Cerebral Venous Thrombosis Associated With
Pregnancy and Puerperium
Review of 67 Cases
Carlos Cantu, MD; Fernando Barinagarrementeria, MD
Background and Purpose: Cerebral venous thrombosis is characterized by its clinical pleomorphism and
pathogenetic variability. We studied 67 patients with cerebral venous thrombosis associated with
pregnancy and puerperium and compared them with 46 other cases unrelated to obstetric causes to
disclose differences in their clinical presentation, neuroradiological findings, clinical course, and
Methods: In this retrospective study, we analyzed the clinical, laboratory, and neuroimaging findings of
113 patients collected consecutively at our institute. The diagnosis of cerebral venous thrombosis was
confirmed by angiography, magnetic resonance imaging, or neuropathological study.
Results: Patients with cerebral venous thrombosis associated with pregnancy and puerperium were
younger (average age, 26 versus 36 years), and in most, the onset of symptoms was acute (82% versus 54%;
P=.003). The evolution of symptoms reached a plateau within 10 days in 70% of patients with thrombosis
from obstetric causes, compared with only 45% in those from other causes (P=.01). Anemia was more
frequent in the obstetric group (64% versus 26%; P=.00001). There were no differences regarding
neurological and neuroradiological findings. Although the initial severity of illness was similar in both
groups, the final outcome was considered good in 80%, of patients with obstetric causes, compared with
58% of patients with other causes (P=.01); mortality rates were 9% and 33%, respectively (P=.002).
Conclusions: Cerebral venous thrombosis associated with pregnancy and puerperium has a more acute
onset and a better prognosis than thrombosis due to other causes. These findings might be helpful in the
diagnostic and therapeutic strategies for patients with cerebral venous thrombosis. (Stroke. 1993;24:
KEY WoRDs * pregnancy * puerperium * thrombosis * women
In 1828, Abercrombiel published the first clinico- 1. Angiographic evidence of CVT5,6: total or partial
pathological description of a case of puerperal lack of filling of at least one sinus on at least two
cerebral venous thrombosis (CVT). Since then, projections. When the lack of filling involved only the
the association of CVT with pregnancy and puerperium anterior portion of the superior sagittal sinus or only
has been recognized, especially in developing coun- one lateral sinus, additional signs of CVT were re-
tries.2-4 Of 113 cases of aseptic CVT seen in the past 20 quired, such as evidence of delayed venous emptying or
years at the National Institute of Neurology and Neu- development of collateral circulation.
rosurgery in Mexico City, 67 were diagnosed in the 2. Evidence of CVT via neuropathological study:
puerperal period, 5 during pregnancy, and in 1 patient presence of thrombosis of venous sinuses or cerebral
after abortion. The aim of this paper is to analyze these veins, with or without associated venous infarction.
67 cases and compare them with the remaining 46 cases 3. Magnetic resonance imaging (MRI) evidence of
of CVT, which were not related to obstetric causes, to CVT7: absence of voiding on TI weighted images with
disclose some clinical, paraclinical, or prognostic fea- isointensity or hypointensity on T2 weighted images
tures that may be distinctive for patients with CVT during the first few days and hyperintensity of the
lumen of the sinus on T1 and T2 weighted images from
associated with pregnancy and puerperium. about 1 week to 1 month after thrombosis.
Subjects and Methods Patients were excluded when (1) clinical or radiolog-
ical records were incomplete, (2) radiological studies
Patients with a diagnosis of aseptic CVT admitted to were inconclusive, (3) there was cavernous sinus throm-
our hospital were included in this study when they bosis, or (4) CVT was associated with sepsis. Of 67
satisfied any of the following criteria: patients with CVT associated with pregnancy and pu-
erperium (group 1), 50 were included on the basis of
angiographic criteria, 10 by MRI criteria, and 7 by
Received March 24, 1993; final revision received July 18, 1993; neuropathological criteria. Of the 46 cases with CVT
accepted August 3, 1993. unrelated to obstetric events (group 2), 26 were in-
From the Stroke Clinic, National Institute of Neurology and
Neurosurgery, Mexico City. cluded on the basis of angiographic criteria, 11 by
Correspondence to Dr Carlos Cantu, Instituto Nacional de neuropathological criteria, and 9 by MRI criteria.
Neurologia y Neurocirugia, Insurgentes Sur 3877, Tlalpan, Mexico The following characteristics were compared between
14269 DF, Mexico. the two groups:
Canti and Barinagarrementeria Cerebral Venous Thrombosis 1881
TABLE 1. Clinical Profile and Evolution of Cerebral Venous Thrombosis
Group 1 (n=67) Group 2 (n=46)
n % n % P
Acute 55 82.1 25 54.3 .003
Insidious 12 17.9 21 45.7
Progressive 48 71.6 24 52.2 .03
Nonprogressive 19 28.4 22 47.8
Time elapsed between onset and
stabilization of symptoms
<10 days 47 70.1 21 45.7 .01
>10 days 20 29.9 25 54.3
Time elapsed between onset and
<1 week 38 56.7 10 21.7 .0002
1-3 weeks 25 37.3 16 34.8
>3 weeks 4 6.0 20 43.5
1. Clinical: mode of onset, classified as "acute" Treatment for both groups was mainly conservative.
(symptoms appeared suddenly or developed in less than Only five patients with puerperal CVT received antico-
48 hours) or "insidious" (gradual development of symp- agulation therapy, and their data were excluded from
toms over several days or weeks); presenting symptoms; the final outcome evaluation tables because it is well
clinical course, classified as "progressive" (development known that anticoagulation in the early stages of CVT
of additional symptoms and/or worsening of the pre- modifies its natural history.8'9
senting symptoms) or "nonprogressive" (presenting Differences between groups were tested for statistical
symptoms remained stable or tended to subside without significance using multivariate analysis with 2x 2 contin-
development of additional symptoms); neurological gency tables, x2 test, and Fisher's exact test.
signs and symptoms; time elapsed between the onset Results
and stabilization of symptoms; and time elapsed be-
tween onset and diagnosis. The mean age of group 1 patients was 26 years (range,
2. Hematologic: anemia (hemoglobin <12 g%), leu- 16 to 44 years) and of group 2, 36 years (range, 15 to 77
kocytosis (white blood count >10 000/mm3), thrombo- years). Thirteen male patients were included in group 2.
cytosis (platelet count >450 000/mm3), erythrocyte sed- Regarding puerperal patients, 21 cases of CVT
imentation rate (ESR) (normal 0 to 20 mm/h), and (34.4%) occurred during the first week after delivery (9
polycythemia (hematocrit >60%). within the first 48 hours) and 36 (59%) during weeks 2
3. Cerebrospinal fluid (CSF) data: opening pressure and 3 postpartum. Pregnant patients were affected in
(normal pressure <200 mm water), total protein (nor- any trimester (1 in the first, 2 in the second, and 2 in the
mal value 15 to 45 mg/dL), and cell count (normal value third trimester). In group 1, there were 43 multiparas
<5 cells/mm3). (64.1%), 36 women (53.7%) from low or rural social
4. Neuroimaging: evidence of CVT and associated stratum, and 21 women (31.3%) whose delivery was
conducted at home by midwives and who did not have
parenchymal lesions detected by computerized tomogra- proper prenatal care. Three patients had a history of
phy (CT) and MRI. Abnormalities considered as direct venous thrombosis outside the central nervous system
signs of CVT on CT scan included the cord sign, the during previous pregnancies (two instances of pelvic
dense triangle, and the empty delta sign. Hemorrhagic venous thrombosis and 1 case of pulmonary embolism).
infarction was defined as patchy areas of high attenuation Patients from group 2 had the following predisposing
with indistinct margins and a speckled or mottled appear- factors for CVT: use of oral contraceptives (7 patients),
ance, whether separate or coalescent, within a venous thrombocytosis in 4 patients, secondary polycythemia in
infarction. Intracerebral hemorrhage was defined as a 3 cases, circulating lupus anticoagulant in 2, and protein
homogeneous region of high attenuation. S deficiency in 2. In addition, an arteriovenous malfor-
5. Sites of thrombosis within the cerebral venous mation was present in 2 patients and metastatic carci-
system. noma in 1. In the remaining 25 cases, no predisposing
6. Clinical evaluation at the time of discharge from factors were found.
hospital: classified as "good outcome" when patients Clinical, laboratory, and neuroradiological findings
recovered fully or had partial functional recovery with are displayed in Tables 1 through 5. Results of the final
only mild sequelae and "poor outcome" when sequelae clinical evaluation are shown in Table 6. Several differ-
were serious or the patient died. ences were relevant: (1) Patients of group 1 had an
1882 Stroke Vol 24, No 12 December 1993
TABLE 2. Presenting Symptoms in 113 Patients With TABLE 4. Hematologic and Cerebrospinal Fluid
Cerebral Venous Thrombosis Findings in Cerebral Venous Thrombosis
Group 1 Group 2 Group 1 Group 2
(n=67) (n=46) (n=67) (n=46)
n % n % n % n %
Headache 49 73.1 28 60.8 Hematologic
Seizures Anemia* 43 64.1 12 26.0
Generalized 1 1.4 5 10.8 Leukocytosis 24 35.8 20 43.4
Focal 7 8.9 2 4.3 Increased erythrocyte
sedimentation ratet 26/47 55.3 12/41 29.2
Thrombocytosis 2 2.9 6 13.0
Motor 1 1.5 5 10.9
Sensitive 6 9.0 4 8.7 Cerebrospinal fluid
Normal 13/46 28.2 6/27 22.2
consciousness 3 4.5 2 4.3 Increased opening
pressure 19/46 41.3 15/27 55.5
acute onset more often (P=.003); (2) although the exam 29/46 43.4 11/27 20.7
course of illness was progressive in more patients of Hemorrhagic 11/46 23.9 9/27 33.3
group 1 (P=.03), symptoms tended to become stable or
began to subside in a shorter interval (P=.01); (3) *P=.00001, tP=.01.
diagnosis was established within a week of onset in a
higher proportion of patients of group 1 (P=.0002); (4) better outcome (P=.01) and a lower mortality rate
concomitant extraneurological thrombosis was more (P=.002).
frequent in group 1 (22.3% versus 6.5%; P=.01); (5) There were no differences between groups regarding
anemia was more often present in group 1 (P=.00001) sites of thrombosis within the cerebral venous system
and was usually related to an inadequate supply of iron (Table 7).
during pregnancy with or without blood loss during
delivery and puerperium (hypochromic, 26/43, 60.5%; Discussion
normochromic, 17/43, 39.5%); (6) although the severity A wide variety of well-known conditions may cause or
of illness was similar in both groups, with comparable predispose to CVT, and their relative importance may
neurological and neuroradiological findings, CVT asso- vary in different areas of the world.9-14 In the present
ciated with pregnancy and puerperium had a much study. 60% of cases were associated with pregnancy and
puerperium (ratio group 1/group 2, 1.5). This complica-
TABLE 3. Neurological Findings in CVT tion is also common in India, with a prevalence of
4.5/1000 obstetric admissions.2.15 In accordance with an
Group 1 Group 2 obstetric CVT series,16 symptoms of CVT appeared in
(n=67) (n=46) the first 3 weeks after delivery in the majority of cases
Findings n % n % and, as in the Indian series, women who had home
deliveries and poor prenatal care were more often
Headache 59 88.0 32 69.5 affected. CVT should be suspected in any woman who
Focal signs 53 79.1 35 76.0 develops neurological symptoms in the immediate post-
Motor 52 77.6 33 71.7 partum period, since nearly 15% of our cases occurred
in the first 2 days after a normal childbirth. CVT
Sensitive 25 37.3 13 28.2 occurred 13 times more often during puerperium than
Aphasia 17 25.3 10 21.7 during pregnancy. Although in the Indian population,
Disorders of consciousness 42 62.6 27 58.6 multiparas are more often affected than primiparas, in a
Somnolence 24 35.8 9 19.5
proportion of 4: 1,17 in our population this difference
was not so striking (1.8:1).
Stupor/coma 16 23.9 15 32.6 In this study, we found a high frequency (70%) of
Confusion 2 2.9 3 6.5 parenchymal lesions (venous infarcts and intracerebral
Seizures 40 59.7 29 63.0 hemorrhages) that manifested as focal neurological signs,
epileptic seizures, disorders of consciousness, intracranial
Generalized 18 26.9 16 34.8 hypertension, nuchal rigidity, and hemorrhagic CSF.
Focal 22 32.8 13 28.2 These findings also explain the high mortality rate in our
Bilateral pyramidal signs 28 41.7 18 39.1
whole series (18.5%). This is in contrast with reports from
other countries18-20 and illustrates the great clinical and
Papilledema 27 40.2 24 52.1 neuroradiological pleomorphism of this condition world-
Nuchal rigidity 22 32.8 12 26.0 wide. Our series is characterized by a low frequency of
Isolated intracranial isolated intracranial hypertension and absence of Behcet's
hypertension 5 7.4 8 17.3 disease as a cause of CVT. Even when we consider these
variations, there seem to be important differences be-
Cantu and Barinagarrementeria Cerebral Venous Thrombosis 1883
TABLE 5. Neuroradiological Findings
Computed Tomographic Scan Magnetic Resonance Imaging
Group 1 Group 2 Group 1 Group 2
(n=59) (n=36) (n=19) (n=20)
Findings n % n % n % n %
Normal 5 8.4 4 11.1 0 0 0 0
Signs of cerebral venous thrombosis 19 32.2* 13 36.1* 17 89.4 19 95
Nonhemorrhagic venous infarct 16 27.1 7 19.4 3 15.7 2 10
Hemorrhagic venous infarct 21 35.5 12 33.3 10 52.6 11 55
Intracerebral hemorrhage 6 10.1 5 13.8 2 10.5 4 20
Unilateral lesion 25 42.3 15 41.6 8 42.1 12 60
Bilateral lesions 18 30.5 9 25 7 36.8 5 25
*Delta sign, dense triangle, or empty delta sign.
tween CVT associated with pregnancy and puerperium role, such as protein S deficiency, which is common
and that with nonobstetric causes. The clinical picture of during pregnancy and puerperium.23 25 Deficiency of
the former featured a disease with sudden or acute onset, protein S has been reported in previous cases of CVT in
with a progressive course that tends to become stable or obstetric patients.11,26.27 The higher prevalence of extra-
begin to subside in a short time (5 to 10 days). In neural thrombosis in CVT associated with pregnancy
nonobstetric cases, the onset was insidious in about half and puerperium may represent the systemic repercus-
the patients; the course could be progressive or not and sion of a hypercoagulable state.
tended to be stable after 10 days in >50% of patients. Although the disease was of similar severity in both
These findings suggest that, in obstetric cases, the patho- groups, the outcome was more favorable in obstetric
physiological process leading to venous occlusion develops patients. Possibly, CVT in obstetric subjects is a more
faster but is usually self-limited and resolves itself in a benign entity than the one affecting other types of
shorter time. Further evidence for this is the shorter patients because of a more limited and transient occlu-
period between onset of symptoms and diagnosis in ob- sion, with rapid sinovenous recanalization by spontane-
stetric patients; however, it should be born in mind that ous thrombolysis or development of collaterals.
puerperium represents an important clue for suspecting
CVT. TABLE 7. Site of Venous Occlusion in 113 Patients With
Another important difference was the high frequency Cerebral Venous Thrombosis
of anemia in obstetric patients, probably reflecting a
poor diet during pregnancy and/or blood loss during Group 1 Group 2
delivery. The high prevalence of anemia has been (n=67) (n=46)
recognized in most obstetric series of CVT23 and de- n % n %
serves further study to clarify its role in the pathogen-
esis of this entity. SSS 60 89.5 45 97.8
Traditionally, the most frequently postulated mecha- 22 32.8* 11 23.9*
nism involved in CVT in obstetric patients is a hyper- LS 23 34.3 20 43.4
coagulable state2122 associated with dehydration and
anemia. However, other factors may play an important 1 1.4* 1 2.1*
DVS 17 25.3 10 21.7
TABLE 6. Final Outcome in 108 Cases of Cerebral 4 7.4* 0 ...
CCV 13 19.4 14 30.4
Group 1 Group 2 SSS+LS 16 23.8 15 32.6
(n=62)* (n =46)
SSS+CCV 9 13.4 8 17.3
n % n % P
SSS+DVS 6 8.9 5 10.8
Good outcome .01
SSS+LS+DVS 4 5.9 2 4.3
Total recovery 31 50.0 24 52.1
SSS+CCV+DVS 2 2.9 2 4.3
Mild sequelae 19 30.6 3 6.5 1.4 1 2.1
Poor outcome SSS+CCV+LS+DVS 0 ... 1 2.1
Severe sequelae 6 9.7 4 8.7
CCV+LS 1 ...
Death 6 9.7 15 32.6 .002
SSS indicates superior sagittal sinus; LS, lateral sinus; DVS,
*Five patients who had good outcome received anticoagula- deep venous system; and CCV, cortical cerebral veins.
tion and were excluded from statistical analysis. *Isolated venous occlusion involvement.
1884 Stroke Vol 24, No 12 December 1993
Prognosis and treatment are the most controversial 13. Averback P. Primary cerebral venous thrombosis in young adults:
issues of cerebral venous thrombosis. Prognosis is quite the diverse manifestations of an underrecognized disease. Ann
variable. The outcome can range from total recovery to 14. Banerjee AK, Varma M, Vasista RK, Chopra JS. Cerebrovascular
death. Since the introduction of MRI, an increasing disease in North-West India: a study of necropsy material. J Velurol
number of patients with a benign course of CVT have Neurosurg Psychiatry. 1989;52:512-515.
been diagnosed.28'29 Conversely, the benefit of anticoag- 15. Srinivasan K. Puerperal cerebral venous and arterial thrombosis.
ulation now appears well established.89 It would be of Semin Neurol 1988;8:222-225.
16. Dibakdsib JO. Cerebrovascular disease. In: Donaldson JO, ed.
paramount importance to be able to identify the benign Neurology of Pregnancy. Philadelphia. Pa: WB Saunders: 1978:
cases that will recover completely whatever the treat- 135-144.
ment and the severe cases that would justify the risk of 17. Chopra JS, Banerjee AK. Primary intracranial sinovenous
anticoagulation. Based on the experience acquired at occlusions in youth and pregnancy. In: Vinken PJ, Bruyn GW,
our institution, we do not recommend anticoagulation Klawans HL, eds. Handbook of Clinical Neurology, Vol 10: Vascular
Diseases, Part II. New York/Amsterdam: Elsevier Science Pub-
in patients without the factors associated with a bad lishing Co Inc; 1989:425-452.
prognosis.30 18. Bousser M-G, Chiras J, Bories J, Castaigne P. Cerebral venous
thrombosis: a review of 38 cases. Stroke. 1985;16:199-221.
Acknowledgment 19. Thron A, Wessel K, Linden D, Schroth G, Dichgans J. Superior
We are grateful to Dr Carlos Marquez for his assistance. sagittal sinus thrombosis: neuroradiological evaluation and clinical
findings. J Neurol. 1986;233:283 -288.
20. Wechsler B, Vidhaihet M, Piette JC, Bousser MG, Dell Isola B,
References Bletry 0, Godeau P. Cerebral venous thrombosis in Behget's
1. Abercrombie J. Pathological and practical researches on diseases disease: clinical study and long-term follow-up of 25 cases. NVeu-
of the brain and spinal cord. Edinburgh: 1828:83-85. Cited by: rology. 1992;42:614-618.
Donaldson JO, ed. Neurology of Pregnancy. Philadelphia, Pa: WB 21. Strauss HS, Diamond LK. Elevation of factor VIII during
Saunders; 1978. pregnancy in normal persons and in a patient with von Wille-
2. Bansal BC, Gupta RR, Prakash C. Stroke during pregnancy and brand's disease. N Engl J Med. 1963;269:1251-1252.
puerperium in young females below the age of 40 as a result of 22. Inglis TCM, Stuart J, George AJ, Davies AJ. Haemostatic and
cerebral venous/venous sinus thrombosis. Jpn Heart J. 1980;21: rheological changes in normal pregnancy and pre-eclampsia. Br J
171-183. Haematol. 1982;50:461-465.
3. Srinivasan K. Cerebral venous and arterial thrombosis in pregnancy 23. Comp PC, Thurnau GR, Welsh J. Esmon CT. Functional and
and puerperium, a study of 135 patients. Angiology. 1983;34:733-746. immunologic protein S levels are decreased during pregnancy.
4. Estaniol B, Rodriguez A, Conte G, Aleman JM, Loyo M, Pizzuto J. Blood. 1986;68:881- 885.
Intracranial venous thrombosis in young women. Stroke. 1979;10: 24. Boerger LM, Morris PC, Thurnau GR, Esmon CT, Comp PC. Oral
680- 684. contraceptives and gender affect protein S status. Blood. 1987;69:
5. Vines FS, Davis DO. Clinical-radiological correlation in cerebral 692-694.
venous occlusive disease. Radiology. 1971;98:9-22. 25. Mackinnon S, Walker ID, Davidson JF, Walker JJ. Plasma fibri-
6. Kalbag RM. Cerebral venous thrombosis. In: Kapp JP, Schmidek nolysis during and after normal childbirth. Br J Haematol. 1987;65:
HH, eds. The Cerebral Venous System and Its Disorders. Orlando, 339-342.
Fla: Grune and Stratton; 1986:527-529. 26. Cros D, Comp PC, Beltran G, Gum G. Superior sagittal sinus
7. Zimmerman RD, Ernst RJ. Neuroimaging of cerebral venous thrombosis in a patient with protein S deficiency. Stroke. 1990:21:
thrombosis. Neuroimaging Clin North Am. 1992;2:463-485. 633- 636.
8. Einhaupl HM, Villringer A, Meister W, Mehraein S, Garner C, 27. Moreb J, Kitchens CS. Acquired functional protein S deficiency,
Pellkofer M, Haberl RL, Pfister H-W, Schmiedek P. Heparin cerebral venous thrombosis and coumarin skin necrosis in asso-
treatment in sinus venous thrombosis. Lancet. 1991;338:597-600. ciation with antiphospholipid syndrome: report of two cases. Am J
9. Ameri A, Bousser M-G. Cerebral venous thrombosis. Neurol Clin. Med. 1989;87:207-210.
1992; 10:87-1 11. 28. Hulcelle PJ, Dooms GC, Mathurin P, Cornelis G. MRI assessment
10. Krayenbuhl H. Cerebral venous and sinus thrombosis. Clin Neu- of unsuspected dural sinus thrombosis. Neuroradiology. 1989:31:
rosurg. 1967;14:1-24. 217-221.
11. Enevoldson TP, Ross-Russell RW. Cerebral venous thrombosis: 29. Sze G. Simmons B, Krol G, Walker R, Zimmerman RD, Deck
new causes for an old syndrome? Q J Med. 1990:77:1255-1275. MDF. Dural sinus thrombosis: verification with spin echo tech-
12. Lefkowitz D. Cortical thrombophlebitis and sinovenous disease. niques. AJNR Am J Neuroradiol. 1988;9:679-686.
In: Vinken PJ, Bruyn GW, Klawans HL, eds. Handbook of Clinical 30. Barinagarrementeria F, Cantu C, Arredondo H. Aseptic cerebral
Neurology, Vol 10: Vascular Diseases, Part II. New York/ venous thrombosis: proposed prognostic scale. J Stroke Cerebrovasc
Amsterdam: Elsevier Science Publishing Co Inc. 1989:395-423. Dis. 1992;2:34-39.