SURGICAL INFECTIONS Volume 15, Number 3, 2014 ª Mary Ann Liebert, Inc. DOI: 10.1089/sur.2013.031
Risk Factors for Post-Operative Pulmonary Complications after Gastrectomy for Gastric Cancer Mikito Inokuchi,1 Kazuyuki Kojima,2 Keiji Kato,1 Hirofumi Sugita,1 and Kenichi Sugihara1
Abstract
Background: Post-operative pulmonary complications (PPCs) negatively affect patients’ quality of life and can be life-threatening. Predictors of PPCs have been evaluated in patients who underwent various operations, but few studies have specifically focused on gastrectomy. Methods: We retrospectively studied 1,053 patients with gastric adenocarcinoma who underwent radical gastrectomy with lymphadenectomy in our hospital between 1999 and 2011. Post-operative pulmonary complications were defined as conditions such as pneumonia, macroscopic atelectasis, pneumothorax, and acute respiratory distress syndrome that developed within 30 d after surgery. We evaluated the relations between PPCs and pre-operative or intra-operative factors and assessed risk factors for PPCs after gastrectomy. Result: A total of 49 (4.7%) patients had PPCs. On univariate analysis, PPCs were significantly associated with male gender (p = 0.024), predicted vital capacity (VC) (p = 0.020), a lower pre-operative serum albumin concentration (p = 0.023), open surgery (p = 0.007), total gastrectomy (p < 0.001), combined resection of another organ (p = 0.001), extended operating time (p < 0.001), higher operative bleeding volume (p < 0.001), intra-operative or post-operative blood transfusion (p = 0.009), and pathologic tumor stage (p = 0.003). On multivariable analysis, extended operating time (odds ratio [OR], 3.21, 95% confidence interval [CI] 1.46– 7.07; p = 0.004), total gastrectomy (OR, 2.65, 95% CI 1.25–5.59; p = 0.011) and predicted VC (OR, 2.42, 95% CI 1.01–5.85; p = 0.049) were independent risk factors. These three factors also were independent risk factors for post-operative pneumonia (total gastrectomy OR, 2.64, 95% CI 1.32–5.30; p = 0.006); extended operating time OR, 2.54, 95% CI 1.24–5.19; p = 0.011; and predicted VC OR, 2.41, 95% CI 1.01–5.75; p = 0.048). Conclusion: Extended operating time, total gastrectomy, and predicted VC were independent predictors of PPCs, particularly pneumonia, in patients with gastric cancer who underwent gastrectomy. In patients with restrictive pulmonary dysfunction who are scheduled to undergo total gastrectomy, reduced lymphadenectomy or the avoidance of combined resection should be considered to shorten the operating time.
P
ost-operative pulmonary complications (PPCs) negatively affect patients’ quality of life and can even be life-threatening. Post-operative pulmonary complications increase hospital morbidity, prolong the hospital stay, and contribute to additional healthcare costs [1]. Identification of risk factors for PPCs might contribute to a reduction in PPCs. Arozullah et al. performed a large-scale study of approximately 160,000 patients and proposed a post-operative pneumonia risk index that included many factors related to general health, immune status, respiratory status, neurologic status, and type of surgery [2]. However, their subjects underwent various operations, with the exclusion of cardiac surgery. In addition, the risk index they proposed was complex because it included many risk factors with variable ranges. 1
Gastric cancer is the fourth most common cancer worldwide [3]. Radical gastrectomy with lymphadenectomy is a standard procedure used widely to treat gastric cancer. Gastrectomy is one of the high-risk operations for PPCs in nonthoracic surgery. Upper abdominal procedures, prolonged duration of anesthesia, and placement of nasogastric tube are essential factors for gastrectomy. These factors also are independent risk factors for PPCs after non-thoracic surgery [2,4–6]. However, there are few reports on risk factors for various PPCs after radical gastrectomy with lymphadenectomy. The incidences of pneumonia after gastrectomy with D2 or more radical lymphadenectomy were about 2% to 3% in Japanese and Korean studies [7,8]. In open gastrectomy, intra-operative or post-operative blood transfusion and
Department of Surgical Oncology, 2Department of Minimally Invasive Surgery, Tokyo Medical and Dental University, Tokyo, Japan.
314
POST-OPERATIVE PULMONARY COMPLICATIONS AFTER GASTRECTOMY
male gender were independent risk factors for post-operative pneumonia [9]. Total gastrectomy may increase the risk of PPCs because of the extended operating time, voluminous blood loss, and intense pain from the incision. By contrast, laparoscopy-assisted gastrectomy has been established as a minimally invasive procedure for gastric cancer [10,11], although carbon dioxide used to create pneumoperitoneum and extended operating time might have unfavorable cardiopulmonary effects [12]. In colorectal surgery, laparoscopic colectomy is associated with a significantly lower risk of pneumonia than open colectomy [13]. Whether laparoscopic gastrectomy affects the incidence of PPCs remains a matter of debate, although it seems to be feasible even in patients with chronic obstructive pulmonary disease (COPD) [14,15]. We performed this retrospective cohort study to identify risk factors for PPCs after radical gastrectomy for gastric cancer, including both open and laparoscopic procedures. Patients and Methods
We identified retrospectively 1,053 consecutive patients from who underwent radical gastrectomy (R0 or R1) with lymphadenectomy for gastric adenocarcinoma in our hospital between 1999 (the year we started to perform laparoscopyassisted gastrectomy) and 2011. Patients who underwent thoracoabdominal incision or emergency surgery were excluded. Patient characteristics are shown in Table 1. All smokers were instructed to stop smoking at least two weeks before surgery. All patients with pulmonary dysfunction were assessed by spirometric testing performed pre-operatively. The surgical approach was classified as laparoscopic or open, and the extent of lymph node dissection was classified as D1, D1 + (a or b), or D2 in accordance with the treatment guidelines of the Japanese Gastric Cancer Association [16]. In principle, early-stage gastric cancer was treated by laparoscopy-assisted gastrectomy. Splenectomy was performed principally in patients with advanced cancer arising in the upper third of the stomach. In patients who underwent laparoscopic surgery, carbon dioxide pneumoperitoneum was maintained at 10 mm Hg, and a 4- to 5-cm incision was made in the upper abdomen or navel to remove tissue specimens and conduct anastomosis. For laparotomy, a midline or subcostal incision was made. All patients received systemic antibiotics (a first-generation cephalosporin) several times on the day of surgery. All patients received continuous epidural anesthesia for 48 h post-operatively whenever possible. The nasogastric tube was left in place until post-operative day 1. All patients pre-operatively underwent chest radiography, venous blood analysis, and pulmonary function testing, including vital capacity (VC), forced expiratory volume in 1 sec (FEV1), and forced vital capacity (FVC). Other evaluations of pulmonary function, such as arterial blood gas analysis, were performed in selected patients as required. These data were obtained from our medical records along with other information, including recent history of smoking (within 60 d before surgery), regular use of steroids, and the presence of gait disturbance. The following other parameters were obtained from our prospective database: Patient age and gender; body mass index; concurrent illness; tumor characteristics; extent of lymph node dissection; operation time; estimated blood loss; peri-operative blood transfusion; and
315
Table 1. Patient Characteristics n (%) Gender Male Female Age (mean – SD) Body mass index (mean – SD) Current smoking Predicted VC (%, mean – SD) FEV1/ FVC (%, mean – SD) Anemia Serum albumin (g/dL mean – SD) Other comorbidities Diabetes mellitus Cardiac disease Cerebral disease Liver disease Elevated pre-operative creatinine Regular use of steroids Gait disturbance Approach Open Laparoscopic Type of resection Total Proximal Distal Extent of LN dissection D1 D1 + D2 Combined resection Spleen Gallbladder Intestine/colon Pancreas body and tail Operating time (min, mean – SD) Bleeding (mL, mean – SD) Blood transfusion Intra-/post-operative Pre-operative Pathologic tumor stage I II III IV (CY1)
778 (74) 275 (26) 65.4 – 11.3 22.6 – 3.2 321 (30) 106.6 –17.0 74.7 – 9.7 486 (46) 4.1 – 0.4 146 121 85 39 111 9 36
(14) (12) ( 8.1) ( 3.7) (10.5) ( 0.9) ( 3.4)
469 584
(44.5) (55.5)
238 56 759
(22.6) ( 5.3) (72.1)
24 699 330
( 2.3) (66.4) (31.3)
112 107 27 7 279 329
(10.6) (10.2) ( 2.6) ( 0.7) – 74 – 436
66 13
( 6.3) ( 1.2)
664 172 209 8
(63.1) (16.3) (19.8) ( 0.8)
SD = standard deviation; VC = vital capacity; FEV1 = forced expiratory volume in 1 sec; FVC = forced vital capacity; LN = lymph node; CY1 = positive peritoneal cytology.
post-operative complications. Post-operative pulmonary complications were defined as the development of pneumonia, macroscopic atelectasis, pneumothorax, prolonged mechanical ventilation (longer than 24 h), mass pleural effusion requiring puncture, or acute respiratory distress syndrome (ARDS) within 30 d after surgery. We performed chest radiography in all patients on post-operative days zero, one, three or four, and seven. Pneumonia was diagnosed on the basis of symptoms and the results of blood analysis, chest radiography, or computed tomography. Pneumonia met the criteria proposed by the U. S. Centers for Disease Control and Prevention [17]. Macroscopic atelectasis was diagnosed only
316
by chest radiography. We evaluated the relations between PPCs and various pre-operative or intra-operative factors to identify risk factors. Statistical analysis
Risk factors related to complications were investigated by univariate and multi-variable analyses. Pearson chi-square test or Fisher exact test were used to estimate p values, according to the expected count. A p value of < 0.05 was considered to indicate statistical significance. Predictive factors for complications were evaluated by binary logistic multiple regression analysis using dummy variables with p < 0.10. No variables with strong linear correlations were confirmed by the coefficient of association before multi-variable analysis. Seventeen patients (1.6%) were excluded from the multivariable analysis because of missing data. All analyses were performed with the statistical software package SPSS 17 (SPSS Japan Inc., Tokyo, Japan). Results
A total of 49 (4.7%) patients had PPCs. There were no PPC-related deaths. Pneumonia, macroscopic atelectasis, and acute respiratory distress syndrome developed in 23 (2.2%), 20 (1.9%), and four (0.4%) patients, respectively. Patients who only had macroscopic atelectasis inhaled oxygen to treat hypoxemia or a nebulizer was used to facilitate the expectoration of sputum. Another patient had massive pleural effusion and underwent drainage. An additional patient received mechanical ventilation for more than 24 h after surgery, until the tidal volume was high enough to allow liberation. All PPCs and clinical factors
On univariate analysis, PPCs were associated significantly with male gender (p = 0.024), lower predicted (VC < 80 %; p = 0.020), a lower pre-operative serum albumin concentration ( < 3.5 g/dL; p = 0.023), open surgery (p = 0.007), total gastrectomy (p < 0.001), combined resection of other organ (excluding gallbladder, p = 0.001), extended operating time ( ‡ 6 h; p < 0.001), higher operative blood loss ( ‡ 500 mL; p < 0.001), intra-operative or post-operative blood transfusion (p = 0.009), and advanced pathologic tumor stage ( ‡ II; p = 0.003; Table 2). Current smoking was not significantly related to PPCs (p = 0.058). We evaluated independent risk factors for PPCs using a multi-variable model adjusted for the following variables: Gender, predicted VC, operative approach, type of resection, combined resection, operating time, operative blood loss, blood transfusion, pathologic tumor stage, current smoking, and pre-operative serum albumin concentration. The largest absolute value of the phi co-efficient among these variables was 0.55 between the type of resection and combined resection, and no strong correlation was found among the dummy variables (Table 3). Three independent risk factors were significantly related to PPCs: Extended operating time (odds ratio [OR], 3.21, 95% confidential interval [CI] 1.46–7.07; p = 0.004), total gastrectomy (OR, 2.65, 95% CI 1.25–5.59; p = 0.011) and predicted VC (OR, 2.42, 95% CI 1.01–5.85; p = 0.049; Table 4).
INOKUCHI ET AL. Post-operative pneumonia and clinical factors
On univariate analysis, post-operative pneumonia was associated significantly with male gender (p = 0.004), older age ( ‡ 80 years; p = 0.031), lower predicted VC (p < 0.001), total gastrectomy (p < 0.001), extended operating time (p = 0.012), high operative blood loss (p = 0.026), and pathologic tumor stage (p < 0.001). Anemia was not related to post-operative pneumonia (p = 0.064;Table 2). On multivariable analysis, independent risk factors for post-operative pneumonia were same as those for PPCs: extended operating time (OR, 2.64, 95% CI 1.32–5.30; p = 0.006), total gastrectomy (OR, 2.54, 95% CI 1.24–5.19; p = 0.011) and predicted VC (OR, 2.41, 95% CI 1.01–5.75; p = 0.048;Table 4). Post-operative macroscopic atelectasis and clinical factors
On univariate analysis, post-operative macroscopic atelectasis was associated significantly with higher body mass index ( ‡ 25; p = 0.023), a lower pre-operative serum albumin concentration (p = 0.026), open surgery (p = 0.006), total gastrectomy (p = 0.027), combined resection of an other organ (p = 0.037), higher operative blood loss (p = 0.016), and intraoperative or post-operative blood transfusion (p =0.001; Table 2). On multivariable analysis, total gastrectomy was the only independent risk factor for post-operative macroscopic atelectasis (OR, 3.14, 95% CI 1.54–6.40, p = 0.002; Table 4). Discussion
Our results showed that extended operating time ( ‡ 6 h), total gastrectomy, and lower predicted VC ( < 80%) were independent risk factors for PPCs, after gastrectomy particularly pneumonia. The results of other studies that included non-thoracic surgery are similar to ours and showed that extended anesthesia or operative time was an independent risk factor [2,4,5]. In open gastrectomy, two studies showed that an extended operative time (5–6 h or longer) is an independent risk factor for various post-operative complications [7,18]. Risk factors for PPCs should be evaluated independently for each type of surgery but few large studies of gastrectomy have been performed to date. Total gastrectomy was not significantly related to post-operative pneumonia in another study [9]. In total gastrectomy, incisional pain might decrease pulmonary volume, or dissection around the esophageal hiatus and mediastinum might reduce diaphragm mobility, potentially leading to PPCs. Pneumonia after gastrectomy has been attributed not only to the direct effect of operative stress or immunosuppression but also to the aspiration of esophageal reflux contents [19]. We suspected that most cases of post-operative pneumonia were caused by aspiration in our study, although the differential diagnosis of aspiration pneumonia is often challenging. On the other hand, combined resection including concurrent splenectomy did not increase appreciably the risk ratio for PPCs in patients who underwent total gastrectomy in the present study. This finding is consistent with the results of other studies showing that adjacent organ resection, such as splenectomy, does not increase the risk of pneumonia [9] and that splenectomy is not an independent risk factor for any post-operative complication [7]. Sano et al. showed that extended lymphadenectomy did not increase post-operative
317
Gender Male 778 Female 275 Age ‡ 80 84 < 80 969 Body mass index (kg/m2) ‡ 25 226 < 25 825 Current smoking Yes 321 No 703 Unknown 29 Predicted VC (%) < 80 76 ‡ 80 962 Not evaluated 15 FEV1/ FVC (%) < 70 260 ‡ 70 778 Not evaluated 15 Anemia Yes 486 No 567 Preoperative serum albumin (g/dL) < 3.5 127 S3.5 921 Not evaluated 5 Diabetes mellitus Yes 146 No 907 Cardiac disease Yes 121 No 932 Cerebral disease Yes 85 No 968
n
2.20 (1.10–4.43) 1 0.86 (0.36–2.06) 1 0.87 (0.34–2.24) 1 1.98 (0.86–4.55) 1.00
11 ( 8.7) 38 ( 4.1) 0 6 ( 4.1) 43 ( 4.7) 5 ( 4.1) 44 ( 4.7) 7 ( 8.2) 42 ( 4.3)
2.71 (1.22–6.02) 1
8 (11.00) 40 ( 4.2) 1
1.46 (0.82–2.59) 1.00
1.75 (0.98–3.15) 1
21 ( 6.5) 27 ( 3.8) 1
27 ( 5.6) 22 ( 3.9)
1.34 (0.70–2.57) 1
13 ( 5.8) 36 ( 4.4)
1.38 (0.74–2.59) 1
1.66 (0.68–4.01) 1
6 ( 7.1) 43 ( 4.4)
15 ( 5.8) 33 ( 4.2) 1
2.62 (1.10–6.23) 1
OR (95%CI)
43 ( 5.5) 6 ( 2.2)
n (%)
All PPCs
0.10
0.77
0.74
0.023
0.20
0.31
0.020
0.058
0.38
0.27
0.024
p
4 (4.7) 19 (2.0)
4 (3.3) 19 (2.0)
5 (3.4) 18 (2.0)
4 (3.1) 19 (2.1) 0
15 ( 3.1) 8 (1.4)
5 (1.9) 17 (2.2) 1
6 (7.7) 16 (1.7) 1
10 (3.1) 13 (1.8) 0
3 (1.3) 20 (2.4)
5 (6.0) 18 (1.9)
23 (3.0) 0 (0)
n (%)
2.47 (0.82– 7.42) 1
1.64 (0.55– 4.91) 1
1.75 (0.64– 4.79) 1
1.54 (0.52– 4.61) 1
2.23 (0.-4– 5.30) 1
0.88 (0.32– 2.40) 1
5.07 (1.92–13.4) 1
1.71 (0.74– 3.93) 1
0.54 (0.16– 1.84) 1
3.34 (1.21– 9.25) 1
1.03 (1.02– 1.04) 1
OR (95%CI)
Pneumonia
0.11
0.21
0.24
0.51
0.064
1 (1.2) 19 (2.0)
1 (0.8) 19 (2.0)
0 (0.0) 20 (2.2)
6 (4.7) 14 (1.5) 0
10 (2.1) 10 (1.8)
7 (2.7) 13 0
2 (2.6) 18 (1.9) 0
< 0.001
0.80
8 (2.5) 11 (1.6) 0
9 (4.0) 11 (1.3)
1 (1.2) 19 (2.0)
15 (1.9) 5 (1.8)
n (%)
1.00
0.72
0.10
0.026
0.73
0.30
0.65
0.31
0.023
1.00
0.91
p
(continued)
0.60 (0.08– 4.50) 1
0.40 (0.05– 3.02) 1
0.98 (0.97– 0.99) 1
3.21 (1.21– 8.52) 1
1.17 (0.48– 2.84) 1
1.63 (0.64– 4.13) 1 (1.7)
1.42 (0.32– 6.23) 1
1.61 (0.64– 4.04) 1
3.07 (1.26– 7.50) 1
0.60 (0.80– 4.56) 1
1.06 (0.38– 2.95) 1
OR (95%CI)
Macroscopic atelectasis
0.21
0.44
0.031
0.004
p
Table 2. Uni-Variate Analysis of All PPCs, Pneumonia, and Macroscopic Atelectasis
318
0.95 (0.94–0.97) 1 1.44 (0.63–3.29) 1 0.95 (0.94–0.97) 1 1.21 (0.28–5.21) 1 2.22 (1.23–4.03) 1 4.22 (2.36–7.55) 1 1.07 (0.58–1.96) 1 2.84 (1.49–5.42) 1 3.28 (1.71–6.30) 1 3.00 (1.64–5.46) 1 3.18 (1.43–7.10) 1 2.38 (1.33–4.24) 1
7 ( 6.3) 42 ( 4.5) 0 ( 0.0) 49 ( 4.7) 2 ( 5.6) 47 ( 4.6) 31 ( 6.6) 18 ( 3.1) 26 (12) 23 ( 2.8) 16 ( 4.8) 33 ( 4.6) 14 (10) 35 ( 3.8) 14 (11) 35 ( 3.8) 19 ( 9.5) 29 ( 3.4) 1 8 (12) 41 ( 4.2) 28 ( 7.2) 21 ( 3.2)
OR (95%CI)
0 ( 0.0) 49 ( 4.8)
n (%)
9 (4.5) 14 (1.6) 0
< 0.001
0.003
17 (4.4) 6 (0.9)
2 (3.0) 21 (2.1)
7 (5.7) 16 (1.7)
< 0.001
0.009
4 (2.9) 19 (2.1)
0.001
8 (2.4) 15 (2.1)
13 (5.5) 10 (1.2)
< 0.001 0.84
13 (2.8) 10 (1.7)
2 (5.6) 21 (2.1)
0 (0.0) 23 (2.2)
4 (3.6) 19 (2.0)
0 (0.0) 23 (2.3)
n (%)
0.007
0.79
1.00
0.38
0.15
p
5.01 (1.96–12.8) 1
1.44 (0.33– 6.27) 1
2.84 (1.21– 6.65) 1
3.45 (1.39– 8.56) 1
1.41 (0.47– 4.20) 1
1.17 (0.49– 2.79) 1
4.65 (2.01–10.7) 1
1.64 (0.71– 3.77) 1
2.79 (0.63–12.4) 1
0.98 (0.97– 0.99) 1
1.82 (0.61– 5.44) 1
0.98 (0.97– 0.99) 1
OR (95%CI)
Pneumonia
6 (9.1) 14 (1.4) 9 (2.3) 11 (1.7)
< 0.001
8 (4.0) 11 (1.3) 1
4 (3.3) 16 (1.7)
6 (4.3) 14 (1.5)
6 (1.8) 14 (1.9)
9 (3.8) 11 (1.3)
15 (4.8) 5 (0.9)
0 20 (2.0)
0 20 (1.9)
0 20 (2.1)
0 20 (2.0)
n (%)
0.65
0.026
0.012
0.53
0.72
< 0.001
0.24
0.18
1.00
0.29
1.00
p
1.41 (0.58– 3.42) 1
6.95 (2.58–18.7) 1
3.20 (1.27– 8.07) 1
1.92 (0.63– 5.84) 1
2.93 (1.11– 7.75) 1
0.94 (0.36– 2.46) 1
2.87 (1.18– 7.02) 1
3.83 (1.38–10.6) 1
0.98 (0.97– 0.99) 1
0.98 (0.97– 0.99) 1
0.98 (0.97– 0.99) 1
0.98 (0.97– 0.99) 1
OR (95%CI)
Macroscopic atelectasis
PPCs = post-operative pulmonary complication; OR = odds ratio; CI = confidence interval; VC = vital capacity; FEV1 = forced expiratory volume in 1 sec; FVC = forced vital capacity.
Liver disease Yes 39 No 1014 Elevated pre-operative creatinine Yes 111 No 942 Regular use of steroids Yes 9 No 1044 Gait disturbance Yes 36 No 1017 Operative approach Open 469 Laparoscopic 584 Type of resection Total 238 Partial 815 Extent of lymph node dissection D2 330 < D2 723 Combined resection Yes 138 None or only gallbladder 915 Surgery duration (h) ‡6 123
Risk Factors for Post-Operative Pulmonary Complications after Gastrectomy for Gastric Cancer Mikito Inokuchi,1 Kazuyuki Kojima,2 Keiji Kato,1 Hirofumi Sugita,1 and Kenichi Sugihara1
Abstract
Background: Post-operative pulmonary complications (PPCs) negatively affect patients’ quality of life and can be life-threatening. Predictors of PPCs have been evaluated in patients who underwent various operations, but few studies have specifically focused on gastrectomy. Methods: We retrospectively studied 1,053 patients with gastric adenocarcinoma who underwent radical gastrectomy with lymphadenectomy in our hospital between 1999 and 2011. Post-operative pulmonary complications were defined as conditions such as pneumonia, macroscopic atelectasis, pneumothorax, and acute respiratory distress syndrome that developed within 30 d after surgery. We evaluated the relations between PPCs and pre-operative or intra-operative factors and assessed risk factors for PPCs after gastrectomy. Result: A total of 49 (4.7%) patients had PPCs. On univariate analysis, PPCs were significantly associated with male gender (p = 0.024), predicted vital capacity (VC) (p = 0.020), a lower pre-operative serum albumin concentration (p = 0.023), open surgery (p = 0.007), total gastrectomy (p < 0.001), combined resection of another organ (p = 0.001), extended operating time (p < 0.001), higher operative bleeding volume (p < 0.001), intra-operative or post-operative blood transfusion (p = 0.009), and pathologic tumor stage (p = 0.003). On multivariable analysis, extended operating time (odds ratio [OR], 3.21, 95% confidence interval [CI] 1.46– 7.07; p = 0.004), total gastrectomy (OR, 2.65, 95% CI 1.25–5.59; p = 0.011) and predicted VC (OR, 2.42, 95% CI 1.01–5.85; p = 0.049) were independent risk factors. These three factors also were independent risk factors for post-operative pneumonia (total gastrectomy OR, 2.64, 95% CI 1.32–5.30; p = 0.006); extended operating time OR, 2.54, 95% CI 1.24–5.19; p = 0.011; and predicted VC OR, 2.41, 95% CI 1.01–5.75; p = 0.048). Conclusion: Extended operating time, total gastrectomy, and predicted VC were independent predictors of PPCs, particularly pneumonia, in patients with gastric cancer who underwent gastrectomy. In patients with restrictive pulmonary dysfunction who are scheduled to undergo total gastrectomy, reduced lymphadenectomy or the avoidance of combined resection should be considered to shorten the operating time.
P
ost-operative pulmonary complications (PPCs) negatively affect patients’ quality of life and can even be life-threatening. Post-operative pulmonary complications increase hospital morbidity, prolong the hospital stay, and contribute to additional healthcare costs [1]. Identification of risk factors for PPCs might contribute to a reduction in PPCs. Arozullah et al. performed a large-scale study of approximately 160,000 patients and proposed a post-operative pneumonia risk index that included many factors related to general health, immune status, respiratory status, neurologic status, and type of surgery [2]. However, their subjects underwent various operations, with the exclusion of cardiac surgery. In addition, the risk index they proposed was complex because it included many risk factors with variable ranges. 1
Gastric cancer is the fourth most common cancer worldwide [3]. Radical gastrectomy with lymphadenectomy is a standard procedure used widely to treat gastric cancer. Gastrectomy is one of the high-risk operations for PPCs in nonthoracic surgery. Upper abdominal procedures, prolonged duration of anesthesia, and placement of nasogastric tube are essential factors for gastrectomy. These factors also are independent risk factors for PPCs after non-thoracic surgery [2,4–6]. However, there are few reports on risk factors for various PPCs after radical gastrectomy with lymphadenectomy. The incidences of pneumonia after gastrectomy with D2 or more radical lymphadenectomy were about 2% to 3% in Japanese and Korean studies [7,8]. In open gastrectomy, intra-operative or post-operative blood transfusion and
Department of Surgical Oncology, 2Department of Minimally Invasive Surgery, Tokyo Medical and Dental University, Tokyo, Japan.
314
POST-OPERATIVE PULMONARY COMPLICATIONS AFTER GASTRECTOMY
male gender were independent risk factors for post-operative pneumonia [9]. Total gastrectomy may increase the risk of PPCs because of the extended operating time, voluminous blood loss, and intense pain from the incision. By contrast, laparoscopy-assisted gastrectomy has been established as a minimally invasive procedure for gastric cancer [10,11], although carbon dioxide used to create pneumoperitoneum and extended operating time might have unfavorable cardiopulmonary effects [12]. In colorectal surgery, laparoscopic colectomy is associated with a significantly lower risk of pneumonia than open colectomy [13]. Whether laparoscopic gastrectomy affects the incidence of PPCs remains a matter of debate, although it seems to be feasible even in patients with chronic obstructive pulmonary disease (COPD) [14,15]. We performed this retrospective cohort study to identify risk factors for PPCs after radical gastrectomy for gastric cancer, including both open and laparoscopic procedures. Patients and Methods
We identified retrospectively 1,053 consecutive patients from who underwent radical gastrectomy (R0 or R1) with lymphadenectomy for gastric adenocarcinoma in our hospital between 1999 (the year we started to perform laparoscopyassisted gastrectomy) and 2011. Patients who underwent thoracoabdominal incision or emergency surgery were excluded. Patient characteristics are shown in Table 1. All smokers were instructed to stop smoking at least two weeks before surgery. All patients with pulmonary dysfunction were assessed by spirometric testing performed pre-operatively. The surgical approach was classified as laparoscopic or open, and the extent of lymph node dissection was classified as D1, D1 + (a or b), or D2 in accordance with the treatment guidelines of the Japanese Gastric Cancer Association [16]. In principle, early-stage gastric cancer was treated by laparoscopy-assisted gastrectomy. Splenectomy was performed principally in patients with advanced cancer arising in the upper third of the stomach. In patients who underwent laparoscopic surgery, carbon dioxide pneumoperitoneum was maintained at 10 mm Hg, and a 4- to 5-cm incision was made in the upper abdomen or navel to remove tissue specimens and conduct anastomosis. For laparotomy, a midline or subcostal incision was made. All patients received systemic antibiotics (a first-generation cephalosporin) several times on the day of surgery. All patients received continuous epidural anesthesia for 48 h post-operatively whenever possible. The nasogastric tube was left in place until post-operative day 1. All patients pre-operatively underwent chest radiography, venous blood analysis, and pulmonary function testing, including vital capacity (VC), forced expiratory volume in 1 sec (FEV1), and forced vital capacity (FVC). Other evaluations of pulmonary function, such as arterial blood gas analysis, were performed in selected patients as required. These data were obtained from our medical records along with other information, including recent history of smoking (within 60 d before surgery), regular use of steroids, and the presence of gait disturbance. The following other parameters were obtained from our prospective database: Patient age and gender; body mass index; concurrent illness; tumor characteristics; extent of lymph node dissection; operation time; estimated blood loss; peri-operative blood transfusion; and
315
Table 1. Patient Characteristics n (%) Gender Male Female Age (mean – SD) Body mass index (mean – SD) Current smoking Predicted VC (%, mean – SD) FEV1/ FVC (%, mean – SD) Anemia Serum albumin (g/dL mean – SD) Other comorbidities Diabetes mellitus Cardiac disease Cerebral disease Liver disease Elevated pre-operative creatinine Regular use of steroids Gait disturbance Approach Open Laparoscopic Type of resection Total Proximal Distal Extent of LN dissection D1 D1 + D2 Combined resection Spleen Gallbladder Intestine/colon Pancreas body and tail Operating time (min, mean – SD) Bleeding (mL, mean – SD) Blood transfusion Intra-/post-operative Pre-operative Pathologic tumor stage I II III IV (CY1)
778 (74) 275 (26) 65.4 – 11.3 22.6 – 3.2 321 (30) 106.6 –17.0 74.7 – 9.7 486 (46) 4.1 – 0.4 146 121 85 39 111 9 36
(14) (12) ( 8.1) ( 3.7) (10.5) ( 0.9) ( 3.4)
469 584
(44.5) (55.5)
238 56 759
(22.6) ( 5.3) (72.1)
24 699 330
( 2.3) (66.4) (31.3)
112 107 27 7 279 329
(10.6) (10.2) ( 2.6) ( 0.7) – 74 – 436
66 13
( 6.3) ( 1.2)
664 172 209 8
(63.1) (16.3) (19.8) ( 0.8)
SD = standard deviation; VC = vital capacity; FEV1 = forced expiratory volume in 1 sec; FVC = forced vital capacity; LN = lymph node; CY1 = positive peritoneal cytology.
post-operative complications. Post-operative pulmonary complications were defined as the development of pneumonia, macroscopic atelectasis, pneumothorax, prolonged mechanical ventilation (longer than 24 h), mass pleural effusion requiring puncture, or acute respiratory distress syndrome (ARDS) within 30 d after surgery. We performed chest radiography in all patients on post-operative days zero, one, three or four, and seven. Pneumonia was diagnosed on the basis of symptoms and the results of blood analysis, chest radiography, or computed tomography. Pneumonia met the criteria proposed by the U. S. Centers for Disease Control and Prevention [17]. Macroscopic atelectasis was diagnosed only
316
by chest radiography. We evaluated the relations between PPCs and various pre-operative or intra-operative factors to identify risk factors. Statistical analysis
Risk factors related to complications were investigated by univariate and multi-variable analyses. Pearson chi-square test or Fisher exact test were used to estimate p values, according to the expected count. A p value of < 0.05 was considered to indicate statistical significance. Predictive factors for complications were evaluated by binary logistic multiple regression analysis using dummy variables with p < 0.10. No variables with strong linear correlations were confirmed by the coefficient of association before multi-variable analysis. Seventeen patients (1.6%) were excluded from the multivariable analysis because of missing data. All analyses were performed with the statistical software package SPSS 17 (SPSS Japan Inc., Tokyo, Japan). Results
A total of 49 (4.7%) patients had PPCs. There were no PPC-related deaths. Pneumonia, macroscopic atelectasis, and acute respiratory distress syndrome developed in 23 (2.2%), 20 (1.9%), and four (0.4%) patients, respectively. Patients who only had macroscopic atelectasis inhaled oxygen to treat hypoxemia or a nebulizer was used to facilitate the expectoration of sputum. Another patient had massive pleural effusion and underwent drainage. An additional patient received mechanical ventilation for more than 24 h after surgery, until the tidal volume was high enough to allow liberation. All PPCs and clinical factors
On univariate analysis, PPCs were associated significantly with male gender (p = 0.024), lower predicted (VC < 80 %; p = 0.020), a lower pre-operative serum albumin concentration ( < 3.5 g/dL; p = 0.023), open surgery (p = 0.007), total gastrectomy (p < 0.001), combined resection of other organ (excluding gallbladder, p = 0.001), extended operating time ( ‡ 6 h; p < 0.001), higher operative blood loss ( ‡ 500 mL; p < 0.001), intra-operative or post-operative blood transfusion (p = 0.009), and advanced pathologic tumor stage ( ‡ II; p = 0.003; Table 2). Current smoking was not significantly related to PPCs (p = 0.058). We evaluated independent risk factors for PPCs using a multi-variable model adjusted for the following variables: Gender, predicted VC, operative approach, type of resection, combined resection, operating time, operative blood loss, blood transfusion, pathologic tumor stage, current smoking, and pre-operative serum albumin concentration. The largest absolute value of the phi co-efficient among these variables was 0.55 between the type of resection and combined resection, and no strong correlation was found among the dummy variables (Table 3). Three independent risk factors were significantly related to PPCs: Extended operating time (odds ratio [OR], 3.21, 95% confidential interval [CI] 1.46–7.07; p = 0.004), total gastrectomy (OR, 2.65, 95% CI 1.25–5.59; p = 0.011) and predicted VC (OR, 2.42, 95% CI 1.01–5.85; p = 0.049; Table 4).
INOKUCHI ET AL. Post-operative pneumonia and clinical factors
On univariate analysis, post-operative pneumonia was associated significantly with male gender (p = 0.004), older age ( ‡ 80 years; p = 0.031), lower predicted VC (p < 0.001), total gastrectomy (p < 0.001), extended operating time (p = 0.012), high operative blood loss (p = 0.026), and pathologic tumor stage (p < 0.001). Anemia was not related to post-operative pneumonia (p = 0.064;Table 2). On multivariable analysis, independent risk factors for post-operative pneumonia were same as those for PPCs: extended operating time (OR, 2.64, 95% CI 1.32–5.30; p = 0.006), total gastrectomy (OR, 2.54, 95% CI 1.24–5.19; p = 0.011) and predicted VC (OR, 2.41, 95% CI 1.01–5.75; p = 0.048;Table 4). Post-operative macroscopic atelectasis and clinical factors
On univariate analysis, post-operative macroscopic atelectasis was associated significantly with higher body mass index ( ‡ 25; p = 0.023), a lower pre-operative serum albumin concentration (p = 0.026), open surgery (p = 0.006), total gastrectomy (p = 0.027), combined resection of an other organ (p = 0.037), higher operative blood loss (p = 0.016), and intraoperative or post-operative blood transfusion (p =0.001; Table 2). On multivariable analysis, total gastrectomy was the only independent risk factor for post-operative macroscopic atelectasis (OR, 3.14, 95% CI 1.54–6.40, p = 0.002; Table 4). Discussion
Our results showed that extended operating time ( ‡ 6 h), total gastrectomy, and lower predicted VC ( < 80%) were independent risk factors for PPCs, after gastrectomy particularly pneumonia. The results of other studies that included non-thoracic surgery are similar to ours and showed that extended anesthesia or operative time was an independent risk factor [2,4,5]. In open gastrectomy, two studies showed that an extended operative time (5–6 h or longer) is an independent risk factor for various post-operative complications [7,18]. Risk factors for PPCs should be evaluated independently for each type of surgery but few large studies of gastrectomy have been performed to date. Total gastrectomy was not significantly related to post-operative pneumonia in another study [9]. In total gastrectomy, incisional pain might decrease pulmonary volume, or dissection around the esophageal hiatus and mediastinum might reduce diaphragm mobility, potentially leading to PPCs. Pneumonia after gastrectomy has been attributed not only to the direct effect of operative stress or immunosuppression but also to the aspiration of esophageal reflux contents [19]. We suspected that most cases of post-operative pneumonia were caused by aspiration in our study, although the differential diagnosis of aspiration pneumonia is often challenging. On the other hand, combined resection including concurrent splenectomy did not increase appreciably the risk ratio for PPCs in patients who underwent total gastrectomy in the present study. This finding is consistent with the results of other studies showing that adjacent organ resection, such as splenectomy, does not increase the risk of pneumonia [9] and that splenectomy is not an independent risk factor for any post-operative complication [7]. Sano et al. showed that extended lymphadenectomy did not increase post-operative
317
Gender Male 778 Female 275 Age ‡ 80 84 < 80 969 Body mass index (kg/m2) ‡ 25 226 < 25 825 Current smoking Yes 321 No 703 Unknown 29 Predicted VC (%) < 80 76 ‡ 80 962 Not evaluated 15 FEV1/ FVC (%) < 70 260 ‡ 70 778 Not evaluated 15 Anemia Yes 486 No 567 Preoperative serum albumin (g/dL) < 3.5 127 S3.5 921 Not evaluated 5 Diabetes mellitus Yes 146 No 907 Cardiac disease Yes 121 No 932 Cerebral disease Yes 85 No 968
n
2.20 (1.10–4.43) 1 0.86 (0.36–2.06) 1 0.87 (0.34–2.24) 1 1.98 (0.86–4.55) 1.00
11 ( 8.7) 38 ( 4.1) 0 6 ( 4.1) 43 ( 4.7) 5 ( 4.1) 44 ( 4.7) 7 ( 8.2) 42 ( 4.3)
2.71 (1.22–6.02) 1
8 (11.00) 40 ( 4.2) 1
1.46 (0.82–2.59) 1.00
1.75 (0.98–3.15) 1
21 ( 6.5) 27 ( 3.8) 1
27 ( 5.6) 22 ( 3.9)
1.34 (0.70–2.57) 1
13 ( 5.8) 36 ( 4.4)
1.38 (0.74–2.59) 1
1.66 (0.68–4.01) 1
6 ( 7.1) 43 ( 4.4)
15 ( 5.8) 33 ( 4.2) 1
2.62 (1.10–6.23) 1
OR (95%CI)
43 ( 5.5) 6 ( 2.2)
n (%)
All PPCs
0.10
0.77
0.74
0.023
0.20
0.31
0.020
0.058
0.38
0.27
0.024
p
4 (4.7) 19 (2.0)
4 (3.3) 19 (2.0)
5 (3.4) 18 (2.0)
4 (3.1) 19 (2.1) 0
15 ( 3.1) 8 (1.4)
5 (1.9) 17 (2.2) 1
6 (7.7) 16 (1.7) 1
10 (3.1) 13 (1.8) 0
3 (1.3) 20 (2.4)
5 (6.0) 18 (1.9)
23 (3.0) 0 (0)
n (%)
2.47 (0.82– 7.42) 1
1.64 (0.55– 4.91) 1
1.75 (0.64– 4.79) 1
1.54 (0.52– 4.61) 1
2.23 (0.-4– 5.30) 1
0.88 (0.32– 2.40) 1
5.07 (1.92–13.4) 1
1.71 (0.74– 3.93) 1
0.54 (0.16– 1.84) 1
3.34 (1.21– 9.25) 1
1.03 (1.02– 1.04) 1
OR (95%CI)
Pneumonia
0.11
0.21
0.24
0.51
0.064
1 (1.2) 19 (2.0)
1 (0.8) 19 (2.0)
0 (0.0) 20 (2.2)
6 (4.7) 14 (1.5) 0
10 (2.1) 10 (1.8)
7 (2.7) 13 0
2 (2.6) 18 (1.9) 0
< 0.001
0.80
8 (2.5) 11 (1.6) 0
9 (4.0) 11 (1.3)
1 (1.2) 19 (2.0)
15 (1.9) 5 (1.8)
n (%)
1.00
0.72
0.10
0.026
0.73
0.30
0.65
0.31
0.023
1.00
0.91
p
(continued)
0.60 (0.08– 4.50) 1
0.40 (0.05– 3.02) 1
0.98 (0.97– 0.99) 1
3.21 (1.21– 8.52) 1
1.17 (0.48– 2.84) 1
1.63 (0.64– 4.13) 1 (1.7)
1.42 (0.32– 6.23) 1
1.61 (0.64– 4.04) 1
3.07 (1.26– 7.50) 1
0.60 (0.80– 4.56) 1
1.06 (0.38– 2.95) 1
OR (95%CI)
Macroscopic atelectasis
0.21
0.44
0.031
0.004
p
Table 2. Uni-Variate Analysis of All PPCs, Pneumonia, and Macroscopic Atelectasis
318
0.95 (0.94–0.97) 1 1.44 (0.63–3.29) 1 0.95 (0.94–0.97) 1 1.21 (0.28–5.21) 1 2.22 (1.23–4.03) 1 4.22 (2.36–7.55) 1 1.07 (0.58–1.96) 1 2.84 (1.49–5.42) 1 3.28 (1.71–6.30) 1 3.00 (1.64–5.46) 1 3.18 (1.43–7.10) 1 2.38 (1.33–4.24) 1
7 ( 6.3) 42 ( 4.5) 0 ( 0.0) 49 ( 4.7) 2 ( 5.6) 47 ( 4.6) 31 ( 6.6) 18 ( 3.1) 26 (12) 23 ( 2.8) 16 ( 4.8) 33 ( 4.6) 14 (10) 35 ( 3.8) 14 (11) 35 ( 3.8) 19 ( 9.5) 29 ( 3.4) 1 8 (12) 41 ( 4.2) 28 ( 7.2) 21 ( 3.2)
OR (95%CI)
0 ( 0.0) 49 ( 4.8)
n (%)
9 (4.5) 14 (1.6) 0
< 0.001
0.003
17 (4.4) 6 (0.9)
2 (3.0) 21 (2.1)
7 (5.7) 16 (1.7)
< 0.001
0.009
4 (2.9) 19 (2.1)
0.001
8 (2.4) 15 (2.1)
13 (5.5) 10 (1.2)
< 0.001 0.84
13 (2.8) 10 (1.7)
2 (5.6) 21 (2.1)
0 (0.0) 23 (2.2)
4 (3.6) 19 (2.0)
0 (0.0) 23 (2.3)
n (%)
0.007
0.79
1.00
0.38
0.15
p
5.01 (1.96–12.8) 1
1.44 (0.33– 6.27) 1
2.84 (1.21– 6.65) 1
3.45 (1.39– 8.56) 1
1.41 (0.47– 4.20) 1
1.17 (0.49– 2.79) 1
4.65 (2.01–10.7) 1
1.64 (0.71– 3.77) 1
2.79 (0.63–12.4) 1
0.98 (0.97– 0.99) 1
1.82 (0.61– 5.44) 1
0.98 (0.97– 0.99) 1
OR (95%CI)
Pneumonia
6 (9.1) 14 (1.4) 9 (2.3) 11 (1.7)
< 0.001
8 (4.0) 11 (1.3) 1
4 (3.3) 16 (1.7)
6 (4.3) 14 (1.5)
6 (1.8) 14 (1.9)
9 (3.8) 11 (1.3)
15 (4.8) 5 (0.9)
0 20 (2.0)
0 20 (1.9)
0 20 (2.1)
0 20 (2.0)
n (%)
0.65
0.026
0.012
0.53
0.72
< 0.001
0.24
0.18
1.00
0.29
1.00
p
1.41 (0.58– 3.42) 1
6.95 (2.58–18.7) 1
3.20 (1.27– 8.07) 1
1.92 (0.63– 5.84) 1
2.93 (1.11– 7.75) 1
0.94 (0.36– 2.46) 1
2.87 (1.18– 7.02) 1
3.83 (1.38–10.6) 1
0.98 (0.97– 0.99) 1
0.98 (0.97– 0.99) 1
0.98 (0.97– 0.99) 1
0.98 (0.97– 0.99) 1
OR (95%CI)
Macroscopic atelectasis
PPCs = post-operative pulmonary complication; OR = odds ratio; CI = confidence interval; VC = vital capacity; FEV1 = forced expiratory volume in 1 sec; FVC = forced vital capacity.
Liver disease Yes 39 No 1014 Elevated pre-operative creatinine Yes 111 No 942 Regular use of steroids Yes 9 No 1044 Gait disturbance Yes 36 No 1017 Operative approach Open 469 Laparoscopic 584 Type of resection Total 238 Partial 815 Extent of lymph node dissection D2 330 < D2 723 Combined resection Yes 138 None or only gallbladder 915 Surgery duration (h) ‡6 123
