Underlying prothrombotic states in pregnancy associated cerebral venous thrombosis
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.68676
Source of Support: None, Conflict of Interest: None
Background : The exact pathogenesis of pregnancy associated cerebral venous thrombois is still unsettled. Aims : To identify possible inherited and acquired prothrombotic risk factors and also identify the factors associated with mortality in pregnancy associated CVT. Settings and Design : Prospective cohort study to identify prothrombotic risk factors and case control study of influence of local traditional practice of puerperal water restriction on postpartum CVT. Materials and Methods : Consecutive patients with pregnancy associated CVT seen over a period of three years. Thrombotic workup included genetic markers, protein assays, and other factors. Statistical Analysis : Univariate and chi-square analysis. Results: Of the 41 patients studied during the study period, 71% of patient had a single and 34% had multiple prothrombotic risk factors. Methylene tetrahydro-folate reductase (MTHFR) heterozygosity (19.5%) and factor V Leiden heterozygous (7.3%) were the commonest genetic markers. Hyperhomocysteinemia (34%) and elevated factor VIII levels (14.6%) were the other important risk factors. In this cohort the mortality was 17%. Mortality increased by odds of 1.3 for every additional prothrombotic marker. The factors associated with increased mortality included: status epileptics (P = 0.05, OR 13.2, 95% CI 1.002 - 173), deep venous system involvement (P = 0.016, OR 9.64, 95% CI 1.53 - 60.6), presence of midline shift (P = 0.012, OR 24.7, 95% CI 2.05 - 29.8) and diffuse cerebral edema (P = 0.006, OR 14.5, 95% CI 2.18- 96.4). The traditional practice of decrease intake of water during puerperium was significant in woman with pregnancy associated CVT when compared to control subjects (P < 0.02). Conclusion : In patients with pregnancy associated CVT, prothrombotic markers can be multiple and are associated with increased odds of mortality. Deep venous system involvement, presence of midline shift and diffuse cerebral edema increased mortality. Peuperial water restriction may be a modifiable risk factor.
Keywords: Cerebral venous thrombosis, factor V leiden, factor VIII, hyperhomocysteinemia, MTHFR, protein C, protein S
The first epidemiological study of hemiplegia due to stroke in south India  showed that the majority of strokes in young women occurred in the puerperal period. In a subsequent study, cerebral venous thrombosis (CVT) was found to be the second common cause of strokes in young women.  Stroke was the most common indirect obstetric cause for maternal mortality.  Srinivasan  suggested that puerperial CVT is seen more frequent in India compared to western countries. The reason for this high incidence is not clear. There has been many studies form India on post partum CVT. ,, But only very few of the studies  looked at the possible underlying prothrombotic states.
The aim of this study was to identify the possible inherited and acquired prothrombotic risk factors in pregnancy related CVT and also to study the influence of the traditional practices followed in the post partum period in the causation of post partum CVT by a case control study.
This prospective study was carried out over a period of 3 years in the Neurology Unit of a quaternary level teaching hospital in south India to identify the inherited and acquired prothrombotic risk factors in pregnancy associated CVT. A case control study looked at the influence of a local traditional practice of water restriction in the puerperal period in the causation of post partum CVT.
Consecutive patients admitted with pregnancy and puerperium associated CVT confirmed by magnetic resonance imaging (MRI), MR venography, or plain and contrast CT scan were included in the study. Patients who were not permanent residents (with the same racial and cultural background) within a radius of 200 kilometers from the hospital, where the study was conducted, were excluded. Data collected included: demographics, detailed clinical profile, and traditional practice of restricting water intake in the post partum period. The history of water restriction in the post partum period was taken from the patient and/or the post partum caregiver (usually the mother or the mother in-law). Controls were matched for age and parity and women from the same geographical area with normal post partum period (interviewed at one month post partum). Questions were addressed in an unbiased manner to both the groups.
All the patients had thrombotic workup, which included: (i) genetic markers: factor V Leiden G1691A mutation, methylene tetrahydro-folate reductase (MTHFR) C677T polymorphism, and prothrombin gene PTG G20210A polymorphism; (ii) protein assays: protein C, activated protein C resistance (APCR), protein S bound and free, antithrombin III, factor VIII, lupus anticoagulant, and anticardiolipin antibody; (iii) other tests: serum homocysteine, antinuclear antibody, haemoglobinuria, and sickle cell preparation. Markers for thrombosis, which could be influenced by the acute thrombotic event (protein assays) if abnormal, were repeated after six months, to avoid a false positive result. This was done after stopping oral anticoagulants for a minimum period of four weeks. Outcome was assessed using modified Rankin Scale (mRs). 
Univariate analysis and chi-square were the statistical tests used for data analysis. Statistical analysis was done using a Statistical Package STATA. In a pilot study,  of the patients with post partum CVT interviewed, 83% gave history of restricted water intake during the post partum period. From this data to answer the research question, "does less water consumption in the post partum period cause CVT?" the sample size required was 32 (alpha error of 0.05 and a beta error of 0.20).
Of the 41 patients (mean age 23.5 + 4.3 years, range 22 - 40 years, 93% below 30 years) studied, 34 (93%) patients presented with CVT during puerperium and three (7%) patients in the antepartum period (2 had aborted and in one the pregnancy was medically terminated). All the patients had the same racial and cultural background and were form the same geographical area (within 200kms radius). Majority, 35 (86%) of the patients were from a rural background and only four (10%) had high school education.
Of the 41 women studied, 18 (44%) were primigravids, 14 (34%) were second gravids and the remaining were 3rd and 4th gravids. Seventeen (17%) patients gave history of spontaneous abortions in their earlier pregnancies, of which three were in the first trimester and two each in the second and third trimester. During their previous pregnancies, one patient had an imaging proven lower limb deep venous thrombosis and another had a hemiparesis which resolved completely in a week without treatment. Thirty-six (88%) of the pregnancies were supervised. Eight (19.5%) patients had complications during pregnancy: seven (17%) patients had eclampsia and one had diabetes mellitus. Three (8%) were delivered at home by the village midwife, 35 (92%) had hospital delivery: nine (23%) were delivered by LSCS. Two (5%) had significant post partum hemorrhage requiring blood transfusions. Six (16%) were still births and in three (8%) the baby died soon after birth. Twenty four (63%) patients developed symptoms within days of post partum [Table 1].
Traditional practice of water restriction during puerperal period
Of the 38 patients, 23 (60.5%) patients gave a history of restricting their water intake (10 (43%) patients took less than 500ml per day, 13 (57%) patients took less than 1000ml per day). Of the 100 control subjects, 39 (39%) had followed the traditional practice and had restricted water intake (<1000ml per day) during puerperium. This difference was significant (P< 0.02).
Clinical features and, image findings are given in [Table 2]. Headache and seizures were the most common symptoms. Glasgow Coma Scale (GCS) score was less than eight in 15 patients and 44% had motor deficits.
Thirty-seven (90%) patients were evaluated by MRI/MRV scan and four (10%) patients had plain and contrast CT scan. Twenty-six (63%) patients had involvement of only the superficial cerebral venous system, 11 (27%) patients had involvement of both the superficial and the deep venous system, and four (10%) had only the deep venous system involvement. The commonest pathology on imaging was the presence of venous infarcts in 29 (71%) patients. Bilateral infarcts were seen in 11 (27%) and hemorrhagic transformation of the infarcts was seen in 14 (29%). One patient had hydrocephalus as a result of the edema causing obstruction to the cerebrospinal fluid (CSF) pathway. One patient had features of a chronic cerebral venous sinus thrombosis with a benign intracranial hypertension like presentation.
CSF was done in 33 patients and was abnormal in 20 (60.6%). No patients had evidence for meningitis or hemorrhagic CSF. The commonest CSF abnormality was an elevated protein content.
Of the 41 patients studied, 29 (71%) patients had prothrombotic states: one prothrombotic marker in fifteen (37%), two prothrombotic markers in eight (19%), and three prothrombotic markers in six (15%). Of the gene markers for thrombosis MTHFR heterozygosity was the commonest [Table 3]. Elevated Protein VIII Level was the most common protein abnormality [Table 4]. In all the patients who had their factor VIII levels elevated, it was reconfirmed after repeat test after a gap of two months (as the initial level could have been due to a acute phase reactant effect).
In all the patients, a decrease in protein C, S and AT III levels were confirmed by rechecking them after two months (this was done as the initial fall in level could be due a consumption coagulopathy, also the fall can be a physiological phenomenon occurring in pregnancy). This confirmation could not be carried out in one patient who died, This patient had a very low protein S level 7.1μ/dl. This level was too low to be caused by consumption coagulopathy, and was hence taken as significant. Two patients tested positive for activated protein C resistance (one was homozygous and the other was heterozygous for the Factor V Leiden mutation).
Lupus anticoagulant was positive in five (12%) patients. None of the patients was tested positive for anticardiolipin antibody. One patient had both antinuclear antibody (ANA) and double stranded DNA. Hyperhomocysteinemia was seen in 14 (34%) patients: ten patients had mild (15 to 30μ moles) and four had moderate hyperhomocysteinemia (30 to 100μ moles). All the patients had tested negative for HIV infection. None of the women were smokers
In the acute phase (unless surgical decompression was being considered) the patients were treated with intravenous un-fractionated heparin (target PTT > 11/2 times the control). Oral anticoagulants were started after 24 - 48 hours and once the target INR (2-2.5) was reached, heparin was discontinued. All the 34 patients who survived were started on long-term oral anticoagulation. At the end of six months, 15 (44%) patients were advised long-term anticoagulation based on the thrombotic workup.
Antiepileptic drugs (AED) were started in patients who presented with seizures and as prophylaxis in patients with cortical infarcts. AED medications were continued in 20 (59%) of the 34 patients started on AEDS, for a period of six months. Acetazolamide was initiated in patients with severe headache and cerebral edema and continued for variable periods. One patient with low GCS and midline shift needed decompressive craniectomy.
There were five deaths and two patients got discharged on request in a moribund state and they were presumed to have died. Thus total deaths were seven (17%). Variables found to have a significant influence on the mortality (univariate analysis) were: status epileptics at admission (P= 0.05 (odds ratio (OR) 13.2, 95% confidence interval (CI) 1.002 - 173), deep venous system involvement (P= 0.016, OR 9.64, 95% CI 1.53 - 60.6), presence of midline shift (P= 0.012, OR 24.7, 95% CI 2.05 - 29.8), and diffuse cerebral edema (P= 0.006, OR 14.5, 95% CI 2.18- 96.4). Because of fewer deaths multivariate logistic regression could not be done. The chance of mortality increased by an odds ratio of 1.3 for each additional prothrombotic factor detected in a given patent.
Of the 34 survivors, four (11.7%) patients were lost for follow up, the discharge mRS scores were two in three patients and three in one patient. At six months of follow-up the mRS scores were: mRS 0 in 13 (43%), mRS 1 in eight (27%), mRS 2 in four (13%), mRS 3 in 3 (10%) and mRS 4 in two (7%) patients.
Headache was the most common complaint in the follow up period. This was seen in five (15%) patients and in two (6%) it started after stopping the oral anticoagulation and in one it was associated with withdrawal of the acetazolamide. Two patients had seizures at the end of six months.
In this study one or more thrombotic markers were positive in 71% of the patients. Multiple prothrombotic risk factors were found in 34% of patients. The significance of these risk factors in causing CVT may be debatable as they may be seen with varying frequency even in normal population. , In this study neither we had studied the frequency of these markers in the population nor we had a control group.
Nagaraja et al,  in a case controlled study found prevalence of factor V Leiden (hetrozygous) in 2.3% and MTHFR C677T in 16.3% of patients with pueperal CVT. They did not find significant association with CVT. This study evaluated only genetic mutations and other thrombophilic states like protein deficiencies, hyperhomocystinemia were not studied. In a small series Cakmak et al,  found prothrombotic risk factors positive in 75% of CVT patients with CVT and in this study 38% of patients had multiple factors. Vora et al,  studied antenatal women with deep venous thrombosis and found 68% having two or more thrombophilic risk factors. This study showed association between some of the hereditary and acquired prothrombotic states and pregnancy associated deep vein thrombosis (DVT). Salomen et al,  showed multiple prothrombotic states in 27.5% of the patients evaluated for idiopathic venous thrombo-embolism.
In our study MTHFR mutations were the commonest genetic marker detected. Martinelli et al,  found hyperhomocysteinemia as the most common thrombophelia state in a large study in patients with CVT. In our study we did not find correlation between the MTHFR mutation and hyperhomocystenemia. Similar were the observation in UK Asian Indians.  In our study 19.5% of the patients were heterozygous and 2% were homozygous for factor V Leiden, which is much higher than 5.5% observed in similar Tamilian population.  No prothrombin gene mutations either heterozygous or homozygous was detected in our. study. Till date this mutation has not been reported from India. ,,,,
In our study elevated protein VIII level was the most common protein abnormality seen 14.6%. High factor VIII levels may increase the risk of venous thrombosis by enhanced thrombin formation and/or through the induction of acquired APC resistance. Elevated factor VIII level was the commonest risk factor found in the study in patients with CVT by Cakmak et al. 
In our study 39% of the multigravids had a history of spontaneous abortion or thrombotic events in previous pregnancies. Kupferminc et al,  found underlying thrombotic state in 71% of patients with various types of obstetric complications. Martinelli et al,  had showed the relevance of thrombophilic gene polymorphisms in women with unexplained late fetal loss. Deficiency of antithrombin III, protein C, and protein S  and also factor V Leiden mutation  were also shown to be associated with risk of abortion and stillbirths. In our study 23% of deliveries were by caesarian section. Caesarian delivery is an independent risk factor for CVT.  Following any surgery protein C levels can decrease  also there is a pregnancy associated resistance to activated protein C. ,
It is a common traditional practice in this geographical region, where women are given restricted water in the post partum period. Regarding this practice there are many believes: (i) water restriction will result in more concentrated breast milk for the baby, (ii) less water intake will prevent loss of abdominal muscle tone and resultant pendulous abdomen, and (iii) restricted water intake during labour can prevent excessive bleeding during delivery. In this study woman with pregnancy associated CVT consumed significantly less water when compared to women with no pregnancy associated CVT, thus suggesting possible role for this practice in the causation of CVT. Although, this traditional practice is followed even in the hot summer months, there was no increase in the incidence of CVT in the summer months.
The outcomes in pregnancy associated CVT are good.  The mortality in our study was 17%, similar to the other Indian studies, 12% to 18%. ,, In this study mortality increased by odds of 1.3 for every additional prothrombotic marker. The factors associated with increased mortality included: status epileptics, deep venous system involvement, presence of midline shift, and diffuse cerebral edema.
Presence of a single prothrombotic factor may not be a significant risk factor in causing CVT. In the setting of physiological state of increased susceptibility for thrombosis like pregnancy and puerperium it is quite possible that the presence of multiple prothrombotic states may shift the balance towards thrombosis. The finding of underlying prothrombotic states will have a significant impact on the long-term management of puerperal CVT especially in regard to the duration of oral anticoagulant use, future pregnancies  and the use of estrogen containing oral contraceptive pills. Our study also suggests that patients should be educated about the importance of adequate hydration during the peuperium. The main limitation of this study is that there was no control group for evaluation of thrombotic markers.
[Table 1], [Table 2], [Table 3], [Table 4]