RESULT
The overall study consisted of 1003 patients; a mean age of 62.9 ± 10.8 years, of whom 755 (75.3%) were male and 248 (24.7%) were female. Hypertension was the most identified risk factor in 452 patients (45.1%), and 279 of all patients (27.8%) had diabetes mellitus. In-hospital mortality occurred in 69 patients (6.87%).
Patients were divided into two groups according to in-hospital mortality. 934 patients without in-hospital mortality were defined as Group I, and 69 patients who died in the hospital were defined as Group II. Compared to Group I, Group II patients were significantly older (62.5 ± 10.8 vs. 67.45 ± 10.1, p=0.001). Comparisons of the groups with respect to the laboratory data revealed that Group II had significantly lower hemoglobin levels, albumin levels, (p=0.001). Preoperative serum creatinine levels were significant differences between Group I and Group II (1.0 ± 0.02 vs. 1.3 ± 0.1; p=0.01, respectively). Compared to Group I, PNI level was significantly lower in the Group II (48.34 ± 6.71 vs. 44.76 ± 7.63, respectively, p=0.001). The mean BMI was similar between the groups (26.70 ± 2.87 vs. 26.54 ± 2.60 p=0.66). Preoperative LVEF was statistically lower in the Group II rather than the Group I (51.6 ± 0.3 vs. 44.5 ± 1.2; P=0.001). Since the infection was broadly categorized in our study, we compared hospital-acquired infection between groups. There was a significantly different a higher tendency of hospital-acquired infection in the Group II (151; 16.2% vs. 44; 63.8%; p=0.001). Postoperative stroke significantly higher in the Group II (35; 3.7% vs. 62; 89.9%; p=0.001). Multivariate analysis revealed that X-clamp time (p<0.005), CPB time (p=0.001), IABP usage (p=0.001), and ICU stay (p=0.001) significantly higher in the Group II.
The demographic, clinical, echocardiographic, biochemical data of the two groups were shown in Tables 2 and 3.
In the receiver-operating characteristic analysis, PNI lower than 42.6 predicted in-hospital mortality with 82% sensitivity and 42.6 % specificity [area under the curve 0.64, 95% confidence interval 0.572 - 0.714; p=0.001] (Figure1). Multivariate analysis revealed that male sex (95% confidence interval 1.29 - 11.91, p=0.016), age (95% confidence interval 1.11 - 10.15, p=0.032), and the PNI (95% confidence interval 0.18 - 0.90, p=0.026) were independently associated with postoperative survival.