Association of HF medical treatment with VA
Comparing patients with documented VA to those without VA, revealed similar baseline characteristics, except for lower prevalence of Diabetes Mellitus (DM) and longer follow-up among the VA group (Table 1). Crude medication prescription was not associated with VA, nor was the number of guideline-based medications (2.3±0.83 for VA vs. 2.25±0.7 without VA; p =0.8). Patients taking all 3 guideline-recommended medication groups did not have less VA (p 0.33). The patients with VA were treated with significantly lower doses of BB compared to those without VA (23.9±19% versus 35.5±27% target dose; p=0.012). There was no significant difference in AA or MRA doses between the VA and no VA groups (Table 1).
Incident VA was significantly less common among patients treated by >25% target dose of BB as compared to those treated with ≤ 25% target dose (17.6% vs. 33.9%, p 0.017). This was not observed in patients taking >25% AngA (30% vs. 26% p 0.55) or MRA (29.5% vs. 26.5% p=0.64) compared to those treated by ≤ 25% target dose of these medications. Kapkan-Myer (KM) analysis for survival without VA according to each medication group dose, supported reduced VA among patients receiving > median dose of BB (Figure 2).
Univariate parameters found to be significantly associated with incident VA were: heart rate at admission (HR 1.02; 95% CI 1.00-1.04; p=0.02), DM (HR 0.42; 95% CI 0.23-0.78; p=0.006), and BB >25% target dose (HR 0.51; 95% CI 0.27-0.98; p=0.04). In Cox multivariable model for VA including age, gender, DM, medication dosage (>25% target dose), and heart rate, both BB dose > median dose (HR 0.443, 95% CI 0.222-1.022; p=0.021) and DM (HR 0.454, 96% CI 0.237-0.868; p=0.017) were significantly and independently associated with lower incidence of VA; while increased heart rate was significantly associated with VA (HR 1.03, 95% CI 1.009-1.049; p=0.004) (Table 3).