Skip to main content
. 2014 Dec 1;9(12):e113656. doi: 10.1371/journal.pone.0113656

Table 3. Results of logistic regression model to predict postoperative pulmonary complications (PPC) and assigned points based on regression coefficient to develop the PPC scoring index.

Univariable analysis Multivariable analysis
Variables With PPC (n = 140) Without PPC (n = 1919) OR (95% CI) p Regression coefficient OR (95% CI) p Assigned points
Age ≥70 yr 82 (59) 822 (43) 1.89 (1.33–2.67) <0.001 0.708 2.03 (1.37–3.00) <0.001 2
Sex, male 102 (73) 1186 (62) 1.66 (1.13–2.44) 0.010
Body mass index
<18.5 kg/m2 12 (9) 114 (6) 1.40 (0.68–2.89) 0.594
18.5–25.0 kg/m2 87 (62) 1158 (60) 1.00 Ref
>25.0 kg/m2 41 (29) 647 (34) 0.84 (0.54–1.31) 0.768
Current smoker within 2 months 33 (24) 267 (14) 1.91 (1.27–2.88) 0.002 0.482 1.62 (1.03–2.54) 0.036 1
Alcoholics 4 (3) 73 (4) 0.75 (0.27–2.08) 0.580
Malignancy 78 (56) 766 (40) 1.89 (1.34–2.68) <0.001
CHF 7 (5) 78 (4) 1.24 (0.56–2.75) 0.592
Confused mentality 2 (1) 33 (2) 0.83 (0.20–3.49) 0.797
Recent loss of body weight 11 (8) 65 (3) 2.44 (1.26–4.73) 0.009
ASA physical status ≥ class 2 109 (78) 1201 (63) 2.10 (1.40–3.17) <0.001 0.466 1.59 (1.01–2.52) 0.046 1
Presence of airflow limitation 68 (49) 545 (28) 2.38 (1.69–3.37) <0.001 0.558 1.75 (1.18–2.58) 0.005 1
Serum albumin <4.0 g/dl 61 (44) 547 (29) 1.95 (1.37–2.77) <0.001 0.398 1.49 (1.00–2.21) 0.049 1
Serum hemoglobin <10 g/dl 16 (11) 129 (7) 1.79 (1.03–3.10) 0.039
Active lesion on chest radiography 18 (13) 192 (10) 1.33 (0.79–2.22) 0.284
General anesthesia 129 (92) 1640 (85) 2.00 (1.06–3.74) 0.031
Emergency surgery 10 (7) 74 (4) 1.92 (0.97–3.80) 0.062 0.768 2.16 (1.00–4.65) 0.050 2
Surgical specialty
Others 32 (23) 969 (50) 1.00 Ref 1.00
Cardiac/aortic aneurysm repair/abdominal open surgery 89 (64) 415 (22) 6.49 (3.89–10.85) <0.001 1.949 7.02 (4.04–12.19) <0.001 4
Laparoscopic abdominal 12 (9) 219 (11) 1.66 (0.72–3.81) 0.432 0.551 1.74 (0.74–4.09) 0.369
Neurosurgery 7 (5) 316 (16) 0.67 (0.24–1.84) 1.000 −0.268 0.77 (0.25–2.30) 1.000

A multivariable analysis using a logistic regression model was conducted with backward step-wise selection with p<0.05 for inclusion of variables and p>0.10 for removal of variables from all variables used for Univariable analysis and only significant variables in a multivariable analysis were indicated in the table.

PPC, postoperative pulmonary complications; OR, odds ratio; CI, confidence interval; CHF, congestive heart failure; ASA, American Society of Anesthesiologists.

Hosmer-Lemeshow test, χ2 = 7.653, p = 0.468.