Follow-up Loss After Curative Gastrectomy for Gastric Cancer: Incidence, Contributing Factors, and Survival Impact

Abstract

Background/Aims

This study examined the incidence, causes, and survival outcomes of follow-up loss (FUL) after a gastrectomy for gastric cancer.

Methods

Patients who underwent a curative gastrectomy between January 2016 and May 2019 at a regional tertiary hospital were divided into two groups based on their follow-up (FU) adherence. Patients who maintained a regular FU throughout the five-year period were classified as the FU group, and those who failed to attend their scheduled visits for more than 12 consecutive months were grouped as the FUL group. Telephone interviews were conducted to identify the reasons for FU discontinuation and survival status. The sociodemographic and clinical variables were compared, and the independent predictors and survival outcomes were compared.

Results

Among the 435 patients, 137 (31.5%) were in the FUL group, and contact was successful in 131 patients (95.6%). The leading cause of FUL was death from non-gastric cancer causes (40.1%). Independent predictors of FUL were older age (hazard ratio [HR]=1.044, p<0.001), lower body mass index (BMI, HR=0.927, p=0.015), absence of familial support (HR=2.666, p=0.005), and total gastrectomy (HR=1.660, p=0.012). The BMI lost significance in sensitivity analysis (p=0.293). The overall survival (OS) was lower in the FUL group (p=0.0370), particularly for the stage I patients (p=0.046). The independent predictors of OS were FUL (HR=2.148, p=0.006) and pathologic stage (p<0.001).

Conclusions

FUL after a gastrectomy was associated with older age, absence of familial support, total gastrectomy, and was related to a poorer OS, particularly in stage I patients.

INTRODUCTION

Follow-up (FU) after a radical gastrectomy in gastric cancer is an essential component of postoperative management, aiming to detect early and late complications, nutritional deterioration, and recurrence.^1,2 Regarding FU, the major guidelines recommend surveillance every three to six months for five years after a gastrectomy, but they differ in their recommended surveillance protocols across regions.^1-4 A recent randomized trial compared different guideline protocols to determine the optimal FU strategy. The nature and clinical implications of follow-up loss (FUL) remain largely uninvestigated.⁵

FUL can affect early detection for recurrence, nutritional deficit, and postoperative complications, and affect the patient’s quality of life. In other malignancies, high FUL rates have been reported with demographic, socioeconomic, and behavioral factors as the key predictors.^6-8

Gastric cancer patients may present with unique clinical features after a gastrectomy and require special attention during FU. Long-term nutritional impairments can influence patient engagement in scheduled FU at dedicated centers because gastric cancer predominantly affects older patients with comorbidities.

This study examined the incidence, reasons, and survival impact of FUL after a curative gastrectomy for gastric cancer in a consecutive patient cohort and investigated its underlying causes.

SUBJECTS AND METHODS

1. Patients

Patients who underwent a gastrectomy for primary gastric cancer at Gyeongsang National University Hospital between January 2016 and May 2019 were collected. After excluding patients who underwent palliative or non-curative surgery or lacked baseline clinical information, 435 patients were studied. The demographic variables, social history, and clinicopathological data were extracted from the electronic medical records. The study, including all data collection, was conducted after gaining approval from the institutional review board (IRB-2020-03-002).

2. Definition of FUL

Postoperative FU was performed according to international guidelines, with outpatient visits every three to six months during the first two years and annually thereafter for up to five years.^1-3 Patients who maintained a regular FU throughout the five-year postoperative period were classified as the FU group.

During the five years of surveillance, patients who failed to attend scheduled visits for more than 12 consecutive months without any documented issues were considered FUL. Patients who continued FU beyond five years after surgery were considered long-term cancer survivors and were classified as the FU group. The analyses were limited to data obtained during the initial five years of the postoperative period.

For patients in the FUL group, structured telephone interviews were conducted to identify the underlying reasons for discontinuing the hospital visits and collect information on the survival status, including the date and cause of death. Family members were reached when direct patient contact was not possible, and an interview was attempted. Based on these interviews and supplementary medical record reviews, the causes of FUL were categorized as follows: 1) patient refusal, 2) relocation, 3) transfer due to (other) comorbid diseases, 4) dissatisfaction with hospital services, 5) death from non-gastric cancer causes, 6) gastric cancer-related death, and 7) uncontactable despite repeated attempts.

3. Studied variables

The following demographic findings were reviewed: age, sex, body mass index (BMI), Charlson comorbidity index, alcohol consumption, smoking status, occupation, level of education, adherence to the National Cancer Program, private insurance, religious affiliation, urban residence, and familial support status (spouse, family other than spouse, and none).

The clinical variables included complications after surgery, surgical approach (minimally invasive vs. conventional open), extent of gastrectomy (distal, total, or function-preserving), change of FU surgeon/physician, pathological stage according to the 8th edition of the AJCC TNM classification, and systemic therapy.

4. Statistical analysis

All statistical analyses were performed using SPSS version 28 (IBM Corp., Armonk, NY, USA). The continuous variables are expressed as mean ± standard deviation and compared using a Student’s t-test. The categorical variables were compared using a chi-square test or Fisher’s exact test, as appropriate.

Univariate analysis was conducted to evaluate the association of the sociodemographic and clinical variables with FUL. The independent predictors associated with FUL after a gastrectomy were identified by entering the variables with p <0.10 in univariate analysis into a Cox proportional hazards regression model using the backward likelihood ratio method. Age, pathological stage, and extent of gastrectomy were also entered into the multivariable model regardless of the univariable significance because of their clinical relevance. Among overlapping or highly correlated variables, the one with greater explanatory power for the outcome was retained to avoid collinearity.

The results are expressed as the hazard ratios (HRs) with 95% confidence intervals (CIs). A two-sided p-value <0.05 was considered significant. The cumulative incidence of FUL over time was plotted using the Kaplan–Meier method. The overall survival (OS) in the FU and FUL groups was compared using the log-rank test, and survival curves were generated using the Kaplan–Meier method.

5. Sensitivity analysis for predictors of FUL

Sensitivity analysis was performed in which FUL was redefined as voluntary discontinuation of a scheduled FU only to determine the presence of any potential misclassification from including deceased patients in the FUL group. Patients whose FU ended due to non–gastric cancer death (n=56) or gastric cancer–related death (n=16) were excluded from the redefined FUL group for this analysis.

RESULTS

1. Patient characteristics

Four hundred and thirty-five patients who underwent a curative gastrectomy for gastric cancer were included in the analysis. The mean age of the study population was 62.7±11.7 years, and 72% were male. The mean BMI was 23.0±3.1 kg/m². Table 1 lists the baseline characteristics.

Table 1.

Baseline Characteristics of the Study Population

2. Univariate analysis

Univariate analysis of demographic and clinical variables associated with FUL (Table 2) showed that patients in the FUL group were significantly older than those in the FU group (67.2±11.1 vs. 60.8±11.4 years, p<0.001), had a lower mean BMI (22.3±3.0 vs. 23.3±3.2, p=0.002), and the male proportion was higher (p=0.038).

Table 2.

Univariate Analysis for the Predictors of Follow-up Loss

Regarding the social factors, adherence in the FU group was associated with occupation (p=0.001), higher education (p=0.002), regular participation in the National Cancer Screening Program (p=0.015), and possession of private insurance (p=0.004). Familial support status indicated that patients with a spouse had a lower rate of FUL, but it did not reach statistical significance (p=0.069).

In terms of the clinical factors, the FU group included a higher proportion of patients who underwent minimally invasive surgery compared to those who underwent conventional open surgery (p=0.036). Within the FUL group, patients who underwent a total gastrectomy had a significantly higher rate of FUL than those who received a distal or function-preserving gastrectomy (p=0.028). A change in the primary physician/surgeon during surveillance was significantly associated with FUL (p=0.013).

3. Cox-proportional hazard model for FUL

In the Cox proportional hazards model, older age (per one-year increase in age) was significantly related to a higher risk of FUL (HR=1.044, 95% CI 1.025–1.063, p<0.001) (Table 3). Each BMI unit (per 1 kg/m² increase) was inversely associated with FUL (HR=0.927, 95% CI 0.872–0.985, p=0.015).

Table 3.

Cox-Proportional Hazard Model for Follow-up Loss

Patients without familial support had a higher risk of FUL than those living with a spouse (HR=2.666, 95% CI 1.341– 5.300, p=0.005). A total gastrectomy was independently associated with an increased risk of FUL (HR=1.660, 95% CI 1.120– 2.461, p=0.012).

Sensitivity analysis excluding all deaths and redefining the endpoint as voluntary FUL among surviving patients revealed the following overall risk factors (Table 3). Age (p=0.008), absence of familial support (p=0.007), and total gastrectomy (p=0.004) were significant. The association between BMI and FUL was no longer statistically significant (p=0.293).

4. FU status and related reasons

During the 60-month FU period, 137 patients (31.5%) experienced FUL after a gastrectomy. The Kaplan–Meier cumulative- incidence curve revealed a gradual increase in the occurrence of FUL throughout the postoperative period, with a relatively steeper rise between 36 and 48 months, followed by a plateau thereafter (Fig. 1). The cumulative incidence of FUL at 12, 24, 36, 48, and 60 months after the gastrectomy was approximately 3%, 7%, 9%, 27%, and 31.5%, respectively.

Fig. 1.

Kaplan–Meier for follow-up loss after a curative gastrectomy for gastric cancer.

Table 4 lists the detailed reasons for FUL. The most common cause was death from causes unrelated to gastric cancer (n=56, 40.1%), followed by patient refusal to return despite being asymptomatic (n=20, 14.6%) and personal transfer or relocation (n=20, 14.6%). In addition, 17 patients (12.4%) reported transfer to another hospital for other comorbid diseases, and death related to gastric cancer at another institution was recorded in 16 cases (11.7%). Six patients (4.4%) were uncontactable, and two patients (1.5%) discontinued the FU due to dissatisfaction with hospital services.

Table 4.

Reasons for Follow-up Loss

5. Patterns of FUL according to clinical and sociodemographic factors

Fig. 2 presents the subgroup analysis of the causes of FUL. Female patients reported higher refusal and relocation rates (25.0% each), whereas male patients were more commonly lost due to non–cancer death (44.2%) (p=0.024). Non-gastric cancer death was highest in the function-preserving gastrectomy group (79.8%) compared to distal gastrectomy (42.4%) and total gastrectomy (32.5%) (p=0.011). Relocation was observed more frequently in stage I patients than stage II–III patients (19.4% vs. 2.6%, p=0.014), whereas gastric cancer-related death was significantly more common in stage II–III than in stage I (28.2% vs. 5.1%, p<0.001).

Fig. 2.

Distribution of the reasons for follow-up loss (FUL) across clinical and sociodemographic subgroups. The heatmap illustrates the proportion (%) of specific reasons for FUL among patients after gastrectomy, stratified by clinical and sociodemographic factors.

6. Survival analysis

Kaplan–Meier analysis revealed a significant difference in overall survival (OS) according to the FU status (Fig. 3). The FU group showed better survival outcomes than the FUL group (log-rank p=0.037) (Fig. 3A). The survival difference was significant in Stage I (p=0.046) (Fig. 3B), whereas Stage II (p=0.143) (Fig. 3C) and Stage III (p=0.257) (Fig. 3D) showed no difference between the groups.

Fig. 3.

Kaplan–Meier overall survival (OS) curves according to the follow-up loss after gastrectomy. OS was compared in all stages (A), stage I (B), stage II (C), and stage III (D).

The Cox proportional hazards model for OS (Table 5) revealed FUL as an independent factor for OS (HR=2.148, 95% CI 1.218– 3.709, p=0.006) following the pathologic stage (p<0.001). Other factors, including age, did not show significance.

Table 5.

Cox-Proportional Hazard Model for Overall Survival

Variable	Beta (coefficient)	Hazard ratio (Exp beta)	95% CI	p-value
Age (per 1-year increase)	0.024	1.020	0.995–1.047	0.118
Stage (8th AJCC)–overall				<0.001
Stage II vs I	2.419	11.233	4.504–23.559	<0.001
Stage III vs I	3.414	30.379	14.586–64.829	<0.001
Follow up loss	0.765	2.148	1.218–3.709	0.006

CI, confidence interval; AJCC, American Joint Committee on Cancer.

DISCUSSION

This study examined the key factors associated with FUL after a curative gastrectomy for gastric cancer. The incidence, causes, and survival impact of FUL in this clinical context were evaluated. Older age, total gastrectomy, and absence of familial support were independently associated with FUL. In addition, FUL was associated with a poorer OS in patients with stage I disease, whereas this association was not evident in stage II–III disease.

Older age and lower BMI were associated with FUL in the primary analysis. Both factors are generally related to frailty and sarcopenia, which can lead to reduced physical activity, cognition, and diminished motivation to maintain consistent medical FU.⁹ Gastric cancer patients with a higher BMI likely maintain better nutritional reserves and sustain engagement in FU care.¹⁰ These findings are consistent with previous reports, suggesting that the nutritional and functional status are the key determinants of health-related behaviors and adherence after gastrectomy.¹¹

Sensitivity analysis revealed a distinction between these factors. While older age remained even after excluding deceased patients, the association between lower BMI and FUL was no longer significant (p=0.293). This disparity suggests that although BMI and older age are linked to frailty, the influence of a lower BMI on FUL is connected primarily to the underlying patient condition and non-cancer-related mortality risk, rather than being a factor driving failure to comply with scheduled visits.

Total gastrectomy was also independently associated with higher FUL. Beyond its correlation with advanced disease and more complex postoperative courses, the physiological consequences of total gastrectomy, including chronic nutritional deficiencies, weight loss, and impaired quality of life, may discourage patients from attending long-term surveillance visits.^12,13 Sensitivity analysis showed that total gastrectomy remained a significant predictor, supporting the strength of this association. This finding aligns with previous reports indicating that nutritional deficiencies and a decreased quality of life after a total gastrectomy can discourage patients from attending long-term visits.^14,15

Familial support is an important determinant of FU adherence. In the present cohort, patients without any caregiver had a markedly higher risk of FUL than those supported by a spouse. Previous studies reported that caregiver involvement reduces complications, readmissions, and psychological distress while improving the functional outcomes and quality of life.^16,17 Spousal caregivers provide continuous emotional and logistical support, which may be related to patients maintaining routine clinic visits. These findings suggest that assessing caregiver availability and providing targeted support to socially isolated patients may help improve adherence. Nevertheless, these findings should be interpreted with caution because of the limited sample size from a single center and the potential of uncontrolled confounding factors inherent in the study design. Therefore, future larger-scale research will be needed to validate these associations.

FUL is generally defined as the discontinuation of patients’ hospital visits. Hence, the exclusion of deceased individuals from this category may be considered reasonable. Nevertheless, this study adopted a different approach. The mortality patterns that may have been influenced by behavioral characteristics associated with FUL could be assessed by including these patients.

Consequently, FUL was associated with the OS in stage I patients but not in stage II or III, which is consistent with the results of a secondary analysis of a well-controlled randomized clinical trial, KLASS-02.¹⁸ The report indicated that while intensive surveillance detected recurrence earlier, it did not improve the OS, suggesting the limited efficacy of treatments for recurrent gastric cancer and the influence of lead-time bias. Similar to these findings, FU may enable the timely detection of relapses, but its survival benefit remains uncertain in advanced disease.

The decreased survival in Stage I FUL patients is likely associated with non–cancer-related conditions or behavioral factors, compared to cancer progression alone. Previous studies have reported that patients with stage I gastric cancer who undergo curative surgery generally achieve comparable survival to that of the general population.^2,19 In this setting, non-cancer-related conditions, such as cardiovascular disease, cerebrovascular events, chronic pulmonary disease, infections, and second primary malignancies, become major competing risks for mortality.^20,21 These results suggest that FUL in Stage I patients may be a marker of broader behavioral or health vulnerabilities. Therefore, individualized survivorship programs for early-stage patients, emphasizing the management of comorbidities and health promotion, may be warranted.

In long-term nationwide studies, socioeconomic factors were related to the survival outcomes in gastrectomy patients.^22,23 These studies focused on patients without active disease progression or ongoing treatment effects, providing insight into the determinants of survival. In the present finding, several socioeconomic variables were significant in univariate analysis, but none retained significance in the regression model. These findings suggest that socioeconomic factors may indirectly influence FU adherence, and their influence can vary depending on the patient cohort and analytic settings. Although these variables did not reach significance in this study, they may be considered candidates for the predictors of the long-term outcomes in future research.

In the present study, structured telephone contact with patients and families achieved a 95.6% response rate, suggesting that proactive communication strategies can minimize data loss and provide opportunities to monitor the patient status. Patient management systems that enable telephone contact with those who miss scheduled visits can help monitor their well-being, or this finding may provide preliminary evidence for the use of telemedicine. Previous studies have shown that mobile or digital health interventions can enhance symptom monitoring and treatment adherence in oncology settings.^24-26 Nevertheless, further evidence is required for long-term effectiveness, acceptability, costs, and equity of access before widespread implementation.^27,28

This study had several limitations. It was conducted at a single tertiary center within a uniform healthcare system, which may limit generalizability to other populations with different cultural or insurance frameworks. The survival and cause-of-death data were obtained primarily through structured interviews, raising the possibility of recall bias and misclassification of the specific causes of death. Psychosocial distress and mental health status were not measured using validated instruments, even though these factors may affect the motivation for FU and healthcare-seeking behavior. Future multicenter studies incorporating standardized psychosocial and quality-of-life assessments, as well as a linkage to national mortality data, will be needed to clarify the behavioral and clinical determinants of FUL and survival.

Despite these limitations, this study is the first to evaluate the clinical and social factors associated with FUL after gastrectomy and explore its stage-specific association with survival. FUL was associated with identifiable clinical (age, BMI, and extent of gastrectomy) and social (familial support) characteristics and with poorer survival among patients with stage I gastric cancer. Implementing risk-adapted surveillance models, strengthening caregiver- and community-based support, and integrating proactive communication strategies may help improve FU adherence and the long-term outcomes in this population.