By Thomas Jenkins
For November 2019, we have selected: Zuurbier et al., Long-term antithrombotic therapy and risk of intracranial haemorrhage from cerebral cavernous malformations: a population-based cohort study, systematic review, and meta-analysis Lancet Neurol 2019; 18: 935-41.
This month’s selected paper addresses the risk of neurological deficits in people with cavernomas treated with antithrombotic medication. Perhaps counter-intuitively, the authors report a lower risk of intracerebral haemorrhage in patients taking antithrombotic medication (mostly antiplatelet therapy). The study therefore provides reassurance for clinicians managing patients with this relatively common, and often incidental, vascular finding that antithrombotic therapy does not appear harmful; conversely, there could be some benefit, but this requires further evaluation in a randomised controlled trial.
Cerebral cavernous malformations (or cavernomas) are the second commonest incidental vascular finding on brain MRI. They are commonly asymptomatic but can occasionally cause stroke through intracranial haemorrhage or non-haemorrhage associated focal neurological deficit if in an eloquent brain location. Around a quarter of patients with cavernomas have an indication for antithrombotic therapy. Prior to this study, expert opinion advised against anticoagulants based on a case report of haemorrhage. However, four previous retrospective hospital-based cohort studies reported non-significant associations between antithrombotic medication and lower risk of haemorrhage. This study is the first prospective population-based study to address this issue. The resulting data were also combined with the four previous studies and additional unpublished data from the Mayo clinic, Rochester, USA to perform a meta-analysis to determine risk of intracerebral haemorrhage in patients with cavernoma treated with antithrombotic medication.
The primary study reports data from the Scottish Audit of Intracranial Vascular Malformations, in which observational data were collected on all patients diagnosed with a cavernoma in Scotland, maximising case ascertainment through an opt-out consent process. To determine subsequent outcomes, hospital and general practitioner records were analysed prospectively annually, and patients contacted (via opt-in consent) and invited to fill in a questionnaire. Antithrombotic treatment was defined as at least 90 days of anticoagulant or antiplatelet therapy between inception into the cohort and last follow-up, and had to occur before any recorded outcome events. The primary outcome was a composite of haemorrhagic stroke confirmed on brain imaging, or new and persistent focal neurological deficit deemed definitely related to anatomical location of the cavernoma. Standard statistical approaches including Kaplan-Maier and Cox proportional hazards models were applied.
Three-hundred and six patients were diagnosed with a cavernoma over a 10 year period. Six patients with an incidental finding of cavernoma at autopsy were excluded. Sixty-one of the final cohort of 300 patients had been treated with antithrombotic therapy (16% with anticoagulants with or without antiplatelet therapy, the remainder antiplatelet therapy alone). Fifty-three percent were already on treatment with antithrombotic agents at the time of diagnosis of the cavernoma. Mean duration of therapy was 7.4 years (standard deviation 5.4 years). The treated group were older, with more vascular comorbidity, and a higher likelihood that the cavernoma was an incidental radiological finding than the non-treated group. Mean follow-up was 11.6 years (standard deviation 5 years), totalling 3634 person years. A single patient taking antiplatelet therapy (1/61; 2%) developed the primary outcome, compared to 29/239 (12%) patients not taking antithrombotic therapy (p=0.011). The hazard ratio, adjusted for age, type of presentation, and cavernoma location was 0.12 (95% confidence interval (CI) 0.02, 0.88, p=0.037).
The meta-analysis comprised 1342 patients, assessed for the primary outcome of intracerebral haemorrhage (non-haemorrhage associated focal neurological deficits were excluded from this analysis to allow uniformity of reporting). Three percent of people used anticoagulants and, in this group, there was 1 haemorrhage in 260 person years of follow-up. Fifteen percent used antiplatelet therapy and 7 haemorrhages in 1221 person years were recorded. In the remaining 81%, no antithrombotic therapy was given and there were 152 haemorrhages in 6214 person years. Antithrombotic therapy was associated with a lower risk of intracerebral haemorrhage (3% versus 14%, incidence rate ratio 0.25 (95%CI 0.13, 0.51, p<0.001).
To explain these results, the authors proposed that haemorrhage events in patients with cavernomas could be precipitated by thrombosis, similar to the mechanisms in other disorders, for example, secondary haemorrhage in venous sinus thrombosis. They acknowledged possible sources of confounding bias between antithrombotic treated and untreated groups, but results survived adjustment for these factors in their analysis. They concluded that there was no evidence of a harmful association between antithrombotic therapy and intracerebral haemorrhage, and recommended that the possibility of benefit should be investigated further in a randomised controlled trial.
“This paper may change the way we think about risk in patients with cerebral cavernous malformations”, said Professor Simona Sacco, University of L’Aquila and Director of Neurology, Avezzano Hospital, Italy, “Decisions in regard to antithrombotic treatment can be difficult in individual patients and there are important implications for clinical practice.”
