TRIP

Minorities with Afib at Greater Risk for Brain Bleeds (CME/CE)

MedPageToday, 2012

randomized more than 14,000 patients with nonvalvular atrial fibrillation who were at high risk for stroke to receive warfarin or rivaroxaban (Xarelto).
The patients in the rivaroxaban arm experienced numerically fewer intracranial hemorrhages and fatal bleeds compared with patients treated with warfarin, but had significantly more major gastrointestinal bleeds.
The overall rate of intracranial hemorrhage was low, at 0.5% per year (n=139 from the on-treatment analysis cohort).
For this substudy, however, Hankey and colleagues included 172 intracranial bleeds, which were from the intention-to-treat analysis cohort (i.e., all patients), Hankey explained to MedPage Today .
"Our statistician was initially mistaken when she performed the statistical model on only those patients who took the medication throughout the trial.
As 20% of people stopped the trial medication at some time (i.e., they bled), she did not include them in the initial analysis.
However, the revised figure includes all patients who were enrolled in the study, whether they took the treatment or not –- that is the correct analysis," Hankey said.
The additional bleeds were mostly intracerebral as opposed to subarachnoid, subdural, and extradural.
They then analyzed the data to determine independent risk factors associated with intracranial bleeds.
The factors with tight 95% confidence intervals included: Creatinine clearance, with an HR of 1.10 for every 10 mL/min decrease (95% CI 1.04 to 1.18) Diastolic blood pressure, with an HR of 1.21 for every 10 mm Hg increase (95% CI 1.05 to 1.41) Low platelet count, with an HR of 1.08 for every 10 x 10 9 /L below 210 x 10 9 /L (95% CI 1.03 to 1.14) Eastern Europe versus Western Europe, North and Latin America, and Asia Pacific (HR 0.34, 95% CI 0.22 to 0.50)
Those factors with wider confidence intervals included baseline warfarin use (which conferred a 38% reduced risk), baseline thienopyridine use (2.5-fold greater risk), low serum albumin (37% increased risk for every 0.5 g/dL decrease), and randomization to rivaroxaban (40% decreased risk).