Long-term anticoagulant therapy in cerebrovascular disease: does bleeding outweigh the benefit?

Abstract

The aim of the present study was to determine the risk of major haemorrhagic complications, stroke and other cardiovascular events, and mortality during long-term anticoagulant therapy (ACT) in patients with cerebrovascular disease not included in any prospective trials. The data were collected retrospectively. All patients with symptomatic cerebrovascular disease discharged from the Stroke Unit, Aker University Hospital, Oslo, with ACT (warfarin) during 1983 through to 1986 were included. The material consists of 161 patients with a mean age of 67.8 (range 40-90) years. The reason for initiating ACT was frequent transient ischaemic attacks (TIAs) in 52 patients, stroke in progression (SIP) in 33 patients, and probable embolic stroke in 76 patients. International normalized ratio (INR) of 4.2-2.8 was aimed at. Major haemorrhagic complications, recurrent stroke and survival was determined for the total material, and in the subgroups non-valvular atrial fibrillation (NVAF, n = 49), TIAs, and SIP. The mean duration of ACT was 21.1 (range 0.5-60.2) months with a total of 282.9 patient-years. The rate of major (including fatal) haemorrhagic complications was 4.6% per year, and the rate of fatal haemorrhagic complications was 1.4% per year. The complication rates in the subgroups of patients did not differ significantly from that in the total material. Only two out of the 13 major haemorrhagic complications occurred during the initial 6 months of ACT. No strokes occurred in the TIA subgroup. The rate of recurrent stroke (excluding intracranial haemorrhage) was 3.9% per year for all patients, 4.7% per year for the patients with NVAF, and 4.2% per year for the patients with SIP. The total results suggest a positive net effect of ACT in patients with NVAF and TIAs. Without comparable data, no definite conclusions concerning the effect of ACT on patients with SIP can be drawn. The rate of bleeding complications was similar to that in other studied materials and is not negligible. In patients with SIP and TIAs, ACT beyond 6 months should probably only be continued if aspirin is not tolerated or has proven ineffective in the particular patient.