Cerebral venous thrombosis

Abstract

Background

Plasma cardiac troponin (cTn) elevation occurs in acute ischemic stroke and intracranial hemorrhage and can suggest a poor prognosis. Because acute cerebral venous thrombosis (CVT) might lead to venous stasis, which could result in cardiac stress, it is important to evaluate whether cTn elevation occurs in patients with CVT.

Methods

Inpatients at Johns Hopkins Hospital from 2005 to 2015 meeting the following criteria were included: CVT (ICD-9 codes with radiologic confirmation) and available admission electrocardiogram (ECG) and cTn level. In regression models, presence of ECG abnormalities and cTn elevation (>0.06 ng/mL) were evaluated as dependent variables in separate models, with location and severity of CVT involvement as independent variables, adjusted for age, sex, and hypertension.

Results

Of 81 patients with CVST, 53 (66%) met the inclusion criteria. Participants were, on average, aged 42 years, white (71%), and female (66%). The left transverse sinus was most commonly thrombosed (47%), with 66% having >2 veins thrombosed. Twenty-two (41%) had cTn elevation. Odds of cTn elevation increased per each additional vein thrombosed (adjusted OR 2.79, 95% CI [1.08–7.23]). Of those with deep venous involvement, 37.5% had cTn elevation compared with 4.4% without deep clots (p = 0.02). Venous infarction (n = 15) was associated with a higher mean cTn (0.14 vs 0.02 ng/mL, p = 0.009) and was predictive of a higher cTn in adjusted models (β = 0.15, 95% CI [0.06–0.25]).

Conclusions

In this single-center cohort study, markers of CVT severity were associated with increased odds of cTn elevation; further investigation is needed to elucidate causality and significance.

Cerebral venous thrombosis (CVT) is distinct in the field of cerebrovascular neurology, given its predilection for young adults with minimal comorbidity. The pathophysiology of brain injury from CVT is generally from (1) increased intracranial pressure (ICP) from impaired venous drainage resulting in neurologic dysfunction or (2) venous ischemia and/or hemorrhage leading to focal brain parenchymal damage.¹ Venous stasis can lead to elevated ICP, venous infarct, or hemorrhage, suggesting that the 2 mechanisms are likely synergistic, but the clinical deficits resulting from either mechanism can lie along a spectrum ranging from mild disease to profound morbidity.

There is increasing recognition and research regarding the importance of the heart-brain axis in stroke prognosis. Patients with arterial ischemic stroke with elevated plasma cardiac troponin (cTn), compared with those without elevated cTn, have increased mortality,² as do those with intracranial hemorrhage, independent of hemorrhage volume, sex, and age.³ In-hospital electrocardiogram (ECG) changes portend a poorer prognosis for both stroke types.^4,5 Single measures of cTn are associated with incident venous thromboembolism,⁶ but whether such changes are seen in patients with CVT is unknown. As cTn is a marker of cardiac injury, it is plausible that turbulent and impaired venous flow occurring in more severe CVT cases, such as those with higher degrees of clot burden, would have elevated cTn. The aim of this study is to describe possible cardiac changes, specifically ECG characteristics and cTn levels, seen in patients with CVT. We hypothesize that more severe features of CVT will be associated with abnormal ECG findings or cTn elevation.

Methods

Study population

Study inclusion criteria were admission to Johns Hopkins Hospital during a 10-year period (2005–2015) with a diagnosis of CVT (ICD-9: 325, 437.6) confirmed by inpatient cerebral imaging, with an available inpatient ECG and cTn drawn during the hospitalization. ECG is standard of care, but cTn was measured at the discretion of the clinical team. Cerebral imaging used to confirm diagnosis was either cerebral tomography angiography or MR angiography. The presence of venous infarction and venous hemorrhage was determined using radiology reports. Age was defined in years at the time of admission. Race and ethnicity were self-reported at the time of admission. Hypertension was defined as a medical history of hypertension. The study was single centered to enable consistent methodology.

ECG measures

The following continuous ECG variables were collected: QT interval (ms), QRS duration (ms), ventricular rate (bpm), atrial rate (bpm), P axis (degrees), and R axis (degrees). The presence or absence of T-wave inversion, ST elevation, U waves, Q waves, premature ventricular contractions, premature atrial contractions, right bundle branch block, and/or left bundle branch block was also determined. These variables were collected, by an investigator masked to CVT characteristics, from the clinical reports from during the inpatient hospitalization, with additional direct review and confirmation of the ECG tracings by the study physician (M.C.J.).

Cardiac troponin

cTn was drawn during the inpatient stay, with the level measured by the Johns Hopkins clinical laboratory. Levels were recorded continuously and a positive cTn defined as ≥0.06 ng/mL.

CVT characteristics

Imaging reports were reviewed masked to ECG data to determine the thrombus location and the individual clot count per patient. Specifically, the location of the venous clot was categorized as present or absent using the following designations: superior sagittal sinus, left transverse sinus (LTS), right transverse sinus, left sigmoid sinus, right sigmoid sinus, left internal jugular, right internal jugular, inferior sigmoid sinus, straight sinus (SS), cortical vein thrombosis (defined as clot in the vein of Labbe or Trolard), deep venous thrombosis (defined as clot in the medullary and/or subependymal veins), and cavernous sinus thrombosis. A large thrombus extending from one venous structure to the adjacent venous structure was counted as present in both structures. Imaging reports were reviewed to determine the presence or absence of insular stroke, venous infarction, and venous hemorrhage, each. If the patient had more than one imaging procedure performed during inpatient admission, the scan resulting in diagnosis (usually the first) was the one considered for analysis and definition of CVT characteristics.

Covariates

Demographics and risk factor data were obtained from the medical record. Demographics considered were race and sex and age at diagnosis of CVT. Risk factor data collected were based on the literature and included a documented (electronic medical record) medical history of type 2 diabetes, high cholesterol, coronary artery disease, atrial fibrillation, previous history of stroke, previous history of intracerebral hemorrhage, sickle cell disease, and oral contraceptive use at the time of CVT event.

Statistical analysis

Categorical variables are reported as percentages, whereas continuous variables are reported as mean values. Two-sample t tests with equal variance and χ² method were used for univariable analysis for continuous and categorical variables, respectively. Multivariable linear regression was used to estimate the associations between continuous ECG parameters (dependent variable) and markers of CVT severity (clot location, clot count, presence of thrombus in deep venous structures, insular stroke, venous hemorrhage, and venous infarct). Multivariable logistic regression was used to estimate the associations between dichotomous ECG parameters and markers of CVT severity. Similarly, regression models were used to estimate the associations between continuous cTn or dichotomous positive cTn (dependent variables) and markers of CVT severity. Regression models were then adjusted for age and sex (model 1) and additionally for hypertension (model 2). Given the small number of observations, only those demographics for which there was the highest suspicion for confounding were included in the adjustment models.

Standard protocol approvals, registrations, and patient consents

For this retrospective analysis, patient data were obtained from hospital records in compliance with the institutional ethical standards committee.

Data availability

Anonymized data not published within this article will be made available by request from any qualified investigator.

Results

Demographics

Fifty-three patients met the inclusion criteria; they were predominantly white females who were middle aged at the time of CVT diagnosis (table 1). Patients had little cardiac comorbidity, with the most common risk factor being a history of hypertension (26%). Seven patients (13%) were using oral contraceptive pills at the time of diagnosis. The LTS was the most common location for thrombus (47%), with 10 patients having hemorrhage associated with the CVT and 8 having an insular stroke (table 1). ECG characteristics fell within the normal range with an average ventricular rate for patients of 78 (±16.0) bpm at the time of ECG. Seven patients had evidence of T-wave inversion. Twenty-two patients had a positive troponin with the overall range of 0.06–0.93 ng/mL (table 1).

Table 1.

Patient characteristics (N = 53)

ECG analysis

There was no difference when comparing mean values or percentages of either the continuous or dichotomous ECG variables when considering these markers across patients with CVT with insular stroke, venous hemorrhage, or venous infarct. When considering the total clot count of patients with CVT, none of the ECG parameters were found to be significantly associated in either univariable or adjusted models (age, sex, and hypertension) (tables 2 and 3).

Table 2.

Logistic regression analysis of the associations between the number of venous structures thrombosed^a and dichotomous ECG parameters (dependent variable)

Table 3.

Linear regression analysis of the associations between the number of venous structures thrombosed^a and continuous ECG parameters (dependent variable)

cTn analysis

Of the 53 participants, 22 met the criteria for cTn elevation (41%). When considering the continuous troponin level, there was no difference in the mean troponin level among those with venous hemorrhage or insular stroke. Those with venous infarction (n = 15) had a higher mean cTn (0.14 vs 0.02, p = 0.009) than did individuals without venous infarction. Venous infarct was associated with a higher troponin level among those of the same age, sex, and hypertension history (β = 0.15, 95% CI [0.06–0.25]) (table 4). Of those with deep venous involvement (deep veins, SS; n = 8), 37.5% had cTn elevation compared with 4.4% without deep clots (p = 0.02). Patients with a higher number of thrombosed deep venous structures had a higher mean troponin in the final adjusted model (β = 0.10 per additional thrombosed deep venous structure, 95% CI [0.01–0.19]); however, there was no association between the presence of thrombus (versus no thrombus) in any individual deep vein and a higher mean troponin. Those with an insular stroke had a higher troponin level compared with those without, but only after adjusting for age, sex, and hypertension history (β = 0.12, 95% CI [0.00–0.24]). Those with a venous infarct or venous hemorrhage had approximately 20 times the odds of a positive troponin compared with those who did not have these CVT characteristics, but this estimate, while significant, is imprecise. Odds of cTn elevation increased per each additional vein thrombosed (OR 2.79, 95% CI [1.08–7.23]).

Table 4.

Regression analyses of the associations between markers of CVT severity and cTn (dependent variable)

Discussion

Among patients admitted to an academic center with a diagnosis of CVT, characteristics of more severe CVT such as venous infarction, venous hemorrhage, involvement of the deep venous structures, or increasing clot count were associated with an elevation of cTn.

The pathophysiology of CVT is unique and distinct and recognized to be different from that of arterial thrombosis. Vasogenic edema occurs earlier in venous strokes than in arterial strokes⁷ and is thought to be due to the early increase in venous pressure that occurs from venous occlusion. Continued increases in venous pressure lead to venous rupture, resulting in hemorrhage, or alternatively, progressive drops in cerebral perfusion pressure cross the ischemic threshold and result in strokes.

Troponin elevation has been repeatedly described in strokes of all etiologies, with a meta-analysis suggesting an 18% overall prevalence.⁸ Troponin release occurs in the setting of cardiac cell necrosis, and hence, it has a role in measuring myocardial ischemia.⁹ However, a study involving patients with subarachnoid hemorrhage and with a positive cTn and wall motion abnormalities on echocardiogram failed to demonstrate any abnormalities during coronary angiogram, suggesting that myocardial ischemia alone is not sufficient to describe what leads to cTn elevation in these patients.^10,11

Adequate venous return to the heart is pivotal for sustained cardiac output¹² and involves the careful balance of mean systemic pressure and right atrial pressure. Mean systemic pressure is affected by blood volume and vascular tone, with the sympathetic nervous system playing an important role in regulation of vascular tone. Cerebral venous drainage is complex with the assumption that increases in right arterial pressure are transmitted through the jugular venous channel, but how this relates to systemic venous return is not well described.¹³ It is recognized that autoregulatory mechanisms protect the brain from alterations in cerebral blood flow, but how much this model applies in the event of a disrupted blood-brain barrier, such as what can occur in CVT, is unknown. Imaging studies suggest that patients with CVT have alterations of venous hemodynamics with increased mean blood flow velocity and turbulence of venous flow in attempting to recruit venous collaterals.¹⁴ It is possible that CVT through this altered turbulent venous return could result in either early myocyte necrosis or some degree of myocardial stretch resulting in a positive troponin.

Our associations with cTn were the strongest among those with deep venous clots. Cerebral deep venous thrombosis is strongly associated with dysfunction of the diencephalon, which can manifest clinically with symptoms of decreased consciousness or other brain stem signs.¹⁵ The presence of a clot in the deep venous system has been associated with an increased risk of death and dependency.¹⁶ It is possible that our reported association represents not only a flow phenomenon but also a disruption to the diencephalon, which modulates the autonomic system and neurohormonal activation and results in nonischemic myocyte necrosis. Autonomic changes secondary to perturbations of particular areas of the brain have been suggested as one potential mechanism for cTn elevation in ischemic stroke and hemorrhage. Although the number of participants with insular stroke was low, we notably did not find an association between insular involvement and odds of elevated cTn or the mean level of patients' cTn.

We have postulated some of the mechanisms by which positive cTn may be seen in CVT, but we recognize that there are limitations to this study. The small size of the cohort may limit our ability to draw conclusions. This study being observational was conducted at a single center to have consistent methodology in terms of the diagnostic evaluation of these patients. The data presented are useful to generate hypotheses that will inform future research. A larger study is warranted to draw conclusions. The ECG obtained on admission and the first image study after admission were used to define study parameters in patients with multiple tests over the duration of the admission. It may be that associations would change should later ECGs or imaging be used. The included patients were also a population in whom a troponin level was obtained at the discretion of the admission team, and this may introduce selection bias. As a result of the retrospective nature of this work, clinical neurologic deficits as documented were not included, so this work is unable to speak to the relationship between clinical deficits and cTn. Furthermore, because of the observational nature of this study, we cannot draw conclusions about causality, and it is likely that there are some residual confounders.

In conclusion, we have demonstrated that among patients with acute CVT with more severe characteristics, there is an increased odds of cTn and that this increases per additional venous structure involved and with deep venous involvement. Although the exact mechanism and clinical ramifications of such a finding remain to be elucidated, these associations may possibly suggest alterations of flow dynamics that have a systemic effect.

Author contributions

M.C. Johansen: drafting/revising the manuscript, data acquisition, study concept or design, and analysis or interpretation of data. R.F. Gottesman: drafting/revising the manuscript, study concept or design, and analysis or interpretation of data. V.C. Urrutia: drafting/revising the manuscript and study supervision.

Study funding

M.C. Johansen: American Heart Association Mentored Clinical and Population Research Award #16MCPRP30350000, National Institutes of Health/ICTR (KL2); Rebecca F. Gottesman: National Institute on Aging (K24 AG052573).

Disclosure

M.C. Johansen reports no disclosures. R.F. Gottesman serves as associate editor of Neurology. V.C. Urrutia-Member serves on an advisory board for Genentech, Inc and site PI for TIMELESS (Sponsor: Genentech, Inc). Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/cp.