Background: To compare adjuvant chemotherapy completion rates in patient’s post-pancreaticoduodenectomy (PD) with and without exposure to the Enhanced Recovery After Surgery (ERAS) protocol.
Methods: A retrospective cohort study was conducted in all PD patients with adenocarcinomas, who commenced adjuvant chemotherapy and either had the pre-ERAS (n=19) or ERAS (n=66) protocol between 2009 and 2017. Data was analysed via Chi-square, logistic regression and Kaplan–Meier analysis.
Results: Pre-ERAS and ERAS patients displayed similar chemotherapy completion rates (79%). However, ERAS patients had higher America Society of Anaesthesiology (ASA) Physical Status scores (p=0.04) and Charlson Comorbidity Index (CCI) scores (p=0.004) as well as approaching statistical significantly different Clavien-Dindo (CD) scores (p=0.052), compared to pre-ERAS patients. CD score was the only independent variable that statistically influenced chemotherapy completion (p=0.02). Chi-square cross tabulation showed the ERAS group had a higher proportion of CD minor and major complications completing chemotherapy (86.4 and 63.6%, respectively), compared to non-completion (13.6% and 36.4%, respectively) (p=0.033). On Kaplan-Meier analysis, ERAS showed a trend of increased 5-year survival (56%) when compared to pre-ERAS (36%) (p=0.36).
Conclusions: This is the first study that shows ERAS patients displaying similar chemotherapy completion rates compared to the pre-ERAS group (79%), despite having higher ASA, CCI and CD scores. While CD score was found to be the only independent variable influencing chemotherapy completion, further studies are required. The ERAS protocol may help significantly more complex patients complete multimodal therapy including chemotherapy. Overall, the ERAS protocol implemented at FSH hospital is worth considering as a standardised protocol in future trials.
Keywords: Chemotherapy completion; ERAS; Pancreaticoduodenectomy
List of Abbreviations: ASA: American Society of Anaesthesiologists; AJCC: The American Joint Committee on Cancer; BMI: Body Mass Index; CCI: Charlson Comorbidity Index; CD: Clavien-Dindo Post-Operative Complications; ERAS: Enhance Recovery After Surgery; PSC: Physical Status Score; PD: Pancreaticoduodenectomy
Pancreatic cancer is one of the leading causes of cancer relate death in Australia. In 2018 more than 3000 death occurred ranking 4th in all age groups . The overall 5- year survival rate for all stages pancreatic malignancy is only 9 % . Pancreaticoduodenectomy (PD) remains the only curative treatment option for head of pancreas lesions . The survival advantage of Multimodal Therapy (MMT) which includes surgery, chemotherapy and in some cases, radiotherapy has been consistently demonstrated in multiple randomised trials and prospective analysis [4-12]. There are multiple factors precluding patients from completion of MMT. Postoperative complications and frailty are the leading causes for patients no initiating or completing chemotherapy [13-15]. Identification of modifiable factors that lower surgical morbidity, increase chemotherapy completion rates and potentially increase overall survival are a major focus in the field of pancreatic cancer surgery.
Enhanced Recovery After Surgery (ERAS) is a package of interventions aimed at decreasing surgical morbidity, lowering complication rates and decreasing length of stay in hospital. These protocols aim to minimize surgical trauma and postoperative pain, reduce complications, improve outcomes, and decrease LOS, while expediting recovery following elective procedures [16,17]. Currently, there is no unified ERAS protocol apart from general guidelines for PD patients . Since ERAS aims to expedite recovery and reduce morbidity, this may aid in successful completion of adjuvant chemotherapy. We hypothesise that the FSH ERAS protocol may help patients post pancreaticoduodenectomy complete adjuvant chemotherapy.
Materials and Methods
The study was initiated by The Upper Gastrointestinal Surgical Unit, South Metropolitan Health Service, Perth, WA. Ethical approval was obtained from the Human Research Ethics Committees of the South Metropolitan Health Service (HREC Ref: 15-040-1) and University of Notre Dame Australia (HREC Ref: 018072F).
Patient Selection and Data Collection
A retrospective cohort study was conducted in all patients who underwent Pancreaticoduodenectomy (PD) through the Upper Gastrointestinal unit of the Western Australian South Metropolitan Health Service from 2009 to 2017 (n=169) across two sites (Fremantle hospital prior to 2015, and Fiona Stanley hospital form 2015 on wards). The pre-ERAS group (n=29) consisted of patients at Fremantle Hospital between January 2009 to December 2011. In accordance with the ERAS society guidelines (see Supplementary Figure S1, Kristoffer, 2013), the care of patients undergoing PD were modified and adapted as the early-ERAS programme. A total of 140 consecutive patients were included in the analysis (67 from Fremantle hospital and 73 from Fiona Stanley hospital). All operations were performed by the same team of pancreatic surgeons in these hospitals, comprising two fellowship-trained hepatobiliary consultant surgeons at Fremantle Hospital with the addition of a third hepatobiliary surgeon at Fiona Stanley Hospital.
Only patients with resectable pancreatic adenocarcinoma were included in our study. Other histological variants were excluded such as pancreatic neuroendocrine tumours and intraductal papillary mucinous neoplasm. Other inclusion criteria included being referred and commencement adjuvant chemotherapy and being subject to the pre-ERAS or ERAS protocol (summarised in Table 1). Any chemotherapy agents were part of the inclusion criteria. These were categorised under Gemcitabine (Gemzar), 5-Fluorouracil (5-FU), oxaliplatin (Eloxatin), Albumin-bound paclitaxel (Abraxane), Capecitabine (Xeloda), Cisplatin, Paclitacel (Taxol), Leucovirin, Irinotecan (Onivyde), Folic acid or other. Patients who were not referred or referred but did not commence adjuvant chemotherapy were excluded.
PPI: Proton Pump Inhibitor; IDDM: Insulin Dependent Diabetes Mellitus; NGT: Nasogastric Tube; NJT: Nasojejunal Tube; PCA: Patient-Controlled Analgesia; IV: Intravenous; CSL: Compound Sodium Lactate; HDU: High Dependency Unit; BD: Bis in Die (twice daily); NF: Nourishing Fluids; APS: Acute Pain Service; POPF: Postoperative Pancreatic Fistula. * If no sign of ileus .
Clinical factors were obtained and documented for each patient. These included age, gender, body mass index (BMI), American Society of Anaesthesiologists (ASA) physical status score , Charlson Comorbidity Index (CCI) score , and tumour histopathology. Neoadjuvant chemotherapy was not recorded in our analysis, however, there were limited numbers in both the pre-ERAS (n=0) and ERAS (n=12) groups.Postoperative variables were assessed for each patient, including Clavien-Dindo post-operative complications (CD) (Dindo + Clavien, 2004), the American Joint Committee on Cancer (AJCC) staging system [22,23], chemotherapy agents, chemotherapy completion rates and the adherence to the ERAS protocol (Supplementary Figure S1) based on the goals of care. A CD score of 3 or more was considered to be a major complication
Primary and Secondary Outcomes
The primary outcome was the completion of all cycles of chemotherapy. A cycle of chemotherapy was defined by type of chemotherapy protocol. The agents and number of doses which defined a chemotherapy cycle varied thorough the duration. This was due to evolving protocols based on evidence based medicine throughout the study. The secondary outcomes included factors relating to chemotherapy completion. These included ASA scores, CCI scores, CD scores, AJCC staging system, referral or non-referral for chemotherapy (including reasons and numbers of incomplete cycles). The reasons for incompletion include patient factors, physician factors, progression of disease and unknown. Post-operative complications were associated with physician and patient factors.
Descriptive analyses were conducted including frequencies and percentages for categorical variables and mean and Standard Deviation (SD) for continuous variables. Patient demographics, clinical features and chemotherapy completion were compared between the pre-ERAS and ERAS groups. Continuous variables were compared by t-tests and categorical variables by Chi-square tests. Additional analysis (t-tests and Chi-square tests) for other factors associated with chemotherapy completion were performed. Logistic regression was performed to adjust for covariates and confounders between the chemotherapy completion and chemotherapy non-completion. The variables included were age, gender, ASA score, CCI score, CD score, AJCC classification and pre-ERAS or ERAS protocol. All variables were entered simultaneously. Survivability between pre-ERAS and ERAS was performed via Kaplan– Meier analyses.
All statistical tests were two-tailed and P<0.05 were considered statistically significant. All statistical analyses were performed using Statistical Package for the Social Sciences, version 16.0 (SPSS).
A total of 169 patients had undergone pancreaticoduodenectomy between 2009 and 2017. Only 118 patients had pancreatic adenocarcinomas, with 23 in pre-ERAS and 95 in ERAS groups. Chemotherapy data was not available for 16 patients electing to have chemotherapy in out of district or private institutions or not at all due to side effects (Figure 1). The final patient data for analysis comprised 19 and 66 patients in the pre-ERAS and ERAS groups, respectively. Adjuvant chemotherapy completion rates were similar between pre-ERAS and ERAS groups, being 79%(p=0.98) (Figure 1). Compliance to the ERAS protocol in the ERAS group, was confirmed in a previous study . The Pre ERAS group had a lower ASA grade as compared to the ERAS group. 15.8% of pre-ERAS patients had ASA scores of 3-4 as compared to 52% of ERAS patients (p=0.04). Furthermore, 31.58% of the pre-ERAS patients had CCI scores of 4-6, compared to 80.29% of ERAS patients (p=0.004) (Table 2).
There was a lower rate of major complications in Pre ERAS group however this did not reach statistical significance (10.53% pre-ERAS, 33.33% of ERAS group, p=0.052) (Table 2). Patients who suffered a major post-operative complication had a lower chemotherapy completion rate (62.5%) as compared to those who had minor complication (85.3%) (p=0.02) (Table 3). Multivariable analysis using logistic regression found CD complications to be the only independent variable that statistically influenced chemotherapy completion (p=0.02) (Table 4). The ERAS protocol was not shown to influence chemotherapy completion, along with age, gender, ASA score and CCI score.
On further analysis of CD complications, Chi-square cross tabulation showed a relationship between chemotherapy completion, ERAS and CD complication (Table 5). There was an association between the implementation of the ERAS protocol and higher chemotherapy completion rates, when comparing both the minor and major CD complication groups. The ERAS group had a higher proportion of CD minor and major complications completing chemotherapy (86.4 and 63.6%, respectively), compared to non-completion (13.6% and 36.4%, respectively) (p=0.033). Reasons for not completing chemotherapy are listed in Table 6. A high number of patients in the ERAS group declined further chemotherapy due to patient preference. Patient reasons were often variable (data not shown) however commonly involved side effects being intolerable.
Recurrence before 3 months are displayed in Table 7. ERAS patients had lower levels of recurrence <3 months (p=0.001), when compared to the pre-ERAS group. The ERAS group showed a trend of increased 5-year (60 months) survival and median survival compared to the pre-ERAS group. The ERAS group 5-year survival and median survival were 56% and 4.36 years, respectively. This was compared to the pre-ERAS group 5-year survival and median survival of 36% and 3.99 years, respectively (displayed in Figure 2). These differences in survival were not shown to be statistically significant (p=0.37).
With suboptimal 5-year survival rates post pancreaticoduodenectomy of 15-20% [24,25], multimodal therapy is the mainstay treatment of pancreatic cancer. These modalities include surgery, adjuvant chemotherapy, radiotherapy and a variety of other protocols. This study aimed to compare adjuvant chemotherapy completion rates in patients post-pancreaticoduodenectomy (PD) with and without exposure to the ERAS protocol. This is the first study to our knowledge that explores the relationship between ERAS and chemotherapy completion rates following PD.
ERAS protocol implements multimodal strategies that aim to maximise and restore functional capacity after surgery. Morbidity is minimised by reducing surgical stress, post-operative complications and post-operative pain as well as early oral diet and early mobilisation . The beneficial effects of the ERAS protocol have been extensively studied in colorectal cancer and were associated with stepwise reductions in complication rates, LOS, readmissions, the surgical stress response and increased 5-year survivability [26-29]. While, the beneficial effects of the ERAS protocol have been demonstrated after pancreaticoduodenectomy in a number of studies [30-35] no unified protocol exists apart from general guidelines . We hypothesised that our ERAS protocol improved adjuvant chemotherapy completion rates and thus overall survival (summarised in Table 1).
In our study, the pre-ERAS and ERAS chemotherapy completion rates were high (79%) in comparison to world standards (Table 2). The ESPAC-3 and CONKO-001 displayed completion rates of 68% and 62%, respectively [9,10,36]. Our high chemotherapy completion rates may reflect good case selection for pancreaticoduodenectomy. The pre-ERAS and ERAS groups displayed similar completion rates, despite ERAS patients being considerably more complicated as reflected in higher ASA scores, CCI scores and PD post-operative complications approaching statistical significance.
Numerous phase 3 clinical trials have demonstrated a clear relationship between completion of adjuvant chemotherapy and improved overall survival after pancreaticoduodenectomy. In 2013, the German CONKO-001 trial, demonstrated that among patients with macroscopic complete removal of pancreatic cancer, the use of adjuvant gemcitabine for 6 months compared with observation alone resulted in increased 5-year overall survival of 20.7% compared to 10.4%, respectively (p=0.01) . The ESPAC-3 study showed that overall survival favoured patients who completed the full six courses of either gemcitabine or fluorouracil treatment versus those who did not (p<0.001) . A Mayo Clinic clinical trial additionally showed overall survival was longer for recipients of adjuvant chemoradiotherapy versus surgery alone (OS 44.7 vs. 34.6%, p<0.001) .
Our study additionally displayed high initiation of chemotherapy rates, when compared to world standards. Pre-ERAS and ERAS initiation of chemotherapy rates were 82.6% and 84.6%, respectively (Figure 1). With the exception of a few centres of excellence with aggressive adjuvant chemotherapy initiation , in the majority of institutions, only 47-60 % of patients receive any adjuvant therapy after resection [15,38,39]. Reasons for this included advanced age, age-adjusted Charlson comorbidity index, side effects of chemotherapy and post-operative complications [13,14]. CD complications was the only independent variable that statistically influenced chemotherapy completion via logistic regression (p=0.02) (Table 4). Our findings are corroborated in a study by Labori et al which found major postoperative complications to negatively impact chemotherapy completion . In that study, of patients with major postoperative complications (CD complication≥3), only nine completed chemotherapy, with 32 not completing chemotherapy, and this was statistically significant. Major postoperative complications have been suggested to impair cellular immunity, increase the patient’s predisposition to early cancer recurrence and reduced survival .
Interesting, Chi-square cross tabulation analysis found the ERAS protocol to influence CD post-operative complications and chemotherapy completion. The ERAS group had a higher proportion of CD minor and major complications completing chemotherapy compared to non-completion. This is opposite to the majority of studies, where high rates of post-operative complications correlate to lower initiation and completion of chemotherapy rates [13,40,41]. It could be deduced that ERAS significantly helps more complex patients with post-operative complications complete chemotherapy. The ERAS group has significantly lower cancer recurrence before 3 months, compared to the pre-ERAS group (Table 6). This may be due to improved patient selection as the ERAS protocol for pre-op evaluation has changed during this period involving routine use of PET scans and selective use of staging laparoscopy. However, greater than 3-month recurrence rates would be more meaningful results with data currently being collected. Implementation of the ERAS protocol showed a trend to improve 5-year survival, compared to pre-ERAS. Rates were approximately 56% and 36%, respectively between the two groups (Figure 2). This is considerably higher than 5-year survival world standards of 15-25% [24,25,36,42]. Interestingly, the presence of CD major complications has been shown to negatively influence long-term survival in PD patients [43,44]. Some studies have demonstrated the presence of a postoperative complication correlated with delays in time to initiate chemotherapy. It is possible that delayed recovery from surgery due to postoperative complications might influence whether a patient receives adjuvant therapy [41,45].
The need to identify and reduce PD patient risk of failing to initiate and complete chemotherapy is of paramount importance and thus the ERAS protocol continues to evolve and develop. While many studies focus on the short-term benefits of ERAS, more research is required to explore the potential long-term benefits of ERAS, such as chemotherapy completion rates. The ERAS protocol for PD patients is to a large extent unexplored. These findings support the continued development of ERAS in PD patients as an opportunity to standardise care, reduce morbidity and the potential to reduce health care burden.
The limitations of this study are related to its retrospective nature. Firstly, small sample size reduced statistical significance of our results. However, a tertiary centre completing five pancreaticoduodenectomies a year is considered to be a high output centre [46-49]. Secondly, incomplete data regarding adjuvant chemotherapy were omitted from our analysis. Thirdly, we were unable to control the adjuvant chemotherapy regimen including agents used, dose, route of delivery and duration between pre-ERAS and ERAS groups. Nevertheless, the optimal regimen of either gemcitabine or 5-FU based chemotherapy were most often used. Furthermore, our knowledge of best practice has changed over the years [10,36] and is currently still developing . Additionally, we did not analyse the potential effects of neoadjuvant chemotherapy as well as radiotherapy on chemotherapy completion rates. Lastly, the ERAS protocol with closer consultant review may have increased the rate of detection, diagnosis and treatment of complications and thus lessen their impact.
This is the first study to our knowledge that has investigated the relationship between ERAS and chemotherapy completion rates. Our ERAS protocol has shown no deleterious effects on patient morbidity or mortality. Pre-ERAS and ERAS patients displayed similar chemotherapy completion rates of 79%, despite ERAS patients having higher ASA, CCI and CD postoperative scores. In the literature, higher postoperative complication rates correlate with lower chemotherapy completion and overall survival. In contrast, our study showed that high complication rates had higher chemotherapy completion rates, when the ERAS protocol was implemented. Overall, the ERAS protocol implemented at FSH hospital, is worth considering as a standardised protocol in future trials, with higher than world standard chemotherapy completion rates, despite significantly more complex patients.