Korean Journal of Clinical Oncology 2022;18:36-46
https://doi.org/10.14216/kjco.22005
pISSN 1738-8082 ∙ eISSN 2288-4084
Original
Article
Local complications are related to poor long-term
outcome in patients undergoing curative
gastrectomy for advanced gastric cancer
Tae Sun Ha1, Gyu Seok Cho1, Eung Jin Shin1, Seung Wan Ryu2, Keun Won Ryu3, Min Chan Kim4, Woo Jin Hyung5, Chan Young Kim6,
Hyuk-Joon Lee7, Dong Woo Shin8, Jun Ho Lee9
1
Department of Surgery, Soonchunhyang University Bucheon Hospital, Bucheon;
Department of Surgery, Keimyung University School of Medicine, Daegu;
3
Gastric Cancer Branch, National Cancer Center, Goyang;
4
Department of Surgery, Dong-A University College of Medicine, Busan;
5
Department of Surgery, Yonsei University College of Medicine, Seoul;
6
Department of Surgery, Jeonbuk National University Medical School, Jeonju;
7
Department of Surgery, Seoul National University College of Medicine, Seoul;
8
Department of Surgery, Hallym University Dongtan Sacred Heart Hospital, Hwaseong;
9
Department of Surgery, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea
2
Purpose: The present study was performed to investigate the effects of local complications (LC) on long-term survival and cancer recurrence
in patients undergoing curative gastrectomy for gastric cancer.
Methods: We analyzed 2,627 patients after curative gastrectomy for gastric cancer between January 2001 and December 2006. Patients
were classified into groups no complications (NC), LC, or systemic complications (SC).
Results: Among the 2,627 patients, 475 patients developed complications (LC group [n=374, 14.2%] and SC group [n=101, 3.9%]). The
5-year cancer-specific survival rate was significantly poorer in the LC group compared to the NC and SC groups (LC, 78.0%; NC, 85.4%; SC,
80.2%; P=0.007). The occurrence of LC was identified as a significant independent prognostic factor for overall and cancer-specific survival
(hazard ratio [HR], 2.08; 95% confidence interval [CI], 1.46–2.97; P=0.001 and HR, 1.77; 95% CI, 1.12–2.81; P=0.015). The tumor recurrence
rates were higher in the LC group than the in other two groups (LC, 23.5%; NC, 15.4%; SC, 15.8%; P<0.001). The occurrence of LC was an
independent predictor of tumor recurrence in patients undergoing curative gastrectomy for gastric cancer (HR, 1.55; 95% CI, 1.11–2.17;
P=0.011).
Conclusion: LC are associated with adverse long-term outcomes in patients after curative gastrectomy for advanced gastric cancer.
Keywords: Stomach neoplasms, Postoperative complication, Tumor recurrence
Received: Jan 19, 2022 Revised: Jun 13, 2022 Accepted: Jun 15, 2022
Correspondence to: Eung Jin Shin
Department of Surgery, Soonchunhyang University Bucheon Hospital,
170 Jomaru-ro, Wonmi-gu, Bucheon 14584, Korea
Tel: +82-32-621-5834, Fax: +82-32-621-5016
E-mail: applips@schmc.ac.kr
ORCID: Tae Sun Ha (https://orcid.org/0000-0003-3683-6929), Gyu Seok Cho
(https://orcid.org/0000-0003-0866-0946), Eung Jin Shin (https://orcid.org/00000002-2029-4136), Seung Wan Ryu (https://orcid.org/0000-0001-6429-5926), Keun
Won Ryu (https://orcid.org/0000-0002-5935-9777), Min Chan Kim (https://orcid.
org/0000-0003-1905-3316), Woo Jin Hyung (https://orcid.org/0000-0002-85939214), Chan Young Kim (https://orcid.org/0000-0001-6656-9425), Hyuk-Joon Lee
(https://orcid.org/0000-0002-9530-647X), Dong Woo Shin (https://orcid.org/00000003-0373-5590), Jun Ho Lee (https://orcid.org/0000-0002-1641-4772)
Copyright © 2022 Korean Society of Surgical Oncology
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which
permits unrestricted non-commercial use, distribution, and reproduction in any medium,
provided the original work is properly cited.
36
Korean Journal of Clinical Oncology
INTRODUCTION
Despite its decreasing incidence rate, gastric cancer remains a common malignancy worldwide. In particular, gastric cancer is the second most common cause of cancer death in Korea, although the
number of patients diagnosed with early gastric cancer has increased due to the development of gastroscopy and mass screening
tests [1]. Age, tumor size, tumor location, histological classification,
depth of invasion, lymph node invasion, lymphatic and vascular invasion, and distant metastasis have been identified as prognostic
factors that influence the treatment outcome of gastric cancer [1-3].
Of these, D2 lymphadenectomy is the sole treatment option depending on the skill of the attending physician [4].
Anastomotic leakage, which is one of the most serious compli-
Tae Sun Ha et al. • Local complications exacerbate long-term outcome
cations after gastrointestinal surgery, is related to poor outcome
despite improvements in surgical techniques and perioperative
management of cancer surgery [5-7]. In esophageal and colon cancer surgery, several studies have reported that negative effects of
complications, including anastomotic leakage, are associated with
postoperative mortality, long-term outcome, and recurrence [5-7].
Some recent studies have investigated the relationship between
postoperative complications and long-term oncological outcome
in cases of stomach cancer surgery [8,9]. However, the impact of
complications other than anastomotic leakage and intraabdominal
infection were not assessed, and studies have been limited to cases
in which only total gastrectomy was performed or in cases of advanced but not early gastric cancer [8,10].
This study was performed to investigate the impact of postoperative local and systemic complications (SC) on long-term survival
and recurrence rates in patients following curative gastrectomy for
gastric cancer.
METHODS
This was a retrospective multicenter observational cohort study of
patients who underwent gastrectomy for gastric cancer between
January 2001 and December 2006 at nine surgical centers: Soonchunhyang University Bucheon Hospital, Keimyung University
Hospital, National Cancer Center, Dong-A University Hospital,
Yonsei University Hospital, Chonbuk National University Hospital,
Seoul National University Hospital, Hallym University Dongtan
Sacred Heart Hospital, Samsung Medical Center, and Sungkyunkwan University. Consecutive patients who underwent gastrectomy
for gastric cancer were prospectively registered over the study period. This study was granted Institutional Review Board approval
(IRB No. SCHBC 2014-07-011) and informed consent was waived
because of the retrospective observational design and the anonymous nature of the study.
Study population
The inclusion criteria required patients to be histologically diagnosed with gastric cancer and to be receiving surgical treatment
during the study period. Patients diagnosed with synchronous primary malignancy, those undergoing non-curative surgery, and
those receiving preoperative chemotherapy were excluded.
Definitions
Postoperative complications were defined by a combination of
clinical findings based on symptoms and physical examination of
patients, as well as the results of laboratory and radiological studies,
including X-ray, ultrasonography, and computed tomography
(CT). In this study, patients were divided into groups exhibiting no
complications (NC), local complications (LC), or SC. LC included
wound infection, fluid collection/abscess, intraabdominal bleeding, intraluminal bleeding, intestinal obstruction, ileus, stenosis,
anastomotic leakage, pancreatitis, wound dehiscence, and cholecystitis. SC consisted of pulmonary, hepatic, cardiac, urinary tract
infections, splenic infarction, transient ischemic attack, and delirium. The Clavien-Dindo classification was used to assess the severity of surgical complications, using the following definitions based
on its contracted form: grade I: any complication not requiring
medical or surgical treatment; grade II: a complication requiring
pharmacological treatment but no active intervention; grade III: a
complication requiring surgical, radiological, or endoscopic treatment; grade IV: complications that are potentially life-threatening
and require intensive care. In this study, we proceeded to record
the complications observed in our patients, distribute them according to the type of treatment required, and classify them on this
scale during hospitalization after surgery. In case of wound repair,
when the wound was approximated at bedside, it was classified as
grade I, when repaired under local anesthesia in an operation theater as grade III. Severe complications were defined as those graded as > III [11]. Surgical methods were classified as subtotal gastrectomy (distal or proximal gastrectomy) and total gastrectomy
(total or completion gastrectomy), carried out by standard radical
gastrectomy with D1+ or more. No patients underwent laparoscopic gastrectomy in the present study. According to the Japanese
Classification of Gastric Cancer second English edition [12], regional lymph node dissection was defined as follows: D1, dissection of perigastric lymph node (i.e., stations 1–6); D2, D1 dissection and additional removal of the nodes along the left gastric,
common hepatic, splenic, and left hepatoduodenal arteries (i.e.,
stations 7–11). If gross residual tumor and microscopic tumor
were absent in the primary tumor bed, the surgery was considered
curative resection. T stage was defined according to the International Union Against Cancer TNM classification of malignant tumors seventh edition [13] as local tumor growth (T1, invasion of
lamina propria or submucosa; T2, invasion of muscularis propria
or subserosa; T3, penetration serosa without invasion of adjacent
structures; and T4, invasion of adjacent structures). Definitions for
N stages were as follows: N0, no regional lymph node metastasis;
N1, metastasis in 1–6 regional lymph nodes: N2, metastasis in
7–15 regional lymph nodes; and N4, metastasis in more than 15
regional lymph nodes. The tumor biopsy specimens were histologically classified as differentiated (papillary adenocarcinoma, tubular adenocarcinoma of well or moderately differentiated type) or
undifferentiated (poorly differentiated adenocarcinoma, signet-ring cell carcinoma, mucinous adenocarcinoma) carcinoma.
www.kjco.org
37
Vascular invasion was defined as groups of tumor cells attached to
the blood vessel wall and projecting into the lumen of blood vessels
in the capsule or outside the tumor. Lymphatic invasion was defined as the presence of tumor emboli or tumor masses within
lymphatic vessels. Body mass index (BMI) was defined as the patient’s weight in kilograms divided by height (in meters) squared.
Patients were classified into two groups: normal weight (BMI < 25
kg/m2) or overweight and obese (BMI > 25 kg/m2) [14]. Recurrence patterns were classified as locoregional, peritoneal, and distant recurrence. Locoregional recurrence was defined as any cancer recurrence at the resection margin, operation bed within the
region of the resection, or regional lymph node metastasis. Peritoneal recurrence was defined as peritoneal carcinomatosis or
Krukenberg tumor. Distant recurrence was defined as metastasis
to non-regional lymph node or systemic organs, such as the liver,
lung, bone, brain, and ovaries.
ranges (IQRs) or as numbers (percentages) of patients because the
majority of the data did not follow a normal distribution. The
Mann-Whitney U test and Kruskal-Wallis one-way analysis of
variance were used for comparison of continuous variables, and
comparisons of categorical data were performed using the chisquared or Fisher exact test where appropriate. Multivariate Cox
regression analysis was used to determine independent predictors
of the 5-year overall survival rate and the 5-year cancer recurrence
rate in a forward stepwise manner. Covariates included univariate
predictors with P < 0.20 and those that we considered clinically relevant. Survival curves were constructed using Kaplan-Meier estimates and compared with the log-rank test. All tests were twotailed and P < 0.05 was considered statistically significant. Statistical tests were performed using SPSS software, version 18.0 for
Windows (SPSS Inc., Chicago, IL, USA).
RESULTS
Adjuvant chemotherapy and patient follow-up
The follow-up schedule and indications for adjuvant chemotherapy were not standardized among the various institutions. However,
most patients with advanced gastric cancer at TNM stage II or
higher received fluorouracil-based adjuvant treatment. No patients
were given postoperative or perioperative radiotherapy. Follow-up
observations were performed at 3- or 6-month intervals for the
first 2 years, at 6-month intervals for the next 3 years, and then annually until the patient’s death. The follow-up program consisted
of physical examination, laboratory blood tests, endoscopy, CT,
and positron emission tomography-CT (PET-CT). Recurrence
was diagnosed from imaging findings, including CT, magnetic resonance imaging, PET-CT, endoscopy, biopsies, or cytology.
During the study period, a total of 3,284 patients who underwent
gastrectomy for gastric cancer were entered into the registry. Of
these, 417 patients were excluded according to the exclusion criteria: palliative gastrectomy (n = 287), insufficient medical records
(n = 126), and reasons unrelated to cancer death (suicide and death
due to traffic accidents, n = 4). Ultimately, a total of 2,867 patients
Patients receiving gastrectomy
January 2001 to February 2006
(n = 3,284)
Exclusion:
Palliative gastrectomy (n = 287)
Insufficient medical records (n = 126)
Unrelated to cancer death (n = 4)
Data collection
The following data were obtained from electronic medical records
and the prospectively collected gastric cancer registry from each
hospital: general characteristics of the patients, including demographic data, tumor site and size, type of surgical procedure, and
extent of lymph node dissection. Overall survival was calculated
from the time of surgery to death due to any cause, or to the last
follow-up that ended without death. Cancer-specific survival was
calculated as the date of a diagnosis of stomach cancer until date of
death, and was censored if the primary or underlying cause of
death was not stomach cancer. The primary outcome was the
5-year survival rate and the secondary outcome was the 5-year
cancer recurrence rate.
Gastric cancer treated with
curative intent
(n = 2,867)
Died of other diseases (excluded
when analyzing cancer-specific
survival) (n = 240)
No complications
(n = 2,152)
Statistical analysis
All results and tables are presented as medians and interquartile
38
Korean Journal of Clinical Oncology
Fig. 1. Patient flowchart.
Local
complications
(n = 374)
Systemic
complications
(n = 101)
Tae Sun Ha et al. • Local complications exacerbate long-term outcome
Table 1. Clinicopathological characteristics of 2,867 patients undergoing curative gastrectomy for gastric cancer
Characteristics
Male sex
Age ≥ 65 yr
Body mass index ≥ 25 kg/m2
Co-morbidity
None
Hypertension
Diabetes mellitus
Pulmonary disease
Heart disease
Liver disease
Cerebrovascular disease
Renal disease
Others
Surgical procedure
Subtotal
Total
Lymph node dissection ≥ D2
Combine resection
Tumor size ≥ 5 cm
Tumor site
Upper
Middle
Lower
Whole
Histology, undifferentiated
Tumor depth
T1
T2
T3
T4
Nodal involvement
N0
N1
N2
N3
TNM stage
I
II
III
IV
Adjuvant chemotherapy
Cancer recurrence
No complications
(n= 2,324)
Local complications
(n= 420)
Systemic complications
(n= 123)
1,569 (67.5)
792 (34.1)
572 (24.8)
307 (73.1)
145 (34.5)
137 (32.9)
84 (68.3)
56 (45.5)
30 (25.0)
1,438 (61.9)
446 (19.2)
123 (5.3)
86 (3.7)
31 (1.3)
86 (3.7)
17 (0.7)
5 (0.2)
92 (4.0)
232 (55.2)
90 (21.4)
29 (6.9)
12 (2.9)
11 (2.6)
20 (4.8)
3 (0.7)
2 (0.5)
21 (5.0)
70 (56.9)
29 (23.6)
9 (7.3)
3 (2.4)
1 (0.8)
3 (2.4)
2 (1.6)
0
6 (4.9)
1,694 (72.9)
630 (27.1)
2,010 (86.8)
314 (13.5)
960 (41.3)
268 (63.8)
152 (36.2)
374 (89.3)
74 (17.6)
212 (50.6)
81 (65.9)
42 (34.1)
109 (90.1)
22 (17.9)
54 (43.9)
318 (13.7)
676 (29.1)
1,297 (55.8)
33 (1.4)
1,336 (57.5)
87 (20.7)
111 (26.4)
204 (48.6)
18 (4.3)
245 (58.3)
23 (18.7)
42 (34.1)
56 (45.5)
2 (1.6)
70 (56.9)
1,041 (44.8)
320 (13.8)
482 (20.7)
478 (20.6)
144 (34.3)
45 (10.7)
113 (26.9)
117 (27.9)
52 (42.3)
19 (15.4)
26 (21.1)
26 (21.1)
1,251 (53.9)
296 (12.7)
324 (13.9)
452 (19.5)
193 (46.0)
56 (13.3)
62 (14.8)
109 (26.0)
64 (52.0)
17 (13.8)
18 (14.6)
24 (19.5)
1,131 (48.7)
468 (20.1)
689 (29.6)
33 (1.4)
1,002 (51.1)
375 (16.3)
159 (37.9)
83 (19.8)
168 (40.0)
9 (2.1)
200 (55.4)
109 (26.1)
61 (49.6)
21 (17.1)
39 (31.7)
2 (1.6)
40 (40.8)
24 (19.5)
P-value
0.077
0.034
0.002
0.031
0.001
0.193
0.044
0.002
0.023
0.948
< 0.001
0.004
< 0.001
0.023
< 0.001
Values are presented as number (%).
who underwent curative gastrectomy for gastric cancer were included in this study. Of the eligible patients, 240 (8%) who died of
other diseases were excluded when analyzing cancer-specific sur-
vival. A total of 2,627 patients were categorized according to postoperative complication type into the NC group (n = 2,152, 81.9%),
the LC group (n = 374, 14.2%), and the SC group (n = 101, 3.8%)
www.kjco.org
39
No. of patients
Grade ≥ III
421
141 (4.9)
70 (2.4)
20 (0.7)
14 (0.5)
25 (0.9)
43 (1.5)
39 (1.4)
30 (1.0)
14 (0.5)
25 (0.9)
122
60 (2.1)
15 (0.5)
5 (0.2)
3 (0.1)
5 (0.2)
4 (0.1)
2 (0.1)
9 (0.3)
19 (0.7)
174
39 (1.4)
31 (1.1)
9 (0.3)
4 (0.1)
16 (0.6)
5 (0.2)
34 (1.2)
28 (1.0)
2 (0.1)
6 (0.1)
32
11 (0.4)
3 (0.1)
1 (0.0)
1 (0.0)
3 (0.1)
2 (0.1)
0
0
11 (0.4)
Local complications
Wound problems
Fluid collection/abscess
Intraabdominal bleeding
Intraluminal bleeding
Intestinal obstruction
Ileus
Stenosis
Anastomotic leakage
Pancreatitis
Others
Systemic complications
Pulmonary
Hepatic
Cardiac
Renal
Splenic infarction
Colitis
Transient ischemic attack
Delirium
Others
Values are presented as number (%).
(Fig. 1).
The clinicopathological characteristics of the 2,867 patients at
the time of curative gastrectomy for gastric cancer are presented in
Table 1. Overall, the patient population included 1,960 males
(68.4%) and 907 females (31.6%), with a median age of 60 years
(IQR, 50–67 years). The ratio of patients over age 65 was higher in
the SC group. There were no significant differences between the
three groups with respect to histological differentiation or lymph
node dissection ≥ D2. Total gastrectomy and combined resection
were significantly more frequent in both the LC and the SC groups
than in the NC group. BMI ≥ 25 kg/m2, tumor size ≥ 5 cm, advanced TNM stage, and recurrence of cancer were markedly higher in the LC group than in the other two groups.
Of the 2,867 patients who underwent gastrectomy for gastric
cancer, 543 patients (18.9%) developed complications. The numbers of Clavien-Dindo grade I, II, III, and IV complications were
155 (5.4%), 182 (6.3%), 132 (4.6%), and 74 (2.6%), respectively. Table 2 shows incidence of local and SC and grade ≥ III according to
the Clavien-Dindo classifications after curative gastrectomy.
Among LC of grade III or over, wound problems, fluid collection/
abscess, stenosis, and anastomotic leakage were relatively common.
40
Korean Journal of Clinical Oncology
Cumulative survival (%)
Complications
100
80
NC
60
P < 0.001
LC
SC
40
0
20
40
60
A
Time (mo)
100
Cumulative survival (%)
Table 2. Incidence of local and systemic complications and grade
≥III according to the Clavien-Dindo classification of 2,867 patients
undergoing curative gastrectomy for gastric cancer
80
NC
60
P = 0.007
LC
SC
40
0
20
40
Time (mo)
60
B
Fig. 2. Comparison of 5-year overall survival and cancer-specific
survival rates among groups with no complications (NC), local complications (LC), and systemic complications (SC). (A) Overall survival.
(B) Cancer-specific survival.
Most cases of anastomotic leakage required intervention (22/30
cases). Reoperation under general anesthesia was performed on 34
patients (19.5%) at grade III of LC group: intestinal obstruction
(n = 12), wound problems (n=8), anastomotic leakage (n=6), Fluid
collection/abscess (n =3), intraabdominal bleeding (n = 3), stasis/
stenosis (n = 2). As for SC, pulmonary and hepatic complications
were common (2.1% and 0.5%, respectively).
The median follow-up period of the surviving patients was 59.0
months (IQR, 48.0–73.0 months), and ranged from 1 to 106
months. The 5-year overall survival rate of the SC group was lower
than those of the LC group and NC group (NC, 79.4%, LC, 70.7%;
P < 0.001, SC, 66.7%; P < 0.001). The 5-year cancer-specific survival rates of NC, LC, and SC groups were 85.4%, 78.3% (P < 0.001),
and 80.2% (P = 0.159), respectively, indicating that the occurrence
of LC is related to poor survival (P = 0.007) (Fig. 2). Fig. 3 shows a
comparison of the 5-year cancer-specific survival rates among NC,
LC, and SC groups according to TNM stage. The 5-year cancer-specific survival rates were 97.5% in the NC group, 96.7%
(P = 0.489) in the LC group, and 92.9% (P = 0.062) in the SC group
for patients with stage I disease; the differences among the three
Tae Sun Ha et al. • Local complications exacerbate long-term outcome
100
Cumulative survival (%)
Cumulative survival (%)
100
80
LC (P = 0.489)
SC (P = 0.062)
NC
60
LC
80
SC
SC
40
40
0
20
40
60
0
20
A
Time (mo)
40
60
B
Time (mo)
100
100
LC (P = 0.021)
SC (P = 0.244)
Cumulative survival (%)
Cumulative survival (%)
LC (P = 0.011)
SC (P = 0.472)
NC
LC
60
80
NC
60
LC
SC
80
60
40
NC
LC
20
SC
0
40
0
20
40
60
0
20
C
Time (mo)
40
60
D
Time (mo)
100
100
80
80
Cumulative survival (%)
Cumulative survival (%)
Fig. 3. Comparison of the 5-year cancer-specific survival rates among groups with no complications (NC), local complications (LC) and systemic complications (SC) according to stage. (A) Stage I. (B) Stage II. (C) Stage III. (D) Stage IV.
60
P < 0.001
40
20
No complications
Clavin-Dindo II
Clavin-Dindo III
Clavin-Dindo IV
0
60
P < 0.001
40
20
No complications
Clavin-Dindo II
Clavin-Dindo III
Clavin-Dindo IV
0
20
40
Time (mo)
60
20
A
40
60
B
Time (mo)
Fig. 4. Comparison of survival outcomes based on the Clavien-Dindo classification. (A) Overall survival. (B) Cancer-specific survival.
groups were not statistically significant (Fig. 3A). In stage II disease,
the 5-year cancer-specific survival rates were 89.4%, 83.3%, and
84.6% in the NC, LC, and SC groups, respectively (P = 0.011 and
P = 0.472), and those in stage III disease were 62.7%, 51.9%, and
59.0%, respectively (P = 0.021 and P = 0.244) (Fig. 3B and C). Thus,
LC were related to significantly poor survival in patients receiving
curative gastrectomy for advanced gastric cancer. In stage IV disease, the 5-year cancer-specific survival rates were 28.6% in the NC
group and 0% in the LC group, and were not significantly different
due to the small sample size (Fig. 3D).
In the analysis of patients in different severity of complications,
increasing Clavien-Dindo scores from II to IV compared to no
complication was significantly associated with a corresponding
decrease in overall survival (80.4%, 71.4%, 66.5%, 29.5%, P < 0.001)
and cancer-specific survival (86.5%, 78.4%, 74.3%, 39.6% P <
0.001) (Fig. 4). The subgroup analysis showed that increasing Clawww.kjco.org
41
Table 3. Multivariate analyses of 5-year overall survival and cancer-specific survival rate after curative gastrectomy for gastric cancer
Overall survival
Variable
Age (compared to < 65 yr)
Tumor size (compared to < 5 cm)
Surgical procedure, total (compared to subtotal)
Lymph node dissection (compared to < D2)
Cancer stage (compared to I)
II
III
IV
Vascular invasion
Lymphatic invasion
Complications (compared to no complications)
Local complications
Systemic complications
Adjuvant chemotherapy
Cancer-specific survival
HR (95% CI)
P-value
HR (95% CI)
P-value
1.57 (1.32–1.87)
1.26 (1.02–1.57)
1.59 (1.27–1.84)
0.67 (0.51–0.88)
< 0.001
0.034
< 0.001
0.005
< 0.001
< 0.001
< 0.001
< 0.001
0.001
< 0.001
< 0.001
0.001
0.061
0.024
1.48 (1.21–1.82)
1.30 (1.01–1.66)
1.35 (1.10–1.166
0.71 (0.51–0.99)
< 0.001
0.041
0.005
0.048
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
0.001
0.016
0.015
0.076
0.021
1.95 (1.36–2.79)
6.12 (4.43–8.47)
14.23 (8.80–22.99)
1.57 (1.22–2.03)
1.63 (1.24–2.12)
2.08 (1.46–2.97)
1.21 (0.97–1.52)
1.29 (1.03–1.61)
2.77 (1.75–4.39)
9.40 (6.17–14.32)
25.67 (14.39–45.80)
1.68 (1.27–2.23)
1.69 (1.24–2.32)
1.77 (1.12–2.81)
1.26 (0.98–1.63)
1.35 (1.05–1.73)
HR, hazard ratio; CI, confidence interval.
Table 4. Initial recurrence site according to complications in 2,627 patients undergoing curative gastrectomy for gastric cancer
Variable
No complications (n= 2,152)
No recurrence
Locoregional recurrence
Peritoneal recurrence
Distant recurrence
Local complications (n= 374) Systemic complications (n= 101)
1,820 (84.6)
91 (4.2)
132 (6.1)
109 (5.0)
286 (76.5)
16 (4.3)
33 (8.8)
39 (10.4)
85 (84.2)
1 (1.0)
10 (9.9)
5 (5.0)
P-value
< 0.001
Values are presented as number (%).
40
Cancer recurrence (%)
NC
30
P = 0.017
LC
SC
20
10
0
20
40
60
Time (mo)
Fig. 5. Comparison of 5-year cancer recurrence rates among groups
with no complications (NC), local complications (LC) and systemic
complications (SC) after curative gastrectomy for gastric cancer.
vien-Dindo scores was associated with decreased overall survival
(96.5%, 93.5%, 86.2%, 49.1%, P < 0.001) and cancer-specific survival (98.5%, 96.2%, 90.8%, 74.7%, P < 0.001) for stage II patients; decreased overall survival (82.7%, 76.3%, 70.3%, 39.5%, P < 0.001)
and cancer-specific survival (89.4%, 82.9%, 76.0%, 52.1%, P < 0.001)
42
Korean Journal of Clinical Oncology
for stage III patients; decreased overall survival (54.9%, 47.0%,
41.9%, 17.0%) and cancer-specific survival (62.6%, 55.9%, 49.2%,
P < 0.001) for stage IV patients.
In the univariate analysis of the 5-year cancer-specific survival in
2,627 patients after curative gastrectomy for gastric cancer, age, total gastrectomy, combine resection, lymph node dissection ≥ D2,
tumor size, tumor in the whole stomach, histological differentiation, T category, N category, stage, lymphatic invasion, vascular invasion, adjuvant chemotherapy, and complication occurrence
were identified as prognostic factors. Multivariate analysis of
5-year overall survival and cancer-specific survival rate after curative gastrectomy for gastric cancer are presented in Table 3. When
multivariate analysis was performed for overall and cancer-specific
survival rates, the common independent poor prognostic factors
were patient age ≥ 65 years, tumor size ≥ 5 cm, total gastrectomy,
lymph node dissection ≥ D2, lymphatic invasion, vascular invasion, adjuvant chemotherapy, and advanced stage. In particular,
the occurrence of LC was identified as a poor prognostic factor in
the multivariate analysis of both overall and cancer-specific survival rates (overall survival: OR, 2.08; 95% CI, 1.46–2.97; P = 0.001;
cancer-specific survival: OR, 1.77; 95% CI, 1.12–2.81; P = 0.015),
Tae Sun Ha et al. • Local complications exacerbate long-term outcome
Table 5. Univariate and multivariate analyses of factors associated with cancer recurrence in 2,627 patients after curative gastrectomy for
gastric cancer
Variable
Sex (male)
Age (compared to < 65 yr)
Surgical procedure (compared to subtotal)
Combine resection
Tumor size (compared to < 5 cm)
Tumor site (compared to lower)
Upper
Middle
Whole
Histology, undifferentiated
Cancer stage (compared to I)
II
III
IV
Lymphatic invasion
Vascular invasion
Adjuvant chemotherapy
Complications (compared with no complications)
Local complications
Systemic complications
Univariate analysis
HR (95% CI)
1.01 (0.81–1.26)
1.19 (0.96–1.48)
2.48 (2.01–3.07)
1.65 (1.25–2.16)
4.19 (3.36–5.23)
0.99 (0.78–1.27)
1.35 (1.01–1.80)
6.65 (3.56–12.41)
2.04 (1.64–2.55)
5.08 (3.42–7.56)
18.98 (13.49–26.70)
53.19 (31.54–89.70)
7.41 (5.68–9.66)
4.98 (3.80–6.52)
4.92 (3.81–6.36)
1.68 (1.29–2.19)
1.03 (0.60–1.78)
P-value
0.92
0.11
< 0.001
< 0.001
< 0.001
< 0.001
0.96
0.51
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
0.001
< 0.001
0.92
Multivariate analysis
HR (95% CI)
P-value
1.37 (1.04–1.80)
0.024
3.27 (2.09–5.11)
10.13 (6.70–15.34)
27.64 (15.31–49.91)
1.86 (1.37–2.54)
1.53 (1.17–1.99)
1.44 (1.10–1.87)
1.55 (1.11–2.17)
1.69 (0.86–3.35)
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
0.002
0.007
0.017
0.011
0.126
HR, hazard ratio; CI, confidence interval.
but the presence of SC was not a significant factor (overall survival:
OR, 1.21; 95% CI, 0.97–1.52; P = 0.061; cancer-specific survival:
OR, 1.26; 95% CI, 0.98–1.63; P = 0.076).
Recurrence rates were higher in the LC group (n = 88, 23.5%)
than in the NC and SC groups (n = 332, 15.4% and n = 16, 15.8%;
P = 0.017) (Fig. 5). Details of the initial recurrence site according to
complication groups following curative gastrectomy are shown in
Table 4. In the LC group, distant recurrence was the most frequent
pattern, whereas peritoneal recurrence was the most frequently
observed site of recurrence in the SC group. There were significant
differences in patterns of recurrence between these three groups
(P < 0.001). Table 5 shows the results of univariate and multivariate
analyses of factors associated with cancer recurrence. Multivariate
analysis of the factors related to recurrence showed that age ≥ 65
years, advanced stage, lymphatic and vascular invasion, adjuvant
chemotherapy, and LC were independent prognostic factors of recurrence.
DISCUSSION
This is the largest multicenter study to examine the impact of postoperative complications on overall and cancer-specific survival
rates in patients undergoing curative resection for gastric cancer.
The results of the present study, along with previous reports in the
literature, showed that LC following curative gastrectomy for gastric cancer are associated with high morbidity and mortality compared to other groups. In addition, the increase in the ClavienDindo classification proportionally deteriorate survival outcomes.
Prognostic factors of the 5-year overall and cancer-specific survival
in patients undergoing curative gastrectomy included LC as well as
patient age, tumor size, total gastrectomy as the surgical procedure,
more advanced stage of gastric cancer, vascular and lymphatic invasion, and adjuvant chemotherapy. Among the three groups, only
the LC group was associated with higher recurrence rates, and the
occurrence of LC was identified as one of the major independent
prognostic factors of the 5-year cancer recurrence.
A number of recent studies have reported that the development
of postoperative complications, especially intraabdominal sepsis,
results in poorer long-term survival and increased risk of recurrence in patients undergoing resection of esophageal and colorectal cancer [6,15-18]. Sierzega et al. [8] showed for the first time in
stomach cancer that anastomotic leakage has a negative impact on
long-term survival after total gastrectomy for gastric cancer. However, their study included only total gastrectomy as the surgical
www.kjco.org
43
procedure and also included the cases of non-curative gastrectomy,
and did not investigate the impact of other postoperative complications, unlike our study. Tsujimoto et al. [9] reported a direct correlation between complications due to postoperative infection and
unfavorable outcome, and anastomotic leakage was only a prognostic factor of cancer-specific survival although they limited infectious complications to pneumonia, enterocolitis, cholecystitis,
anastomotic leakage, and intraabdominal abscess. In a recent
study, Tokunaga et al. [19] reported poor long-term overall and relapse-free survival rates, and suggested that intraabdominal complications after curative gastrectomy increase mortality and exacerbate recurrence rates, even if infectious complications are restricted to anastomotic leakage, pancreas-related complications,
and intraabdominal abscess. In the present study, however, we
evaluated the relationships between all complications and 5-year
overall and cancer-specific survival in patients after curative gastrectomy for gastric cancer of all stages. The results indicated that
LC, but not SC, were significant predictors of poor outcome with
regard to survival in patients who received curative gastrectomy
for advanced but not early gastric cancer.
The incidence of postoperative complications is an important
surrogate marker for measuring the quality of surgery. Dindo et al.
[11] defined a complication as any deviation from the normal
postoperative course and classified postoperative complications
based on the type of therapy needed to correct the complication.
The Clavien-Dindo classification system, which has proved to be a
reliable tool for quality assessment in surgery, is adopted to overcome a difficult evaluation of postoperative complications in gastrointestinal surgeries. The inaccurate and often confusing term,
major and minor complications is replaced by the well-standardized Clavien-Dindo classification. Using this standardized system,
we analyzed and classified the postoperative complications after
curative gastrectomy for gastric cancer. However, due to the limited number of complications in grade or higher, we did not perform specific analysis in each Clavien-Dindo grading. Li et al. [20]
showed no significant difference in the 5-year overall survival and
disease-specific survival between Clavien-Dindo II group and NC
group after total gastrectomy for advanced gastric cancer. However, Duraes et al. [21] showed that complications of minor severity,
such as the Clavien-Dindo classification, can adversely affect oncologic outcomes after colorectal cancer resection. In this study, we
also found that the increase of Clavien-Dindo grading, including
Clavien-Dindo II group, adversely affect the 5-year overall survival
and cancer-specific survival after curative gastrectomy for gastric
cancer. Because the relationship between the Clavien-Dindo classification and long-term survival outcomes was not linear and uniform, patients with Clavien-Dindo IV complications had worse
44
Korean Journal of Clinical Oncology
survival outcomes compared with patients with Clavien-Dindo III
complications.
As mentioned above, several previous studies pertaining to gastrointestinal malignancies have indicated that postoperative complications increase cancer-specific mortality rates [8-10,19]. However, the mechanism relating long-term survival of patients with
postoperative complications remains unclear. McMillan et al. [22]
showed that systemic inflammatory response syndrome is associated with poor survival in patients with curative treatment of colorectal cancer. That is, various mediators of exaggerated systemic
inflammatory response caused by postoperative complications
may induce release of proinflammatory cytokines [23,24]. The release of proinflammatory cytokines, including tumor necrosis factor-alpha, interleukin (IL)-1beta, IL-6, and IL-18, and oxygen free
radicals may alter host defense and promote the motility, invasiveness, and proliferation of residual or implanted tumor cells [25].
Furthermore, cell-mediated immunity, involving cells such as natural killer cells and cytotoxic T lymphocytes, is related to the extent
of surgical trauma and tissue damage [26,27]. Systemic infection
and sepsis due to postoperative complications may induce a period
of immunosuppression, which is enhanced after surgical trauma,
and contributes to the proliferation of occult or dormant cancer
cells, resulting in decreased survival [28,29]. Consequently, differences in the degree of immune suppression and systemic inflammatory response affect survival according to type of complications.
We classified complications as local or SC, and assessed the correlation between type of complications and the 5-year survival in
patients undergoing curative gastrectomy for gastric cancer. The
observations of our study indicating a low survival rate and high
cancer recurrence rate in the LC group due to induced immune
suppression and systemic inflammatory response provide additional support for the results discussed above.
Subset analyses indicated no significant differences in cancer-specific survival in stage I, but significant differences between
the LC group and the other groups were identified in advanced
stage II and III disease. In early gastric cancer, tumor cells are confined to the mucosal layer, and tumor cells are mostly removed by
curative radical gastrectomy. However, spreading or spillage of viable exfoliated tumor cells occurs frequently in advanced compared
to early gastric cancer [17,30]. When LC develop in advanced gastric cancer, therefore, the resulting effects on the degree of inflammatory immune response and immune suppression may produce
a greater number of adverse outcomes with respect to cancer-specific survival and recurrence rates in comparison with NC and SC
groups. Tsujimoto et al. [9] also demonstrated a negative impact of
infectious complications on gastric cancer recurrence.
Our study had several limitations. First, this was a non-random-
Tae Sun Ha et al. • Local complications exacerbate long-term outcome
ized observational study and had a retrospective design. As such
studies cannot be based on a randomized controlled experiment
due to ethical issues, it is essential to perform a multicenter observational study. However, all consecutive patients undergoing curative gastrectomy for gastric cancer were prospectively included in a
multicenter registry, and continuing data collection will improve
the generalizability of our findings. Second, with respect to type of
complications, we could not investigate the postoperative biological status, such as proinflammatory cytokines, including tumor
necrosis factor-alpha, IL-1beta, IL-6, and IL-18, and we could not
estimate the degree or duration of immune suppression in this
study. Third, grade I complications present any complication not
requiring medical or surgical treatment, such as nausea, vomiting,
and fever. In fact, since grade I complications were not always considered as complications in clinical practice, data of grade I complications would be unreliable and the overall incidence of grade I
complications could be underestimated. Thus, we did not evaluate
grade I complications after curative gastrectomy for gastric cancer.
Finally, LC delay or prohibit patients from receiving adjuvant chemotherapy to a greater extent than SC. This may account for worse
survival in the LC group, but our study could not investigate the
percent of patients in each group who delayed or failed to receive
adjunctive chemotherapy.
In conclusion, the postoperative occurrence of LC was an independent prognostic factor associated with poor overall and cancer-specific survival, and a high recurrence rate in patients undergoing curative gastrectomy for advanced gastric cancer, especially
stage II and III diseases. The negative effect on survival outcomes
were also increased with higher Clavien-Dindo grades. These
findings suggest that surgeons should implement surgical procedures with careful perioperative care to minimize the occurrence
of postoperative complications.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
CONFLICT OF INTEREST
No potential conflict of interest relevant to this article was reported.
14.
REFERENCES
1. Jeong O, Park YK. Clinicopathological features and surgical treatment of gastric cancer in South Korea: the results of 2009 nationwide survey on surgically treated gastric cancer patients. J Gastric
Cancer 2011;11:69-77.
2. Sakuramoto S, Sasako M, Yamaguchi T, Kinoshita T, Fujii M,
Nashimoto A, et al. Adjuvant chemotherapy for gastric cancer with
S-1, an oral fluoropyrimidine. N Engl J Med 2007;357:1810-20.
3. Shiraishi N, Sato K, Yasuda K, Inomata M, Kitano S. Multivariate
15.
16.
17.
prognostic study on large gastric cancer. J Surg Oncol 2007;96:148.
Kim JP. Current status of surgical treatment of gastric cancer. J Surg
Oncol 2002;79:79-80.
Walker KG, Bell SW, Rickard MJ, Mehanna D, Dent OF, Chapuis
PH, et al. Anastomotic leakage is predictive of diminished survival
after potentially curative resection for colorectal cancer. Ann Surg
2004;240:255-9.
McArdle CS, McMillan DC, Hole DJ. Impact of anastomotic leakage on long-term survival of patients undergoing curative resection
for colorectal cancer. Br J Surg 2005;92:1150-4.
Vignali A, Fazio VW, Lavery IC, Milsom JW, Church JM, Hull TL,
et al. Factors associated with the occurrence of leaks in stapled rectal anastomoses: a review of 1,014 patients. J Am Coll Surg 1997;
185:105-13.
Sierzega M, Kolodziejczyk P, Kulig J; Polish Gastric Cancer Study
Group. Impact of anastomotic leakage on long-term survival after
total gastrectomy for carcinoma of the stomach. Br J Surg 2010;
97:1035-42.
Tsujimoto H, Ichikura T, Ono S, Sugasawa H, Hiraki S, Sakamoto
N, et al. Impact of postoperative infection on long-term survival
after potentially curative resection for gastric cancer. Ann Surg Oncol 2009;16:311-8.
Yoo HM, Lee HH, Shim JH, Jeon HM, Park CH, Song KY. Negative impact of leakage on survival of patients undergoing curative
resection for advanced gastric cancer. J Surg Oncol 2011;104:73440.
Dindo D, Demartines N, Clavien PA. Classification of surgical
complications: a new proposal with evaluation in a cohort of 6336
patients and results of a survey. Ann Surg 2004;240:205-13.
Japanese Gastric Cancer Association. Japanese classification of gastric carcinoma: 2nd English edition. Gastric Cancer 1998;1:10-24.
Sobin LH, Gospodarowicz MK, Wittekind C; International Union
Against Cancer. TNM classification of malignant tumours. 7th ed.
Hoboken, NJ: Wiley-Blackwell; 2009.
Wong J, Rahman S, Saeed N, Lin HY, Almhanna K, Shridhar R, et
al. Effect of body mass index in patients undergoing resection for
gastric cancer: a single center US experience. J Gastrointest Surg
2014;18:505-11.
Law WL, Choi HK, Lee YM, Ho JW, Seto CL. Anastomotic leakage
is associated with poor long-term outcome in patients after curative colorectal resection for malignancy. J Gastrointest Surg 2007;
11:8-15.
Branagan G, Finnis D; Wessex Colorectal Cancer Audit Working
Group. Prognosis after anastomotic leakage in colorectal surgery.
Dis Colon Rectum 2005;48:1021-6.
Bell SW, Walker KG, Rickard MJ, Sinclair G, Dent OF, Chapuis PH,
www.kjco.org
45
18.
19.
20.
21.
22.
23.
24.
46
et al. Anastomotic leakage after curative anterior resection results
in a higher prevalence of local recurrence. Br J Surg 2003;90:12616.
Merkel S, Wang WY, Schmidt O, Dworak O, Wittekind C, Hohenberger W, et al. Locoregional recurrence in patients with anastomotic leakage after anterior resection for rectal carcinoma. Colorectal Dis 2001;3:154-60.
Tokunaga M, Tanizawa Y, Bando E, Kawamura T, Terashima M.
Poor survival rate in patients with postoperative intra-abdominal
infectious complications following curative gastrectomy for gastric
cancer. Ann Surg Oncol 2013;20:1575-83.
Li Z, Bai B, Ji G, Li J, Zhao Q. Relationship between Clavien-Dindo
classification and long-term survival outcomes after curative resection for gastric cancer: a propensity score-matched analysis. Int J
Surg 2018;60:67-73.
Duraes LC, Stocchi L, Steele SR, Kalady MF, Church JM, Gorgun E,
et al. The relationship between Clavien-Dindo morbidity classification and oncologic outcomes after colorectal cancer resection.
Ann Surg Oncol 2018;25:188-96.
McMillan DC, Canna K, McArdle CS. Systemic inflammatory response predicts survival following curative resection of colorectal
cancer. Br J Surg 2003;90:215-9.
Mantovani A, Allavena P, Sica A, Balkwill F. Cancer-related inflammation. Nature 2008;454:436-44.
Smyth MJ, Cretney E, Kershaw MH, Hayakawa Y. Cytokines in
cancer immunity and immunotherapy. Immunol Rev 2004;202:
Korean Journal of Clinical Oncology
275-93.
25. Majima T, Ichikura T, Chochi K, Kawabata T, Tsujimoto H, Sugasawa H, et al. Exploitation of interleukin-18 by gastric cancers for
their growth and evasion of host immunity. Int J Cancer 2006;118:
388-95.
26. Santarelli RL, Naud N, Tache S, Gueraud F, Vendeuvre JL, Zhou L,
et al. Calcium inhibits promotion by hot dog of 1,2-dimethylhydrazine-induced mucin-depleted foci in rat colon. Int J Cancer
2013;133:2533-41.
27. Goldfarb Y, Sorski L, Benish M, Levi B, Melamed R, Ben-Eliyahu S.
Improving postoperative immune status and resistance to cancer
metastasis: a combined perioperative approach of immunostimulation and prevention of excessive surgical stress responses. Ann
Surg 2011;253:798-810.
28. Aosasa S, Ono S, Mochizuki H, Tsujimoto H, Osada S, Takayama E,
et al. Activation of monocytes and endothelial cells depends on the
severity of surgical stress. World J Surg 2000;24:10-6.
29. Mokart D, Capo C, Blache JL, Delpero JR, Houvenaeghel G, Martin C, et al. Early postoperative compensatory anti-inflammatory
response syndrome is associated with septic complications after
major surgical trauma in patients with cancer. Br J Surg 2002;89:
1450-6.
30. Denis MG, Lipart C, Leborgne J, LeHur PA, Galmiche JP, Denis M,
et al. Detection of disseminated tumor cells in peripheral blood of
colorectal cancer patients. Int J Cancer 1997;74:540-4.