Perioperative outcomes and survival in elderly patients aged ≥ 75 years undergoing gastrectomy for gastric cancer: an 18-year retrospective analysis in a single Western centre

Purpose Surgical resection for elderly patients with gastric cancer is controversial. This study aims to evaluate the preoperative features and postoperative short- and long-term outcomes of elderly patients following surgical resection for gastric adenocarcinoma. Methods Between January 2000 and May 2018, a total of 177 consecutive patients underwent curative gastrectomy for gastric adenocarcinoma was retrospectively reviewed. Propensity score matching (PSM) analysis was used to balance confounding covariates between the elderly and non-elderly groups. Clinicopathological characteristics, intraoperative characteristics, postoperative complications and long-term survival outcomes including overall survival (OS) and Disease Specific Survival (DSS) were compared and analysed using the Kaplan-Meier log-rank test. Multivariate cox proportional hazards regression analysis of clinicopathological factors influencing survival were evaluated. Results There were 50 patients in the elderly group (age >= 75 years) and 127 patients in the non-elderly group (age < 75 years). Elderly patients had more comorbid conditions (p < 0.001), lower albumin concentration (p = 0.034), lower haemoglobin levels (p = 0.001), and poorer renal function (p = 0.043). TNM stage was similar between both groups (p = 0.174); however, lymphatic invasion (p = 0.006) and lymph node metastasis (p = 0.029) were higher in the elderly group. Elderly patients were much less likely to receive any chemo- (p < 0.001) or radiotherapy treatment (p = 0.007) with surgical treatment. After PSM, there were 50 patients in each group. Elderly patients were more likely to develop complications (Clavien Dindo >= 2: 50% vs. 26%, p = 0.003). The most common postoperative complications were pneumonia (12% vs. 6%, p = 0.498) and delirium (10% vs. 0%, p = 0.066). Elderly patients had a longer median length of hospital stay (median (IQR): 15.6(9.5) vs. 11.3 (9.9), p = 0.030). There were no differences in 30-day mortality (elderly vs. non-elderly: 1% vs. 1%, p = 0.988). Before and after PSM, age remains an independent predictor of postoperative complications. Before PSM, the estimated mean OS for the elderly and non-elderly patients were 108 months (95%CI, 72.5-143.5) and 143 months (95%CI, 123.0-163.8), respectively (p = 0.264). After PSM, the estimated mean OS for the elderly and non-elderly patients were 108 months (95%CI, 72.5-143.5) and 140 months (95%CI, 112.1-168.2), respectively, (p = 0.360). Before PSM, the estimated mean DSS for the elderly and non-elderly patients were 94 months (95%CI, 61.9-127.5) and 121 months (95%CI, 100.9-141.0), respectively (p = 0.405). After PSM, the estimated mean DSS for the elderly and non-elderly patients were 94 months (95%CI, 61.9-127.5) and 115 months (95%CI, 87.3-143.3), respectively (p = 0.721). Age was not an independent predictor of mortality following gastrectomy for gastric cancer in both PSM matched and unmatched cohort. Conclusion Chronological age alone is not a contraindication to curative resection of gastric adenocarcinoma in elderly patients with acceptable risk. Whilst age affects perioperative complications, the incidence of postoperative mortality and overall survival were not significantly different between elderly and non-elderly gastric cancer patients treated with curative surgery. Gastrectomy with D2 lymphadenectomy can also be performed in carefully selected elderly patients by surgeons with expertise in gastric resection along with appropriate perioperative management.