Long‐term impact of postoperative pneumonia after curative gastrectomy for elderly gastric cancer patients

Abstract With the extension of life expectancy, cancer has been increasing in elderly populations. Postoperative pneumonia can negatively influence immediate mortality following gastrectomy for elderly patients, but its impact on long‐term survival remains unclear. We retrospectively reviewed the cases of patients aged ≥75 years who underwent curative gastrectomy for gastric cancer from 2000 to 2014 to determine the long‐term effects of postoperative pneumonia and to identify independent risk factors along with physical status and surgical procedure. Of 250 patients, 32 (12.8%) developed postoperative pneumonia. Patients with postoperative pneumonia had significantly worse overall survival (OS) than those without postoperative pneumonia (P<.001). A multivariate analysis identified postoperative pneumonia as a prognostic factor for OS (hazard ratio, 2.06; 95% CI, 1.05‐3.75; P=.036). Significant risk factors for the development of postoperative pneumonia were male gender (P=.026) and D2 lymphadenectomy (P<.001). D2 lymphadenectomy was associated with poorer OS than D1 or D1+lymphadenectomy in patients with an American Society of Anesthesiologists (ASA) score 3 (P=.026), but did not influence OS negatively in patients with an ASA score ≤2. Limited lymphadenectomy did not affect the cancer‐specific survival of elderly patients with ASA score 3. Postoperative pneumonia following gastrectomy has an adverse impact on the long‐term survival of elderly gastric cancer patients. A limited lymphadenectomy during curative resection should be considered to prevent postoperative pneumonia in frail elderly patients with ASA score 3. Postoperative pneumonia following gastrectomy has an adverse impact on the long‐term survival of elderly gastric cancer patients. Extent of lymph node dissection during curative resection should be limited to prevent postoperative pneumonia, based on the patient's frailty.


Abstract
With the extension of life expectancy, cancer has been increasing in elderly populations. Postoperative pneumonia can negatively influence immediate mortality following gastrectomy for elderly patients, but its impact on long-term survival remains unclear. We retrospectively reviewed the cases of patients aged ≥75 years who underwent curative gastrectomy for gastric cancer from 2000 to 2014 to determine the long-term effects of postoperative pneumonia and to identify independent risk factors along with physical status and surgical procedure. Of 250 patients, 32 Society of Anesthesiologists (ASA) score 3 (P=.026), but did not influence OS negatively in patients with an ASA score ≤2. Limited lymphadenectomy did not affect the cancer-specific survival of elderly patients with ASA score 3. Postoperative pneumonia following gastrectomy has an adverse impact on the long-term survival of elderly gastric cancer patients. A limited lymphadenectomy during curative resection should be considered to prevent postoperative pneumonia in frail elderly patients with ASA score 3. Postoperative pneumonia following gastrectomy has an adverse impact on the long-term survival of elderly gastric cancer patients. Extent of lymph node dissection during curative resection should be limited to prevent postoperative pneumonia, based on the patient's frailty.

| INTRODUCTION
The populations of many countries are aging rapidly and, in Japan, it is estimated that 26% of the population is now ≥65 years old. 1 The number of elderly patients with gastric cancer has also increased annually and, in 2011, 27% of gastric cancer patients in Japan were over 80 years old. 2 Development of therapeutic strategies for this segment of the population presents a challenge, as elderly patients usually have comorbidities and poor physical status that adversely affect their surgical outcomes and prolong their postoperative disabilities. [3][4][5] Surgical resection is the main therapeutic modality for malignancies, but surgery has the potential to result in significant worsening of an elderly patient's quality of life. Postoperative complications also seriously impair physical status and directly result in poor outcomes in elderly patients. 6 These disadvantages of surgery can worsen the prognosis of elderly patients per se.
Surgical treatment must be planned on the basis of tolerability, curability and future quality of life. We need to understand the causes of reduced survival in order to improve the surgical outcomes of elderly patients.
Pneumonia is one of the main causes of death in the elderly, 7 and it is a relatively common postoperative complication. Among abdominal surgeries, gastrectomy is especially associated with postoperative pneumonia, with the second highest incidence (16-17%) following thoracic surgery (15-37.5%). 8,9 Several studies have reported that pulmonary complications negatively affect not only short-term mortality but also long-term survival after thoracic surgeries such as resection of esophageal and lung cancer. [10][11][12][13] Although a previous study reported that patients with postoperative pneumonia following gastrectomy had high postoperative mortality, 14 the potential long-term impact of postoperative pneumonia has not been investigated. Identification of risk factors for postoperative pneumonia may have a crucial role in the development of therapeutic strategies to reduce the incidence and impact of pneumonia in elderly patients.
We conducted the present study to evaluate the impact of postoperative pneumonia after curative gastrectomy on long-term survival in elderly gastric cancer patients and to identify potential independent risk factors. Clinical characteristics, surgical outcomes, pathological findings and follow-up data were extracted from medical records. Physical status was assessed based on the following factors: age, gender, smoking history, American Society of Anesthesiologists (ASA) score, Charlson comorbidity index (CCI), presence or absence of chronic obstructive pulmonary disease (COPD), body mass index (BMI), and prognostic nutritional index (PNI). PNI was calculated as 109 serum albumin value +0.0059 total lymphocyte count in peripheral blood just before resection. 15 Tumor status was diagnosed according to the Japanese Gastric Cancer Association classification system. 16

| Surgical procedure
Extents of gastrectomy and lymph node dissection were defined based on tumor location and depth of invasion according to the Japanese Gastric Cancer Treatment Guidelines. 17 D2 lymphadenectomy is recommended for tumors with more than clinical T2 or N1 disease, whereas D1+lymphadenectomy is allowable for clinical T1N0 tumors. D1 lymphadenectomy removed perigastric lymph nodes plus the lymph node at the base of the left gastric artery.
D1+lymphadenectomy was defined as D1 lymphadenectomy plus nodal dissection around the celiac artery, along the common hepatic artery, and at the base of the splenic artery particularly for total and proximal gastrectomies. D2 lymphadenectomy was defined as D1+lymphadenectopmy plus nodal dissection along the proper hepatic artery, and around the distal splenic artery and splenic hilar with a splenectomy (particularly for total gastrectomy). For clinical stage I disease, laparoscopic gastrectomy was carried out.

| Definition of pneumonia
Postoperative pneumonia was defined as higher than grade II on the Clavien-Dindo classification. 18 Pneumonia was diagnosed based on the findings of consolidation by chest X-ray or computed tomography (CT), and on the clinical findings (ie presenting one or more of the following factors: temperature ≥38°C, cough productive of purulent sputum, and raised white cell count [≥10 000/mm 3 ]).

| Statistical analyses
We assessed the results of a univariate analysis of risk factors for postoperative pneumonia by using the Mann-Whitney U-test and Student's t-test for continuous variables and with Pearson's chisquared test and Fisher's exact test for categorical variables as appropriate. Variables found to have a significant association with postoperative pneumonia in the univariate analysis were subjected to a multivariate logistic regression analysis. Cumulative survival rate was calculated by the Kaplan-Meier method, and survival curves were compared using the log-rank test. Cox proportional hazards model was constructed to identify independent prognostic factors. All statistical analyses were carried out using JMP statistical software (ver. 11: SAS, Cary, NC, USA). P-values <.05 were considered significant.

| Comparison of survival between the extents of lymph node dissection
To determine which patients were unfit for D2 lymphadenectomy, we compared the OS curve stratified by ASA score between D2 lymphadenectomy and D1 or D1+lymphadenectomy. D2 lymphadenectomy significantly decreased OS rate compared to D1 or D1+lymphadenectomy in the 54 elderly patients who had an ASA score of 3 (log rank P=.026, Figure 2A), whereas D2 lymphadenectomy did not influence OS negatively in 196 patients with an ASA score ≤2 ( Figure 2B). Regarding patients with ASA score 3, significant differences between guideline recommended lymph node dissection and less than recommended lymph node dissection were not

| DISCUSSION
Several studies have shown a relationship between postoperative complications and long-term survival in various malignancies. 10 ASA, American Society of Anesthesiologists; BMI, body mass index; CCI, Charlson comorbidity index; C-D, Clavien-Dindo; PNI, prognostic nutritional index.
overall survival in elderly patients undergoing curative-intent gastrectomy, and our analysis showed that postoperative pneumonia was also an independent predictor for worse survival. These findings mean that postoperative pneumonia affects prognosis separately from preoperative health status, operative procedure, and disease stage in elderly patients.
A potential explanation for these findings is that postoperative pneumonia prolongs patient hospital stay, worsens postoperative disability, and reduces activities of daily living. Impaired physical status may make elderly patients more susceptible to cancer-related death and cancer-unrelated death. An additional possible explanation is host immunosuppression induced by inflammatory responses to complications. As an inflammatory disease, pneumonia leads to immunomodulatory impairment. These circumstances may facilitate the progression of residual microscopic disease that manifest as a recurrence or death. Further investigations such as nationwide studies are necessary to elucidate the mechanisms by which postoperative complications worsen prognoses.
Frequency of pneumonia after gastrectomy in elderly patients aged ≥75 years has been reported as 5.1-13.3%, which is significantly higher than that of younger patients (ie 45-65 years old). 14,24,25 Pneumonia developed in 12.8% of the present patients, which is similar to findings of previous studies, but advanced age (≥80 years) did not increase the incidence of pneumonia in our study cohort. One possible reason for this result is that we chose patients ≥75 years age, which is a heterogeneous cohort with substantial differences in health status, and thus the age-related changes do not necessarily correspond to chronological age. Reported risk factors for the development of postoperative pneumonia in patients aged ≥75 years are gender, ASA-PS, PS (performance status), impaired respiratory function, diabetes mellitus and blood transfusion, but these factors varied among all three studies. 14,24,25 This is likely to be a reflection of the small sample size in each study.
The present study's multivariate analysis identified radical lymphadenectomy as an independent risk factor for the development of postoperative pneumonia. Gastrectomy with D2 lymphadenectomy has been a standard procedure for gastric cancer, but there is no available prospective study which confirms the feasibility of D2 lymphadenectomy for patients >75 years old. Karasaki et al. 26  To modulate the extent of lymph node dissection on the basis of curative intent and physical fitness, we stratified the elderly patients according to ASA grade to identify the patients who were not fit for radical lymphadenectomy. We found that radical lymphadenectomy was associated with a poorer overall survival of the patients who had higher ASA scores. Limited lymph node dissection seems not to influence the CSS in ASA score 3 elderly patients. Inferiority of less than D2 lymphadenectomy for frail patients with advanced stage Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved: T. Oshikiri and T. Nakamura.