Predictors for intra cranial haemorrhage in frail elderly patients with frequent falls using antithrombotic medication

Abstract

Background Physicians can be reluctant to prescribe antithrombotic agents in frail elderly patients with frequent falls due to the fear for severe bleeding, mainly for intracranial haemorrhage (ICH). Presently, there is only a limited amount of inconclusive data available on the topic. Purpose Identification of risk factors for ICH within a cohort of geriatric patients with repeated falls. Methods All patients of 65 years of age and older with repeated falls at our day clinic were eligible. If an MRI of the brain was performed as part of the assessment, patients were included in this analysis. Baseline characteristics including medical, functional, and cognitive state were collected, a Frailty Index (FI) was calculated [1,2]. Cerebral small vessel disease was described and evaluated as proposed in a position paper in 2013 [3]. Follow-up data concerning major bleeding events were retrieved from the electronic medical files. Odds ratios (OR) with confidence intervals (CI) were calculated. Results 670 patients were eligible; an MRI was performed in 486 patients. The average age was 80 years, 50% was severely frail at the time of inclusion. 83 patients (17%) used OAC (mainly Vitamin K antagonists prescribed for atrial fibrillation), 165 patients (34%) used anti platelet agents (APA), 1 patient used both OAC and APA. In total, 29 major bleeding events (MB) occurred, of which 13 were ICH. Among patients using OAC, 8 MB occurred, of which 2 were ICH. The patient with both OAC and APA did not experience a bleeding event. Well known risk factors for ICH such as hypertension, diabetes mellitus and cognitive impairment were not predictive for ICH in this cohort, nor were the use APA (OR 0.86, 95% CI 0.26–2.84), or vitamin K antagonists (OR 0.88, 95% CI 0.19–4.05). However, a composite factor of using either APA or OAC, heightened the risk for MB (OR 3.24, 95% CI 1.35–7.74), but not for ICH (OR 0.83, 95% 0.27–2.49). Of cerebral small vessel disease, predictive factors for ICH were the presence of lacunes (OR 3.81, 95% CI 1.25–11.56), and relevant white matter hyperintensities (WMH) (defined as a Fazekas score of 2 or more) (OR 11.3, 95% CI 1.45–87.3). Furthermore, cognitive decline defined as an MMSE score of ≤26 heightened the risk of MB (OR 2.28, 95% CI 1.05–4.96). The low number of ICH did not allow for a multivariate analysis. Conclusion This analysis has several important findings. First, despite the long follow up of a cohort of severely frail patients that frequently fall, a low number of MB and ICH was observed. Second, well known risk factors for MB do not seem predictive of ICH in this cohort of very elderly patients. Finally, cognitive decline was predictive for MB, and WMH and lacunes were predictive for ICH. Adding cognitive screening and brain imaging to the diagnostic work up of patients with an indication for OAC could be of value when assessing the future risk for major bleeding events. FUNDunding Acknowledgement Type of funding sources: None.