‘Risky’ view of thrombolysis in subtherapeutic warfarin patients disputed

By Eleanor McDermid

A study in JAMA challenges the view that intravenous thrombolysis is a particularly risky procedure in stroke patients with subtherapeutic international normalized ratios (INRs).

The findings from the Get With The Guidelines (GWTG)-Stroke project conflict with those of a meta-analysis, published less than a month ago, which found an overall increased hemorrhage risk among warfarin-treated patients with an INR of 1.7 or less.

Judith Harrer (Caritas Medical Center Saarbrücken, Germany), who wrote an editorial accompanying the previous meta-analysis, remains cautious in the light of the contradictory findings, telling MedWire News that "the jury is still out on this subject."

However, she described the JAMA study as "very convincing," given its large size. It involved 23,437 stroke patients who underwent thrombolysis in GWTG-Stroke hospitals. Of these, 7.7% were taking warfarin at the time, but had a subtherapeutic INR (median 1.2).

Symptomatic intracranial hemorrhage (ICH), by the National Institute of Neurological Disorders and Stroke definition, occurred in significantly more warfarin-treated patients than other patients, at 5.7% versus 4.6%. But warfarin-treated patients were generally older than the other patients, more often had a history of stroke or cardiac or vascular disease, and had more severe stroke.

After accounting for these and other baseline differences, the odds ratio for symptomatic ICH in warfarin-treated patients was nonsignificant, at just 1.01. There were also no between-group differences for other outcomes (systemic hemorrhage, thrombolysis complications, and in-hospital mortality).

When assessed as a continuous variable, there was a nonsignificant 10% rise in symptomatic ICH risk for each 0.1 increase in INR (confidence interval 1.00-1.20; p=0.06). Eric Peterson (Duke Clinical Research Institute, Durham, North Carolina, USA) and team also conducted an exploratory analysis of all patients with subtherapeutic INRs (ie, <2.0), again finding no significant increase in ICH risk.

However, this analysis included just 33 patients with INRs between 1.7 and 2.0, one of whom had symptomatic ICH. And Harrer commented: "Since we have conflicting data on treatment with lower INR levels I would hesitate to further extend studies to INR ≥ 1.7."

Notably, the findings offer no guidance on treating patients who are taking the newer alternatives to warfarin, such as dabigatran or rivaroxaban.

Currently, "there is no adequate evidence to safely judge the degree of anticoagulation of patients with new oral anticoagulants," said Harrer.

She said that if standard tests such as thrombin time and anti-factor-Xa assay are normal, then "a more or less 'normal' bleeding risk may probably be assumed." But she stressed: "The patient has to be informed distinctively, of course."

Licensed from medwireNews with permission from Springer Healthcare Ltd. ©Springer Healthcare Ltd. All rights reserved. Neither of these parties endorse or recommend any commercial products, services, or equipment.