Targeted use of heparin, heparinoids, or low-molecular-weight heparin to improve outcome after acute ischaemic stroke: an individual patient data meta-analysis of randomised controlled trials

Abstract

Background: Many international guidelines on the prevention of venous thromboembolism recommend targeting heparin treatment at patients with stroke who have a high risk of venous thrombotic events or a low risk of haemorrhagic events. We sought to identify reliable methods to target anticoagulant treatment and so improve the chance of avoiding death or dependence after stroke.

Methods: We obtained individual patient data from the five largest randomised controlled trials in acute ischaemic stroke that compared heparins (unfractionated heparin, heparinoids, or low-molecular-weight heparin) with aspirin or placebo. We developed and evaluated statistical models for the prediction of thrombotic events (myocardial infarction, stroke, deep vein thrombosis, or pulmonary embolism) and haemorrhagic events (symptomatic intracranial or significant extracranial) in the first 14 days after stroke. We calculated the absolute risk difference for the outcome "dead or dependent" in patients grouped by quartiles of predicted risk of thrombotic and haemorrhagic events with random effect meta-analysis.

Findings: Patients with ischaemic stroke who were of advanced age, had increased neurological impairment, or had atrial fibrillation had a high risk of both thrombotic and haemorrhagic events after stroke. Additionally, patients with CT-visible evidence of recent cerebral ischaemia were at increased risk of thrombotic events. In evaluation datasets, the area under a receiver operating curve for prediction models for thrombotic events was 0·63 (95% CI 0·59-0·67) and for haemorrhagic events was 0·60 (0·55-0·64). We found no evidence that the net benefit from heparins increased with either increasing risk of thrombotic events or decreasing risk of haemorrhagic events.

Interpretation: There was no evidence that patients with ischaemic stroke who were at higher risk of thrombotic events or lower risk of haemorrhagic events benefited from heparins. We were therefore unable to define a targeted approach to select the patients who would benefit from treatment with early anticoagulant therapy. We recommend that guidelines for routine or selective use of heparin in stroke should be revised.

Funding: MRC.

Figure 1

Association of baseline variables with thrombotic events—myocardial infarction, recurrent ischaemic stroke, deep vein thrombosis, and pulmonary embolism Each square represents the point estimate from a random effects meta-analysis across trials, and the horizontal line the 95% CI. N=number of patients. n=number of events in each meta-analysis. NIHSS=National Institutes of Health stroke scale.

Figure 2

Association of baseline variables with haemorrhagic events (intracranial or extracranial haemorrhage) Each square represents the point estimate from a random effects meta-analysis across trials, and the horizontal line the 95% CI. N=number of patients, n=number of events in each meta-analysis.