Nutritional assessment and surgical outcomes in very elderly patients undergoing pancreaticoduodenectomy: a retrospective study

To evaluate and compare the nutritional factors and clinical outcomes of pancreaticoduodenectomy between elderly and non-elderly patients. This retrospective study evaluated 122 consecutive patients who underwent pancreaticoduodenectomy from April 2008 to April 2020. Preoperative and postoperative nutritional factors (prognostic nutritional index), complication rates, and survival rates were compared between the elderly (≥ 80 years) and non-elderly (< 80 years) patient groups. Changes in nutrition markers were evaluated before surgery to 1 year after surgery. A total of 20 elderly patients (16.4%) and 102 non-elderly patients (83.6%) underwent pancreaticoduodenectomy. Elderly patients had a significantly lower preoperative prognostic nutritional index than did non-elderly patients. At 3 months postoperatively, elderly patients had a lower albumin level and prognostic nutritional index. The median length of hospital stay was significantly longer (39.9 vs. 27 days, P = 0.004), the rate of death due to other diseases was higher, and the overall survival rate was significantly lower (1-/3-/5 year overall survival rates: 78.1%/26.7%/13.3% vs. 87.1%/54.4%/46.7%; log-rank test, P = 0.003) in the elderly group than in the non-elderly group. The results suggest that careful patient selection and optimal perioperative care are necessary to determine whether pancreaticoduodenectomy is indicated for elderly patients.


Introduction
Pancreaticoduodenectomy (PD) offers a chance for cure in patients with pancreatic cancer involving the head of the pancreas and other periampullary malignancies [1,2]. Furthermore, PD is a valid treatment option for selected patients with benign tumors and non-cancerous conditions, including chronic pancreatitis. The overall morbidity and mortality associated with PD range from 45 to 52% [3,4] and from 0 to 5% [5,6], respectively.
The population of individuals aged > 80 years is increasing globally, accounting for approximately 1% of the total population, and it is predicted to increase four-fold by 2050 [7]. There will be an increasing need to consider patients older than 80 years of age for PD, particularly for the treatment of pancreatic cancer, for which the incidence increases with age [7].
The outcomes of PD have improved with better surgical techniques and postoperative care of patients [7]. However, performing this surgery in elderly patients is still a controversial issue. In many studies regarding PD, the elderly cohort accounted for a very small proportion of the study population. There may be some selection bias for the indication of PD. Compared to non-elderly patients, elderly patients are usually frail and malnourished and have more common diseases.
The morbidity and mortality after PD are associated with malnutrition. Nutritional support may reduce postoperative complications in patients undergoing PD [8]. A low preoperative albumin level has been associated with an increased risk of morbidity and mortality in patients undergoing PD [9]. The prognostic nutritional index (PNI) and platelet-to-lymphocyte ratio are useful prognostic indicators for pancreatic cancer patients after PD [10]. However, the association between an elderly age and the nutrition status is unclear during the perioperative period. There exist few reports on the postoperative changes in the nutritional indexes in elderly patients.
This study aimed to evaluate and compare the nutritional factors and clinical outcomes of PD between elderly (age ≥ 80 years) and non-elderly (age < 80 years) patients as well as to examine the effect of PD on changes in nutritional indexes to monitor patients during postoperative follow-up. As appropriate preoperative evaluation of elderly patients will lead to their safe management, we investigated the nutritional parameters, including the albumin level, hemoglobin level, platelet count, lymphocyte count, and PNI, before surgery to 1 year after PD [11].

Ethical statements
The patients were not required to provide their informed consent for the study, because the analysis used anonymous data that were obtained after patients agreed to treatment by written consent. All human studies have been approved by the appropriate ethics committee (approval number: 0138) and have, therefore, been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.

Patients, data collections, and definitions
This retrospective study included 122 consecutive patients who underwent PD in a single institution from April 2008 to April 2020.
We compared preoperative and postoperative nutritional factors, PNI, complication rates, and survival rates between the elderly (age ≥ 80 years) and non-elderly (age < 80 years) patient groups. Preoperative blood samples were obtained within 2 weeks before surgery. The following preoperative patient data were collected: sex, age, hypertension, diabetes, alcohol consumption, smoking, body mass index (BMI), use of anticoagulants, indication for surgery, main pancreatic duct (MPD) diameter by preoperative computed tomography or magnetic resonance cholangiopancreatography, blood transfusion, blood loss, operative time, pancreaticojejunostomy technique, and tumor stage as per the Union for international cancer control (UICC) classification [12].
Regarding the nutritional assessment, perioperative timedependent changes in the nutrition markers, including the serum albumin level, hemoglobin level, lymphocyte count, and PNI score, were evaluated from before surgery to 1 year after surgery (at 1 week, 1 month, 3 months, and 1 year postoperatively) and compared between the two groups. PNI was calculated according to the following formula: 10 × serum albumin level (g/dL) + 0.05 × total lymphocyte count (per mm 3 ).
All specimens were histopathologically identified as and categorized into either malignant or benign tumors. A microscopic positive margin (R1) was defined as the presence of a tumor at the surgical margin at the time of histological examination.
Postoperative complications were graded according to the Clavien-Dindo classification [13], which was validated in pancreatic surgery. Grade 3 or 4 complications were considered to be major complications. A pancreatic fistula was diagnosed and graded in accordance with the International Study Group on Pancreatic Fistula classification [14,15]. Perioperative mortality was defined as death within 30 days from operation or during hospitalization.
We analyzed whether the preoperative variables including the nutritional status might be related to the development of postoperative complication after PD. Furthermore, we analyzed the prognostic factors including the nutrition status in all patients and in elderly patients, respectively.

Operative procedure
The patients underwent subtotal stomach-preserving PD, which was performed by two surgeons who specialized in pancreatic surgery. All operations were performed via an open approach, and the degree of locoregional lymphadenectomy was determined according to the patient's preoperative diagnosis. Surgical reconstruction was performed using a modification of the Child method [16]. The proximal jejunal stump was passed through the retrocolic pathway, and pancreaticojejunostomy, biliojejunostomy, and gastrojejunostomy were subsequently performed. Pancreaticojejunostomy was performed using the modified Kakita anastomosis (n = 47; April 2008 to May 2013) or the modified Blumgart anastomosis (n = 75; June 2013 to present) [17]. During the procedure, the surgeon decided whether plastic stents for internal drainage needed to be inserted into the MPD. Two or three abdominal drains were placed anteriorly or posteriorly to the pancreaticojejunostomy anastomosis and hepaticojejunostomy anastomosis.

Statistical analysis
Continuous data are expressed as the mean ± standard deviation. Statistical analyses were conducted using the unpaired Student's t test and the chi-squared test with Fisher's exact test. Overall patient survival was evaluated using the Kaplan-Meier method and compared with the log-rank test.

3
The optimal cutoff values were determined by maximizing the Youden index (sensitivity + specificity − 1) [18]. Univariate and multivariate analyses were performed to determine the prognostic factors using the Cox proportional hazard model, and significantly different variables analyzed by the univariate analysis were further analyzed by the multiple Cox proportional hazards model. A P value < 0.05 was considered to be statistically significant. The statistical analysis was performed using the JMP software program version 9 (SAS Institute, Cary, NC, USA).

Patients' characteristics
The patients' clinicopathologic characteristics according to age are summarized in Table 1. Of the 122 patients who underwent PD, 20 (16.4%) were 80 years or older of age. The elderly group had more poor-risk patients with respect to the American Society of Anesthesiologists (ASA) score (P = 0.002). Elderly patients were significantly more likely to have cardiac disease including coronary artery disease and arrhythmia than were non-elderly patients (P = 0.038). No significant differences in sex, BMI, drinking history, smoking history, and use of anticoagulants were found between the two groups. Regarding the pathologic results, all elderly patients undergoing PD had malignancy, whereas 78.4% of non-elderly patients had malignancy (P = 0.003). Furthermore, more elderly patients underwent R1 resection than did non-elderly patients (P = 0.027).

Operative outcomes
No significant differences in intraoperative factors (surgical procedures, blood loss, operative time, and transfusion) were noted between the groups ( Table 2). Portal vein resection was performed in two patients (10%), both of whom had pancreatic cancer.

Postoperative outcome
The postoperative complications are presented in Table 2. There was no significant difference in complications and major complications (Clavien-Dindo classification ≥ III) between the groups. The incidence of postoperative pancreatic fistula (POPF) was similar between the groups. There was no significant association between preoperative nutritional status and postoperative complications ( Table 3). The duration of in-hospital stay was significantly longer in the elderly group than in the non-elderly group (P = 0.004). UICC staging distribution was not significantly different between the groups (P = 0.554). In the elderly group, one patient (5%) died due to abdominal bleeding associated with POPF within 30 days after surgery. On the other hand, three patients (2.9%) died within 30 days after surgery in the nonelderly group: two patients died due to aspiration pneumonia and sepsis associated with POPF and one patient died due to pulmonary thromboembolism. There was no significant difference in the mortality rate between the groups (P = 0.636). The rate of postoperative adjuvant chemotherapy for pancreatic cancer was lower in the elderly group than in the nonelderly group, although this was not significant [4 (57.1%) vs. 36 (80.0%); P = 0.182].

Change in the nutritional indexes
The preoperative albumin level and PNI were significantly lower in the elderly group than in the non-elderly group ( Table 1). No differences in the preoperative hemoglobin level and platelet count were observed between the groups. The change in the nutritional indexes examined before and after surgery is shown in Fig. 1a-f. Compared to the nonelderly group, the elderly group showed significant decreases in the lymphocyte count before surgery; albumin level before surgery and at 1 week, 1 month, 3 months, 6 months, and 1 year postoperatively; and PNI before surgery and at 3 months postoperatively. Those results may indicate that elderly patients had little protein reserves and impaired immune defenses compared to non-elderly patients.

Long-term survival
The median overall survival was 17.5 months in the elderly group and 45.7 months in the non-elderly group. Additionally, the 1-, 3-, and 5 year survival rates were 78.1%, 26.7%, and 13.3%, respectively, in the elderly group and 87.3%, 54.4%, and 46.7%, respectively, in the non-elderly group. Elderly patients had a significantly lower overall survival rate than did non-elderly patients (log-rank test, P = 0.003) (Fig. 2). When the population was categorized according to malignant tumors, including pancreatic cancer, bile duct cancer, ampullary cancer, and gallbladder cancer, survival did not significantly differ between the groups (P = 0.099) (Fig. 3). The median overall survival times were 17.5 months for the elderly patients with malignant tumors and 30.0 months for the non-elderly patients with malignant tumors. Table 4 shows the relationship between the clinicopathological variables and the overall survival rate following PD in all patients. According to a univariate analysis, the overall survival was significantly worse in patients aged ≥ 80 years (P = 0.003) and those with PNI < 45 (P = 0.002), operative time ≥ 500 min (P = 0.005), and microscopic positive margin (P < 0.001). According to a multivariate analysis, PNI < 45 (P = 0.028), operative time ≥ 500 min (P = 0.005), and microscopic positive margin 1 3 (P = 0.004) were independent and significant predictors of overall survival. Table 5 shows the relationship between the clinicopathological variables and overall survival rate following PD in elderly patients. In the univariate analysis, overall survival was significantly worse in patients with PNI < 46 (P < 0.001).
In the multivariate analysis, PNI < 46 (P = 0.011) was found to be an independent and significant predictor of overall survival.

Causes of death
We also analyzed the causes of death. Of the 20 elderly patients, nine died of disease recurrence, whereas six  (Table 2). Of the 102 non-elderly patients, 32 died of disease recurrence, whereas 14 died of other causes. The rate of death due to other causes was higher in the elderly group than in the non-elderly group, although this was not significant [6 (30.0%) vs. 14 (13.7%); P = 0.072]. Other causes of death in the elderly group were surgical complications including malnutrition in two patients, abdominal bleeding in one patient, cholangitis in one patient, comorbidities or non-specific complications including myocardial infarction in one patient, and unknown complications in one patient. Other causes of death in the non-elderly patients were surgical complications including aspiration pneumonia in six patients, sepsis in one patient, pulmonary thromboembolism in one patient, portal vein obstruction in one patient, dyspepsia in one patient, malnutrition in one patients, cholangitis caused by stenosis of biliojejunostomy in one patient, comorbidities or non-specific complications including myocardial infarction in one patient, and other cancer in one patient.
We analyzed whether there were any differences in the preoperative nutritional markers between the patients with non-cancer-specific death and those with cancer-specific death ( Table 6). The patients with non-cancer-specific death had a significantly lower serum albumin level, lower PNI, and higher CRP level than those with cancer-specific death.

Discussion
There have been an increasing number of publications reporting on the outcomes of PD in patients aged 80 years and older instead of younger patients [3,[19][20][21][22]. A systematic review and meta-analysis indicated that compared to younger patients, elderly patients had a two-fold increased mortality rate [23]. The cause of death was more closely related to the patients' comorbidities or non-specific surgical complications than other causes [7]. In our study, elderly patients had significantly more cardiac disease than did nonelderly patients. The cause of death was more closely related to the patients' comorbidities or non-specific surgical complications in elderly patients than in non-elderly patients, although this was not significant ( Table 2). Patients with non-cancer-specific death had a significantly lower nutritional status than those with cancer-specific death ( Table 6). This finding suggests that a poor functional status, such as one's nutritional status and medical comorbidities, rather than age alone, contributes to the increased postoperative mortality in elderly patients.
The morbidity associated with PD that is related to specific and non-specific surgical complications remains high. A systematic review showed that patients aged 80 years or older were 1.5 fold more likely to have a complication than their younger counterparts [23]. However, with regard to complications including POPF and operative factors, we found no differences between the groups, which is in line with the findings reported in other published studies [21,24,25]. The preoperative nutritional status, which was poor in elderly patients, was not associated with major postoperative complication. Our experience shows that PD can be performed safely in elderly patients. The perioperative mortality rate in the elderly group might have been relatively high, although there was no difference between the two groups. Supice et al. described that PD in elderly patients appeared to be associated with an increased risk of postoperative death caused by postoperative hemorrhage due to POPF [26]. The reason for the relatively high mortality rate in our study, despite similar rates of POPF, is unclear. The presence of comorbidities could have contributed to this high mortality rate: all patients who died within 30 days after surgery had comorbidities; however, there was no significant association between complications and comorbidity because of the small sample size. The long-term survival was compared between patients aged > 80 years and younger patients in several previous studies. There appears to be a major variation in survival after PD in these studies, which may be related to the inclusion of different pathologies in the analysis [6,27]. Other studies have reported a higher mortality associated with PD in patients aged 80 years and older, which often reflects the higher ASA status [7,28]. In our study, overall survival was significantly decreased in elderly patients, as compared to that in non-elderly patients. In all patients, the independent prognostic factors included low PNI as a nutritional marker, longer operative time, and microscopic positive margin, though age was not a significant factor. Overall survival also decreased in elderly patients treated for a malignant tumor, although this was not significant, because more elderly patients underwent R1 resection than Fig. 2 Overall survival curves for the entire study population comparing elderly patients undergoing pancreaticoduodenectomy with non-elderly patients. The median overall survival was 17.5 months in the elderly group and 45.7 months in the non-elderly group. Additionally, the 1-, 3-, and 5 year survival rates in the elderly group were 78.1%, 26.7%, and 13.3%, respectively, and 87.3%, 54.4%, and 46.7%, respectively, in the non-elderly group. Elderly patients had a significantly lower overall survival rate than non-elderly patients (logrank test, P = 0.008)  did non-elderly patients. Additionally, elderly patients had a lower rate of postoperative adjuvant chemotherapy, which is known to delay the recovery after PD, than did non-elderly patients, although this was not significant. Additionally, patients with malignant disease underwent dissection of the pancreatic nerve plexus around the superior mesenteric artery. As all patients in the elderly group had malignant diseases, this factor might have influenced the slow recovery of the nutritional status. We showed the preoperative PNI as a nutritional marker to be an independent prognostic factor in elderly patients. This result revealed that an improvement of the nutritional status was one of the factors of a prolonged survival after PD. A reduction in survival among elderly patients may be related to the age-related predicted lower life expectancy of those aged 80 years and older. The average life expectancy is increasing worldwide, with men and women aged 80 years and older being expected to live for another 10 years; furthermore, the life expectancy of people aged 65 years is double that of those aged 80 years and older. Considering those reasons as to why the overall survival decreased in elderly patients, PD may be a therapeutic option, especially in elderly patients with a good nutritional status.
In several studies, the length of hospital stay was not different between the elderly and younger groups [24,28,29]. It should be noted that elderly patients were far more likely to be discharged to a rehabilitation facility instead of home than were younger patients [30,31]. In this study, the length of hospital stay was significantly longer in the elderly group than in the non-elderly group; nonetheless, the postoperative complication rate was similar between the two groups. This finding may be reflective of fewer rehabilitation facilities in our rural region than in an urban region.
The nutritional assessment of our study showed that there was a significant decrease in the albumin level, PNI, and delay of recovery after PD in elderly patients compared to non-elderly patients. Regarding the preoperative status, the serum albumin level and PNI were already found to be significantly worse in the elderly group than in the non-elderly group. Those results showed that elderly patients had small protein reserves and impaired immune systems. Hypoalbuminemia is linked to poor tissue healing of surgical wounds or anastomoses, decreased collagen synthesis, delayed return of bowel function, and impairment of the cell-mediated immune response [32,33]. Therefore, wound infection and remote infections, such as pneumonia and anastomotic leakage, are commonly observed in hypoalbuminemic patients. In addition, the serum albumin concentration is related to the suppression of the systemic inflammatory response [34,35]. The postoperative systemic immune response induced by the invasiveness of PD includes many cytokines. As a result, an immunonutritional disorder causes a decrease in the albumin concentration and total lymphocyte count [36][37][38][39]. Therefore, in elderly patients with small protein reserves and impaired immune systems, a compromised immunonutritional status is an important factor that can lead to increased postoperative complications after PD.
At 3 months after PD, PNI was significantly lower in the elderly group than in the non-elderly group. However, Miyazaki et al. reported that with regard to the change in PNI during the postoperative period, there was no difference between elderly patients (> 70 years) and younger patients, although the definition of elderly patients in their study was different from that in our study [40]. Yamashita et al. reported that there was no significant difference in the changes in the serum albumin level at 3 months after PD between the elderly patients (> 75 years) and younger patients, but the recovery of the serum prealbumin level from 1 to 3 months after PD in patients aged > 75 years was significantly delayed compared to that in younger patients [41]. Furthermore, we assessed the change in nutritional status in the long term. At 1 year after PD, the albumin level  was significantly lower in elderly patients than in non-elderly patients. Elderly patients carry the risk for a prolonged recovery or malnutrition. Nevertheless, there was no difference between the groups with respect to other nutritional and immune markers in elderly patients; thus, further analysis is necessary.
Elderly patients have a lower physiologic reserve than do younger patients [42]. Moreover, this decrease is more likely to be related to the higher incidence of comorbidities in elderly patients than in younger patients. Recently, the enhanced recovery after surgery (ERAS) program has been reported to be safe and it has contributed to decreasing the total complication rates and length of hospital stay in the hepatobiliary field [43,44]. The underlying principle of the ERAS program is a multimodal perioperative protocol to attenuate the inflammatory response and potentiate patient rehabilitation after surgery [45,46]. The adaptation of the ERAS program may be effective in elderly patients who have a low nutritional status to prevent a decline in their nutritional status and activities of daily living after surgery. In patients undergoing PD, the ERAS program was reported to be safe, as it decreased the total complications ASA American Society of Anesthesiologists, BMI body mass index, CI confidence interval, CRP C-reactive protein, CT computed tomography, HR hazard ratio, MPD main pancreatic duct, PLR platelet-to-lymphocyte ratio, PNI prognostic nutritional index and improved the clinical outcomes [47,48]. Additionally, in elderly patients undergoing PD, ERAS seems to be feasible and safe [49]. In this study, the elderly patients started a solid food diet by postoperative day 4. The ERAS program involves early oral feeding that is not always feasible after PD [50]. Delayed gastric emptying or intestinal paralysis may lead to insufficient postoperative caloric intake after PD. To compensate for deficiencies in postoperative caloric intake, early combined parenteral and enteral nutrition for PD has been recommended [51]. This method is safe and improves outcomes. Considering the poor survival benefit of elderly patients, it is essential to carefully select candidates for PD among elderly patients based on strict indications. As a result of an independent prognostic factor analysis, we clarified PNI ≥ 46 as a benefit of the selection criteria. Our results may be in part explained by some selection bias: we excluded patients with a very high surgical risk (ASA 4), especially those with cardiovascular or pulmonary disease, from PD. However, there are currently no clear guidelines indicating the appropriate selection of elderly patients for PD.
The present study is associated with some limitations attributable to its small sample size and the inherent nature of retrospective analysis. Prospective cohort studies conducted at multiple institutions should, therefore, be performed to confirm our study results.

Conclusions
Although age seems to be associated with an increase in mortality, it does not appear to be prohibitive and somewhat reflects a higher grade of pre-existing comorbidities and a low nutritional status. The careful selection of elderly patients for PD based on their preoperative status, including their nutritional status, should, therefore, guide the surgical decision making process for this patient population.