Baseline characteristics of patients
There were 271 never-smokers and 268 ever-smokers in this cohort. Baseline characteristics are listed in Table1. The age of the smoker group is slightly higher than the age of the never-smoker group. Compared with the never-smokers, smokers had a significantly higher ASA classification and a significantly higher percentage of arrhythmia and COPD. Compared with the never-smokers, more patients in the smoking group received statin, steroids, and inhalational agents. Baseline characteristics of different smoking burden groups (pack-years) are listed in TableS1.
Table 1: Baseline characteristics
Variables
|
Never Smoker
(n=271)
|
Smoker
(n=268)
|
Standardize diff.
|
P-value
|
Age, years [mean (SD)]
|
58.88 (15.07)
|
61.96 (12.69)
|
0.22 (0.05, 0.39)
|
0.059
|
Male Gender
|
140 (51.66%)
|
151 (56.34%)
|
0.09 (-0.07, 0.26)
|
0.275
|
BMI, kg/m2 [mean (SD)]
|
32.05 (13.43)
|
31.46 (10.51)
|
0.05 (-0.12, 0.22)
|
0.469
|
Poor Functional Status (<4 MET)
|
133 (49.08%)
|
137 (51.12%)
|
0.04 (-0.13, 0.21)
|
0.635
|
ASA Classification
|
|
|
0.27 (0.10, 0.44)
|
0.009
|
II
|
28 (10.33%)
|
15 (5.60%)
|
|
|
III
|
192 (70.85%)
|
177 (66.04%)
|
|
|
IV
|
51 (18.82%)
|
76 (28.36%)
|
|
|
PHTN Severity Classification
|
|
|
0.11 (-0.07, 0.28)
|
0.479
|
Mild
|
118 (46.09%)
|
108 (41.38%)
|
|
|
Moderate
|
113 (44.14%)
|
129 (49.43%)
|
|
|
Severe
|
25 (9.77%)
|
24 (9.20%)
|
|
|
Surgical Characteristics
|
|
|
|
|
Length of Surgery [median (IQR)]
|
77.00 (34.00-141.00)
|
94.00 (36.75-159.25)
|
0.14 (-0.03, 0.31)
|
0.168
|
Open Surgical Approach
|
153 (56.46%)
|
131 (48.88%)
|
0.15 (-0.02, 0.32)
|
0.078
|
Intraabdominal
|
52 (19.19%)
|
60 (22.39%)
|
0.08 (-0.09, 0.25)
|
0.360
|
Intrathoracic
|
14 (5.17%)
|
13 (4.85%)
|
0.01 (-0.15, 0.18)
|
0.867
|
Vascular
|
6 (2.21%)
|
13 (4.85%)
|
0.14 (-0.03, 0.31)
|
0.097
|
Comorbidities
|
|
|
|
|
Systemic Hypertension
|
172 (63.47%)
|
187 (69.78%)
|
0.13 (-0.03, 0.30)
|
0.121
|
Coronary Artery Disease
|
88 (32.84%)
|
93 (34.83%)
|
0.04 (-0.13, 0.21)
|
0.626
|
Arrhythmia
|
136 (50.18%)
|
105 (39.18%)
|
0.22 (0.05, 0.39)
|
0.010
|
Angina
|
16 (5.90%)
|
23 (8.58%)
|
0.10 (-0.07, 0.27)
|
0.230
|
Asthma
|
44 (16.24%)
|
31 (11.61%)
|
0.13 (-0.04, 0.30)
|
0.121
|
COPD
|
10 (3.70%)
|
61 (22.85%)
|
0.59 (0.42, 0.76)
|
<0.001
|
Diabetes
|
77 (28.41%)
|
78 (29.21%)
|
0.02 (-0.15, 0.19)
|
0.838
|
Renal Failure
|
75 (27.68%)
|
60 (22.39%)
|
0.12 (-0.05, 0.29)
|
0.157
|
Medications
|
|
|
|
|
Anticoagulant
|
82 (30.26%)
|
65 (24.25%)
|
0.14 (-0.03, 0.30)
|
0.118
|
Antiplatelet
|
9 (3.32%)
|
9 (3.36%)
|
0.00 (-0.17, 0.17)
|
0.981
|
Statin
|
110 (40.59%)
|
132 (49.25%)
|
0.17 (0.01, 0.34)
|
0.043
|
Steroids
|
39 (14.39%)
|
61 (22.76%)
|
0.22 (0.05, 0.39)
|
0.012
|
Atropine
|
2 (0.74%)
|
3 (1.13%)
|
0.04 (-0.13, 0.21)
|
0.683
|
Inhalational Agents
|
137 (50.93%)
|
168 (63.40%)
|
0.25 (0.08, 0.42)
|
0.004
|
Isoflurane
|
6 (2.21%)
|
12 (4.48%)
|
0.13 (-0.04, 0.30)
|
0.144
|
Sevoflurane
|
111 (40.96%)
|
136 (50.75%)
|
0.20 (0.03, 0.37)
|
0.023
|
Delirium
|
5 (1.85%)
|
11 (4.10%)
|
0.13 (-0.04, 0.30)
|
0.122
|
Mortality
|
3 (1.11%)
|
4 (1.49%)
|
0.03 (-0.13, 0.20)
|
0.724
|
BMI body mass index, MET metabolic equivalent of task, ASA American Society of Anesthesiologists, PHTN pulmonary hypertension, IQR interquartile range, COPD chronic obstructive pulmonary disease
The relationship between the smoking status and POD
The univariate analysis showed that only age and intrathoracic surgery were correlated with POD (TableS2). To accurately evaluate the association between smoking status and POD, we exhibited the unadjusted and adjusted models in Table2. In the unadjusted model, tobacco smoking did not significantly correlate with POD (OR=2.28, 95% CI: 0.78 to 6.64, P=0.1320). However, smoking showed a positive correlation with POD in the minimally adjusted (OR=4.06, 95% CI: 1.26 to 13.07, P=0.0189) and fully adjusted models (OR=5.61, 95% CI: 1.14 to 27.51, P=0.0334).
Table 2: Relationship between smoking status and POD in different models
Variables
|
Unadjusted model (OR 95%CI P)
|
Minimally adjusted model (OR 95%CI P)
|
Fully adjusted model (OR 95%CI P)
|
Tobacco Smoking
|
|
|
|
Never smoker
|
ref
|
ref
|
ref
|
Smoker
|
2.28 (0.78, 6.64) 0.1320
|
4.06 (1.26, 13.07) 0.0189
|
5.61 (1.14, 27.51) 0.0334
|
Unadjusted model: We do not adjust for other covariances
Minimally adjusted model: We adjust for COPD, coronary artery disease, vascular surgery, systemic hypertension
Fully adjusted model: We adjust for age, gender, BMI, poor functional status, PHTN severity class, length of surgery, open surgical approach, intraabdominal surgery, intrathoracic surgery, vascular surgery, COPD, asthma, arrhythmia, systemic hypertension, angina, coronary artery disease, renal failure, diabetes, steroids, anticoagulant, atropine, antiplatelet, statin, anesthesia inhalational agents, isoflurane, sevoflurane
The relationship between the smoking burden and POD
Next, we compared the incidence of POD in different smoking burden groups. Compared with never-smokers (smoking burden=0), heavy smokers (smoking burden>50) show a significant increase in POD incidence both in the unadjusted and adjusted models (Table3).
Table 3: Relationship between smoking burden and POD in different models
Variables
|
Unadjusted model (OR 95%CI P)
|
Minimally adjusted model (OR 95%CI P)
|
Fully adjusted model (OR 95%CI P)
|
Pack-years
|
|
|
|
0
|
ref
|
ref
|
ref
|
>0, <=50
|
1.75 (0.50, 6.14) 0.3823
|
3.16 (0.48, 20.82) 0.2326
|
3.63 (0.44, 29.78) 0.2303
|
>50, <=100
|
3.32 (0.62, 17.85) 0.1611
|
15.59 (1.29, 188.07) 0.0307
|
12.30 (0.74, 204.90) 0.0803
|
>100
|
26.60 (2.06, 343.51) 0.0120
|
58.97 (0.87, 3990.03) 0.0580
|
90.08 (0.79, 10266.01) 0.0625
|
Unadjusted model: We do not adjust for other covariances
Minimally adjusted model: We adjust for age, gender, PHTN severity, length of surgery, open surgical approach, intrathoracic surgery, vascular surgery, COPD, asthma, coronary artery disease, angina, systemic hypertension, anticoagulant, anesthesia inhalational agents, isoflurane, sevoflurane
Fully adjusted model: We adjust for age, gender, BMI, poor functional status, PHTN severity, length of surgery, open surgical approach, intraabdominal surgery, intrathoracic surgery, vascular surgery, COPD, asthma, arrhythmia, systemic hypertension, angina, coronary artery disease, renal failure, diabetes, steroids, anticoagulant, atropine, antiplatelet, statin, anesthesia inhalational agents, isoflurane, sevoflurane
Since the smoking burden is continuous data, we explore the curve relationship between pack-years and POD. After adjusting several cofounders, we found a curvilinear relationship between smoking burden and POD is close to a linear relationship (Fig.1).