Highlights from #AASTAnnualMeeting - Warfarin use and head trauma

Former Resident/Fellow Heena Santry presented her work on coumadin use and head trauma in the elderly.

IMPACT OF PRE-INJURY WARFARIN USE AMONG MEDICARE BENEFICIARIES WITH HEAD TRAUMA

Courtney E. Collins MD, Elan R. Witkowski MD, Julie M. Flahive MS, Timothy A. Emhoff* MD, Fred A. Anderson Sr., Ph.D., Heena P. Santry MD, University of Massachusetts

Introduction: The effect of warfarin on outcomes of head injured patients remains controversial. Yet more than 2 million Americans, many of them elderly, are started on warfarin annually resulting in more than 30 million prescriptions per year. Meanwhile, with the aging US population, elderly Americans are becoming an increasingly large proportion of head injured patients. We studied a national cohort of Medicare beneficiaries with head injuries to determine the effects of pre-injury warfarin on outcomes.

Methods: A retrospective review of a 5% random sample of Medicare claims data (2009-2010) was performed for enrollees with at least 1 year of Medicare eligibility and Part D prescription drug claims available. Head injury cases were identified using ICD-9 codes for intracranial hemorrhage with or without accompanying skull fractures. Patients with isolated skull fractures or concussions without mention of hemorrhage were excluded. Using Part D prescription drug claims, warfarin exposure was defined as two or more warfarin prescriptions filled within 60 days prior to injury. Characteristics (age, sex, race, co-morbidities) and outcomes (mortality, length of stay (LOS), ICU LOS) between warfarin patients and patients not on warfarin (non-users) were compared using univariate tests of association. Multivariable models adjusting for patient characteristics, concomitant torso injuries and long-bone fractures, and need for ICU care were conducted to measure the independent effect of warfarin on in-hospital mortality.

Results: Of 773,389 eligible Medicare beneficiaries, we identified 3,420 head injured patients (0.4%), 6.6% of whom were treated with warfarin. While warfarin users and non-users were similar in race and co-morbidities, warfarin users were more likely to be female (74.2% vs. 65.6%, p<0.01), and older (median age 83, IQR 78-88 vs. 82, IQR 75-87, p=0.04) than non-users. Warfarin users had higher in- hospital mortality compared to non-users (16.9 vs. 10.2%, p < 0.01). In multivariable analyses, only torso trauma and ICU stay were found to be significant independent predictors of mortality. Warfarin users had 1.9 times the odds (95% CI 1.3-2.7) of dying in the hospital compared to non-users when adjusting for confounders. Eighty-nine percent of patients (N=3,055) survived hospitalization, but warfarin use did not predict ICU admission, ICU LOS, or overall LOS among these survivors.

Conclusion: Anticoagulation with warfarin increases risk of mortality after head injury nearly two fold in Medicare beneficiaries even after adjusting for other risk factors. As new, more difficult to reverse, agents are introduced for chronic anticoagulation this problem may be exacerbated. Physicians should exercise caution when initiating chronic anticoagulation in patients over the age of 65.