Predictors of Bleeding in Patients with Symptomatic Peripheral Artery Disease: A Cohort Study Using the Health Improvement Network (THIN) in the UK

Abstract

Introduction: Cardiovascular risk reduction guidelines recommend antiplatelets and statins for all individuals with symptomatic peripheral artery disease (PAD), and antihypertensives for those with concomitant hypertension. Antiplatelet use is associated with an increased bleeding risk. We aimed to assess potential risk factors for intracranial bleeding (ICB) and gastrointestinal bleeding (GIB) in patients with symptomatic PAD. Methods: Patients aged 50–89 years diagnosed with symptomatic PAD in 2000–2010 were identified using The Health Improvement Network (THIN) UK primary care database (N = 28 484). The main exclusion criteria were: dementia; ischaemic stroke; anticoagulant or dual antiplatelet therapy (past 3 months); and previous ICB or GIB (past 6 months). Patients were followed up until they reached an endpoint (ICB or GIB), death or end of follow-up (ICB: June 2013; GIB: March 2012). Follow-up for GIB was restricted to THIN practices linked to hospital episode statistics data. Potential comorbid, demographic and therapeutic risk factors for ICB (153 cases) and GIB (506 cases) were assessed using a nested case–control analysis, with 1000 controls for ICB and 5000 controls for GIB, case-matched by age, sex and calendar year. Odds ratios (ORs) and 95% confidence intervals (CIs), adjusted for potential confounders, were calculated. Results: The incidence of ICB was 7.9 (95% CI: 6.8–9.3) per 10 000 person-years (median follow-up: 6.4 years). Previous ICB was a major predictor of ICB (OR: 3.85; 95% CI: 1.33–11.13). Regarding other potential risk factors for ICB, the OR (95% CI) was 0.90 (0.61–1.34) for treated hypertension, 1.59 (0.65–3.87) for untreated hypertension, 1.38 (0.80–2.36) for current smoking, 0.73 (0.40–1.33) for body mass index (BMI) ≥ 30 kg/m2 and 2.20 (1.12–4.34) for BMI 15–19 kg/m2. Compared with patients not using acetylsalicylic acid (ASA), clopidogrel or warfarin in the previous year (49 patients), the OR (95% CI) for ICB was 0.78 (0.50–1.21) with ASA monotherapy (55 patients), 0.40 (0.09–1.82) with clopidogrel monotherapy (2 patients) and 1.27 (0.47–3.47) with warfarin monotherapy (8 patients). The incidence of GIB was 9.6 (95% CI: 8.8–10.5) per 1000 person-years (median follow-up: 6.0 years). Major predictors (OR; 95% CI) of GIB were peptic ulcer disease (1.40; 1.05–1.86), dual antiplatelet therapy (3.20; 1.81–5.64) and use of non-steroidal anti-inflammatory drugs (1.96; 1.46–2.64). Regarding other potential risk factors for GIB, the OR (95% CI) was 1.32 (1.05–1.65) for treated hypertension, 1.35 (0.77–2.35) for untreated hypertension, 0.98 (0.72–1.33) for smoking, 1.26 (0.92–1.72) for BMI ≥ 30 kg/m2 and 0.67 (0.44–1.01) for BMI 15–19 kg/m2. Compared with individuals not using ASA, clopidogrel or warfarin in the previous year (124 patients), the OR (95% CI) for GIB was 1.78 (1.39–2.30) with ASA monotherapy (245 patients), 2.03 (1.05–3.93) with clopidogrel monotherapy (15 patients) and 1.25 (0.72–2.16) with warfarin monotherapy (19 patients). Conclusion: Incidence of GIB was about 10-fold higher than incidence of ICB in patients with symptomatic PAD in UK primary care. Antiplatelet therapy increased the risk of GIB.