Oncological outcomes after pancreatoduodenectomy for pancreatic ductal adenocarcinoma in octogenarians: case-control study

Abstract

Background

By the end of this decade, 70 per cent of all diagnosed pancreatic ductal adenocarcinomas will be in the elderly. Surgical resection is the only curative option. In the elderly perioperative mortality is higher, while controversy still exists as to whether aggressive treatment offers any survival benefit. This study aimed to assess the oncological benefit of pancreatoduodenectomy in octogenarians with pancreatic ductal adenocarcinoma.

Method

Retrospective multicentre case-control study of octogenarians and younger controls who underwent pancreatoduodenectomy for pancreatic ductal adenocarcinoma between 2008 and 2017. The primary endpoint was overall survival and the secondary endpoint was disease-free survival.

Results

Overall, 220 patients were included. Although the Charlson co-morbidity index was higher in octogenerians, Eastern Cooperative Oncology Group performance status, ASA and pathological parameters were comparable. Adjuvant therapy was more frequently delivered in the younger group (n = 80, 73 per cent versus n = 58, 53 per cent, P = 0.006). There was no significant difference between octogenarians and controls in overall survival (20 versus 29 months, P = 0.095) or disease-free survival (19 versus 22 months, P = 0.742). On multivariable analysis, age was not an independent predictor of either oncological outcome measured.

Conclusion

Octogenarians with pancreatic ductal adenocarcinoma of the head and uncinate process may benefit from comparable oncological outcomes to younger patients with surgical treatment. Due to the age- and disease-related frailty and co-morbidities, careful preoperative assessment and patient selection is of paramount importance.

By the end of this decade, 70 per cent of all diagnosed pancreatic ductal adenocarcinomas (PDACs) will be in the elderly. Pancreatoduodenectomy is the only curative option, however, the benefit of curative resection in the elderly is difficult to determine from the literature. Octogenarians with PDAC of the head and uncinate process may benefit from comparable oncological outcomes to younger patients with surgical treatment after careful preoperative assessment and patient selection.

Introduction

Pancreatic ductal adenocarcinoma (PDAC) will be the second leading cause of cancer-related mortality by the end of this decade¹. Surgical resection is the only curative option for patients with PDAC, however, only 15–20 per cent are eligible on diagnosis². Oncological outcomes can be substantially improved with completion of the treatment pathway (neoadjuvant and/or adjuvant treatment and surgery)^3,4. Nonetheless, 5-year survival is only 20–25 per cent due to local or metastatic disease recurrence^5,6. Historically, the mortality associated with pancreatoduodenectomy (PD) was high and the concern of chronic co-morbidities and frailty amongst the elderly resulted in their exclusion from surgical treatment⁷. Similarly, the perception of elderly patients being poor candidates to receive adjuvant chemotherapy may prevent them being offered a chance of curative treatment pathways⁸.

Following the population surge after World War II, the ‘baby boomers’ have now become octogenarians⁹ and a combination of better general health and hence longer life expectancy has seen a rise in the age-specific trend of the incidence of PDAC¹⁰. By the end of this decade, 70 per cent of all diagnosed PDAC will be in the elderly¹¹. The benefit of curative resection in this population is difficult to determine from the literature as it is underrepresented within clinical trials. Most studies on the safety and benefit of the elderly population undergoing PD refer to an age range of 70–75 years, but over the next 50 years the proportion of octogenarians within the population in Europe will double from 6 to 13 per cent¹². The advances in surgical technique, technology, perioperative management and centralization have resulted in a substantial improvement in mortality (from 4.1 to 2.4 per cent) and failure to rescue (from 13 to 7.4 per cent) following PD^13,14. Although in a previous study of this group, 90-day mortality of octogenarians after PD was double that of younger controls matched on extent of surgery for periampullary malignancies (9 versus 3 per cent), age was not an independent predictor of mortality¹⁵. Careful patient selection and assessment of co-morbidities are of paramount importance in preoperative planning for these patients. Nonetheless, the evidence in the literature is still controversial as to whether aggressive treatment offers any survival benefit for octogenarians with PDAC. The aim of this study was to assess the oncological benefit of PD for octogenarians with PDAC.

Methods

Study design

A multicentre retrospective case-control analysis of prospectively maintained databases was performed, including PD undertaken over a 10-year interval between January 2008 and December 2017. Octogenarians who underwent PD were matched with consecutively operated younger patients (control group) with a 1:1 ratio, based on extent of surgery (venous, arterial or additional organ resection). An invitation to participate in this study was sent out to all specialist pancreatic centres across the UK. Six centres agreed to participate, resulting in data from a total of seven centres included. Institutional board approval was sought and obtained by each centre separately. Data collection was carried out by each centre using a standardized proforma. The primary outcome set for this study was overall survival (OS; defined as time from diagnosis to death) and the secondary outcome disease-free survival (DFS; defined as time from surgery to diagnosis of recurrence). Reporting of results was performed in line with the STROBE statement¹⁶.

Data collection

The preoperative data collected included: demographics, ASA score, Eastern Cooperative Oncology Group (ECOG) performance status, Charlson co-morbidity index (CCI), co-morbidities, preoperative echocardiogram, pulmonary function test and or cardiopulmonary exercise testing (CPEX), preoperative biliary stenting, steroid use, preoperative haemoglobin, serum albumin, bilirubin and use of neoadjuvant chemotherapy and/or radiotherapy.

Intraoperative data included: type of PD (classic Whipple or pylorus preserving), additional organ resection or venous resection.

Postoperative data included: histological type of tumour, lymph node ratio, resection margin status (defined as R0: no tumour cells for at least 1 mm from margin, R1: tumour cells within 1 mm)¹⁷, presence of perineural and intravascular invasion, 30- and 90-day mortality, complications categorized using the Clavien–Dindo classification, 30-day re-admission, OS and DFS.

Statistical methods

Chi squared with exact statistics and ANOVA were used as appropriate to compare variables and outcomes between the two groups, with statistical significance set at P < 0.050. Survival analysis (OS and DFS) was performed using the Kaplan–Meier method and the log-rank test for comparing differences between survival curves. Univariable and multivariable time to event analyses were performed using the Cox proportional hazard model for OS and DFS. Variables were subjected to a univariable analysis first and those with P < 0.20 were introduced into a multivariable model. Hazard ratios (HR) and associated 95 per cent confidence intervals were calculated. A two-tailed P value <0.050 was considered statistically significant. All statistical analyses were performed using the software package SPSS Statistics for Windows® (version 23.0; SPSS Inc., Chicago, IL, USA).

Results

Cohort characteristics

A total of 220 patients comprising 110 octogenerian and 110 non-octogenerian patients underwent PD (Table 1). The octogenarian cohort was matched on complexity of resection with consecutive patients from a younger cohort where 54 (24.5 per cent) of patients underwent vein resections. Although CCI was higher in octogenerians, ECOG performance status and ASA were comparable. Neoadjuvant therapy was not commonly performed in either cohort, but adjuvant therapy was delivered more commonly in the younger cohort (n = 80 (72.7 per cent) versus n = 58 (52.7 per cent), P = 0.006). There was no difference in the tumour stage, lymph node ratio or resection margin status between the groups. There was no significant difference in the recurrence pattern between the groups.

Table 1.

Comparison of demographic, treatment and pathology characteristics between octogenarians and controls

Demographics	Total n = 220	Controls n = 110	Octogenarians n = 110	P
Age (years), median (range)	79 (36–88)	69 (36–79)	81 (80–86)	<0.001*
Sex				0.135
Male	123 (55.9)	67 (60.9)	56 (50.9)
Female	97 (44.1)	43 (39.1)	54 (49.1)
Charlson co-morbidity index				<0.001*
1–2	6 (2.7)	6 (5.5)	0
3–4	34 (15.5)	34 (30.9)	0
>5	167 (75.9)	62 (56.4)	105 (95.5)
ECOG performance status				0.756
0–1	197 (89.5)	98 (89.1)	99 (90)
2–4	9 (4.1)	4 (3.6)	5 (4.5)
ASA				0.635
I–II	149 (67.7)	77 (70)	72 (65.5)
III–IV	38 (17.3)	18 (16.4)	20 (18.2)
Treatment
Neoadjuvant therapy
Chemotherapy	5 (2.3)	3 (2.7)	2 (1.8)	0.651
Radiotherapy	3 (1.4)	2 (1.8)	1 (0.9)	0.561
Operation				0.786
Whipples	122 (55.5)	60 (54.5)	62 (56.4)
PPPD	98 (44.5)	50 (45.5)	48 (43.6)
Vein resection	54 (24.5)	28 (25.5)	26 (23.6)	0.754
Adjuvant chemotherapy	138 (62.7)	80 (72.7)	58 (52.7)	0.006*
Pathology
T stage				0.271
pT1	8 (3.6)	3 (2.7)	5 (4.6)
pT2	10 (4.5)	7 (6.4)	3 (2.7)
pT3	197 (89.6)	99 (90)	98 (89.1)
pT4	5 (2.3)	1 (0.9)	4 (3.6)
Resection margin				0.499
R0	101 (45.9)	48 (43.6)	53 (48.2)
R1	119 (54.1)	62 (56.4)	57 (51.8)
Lymph node ratio, median (range)	0.15 (0–1)	0.19 (0–1)	0.14 (0–0.75)	0.132
Perineural invasion	187 (85)	93 (84.5)	94 (85.5)	0.278
Perivascular invasion	171 (77.7)	84 (76.4)	87 (79.1)	0.183
Outcomes
30-day mortality	5 (2.3)	2 (1.8)	3 (2.7)	0.636
90-day mortality	13 (5.9)	4 (3.6)	9 (8.2)	0.143

Values are n (%) unless otherwise stated. *P values are significant. ECOG, Eastern Cooperative Oncology Group; PPPD, pylorus-preserving pancreatoduodenectomy.

Survival analysis

There was no significant difference in OS (octogenarians median: 20 months, range: 14–26 months versus controls median: 29 months, range: 24–34 months; P = 0.095) or DFS (octogenarians median: 19 months, range: 13–24 months versus controls median: 22 months, range: 15–29 months; P = 0.742) between octogenarians and controls (Fig. 1).

Fig. 1.

Kaplan–Meier curves a Overall survival and b disease-free survival.

Risk analysis

Overall survival

Table 2 shows the results of the univariable Cox regression analysis for OS. Multivariable analysis identified history of angina/percutaneous coronary intervention/coronary surgery (OR = 3.149; c.i. = 1.351–7.341; P = 0.008), preoperative albumin levels (OR = 0.592; c.i. = 0.383–0.914; P = 0.018) and lymph node ratio (OR = 10.048; c.i. = 3.388–29.801; P < 0.001) as independent predictors of OS. Of note, age was not significant.

Table 2.

Univariable Cox regression analysis for overall survival

Overall survival
Parameters	Univariable analysis: whole cohort		Univariable analysis: octogenarians subgroup
	OR (95% c.i.)	P	OR (95% c.i.)	P
Preoperative
Age  Indicator: control; < 80 years	1.410 (0.935–2.125)	0.101	—	—
Sex  Indicator: female	1.001 (0.663–1.511)	0.996	0.928 (0.529–1.628)	0.793
ASA score  Indicator: class 1	1.326 (0.905–1.945)	0.148	0.993 (0.556–1.774)	0.981
ECOG performance status  Indicator: grade 1	1.126 (0.779–1.629)	0.527	1.040 (0.616–1.756)	0.884
ECOG groups  Indicator: grades 0–1	1.172 (0.428–3.211)	0.757	1.392 (0.335–5.773)	0.649
Diabetes  Indicator: no	1.709 (1.118–2.612)	0.013*	2.054 (1.154–3.654)	0.014*
COPD  Indicator: no	0.502 (0.158–1.594)	0.242	0.507 (0.070–3.688)	0.502
Congestive heart failure  Indicator: no	2.776 (0.864–8.914)	0.086	1.862 (0.244–14.208)	0.549
Myocardial infarction  Indicator: no	1.746 (0.837–3.641)	0.138	1.402 (0.497–3.955)	0.523
Prior PCI/previous coronary surgery/angina  Indicator: no	1.918 (0.918–4.005)	0.083	1.330 (0.318–5.571)	0.696
Hypertension  Indicator: no	1.042 (0.667–1.629)	0.856	1.036 (0.572–1.878)	0.907
Impaired sensorium  Indicator: no	7.052 (0.955–52.057)	0.055	NA	NA
Dementia  Indicator: no	4.778 (0.654–34.885)	0.123	NA	NA
Peripheral vascular disease  Indicator: no	1.680 (0.410–6.889)	0.471	0.048 (0–98692.140)	0.683
TIA / CVA  Indicator: no	2.592 (1.186–5.665)	0.017*	1.683 (0.600–4.723)	0.323
Neurological deficit  Indicator: no	17.073 (2.204–132.244)	0.007*	NA	NA
Connective tissue disease  Indicator: no	0.047 (0–32.697)	0.361	0.046 (0–34.921)	0.363
Peptic ulcer disease  Indicator: no	0.292 (0.040–2.116)	0.223	6.804 (0.890–52.025)	0.065
Liver disease  Indicator: no	0.047 (0–8.606)	0.250	0.046 (0–53.310)	0.393
Hypercoagulability  Indicator: no	2.307 (0.724–7.354)	0.158	1.985 (0.612–6.439)	0.254
Chronic kidney disease  Indicator: stage 1	0.996 (0.927–1.069)	0.902	0.988 (0.908–1.075)	0.780
Biliary stent  Indicator: no	0.813 (0.526–1.256)	0.350	0.485 (0.270–0.872)	0.016*
Steroid use prior to operation  Indicator: no	1.841 (1.011–3.353)	0.046*	4.070 (2.020–8.197)	<0.001*
Preoperative haemoglobin	0.952 (0.846–1.071)	0.414	0.852 (0.698–1.042)	0.118
Preoperative bilirubin	1.001 (0.999–1.003)	0.183	1.001 (0.999–1.003)	0.163
Preoperative albumin	0.668 (0.502–0.890)	0.006*	0.482 (0.323–0.718)	<0.001*
Neoadjuvant chemotherapy  Indicator: no	1.259 (0.309–5.126)	0.748	1.287 (0.176–9.398)	0.803
Neoadjuvant radiotherapy  Indicator: no	0.048 (0–29.994)	0.355	0.049 (0–6709.6999)	0.616
Intraoperative
Classical or PPPD  Indicator: classical	0.819 (0.542–1.239)	0.344	0.635 (0.357–1.129)	0.122
Venous resection  Indicator: no	1.113 (0.687–1.803)	0.664	1.083 (0.563–2.082)	0.811
Additional organ resection  Indicator: no	0.610 (0.224–1.666)	0.335	0.266 (0.036–1.948)	0.192
Histopathological
pT  Indicator: pT1	1.408 (0.792–2.502)	0.244	1.452 (0.692–3.048)	0.324
Lymph node ratio	8.654 (4.038–18.549)	<0.001*	16.928 (4.299–66.656)	<0.001*
Resection margin  Indicator: R0	1.790 (1.161–2.762)	0.008*	2.291 (1.259–4.171)	0.007*
Perineural invasion  Indicator: no	3.055 (1.239–7.533)	0.015*	2.881 (0.698–11.892)	0.143
Intravascular invasion  Indicator: no	1.960 (1.087–3.532)	0.025*	3.353 (1.040–10.808)	0.043*
Postoperative
Postoperative complications  Indicator: no	1.125 (0.746–1.696)	0.575	0.933 (0.529–1.644)	0.810
Clavien–Dindo classification based on higher category recorded  Indicator: grade I	0.989 (0.832–1.175)	0.897	0.926 (0.720–1.191)	0.551
Adjuvant chemotherapy  Indicator: yes	1.763 (1.120–2.775)	0.014*	2.052 (1.115–3.775)	0.021*

*P values are significant. ECOG, Eastern Cooperative Oncology Group; COPD, chronic obstructive pulmonary disease; PCI, percutaneous coronary intervention; TIA, transient ischaemic attack; CVA, cerebrovascular accident; PPP, pylorus preserving pancreatoduodenectomy; NA, not applicable.

In an effort to identify predictors of mortality that can be used to council octogenarian patients, separate multivariable analyses were performed within that subgroup for all parameters and for preoperative parameters only as shown in Table S1. Amongst preoperative parameters history of peptic ulcer disease (OR = 18.502; c.i. = 2.205–155.284; P = 0.007), preoperative use of steroids (OR = 2.440; c.i. = 1.105–5.389; P = 0.027) and preoperative albumin levels (OR = 0.528; c.i. = 0.330–0.844; P = 0.008) were independent predictors of OS. When the postoperative parameters were also included in the model, preoperative haemoglobin levels (OR = 0.653; c.i. = 0.475–0.898; P = 0.009), lymph node ratio (OR = 16.300; c.i. = 3.039–87.419; P = 0.001), type of resection (OR = 0.234; c.i. = 0.100–0.546; P = 0.001) and adjuvant chemotherapy (OR = 2.227; c.i. = 1.004–4.939; P = 0.049) were also significant predictors.

Disease-free survival

Table 3 shows the results of the univariable Cox regression analysis for DFS. Multivariable analysis identified history of angina/percutaneous coronary intervention (PCI)/coronary surgery (OR = 2.394; c.i. = 1.173–4.884; P = 0.016), preoperative albumin levels (OR = 0.625; c.i. = 0.446–0.876; P = 0.006), lymph node ratio (OR = 6.383; c.i. = 2.713–15.020; P < 0.001) and perineural invasion (OR = 7.022; c.i. = 1.684–29.281; P = 0.007) as independent predictors of DFS. Of note, age was not significant.

Table 3.

Univariable Cox regression analysis for disease-free survival

Disease-free survival
Parameters	Univariable analysis: whole cohort		Univariable analysis: octogenarians subgroup
	OR (95% c.i.)	P	OR (95% c.i.)	P
Preoperative
Age  Indicator: control; < 80 years	1.070 (0.710–1.612)	0.746	—	—
Sex  Indicator: female	0.931 (0.618–1.404)	0.735	1.202 (0.660–2.191)	0.548
ASA score  Indicator: class 1	1.044 (0.720–1.515)	0.819	0.638 (0.347–1.172)	0.147
ECOG performance status  Indicator: grade 1	1.086 (0.744–1.586)	0.668	1.160 (0.647–2.078)	0.618
ECOG groups  Indicator: grades 0–1	2.217 (0.895–5.491)	0.085	14.351 (3.707–55.552)	<0.001*
Diabetes  Indicator: no	1.496 (0.957–2.337)	0.077	1.637 (0.838–3.199)	0.149
COPD  Indicator: no	0.953 (0.414–2.191)	0.909	0.578 (0.079–4.232)	0.590
Congestive heart failure  Indicator: no	2.267 (0.710–7.234)	0.167	3.056 (0.384–24.307)	0.291
Myocardial infarction  Indicator: no	1.539 (0.707–3.349)	0.277	0.960 (0.295–3.125)	0.946
Prior PCI / previous coronary surgery / angina  Indicator: no	2.346 (1.167–4.717)	0.017*	2.071 (0.632–6.787)	0.229
Hypertension  Indicator: no	1.357 (0.892–2.065)	0.154	1.218 (0.662–2.239)	0.526
Impaired sensorium  Indicator: no	2.955 (0.408–21.415)	0.284	NA	NA
Dementia  Indicator: no	19.769 (2.504–156.043)	0.005*	NA	NA
Peripheral vascular disease  Indicator: no	1.485 (0.469–4.708)	0.501	1.724 (0.407–7.291)	0.459
TIA / CVA  Indicator: no	1.965 (0.853–4.527)	0.113	2.067 (0.731–5.843)	0.171
Neurological deficit  Indicator: no	0.049 (0–6E + 014)	0.873	NA	NA
Connective tissue disease  Indicator: no	0.550 (0.076–3.975)	0.553	0.727 (0.098–5.371)	0.755
Peptic ulcer disease  Indicator: no	0.295 (0.041–2.119)	0.225	0.049 (0–8E + 011)	0.846
Liver disease  Indicator: no	0.639 (0.157–2.598)	0.531	0.046 (0–43.892)	0.379
Hypercoagulability  Indicator: no	2.287 (0.835–6.266)	0.108	1.808 (0.556–5.876)	0.325
Chronic kidney disease  Indicator: stage 1	1.000 (0.975–1.025)	0.982	0.999 (0.975–1.025)	0.961
Biliary stent  Indicator: no	1.099 (0.718–1.684)	0.663	0.608 (0.330–1.123)	0.112
Steroid use prior to operation  Indicator: no	1.030 (0.473–2.245)	0.940	2.722 (1.128–6.569)	0.026*
Preoperative haemoglobin	0.911 (0.799–1.038)	0.161	0.938 (0.765–1.151)	0.540
Preoperative bilirubin	1.001 (0.999–1.002)	0.568	1.000 (0.998–1.003)	0.736
Preoperative albumin	0.708 (0.536–0.936)	0.015*	0.632 (0.410–0.972)	0.037*
Neoadjuvant chemotherapy  Indicator: no	1.039 (0.256–4.228)	0.957	1.275 (0.174–9.336)	0.811
Neoadjuvant radiotherapy  Indicator: no	0.457 (0.064–3.287)	0.437	0.048 (0–7921.938)	0.621
Intraoperative
Classical or PPPD  Indicator: classical	0.687 (0.453–1.042)	0.078	0.807 (0.441–1.475)	0.486
Venous resection  Indicator: no	1.057 (0.652–1.713)	0.823	0.741 (0.341–1.607)	0.447
Additional organ resection  Indicator: no	0.656 (0.241–1.787)	0.410	0.913 (0.220–3.798)	0.901
Histopathological
pT  Indicator: pT1	1.431 (0.829–2.473)	0.199	1.130 (0.575–2.221)	0.722
Lymph node ratio	7.881 (3.725–16.674)	<0.001*	20.937 (4.727–92.741)	<0.001*
Resection margin  Indicator: R0	1.733 (1.129–2.660)	0.012*	0.807 (0.441–1.475)	0.486
Perineural invasion  Indicator: no	4.202 (1.541–11.455)	0.005*	5.200 (0.714–37.868)	0.104
Intravascular invasion  Indicator: no	2.599 (1.381–4.890)	0.003*	2.334 (0.912–5.973)	0.077
Postoperative
Postoperative complications  Indicator: no	1.288 (0.853–1.944)	0.228	1.201 (0.657–2.196)	0.552
Clavien–Dindo classification based on higher category recorded  Indicator: grade I	1.143 (0.964–1.355)	0.124	1.162 (0.902–1.497)	0.245
Adjuvant chemotherapy  Indicator: yes	1.090 (0.672–1.767)	0.727	0.791 (0.407–1.540)	0.491

*P values are significant. ECOG, Eastern Cooperative Oncology Group; COPD, chronic obstructive pulmonary disease; PCI, percutaneous coronary intervention; TIA, transient ischaemic attack; CVA, cerebrovascular accident; PPP, pylorus preserving pancreatoduodenectomy; NA, not applicable.

In an effort to identify predictors of mortality that can be used to council octogenarian patients, separate multivariable analyses were performed within that subgroup for all parameters and for preoperative parameters only. Amongst preoperative parameters low ECOG (OR = 15.053; c.i. = 3.552–63.794; P < 0.001) and preoperative albumin levels (OR = 0.588; c.i. = 0.381–0.907; P = 0.016) were independent predictors of DFS. When the postoperative parameters were also included in the model, lymph node ratio (OR = 10.704; c.i. = 2.148–53.338; P = 0.004) was also a significant predictor as shown in Table S2.

Discussion

The elderly comprises the fastest expanding portion of Western society and this shift in the population demographics is depicted in the patients being diagnosed with and assessed for treatment for PDAC. Cancer outcomes have taken large strides as a result of centralization and high-volume units bringing together multidisciplinary expertise and advances in perioperative and oncological management^18,19. These overall improvements in time-dependent mortality have also been observed in the elderly, where risk of mortality before and after 2000 has almost halved²⁰. Nonetheless, the elderly population appears to be disadvantaged due to selection bias for aggressive oncological treatment pathways. The reason behind this is the concern over the fitness of this subset of patients and their ability to withstand the effects of systemic treatment and the stress of surgical resection. Furthermore, even if this is achieved, doubt still rests in any oncological benefit that may be produced in respect to the patients’ life expectancy in the 9th decade of their life. The elderly are at a higher risk of mortality from both non-PDAC cancers and non-cancer-related causes^21,22. Therefore, the propriety of curative treatment needs to take the patient's remnant life expectancy, as well as quality of life, into consideration²³. Discerning any survival benefit to octogenarians treated for PDAC from the current literature is difficult. This is due to the small proportion of the elderly patients being included within studies or where the age cut-off excludes a significant proportion of elderly patients^24,25. A vast majority of the literature uses age cut-offs much lower than 80 years to define the ‘elderly’ and yet the median age of patients presenting with PDAC is 72 years, where only 7 per cent are under 50 years^26–28.

The elderly are frequently affected by reduced physiological reserve from chronic co-morbidities^29–31. Compounded by PDAC-induced malnutrition, this may result in frailty syndrome and thus a reduced capacity to withstand major stress such as undergoing a PD³². Evidence suggests that in well selected octogenarians, perioperative complications^15,21–23 and 30-day and index admission mortality are comparable with younger controls¹⁵. Ninety-day mortality though was higher (9 per cent versus 3 per cent) and co-morbidities such as previous cerebrovascular event or history of dementia were identified as independent predictors, pointing to a general decline with possible failure to reverse this in the community after hospital discharge. On the contrary, age was not an independent predictor of mortality in any multivariable model¹⁵. Therefore, meticulous assessment and patient selection is of paramount importance when surgical treatment is considered³³. Additional specialized tests, such as pulmonary function tests, cardiopulmonary exercise testing and echocardiography^33,34, use of frailty scoring systems³⁵ and assessment of sarcopenia^36,37 should all be carefully considered in the preoperative assessment, accepting their limitations.

With regards to oncological outcomes, OS has been reported as shorter in elderly patients undergoing pancreatic resections, however, the treatment selection bias with regards to receiving standard therapies, such as venous resection and adjuvant chemotherapy, was also highlighted²⁰. On the other hand, in subgroup analysis of cohorts that receive adjuvant treatment, no difference in survival was reported³⁸. A study of nonagenarians showed that an OS of 20.4 months could be achieved with multimodal therapy, however, 70 per cent of the study group did not receive multimodal therapy³⁹. In this study, OS was defined as the time from the date of diagnosis to death rather than from the date of surgery to death. The rationale behind this choice stands in the inclusion of patients who received preoperative chemotherapy, as well as taking into account the differences in the time between diagnosis and surgery or first treatment among the patients. Any disease progression or stability during this time is also a measure of disease biology that primarily affects OS. There was no significant difference in OS or DFS between octogenarians and controls. The recorded OS in the octogenarian subgroup of 20 months is consistent with the range of 15–30 months in the published literature^40,41. Similarly, there is also no demonstrable difference in DFS between the elderly and younger groups where use of adjuvant therapy has been shown to be an independent prognostic variable^41,42. On risk analysis, history of cardiac co-morbidities and low preoperative albumin were identified as independent predictors of both OS and DFS. High lymph node ratio and perineural were also identified as predictors of oncological outcomes (OS and DFS respectively). The significance of low preoperative albumin levels persisted after analysis of only preoperative variables. Low albumin levels may indicate a status of relative malnutrition which would predispose the patient to prolonged hospitalization and recovery⁴³ and in turn reduce the opportunity to be offered any adjuvant treatment. Similarly, cardiac history may prevent or limit the use of systemic treatment due to the described chemotherapy-related cardiotoxicities⁴⁴. The fact that elderly patients are in general less likely to receive any systemic treatment has been documented by various studies⁴⁵ (possibly due to the concerns regarding tolerability of multiagent regimens), even though evidence supports their safety and efficacy resulting in comparable survival to younger cohorts^8,46,47. Elderly patients are more likely to receive reduced doses of adjuvant chemotherapy and only in the presence of lymph-node-positive disease⁴⁰, a treatment selection bias which is not present in younger patients²².

Limitations of this study include its retrospective nature and the consequent inability to also assess parameters that have not been captured such as frailty and quality of life, type of systemic treatment and follow-up. Details of cause of death after hospital discharge were lacking, hence the inability for any non-cancer-related mortality to be assessed. Nonetheless, this is a multicentre study over a 10-year interval on a large cohort of patients that are underrepresented in the published surgical literature for the management of PDAC. Matching for the extent of surgery was also utilized to account for any possible intraoperative selection bias for utilizing a more radical surgical technique.

In summary, octogenarians with PDAC of the head and uncinate process benefit from comparable oncological outcomes to younger patients with surgical treatment after careful preoperative assessment and patient selection, which should be the focus of future studies.

Funding

The authors have no funding to declare.

Disclosure

The authors declare no conflict of interest.

Supplementary material

Supplementary material is available at BJS Open online.

Data availability

The data sets generated during and/or analysed during the present study are available from the corresponding author on reasonable request.

Author contributions

Rupaly Pande (Data curation, Formal analysis, Writing—original draft), Joseph Attard (Data curation, Formal analysis, Writing—review & editing), Bilal Al-Sarireh (Data curation, Writing—review & editing), Ricky Bhogal (Data curation, Writing—review & editing), Alexia Farrugia (Data curation, Writing—review & editing), Giuseppe Fusai (Data curation, Writing—review & editing), Simon Harper (Data curation, Writing—review & editing), Camila Hidalgo Salinas (Data curation, Writing—review & editing), Asif Jah (Data curation, Writing—review & editing), Gabriele Marangoni (Data curation, Writing—review & editing), Matthew Mortimer (Data curation, Writing—review & editing), Michael Pizanias (Data curation, Writing—review & editing), Andreas Prachalias (Data curation, Writing—review & editing), Keith Roberts (Data curation, Writing—review & editing), Chloe Sew Hee (Data curation, Writing—review & editing), Fiammetta Soggiu (Data curation, Writing—review & editing), Parthi Srinivasan (Data curation, Writing—review & editing) and Nikolaos Chatzizacharias (Conceptualization, Data curation, Formal analysis, Methodology, Supervision, Writing—review & editing).