- Email: [email protected]
Urgent Surgery for Gastric Adenocarcinoma: A Study of the National Cancer Database
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 2 0 ( 2 4 5 ) 6 1 9 e6 2 8
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.JournalofSurgicalResearch.com
Association for Academic Surgery
Urgent Surgery for Gastric Adenocarcinoma: A Study of the National Cancer Database Benjamin W. Fisher, MD,* Marcus Fluck, BS, Katelyn Young, BS, Mohsen Shabahang, MD, PhD, Joseph Blansfield, MD, and Tania K. Arora, MD a
Department of Surgery, Geisinger Medical Center, Danville, Pennsylvania
article info
abstract
Article history:
Background: Gastric adenocarcinoma is a leading cause of cancer death worldwide and, in
Received 4 March 2019
the United States, can present emergently with upper GI hemorrhage, obstruction, or
Received in revised form
perforation. No large studies have examined how urgent surgery affects patient outcomes.
13 July 2019
This study examines the outcomes of urgent versus elective surgery for gastric cancer.
Accepted 19 July 2019
Materials and methods: Patients with gastric adenocarcinoma from the National Cancer
Available online xxx
Database from 2004 to 2015 were examined retrospectively. Patients with metastatic disease or incomplete data were excluded. Urgent surgery was defined as definitive surgery
Keywords:
within 3 d of diagnosis. Univariate and multivariate analysis of patient factors, surgical
Gastric adenocarcinoma
outcomes, and oncologic data was performed. P-values <0.05 were statistically significant.
Gastric cancer
Results: Of 26,116 total patients, 2964 had urgent surgery and 23,468 had elective surgery.
Urgent surgery
Urgent surgery patients were significantly older, were female, were nonwhite, had higher
Emergent surgery
pathologic stage, and were treated at a low-volume center. Urgent surgery was associated
National Cancer Database
with decreased quality lymph node harvest (odds ratio [OR] 0.68 95% confidence interval
NCDB
{CI} [0.62, 0.74]), increased positive surgical margin (OR 1.48, 95% CI [1.32, 1.65]), increased 30-d mortality (OR 1.38, 95% CI [1.16, 1.65]), increased 90-d mortality (OR 1.30, 95% CI [1.14, 1.49]), and decreased overall survival (hazard ratio 1.21, 95% CI [1.15, 1.27]). Conclusions: Urgent surgery for gastric cancer is associated with significantly worse outcomes than elective surgery. Stable patients requiring urgent surgical resection for gastric cancer may benefit from referral to a high-volume center for resection by an experienced surgeon. Patients undergoing urgent resection for gastric cancer should be referred to surgical and medical oncologists to ensure they receive appropriate adjuvant therapy and surveillance. ª 2019 Elsevier Inc. All rights reserved.
Introduction Gastric cancer is one of the leading causes of death worldwide, representing nearly one million new cases annually and
ranking as the third leading cause of cancer death (723,000 yearly).1-3 Although the incidence is much lower in the United States, it still affects a significant number of patients, with 27,510 new diagnoses and 11,140 deaths estimated for 2019.4
* Corresponding author. Geisinger Medical Center, General Surgery, 100 N Academy Avenue, Danville, PA, 17822. Tel.: þ1 570 271 6361; fax: þ1 570-271-8324. 0022-4804/$ e see front matter ª 2019 Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.jss.2019.07.073
620
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 2 0 ( 2 4 5 ) 6 1 9 e6 2 8
Prognosis varies based on stage at diagnosis; 5-y survival after treatment drops precipitously from 88% to 94% for stage IA-B to only 18%-54% for stage IIIA-C disease.5 In 2001, Macdonald et al. demonstrated that postoperative chemoradiation improved median survival by almost 50% when compared to surgical treatment alone.6 The MAGIC trial in 2006 demonstrated that perioperative chemotherapy reduced tumor size and stage at the time of surgery, as well as improving disease-free and overall survival.7 D2 lymph node dissection has now been accepted as the standard of care with reduced rates of local and regional recurrence.8-11 Improved treatment algorithms in the past decade have shown promise in decreasing mortality and extending survival, with gastric cancer mortality decreasing by 9% worldwide over the last 10 y3; however, there remains additional improvement to be made. There is still no consensus regarding the best treatment, with ongoing research into the optimal regimen and timing of therapy.10,12-16 Although research continues to make progress in elective treatment of gastric cancer, there is a relative paucity of data regarding outcomes for three of the well-known complication of gastric cancer: perforation, obstruction, and hemorrhage. It is unclear if urgent surgery for these complications changes adherence to standard treatment protocols and potentially compromises the quality of oncologic resection. This study compares patients undergoing elective surgical resection with patients undergoing urgent surgery to evaluate and characterize the differences in surgical and oncological outcomes.
Materials and methods The National Cancer Database (NCDB) is a joint project of the Commission on Cancer of the American College of Surgeons and the American Cancer Society. The database is a nationwide, facility-based, comprehensive clinical surveillance resource oncology data set; it captures information from approximately 1500 Commission on Cancer accredited hospitals and more than 70% of all newly diagnosed malignancies in the United States. This research was deemed exempt before initiation of study activities. A retrospective review of patients undergoing surgical resection for gastric adenocarcinoma from 2004 to 2015 was performed on the NCDB Stomach participant user file. The TNM pathological stage groups I-III were included. Adenocarcinomas were identified by the ICD-03 histology codes: 8490, 8245, 8210, 8263, 8261, 8572, 8255, 8574, 8570, 8144, 8140, 8560, 8021, 8145, 8010, 8020, 8310, 8243, 8576, 8012, 8575, 8244, 8323, 8480, 8481, 8000, 8260, 8050, 8141, 8142, 8230, 8143, 8211, 8051, 8262. Records missing days from diagnosis to definitive surgery were excluded as well as patients with evidence of metastatic disease. High facility volume cutoff is defined as a facility count of all stomach cancer resections 15 in a given year.17 Quality lymph node dissection was defined as 15 regional nodes examined. Patients with unknown survival or follow-up data were excluded. Urgent surgery was defined as completion of definitive surgery within 3 d of diagnosis. Elective surgery was defined as completion of definitive surgery 4 d or more after diagnosis. Definitive surgery included procedure codes with total or
partial gastric resection, with or without resection of surrounding organs or tissues. Codes for local tumor destruction were excluded from analysis. The primary outcome was the comparison of overall survival between urgent and elective cohorts. Secondary outcomes analyzed comparing these cohorts include 30- and 90-d mortality, 30-d unplanned readmission, surgical margin positivity, lymph node positivity and yield, and length of hospital stay. Univariate and multivariate analysis of surgical and oncologic outcomes was performed, with consideration of patient factors. Cohorts were compared using the chisquare test for categorical variables and the rank-sum test for continuous variables. Overall survival was analyzed using Cox proportional hazard analysis and the KaplanMeier Estimate. Multivariate logistic regression with stepwise selection was used to assess the study’s outcomes. Results are reported as frequency (%), hazard ratios (HRs), and odds ratios (ORs) with 95% confidence intervals (CIs), unless specified otherwise. All analysis was conducted using SAS version 9.4 (SAS Institute Inc, Cary, NC). All tests were 2-sided, and P values < 0.05 were considered statistically significant.
Results Univariate analysis: patient characteristics A total of 26,432 records were analyzed from patients diagnosed from 2004 through 2015, of which 2964 (11.2%) had urgent surgery and 23,468 (88.8%) underwent elective surgery. Univariate analysis comparing clinicodemographic factors is shown in Table 1. Overall, most patients were male (63.9%), and there was a significantly higher proportion of males in the elective cohort (64.5% versus 59.4%, P < 0.0001). No significant difference in patient comorbidity scores was observed (P ¼ 0.958), and 65.6% of patients overall had no Charlson comorbid conditions. Overall, most patients were white (71.7%), and there was a significantly lower percentage of white patients among the urgent cohort (72.2% versus 67.3%, P < 0.0001). Urgent patients tended to live in areas with lower education levels compared to elective patients (P < 0.0001). Table 1 also shows the urgent cohort had a higher proportion of patients treated at community or comprehensive community facility and were more likely to be at a facility with a low volume of gastric resections (P < 0.0001). Regarding tumor characteristics and surgical intervention, patients in the urgent cohort had larger, more welldifferentiated tumors with higher pathologic stage and were more likely to be in the gastric antrum and pylorus rather than the cardia (all P values < 0.0001). Not surprisingly, the urgent cohort was also more likely to undergo adjuvant chemotherapy (35.8% versus 26.6%, P < 0.0001). Unadjusted outcomes data in Table 2 show increased 30- and 90-d mortality, increased positive surgical margins, increased incidence of positive lymph nodes, and decreased quality lymph node dissections for patients in the urgent surgery cohort (all P values < 0.0001). There was no significant difference in 30d readmission or length of stay between the groups (P ¼ 0.7864, P ¼ 0.699, respectively).
621
fi s h e r e t a l u r g e n t g a s t r i c c a n c e r s u r g e r y o u t c o m e s
Table 1 e Elective versus urgent patient demographics. Variable
Diagnosis to definitive surgery, days
Elective
Urgent
n ¼ 23,468
n ¼ 2964
38 (19, 92)
0 (0, 2)
18-29
104 (0.4)
30-39
529 (2.3)
78 (2.6)
40-49
1969 (8.4)
213 (7.2)
50-59
4414 (18.8)
500 (16.9)
60-69
6385 (27.2)
663 (22.4)
70-79
6512 (27.8)
853 (28.8)
80þ
3555 (15.2)
22 (0.7)
635 (21.4)
8325 (35.5)
1204 (40.6)
15,143 (64.5)
1760 (59.4)
16,954 (72.2)
1994 (67.3)
Black
3660 (15.6)
573 (19.3)
Other
2854 (12.2)
397 (13.4)
0
15,392 (65.6)
1936 (65.3)
1
5981 (25.5)
761 (25.7)
2þ
2095 (8.9)
267 (9.0)
736 (3.1)
137 (4.6)
<0.0001
Sex
Male
<0.0001
Race White
Charlson Comorbidity Index
0.9582
<0.0001
Insurance status Not insured Private
8497 (36.2)
909 (30.7)
Medicaid
1624 (6.9)
217 (7.3)
Medicare
12,059 (51.4)
1614 (54.5)
Other government
213 (0.9)
22 (0.7)
Not reported
339 (1.4)
65 (2.2)
492 (21.0)
716 (24.2)
2nd
6155 (26.2)
820 (27.7)
3rd
7176 (30.6)
903 (30.5)
4thehighest
4848 (20.7)
488 (16.5)
Unreported
361 (1.5)
37 (1.3)
<0.0001
High school diploma quartile 1stelowest
<0.0001
Facility program type Community
1479 (6.3)
342 (11.5)
Comprehensive community
8087 (34.5)
1324 (44.7)
Academic/research
11,170 (47.6)
938 (31.7)
Integrated network
2732 (11.6)
360 (12.2) <0.0001
Facility gastric surgery volume High volume
9687 (41.3)
762 (25.7)
Low volume
13,781 (58.7)
2202 (74.3)
Tumor size (mm)
40 (24, 60)
45 (30, 67.5)
<0.0001 <0.0001
TNM pathologic T stage 0/X
1312 (5.6)
152 (5.1)
1
2597 (11.1)
287 (9.7)
2
2642 (11.3)
240 (8.1)
3
8003 (34.1)
1056 (35.6)
4
316 (1.4)
69 (2.3)
8598 (36.6)
1160 (39.1)
Not reported
<0.0001
Age group
Female
P value
(continued)
622
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 2 0 ( 2 4 5 ) 6 1 9 e6 2 8
Table 1 e (continued ) Variable
Elective
Urgent
n ¼ 23,468
n ¼ 2964 <0.0001
TNM pathologic N stage 0/X
11,143 (47.5)
1255 (42.3)
1
6628 (28.2)
887 (29.9)
p2
3596 (15.3)
516 (17.4)
p3
658 (2.8)
71 (2.4)
p3A
983 (4.2)
163 (5.5)
P3B
460 (2.0)
72 (2.4)
8422 (35.9)
880 (29.7)
<0.0001
TNM pathologic stage group I II
6491 (27.7)
800 (27.0)
III
8555 (36.5)
1284 (43.3)
Tumor grade
0.0002
Unreported Poorly or undifferentiated
973 (4.2)
110 (3.7)
14,360 (61.2)
1712 (57.8)
8135 (34.7)
1142 (38.5)
Body of stomach
1897 (8.1)
209 (7.1)
Cardia, NOS
7873 (33.6)
429 (14.5)
Well to moderately differentiated
<0.0001
Primary site
Fundus
73 (3.1)
95 (3.2)
Gastric antrum
5425 (23.1)
910 (30.7)
Greater curvature
1035 (4.4)
124 (4.2)
Lesser curvature
2378 (10.1)
284 (9.6)
Overlapping lesion
1507 (6.4)
235 (7.9)
688 (2.9)
203 (6.9)
1932 (8.2)
475 (16.0)
Pylorus Stomach, NOS
P value
Chemotherapy initial timing Neoadjuvant chemotherapy
5717 (24.4)
72 (2.4)
<0.0001
Adjuvant chemotherapy
6251 (26.6)
1061 (35.8)
<0.0001
Multivariate logistic regression: factors associated with urgent gastric surgery Table 3 shows the multivariate analysis of factors associated to urgent surgery. The youngest and oldest patient groups showed increased likelihood of urgent surgery compared to
patients aged 50-59 y (OR 1.66, 95% CI [1.02, 2.69], OR 1.27 95% CI [1.09, 1.47], respectively, P < 0.0001). The uninsured were significantly more likely to have urgent surgery compared to those with private insurance, Medicaid, or Medicare (OR 0.74, 95% CI [0.60, 0.90], OR 0.77, 95% CI [0.61, 0.97], OR 0.72, 95% CI [0.59, 0.89], respectively, P ¼ 0.0064). Compared to academic/
Table 2 e Unadjusted perioperative outcomes. Outcome
Elective
Urgent
n ¼ 23,468
n ¼ 2964
30-d mortality
817 (3.5)
171 (5.8)
<0.0001
90-d mortality
1687 (7.2)
316 (10.7)
<0.0001
30-d unplanned readmission
1639 (7.0)
211 (7.1)
Positive surgical margin
2598 (11.1)
532 (18.0)
<0.0001
12,906 (55.0)
1778 (60.0)
<0.0001
Positive node(s) Lymph node yield Quality lymph node dissection (15þ nodes) Surgical inpatient stay, days
15 (9, 22) 12,103 (51.6) 9 (7, 13)
12 (7, 19) 1198 (40.4) 9 (6, 13)
P value
0.7864
<0.0001 <0.0001 0.699
623
fi s h e r e t a l u r g e n t g a s t r i c c a n c e r s u r g e r y o u t c o m e s
Table 3 e Factors associated with urgent gastric surgery. Variable
Odds ratio
95% confidence interval Lower
Upper <0.0001
Age group 18-29
1.658
1.023
2.687
30-39
1.118
0.860
1.454
40-49
0.861
0.724
1.023
50-59
Ref
Ref
Ref
60-69
0.923
0.809
1.052
70-79
1.065
0.925
1.226
80þ
1.265
1.086
1.473
Ref
Ref
Ref
Private
0.735
0.601
0.899
Medicaid
0.768
0.606
0.972
Medicare
0.722
0.586
0.890
Other government
0.644
0.396
1.049
Not reported
1.047
0.751
1.460
Insurance status Not insured
0.0064
High school diploma quartile
0.055
1stelowest
1.223
1.078
1.387
2nd
1.205
1.067
1.360
3rd
1.201
1.067
1.353
4thehighest
Ref
Ref
Ref
Unreported
0.908
0.635
1.298
Community
1.875
1.615
2.177
Comprehensive community
1.676
<0.0001
Facility program type
1.512
1.363
Academic/research
Ref
Ref
Ref
Integrated network
1.355
1.184
1.551
High volume
Ref
Ref
Ref
Low volume
1.410
1.272
1.562
<0.0001
Facility gastric surgery volume
TNM pathologic T stage 0/X
0.0257 Ref
Ref
Ref
1
1.163
0.925
1.462
2
0.878
0.699
1.104
3
1.023
0.841
1.244
4
1.188
0.858
1.646
Not reported
1.108
0.914
1.344
Ref
Ref
Ref
1
0.903
0.798
1.022
p2
0.762
0.643
0.902
p3
0.839
0.629
1.119
p3A
0.719
0.574
0.901
P3B
0.655
0.488
0.879
I
Ref
Ref
Ref
II
1.277
1.112
1.467
III
1.768
1.47
2.122
0.675
0.575
0.793
TNM pathologic N stage 0/X
0.0108
<0.0001
TNM pathologic stage group
<0.0001
Primary site Body of stomach
P value
(continued)
624
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 2 0 ( 2 4 5 ) 6 1 9 e6 2 8
Table 3 e (continued ) Variable
Odds ratio
95% confidence interval Lower
Upper
Cardia, NOS
0.370
0.326
0.420
Fundus
0.775
0.617
0.972
Ref
Ref
Ref
Greater curvature
0.713
0.583
0.872
Lesser curvature
0.722
0.625
0.833
Overlapping lesion
0.923
0.789
1.080
Pylorus
1.672
1.404
1.990
Stomach, NOS
1.454
1.283
1.646
Gastric antrum
P value
OR > 1 indicates increased likelihood of urgent surgery.
research cancer program types, community, comprehensive, and integrated network facilities showed increased likelihood for urgent gastric surgery (OR 1.88, 95% CI [1.62, 2.18], OR 1.51, 95% CI [1.36, 1.68], OR 1.36, 95% CI [1.18, 1.55], respectively, P < 0.0001). Low-volume facilities were also more likely to perform urgent surgeries (OR 1.41, 95% CI [1.27, 1.56], P < 0.0001). Compared to stage I, patients were increasingly more likely to undergo urgent surgery with stage II and stage III tumors (OR 1.28, 95% CI [1.11, 1.47], OR 1.77, 95% CI [1.47, 2.12], respectively, P < 0.0001). The primary site of the gastric tumor was also found to be associated with the need for urgent surgery (P < 0.0001).
Kaplan-Meier curve and Cox proportional hazard model: overall survival Figure 1 shows the Kaplan-Meier curve comparing overall survival between cohorts (P < 0.0001). Median survival for elective surgeries was 47.7 mo versus 30.9 mo for urgent operations. Table 4displays the Cox proportional hazard model for overall survival. After adjustment for confounders, urgent gastric surgeries remained at increased risk of all-cause mortality (HR 1.21, 95% CI [1.15, 1.27], P < 0.0001). Increasing age at diagnosis continued to be a factor in all-cause mortality (P < 0.0001), as did increased comorbidities (P < 0.0001), and males were at more risk than females (HR 1.07, 95% CI [1.03, 1.11], P ¼ 0.0002). Uninsured patients showed no difference in
survival from patients with private insurance or Medicaid (HR 0.94, 95% CI [0.85, 1.04], HR 1.06, 95% CI [0.94, 1.19], respectively); however, patients with Medicare were at increased mortality risk (HR 1.13, 95% [1.02, 1.26], P < 0.0001). Compared to patients from the most educated areas, lower education was associated with worse survival (P < 0.0001). Definitive treatment at low-volume centers had increased risk of mortality compared to high-volume facilities (HR 1.23, 95% CI [1.18, 1.28], P < 0.00001). As pathologic T stage increased, overall survival worsened (P < 0.0001), a pattern followed by increasing N stages (P < 0.0001), and by pathological stage group (P < 0.0001). The primary site of the gastric tumor was also found to be associated with significant differences in survival (P < 0.0001).
Multivariable analysis: urgent gastric surgery outcomes Table 5 shows the effect of urgent gastric surgery on selected outcomes. The rate of quality dissection is decreased (OR 0.70, 95% CI [0.64, 0.76], P < 0.001). The rate of positive surgical margin, 30-d mortality, and 90-d mortality are also increased (OR 1.48, 95% CI [1.32, 1.65], OR 1.38, 95% CI [1.16, 1.65], and OR 1.30, 95% CI [1.14, 1.49], respectively, all P < 0.001). There was no significant change in unplanned readmission or lymph node positivity.
Yearly trend Looking at the trend of data over the study period in Figure 2, the rate of urgent surgery has decreased slightly each year, from 313 cases in 2004 to 238 cases in 2014. The rate of elective surgery has increased by approximately 27% during that same period, from 1941 cases in 2004 to 2473 cases in 2014 (P < 0.0001). Urgent surgeries made up 13.9% of total gastric cancer resections in 2004 and steadily dropped to 8.8% of surgeries in 2014.
Discussion
Fig. 1 e Kaplan-Meier curve comparing urgent surgery versus elective surgery. (Color version of figure is available online.)
Obstruction, perforation, and hemorrhage are well-known complications of gastric cancer,18 leading to unstable patients who may require urgent surgical intervention. However, the data are quite limited regarding outcomes for these
625
fi s h e r e t a l u r g e n t g a s t r i c c a n c e r s u r g e r y o u t c o m e s
Table 4 e Cox proportional hazard model: overall survival. Variable
Urgent (ref: elective)
Hazard ratio
1.207
95% confidence interval Lower
Upper
1.149
1.269
18-29
0.797
0.582
1.090
30-39
0.832
0.732
0.946
40-49
0.892
0.827
0.962
50-59
Ref
Ref
Ref
60-69
1.102
1.043
1.165
70-79
1.405
1.322
1.492
80þ
2.185
2.046
2.334
Ref
Ref
Ref
1.071
1.033
1.109
White
Ref
Ref
Ref
Black
1.024
0.976
1.074
Other
0.718
0.676
0.761
Ref
Ref
Ref
Sex
Male
0.0002
<0.0001
Race
<0.0001
Charlson Comorbidity Index 0 1
1.078
1.038
1.120
2þ
1.367
1.294
1.444
Ref
Ref
Ref
Private
0.936
0.845
1.037
Medicaid
1.061
0.943
1.193
Medicare
1.131
1.018
1.256
<0.0001
Insurance status Not insured
Other government
1.065
0.869
1.305
Not reported
1.097
0.932
1.292
1stelowest
1.090
1.033
1.150
2nd
1.172
1.115
1.231
3rd
1.167
<0.0001
High school diploma quartile
1.113
1.061
4thehighest
Ref
Ref
Ref
Unreported
2.054
1.824
2.313 <0.0001
Facility program type Community
1.151
1.075
1.233
Comprehensive community
1.104
1.058
1.152
Academic/research
Ref
Ref
Ref
Integrated network
1.093
1.034
1.155 <0.0001
Facility gastric surgery volume High volume
Ref
Ref
Ref
Low volume
1.226
1.177
1.277
Ref
Ref
Ref
1
2.934
2.59
3.324
2
2.821
2.496
3.187
3
3.293
2.945
3.681
4
4.217
3.591
4.951
Not reported
3.348
2.993
3.745
<0.0001
TNM pathologic T stage 0/X
<0.0001 <0.0001
Age group
Female
P value
(continued)
626
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 2 0 ( 2 4 5 ) 6 1 9 e6 2 8
Table 4 e (continued ) Variable
Hazard ratio
95% confidence interval Lower
Upper <0.0001
TNM pathologic N stage 0/X
Ref
Ref
Ref
p1
1.344
1.272
1.419
p2
1.516
1.409
1.630
p3
1.729
1.558
1.919
p3A
1.770
1.603
1.954
p3B
2.451
2.171
2.768
I
Ref
Ref
Ref
II
1.536
1.44
1.638
III
2.353
2.161
2.561
<0.0001
TNM pathologic stage group
<0.0001
Primary site Body of stomach
0.943
0.879
Cardia, NOS
1.362
1.295
1.432
Fundus
1.218
1.107
1.340
Gastric antrum
P value
1.011
Ref
Ref
Ref
Greater curvature
0.941
0.863
1.026
Lesser curvature
0.901
0.845
0.961
Overlapping lesion
1.146
1.068
1.229
Pylorus
1.061
0.967
1.164
Stomach, NOS
1.141
1.072
1.215
<0.0001
Hazard ratio > 1 indicates increased risk of all-cause mortality.
patients. In 1980, Hahnloser et al. published a small case series of 13 patients who underwent either 1- or 2-stage procedure for gastric perforation, with more stable patients undergoing immediate gastric resection with reasonable survival.19 Additional studies, primarily performed at single institutions with limited numbers of patients, recommend similar treatment, with the emphasis on stabilizing the patient, with the oncologic treatment addressed either once the patient was stable enough to undergo more extensive treatment or once the correct diagnosis had been achieved.20 Survival was based on pathology of the tumor, and some studies found that patients undergoing emergency gastrectomy had satisfactory results, with stage I and II tumors having similar long-term survival when compared to historic norms.21 Others found that radical gastrectomy was not advised at the index surgery, recommending a second procedure (with lymphadenectomy) once the patient had recovered and had been staged appropriately.22 Kasakura et al. found that patients with early-stage carcinoma who underwent emergency R0 resections had minimal complications,23 whereas Blackshaw et al. showed that patients undergoing emergency surgery for gastric cancer have significantly decreased median survival (6 versus 12 mo).24 Few studies have had enough patients to draw strong conclusions, and even fewer have long-term data on oncologic outcomes for urgent gastric resections. In addition, given the improvement in outcomes seen with adoption of neoadjuvant chemotherapy,7 there is concern that patients undergoing urgent surgery likely have worse overall mortality and morbidity. This study validates that concern,
demonstrating that patients undergoing urgent surgery for gastric cancer have significantly decreased rates of quality resection, with decreased lymph node yield and increased positive margin, as well as increased 30-d, 90-d, and overall mortality when compared patients undergoing elective surgery. The decision to proceed with definitive resection is dependent on preoperative and intraoperative patient factors, including clinical stability and overall level of fitness. As might be expected, larger and more locally advanced tumors were associated with an increased rate of urgent surgery and had significantly higher mortality, even when
Fig. 2 e Yearly trends in urgent and elective gastric resection from 2004 to 2014. (Color version of figure is available online.)
627
fi s h e r e t a l u r g e n t g a s t r i c c a n c e r s u r g e r y o u t c o m e s
Table 5 e Effect of urgent gastric surgery on outcomes. Variable
Odds ratio
95% confidence interval Lower
Upper
P value
Quality lymph node dissection (15 þ nodes)
0.693
0.636
0.755
<0.0001
30-d unplanned readmission
1.017
0.875
1.183
0.8223 <0.0001
Positive surgical margin
1.478
1.320
1.654
30-d mortality
1.379
1.156
1.646
0.0004
90-d mortality
1.298
1.135
1.485
<0.0001
Regional node(s) positive
0.844
0.614
1.161
0.2983
Odds ratio > 1 indicates increased likelihood of outcome compared to elective.
matched stage-for-stage with those undergoing elective surgery. Data from urgent colorectal surgery show similar data, with Ming-Gao et al. showing higher mortality for urgent colon resections,25 and Kim et al. demonstrating higher complication rates in emergent colectomies for obstructing colon cancer.26 More advanced disease may necessitate a larger operation in patients who are already unstable, with the increased mortality likely related to longer operative time, friable tissues, and operative blood loss in patients with already diminished physiologic reserves. One additional area of concern is that there are low rates of adjuvant chemotherapy in the urgent group, with low overall rates of chemotherapy in both groups (only 51% receive neoadjuvant or adjuvant chemotherapy in the elective group compared with 38% in the urgent group). This may represent failure of adherence to recommended guidelines for the treatment of gastric cancer, or it may be that these patients failed to have appropriate follow-up with medical oncologists for initiation of chemotherapy. It is also possible that some patients underwent radiation only (which is not captured in the NCDB) or had adjuvant thermotherapy at a facility which was not recorded in the NCDB. The difference in chemotherapy rates was not significant in multivariate analysis but is still a concerning finding worth mentioning, and further investigation is necessary to understand the cause for such an observed low incidence of adjuvant therapy.
Limitations The NCDB was specifically chosen for this study because it contains long-term patient data with tumor and treatment data, which is not available in many other national databases. However, it has limitations. It is a retrospective database and variables are limited only to previously collected data. Urgent surgery is not specifically coded in the database, and this definition is central to our study question. We followed the example of Xu et al.27 and defined “urgent surgery” as definitive surgical intervention within 3 d of the diagnosis of gastric adenocarcinoma. While this duration may seem arbitrary, the definition was chosen to include both critically ill patients taken emergently to the operating room and less critical patients who may need resuscitation or a brief metastatic workup. Thus, patients undergoing emergent surgery for perforation are included in the data, as are patients presenting with obstruction or hemorrhage who may not be taken immediately to the operating room. It is reasonable to surmise
that if these patients were diagnosed with gastric adenocarcinoma within 3 d of definitive surgery, it was either a biopsy during the workup for GI hemorrhage, or intraoperatively when a specimen was sent for frozen or final surgical pathology. Given the brief interval included in our definition, it is also likely that many of the urgent surgery cohort presented with perforation, as obstruction and hemorrhage may be temporized with other therapies, including advanced endoscopy, radiation, and surgical decompression with gastrostomy or jejunostomy, leaving definitive treatment outside the definition for urgent surgery. However, it is impossible to determine the specific cause for any patient’s urgent surgery within the constraints of this database. While not without limitations, this definition provides a useful tool for investigating outcomes in this patient population. In addition, the NCDB provides hospital data but has no information about the surgeons performing each surgery, so it is impossible to see variation between surgeons at a single facility, or to know if the surgeon performing the resection was a board-certified surgical oncologist. Patients undergoing urgent surgical resection had a decreased rate of adequate lymph node yield and an increased positive surgical margin, which suggests that less-experienced surgeons may be performing many of these emergent operations; however, that level of surgeon information is not available in the database. Finally, the NCDB provides only overall survival, not diseasespecific survival, so it is not possible to demonstrate cause of death for these patients. Despite these limitations, this study demonstrates that urgent surgery has significantly worse outcomes in lymph node yield, positive surgical margins, 30- and 90-d mortality, and overall survival in patients undergoing urgent surgery for gastric adenocarcinoma.
Conclusions Urgent surgery for gastric cancer is associated with significantly worse surgical and oncologic outcomes. Stable patients who require urgent surgical resection for gastric masses may benefit from referral to a high-volume center for resection by an experienced surgeon. If that is not possible, consultation with a surgical oncologist may be beneficial for the surgeon who performs fewer gastric resections. For unstable patients requiring emergent surgical treatment, a damage-control procedure should be considered an individual based according to preoperative and intraoperative factors. Consultation
628
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 2 0 ( 2 4 5 ) 6 1 9 e6 2 8
with or transfer to an experienced surgical oncologist for resection may lead to improved surgical margins and lymph node yield, leading to improved survival. All patients undergoing urgent resection for gastric cancer should also be referred to a surgical oncologist and medical oncologist to ensure that the patient receives appropriate adjuvant therapy and cancer surveillance.
Acknowledgment Authors’ contributions: B.W.F. was the primary researcher and manuscript author. He completed a majority of the writing for the manuscript and also performed the literature review. M.F. performed the statistical analysis and assisted in writing the methods and results section. K.Y. assisted in research mechanics and methodology, in addition to editing the final manuscript. M.S. assisted in refining the research question and assisted in editing the final manuscript. J.B. assisted in defining and refining the research question and assisted in editing the final manuscript. T.A. was the primary faculty research mentor and assisted in refining the research question, in addition to assisting with research methodology, data interpretation, and editing the final manuscript.
Disclosure The authors have no conflicts of interest to report and there was no extra-institutional funding provided for this project.
references
1. Soerjomataram I, Lortet-Tieulent J, Parkin DM, et al. Global burden of cancer in 2008: a systematic analysis of disabilityadjusted life-years in 12 world regions. Lancet. 2012;380:1840e1850. 2. Casamayor M, Morlock R, Maeda H, Ajani J. Targeted literature review of the global burden of gastric cancer. Ecancermedicalscience. 2018;12:883. 3. GBD 2017 Causes of Death Collaborators. Global, regional, and national age-sex-specific mortality for 282 causes of death in 195 countries and territories, 1980-2017: a systematic analysis for the global burden of disease study 2017. Lancet. 2018;392:1736e1788. 4. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2019. CA Cancer J Clin. 2019;69:7e34. 5. Amin MB, Greene FL, Edge SB, et al. The eighth edition AJCC cancer staging manual: Continuing to build a bridge from a population-based to a more “personalized” approach to cancer staging. CA Cancer J Clin. 2017;67:93e99. 6. Macdonald JS, Smalley SR, Benedetti J, et al. Chemoradiotherapy after surgery compared with surgery alone for adenocarcinoma of the stomach or gastroesophageal junction. N Engl J Med. 2001;345:725e730. 7. Cunningham D, Allum WH, Stenning SP, et al. Perioperative chemotherapy versus surgery alone for resectable gastroesophageal cancer. N Engl J Med. 2006;355:11e20.
8. Mihmanli M, Ilhan E, Idiz UO, Alemdar A, Demir U. Recent developments and innovations in gastric cancer. World J Gastroenterol. 2016;22:4307e4320. 9. Chen L, Ma LG, Xi HQ. [The conversion therapy for unresectable gastric cancer]. Zhonghua Wai Ke Za Zhi. 2016;54:169e171. 10. Oh D, Bang Y. Adjuvant and neoadjuvant therapy for gastric cancer. Curr Treat Options Oncol. 2013;14(3):311e320. 11. Tegels JJ, De Maat MF, Hulsewe´ KW, Hoofwijk AG, Stoot JH. Improving the outcomes in gastric cancer surgery. World J Gastroenterol. 2014;20:13692e13704. 12. Smyth EC, Wotherspoon A, Peckitt C, et al. Mismatch repair deficiency, microsatellite instability, and survival: an exploratory analysis of the medical research council adjuvant gastric infusional chemotherapy (MAGIC) trial. JAMA Oncol. 2017;3:1197e1203. 13. Song Z, Wu Y, Yang J, Yang D, Fang X. Progress in the treatment of advanced gastric cancer. Tumour Biol. 2017;39, 1010428317714626. 14. Choi AH, Kim J, Chao J. Perioperative chemotherapy for resectable gastric cancer: MAGIC and beyond. World J Gastroenterol. 2015;21:7343e7348. 15. Aoyama T, Yoshikawa T. Adjuvant therapy for locally advanced gastric cancer. Surg Today. 2017;47:1295e1302. 16. Ignjatovic N, Stojanov D, Djordjevic M, Ignjatovic J, Benedeto Stojanov D, Milojkovic B. Perforation of gastric cancer - what should the surgeon do? Bosn J Basic Med Sci. 2016;16:222e226. 17. Smith DL, Elting LS, Learn PA, Raut CP, Mansfield PF. Factors influencing the volume-outcome relationship in gastrectomies: a population-based study. Ann Surg Oncol. 2007;14:1846e1852. 18. Itano S. Early gastric cancer and its complications: bleeding, perforation and pyloric stenosis. Acta Med Okayama. 1983;37:431e440. 19. Hahnloser P, Bruttin JM, Cavin R, Petropoulos P. [Perforation of cancer or stomach ulcer: a 1- or 2-stage operation]. Helv Chir Acta. 1980;47:565e574. 20. Fox JG, Hunt PS. Management of acute bleeding gastric malignancy. Aust N Z J Surg. 1993;63:462e465. 21. Adachi Y, Mori M, Maehara Y, Matsumata T, Okudaira Y, Sugimachi K. Surgical results of perforated gastric carcinoma: an analysis of 155 Japanese patients. Am J Gastroenterol. 1997;92:516e518. 22. Lehnert T, Buhl K, Dueck M, Hinz U, Herfarth C. Two-stage radical gastrectomy for perforated gastric cancer. Eur J Surg Oncol. 2000;26:780e784. 23. Kasakura Y, Ajani JA, Mochizuki F, Morishita Y, Fujii M, Takayama T. Outcomes after emergency surgery for gastric perforation or severe bleeding in patients with gastric cancer. J Surg Oncol. 2002;80:181e185. 24. Blackshaw GR, Stephens MR, Lewis WG, et al. Prognostic significance of acute presentation with emergency complications of gastric cancer. Gastric Cancer. 2004;7:91e96. 25. Ming-Gao G, Jian-Zhong D, Yu W, You-Ben F, Xin-Yu H. Colorectal cancer treatment in octogenarians: elective or emergency surgery? World J Surg Oncol. 2014;12:386. 26. Kim J, Mittal R, Konyalian V, King J, Stamos MJ, Kumar RR. Outcome analysis of patients undergoing colorectal resection for emergent and elective indications. Am Surg. 2007;73:991e993. 27. Xu Z, Becerra AZ, Aquina CT, et al. Emergent colectomy is independently associated with decreased long-term overall survival in colon cancer patients. J Gastrointest Surg. 2017;21:543e553.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.JournalofSurgicalResearch.com
Association for Academic Surgery
Urgent Surgery for Gastric Adenocarcinoma: A Study of the National Cancer Database Benjamin W. Fisher, MD,* Marcus Fluck, BS, Katelyn Young, BS, Mohsen Shabahang, MD, PhD, Joseph Blansfield, MD, and Tania K. Arora, MD a
Department of Surgery, Geisinger Medical Center, Danville, Pennsylvania
article info
abstract
Article history:
Background: Gastric adenocarcinoma is a leading cause of cancer death worldwide and, in
Received 4 March 2019
the United States, can present emergently with upper GI hemorrhage, obstruction, or
Received in revised form
perforation. No large studies have examined how urgent surgery affects patient outcomes.
13 July 2019
This study examines the outcomes of urgent versus elective surgery for gastric cancer.
Accepted 19 July 2019
Materials and methods: Patients with gastric adenocarcinoma from the National Cancer
Available online xxx
Database from 2004 to 2015 were examined retrospectively. Patients with metastatic disease or incomplete data were excluded. Urgent surgery was defined as definitive surgery
Keywords:
within 3 d of diagnosis. Univariate and multivariate analysis of patient factors, surgical
Gastric adenocarcinoma
outcomes, and oncologic data was performed. P-values <0.05 were statistically significant.
Gastric cancer
Results: Of 26,116 total patients, 2964 had urgent surgery and 23,468 had elective surgery.
Urgent surgery
Urgent surgery patients were significantly older, were female, were nonwhite, had higher
Emergent surgery
pathologic stage, and were treated at a low-volume center. Urgent surgery was associated
National Cancer Database
with decreased quality lymph node harvest (odds ratio [OR] 0.68 95% confidence interval
NCDB
{CI} [0.62, 0.74]), increased positive surgical margin (OR 1.48, 95% CI [1.32, 1.65]), increased 30-d mortality (OR 1.38, 95% CI [1.16, 1.65]), increased 90-d mortality (OR 1.30, 95% CI [1.14, 1.49]), and decreased overall survival (hazard ratio 1.21, 95% CI [1.15, 1.27]). Conclusions: Urgent surgery for gastric cancer is associated with significantly worse outcomes than elective surgery. Stable patients requiring urgent surgical resection for gastric cancer may benefit from referral to a high-volume center for resection by an experienced surgeon. Patients undergoing urgent resection for gastric cancer should be referred to surgical and medical oncologists to ensure they receive appropriate adjuvant therapy and surveillance. ª 2019 Elsevier Inc. All rights reserved.
Introduction Gastric cancer is one of the leading causes of death worldwide, representing nearly one million new cases annually and
ranking as the third leading cause of cancer death (723,000 yearly).1-3 Although the incidence is much lower in the United States, it still affects a significant number of patients, with 27,510 new diagnoses and 11,140 deaths estimated for 2019.4
* Corresponding author. Geisinger Medical Center, General Surgery, 100 N Academy Avenue, Danville, PA, 17822. Tel.: þ1 570 271 6361; fax: þ1 570-271-8324. 0022-4804/$ e see front matter ª 2019 Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.jss.2019.07.073
620
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 2 0 ( 2 4 5 ) 6 1 9 e6 2 8
Prognosis varies based on stage at diagnosis; 5-y survival after treatment drops precipitously from 88% to 94% for stage IA-B to only 18%-54% for stage IIIA-C disease.5 In 2001, Macdonald et al. demonstrated that postoperative chemoradiation improved median survival by almost 50% when compared to surgical treatment alone.6 The MAGIC trial in 2006 demonstrated that perioperative chemotherapy reduced tumor size and stage at the time of surgery, as well as improving disease-free and overall survival.7 D2 lymph node dissection has now been accepted as the standard of care with reduced rates of local and regional recurrence.8-11 Improved treatment algorithms in the past decade have shown promise in decreasing mortality and extending survival, with gastric cancer mortality decreasing by 9% worldwide over the last 10 y3; however, there remains additional improvement to be made. There is still no consensus regarding the best treatment, with ongoing research into the optimal regimen and timing of therapy.10,12-16 Although research continues to make progress in elective treatment of gastric cancer, there is a relative paucity of data regarding outcomes for three of the well-known complication of gastric cancer: perforation, obstruction, and hemorrhage. It is unclear if urgent surgery for these complications changes adherence to standard treatment protocols and potentially compromises the quality of oncologic resection. This study compares patients undergoing elective surgical resection with patients undergoing urgent surgery to evaluate and characterize the differences in surgical and oncological outcomes.
Materials and methods The National Cancer Database (NCDB) is a joint project of the Commission on Cancer of the American College of Surgeons and the American Cancer Society. The database is a nationwide, facility-based, comprehensive clinical surveillance resource oncology data set; it captures information from approximately 1500 Commission on Cancer accredited hospitals and more than 70% of all newly diagnosed malignancies in the United States. This research was deemed exempt before initiation of study activities. A retrospective review of patients undergoing surgical resection for gastric adenocarcinoma from 2004 to 2015 was performed on the NCDB Stomach participant user file. The TNM pathological stage groups I-III were included. Adenocarcinomas were identified by the ICD-03 histology codes: 8490, 8245, 8210, 8263, 8261, 8572, 8255, 8574, 8570, 8144, 8140, 8560, 8021, 8145, 8010, 8020, 8310, 8243, 8576, 8012, 8575, 8244, 8323, 8480, 8481, 8000, 8260, 8050, 8141, 8142, 8230, 8143, 8211, 8051, 8262. Records missing days from diagnosis to definitive surgery were excluded as well as patients with evidence of metastatic disease. High facility volume cutoff is defined as a facility count of all stomach cancer resections 15 in a given year.17 Quality lymph node dissection was defined as 15 regional nodes examined. Patients with unknown survival or follow-up data were excluded. Urgent surgery was defined as completion of definitive surgery within 3 d of diagnosis. Elective surgery was defined as completion of definitive surgery 4 d or more after diagnosis. Definitive surgery included procedure codes with total or
partial gastric resection, with or without resection of surrounding organs or tissues. Codes for local tumor destruction were excluded from analysis. The primary outcome was the comparison of overall survival between urgent and elective cohorts. Secondary outcomes analyzed comparing these cohorts include 30- and 90-d mortality, 30-d unplanned readmission, surgical margin positivity, lymph node positivity and yield, and length of hospital stay. Univariate and multivariate analysis of surgical and oncologic outcomes was performed, with consideration of patient factors. Cohorts were compared using the chisquare test for categorical variables and the rank-sum test for continuous variables. Overall survival was analyzed using Cox proportional hazard analysis and the KaplanMeier Estimate. Multivariate logistic regression with stepwise selection was used to assess the study’s outcomes. Results are reported as frequency (%), hazard ratios (HRs), and odds ratios (ORs) with 95% confidence intervals (CIs), unless specified otherwise. All analysis was conducted using SAS version 9.4 (SAS Institute Inc, Cary, NC). All tests were 2-sided, and P values < 0.05 were considered statistically significant.
Results Univariate analysis: patient characteristics A total of 26,432 records were analyzed from patients diagnosed from 2004 through 2015, of which 2964 (11.2%) had urgent surgery and 23,468 (88.8%) underwent elective surgery. Univariate analysis comparing clinicodemographic factors is shown in Table 1. Overall, most patients were male (63.9%), and there was a significantly higher proportion of males in the elective cohort (64.5% versus 59.4%, P < 0.0001). No significant difference in patient comorbidity scores was observed (P ¼ 0.958), and 65.6% of patients overall had no Charlson comorbid conditions. Overall, most patients were white (71.7%), and there was a significantly lower percentage of white patients among the urgent cohort (72.2% versus 67.3%, P < 0.0001). Urgent patients tended to live in areas with lower education levels compared to elective patients (P < 0.0001). Table 1 also shows the urgent cohort had a higher proportion of patients treated at community or comprehensive community facility and were more likely to be at a facility with a low volume of gastric resections (P < 0.0001). Regarding tumor characteristics and surgical intervention, patients in the urgent cohort had larger, more welldifferentiated tumors with higher pathologic stage and were more likely to be in the gastric antrum and pylorus rather than the cardia (all P values < 0.0001). Not surprisingly, the urgent cohort was also more likely to undergo adjuvant chemotherapy (35.8% versus 26.6%, P < 0.0001). Unadjusted outcomes data in Table 2 show increased 30- and 90-d mortality, increased positive surgical margins, increased incidence of positive lymph nodes, and decreased quality lymph node dissections for patients in the urgent surgery cohort (all P values < 0.0001). There was no significant difference in 30d readmission or length of stay between the groups (P ¼ 0.7864, P ¼ 0.699, respectively).
621
fi s h e r e t a l u r g e n t g a s t r i c c a n c e r s u r g e r y o u t c o m e s
Table 1 e Elective versus urgent patient demographics. Variable
Diagnosis to definitive surgery, days
Elective
Urgent
n ¼ 23,468
n ¼ 2964
38 (19, 92)
0 (0, 2)
18-29
104 (0.4)
30-39
529 (2.3)
78 (2.6)
40-49
1969 (8.4)
213 (7.2)
50-59
4414 (18.8)
500 (16.9)
60-69
6385 (27.2)
663 (22.4)
70-79
6512 (27.8)
853 (28.8)
80þ
3555 (15.2)
22 (0.7)
635 (21.4)
8325 (35.5)
1204 (40.6)
15,143 (64.5)
1760 (59.4)
16,954 (72.2)
1994 (67.3)
Black
3660 (15.6)
573 (19.3)
Other
2854 (12.2)
397 (13.4)
0
15,392 (65.6)
1936 (65.3)
1
5981 (25.5)
761 (25.7)
2þ
2095 (8.9)
267 (9.0)
736 (3.1)
137 (4.6)
<0.0001
Sex
Male
<0.0001
Race White
Charlson Comorbidity Index
0.9582
<0.0001
Insurance status Not insured Private
8497 (36.2)
909 (30.7)
Medicaid
1624 (6.9)
217 (7.3)
Medicare
12,059 (51.4)
1614 (54.5)
Other government
213 (0.9)
22 (0.7)
Not reported
339 (1.4)
65 (2.2)
492 (21.0)
716 (24.2)
2nd
6155 (26.2)
820 (27.7)
3rd
7176 (30.6)
903 (30.5)
4thehighest
4848 (20.7)
488 (16.5)
Unreported
361 (1.5)
37 (1.3)
<0.0001
High school diploma quartile 1stelowest
<0.0001
Facility program type Community
1479 (6.3)
342 (11.5)
Comprehensive community
8087 (34.5)
1324 (44.7)
Academic/research
11,170 (47.6)
938 (31.7)
Integrated network
2732 (11.6)
360 (12.2) <0.0001
Facility gastric surgery volume High volume
9687 (41.3)
762 (25.7)
Low volume
13,781 (58.7)
2202 (74.3)
Tumor size (mm)
40 (24, 60)
45 (30, 67.5)
<0.0001 <0.0001
TNM pathologic T stage 0/X
1312 (5.6)
152 (5.1)
1
2597 (11.1)
287 (9.7)
2
2642 (11.3)
240 (8.1)
3
8003 (34.1)
1056 (35.6)
4
316 (1.4)
69 (2.3)
8598 (36.6)
1160 (39.1)
Not reported
<0.0001
Age group
Female
P value
(continued)
622
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 2 0 ( 2 4 5 ) 6 1 9 e6 2 8
Table 1 e (continued ) Variable
Elective
Urgent
n ¼ 23,468
n ¼ 2964 <0.0001
TNM pathologic N stage 0/X
11,143 (47.5)
1255 (42.3)
1
6628 (28.2)
887 (29.9)
p2
3596 (15.3)
516 (17.4)
p3
658 (2.8)
71 (2.4)
p3A
983 (4.2)
163 (5.5)
P3B
460 (2.0)
72 (2.4)
8422 (35.9)
880 (29.7)
<0.0001
TNM pathologic stage group I II
6491 (27.7)
800 (27.0)
III
8555 (36.5)
1284 (43.3)
Tumor grade
0.0002
Unreported Poorly or undifferentiated
973 (4.2)
110 (3.7)
14,360 (61.2)
1712 (57.8)
8135 (34.7)
1142 (38.5)
Body of stomach
1897 (8.1)
209 (7.1)
Cardia, NOS
7873 (33.6)
429 (14.5)
Well to moderately differentiated
<0.0001
Primary site
Fundus
73 (3.1)
95 (3.2)
Gastric antrum
5425 (23.1)
910 (30.7)
Greater curvature
1035 (4.4)
124 (4.2)
Lesser curvature
2378 (10.1)
284 (9.6)
Overlapping lesion
1507 (6.4)
235 (7.9)
688 (2.9)
203 (6.9)
1932 (8.2)
475 (16.0)
Pylorus Stomach, NOS
P value
Chemotherapy initial timing Neoadjuvant chemotherapy
5717 (24.4)
72 (2.4)
<0.0001
Adjuvant chemotherapy
6251 (26.6)
1061 (35.8)
<0.0001
Multivariate logistic regression: factors associated with urgent gastric surgery Table 3 shows the multivariate analysis of factors associated to urgent surgery. The youngest and oldest patient groups showed increased likelihood of urgent surgery compared to
patients aged 50-59 y (OR 1.66, 95% CI [1.02, 2.69], OR 1.27 95% CI [1.09, 1.47], respectively, P < 0.0001). The uninsured were significantly more likely to have urgent surgery compared to those with private insurance, Medicaid, or Medicare (OR 0.74, 95% CI [0.60, 0.90], OR 0.77, 95% CI [0.61, 0.97], OR 0.72, 95% CI [0.59, 0.89], respectively, P ¼ 0.0064). Compared to academic/
Table 2 e Unadjusted perioperative outcomes. Outcome
Elective
Urgent
n ¼ 23,468
n ¼ 2964
30-d mortality
817 (3.5)
171 (5.8)
<0.0001
90-d mortality
1687 (7.2)
316 (10.7)
<0.0001
30-d unplanned readmission
1639 (7.0)
211 (7.1)
Positive surgical margin
2598 (11.1)
532 (18.0)
<0.0001
12,906 (55.0)
1778 (60.0)
<0.0001
Positive node(s) Lymph node yield Quality lymph node dissection (15þ nodes) Surgical inpatient stay, days
15 (9, 22) 12,103 (51.6) 9 (7, 13)
12 (7, 19) 1198 (40.4) 9 (6, 13)
P value
0.7864
<0.0001 <0.0001 0.699
623
fi s h e r e t a l u r g e n t g a s t r i c c a n c e r s u r g e r y o u t c o m e s
Table 3 e Factors associated with urgent gastric surgery. Variable
Odds ratio
95% confidence interval Lower
Upper <0.0001
Age group 18-29
1.658
1.023
2.687
30-39
1.118
0.860
1.454
40-49
0.861
0.724
1.023
50-59
Ref
Ref
Ref
60-69
0.923
0.809
1.052
70-79
1.065
0.925
1.226
80þ
1.265
1.086
1.473
Ref
Ref
Ref
Private
0.735
0.601
0.899
Medicaid
0.768
0.606
0.972
Medicare
0.722
0.586
0.890
Other government
0.644
0.396
1.049
Not reported
1.047
0.751
1.460
Insurance status Not insured
0.0064
High school diploma quartile
0.055
1stelowest
1.223
1.078
1.387
2nd
1.205
1.067
1.360
3rd
1.201
1.067
1.353
4thehighest
Ref
Ref
Ref
Unreported
0.908
0.635
1.298
Community
1.875
1.615
2.177
Comprehensive community
1.676
<0.0001
Facility program type
1.512
1.363
Academic/research
Ref
Ref
Ref
Integrated network
1.355
1.184
1.551
High volume
Ref
Ref
Ref
Low volume
1.410
1.272
1.562
<0.0001
Facility gastric surgery volume
TNM pathologic T stage 0/X
0.0257 Ref
Ref
Ref
1
1.163
0.925
1.462
2
0.878
0.699
1.104
3
1.023
0.841
1.244
4
1.188
0.858
1.646
Not reported
1.108
0.914
1.344
Ref
Ref
Ref
1
0.903
0.798
1.022
p2
0.762
0.643
0.902
p3
0.839
0.629
1.119
p3A
0.719
0.574
0.901
P3B
0.655
0.488
0.879
I
Ref
Ref
Ref
II
1.277
1.112
1.467
III
1.768
1.47
2.122
0.675
0.575
0.793
TNM pathologic N stage 0/X
0.0108
<0.0001
TNM pathologic stage group
<0.0001
Primary site Body of stomach
P value
(continued)
624
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 2 0 ( 2 4 5 ) 6 1 9 e6 2 8
Table 3 e (continued ) Variable
Odds ratio
95% confidence interval Lower
Upper
Cardia, NOS
0.370
0.326
0.420
Fundus
0.775
0.617
0.972
Ref
Ref
Ref
Greater curvature
0.713
0.583
0.872
Lesser curvature
0.722
0.625
0.833
Overlapping lesion
0.923
0.789
1.080
Pylorus
1.672
1.404
1.990
Stomach, NOS
1.454
1.283
1.646
Gastric antrum
P value
OR > 1 indicates increased likelihood of urgent surgery.
research cancer program types, community, comprehensive, and integrated network facilities showed increased likelihood for urgent gastric surgery (OR 1.88, 95% CI [1.62, 2.18], OR 1.51, 95% CI [1.36, 1.68], OR 1.36, 95% CI [1.18, 1.55], respectively, P < 0.0001). Low-volume facilities were also more likely to perform urgent surgeries (OR 1.41, 95% CI [1.27, 1.56], P < 0.0001). Compared to stage I, patients were increasingly more likely to undergo urgent surgery with stage II and stage III tumors (OR 1.28, 95% CI [1.11, 1.47], OR 1.77, 95% CI [1.47, 2.12], respectively, P < 0.0001). The primary site of the gastric tumor was also found to be associated with the need for urgent surgery (P < 0.0001).
Kaplan-Meier curve and Cox proportional hazard model: overall survival Figure 1 shows the Kaplan-Meier curve comparing overall survival between cohorts (P < 0.0001). Median survival for elective surgeries was 47.7 mo versus 30.9 mo for urgent operations. Table 4displays the Cox proportional hazard model for overall survival. After adjustment for confounders, urgent gastric surgeries remained at increased risk of all-cause mortality (HR 1.21, 95% CI [1.15, 1.27], P < 0.0001). Increasing age at diagnosis continued to be a factor in all-cause mortality (P < 0.0001), as did increased comorbidities (P < 0.0001), and males were at more risk than females (HR 1.07, 95% CI [1.03, 1.11], P ¼ 0.0002). Uninsured patients showed no difference in
survival from patients with private insurance or Medicaid (HR 0.94, 95% CI [0.85, 1.04], HR 1.06, 95% CI [0.94, 1.19], respectively); however, patients with Medicare were at increased mortality risk (HR 1.13, 95% [1.02, 1.26], P < 0.0001). Compared to patients from the most educated areas, lower education was associated with worse survival (P < 0.0001). Definitive treatment at low-volume centers had increased risk of mortality compared to high-volume facilities (HR 1.23, 95% CI [1.18, 1.28], P < 0.00001). As pathologic T stage increased, overall survival worsened (P < 0.0001), a pattern followed by increasing N stages (P < 0.0001), and by pathological stage group (P < 0.0001). The primary site of the gastric tumor was also found to be associated with significant differences in survival (P < 0.0001).
Multivariable analysis: urgent gastric surgery outcomes Table 5 shows the effect of urgent gastric surgery on selected outcomes. The rate of quality dissection is decreased (OR 0.70, 95% CI [0.64, 0.76], P < 0.001). The rate of positive surgical margin, 30-d mortality, and 90-d mortality are also increased (OR 1.48, 95% CI [1.32, 1.65], OR 1.38, 95% CI [1.16, 1.65], and OR 1.30, 95% CI [1.14, 1.49], respectively, all P < 0.001). There was no significant change in unplanned readmission or lymph node positivity.
Yearly trend Looking at the trend of data over the study period in Figure 2, the rate of urgent surgery has decreased slightly each year, from 313 cases in 2004 to 238 cases in 2014. The rate of elective surgery has increased by approximately 27% during that same period, from 1941 cases in 2004 to 2473 cases in 2014 (P < 0.0001). Urgent surgeries made up 13.9% of total gastric cancer resections in 2004 and steadily dropped to 8.8% of surgeries in 2014.
Discussion
Fig. 1 e Kaplan-Meier curve comparing urgent surgery versus elective surgery. (Color version of figure is available online.)
Obstruction, perforation, and hemorrhage are well-known complications of gastric cancer,18 leading to unstable patients who may require urgent surgical intervention. However, the data are quite limited regarding outcomes for these
625
fi s h e r e t a l u r g e n t g a s t r i c c a n c e r s u r g e r y o u t c o m e s
Table 4 e Cox proportional hazard model: overall survival. Variable
Urgent (ref: elective)
Hazard ratio
1.207
95% confidence interval Lower
Upper
1.149
1.269
18-29
0.797
0.582
1.090
30-39
0.832
0.732
0.946
40-49
0.892
0.827
0.962
50-59
Ref
Ref
Ref
60-69
1.102
1.043
1.165
70-79
1.405
1.322
1.492
80þ
2.185
2.046
2.334
Ref
Ref
Ref
1.071
1.033
1.109
White
Ref
Ref
Ref
Black
1.024
0.976
1.074
Other
0.718
0.676
0.761
Ref
Ref
Ref
Sex
Male
0.0002
<0.0001
Race
<0.0001
Charlson Comorbidity Index 0 1
1.078
1.038
1.120
2þ
1.367
1.294
1.444
Ref
Ref
Ref
Private
0.936
0.845
1.037
Medicaid
1.061
0.943
1.193
Medicare
1.131
1.018
1.256
<0.0001
Insurance status Not insured
Other government
1.065
0.869
1.305
Not reported
1.097
0.932
1.292
1stelowest
1.090
1.033
1.150
2nd
1.172
1.115
1.231
3rd
1.167
<0.0001
High school diploma quartile
1.113
1.061
4thehighest
Ref
Ref
Ref
Unreported
2.054
1.824
2.313 <0.0001
Facility program type Community
1.151
1.075
1.233
Comprehensive community
1.104
1.058
1.152
Academic/research
Ref
Ref
Ref
Integrated network
1.093
1.034
1.155 <0.0001
Facility gastric surgery volume High volume
Ref
Ref
Ref
Low volume
1.226
1.177
1.277
Ref
Ref
Ref
1
2.934
2.59
3.324
2
2.821
2.496
3.187
3
3.293
2.945
3.681
4
4.217
3.591
4.951
Not reported
3.348
2.993
3.745
<0.0001
TNM pathologic T stage 0/X
<0.0001 <0.0001
Age group
Female
P value
(continued)
626
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 2 0 ( 2 4 5 ) 6 1 9 e6 2 8
Table 4 e (continued ) Variable
Hazard ratio
95% confidence interval Lower
Upper <0.0001
TNM pathologic N stage 0/X
Ref
Ref
Ref
p1
1.344
1.272
1.419
p2
1.516
1.409
1.630
p3
1.729
1.558
1.919
p3A
1.770
1.603
1.954
p3B
2.451
2.171
2.768
I
Ref
Ref
Ref
II
1.536
1.44
1.638
III
2.353
2.161
2.561
<0.0001
TNM pathologic stage group
<0.0001
Primary site Body of stomach
0.943
0.879
Cardia, NOS
1.362
1.295
1.432
Fundus
1.218
1.107
1.340
Gastric antrum
P value
1.011
Ref
Ref
Ref
Greater curvature
0.941
0.863
1.026
Lesser curvature
0.901
0.845
0.961
Overlapping lesion
1.146
1.068
1.229
Pylorus
1.061
0.967
1.164
Stomach, NOS
1.141
1.072
1.215
<0.0001
Hazard ratio > 1 indicates increased risk of all-cause mortality.
patients. In 1980, Hahnloser et al. published a small case series of 13 patients who underwent either 1- or 2-stage procedure for gastric perforation, with more stable patients undergoing immediate gastric resection with reasonable survival.19 Additional studies, primarily performed at single institutions with limited numbers of patients, recommend similar treatment, with the emphasis on stabilizing the patient, with the oncologic treatment addressed either once the patient was stable enough to undergo more extensive treatment or once the correct diagnosis had been achieved.20 Survival was based on pathology of the tumor, and some studies found that patients undergoing emergency gastrectomy had satisfactory results, with stage I and II tumors having similar long-term survival when compared to historic norms.21 Others found that radical gastrectomy was not advised at the index surgery, recommending a second procedure (with lymphadenectomy) once the patient had recovered and had been staged appropriately.22 Kasakura et al. found that patients with early-stage carcinoma who underwent emergency R0 resections had minimal complications,23 whereas Blackshaw et al. showed that patients undergoing emergency surgery for gastric cancer have significantly decreased median survival (6 versus 12 mo).24 Few studies have had enough patients to draw strong conclusions, and even fewer have long-term data on oncologic outcomes for urgent gastric resections. In addition, given the improvement in outcomes seen with adoption of neoadjuvant chemotherapy,7 there is concern that patients undergoing urgent surgery likely have worse overall mortality and morbidity. This study validates that concern,
demonstrating that patients undergoing urgent surgery for gastric cancer have significantly decreased rates of quality resection, with decreased lymph node yield and increased positive margin, as well as increased 30-d, 90-d, and overall mortality when compared patients undergoing elective surgery. The decision to proceed with definitive resection is dependent on preoperative and intraoperative patient factors, including clinical stability and overall level of fitness. As might be expected, larger and more locally advanced tumors were associated with an increased rate of urgent surgery and had significantly higher mortality, even when
Fig. 2 e Yearly trends in urgent and elective gastric resection from 2004 to 2014. (Color version of figure is available online.)
627
fi s h e r e t a l u r g e n t g a s t r i c c a n c e r s u r g e r y o u t c o m e s
Table 5 e Effect of urgent gastric surgery on outcomes. Variable
Odds ratio
95% confidence interval Lower
Upper
P value
Quality lymph node dissection (15 þ nodes)
0.693
0.636
0.755
<0.0001
30-d unplanned readmission
1.017
0.875
1.183
0.8223 <0.0001
Positive surgical margin
1.478
1.320
1.654
30-d mortality
1.379
1.156
1.646
0.0004
90-d mortality
1.298
1.135
1.485
<0.0001
Regional node(s) positive
0.844
0.614
1.161
0.2983
Odds ratio > 1 indicates increased likelihood of outcome compared to elective.
matched stage-for-stage with those undergoing elective surgery. Data from urgent colorectal surgery show similar data, with Ming-Gao et al. showing higher mortality for urgent colon resections,25 and Kim et al. demonstrating higher complication rates in emergent colectomies for obstructing colon cancer.26 More advanced disease may necessitate a larger operation in patients who are already unstable, with the increased mortality likely related to longer operative time, friable tissues, and operative blood loss in patients with already diminished physiologic reserves. One additional area of concern is that there are low rates of adjuvant chemotherapy in the urgent group, with low overall rates of chemotherapy in both groups (only 51% receive neoadjuvant or adjuvant chemotherapy in the elective group compared with 38% in the urgent group). This may represent failure of adherence to recommended guidelines for the treatment of gastric cancer, or it may be that these patients failed to have appropriate follow-up with medical oncologists for initiation of chemotherapy. It is also possible that some patients underwent radiation only (which is not captured in the NCDB) or had adjuvant thermotherapy at a facility which was not recorded in the NCDB. The difference in chemotherapy rates was not significant in multivariate analysis but is still a concerning finding worth mentioning, and further investigation is necessary to understand the cause for such an observed low incidence of adjuvant therapy.
Limitations The NCDB was specifically chosen for this study because it contains long-term patient data with tumor and treatment data, which is not available in many other national databases. However, it has limitations. It is a retrospective database and variables are limited only to previously collected data. Urgent surgery is not specifically coded in the database, and this definition is central to our study question. We followed the example of Xu et al.27 and defined “urgent surgery” as definitive surgical intervention within 3 d of the diagnosis of gastric adenocarcinoma. While this duration may seem arbitrary, the definition was chosen to include both critically ill patients taken emergently to the operating room and less critical patients who may need resuscitation or a brief metastatic workup. Thus, patients undergoing emergent surgery for perforation are included in the data, as are patients presenting with obstruction or hemorrhage who may not be taken immediately to the operating room. It is reasonable to surmise
that if these patients were diagnosed with gastric adenocarcinoma within 3 d of definitive surgery, it was either a biopsy during the workup for GI hemorrhage, or intraoperatively when a specimen was sent for frozen or final surgical pathology. Given the brief interval included in our definition, it is also likely that many of the urgent surgery cohort presented with perforation, as obstruction and hemorrhage may be temporized with other therapies, including advanced endoscopy, radiation, and surgical decompression with gastrostomy or jejunostomy, leaving definitive treatment outside the definition for urgent surgery. However, it is impossible to determine the specific cause for any patient’s urgent surgery within the constraints of this database. While not without limitations, this definition provides a useful tool for investigating outcomes in this patient population. In addition, the NCDB provides hospital data but has no information about the surgeons performing each surgery, so it is impossible to see variation between surgeons at a single facility, or to know if the surgeon performing the resection was a board-certified surgical oncologist. Patients undergoing urgent surgical resection had a decreased rate of adequate lymph node yield and an increased positive surgical margin, which suggests that less-experienced surgeons may be performing many of these emergent operations; however, that level of surgeon information is not available in the database. Finally, the NCDB provides only overall survival, not diseasespecific survival, so it is not possible to demonstrate cause of death for these patients. Despite these limitations, this study demonstrates that urgent surgery has significantly worse outcomes in lymph node yield, positive surgical margins, 30- and 90-d mortality, and overall survival in patients undergoing urgent surgery for gastric adenocarcinoma.
Conclusions Urgent surgery for gastric cancer is associated with significantly worse surgical and oncologic outcomes. Stable patients who require urgent surgical resection for gastric masses may benefit from referral to a high-volume center for resection by an experienced surgeon. If that is not possible, consultation with a surgical oncologist may be beneficial for the surgeon who performs fewer gastric resections. For unstable patients requiring emergent surgical treatment, a damage-control procedure should be considered an individual based according to preoperative and intraoperative factors. Consultation
628
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 2 0 ( 2 4 5 ) 6 1 9 e6 2 8
with or transfer to an experienced surgical oncologist for resection may lead to improved surgical margins and lymph node yield, leading to improved survival. All patients undergoing urgent resection for gastric cancer should also be referred to a surgical oncologist and medical oncologist to ensure that the patient receives appropriate adjuvant therapy and cancer surveillance.
Acknowledgment Authors’ contributions: B.W.F. was the primary researcher and manuscript author. He completed a majority of the writing for the manuscript and also performed the literature review. M.F. performed the statistical analysis and assisted in writing the methods and results section. K.Y. assisted in research mechanics and methodology, in addition to editing the final manuscript. M.S. assisted in refining the research question and assisted in editing the final manuscript. J.B. assisted in defining and refining the research question and assisted in editing the final manuscript. T.A. was the primary faculty research mentor and assisted in refining the research question, in addition to assisting with research methodology, data interpretation, and editing the final manuscript.
Disclosure The authors have no conflicts of interest to report and there was no extra-institutional funding provided for this project.
references
1. Soerjomataram I, Lortet-Tieulent J, Parkin DM, et al. Global burden of cancer in 2008: a systematic analysis of disabilityadjusted life-years in 12 world regions. Lancet. 2012;380:1840e1850. 2. Casamayor M, Morlock R, Maeda H, Ajani J. Targeted literature review of the global burden of gastric cancer. Ecancermedicalscience. 2018;12:883. 3. GBD 2017 Causes of Death Collaborators. Global, regional, and national age-sex-specific mortality for 282 causes of death in 195 countries and territories, 1980-2017: a systematic analysis for the global burden of disease study 2017. Lancet. 2018;392:1736e1788. 4. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2019. CA Cancer J Clin. 2019;69:7e34. 5. Amin MB, Greene FL, Edge SB, et al. The eighth edition AJCC cancer staging manual: Continuing to build a bridge from a population-based to a more “personalized” approach to cancer staging. CA Cancer J Clin. 2017;67:93e99. 6. Macdonald JS, Smalley SR, Benedetti J, et al. Chemoradiotherapy after surgery compared with surgery alone for adenocarcinoma of the stomach or gastroesophageal junction. N Engl J Med. 2001;345:725e730. 7. Cunningham D, Allum WH, Stenning SP, et al. Perioperative chemotherapy versus surgery alone for resectable gastroesophageal cancer. N Engl J Med. 2006;355:11e20.
8. Mihmanli M, Ilhan E, Idiz UO, Alemdar A, Demir U. Recent developments and innovations in gastric cancer. World J Gastroenterol. 2016;22:4307e4320. 9. Chen L, Ma LG, Xi HQ. [The conversion therapy for unresectable gastric cancer]. Zhonghua Wai Ke Za Zhi. 2016;54:169e171. 10. Oh D, Bang Y. Adjuvant and neoadjuvant therapy for gastric cancer. Curr Treat Options Oncol. 2013;14(3):311e320. 11. Tegels JJ, De Maat MF, Hulsewe´ KW, Hoofwijk AG, Stoot JH. Improving the outcomes in gastric cancer surgery. World J Gastroenterol. 2014;20:13692e13704. 12. Smyth EC, Wotherspoon A, Peckitt C, et al. Mismatch repair deficiency, microsatellite instability, and survival: an exploratory analysis of the medical research council adjuvant gastric infusional chemotherapy (MAGIC) trial. JAMA Oncol. 2017;3:1197e1203. 13. Song Z, Wu Y, Yang J, Yang D, Fang X. Progress in the treatment of advanced gastric cancer. Tumour Biol. 2017;39, 1010428317714626. 14. Choi AH, Kim J, Chao J. Perioperative chemotherapy for resectable gastric cancer: MAGIC and beyond. World J Gastroenterol. 2015;21:7343e7348. 15. Aoyama T, Yoshikawa T. Adjuvant therapy for locally advanced gastric cancer. Surg Today. 2017;47:1295e1302. 16. Ignjatovic N, Stojanov D, Djordjevic M, Ignjatovic J, Benedeto Stojanov D, Milojkovic B. Perforation of gastric cancer - what should the surgeon do? Bosn J Basic Med Sci. 2016;16:222e226. 17. Smith DL, Elting LS, Learn PA, Raut CP, Mansfield PF. Factors influencing the volume-outcome relationship in gastrectomies: a population-based study. Ann Surg Oncol. 2007;14:1846e1852. 18. Itano S. Early gastric cancer and its complications: bleeding, perforation and pyloric stenosis. Acta Med Okayama. 1983;37:431e440. 19. Hahnloser P, Bruttin JM, Cavin R, Petropoulos P. [Perforation of cancer or stomach ulcer: a 1- or 2-stage operation]. Helv Chir Acta. 1980;47:565e574. 20. Fox JG, Hunt PS. Management of acute bleeding gastric malignancy. Aust N Z J Surg. 1993;63:462e465. 21. Adachi Y, Mori M, Maehara Y, Matsumata T, Okudaira Y, Sugimachi K. Surgical results of perforated gastric carcinoma: an analysis of 155 Japanese patients. Am J Gastroenterol. 1997;92:516e518. 22. Lehnert T, Buhl K, Dueck M, Hinz U, Herfarth C. Two-stage radical gastrectomy for perforated gastric cancer. Eur J Surg Oncol. 2000;26:780e784. 23. Kasakura Y, Ajani JA, Mochizuki F, Morishita Y, Fujii M, Takayama T. Outcomes after emergency surgery for gastric perforation or severe bleeding in patients with gastric cancer. J Surg Oncol. 2002;80:181e185. 24. Blackshaw GR, Stephens MR, Lewis WG, et al. Prognostic significance of acute presentation with emergency complications of gastric cancer. Gastric Cancer. 2004;7:91e96. 25. Ming-Gao G, Jian-Zhong D, Yu W, You-Ben F, Xin-Yu H. Colorectal cancer treatment in octogenarians: elective or emergency surgery? World J Surg Oncol. 2014;12:386. 26. Kim J, Mittal R, Konyalian V, King J, Stamos MJ, Kumar RR. Outcome analysis of patients undergoing colorectal resection for emergent and elective indications. Am Surg. 2007;73:991e993. 27. Xu Z, Becerra AZ, Aquina CT, et al. Emergent colectomy is independently associated with decreased long-term overall survival in colon cancer patients. J Gastrointest Surg. 2017;21:543e553.