We initially enrolled 225 patients who presented with severe LEAD and excluded three patients who refused surgical amputation. The study cohort consisted of 222 patients aged 74±11 years (54% male), and 12.6% had acute limb ischemia. The patients with atrial fibrillation were significantly older than those without atrial fibrillation (80.6±9.1 vs. 72.6±11.5 years, P = 0.001), were more likely to have a history of ischemic stroke (28.6% vs. 13.9%, P = 0.056), were more likely to have acute limb ischemia (25.0% vs. 10.8%, P = 0.035), had higher CHADS2 scores (2.9±1.3 vs. 2.2±1.2, P = 0.004) and higher serum uric acid levels (6.8±2.9 vs. 5.7±2.2 mg/dl, P = 0.022). The patients with atrial fibrillation had lower prevalences of chronic kidney disease and end-stage renal disease and lower serum creatinine levels than those without atrial fibrillation (1.8±1.1 vs. 3.7±3.5 mg/dl, P < 0.001). Table 1 describes the details of the baseline characteristics of the study population.
During the study period, the patients with atrial fibrillation had higher incidences of all-cause mortality (42.9% vs. 20.1%, P = 0.014), cardiac-related mortality (21.4% vs. 10.3%, P = 0.111), MACEs (32.1% vs. 14.4%, P = 0.028), MALEs (17.9% vs. 14.9%, P = 0.778) and in-hospital mortality (14.3% vs. 6.7%, P = 0.242) than those without atrial fibrillation. In univariate logistic regression analyses, atrial fibrillation was significantly associated with increased risks of all-cause mortality (crude hazard ratio (cHR): 2.435, 95% CI: 1.274-4.654, P = 0.007) and MACEs (cHR: 2.479, 95% CI: 1.169-5.257, P = 0.018). There was a tendency towards a significant association between atrial fibrillation and cardiac-related mortality (cHR: 2.436, 95% CI: 0.977-6.072, P = 0.056). No significant associations of atrial fibrillation with MALEs (cHR: 1.280, 95% CI: 0.495-3.307, P = 0.610) and in-hospital mortality (cHR: 2.120, 95% CI: 0.691-6.503, P = 0.189) were found. Acute limb ischemia and Rutherford stage were both associated with all-cause mortality (cHR: 4.133, 95% CI: 2.257-7.567, P = 0.003 and cHR: 2.073, 95% CI: 1.330-3.233, P = 0.001); acute limb ischemia was also associated with cardiac-related mortality (cHR: 3.769, 95% CI: 1.579-8.998, P = 0.003) and MACEs (cHR: 2.592, 95% CI: 1.222-5.497, P = 0.013), and Rutherford stage was associated with MALEs (cHR: 4.227, 95% CI: 2.339-7.636, P < 0.001). We show the other confounders associated with the study outcomes in Table 2.
In multivariate logistic regression analyses, atrial fibrillation remained significantly associated with all-cause mortality (adjusted HR (aHR): 2.193, 95% CI: 1.109-4.336, P = 0.024) and MACEs (aHR: 2.322, 95% CI: 1.045-5.158, P = 0.039), independent of acute limb ischemia and Rutherford stage. We did not find a significant association between atrial fibrillation and cardiac-related mortality (aHR: 2.146, 95% CI: 0.819-5.626, P = 0.120) or between atrial fibrillation and MALEs (aHR: 1.757, 95% CI: 0.664-4.789, P = 0.271) after adjusting for the confounders. Figure 1 shows a Kaplan-Meier curve. Acute limb ischemia (aHR: 2.872, 95% CI: 1.491-5.533, P = 0.002) and Rutherford stage (aHR: 1.918, 95% CI: 1.186-3.103, P = 0.008) were significantly associated with all-cause mortality but not cardiac-related mortality or MACEs. Rutherford stage was significantly associated with MALEs (aHR: 5.577, 95% CI: 2.780-11.19, P < 0.001). Heart rate at presentation was significantly associated with all-cause and cardiac-related mortality and MACEs (Table 2).
In the sensitivity analyses, we adjusted for kidney function and body mass index, although these factors were not significantly associated with the study outcomes in univariate logistic regression analyses. The significant association between atrial fibrillation and all-cause mortality remained unchanged after adjustment for kidney function, represented by the creatinine level, creatinine clearance, and the estimated glomerular filtration rate. After we adjusted for body mass index and kidney function in various statistical models, atrial fibrillation was still associated with one-year mortality (aHR: 2.158, 95% CI: 1.012-4.605, P = 0.047 for creatinine clearance and aHR: 2.209, 95% CI: 1.036-4.710, P = 0.040 for the estimated glomerular filtration rate), but we found only a tendency toward a significant association when the creatinine level was used (aHR: 2.080, 95% CI: 0.950-4.555, P = 0.067). Table 3 shows the results of the sensitivity analyses.