Risk of major bleeding by ethnicity and socioeconomic deprivation among 488,107 people in primary care: a cohort study

Abstract

Background: Antithrombotic medications (antiplatelets and anticoagulants) reduce the risk of cardiovascular disease (CVD), but with the disadvantage of increasing bleeding risk. Ethnicity and socioeconomic deprivation are independent predictors of major bleeds among patients without CVD, but it is unclear whether they are also predictors of major bleeds among patients with CVD or atrial fibrillation (AF) after adjustment for clinical variables. Methods: Prospective cohort study of 488,107 people in New Zealand Primary Care (including 64,420 Māori, the indigenous people of New Zealand) aged 30–79 years who had their CVD risk assessed between 2007 and 2016. Participants were divided into three mutually exclusive subgroups: (1) AF with or without CVD (n = 15,212), (2) CVD and no AF (n = 43,790), (3) no CVD or AF (n = 429,105). Adjusted hazards ratios (adjHRs) were estimated from Cox proportional hazards models predicting major bleeding risk for each of the three subgroups to determine whether ethnicity and socioeconomic deprivation are independent predictors of major bleeds in different cardiovascular risk groups. Results: In all three subgroups (AF, CVD, no CVD/AF), Māori (adjHR 1.63 [1.39–1.91], 1.24 [1.09–1.42], 1.57 [95% CI 1.45–1.70], respectively), Pacific people (adjHR 1.90 [1.58–2.28], 1.30 [1.12–1.51], 1.62 [95% CI 1.49–1.75], respectively) and Chinese people (adjHR 1.53 [1.08–2.16], 1.15 [0.90–1.47], 1.13 [95% CI 1.01–1.26], respectively) were at increased risk of a major bleed compared to Europeans, although for Chinese people the effect did not reach statistical significance in the CVD subgroup. Compared to Europeans, Māori and Pacific peoples were generally at increased risk of all bleed types (gastrointestinal, intracranial and other bleeds). An increased risk of intracranial bleeds was observed among Chinese and Other Asian people and, in the CVD and no CVD/AF subgroups, among Indian people. Increasing socioeconomic deprivation was also associated with increased risk of a major bleed in all three subgroups (adjHR 1.07 [1.02–1.12], 1.07 [1.03–1.10], 1.10 [95% CI 1.08–1.12], respectively, for each increase in socioeconomic deprivation quintile). Conclusion: Ethnicity and socioeconomic status should be considered in bleeding risk assessments to guide the use of antithrombotic medication for the management of AF and CVD.