virus has been shown to
persist in the body for se-veral months after the acute illness,
though it has not been shown in the circulating blood. Marburg
virus is relatively resistant to heat but is inactivated in serum
maintained at 60C for 60 minutes.10 The Ebola convalescent
serum was therefore treated at this temperature for-60 minutes
to ensure safety. The serum was also tested for-HBsAg and
HBsAb because carriers are common in many parts of tropical
Africa. During convalescence the patient's blood was found to be
negative for HBsAg and HBsAb.
Blood examination during convalescence showed evidence of
bone-marrow depression with a low haemoglobin concentration
and low white blood cell count. These features were shown during
the original outbreak of Marburg disease and were attributed to
the activity of the virus. Interferon also causes bone-marrow
depression affecting the stem cells of the granulocytes'-13 and
synthesis of haemoglobin."4 Furthermore, interferon causes
immunodepression1' and may have contributed to the severity of

TABLE il-Results of virological investigations throughout course of illness

Day of sample
(from onset of illness)
1
2
3
4
5
6
7
8
9

10, 11, 12, 13

14, 16, 20
23, 27
34
39
61
76
92, 110

Details and remarks

Before transfusion of 450 ml convalescent plasma

11 am, 6 pm, 11 pm

Morning
Morning

Before transfusion of 330 ml convalescent plasma

Morning and afternoon
Morning
Morning
Morning
Morning
Morning
Morning

Morning
Morning
Morning
Morning

Activity of

circulating antibody
(Fluorescent
antibody titre)
<1/2
1/16
1/16
1/16
1/16
1/16
1/8
1/16

1/32
1/64
Not done
1/128
Not done
1/128
1/128
Not done

Recovery of infective virus (guinea-pig intraperitoneal infective

units/ml or g of sample tested)
Positive
Negative

Blood, 10''Blood, 104'6

Blood, 3-10
Blood, 3-10
Blood, 3-10
Blood, 3-10
Blood, 3-10
Blood, 3-10

Seminal fluid, 3-10

Seminal fluid, 3-10

Blood
Blood
Blood, faeces, urine, throat swab
Blood, faeces, urine, throat swab
Blood

Blood, urine

Blood, urine, seminal fluid

Urine, seminal fluid

544

the thrush in our patient. Liver function tests during convalescence showed no evidence of liver damage.
In the early stage of the illness facilities were not available for
conducting haematological or biochemical studies safely, so
efforts were concentrated on establishing the virological
diagnosis; in the late stage of the illness, when provision had
been made for routine tests,16 they were not required for the
management of the patient, though they proved useful for
assessing the extent of damage during convalescence. Fortunately there was no bleeding and the use of prophylactic
heparin was not considered to be necessary.
Once the haemoglobin and white blood cell levels had returned
to normal plasmapheresis was performed to obtain a supply of
convalescent serum.
We thank Professor K Cantell for supplying the interferon and
Professor A J Zuckerman for advising on its use; the World Health
Organisation team in the Southern Sudan and the International
Commission team in Zaire for supplying the convalescent sera used
in treatment; Dr D A Rutter at the Microbiological Research Establishment, Porton, for the haematological and blood chemistry studies; and
Dr Patricia A Webb at the Center for Disease Control, Atlanta, for
some serological studies. Finally we would like to express our apprecia-

BRITISH MEDICAL JOURNAL

27 AUGUST 1977

tion of the support given by the staff of The Royal Free Hospital and
the various health departments.

References
1 Martini, G A, Postgraduate Medical Jfournal, 1973, 49, 542.
2 Gear, J S S, et al, British
Medical,Journal, 1975, 4, 489.

3Weekly Epidemiological Record, 1976, 51, 325.

4Simpson, D, personal communication, 1977.
5 Johnson, K M, et al, Lancet, 1977, 1, 569.
6 Bowen, E T W, et al, Lancet, 1977, 1, 571.
" Pattyn, S, et al, Lancet, 1977, 1, 573.
8 Emond, R T D, Postgraduate Medical Journal, 1976, 52, 563.
9 British Medical Journal, 1976, 1, 64.
10 Bowen, E T W, British J7ournal of Experimental Pathology, 1969, 50, 400.
1 Fleming, W A, McNeill, T A, and Kiuin, T, Immunology, 1973, 23, 429.
12

Nissen, C, et al, Lancet, 1977, 1, 203.

13 McNeill, T A, and Gresser, I, Nature, New Biology, 1973, 244 (II), 173.
14 Falcoff, E, et al, Journal of Virology, 1973, 12, 421.
15 Johnson, H M, Smith, B G, and Baron, S, J'ournal of Immunology, 1975,
114, 403.
16 Rutter, D A, British Medical3Journal, 1977, 2, 24.
(Accepted 223'une 1977)

Prolonged remission maintenance in acute myeloid leukaemia

A S D SPIERS, J M GOLDMAN, D CATOVSKY, CHRISTINE COSTELLO, D A G GALTON,
C S PITCHER
British Medical3Journal, 1977, 2, 544-547

TRAP programme must still be regarded as only

palliative treatment for acute myeloid leukaemia.

Summary
Twenty-five patients with acute myeloid leukaemia were
treated with three quadruple drug combinations in
predetermined rotation: TRAP (thioguanine, daunorubicin, cytarabine, prednisolone); COAP (cyclophosphamide, vincristine, cytarabine, prednisolone); and POMP
(prednisolone, vincristine, methotrexate, mercaptopurine). Fifteen patients (60%) achieved complete
remission and five (20%) partial remission. For maintenance, five-day courses of drugs were administered
every 14 to 21 days and doses were increased to tolerance.
The median length of complete remission was 66 weeks.
-In eight patients remission maintenance treatment was
discontinued and some remained in complete remission
for over two years.
In this series the remission induction rate was comparable with that reported for other regimens and
complete remission lasted longer with this intensive
maintenance regimen than with others. Nevertheless, the

Medical Research Council Leukaemia Unit, Royal Postgraduate

Medical School, London W12 OHS
A S D SPIERS, MD, FRACP, consultant physician (now professor, Section
of Medical Oncology, University Hospital, Boston University Medical
Centre, Boston)
J M GOLDMAN, BM, MRCP, consultant physician
D CATOVSKY, MD, MRCPATH, consultant physician
D A G GALTON, MD, FRcP, honorary director
CHRISTINE COSTELLO, MB, MRcP, research fellow
Stoke Mandeville Hospital, Aylesbury, Buckinghamshire HP21 8AL
C S PITCHER, DM, FRCP, consultant haematologist

Introduction
Many regimens are used in the treatment of acute myeloid
leukaemia (AML), but none has shown unique superiority.'-3
Intensive4 treatments have not proved greatly superior to nonintensive5 regimens. Complex remission maintenance using
multiple drugs6 may be little better than simpler7 regimens.
Remission-induction programmes have incorporated single
drugs8 and combinations of seven9 or eight10 antileukaemic
agents administered simultaneously. The complete remission
rate in adults with AML has varied from 9 5%"" to 79% in
different series.3 11 Higher complete remission rates have been
reported for small groups of patients at specialised centres" 12
than for larger groups treated at many hospitals, where rates have
varied from 9-5%" of 200 adults3 to 34% of 301 adults.'3 The
importance of the choice of drugs and the intensity of treatment
are outweighed by uncontrolled factors including patient
selection and the differing capabilities of different institutions
to give supportive care during the induction of remission. The
complete remission rate in adults with AML has seldom
exceeded 50% in a multicentre study and 65% in a specialised
centre.
Attainment of complete remission in AML slightly improves
survival. In large series the median duration of complete
remission has varied from five to 11 months3 13-15; median
survival has been longer but has seldom exceeded 13 months.3
Immunotherapy administered during remission of AML41 7
seems to prolong the short duration of survival after relapse but
does not prolong the duration of complete remission. No
regimen for remission maintenance in AML is definitely
superior, and the advisability of attempting to maintain remission
at all has been questioned.