Skip to main content
. 2019 Aug 6;71(4):350–355. doi: 10.1016/j.ihj.2019.07.004

Table 4.

Predictive factors for hospital stay after cardiac surgery (n = 362).

Variable Hospital stay
≤7 days
>7 days
Univariate
Multivariate
n (%)/mean (SD) n (%)/mean (SD) p value aOR 95% CI p value
Age (years)
≤50 64 (41.6) 90 (58.4) 0.682
>50 82 (39.4) 126 (60.6)
Sex
Male 80 (36.5) 139 (63.5) 0.068 1.39 0.87–2.24 0.171
Female 66 (46.2) 77 (53.8) 1
Type of surgery
CABG 68 (44.2) 86 (55.8) 0.008 1.79 0.84–3.96 0.150
Valve 56 (32.9) 114 (67.1) 2.44 1.11–5.35 0.026
Combined 22 (57.9) 16 (42.1) 1
DM 53 (42.7) 71 (57.3) 0.500
Hypertension 45 (41.7) 63 (58.3) 0.736
IHD 72 (44.2) 91 (55.8) 0.178
VHD 65 (34.8) 122 (65.2) 0.025 0.73 0.37–1.43 0.364
Asthma/COPD 03 (42.9) 04 (57.1) 0.891
CKD 04 (36.4) 07 (63.6) 0.785
Weight (kg) 59.9 (11.4) 61.8 (12.8) 0.162
LVEF 54.4 (3.5) 54.2 (3.4) 0.574
Preop Hb 12.2 (1.9) 12.5 (1.6) 0.141 1.11 0.97–1.27 0.117
Creatinine 1.16 (0.24) 1.20 (0.28) 0.146 1.47 0.62–3.49 0.380
CPB time 87.2 (26.4) 88.4 (29.3) 0.694
ACC time 51.2 (18.2) 52.2 (19.6) 0.614
Pump flow rate 3.7 (0.5) 3.8 (0.6) 0.666
CPB lactate 4.1 (0.6) 4.2 (1.7) 0.188
ACC lactate 3.9 (1.3) 4.0 (2.3) 0.596
On-pump Hb 6.8 (1.2) 6.9 (1.3) 0.274
ICU lactate 4.4 (1.6) 4.5 (1.8) 0.630
6-hr lactate 4.2 (2.1) 4.5 (2.1) 0.141 1.04 0.93–1.16 0.537
12-hr lactate 3.1 (1.8) 3.2 (1.7) 0.343
24-hr lactate 2.0 (1.0) 2.3 (1.4) 0.015 1.23 1.01–1.50 0.036

ICU: intensive care unit, SD: standard deviation, aOR: adjusted odds ratio, CI: confidence interval, CABG: coronary artery bypass grafting, DM: diabetes mellitus, IHD: ischemic heart disease, VHD: valvular heart disease, COPD: chronic obstructive pulmonary disease, CKD: chronic kidney disease, LVEF: left ventricular ejection fraction, Preop Hb: preoperative hemoglobin, CPB: cardiopulmonary bypass, ACC: aortic cross-clamp.

Model χ2 = 24.269, p = 0.002 and Hosmer and Lemeshow p = 0.686 indicates that the model fits the data. The classification table reports that overall expected model performance is 64.3%, that is, 64.3% of the cases can be expected to be classified correctly by the model.