Abstract
Aim: Gastrectomy as a primary treatment for patients with metastatic gastric cancer (M1) is highly controversial. Herein, a review of the literature was undertaken with the aim of assessing evidence regarding associated morbidity and mortality, overall survival, palliation and quality of life. Materials and Methods: A systematic review of the literature from 1980 to 2013 was undertaken to identify relevant studies. Outcome data were pooled, and combined overall effect sizes were calculated using fixed or random effects models. The search identified 19 non-randomized studies reporting on 2,911 patients. Results: Overall postoperative mortality and morbidity were 14% and 27% and were higher in Western than in Asian patients. In studies published during the past decade postoperative mortality was less than 5%. The weighted 1- and 2-year overall survival rates were 38% and 17%, and were twice as high in Asian versus Western patients. In the meta-analysis, the 1-year overall survival was significantly higher in patients undergoing gastrectomy versus conservative (odds ratio (OR)=4.9, 95% confidence interval (CI)=3.2 to 7.5, p<0.0001) or gastrectomy versus non-resectional treatment (OR=2.6, 95% CI=1.7 to 4.3, p<0.0001). Studies reporting on quality of life and palliation indicate a possible benefit of such palliative gastrectomy. Conclusion: A possible benefit of gastrectomy compared to non-resectional treatment for stage IV gastric cancer in terms of survival and palliation was evident but has to be cautiously interpreted due to potential sources of bias of retrospective non-randomized studies. Several questions regarding the optimal management of these patients remain unanswered and require a properly-designed randomized trial.
Even though the incidence and mortality from gastric cancer has decreased during the past two decades, the proportion of stage IV cancer increased significantly during the same period and comprised more than 40% of total cases (1). The prognosis of gastric cancer with distant metastases is poor and it rarely exceeds 5% at five years (2). These patients may present with severe symptoms due to complications of the tumor, such as bleeding, obstruction or perforation (up to 45%) requiring surgery. Alternatively they may present with minor-to-moderate symptoms not requiring urgent intervention (3). Half of patients with non-resected advanced gastric carcinoma will develop severe tumor-related complications during the remaining period of their life, necessitating operation (4). For the latter group of patients, the role of gastric resection has been a matter of debate during the past four decades.
Up to one-third of patients with distant metastases undergo gastric resection without curative intent and one-third of them undergo non-resectional operations (5, 6). Gastric resections include total, sub-total or distal gastrectomy. Proximal gastrectomy is performed rarely, depending on the anatomical localization of the tumor. Non-resectional operations include bypass procedures, jejunostomy or gastrostomy. Any of these procedures can be considered as palliative according to the WHO statement because they aim to “improve the quality of life (QoL) of the patient (…) through prevention and relief of suffering (…)” (7). The purpose of palliative gastrectomy it is to relieve the symptoms associated with cancer, to prevent subsequent tumor complications which are likely to occur and to improve the quality of life. In addition, by reducing the tumor burden, it may be associated with a prolonged survival time
A systematic review of the literature was undertaken with the aim of assessing evidence regarding the outcomes of gastrectomy as a primary treatment for patients with metastatic (M1) gastric carcinoma by addressing the following objectives: What is the overall survival associated with gastric resection? What is the associated morbidity and mortality of this approach? What is the evidence in terms of QoL of this approach?
Materials and Methods
Criteria for considering studies for this review. Types of studies: All prospective and retrospective clinical studies assessing the outcome of gastrectomy for stage IV gastric cancer were considered. Animal studies, case reports and case series with fewer than 10 patients were excluded. Only full text peer-reviewed studies were included. Types of participants: Adult patients (older than 18 years old) with histologically-proven gastric adenocarcinoma diagnosed with stage IV disease (any T, any N, M1) by means of radiological imaging or laparotomy were included.
Types of interventions: Gastric resection comprising total gastrectomy, subtotal, distal or central gastrectomy as primary treatment for gastric cancer were assessed. Patients having undergone previous surgery or neoadjuvant chemotherapy, radiotherapy or combination were excluded.
Types of outcome measures: The outcome measures were: postoperative mortality and morbidity expressed as a percentage of total patients; overall survival at one and two years postoperatively; postoperative quality of life (standardized or not); and postoperative palliation (namely rate of symptom palliation, need for subsequent palliative interventions, and rate of subsequent tumour complications).
Search methods for identification of studies. The literature search included the period January 1980–April 2013 of the MEDLINE database. The following key words and Medical Subject Heading (MeSH) terms were used in searching: “stomach neoplasms” (MeSH Major Topic) AND “gastrectomy” (MeSH Major Topic) AND (“humans” (MeSH Terms) AND “adult” (MeSH Terms)) AND ((“1980/01/01” (PDAT): “2013/04/01” (PDAT)). Only studies published in English, French, Italian and German were included. Additional articles were searched in the reference lists of the eligible publications.
Data collection and analysis. The abstracts of the citations extracted from this initial search were subsequently screened for potential eligibility. The full text of potentially eligible papers was reviewed, studies to be included were identified and their reference lists were screened for additional eligible articles. Data relevant to the aims of the study were extracted and were entered into a structured database. Whenever necessary, overall survival rates were extracted from respective Kaplan–Meier curves.
Statistical analysis. Descriptive statistics of the data were performed with SPSS V.17.0. The analyses were performed with the use of Comprehensive Meta-Analysis software (version 2.0; Biostat, Englewood, NJ, USA). Synthesis of summary statistics of survival analysis was undertaken because individual patient data from different studies could not be amalgamated. The number of patients in each group along with the 1-year overall survival rate were used for meta-analysis to generate the respective odds ratio (OR) as the summary statistic. Pooled ORs with 95% confidence intervals (CIs) were calculated using the Mantel–Haenszel fixed-effects model, unless evidence of between-study heterogeneity existed, in which case random effects models of DerSimonian and Laird were applied (8). In-between study heterogeneity was examined with the combination of the Cochran's Q (chi-square) test and the I2 statistic; I2 values, <25% indicate low heterogeneity, between 25% and 50% moderate heterogeneity, and >50% high heterogeneity. Publication bias was assessed both visually evaluating the symmetry of funnel plots and formally using the Egger's regression intercept, with a value of less than 0.05 as the level of significance (9).
Results
A flow diagram of the literature search is presented in Figure 1. Nineteen eligible articles were retrieved, representing a total of 2,911 patients. Only one study was prospectively designed and none of them was randomized (5). Summary results are presented in Table I.
Postoperative morbidity and mortality. Twelve studies provided data regarding postoperative morbidity and mortality (Table II) (2, 3, 5, 6, 10-17). Postoperative morbidity following palliative gastric resection was comparable with that after non-resectional operation (2, 3, 5, 12, 15, 18). This latter was explained in part by selection of patients in very poor condition, at high risk of complications, for less extensive interventions (5). Intensity of preoperative symptoms did not correlate with postoperative mortality or morbidity or with postoperative hospital stay (5, 18). Palliative gastrectomy in Asians as well as in older patients was associated with higher risk of surgical complications (17). Postoperative stay following palliative gastrectomy is generally prolonged mainly due to the high frequency of postoperative complications. Most studies report median postoperative hospital stay duration between 13 and 33 days. On the other hand, postoperative stay may not differ between resectional and non-resectional operations (3, 13, 15).
Survival after gastrectomy. Eighteen studies (n=2,881) provided overall survival data (Table III) (2, 5, 6, 11, 12, 15-27). Meta-analysis of 10 studies including 2,075 and 3,698 patients undergoing gastrectomy or non-resectional operation, respectively, showed that gastric resections were associated with a 5-fold higher overall survival rate compared with non-resectional operations (p<0.0001, test for heterogeneity: p<0.001, I2=78%, test for publication bias: p=0.173) (Figure 2A). Meta-analysis of data from two studies including patients undergoing gastrectomy (n=187) or conservative treatment (n=144) showed that gastric resections were associated with a 2.5-fold higher overall survival rate compared with conservative treatment (p<0.0011, test for heterogeneity: p<0.001, I2=78%, test for publication bias: p=0.739) (Figure 2B). The independent predictors of survival included age, performance status, several markers of tumor burden or biological behavior, resection and postoperative chemotherapy. Markers of tumor burden included the site and the number of metastatic lesions, whereas markers of biological behavior were the histological type, extent of tumor burden, presence of lymphatic or venous infiltration, ascites and histological differentiation (5, 16, 18, 20, 25, 27). Patients with multiple bilobar liver metastases, patients with metastasis in more than one organ or patients with extended peritoneal dissemination (P3) had the poorest survival among patients with stage IV disease. Extended lymph node dissection was generally avoided during palliative gastric resections but some authors reported a positive survival effect of D2 or greater dissection compared to limited lymphadenectomy (D0, D1) (25, 27). Emergency operation due to bleeding, obstruction or perforation is associated with significantly shorter survival compared to elective operations (2). In most of the studies, the best overall survival was reported for patients who underwent palliative resection and postoperative chemotherapy (12, 21, 23). This benefit remained in the multivariate analysis adjusting for multiple patient- and tumor-related variables (19, 20). Postoperative chemotherapy was administered to 40-70% of the patients. In the multivariate analysis, the effect of palliative gastrectomy appears to be higher than the effect of chemotherapy adjusted for performance status and markers of tumor burden (5).
Flow diagram of the literature search.
Palliative gastrectomy for stage IV gastric cancer. Summary results of studies included in the systematic review.
QoL and palliation. Seven studies (n=580) provided data on QoL and palliation (2, 3, 5, 11, 13, 14, 21) (Table IV). Two of them used standardized measures of QoL (QLQ-C30), two of them used hospital-free survival as an indirect measure of QoL and the rest used other measures, such as the development of subsequent complications and the duration of palliation. In the prospective study of Kulig et al., parameters for QoL were similar in patients undergoing gastrectomy for metastatic and non-metastatic disease for up to 12 months after the operation (5). In studies including symptomatic patients (bleeding, obstruction), significant palliation expressed as hospital-free survival was also reported (13, 28). Patients undergoing gastric resections experience a lower rate of severe tumor-related complications in their remaining lifespan, whereas the majority of patients undergoing non-resectional operations may experience severe tumor-related complications that require at least one interventional or surgical procedure. In the study of Samarasam et al., patients who had resectional operation were able to follow normal daily activities after surgery, take normal diet postoperatively and had significantly less vomiting, hematemesis and melena than patients undergoing non-resectional operations (29). Postoperative symptom palliation, prevention of further operations until the end of their lives and improved QoL has also been reported for patients undergoing resections with positive surgical margins as well as for patients undergoing extended resection for proximal gastric cancer (11). In the study of Kahlke et al. (3), patients with major symptoms (i.e. bleeding, obstruction or perforation) before the operation perceived their postoperative QoL as being higher than that of patients with minor symptoms and this difference persisted for at least three months after surgery. In that study, the preoperative quality of life, assessed with the QLQ-C30 questionnaire, was similar between patients with major and minor symptoms (17). Other authors observed no difference in the postoperative hospital-free survival between resectional and non-resectional gastric operations (2). Park et al. (21) reported a longer symptom-free interval in patients undergoing resectional operations but also a higher need for operation due to tumor complications, and concluded that palliative resection has no significant value as a preventative measure for anticipated complications (21).
Postoperative mortality morbidity and hospital stay of patients undergoing gastric resection for stage IV gastric cancer.
Studies providing data regarding overall survival of patients undergoing gastric resection for stage IV gastric cancer.
A: Meta-analysis of studies comparing one year overall survival between patients undergoing gastric resection for stage IV gastric cancer and non-resectional surgery B: Meta-analysis of studies comparing one year overall survival between patients undergoing gastric resection for stage IV gastric cancer and conservative treatment. *Metastasis to one organ, **metastasis to multiple organs, #liver metastasis, ##peritoneal dissemination (see also Table III).
Discussion
For a long time, palliative gastric resection has been considered of no clinical value due to the higher risk of postoperative mortality out weighing any potential advantage in terms of survival and quality of life. Published data from 1960s to the 1980s, reported surgical mortality rates often reaching 20% (6). In this review, articles published after the mid-1990s report mortality rates not exceeding 7% and the most recent reports show that these operations can be performed with very low postoperative mortality (5). This progress is mainly due to proper selection of patients, progress in anesthesia and preoperative nutritional support, better preoperative management of associated medical problems, effective postoperative care, especially with the advent of interventional radiological techniques sparing reoperation for several postoperative complications, and advances in technology and surgical equipment.
In this systematic review of the literature, the role of gastric resection for stage IV gastric cancer was assessed. The results showed acceptable postoperative morbidity and mortality, which was sometimes even similar to those for curative resection. Patients undergoing gastric resection had significantly higher likelihood of surviving one year after the operation compared to patients undergoing non-resectional operations or non-operative treatment. Finally, there were several studies to support a role of gastric resection in patients with distant metastasis in terms of palliation, improvement of the QoL and prevention of subsequent tumor complications.
Postoperative palliation and quality of life (QoL) of patients undergoing gastric resection for stage IV gastric cancer.
However, these observations have to be interpreted with caution. Almost all studies were retrospective and none of them was randomized. Thus, several sources of bias exist. Heterogeneity between the studies, as well as significant risk of publication bias, was evident from the results of meta-analysis. Moreover, most of the studies compared gastric resection with non-resectional operations, whereas comparisons with chemotherapy as first-line of treatment or with non-surgical interventions might also have been of great relevance. Evidence from retrospective and prospective studies suggests that surgery is associated with better long-term results and is therefore the treatment of choice in patients with a life expectancy of two months or longer. Stent placement was associated with better short-term outcomes and it is preferable for patients expected to live less than two months (30). Patients included in the studies of this review differed not only in terms of tumor burden and performance status, but also in terms of age and race, all of which are significant determinants of outcome. In addition, patients received different postoperative chemotherapy regimens as there was no internationally accepted standard-of-care. Finally, the selection bias, favoring patients undergoing surgery, indisputably contributed to the survival benefit reported in almost all studies, even though the benefit of palliative gastrectomy persisted in the multivariate analysis when adjusting for multiple confounding factors such as performance status, tumor burden and age.
Thus, many questions regarding the role of gastrectomy as primary treatment for stage IV cancer remain unanswered and require for an appropriately designed prospective study. The Results of the Japanese Clinical Oncology Group Study JCOG 0705 and Korean Gastric Cancer Association Study KGCA01 including patients with a single non-curable factor are awaited, although similarly to other oncological and surgical aspects of gastric cancer, the differences between Western and Asian populations will limit their representativeness (5, 31). Therefore, strong evidence on Western cohorts would be of great value. An appropriate study will have to randomize patients to either gastrectomy or chemotherapy as primary treatment for stage IV gastric cancer and compare the outcomes in terms of overall survival; treatment-related morbidity and mortality, quality of life and associated costs. Until then, the treatment of these patients remains highly individualized and an object for debate.
Acknowledgements
We would like to thank Professor Thomas Lehnert for critically reviewing the final version of the article.
Footnotes
-
↵* K. Lasithiotakis was supported by the Major Training Fellowship Grant of the European Society of Surgical Oncology for the year 2013.
-
This article is freely accessible online.
-
Part of the results of this study was presented at the Second ‘ESSO Advanced Course on Upper GI Malignancy in Poznan, Poland, 2013.
-
Conflicts of Interest
The Authors declare no conflicts of interest with regard to this study.
- Received January 3, 2014.
- Revision received March 2, 2014.
- Accepted March 4, 2014.
- Copyright© 2014 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
