Globally, the population is aging; the World Health Organization predicts that by 2050 around 2 billion people in the world will be ≥60 years of age, and 400 million will be ≥80 years of age.¹ The increasing geriatric population and introduction of mass screening programs, has led to a growing number of early gastric cancer (EGC) cases being diagnosed in the very elderly population (those aged >85 years).

The treatment of EGC is well-established. When a lesion meets the endoscopic treatment criteria, endoscopic resection is the first-line treatment modality and achieves comparable results as surgery concerning patient survival.² However, the management of very elderly patients with EGC is challenging in comparison to elderly or non-elderly patients with EGC owing to age-specific concerns, including a higher incidence of comorbid conditions and a lower life expectancy.^3,4 Reports exist on the long-term outcomes of endoscopic resection in the elderly, but research specific to the very elderly are lacking.^5-9 Because the number of elderly patients with EGC has also increased, and will continue to do so in the future, there is a significant need for evidence-based strategies to determine the most appropriate course of treatment for EGC in the very elderly population.

Therefore, this study aimed to evaluate the long- and short-term outcomes of endoscopic resection for EGC for very elderly patients compared to non-elderly and elderly patients, using data from the National Health Insurance Service (NHIS). We also examined the prognostic factors of clinical outcomes.

1. Patients

As a single insurer, the NHIS covers the entire Korean population. In December 2015, the Korean National Health Insurance Big Data Base was developed, based on the treatment data for claims between the NHIS and healthcare providers in Korea.¹⁰ The Korean National Health Insurance Big Data Base includes information on patient demographics, reimbursement for medical services, the International Classification of Diseases 10th Revision codes, medical history, and mortality,¹¹ which are then linked to the Korean National Statistical Office, who allow access for research purposes.

Based on the Korean National Health Insurance Big Data Base, we constructed a custom database which included patients with gastric cancer in Korea from 2006 to 2017. Patients with gastric cancer were screened by the main or sub-disease code, C16, according to the International Classification of Diseases 10th Revision.¹² We assessed patient characteristics including age, sex, and comorbidities. The exclusion criteria were as follows: (1) those who had an additional cancer code besides C16 or underwent gastrectomy 5 years before enrollment of enrollment (n=8,702); (2) lost to follow-up (n=1,576); and (3) those who did not undergo endoscopic submucosal dissection (ESD) or endoscopic mucosal resection (EMR) (n=55,813). The included patients were divided into three age groups based on World Health Organization and geriatric research guidelines:^13,14 very elderly (≥85 years), elderly (65 to 84 years), and non-elderly (≤64 years) (Fig. 1).

Figure 1. Study flowchart.

In Korea, endoscopic resection for EGC is considered for tumors that fulfill the criteria according to the Korean Practice Guideline for Gastric Cancer 2018.¹⁵

The study protocol was in accordance with the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the Institutional Review Board of the Samsung Medical Center (IRB number: 2019-07-128). The requirement for informed consent was waived due to the retrospective study design.

2. Outcomes

1) Long-term outcomes

All-cause mortality was obtained from the NHIS database. We compared overall survival (OS) and cancer-specific survival (CSS) rates between the three groups. The relationships among OS, CSS, and clinical factors were evaluated. We compared the risk of mortality in the three groups in relation to that in the control group matched by age and sex. OS was defined as the interval from the date of gastric cancer diagnosis to any cause of death. CSS was defined as the interval from the date of the primary diagnosis to the date of gastric cancer-specific death.

2) Short-term outcomes

Short-term outcomes were evaluated in terms of the readmission within 3 months after ESD/EMR (including admission for gastrectomy) and mortality within 3 months after ESD/EMR. The relationship between short-term outcomes and clinical factors was also evaluated.

3. Statistical analysis

Continuous values are presented as mean±standard deviation, while categorical variables are shown as number with proportions. The differences between groups were analyzed using the t-test for continuous variables and the chi-square test for categorical variables. OS and CSS were estimated using the Kaplan-Meier method. Survival rates were compared using the log-rank test. Cox proportional hazards regression was performed to identify hazard ratios (HRs) and 95% confidence intervals (CIs). Differences were considered significant at p-value of <0.05. Statistical analyses were performed using SAS version 9.4.1 (Statistical Analysis Software, SAS Institute Inc, Cary, NC, USA) and the statistical program R (version 4.0; R Foundation for Statistical Computing, Vienna, Austria).

1. Patient characteristics

Of the 74,517 patients diagnosed with EGC between 2006 and 2017, 66,091 patients were excluded and 8,426 patients who underwent endoscopic resection for EGC were eligible for the study. Among the 8,426 patients, 8,350 patients underwent ESD and only 76 patients underwent EMR. Baseline characteristics of the study population are presented in Table 1. The mean patient age was 65.5 years (standard deviation, 10.4), and among them, 118 (1.4%) were classified as very elderly, 4,583 (54.4%) as elderly, and 3,725 (44.2%) as non-elderly. A total of 5,998 patients (71.2%) were men and 2,428 (28.8%) were women. The elderly and non-elderly groups were predominantly males (75.8% and 68.0%, respectively); however, the very elderly group had a slight female predominance (51.7%). The severity of all comorbidities increased significantly with age, with the exception of liver disease, acquired immunodeficiency syndrome/human immunodeficiency virus infection and renal disease.

Table 1 . Baseline Characteristics.

Characteristic	Total	Age group			p-value
		≤64 yr	65–84 yr	≥85 yr
No. of patients	8,426	3,725	4,583	118
Age, yr	65.5±10.4	55.9±6.5	72.8±5.0	87.0±1.9	<0.001
Sex					<0.001
Male	5,998 (71.2)	2,825 (75.8)	3,116 (68.0)	57 (48.3)
Female	2,428 (28.8)	900 (24.2)	1,467 (32.0)	61 (51.7)
Comorbidities
Myocardial infarction	142 (1.7)	40 (1.1)	98 (2.1)	4 (3.4)	<0.001
Congestive heart failure	584 (6.9)	118 (3.2)	449 (9.8)	17 (14.4)	<0.001
Peripheral vascular disease	1,650 (19.6)	453 (12.2)	1,164 (25.4)	33 (28.0)	<0.001
Cerebrovascular disease	1,512 (17.9)	344 (9.2)	1,129 (24.6)	39 (33.1)	<0.001
Dementia	233 (2.8)	19 (0.5)	199 (4.3)	15 (12.7)	<0.001
Chronic pulmonary disease	4,126 (49.0)	1,450 (38.9)	2,599 (56.7)	77 (65.3)	<0.001
Connective tissue disease	547 (6.5)	183 (4.9)	346 (7.5)	18 (15.3)	<0.001
Diabetes	2,657 (31.5)	905 (24.3)	1,700 (37.1)	52 (44.1)	<0.001
Paraplegia/hemiplegia	108 (1.3)	26 (0.7)	80 (1.7)	2 (1.7)	<0.001
Liver disease	2,516 (29.9)	1,111 (29.8)	1,374 (30.0)	31 (26.3)	0.684
AIDS/HIV	3 (0.0)	1 (0.0)	2 (0.0)	0	0.902
Renal disease	204 (2.4)	47 (1.3)	154 (3.4)	3 (2.5)	<0.001

Data are presented as mean±SD or number (%)..

AIDS, acquired immunodeficiency syndrome; HIV, human immunodeficiency virus..

2. Comparison of long-term outcomes

The median follow-up periods after endoscopic resection for the very elderly, elderly, and non-elderly groups were 1,200 days (interquartile range, 950 to 1,642 days), 1,379 days (interquartile range, 1,026 to 1,777 days), and 1,475 days (interquartile range, 1,091 to 1,867 days), respectively. The OS rates of the very elderly group were significantly lower than those of the elderly and non-elderly group (1-, 5-, and 10-year OS rates: very elderly patients: 96.6%, 67.8%, and 62.7%; elderly patients: 99.1%, 88.2%, and 85.3%; and non-elderly patients: 99.8%, 97.4%, and 96.5%, respectively; p<0.001) (Fig. 2A). The CSS rate was also significantly lower in the very elderly group than in the other groups (1-, 5-, and 10- year CSS rates: very elderly patients: 98.3%, 90.7%, and 89.8%; elderly patients: 99.7%, 98.5%, 98.3%; and non-elderly patients: 100.0%, 99.7%, and 99.6%, respectively; p<0.001) (Fig. 2B). However, a significant decrease in the risk of mortality was observed among all age groups (HR=0.23, 0.30, and 0.45, for the very elderly, elderly, and non-elderly groups, respectively; all p<0.001) (Table 2), indicating that patients undergoing endoscopic resection for EGC had a lower mortality rate than the control group, regardless of age.

Table 2 . Risk of Mortality in the Three Age Groups Compared to General Population Matched Age and Sex.

Age group	Endoscopic resection				Control group		p-value
	No. of patients	Person-years	No. of cases		No. of cases	HR (95% CI)
≤64 yr	3,725	15,103	85		1,217	0.45 (0.36–0.55)	<0.001
65–84 yr	4,583	17,693	427		9,633	0.30 (0.27–0.33)	<0.001
≥85 yr	118	404	30		1,008	0.23 (0.15–0.32)	<0.001

HR, hazard ratio; CI, confidence interval..

Figure 2. Overall survival (A) and cancer-specific survival (B) of the three age groups.

The risk factors associated with OS were congestive heart failure (HR, 1.92; 95% CI, 1.52 to 2.44; p<0.001), dementia (HR, 1.94; 95% CI, 1.40 to 2.70; p<0.001), paraplegia/hemiplegia (HR, 2.26; 95% CI, 1.46 to 3.50; p<0.001), renal disease (HR, 2.42; 95% CI, 1.72 to 3.40; p<0.001), and increasing Charlson comorbidity index (CCI) score (HR, 1.80; 95% CI, 1.45 to 3.15; p<0.001); however, the only risk factor associated with CSS was congestive heart failure (HR, 2.86; 95% CI, 1.51 to 5.41; p=0.001) (Table 3).

Table 3 . Risk Factors for Long-term Outcome.

Covariate	Overall survival			Cancer-specific survival
	Adjusted HR (95% CI)	p-value		Adjusted HR (95% CI)	p-value
Sex
Male	Reference			Reference
Female	0.60 (0.49–0.74)	<0.001		1.02 (0.59–1.77)	0.950
Comorbidities
Myocardial infarction	1.03 (0.62–1.71)	0.908		0.00 (0.00–Inf)	0.994
Congestive heart failure	1.92 (1.52–2.44)	<0.001		2.86 (1.51–5.41)	0.001
Peripheral vascular disease	0.98 (0.80–1.20)	0.840		1.32 (0.74–2.33)	0.343
Cerebrovascular disease	1.13 (0.92–1.38)	0.260		0.59 (0.30–1.18)	0.136
Dementia	1.94 (1.40–2.70)	<0.001		1.98 (0.76–5.18)	0.163
Chronic pulmonary disease	1.02 (0.85–1.21)	0.857		0.80 (0.47–1.37)	0.416
Connective tissue disease	1.11 (0.81–1.52)	0.513		1.28 (0.54–3.03)	0.578
Diabetes	0.43 (0.05–3.53)	0.432		0.62 (0.11–3.63)	0.564
Paraplegia/hemiplegia	2.26 (1.46–3.50)	<0.001		2.66 (0.61–11.65)	0.194
Liver disease	1.10 (0.92–1.32)	0.295		0.64 (0.34–1.20)	0.167
AIDS/HIV	0.00 (0.00–Inf)	0.987		0.00 (0.00–Inf)	0.999
Renal disease	2.42 (1.72–3.40)	<0.001		2.55 (0.89–7.27)	0.080
Charlson comorbidity index score	1.80 (1.45–3.15)	<0.001		1.21 (0.45–5.78)	0.469

HR, hazard ratio; CI, confidence interval; AIDS, acquired immunodeficiency syndrome; HIV, human immunodeficiency virus; Inf, infinity..

3. Comparison of short-term outcomes

The short-term outcomes are presented in Table 4. Readmission rate within 3 months after endoscopic resection was significantly higher in the very elderly group than in the elderly and non-elderly groups (11.0%, 5.8%, and 5.3%, respectively, p=0.028). Mortality within 3 months after endoscopic resection was significantly higher in the very elderly group than in the elderly and non-elderly groups (0.8%, 0.1%, and 0%, respectively, p<0.001). The risk factors associated with the short-term outcomes are presented in Table 5. Risk factors associated with readmission within 3 months after endoscopic resection were female sex (HR, 1.22; 95% CI, 1.01 to 1.49; p=0.045), congestive heart failure (HR, 1.49; 95% CI, 1.12 to 2.00; p=0.007), dementia (HR, 1.77; 95% CI, 1.20 to 2.61; p=0.004), paraplegia/hemiplegia (HR, 3.50; 95% CI, 2.26 to 5.42; p<0.001), liver disease (HR, 1.51; 95% CI, 1.25 to 1.82; p<0.001), and acquired immunodeficiency syndrome/human immunodeficiency virus infection (HR, 31.47; 95% CI, 7.70 to 128.57; p<0.001). Increasing CCI score was associated with the risk for readmission within 3 months, but not with 3-month mortality. The only risk factor associated with a 3-month mortality rate was cerebrovascular disease (HR, 5.75: 95% CI, 1.05 to 31.38; p=0.043).

Table 4 . Short-term Outcomes.

Outcome	Age group, No. (%)			p-value
	≤64 yr (n=3,725)	65–84 yr (n=4,583)	≥85 yr (n=118)
Overall
Readmission rate within 3 mo after endoscopic resection	199 (5.3)	264 (5.8)	13 (11.0)	0.028
Mortality within 3 mo after endoscopic resection	0	5 (0.1)	1 (0.8)	<0.001
For endoscopic submucosal dissection
Readmission rate within 3 mo after endoscopic resection	187 (5.1)	261 (5.7)	13 (11.0)	0.023
Mortality within 3 mo after endoscopic resection	0	4 (0.1)	1 (0.8)	<0.001
For endoscopic mucosal resection
Readmission rate within 3 mo after endoscopic resection	2 (5.3)	3 (8.6)	0	0.019
Mortality within 3 mo after endoscopic resection	0	1 (2.9)	0	<0.004

Table 5 . Risk Factors for Short-term Outcome.

Covariate	3 mo readmission			3 mo mortality
	Adjusted HR (95% CI)	p-value		Adjusted HR (95% CI)	p-value
Sex
Male	Reference			Reference
Female	1.22 (1.01–1.49)	0.045		0.67 (0.36–1.25)	0.620
Comorbidities
Myocardial infarction	0.85 (0.44–1.66)	0.633		0.00 (0.00–Inf)	0.999
Congestive heart failure	1.49 (1.12–2.00)	0.007		1.79 (0.21–15.34)	0.594
Peripheral vascular disease	1.15 (0.92–1.42)	0.217		2.69 (0.54–13.32)	0.226
Cerebrovascular disease	1.14 (0.91–1.44)	0.252		5.75 (1.05–31.38)	0.043
Dementia	1.77 (1.20–2.61)	0.004		2.00 (0.23–17.15)	0.526
Chronic pulmonary disease	1.10 (0.91–1.33)	0.314		1.50 (0.28–8.21)	0.638
Connective tissue disease	1.21 (0.88–1.67)	0.234		0.00 (0.00–Inf)	0.998
Diabetes	0.59 (0.28–1.26)	0.176		0.00 (0.00–Inf)	1.000
Paraplegia/hemiplegia	3.50 (2.26–5.42)	<0.001		0.00 (0.00–Inf)	0.999
Liver disease	1.51 (1.25-1.82)	<0.001		0.00 (0.00–Inf)	0.996
AIDS/HIV	31.47 (7.70–128.57)	<0.001		0.00 (0.00–Inf)	1.000
Renal disease	1.16 (0.71–1.90)	0.558		0.00 (0.00–Inf)	0.999
Charlson comorbidity index score	18.57 (6.19–83.57)	<0.001		3.74 (0.35–22.15)	0.698

HR, hazard ratio; CI, confidence interval; AIDS, acquired immunodeficiency syndrome; HIV, human immunodeficiency virus; Inf, infinity..

The major finding of this study was that although the overall and gastric cancer-specific mortality rates were higher in the very elderly group than in the elderly and non-elderly groups, mortality risk was significantly lower for all three groups than in the control group. The survival of the very elderly group indicated a 67% decrease in the mortality rate after endoscopic resection for EGC. These results indicate that in very elderly patients who meet the endoscopic resection criteria, aggressive endoscopic resection for EGC is not harmful and, moreover, it may help patients achieve expected survival rates comparable to those of the control group.

It is often difficult to devise treatment strategies for elderly patients with EGC because of their comorbidities and lower life expectancy. Previous studies have reported that in patients who do not receive timely treatment, EGC progresses to advanced stage and lead to death within 4 to 5 years.^16-18 However, the Organisation for Economic Co-operation and Development reports that the current life expectancy for those ≥80 years living within Organisation for Economic Co-operation and Development countries is approximately 9 years,¹⁹ and the efficacy and safety of ESD in comorbid or elderly patients has been well-reported.^20-28 Kim et al.²¹ used CCI to divide patients into low and high risk groups and compared complication rates in various comorbid disease between the two groups. Complications (such as bleeding or perforation) did not differ significantly between the two groups. Moreover, none of the comorbid disease made a statistically significant contribution to the occurrence of a specific complication.²¹ Kato et al.²⁸ compared the short-term outcomes and incidence of complications in elderly (≥75 years of age) and non-elderly groups. The R0 resection rate and curative resection rate did not differ between the two groups. In terms of complication, the incidence of bleeding and incidence of perforation were similar in the two groups.²⁸ In two studies including very elderly patients (≥85 years of age), there were no significant differences in adverse events or in the en bloc resection or complete resection.^29,30 Therefore, aggressive endoscopic treatment of EGC in elderly patients may contribute to lengthening life expectancy. Consistent with our findings, the results from recent studies investigating endoscopic treatments for elderly patients have demonstrated comparable long-term outcomes.^6,7,29,30 Abe et al.⁶ conducted a multicenter study including 440 patients aged ≥80 years which showed that in patients undergoing ESD, the cumulative survival at 5 years after ESD was better than the expected survival for general population matched by age and sex. Chang et al.⁷ reported the long-term outcomes of patients aged ≥75 years who underwent ESD for EGC, were worse for patients with other risk factors such as history of cancer beside gastric cancer or CCI ≥3, than in those without other risk factors.

Few reports have included very elderly patients, and the long-term outcomes of these patients are not well-known. To the best of our knowledge, only two reports have been published, which included a small number of patients (n=43 and n=85).^29,30 Our study included a relatively larger number (n=118) of very elderly patients with a long-term observation period. In addition, we compared the survival rates of the very elderly group with those of the elderly and non-elderly groups, and the survival rate after endoscopic treatment was compared to that of the control group. We also identified prognostic factors for long- and short-term outcomes. The common risk factors for OS and CSS are congestive heart failure, and the risk factor for 3-month mortality was cerebrovascular disease. These results are consistent with reports that patients aged ≥85 with cardiovascular disease showed statistically significant lower OS rate than patients aged ≥85 without cardiovascular disease.²⁹ Cardiovascular and cerebrovascular diseases are major health problems in the elderly, as a result of living to an age when they are likely to suffer the sequelae of chronic atherosclerosis. One study reported the long-term prognosis after ESD for colorectal tumors in patients aged over 80 years and the most common cause of death was heart disease other than pneumonia.³¹ Because our study stratified comorbid diseases according to the CCI, heart diseases were classified into myocardial infarction and congestive heart failure. While patients who suffer a heart attack often survive their initial events and live into their 80s, many go on to die of chronic heart failure. In fact, age-adjusted mortality rate for heart failure appears to be on the rise.³² In terms of short-term outcomes, bleeding is one of the major complications in endoscopic resection.²¹ Patients with cerebrovascular diseases are likely to be on antiplatelet or anticoagulant therapy, and continuing it may increase the risk of bleeding while discontinuing it may increase the risk of ischemic stroke, contributing to higher short-term mortality.^33,34

Our study has several limitations. First, the study was retrospective and may have a potential bias. For example, the very elderly patients in this study may had a better performance status and less comorbidity compared with general elderly population, which may induce selection bias. Second, the 3-month readmission data included admission for additional surgery after non-curative ESD because we could not discriminate the cause for admission. Third, the follow-up duration among the three age groups was significantly different, likely because of the different life expectancies. Fourth, the data from gastric adenomas were not included in this study. Sometimes, large gastric adenoma are more difficult to perform endoscopic procedures than very small EGCs, and the complication rates could be higher. Fifth, the outcomes related to the procedure such as complete resection rate, recurrence rate, and the incidence of complications during procedure, could not be analyzed because this study is a national population-based study.

In conclusion, our study suggests that aggressive endoscopic resection for patients with EGC can be helpful for very elderly patients who meet the endoscopic resection criteria, with the expected post-procedure survival rate being comparable to that of the control group. Regarding comorbidities, special attention should be paid to very elderly patients with congestive heart failure or cerebrovascular disease.

No potential conflict of interest relevant to this article was reported.

Study conception and design: H.L., T.J.K., J.H.P. Data analysis and interpretation: T.J.K., J.H.P. Drafting of the manuscript: T.J.K., J.H.P. Critical revision of the manuscript for important intellectual content: S.C.C., Y.W.M., B.H.M., J.H.L., P.L.R., M.S., Y.H.C., J.J.K. Study supervision: H.L. Approval of final manuscript: all authors.