Mortality of Cerebral Venous–Sinus Thrombosis in a Large National Sample
Background and Purpose—The purpose of this study was to evaluate the mortality rates associated with cerebral venous–sinus thrombosis in a large national sample.
Methods—A cohort of patients with cerebral venous–sinus thrombosis was identified from the National Inpatient Sample database for the years 2000 to 2007. According to the International Classification of Diseases, 9th Revision, Clinical Modification codes, cerebral venous–sinus thrombosis is categorized into pyogenic and nonpyogenic groups. Multivariate logistic regression analysis was used to assess covariates associated with hospital mortality.
Results—Among 3488 patients, the overall mortality rate was 4.39%, which was nonsignificantly higher among the pyogenic group (4.55% versus 3.52%; OR, 0.76; 95% CI, 0.47–1.23). In the pyogenic cerebral venous–sinus thrombosis group, hematologic disorders were the most frequent predisposing condition (16.2%); whereas systemic malignancy followed by hematologic disorders were most common in the nonpyogenic group (14.08% and 10.04%, respectively). Predictors of mortality included age, intracerebral hemorrhage as well as the predisposing conditions of hematologic disorders, systemic malignancy, and central nervous system infection.
Conclusions—Compared with arterial stroke, CVST harbors a relatively low mortality rate. Death is determined by age, the presence of intracerebral hemorrhage, and certain predisposing conditions.
Recent literature supports a dramatic decrease in the mortality rate associated with cerebral venous–sinus thrombosis (CVST) from 30% to 50% in the middle years of the 20th century to 8.3% in the third millennium.1 This shift in mortality is likely related to the better recognition of the entity, earlier diagnosis with noninvasive imaging, and improvements in available treatment options. There is considerable variation in the mortality rate associated with CVST, perhaps influenced by the patient's demographics as well as the pathological conditions leading to venous thrombosis in the cranial circulation. For the most part, these conditions are typically associated with a nonpyogenic CVST, whereas pyogenic CVST is usually associated with head trauma and local extension of an infection from the cranial sinuses.2 The main objectives of this study is to determine the mortality rates associated with pyogenic and non pyogenic types of CVST in a large sample abstracted from a national database as well as the potential predictors of death.
We extracted our sample from the Nationwide Inpatients Sample data set for the years 2000 through 2007. We obtained an Institutional Review Board exemption for waivers of informed consent and Health Insurance Portability and Accountability Act authorization from St Louis University School of Medicine. The sample was identified by International Classification of Diseases, 9th Revision, Clinical Modification codes for pyogenic and nonpyogenic CVST (Supplemental Table I; http://stroke.ahajournals.org). Pyogenic CVST is defined as intracranial venous sinus phlebitis or septic or suppurative infection of intracranial venous sinuses. Nonpyogenic CVST is defined as thrombosis of the intracranial venous sinus in the absence of infection in the cranial cavity. The Charlson comorbidity index was used to weigh the comorbidities of the chronic conditions.3 To identify the predisposing conditions associated with CVST, we used the Clinical Classification Software for the followings groups: hematologic conditions, central nervous system infections, systemic inflammatory diseases, any malignant cancer, benign brain tumors, traumatic brain injury, pregnancy and puerperium, and dehydration.
In addition to in-hospital mortality, we identified common stroke-associated complications including intracerebral hemorrhage, status epilepticus, respiratory failure needing endotracheal intubation, systemic infection (or sepsis), and gastrointestinal hemorrhage. Two regression analyses were performed; the first was fitted to determine the independent association of significant clinical, demographic, and hospital-associated complications for the outcome of in-hospital mortality. Because the classification software groups multiple conditions under the same category, the second model included these grouped predisposing disease categories for the outcome of in-hospital mortality. The significance level was set a priori at P<0.01.
The analysis included 3488 patients with the mean age of 39.1±21.9 years; 14% were excluded due to double representation or missing death status. Among the 2 types of CVST (Table 1), female and whites were predominant in both groups. Patients with pyogenic CVST were younger by a decade (mean age, 37.55±21.65 versus 46.90±21.48 years; P<0.0001), but they had similar other demographics (P>0.05). Predisposing conditions were found in 43.2% of those with pyogenic CVST and 38.5% of nonpyogenic CVST (Figure). Hematologic disorders and systemic malignancy were the most frequently reported conditions with pyogenic and nonpyogenic CVST, 16.2% and 14.08%, respectively. Among the overall cohort, 153 patients died (4.39%) without a significant difference in the mortality rates between the 2 types of CVST (4.55% versus 3.52%; OR, 0.76; 95% CI, 0.47–1.23). Multivariate analysis (Table 2) showed the following predictors of mortality: age, intracerebral hemorrhage, respiratory failure, and a severe comorbidity index. The OR for death increased with each decade of life but became significant after the fourth decade, and it peaked in the seventh decade (OR, 4.31; 95% CI, 2.03–9.16). Other predictors of mortality shown in the second model include systemic malignancy, hematologic disorders, and central nervous system infection. Sensitivity analysis to examine the potential influence of transfer between hospitals showed no difference in outcomes from the original analysis (Supplemental Table II).
The overall mortality rate in our study was similar to those reported in the International Study on Cerebral Vein and Dural Sinus Thrombosis (ISCVT), which showed a rate of 4.3% at discharge and 6.8% at 6 months follow-up.4 Intracerebral hemorrhage was a predictor of mortality in our cohort, but lesion characteristics could not be further ascertained. Pathological processes leading to CVST were found at a lower rate (42.4%) compared with the ISCVT (87.5%). Nearly half of those in the ISCVT were attributed to oral contraceptives. Oral contraceptives were considered a common predisposing factor for CVST in many studies.4,5 Hence, the low rates of identifying a pathological process in the Nationwide Inpatients Sample is likely due to the lack of clinical information including medications administered before hospitalization. The hypothesis that hematologic disorders are more common among the nonpyogenic group and central nervous system infection is more common among the pyogenic group was confirmed in our cohort. Regardless, the mortality in CVST is dependent on the predisposing pathological process.6 Three conditions were associated with high mortality rates including the presence of hematologic disorders, systemic malignancy, and central nervous system infections. The former 2 conditions can be interrelated because malignant cells expose tissue factors that promote the coagulation cascade and subsequent thromboembolism.7 In addition to these conditions, age is an important determinant of death in this young cohort.8
The strength of this report is the large sample size of a rare syndrome collected from a national database; this sample is nearly 4 times that reported in the ISCVT.4 However, this study is limited by the possible coding errors and the lack of clinical and radiological information in the Nationwide Inpatients Sample data set.
The online-only Data Supplement is available at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.111.635664/-/DC1.
- Received August 9, 2011.
- Accepted August 12, 2011.
- © 2012 American Heart Association, Inc.
- Ferro JM,
- Canhao P,
- Stam J,
- Bousser MG,
- Barinagarrementeria F
- Gosk-Bierska I,
- Wysokinski W,
- Brown RD,
- Karnicki K,
- Grill D,
- Wiste H,
- et al
- Kasthuri RS,
- Taubman MB,
- Mackman N