Indications for pancreaticoduodenectomy affected postoperative outcomes in octogenarians

Abstract Aims The safety and efficacy of pancreaticoduodenectomy (PD) in patients over the age of 80 years remain controversial. We aimed to examine post‐PD outcomes and to determine the age limit for PD. Methods Patients were divided into two subgroups: the younger (<80 years) group and octogenarian (≥80 years) group. We retrospectively evaluated the clinical benefit of PD for periampullary diseases in the younger and octogenarian groups, focusing on short‐ and long‐term outcomes. Results From March 2005 to December 2018, 586 consecutive surgically curable patients with diagnosed periampullary diseases were studied, among whom 122 (20.8%) were ≥80 years old. The general preoperative physical condition (G8 screening, instrumental activities of daily living, and Charlson comorbidity index) and nutritional status were significantly worse in the octogenarian group. However, there were no significant differences between the younger and octogenarian groups in postoperative severe complication rates (34% vs 36%) or perioperative mortality rates (1.5% vs 0.0%). We observed significantly poorer 3‐, 5‐, and 10‐year overall survivals in the octogenarian group than in the younger group (P = .007). In the younger group, the main cause of death (89.6%) was cancer recurrence. However, only 60% of patients in the octogenarian group developed and died from cancer recurrence. Increased neutrophilic/lymphocyte ratio and elevated Controlling Nutritional Status score were associated with worse outcomes. Conclusions It is important to carefully determine the indication for PD in octogenarian patients with periampullary diseases, although patient age over 80 years should not be a contraindication for PD.


| INTRODUC TI ON
Opportunities to perform highly invasive surgeries for the elderly and extend the average life span have increased worldwide by progress of safe operation management. The incidence of various cancers has increased and continues to increase as the mean population age rises in Western countries, as reported recently by the Japanese Ministry of Health, Labour, and Welfare. 1,2 Difficult operations are associated with a high rate of complications, but for pancreatoduodenectomy (PD), safety has greatly improved over the past 30 years for operations for pancreatic fistula, delayed gastric emptying, postoperative abdominal hemorrhage, intra-abdominal abscess, and sepsis. 3,4 Because elderly patients are more likely to have comorbidities compared to young patients, effective PD strategies are essential. Furthermore, elderly people are more easily affected by perioperative adverse events. 5 In recent years, improvements in operative and anesthetic techniques, regionalization to high-volume centers, implementation of standardized recovery pathways, and better understanding and management of common complications have contributed to markedly improve short-term outcomes after PD for the elderly. 6,7 With regard to the safety and effectiveness of complicated surgical resections in patients 80 years of age or older, only a few studies have reported the factors that are associated with poor disease prognoses and surgical outcomes. Furthermore, few studies have evaluated the relative advantages of surgical resection in patients 80 years of age or older with regards to operation outcomes. 8,9 Therefore, we retrospectively examined post-PD outcomes in patients of various ages at our own institution. We also evaluated whether there was an age limit in performing PD.

| Patients
We retrospectively reviewed the surgical pathology database of the Kochi Health Science Center to identify patients who underwent pancreatectomies between April 2005 and December 2018.
Patients who underwent PD during the study period were selected as the study cohort. The file system of the Medicare provider was used to distinguish the inpatient hospitalization requests of the Kochi Health Science Center. Statements that were either from the outpatient or submitted by a non-facilities provider were identified using a standard outpatient file analysis system of the Kochi Health Science Center. Physical status and preoperative laboratory values were obtained within two weeks prior to the initiation of surgery.
The American Society of Anesthesiologists (ASA) physical status classification system, G8 screening tool, instrumental activities of daily living (IADL), and Charlson comorbidity index (CCI) were available for all patients in our series. [10][11][12][13] The body mass index (BMI) was calculated by dividing the body weight in kilograms by the square of the height in meters. The prognostic nutritional index (PNI) was calculated based on the serum albumin and total lymphocyte count, using the following equation: PNI = 10 × serum albumin (g/dL) + 0.005 × total lymphocyte count (/mL). 14

The Controlling
Nutritional Status (CONUT) score is an index calculated from the following factors: serum albumin concentration, total peripheral lymphocyte count, and total cholesterol concentration. 15 Our criteria, based on the proper selection of octogenarian patients with pancreatic cancer, were as follows: patients with (a) no severe comorbidities, (b) no cognitive impairment, (c) preserved preoperative functional status, (d) the expectation that the prognosis would be extended if surgical resection with intent-to-cure could be performed, compared to when surgical treatment was not performed, and (e) the desire for a cure for the pancreatic cancer. 16 The institution's ERAS protocol for perioperative care of PD patients was developed based on guidelines by the ERAS society published in 2012. 17 Our department followed the prognosis of each case and obtained accurate details of the outcomes. The study was approved by the ethics committee of Kochi Health Sciences Center. All patients provided written informed consent.

| Assessment
After curative surgical resection, patients were divided into two groups: a younger group (<80 years of age) and an octogenarian group (80 years of age or older). The primary outcomes of this study were the rate of postoperative complication and overall survival in patients with PD, comparing the younger and octogenarian groups.
Postoperative morbidities, including postoperative intra-abdominal hemorrhage, intra-abdominal abscess formation, postoperative ileus, pneumonia, delirium, pancreatic fistula, and delayed gastric emptying, were defined according to both the classification system of the International Study Group of Pancreatic Surgery (ISGPF) and the Clavien-Dindo (C-D) classification of surgical complications. 18,19 Overall survival was calculated from the date of surgery for periampullary diseases until the date of death in patients with or without recurrent disease. Postoperative complications and recurrence were examined by abdominal ultrasonography and/or computed tomography. Secondary outcomes included causes of early and late morbidity. We evaluated postoperative cholangitis and gastrointestinal bleeding. Cholangitis was defined according to the TG13 diagnostic guidelines and an acute cholangitis severity rating system after the operation. 20 Known confounding variables included BMI, C-reactive protein/albumin (CRP/Alb) ratio, platelet/lymphocyte ratio (PLR), and neutrophilic/lymphocyte ratio (NLR). The PNI variable represented the influence of the physio-biological value before treatment, with respect to the continuous variable of OS. We used a Cox proportion hazard model, after regulating PLR and CONUT score.

| Statistical analysis
The data in this study were prospectively collected and was retrospectively analyzed by a biostatistician (TI). Survival curves were generated using the Kaplan-Meier method and compared using the log-rank test. Patients alive as of December 2018 were censored at the time of follow-up. Qualitative variables were compared using the chi-square test or Fisher's test, while quantitative variables were analyzed using Student's t test or a nonparametric test, and survival data were determined using a stratified log-rank test. G8 screening tool, IADL score, and CCI were done using Mann-Whitney U test.
We made an addressee receiver operating characteristic curve, and the discrimination of the logistic model equation was decided by calculating a coincident indicator. A covariate-adjusted restricted cubic spline regression analysis with three knots was performed to plot the survival after PD by age and to identify the age at which the survival after PD substantially decreased. This was done to determine the relationship between survival after PD and patient age.

| Patients' demographics
Abbreviations: ASA-PS, American Society of Anesthesiologists physical status; CCI, Charlson comorbidity index; CONUT, controlling nutritional status; IADL, instrumental activities of daily living by Lawton MP.

TA B L E 1 Patient demographics
groups. The octogenarian group showed a significantly worse general condition preoperatively than the younger group, according to the ASA physical status classification system (Table 1). Expectedly, the preoperative value of G8 screening tool in the octogenarian group was significantly reduced than that in the younger group (Table 1). And also, the baseline of IADL score and CCI in octogenarian group were significantly higher compared to those in younger group (Table 1). Preoperative serological C-reactive protein/albumin ratio, and platelet/lymphocyte ratio did not differ between the groups, although the median hemoglobin count, PNI level, and CONUT score were significantly lower and the median NLR was significantly increased in the octogenarian group than in the younger group (Table 1).

| Surgery-related characteristics
Surgery-related demographics are represented in Table 2.
Operation time was significantly shorter in the octogenarian group than in the younger group (median, 279 minutes in the younger group and 248 minutes in the octogenarian group, P = .001). Blood loss was not significantly different between the groups (median, 390 mL in the younger group and 320 mL in the octogenarian group; Table 2). The rate of porto-mesenteric vein resection (PVR) was significantly different between the groups.
We evaluated surgery-related characteristics in octogenarian patients who were divided into two groups: the "early period group" composed of patients who were treated between 2005 and 2012, and the "late period group" composed of patients treated between 2013 and 2018. PD for octogenarian patients was more frequently performed in the late period ( Table 2). There was a significant difference between the two groups in octogenarian patients in terms of pathologic demographics. Interestingly, however, surgical duration (median operative time: 247 minutes in the early period group vs 248 minutes in the late period group) and blood loss volumes (median blood loss volume: 350 mL in the early period group vs 320 mL in the late period group) did not differ significantly between the two groups in octogenarian patients who underwent PD (Table 2).

| Short-term outcomes
There were seven in-hospital deaths (1.5%) among the younger group and no in-hospital deaths (0.0%) among the octogenarian patients. All seven patients died during the primary hospital stay (total mortality, 1.2%). All seven deaths followed a surgical complication and were caused either by pancreatic fistula-related bleeding (n = 4) or sepsis-related multiple organ failure (n = 3). When patients were stratified according to younger or octogenarian groups, no significant difference was found with regard to postoperative complications, including intra-abdominal hemorrhage, abscess formation, postoperative ileus, or pneumonia (   Figure 1A). Figure 1B shows the overall survival curves of the two patient groups following whole PD at the time of data analysis.

| Secondary outcomes
All the mortalities after PD are shown in Table 3

| D ISCUSS I ON
This is the first report addressing how old patients with periampullary diseases may benefit from PD and how to assess the indication and age limitations of patients undergoing PD. In the current study, the remedial treatment achieved superior PD for patients F I G U R E 2 A, Overall, the 1-, 3-, and 5-y survival rates in the younger patients with periampullary neoplasms after surgery were 83.1%, 53.6%, and 44.5%, respectively, while 1-, 3-, and 5-y overall survival rates in octogenarian patients with periampullary neoplasms were 77.9%, 44.1%, and 35.9%, respectively. B, The overall survival in patients with pancreatic adenocarcinoma after curative resection with the intent to cure. The 3-and 5-y survival rates for the younger vs octogenarian group were 46.5% and 35.1%, respectively, vs 30.3% and 25.2%, respectively (P = .045) with periampullary disease who were 80 years of age or older.

TA B L E 3 Mortality after pancreaticoduodenectomy
Furthermore, a meaningful increase in postoperative complications was not found for patients aged 80 years old or over compared with that for those younger than 80 years. However, significantly more patients in the octogenarian group experienced postoperative delirium compared to the younger group, although this complication was not considered to be serious. Several publications over the past decade have reported on the outcomes of PD in patients aged 80 years and older compared to younger patients. 7,9,16 Undoubtedly, the current study is a very pertinent topic considering our aging population and the current major question was which patients should be choosen as proper octogenarians for PD for periampullary diseases and how to do it. Obviously, because of improvements in preoperative diagnostic technology, surgical techniques, and perioperative care, patient treatment has dramatically ameliorated in recent years, thereby reducing the mortality rate and improving postoperative outcomes.
Thus, the incidence of periampullary diseases to offer PD is increasing and will continue to increase as the average population age rises.
In general, age is not the primary consideration for surgical risk; performance status and comorbidities of the patient are more important factors than a patient's age. Geriatric assessment for elderly patients is essential to perform such a surgery with many complications. In our series, the preoperative physical status and concurrent comorbidities in octogenarian patients according to the G8 screening tool and CCI were significantly reduced compared to younger patients. However, from the viewpoint of functioning in the community setting, IADL scale in octogenarian patients might be preserved enough high functioning, although IADL scale score in the octogenarian group was significantly lower than that in the younger group.
Overall, our results suggest that age is not a contraindication for PD and that surgical treatment for octogenarians with periampullary diseases according to our subjective criteria in addition to objective evaluation like G8 screening, IADL, and CCI could be of clinical benefit with acceptable long-term survival results. [11][12][13]21,22 The fact that the ranges of G8 and CCI did not differ greatly between the two groups is considered to be evidence that appropriate patient selection has been performed. This study could also contribute to better-informed decision-making for octogenarian patients and their families.
The immunonutritional and physical status is an important prognostic factor for perioperative outcomes. 23 In our subgroup TA B L E 4 Association of overall survival with physio-biologic characteristics in patients who underwent pancreaticoduodenectomy This study has several potential limitations. First, we acknowledge the presence of a selection bias, since patients included in our study were admitted for surgery at a single institute. Moreover, this study was also a retrospective cohort review of patients undergoing PD that included only patients who underwent a resection. Second, our study focused on the short-and long-term outcomes in patients with periampullary diseases after PD; hence, it is important to consider the postoperative quality of life as a treatment outcome in addition to operative mortality and long-term survival rates. This design inherently represents some selection bias, based on the selection of patients chosen to undergo pancreatectomy, particularly for the older than 80 years group. The current study seems to address a very pertinent topic in our aging population and the current major question aims to determine which octogenarians should benefit from PD for periampullary diseases and how to perform PD for such patients. Therefore, it is necessary to evaluate patient comorbidities, cognitive status, preoperative functional state, and frailty. Furthermore, patients who will benefit from surgical resection should also be chosen according to octogenarian periampullary disease. However, despite these limits, we focus on the importance of PD management in the current study of octogenarians, where treatment regimens were chosen based on patient characteristics. The final decision to perform an operation may depend on the patient's preference, and our results may be presented to patients during informed consent.
In conclusion, our results suggest that age over 80 years should not be a contraindication for PD. Therefore, we are careful not to exaggerate our findings, and octogenarian patients should not be denied surgical opportunities when it is thought that the patient is an ideal candidate for the treatment and the agreement of the family is provided.

ACK N OWLED G EM ENTS
Dr Takehiro Okabayashi is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. We would like to thank Editage (www.edita ge.com) for English language editing.