Acta Pñdiatr 92: 254±257. 2003

Cerebral venous thrombosis and Escherichia coli infection in neonates
H Farstad1, P Gaustad2, P Kristiansen3, G Perminov4 and TG Abrahamsen1
Department of Pediatrics1, Institute of Microbiology2, Department of Radiology 3, Rikshospitalet University Hospital, Oslo, Norway;
Department of Pediatrics4, Akershus University Hospital, Nordbyhagen, Norway

Farstad H, Gaustad P, Kristiansen P, Perminov G, Abrahamsen TG. Cerebral venous thrombosis
and Escherichia coli infection in neonates. Acta Pædiatr 2003; 92: 254–257. Stockholm. ISSN
0803-5253
Aim: To present a possible association between cerebral venous thrombosis (CVT) and infection
with Escherichia coli. Methods: Four neonates with deep CVT occurring during an E. coli
infection are presented. Results: In these patients the thrombotic disease was found by Doppler
ultrasonography. The thrombosis involved at least the sagittal sinus and the transverse sinus
according to subsequent MRI scans. The E. coli strains did not produce verotoxin or haemolysin.
Disseminated intravascular coagulation was not demonstrated. Three patients presented with
seizures. At discharge, all of the patients had signs of neurological damage, but two of them have
improved significantly since then. None of the patients has had recurrent (venous) thrombosis.
Conclusion: E. coli infections in neonates may predispose to CVT, a finding that has clinical
implications.
Key words: Cerebral venous thrombosis, Doppler ultrasonography, Escherichia coli infection,
MRI scans, neonates
Hanne Farstad, St. Olavs Hospital, Trondheim University Hospital, NO-7006 Trondheim, Norway
(Tel. ‡47 73 868 213, ‡47 72 555 251 (Home), fax. ‡47 73 867 322, e-mail. hanne.farstad@
rit.no)

Cerebral venous thrombosis (CVT) in neonates is a rare
disorder, but it is usually diagnosed after CT and MRI
scanning (1, 2). Several risk factors have been identified, including infection, dehydration, perinatal hypoxia
(1–5), and in recent years, prothrombotic disorders
(1, 2, 5, 6). We describe four neonates with CVT, all
occurring during an Escherichia coli infection. Two
patients had septicaemia with meningitis and ventriculitis. The other two had urinary tract infections with
suspected, but not verified urosepsis. None of the
patients was known to have had recurrent thrombosis
at follow-up. This prompted the question of a specific
association between E. coli infection and the development of CVT.

Patients and methods
MRI scans were performed with a Siemens Magnetom
1.5 Tesla, Erlangen, Germany, and colour Doppler
ultrasonograms were performed with an Acuson XL
128 machine.
The E. coli isolates were found using routine
microbiological methods in the local hospitals. Blood
cultures were drawn and incubated using the BactAlert
blood culture system (Organon Teknika). The isolates
were tested for their ability to produce haemolysin on
blood agar plates, and were checked for verotoxin or
Shiga-like toxin by the vero cell assay (7). Serotyping
 2003 Taylor & Francis. ISSN 0803-5253

was done in the Escherichia and Klebsiella Centre
(WHO) in Copenhagen, Denmark (8).

Case reports:
Patient 1
This boy, born at term by vaginal delivery, was
transferred to the neonatal ward at the local hospital at
1 d of age because of tachypnoea and hypoglycaemia.
Doppler ultrasonogram showed a partially obstructed
left pulmonary artery, believed to be due to a thrombus.
The patient was discharged at 12 d of age. Readmission
followed 11 d later owing to weight loss, lethargy and
dehydration. C-reactive protein (CRP) was 177 mg/L
and E. coli (more than 100 000 colonies/ml) was
isolated from urine. Antibiotic treatment was started.
Lumbar puncture was unsuccessful. After 3 d the child
had convulsions. A Doppler ultrasonogram indicated
absent flow in the superior sagittal sinus with a
suspected CVT. An MRI at our hospital confirmed the
occurrence of thrombi in several intracranial veins: the
sagittal sinus, the sigmoid sinus, the transverse sinus
and the galenic vein. Although structural changes were
found in both hemispheres on MRI at discharge from
our unit, the patient’s development has been normal so
far, 8 y later. However, he still receives anticonvulsive
treatment. No urinary tract anomaly was found, and

Clinical observations

ACTA PÆDIATR 92 (2003)

255

Table 1. Results of coagulation tests.

Patient 1
Patient 2
Patient 3
Patient 4

aPTT
s

NT
%

FDP
‡/

Fibrinogen
mmol/L

Platelet count
109/l

24
29
29
60/98/31

65/101
117
119
150

‡

7.2
16/7.4
18
14.6/8.2

65/144/211
151/32/165
303/347
129/557

‡
‡

NT: prothrombin time; aPTT: activated partial thromboplastin time, normal activity 70–140%; FDP: fibrin degradation products.

coagulation tests were normal (Table 1). Testing for
prothrombotic disorders included protein C, protein S
and antithrombin III, with normal neonatal values, 58%,
67% and 80%, respectively.
Patient 2
This boy, vaginally delivered at term, was admitted to
the local hospital at 10 d of age after suffering fever,
lethargy and anorexia for two days. CRP was 247 mg/L.
E. coli was cultured from urine and blood, and antibiotic
treatment was started. After five days the patient
became irritable and had convulsions. A Doppler
ultrasonogram showed absent flow in the superior
sagittal sinus. The patient was then transferred to our
hospital. A ventricular puncture revealed an opaque
fluid with 2500 cells per ml and growth of E. coli. The
subsequent six punctures revealed sterile fluid, with a
cell count decreasing to 46 per ml after three weeks. An
MRI scan showed thrombosis in deep cerebral veins, the
sagittal sinus and the transverse sinus (see Fig.). Owing
to increasing hydrocephalus, a ventriculoperitoneal
shunt was inserted. Neurological sequelae with hypotonia and uncertain visual contact had developed at
discharge. After seven years of follow-up, the boy has
delayed motor development and needs a hearing aid.
Image studies of the urinary tract were normal, as were
coagulation tests (Table 1).
Patient 3
This boy, delivered vaginally at term, was admitted to
the local hospital at 8 d of age with fever, irritability and
poor feeding ability. Antibiotic treatment for septicaemia was commenced. E. coli was cultured from blood
while the urine was sterile. A lumbar puncture after one
day of treatment revealed an opaque fluid so dense with
cells and cell aggregates that counting was impossible.
However, no pathogen was isolated. A Doppler ultrasonogram showed absent flow in the superior sagittal
sinus, and a CVT was suspected. After admission to our
department, an MRI scan confirmed this diagnosis.
Thrombi were found in the transverse sinus, the sagittal
sinus and the straight sinus (see Fig.). There were no
convulsions, and EEG was normal. Neurosurgical
evacuation of subdural empyema containing E. coli
was performed 13 d after antibiotics had been started.
Later, a ventriculoperitoneal shunt was inserted. At 7 y

of age the child has mild visual, hearing and psychomotor impairment. Coagulation tests were normal
(Table 1).
Patient 4
This boy was born at term after a vaginal delivery. The
infant was breastfed. At 3 wk of age he had convulsions.

Fig. 1. MRI scans. a. From patient 2: T1-weighted axial view
showing thrombosis in the superior sagittal sinus (arrows). b. From
patient 3: T1-weighted sagittal view demonstrating thrombosis in the
sagittal and straight sini (arrows). c. Also from patient 3: T1weighted coronal view showing thrombosis in the sagittal sinus and
the straight sinus (black arrows) and hydrocephalus development
(white arrow). d. From patient 4: T1-weighted view demonstrating
the thalamic bleeding (black arrow) and thrombosis in the distal
superior sagittal sinus (white arrow).

256

Clinical observations

When he was admitted to the local hospital, it became
apparent that the infant had lost 25% of his birthweight
and was dehydrated. A lumbar puncture revealed blood
in the cerebrospinal fluid (CSF) and a cell count of 220
per ml. The CSF was sterile. A CT scan showed an
intracerebral haematoma located in the thalamus region.
E. coli was cultured from urine, and antibiotic treatment
was commenced. Image studies of the urinary tract were
normal. Coagulation tests (see Table 1 for details)
showed a possible bleeding tendency with a prolonged
activated partial thromboplastin time (aPTT) of 60 s,
increasing to 98 s the next day. Haemophilia was
suspected, and the patient was referred to our hospital.
Here an MRI scan showed thrombi in the sagittal sinus,
the transverse sinus, the straight sinus and the internal
cerebral vein (see Fig.). The haematoma was believed to
be secondary to the thrombus (9), and therefore no
anticoagulant treatment was administered. The aPPT
was normal a week later (31 s). Assessment of
coagulation factors VIII and IX and von Willebrand
factor showed normal activity. Six years later, the
child’s neurological development is normal.

Microbiology results
E. coli grew in blood cultures from two patients, in urine
from three patients, and in CSF from two patients
(Table 2). The saved blood culture and CSF isolates
from patients 2 and 3 were further characterized. They
were negative for verotoxin and haemolysin production.
The blood culture isolate from patient 2 was the
serotype O78:K :H , an extra-intestinal strain. CSF
drawn six days later yielded the same type of E. coli. An
E. coli of the serotype O18ac:K1:H7 was found in the
blood of patient 3, and the same bacterium was cultured
from subdural empyema 13 d later.

Discussion
We describe four boys born at term, with CVT during E.
coli infection. None of the children had a urinary tract
anomaly despite the fact that three had E. coli in urine
cultures. The first three patients came from the same
hospital. However, the isolated E. coli strains in patients
2 and 3 showed different serotypes.
There are several possible mechanisms by which E.
coli can modulate the coagulation system. Toxins such
as verotoxin and haemolysin, which are produced by
certain strains of E. coli, may affect this system.
However, our isolates were not toxin producing.
Furthermore, E. coli releases lipopolysaccharide (LPS)
that may cause a systemic reaction, with life-threatening
shock and disseminated intravascular coagulation (DIC)
(10). None of our patients had measurable DIC, while
CRP, as a parameter of acute phase response, was
highly elevated during the first days of the E. coli
infection. Abnormal coagulation tests strongly support

ACTA PÆDIATR 92 (2003)

Table 2. Isolation sites of E. coli in the four patients.
Urine
Patient 1
Patient 2
Patient 3
Patient 4

‡
‡
‡

Blood

CSF

O78:K :H
O18ac:K1:H7

*
O78:K :H
O18ac:K1:H7**

CSF: cerebrospinal fluid.
* No CSF obtained owing to failed attempts at a lumbar puncture.
** Fluid from a subdural empyema.

the diagnosis of DIC, while normal tests do not exclude
it (11). Recently, Herwald et al. (12) identified surface
proteins on Gram-negative bacteria including E. coli
that specifically interact with proteins of the fibrinolytic
and coagulation cascades.
Gram-negative meningitis, in neonates often caused
by E. coli, generally gives a poorer outcome than Grampositive meningitis. It is known that the presence of K1antigen, found on one of our E. coli isolates, further
increases the mortality and morbidity of the meningitis
(13). However, in large series of neonatal meningitis
(14–17) CVT is not listed among the complications, and
DiNubile et al. (18) documented cortical vein thrombosis in only 1% of adult patients with meningitis.
Our patient no. 4 developed CVT without meningitis,
but with urinary tract infection (UTI). Sterile CSF
pleocytosis during UTI has been reported in neonates
and children (19). Syrogiannopoulos et al. (19) assume
that this is due to meningeal inflammation mediated by
inflammation-inducing molecules derived from bacteria
such as LPS.
The classic approach to the pathogenesis of thrombus
formation by Virchow in 1856 includes endothelial
damage, hypercoagulability and disturbances of blood
flow. Moulding and overlapping of sutures during
normal birth can damage underlying sinuses and induce
thrombus formation (5). Catheters primarily cause
arterial thromboembolism and hepatic and splenic vein
occlusions, and are not a major contributory factor in
cerebral venous thrombosis (5) Thrombophilia as a
contributory factor has been well established (1, 2, 5).
Dehydration (and in neonates, often the resulting
hyperviscosity) causes altered blood flow and was a
co-factor in our patient 4. Sepsis can cause hypotension
and thereby reduced blood flow, and damage the vessel
wall, as can meningeal inflammation (13). All our
patients had one or more of these risk factors.
The pathogenesis of neonatal cerebral venous thrombosis is not well elucidated (1–6). Even after thorough
investigation of known risk factors, such as infection,
dehydration and asphyxia, and prothrombotic disorders,
some cases with CVT remain idiopathic (1–6).
Wu et al. (6) found right-sided jugular vein thrombosis or sigmoid sinus thrombosis in 5% of infants on
extracorporeal membrane oxygenation treatment, possibly caused by the iatrogenic occlusion of the right
internal jugular vein.

ACTA PÆDIATR 92 (2003)

In the literature (1–6) there is very little information
about microbes as cofactors in inducing CVT. The
observation of coexisting E. coli- infection in all our
cases may reflect an increased risk. This possible
association between CVT and E. coli infection has
one important clinical implication: the lethargy and
convulsions often observed in neonatal E. coli infections warrant imaging studies such as colour Doppler
ultrasonography of the head. The results obtained can be
supplemented with an MRI scan, which is the gold
standard of the CVT diagnosis since detailed information about all cerebral veins is obtained.
Acknowledgements.—We thank the Escherichia and Klebsiella Centre
(WHO) in Copenhagen, Denmark for the serotyping and testing of the E.
coli strains from patients 2 and 3.

References
1. deVeber G, Andrew M, Adams A, Bjornson B, Booth F, Buckley
DJ, et al. Cerebral sinovenous thrombosis in children. N Engl J
Med 2001; 345: 417–23
2. Carvalho KS, Bodensteiner JB, Connolly PJ, Garg BP. Cerebral
venous thrombosis in children. J Child Neurol 2001; 16: 574–80
3. Rivkin MJ, Anderson ML, Kaye EM. Neonatal idiopathic
cerebral venous thrombosis: aAn unrecognized cause of transient
seizures or lethargy. Ann Neurol 1992; 32: 51–6
4. Shevell MI, Silver K, O’Gorman AM, Watters GV, Montes JL.
Neonatal dural sinus thrombosis. Pediatr Neurol 1989; 5: 161–5
5. Andrew ME, Monagle P, deVeber G, Chan AK. Thromboembolic disease and antithrombotic therapy in newborns. Hematology (Am Soc Hematol Educ Program) 2001; 358–74. Review
6. Wu YW, Miller SP, Chin K, Collins AE, Lomeli SC, Chuang
NA, et al. Multiple risk factors in neonatal sinovenous
thrombosis. Neurology 2002; 13: 438–40
7. Karmali MA. Laboratory diagnosis of verotoxin producing
Escherichia coli infection. Clin Microbiol N Lett 1987; 9: 65–70

Clinical observations

257

8. Ørskov F, Ørskov I. Serotyping of Escherichia coli. Methods
Microbiol 1984; 14: 43–112
9. Govaert P, Achten E, Vanhaesebrouck P, De Praeter C, Van
Damme J. Deep cerebral venous thrombosis in thalamoventricular hemorrhage of the term newborn. Pediatr Radiol
1992; 22: 123–7
10. Eisenstein BI. Enterobacteriaceae, In: Mandell GL, Bennett JE,
Dolin R, editors. Mandell, Douglas and Bennett’s principles and
practice of infectious diseases. 4th ed. New York: Churchill
Livingstone; 1995. p. 1971
11. Schmidt B, Vegh P, Johnston M, Andrew M, Weitz J. Do
coagulation tests detect increased generation of thrombin and
plasmin in sick newborn infants? Thromb Haemostasis 1993; 69:
418–21
12. Herwald H, Mörgelin M, Olsén A, Rhen M, Dahlbäck B, MüllerEsterl W, et al. Activation of the contact-phase system on
bacterial surfaces—a clue to serious complications in infectious
diseases. Nat Med 1998; 4: 298–302
13. Overall JC. Perinatal bacterial diseases, In: Feigin RF, Cherry
JD, editors. Textbook of infectious diseases. 4th ed. Philadelphia:
WB Saunders; 1998. p. 907–9.
14. de Louvois, J. Acute bacterial meningitis in the newborn. J.
Antimicrob Chemother 1994; 34 Suppl A: 61–73
15. Shattuck KE, Chonmaitree T. The changing spectrum of neonatal
meningitis over a fifteen-year period. Clin Pediatr 1992; 31:
130–6
16. Philip AGS. The changing face of neonatal infection: experience
at a regional medical center. Pediatr Infect Dis J 1994; 13: 1098–
102
17. Adhikari M, Coovadia YM, Singh D. A 4-year study of neonatal
meningitis: clinical and microbiological findings. J Trop Pediatr
1995; 41: 81–5
18. DiNubile MJ, Boom WH, Southwick FS. Septic cortical
thrombophlebitis. J Infect Dis 1990; 161: 1216–20
19. Syrogiannopoulos GA, Grivea IN, Anastassiou ED, Triga MG,
Dimitracopoulos GO, Beratis NG. Sterile cerebrospinal fluid
pleocytosis in young infants with urinary tract infection. Pediatr
Infect Dis J 2001; 20: 927–30
Received Feb. 14, 2002; revision received Sept. 16, 2002; accepted
Sept. 19, 2002