Article Text
Abstract
Background and purpose Few studies have investigated the rates of recanalisation after cerebral venous thrombosis (CVT). Our objective was to investigate the recanalisation rate and to identify predictors of recanalisation in patients with CVT.
Methods We included 102 patients with confirmed first-ever, non-septic CVT. All patients received anticoagulation for 12 months or until complete recanalisation. To assess recanalisation, patients underwent MR venography every 3 months until partial or complete recanalisation or for 12 months after diagnosis. We conducted two parallel analyses of complete recanalisation versus partial and no recanalisation versus any recanalisation. As a secondary objective we explored the influence of recanalisation on outcome and recurrent events. We calculated the probability of recanalisation using Kaplan-Meier analysis and conducted multivariate analysis using a Cox model.
Results The mean age of patients was 33.5±11 years (80 (78.4%) women). Survival analysis indicated that 50% of the patients had any recanalisation (grades I, II and III) by 64 days and complete recanalisation (grade III) by 169 days. Adjusted Cox proportional model revealed that age <50 years (HR=11.5 95% CI=1.58 to 84.46, p=0.01) and isolated superior sagittal sinus thrombosis (HR=0.39, 95% CI=0.14 to 1.04, p=0.05) predict complete recanalisation, while age <50 years (HR=4.79; 95% CI=1.69 to 13.5, p=0.003) predicts any recanalisation. Patients with complete recanalisation had a greater chance of good functional outcome (HR=5.17; 95% CI=2.8 to 9.53, p<0.001).
Conclusions We found that recanalisation occurs over time, until month 11. Complete recanalisation may influence functional outcome.
- CEREBROVASCULAR DISEASE
Statistics from Altmetric.com
Introduction
Cerebral venous thrombosis (CVT) is a relatively uncommon form of stroke that accounts for approximately 1% of all strokes,1 with an associated mortality ranging from 10% to 13%.1 ,2 While several predictors of outcome have been identified, the prognosis of a given individual patient is highly unpredictable. The degree of recanalisation has been proposed as a possible prognostic marker.3 ,4 However, only a few studies have investigated the rates of recanalisation after CVT and the impact of recanalisation on outcome in patients after CVT.3–7 In the majority of patients, partial or complete recanalisation can be demonstrated with MRI. Combining the results of studies with repetitive imaging led to the suggestion that recanalisation only occurs within the first months and not thereafter, irrespective of anticoagulant therapy.7 ,8 It is not known if MRI visualised recanalisation correlates with clinical outcome or recurrence. In this study, we wanted to assess the rate at which recanalisation takes place in patients with CVT and to investigate factors that may predict recanalisation in patients with CVT receiving oral anticoagulants (OA). As a secondary objective, we explored the influence of recanalisation on outcome and recurrent events.
Patients and methods
This is a retrospective analysis of data collected prospectively in our CVT registry. In this analysis, we included consecutive patients with confirmed CVT treated at the Instituto Nacional de Neurología y Neurocirugía (INNN) in Mexico City between 2000 and 2013 who had a first-ever non-septic CVT diagnosed with a baseline MR venogram (MRV) performed within 24 h after presentation to the emergency room and who had at least one follow-up MRV at 3-month intervals for up to 1 year. All patients underwent a standardised diagnostic and follow-up protocol and were treated acutely with intravenous heparin followed by OA for 12 months or until recanalisation was identified on MRV, whichever occurred first, as required by our institutional guidelines (developed based on the clinical experience rather than experimental data). Low-molecular weight heparins were used in pregnant patients until delivery and then changed to OA. The intensity of OA was measured monthly up to a level of international normalised ratio of 2.5 and then every 3 months to complete 1 year of treatment. Patients diagnosed with a specific clotting abnormality are considered to have a high risk of recurrence and received indefinite anticoagulation. Patients with cryptogenic CVT or CVT associated with transient risk factor were treated with aspirin after OA was stopped.
Patient data, including demographics, medical history, CVT onset time, arrival time at the hospital, in-hospital diagnosis studies, time to confirmation of CVT, complications, other treatments and procedures, in-hospital mortality and ambulatory status at discharge were registered in each case. We also investigated specific risk factors for CVT, including current smoking (defined as actively smoking or having quit ≤1 year earlier), admission haemoglobin–haematocrit levels, recent use of hormonal contraceptives (defined as routine use in the year before admission or emergency/occasional use in the previous 3 months), history of abortions or miscarriages, pregnancy or post partum, personal history of thrombophilia or antiphospholipid syndrome, and previous deep venous thrombosis. At our institution, testing for prothrombotic conditions includes protein C, protein S and antithrombin deficiency, antiphospholipid syndrome, prothrombin G20210A mutation and factor V Leiden and is performed 2–4 weeks after completion of anticoagulation.
Clinical presentation was categorised according to the major clinical syndromes that led to emergency department consultation: intracranial hypertension, seizures, motor deficit and others. The modified Rankin scale (mRS) was assessed at hospital discharge and at the time of each follow-up MRV. We classified patients as having good (mRS of 0 and 1) or bad (mRS 2 to 6) functional outcomes. This study was approved by the INNN Ethics Committee. Because of the lack of treatment influence, the study was exempt from informed consent. However, patients, or family members (if the patient was neurologically impaired), verbally agreed to participate in the project.
Imaging diagnosis only Imaging diagnosis
CVT was confirmed via MRI and MRV performed within 24 h after the emergency consultation. MRI and MRV were also used for follow-up. Imaging was performed in a 1.5 or 3.0 T scanner (Signa HDxt systems). The MRI protocol in all cases included the following: sagittal T1-weighted and T2-weighted images; axial T1-weighted and T2-weighted images, fluid attenuated inversion recovery images, T2*-weighted in gradient-echo, and diffusion-weighted and apparent diffusion coefficient images; and coronal T2-weighted images and a three-dimensional time-of-flight MRV.
For this analysis, images from the baseline and all follow-up MRVs were analysed jointly by three experienced stroke neurologists who were blinded to clinical data, including outcome. Readers evaluated the images in two stages and established the diagnosis by consensus. In the first stage, we evaluated MRI sequences to assess parenchymal lesions and signs of CVT. In the second stage, we reviewed MRV and rated each of the following dural sinuses as occluded or not: superior sagittal sinus (SSS), transverse sinus (TS), sigmoid sinus (SS), straight sinus, vein of Galen, inferior sagittal sinus and other deep venous system thrombosis (internal cerebral vein, posterior fossa veins or basal vein). Then, we classified occluded sinuses as SSS alone, unilateral TS/SS, SSS+unilateral TS/SS or ≥3 affected sinuses). Because ipsilateral TS and SS integrate the same functional structure, we considered these two structures as one sinus with the name TS/SS.
Assessment of recanalisation
Recanalisation status was determined by a consensus of three readers using the classification for recanalisation proposed by Qureshi.9 To assess recanalisation, all patients underwent MRI with MRV every 3 months until recanalisation or for 12 months after diagnosis, whichever occurred first. Recanalisation was classified as grade I (partial recanalisation of one or more occluded dural sinuses with improved flow or visualisation of branches), grade II (complete recanalisation of one sinus but persistent occlusion of the other sinuses (A: no residual flow, B: non occlusive flow)), grade III (complete recanalisation) and no recanalisation.
Statistical analysis
Data analysis was performed with the statistical package SPSS V.20 (IBM Corp; Armonk, New York, USA). We conducted two parallel analyses: (1) complete recanalisation (grade III) versus incomplete (grade I and II) or no recanalisation and (2) any recanalisation (grade I, II or III) versus no recanalisation. We calculated the cumulative probability of recanalisation (the complement of the probability of survival) using standard Kaplan-Meier analysis. As we did not know when recanalisation exactly occurred (MRVs were not regularly obtained), for the survival analysis, we considered the time to recanalisation as the midpoint between the last MRV without recanalisation and the first MRV in which recanalisation was noted.10 Variables with p value ≤0.1 were included in the Cox proportional-hazards model to determine which variables were associated with complete or any recanalisation (this model was adjusted to covariates that could influence recanalisation time and pattern: sex, malignancy, acquired or genetic thrombophilia and treatment type).
Results
Over 13 years, 178 patients with CVT were treated at the INNN. For this analysis we excluded 44 who did not meet the inclusion criteria and 32 who did not have follow-up after hospital discharge. The baseline characteristics of included and excluded cases were similar. Our final sample included 102 patients; their characteristics are summarised in table 1. The mean age of the patients was 33.5±11 years and 80 patients (78.4%) were women. Among the women, 30 (37.5%) had recently used hormonal contraceptives, 15 (18.8%) had a history of abortions or miscarriages and 35 (43.8%) were pregnant or in the postpartum period when they had the CVT. Twenty of the patients (19.6%) were current smokers.
We found intracranial hypertension in 40 (39.2%) patients, seizures in 27 (26.5%), acute motor deficit in 17 (16.7%) and other syndromes in 18 (17.6%). The CVT aetiology was cryptogenic in 31 (30.4%) patients, associated with pregnancy or post partum without other cause in 28 (27.5%; in 7 cases associated with pregnancy or post partum, we found other definitive cause for the CVT), due to thrombophilia in 18 (17.7% (4 patients protein S deficiency, 8 protein C deficiency, 4 prothrombin G20210A mutation, 1 antithrombin III deficiency patient and 1 factor V Leiden mutation)) related to hormonal contraceptives in 9 (8.8%), associated with antiphospholipid syndrome in 9 (8.8%) and other causes in 7 (6.9%); only 2 (1.9%) patients were associated with a malignancy and no myeloproliferative syndromes were found. Comprehensive thrombophilia testing was performed in 90 (88.2) patients according to institutional protocol. Diagnostic imaging showed that 25 (24.5%) patients did not have a parenchymal lesion, 23 (22.5%) suffered a single venous infarct, 35 (34.3%) had a single venous haemorrhagic infarct, 11 (10.8%) presented with multiple venous infarcts and 8 (7.8%) had multiple venous haemorrhagic infarcts. The involved sinuses were SSS in 37 (36.3%), TS in 13 (12.7%) SSS+TS in 32 (31.4%) and multiple sinuses in 20 (19.6%) patients. Additionally, the median period from symptom onset to diagnosis was 8 days; most patients (79 patients; 77.5%) were diagnosed within the first 24 h after emergency department presentation. No patients died during the study.
Age was dichotomised for analysis as a nominal variable, with different cut-off values (using the median of its nearest multiple of 5 integer) to evaluate its influence into the recanalisation model. No statistical significance was found among different cut-off values at ages below 40 years. At the age of 50 years or above, statistical significance was found for any type and complete recanalisation (table 2).
After the follow-up MRVs were completed, 3 (2.9%) patients without recanalisation and treated with aspirin had recurrent symptoms of CVT.
The median number of follow-up MRV studies was 2 (48% had 1 follow-up MRV; 39.2% had 2; 10.8% had 3; and 2% had 4). The median time to last MRV was 224 days IQR=144–368) and was longer in patients with partial or no recanalisation than in patients with complete recanalisation (p=0.005). At the last MRV follow-up, recanalisation was complete (grade III) in 67 patients (65.8%) and partial (grade I and II) in 28 (27.5%). In seven patients (6.7%) no recanalisation was noted. Table 3 shows time to recanalisation and the cumulative recanalisation at baseline, and every 3 months.
In the univariate analysis, complete recanalisation was associated with age <50 years, isolated superior sagittal sinus thrombosis (iSSST) and defined cause. Whereas any (grad I, II or III) recanalisation was associated with age <50 years and defined aetiology. Table 2 shows the results for all variables.
The recanalisation profile (partial or complete) had no significant association with long-term headache, long-term seizures or long-term focal deficit. Survival tables indicated that 50% of patients had any recanalisation by 64 days and complete recanalisation by 169 days. Any recanalisation (grades I, II and III) occurred in 93% of the cases. Kaplan-Meier tables for cumulative recanalisation at 365 days are presented in figure 1. In the adjusted Cox proportional-hazard model for covariates, we found that age <50 years (HR=11.56; 95% CI=1.58 to 84.46; p=0.01) and iSSST (HR=0.39; 95% CI=0.14 to 1.04; p=0.05) predicted grade III (complete) recanalisation, while age <50 years (HR=4.79; 95% CI=1.69 to 13.51; p=0.003) predicted recanalisation grades I, II and III. Patients with complete recanalisation had a greater chance of good functional outcome (HR=5.17; 95% CI=2.80 to 9.53; p<0.001) compared to those without recanalisation.
Discussion
We found that most patients with CVT had some degree of recanalisation within 4 months of onset of the thrombosis and a significant proportion of patients with partial recanalisation at 4 months progressed to complete recanalisation over the following months. In our series, only 6.7% of the patients did not have any recanalisation by the time of the last MRV. The cumulative proportion of patients who had any recanalisation was 71% by 3 months, 90% by 6 months and 94% by 9 months.
Our findings are in line with previous studies that show that at 3–6 months, complete recanalisation occurred in 51–57% of patients and partial recanalisation in 19–45%.4–6 ,11 A systematic review of five studies involving 154 patients found that the proportion of patients who had any recanalisation at 3 months and 1 year was 84% and 85%, respectively.8 A retrospective study of 91 patients found that most patients achieved recanalisation, which was complete in a high proportion of cases.6 In these studies, the timing of MRV was variable, and most patients had a single follow-up MRV. In contrast, in our study, patients had multiple MRVs, at least one every 3 months, and this allowed us to determine the recanalisation rate over time and evaluate how the process progressed. Our study suggests that CVT is a relatively dynamic process within the first months after disease onset. Our findings also suggest that, as in other venous pathologies, recanalisation is associated with a better prognosis.4 ,6 We found that in patients with any recanalisation, the HR for having a good functional outcome was increased. In a Finnish study, patients with complete recanalisation were more likely to have an mRS of 0, although this was not statistically significant.6 In our series, a few patients without recanalisation developed new symptoms of CVT after MRV follow-up and under aspirin treatment. The influence of recanalisation on outcome and recurrent symptoms of CVT needs further investigation in larger cohorts.
Other series have found that male gender, age ≥37 years and unknown aetiology were predictors of no recanalisation.6 In our study, however, age was the main predictor of no recanalisation, and the magnitude of the difference could be due to our larger sample size and the fact that we dichotomised age at ≥50 years. Ferro et al12 previously found that the prognosis of patients >65 years was worse than that of younger patients. An explanation of poor prognosis and poor recanalisation in elderly patients could be some of their associated conditions, such as malignancy and dehydration.
As in other studies, we found that patients with a cryptogenic aetiology were less likely to recanalise, although the differences were not statistically significant.3 ,4 ,6 This finding raises a hypothesis that identifying any risk factor, particularly gender-specific risk factors, may lead to better chances of eliminating any potential predisposing factor and thus precipitating recanalisation and better prognosis. A substudy of International Study on Cerebral Vein and Dural Sinus Thrombosis (ISCVT) analysed gender differences and showed that women with gender-specific risk factor, such as oral contraceptive use, hormone replacement therapy, pregnancy or puerperium, have a considerably better prognosis than other patients.13
The characteristics of our population were similar to those from previously reported studies.2 ,14 Our population was slightly younger than in the ISCVT2 ,12 ,14 (33.5 vs 39.1 years); however, the gender distributions in our series and in the ISCVT are very similar (74.5% vs 78.4%). An important difference was that our population had a higher proportion of pregnant and postpartum women, which could have influenced the results.2 ,14 Other differences between our series and the ISCVT included the higher rate of cryptogenic aetiologies (30.4% vs 12.5%) and the thrombophilia diagnoses (26.5% vs 34.1%).2 Our population had slightly better functional outcomes than patients in the ISCVT (mRS 0–1 at last follow-up: 90.2% vs 79.1% in ISCVT). This result may be explained by the inclusion of only patients with completed follow-up, an exclusion of patients who died in the hospital, and resulted in a higher proportion of patients with better prognosis, as was already discussed.
Our study has several strengths, including prospective data collection on a large consecutive series of patients, and the use of a standard protocol for diagnosis, treatment and follow-up. In addition, follow-up imaging was performed multiple times during the first year and with the same equipment. Our study also has limitations. We could not follow-up all patients after the acute phase. Some of our findings could be the results of bias, for example, patients with iSSSt are more likely to be labelled as complete recanalisation since SSS had the highest rate of recanalisation (94%).14 ,15 We did not analyse intracranial collaterals, as we intended to analyse recanalisation as an independent factor, and we excluded patients with severe forms of CVT in whom high-mortality rates are expected. Previously, we found that intracranial venous collaterals are frequently (88%) found in patients with CVT during the acute phase. However, they do not have an independent effect on brain damage, clinical manifestations or prognosis.16
In conclusion, we found that recanalisation occurs over time, until month 11. Complete recanalisation may influence functional outcome. The role of recanalisation on outcome and recurrent CVT needs to be studied in future trials.
References
Footnotes
Twitter Follow Jose Merino at @JG_Merino
Contributors J-CV-G, AA, NA-M, AR-F and JGM were involved in the design of the study. J-CV-G, EM-J, JC, AR-F, NA-M and AQ-C collected the data and bibliographic sources. J-CV-G, NA-M and AA evaluated the radiological studies included in this study. J-CV-G, JGM, MAB and AA were involved in the statistical analysis, and J-CV-G drafted the manuscript. All authors approved the final version of this manuscript.
Funding This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors. The MRVs were obtained as part of routine clinical care.
Competing interests JGM is the US Clinical Research Editor for The BMJ and the journal provides him with salary support. He is also co-editor of the Stroke blog of the journal ‘Stroke’.
Ethics approval Ethics committee of the Instituto Nacional de Neurologia.
Provenance and peer review Not commissioned; externally peer reviewed.