ERAS Versus ERAS Plus Artificial Nutrition in Open Pancreatoduodenectomy

Background

The enhanced recovery after surgery (ERAS) protocol is currently considered the gold-standard clinical pathway for perioperative care (1). This protocol is a bundle of interventions derived from the best evidence-based perioperative treatments (1). The aim of the ERAS protocol is to accelerate patient recovery through the reduction of the dysmetabolic consequences of surgical and anesthesiology injury. When the ERAS protocol is implemented, it has been shown a reduction of the rate of surgery-related complication, length of hospital stay, and health care costs (1).

The achievement of single ERAS items also facilitates a faster recovery of oral food intake and of several key metabolic pathways (2). Appropriate tissue healing and recovery/maintenance of organ function after an operation needs an effective and efficient metabolic response, which in turn necessitates adequate qualitative and quantitative nutritional substrates to be effective. Most of the evidence on the advantages of implementing an ERAS program are derived from trials conducted in patients undergoing colorectal surgery. In this cohort the baseline nutritional risk and the rate of malnutrition is less than 10% (3). Moreover, the growing use of minimally invasive techniques has facilitated a faster recovery of bowel function and optimal tolerance of early food resumption after surgery (4). As a result, patients who receive a laparoscopic colorectal resection easily tolerate at once the reintroduction of normal food after the operation and can achieve the full nutritional requirement within 2/3 days (5).

Conversely, major pancreatic resections represent one of the most complex and challenging abdominal operation that do not fit with laparoscopic approach with exception of very selective cases (6). Furthermore, the proportion of patients having a high nutritional risk or are malnourished at baseline is higher than 80% (7). On the top, pancreatic surgery is characterized by peculiar complications such as primary delayed gastric emptying (DGE) or secondary DGE after the occurrence of a pancreatic fistula. These events are frequent (up to 50%) (8) and deeply compromise the regular resumption of oral feeding posing patients at risk of severe underfeeding. Therefore, beside the fact that early oral feeding after pancreatic surgery seems safe, there are no convincing data on whether the goal of attaining adequate nutritional needs is accomplished by a pure ERAS protocols (9) which endorse the use of artificial nutrition only in selected cases (6). However, the level of evidence is low. In fact, only one trial has specifically addressed this issue (10). In this study, the mean daily calorie and protein intakes in the first 2 weeks were similar in the ERAS group and the other group managed conventionally, despite the fact that during the first five post-operative days, the mean daily intakes of calories and proteins slightly favored the ERAS group. Overall, the authors showed that the total energy goals through oral feeding were quite low in both groups.

Many of the published studies did not analyze the compliance with ERAS protocols and reported incomplete data, particularly on early oral feeding (11,12). Robertson et al. (13) reported compliance rates of 82% for resumption of oral fluids and 86% for tolerance of the diet. Conversely, in another large trial (14), postoperative oral liquids were tolerated by 55% of the patients and solid food in 53%, but compliance dropped substantially in patients with major complications.

Thus, the available evidence suggests that nutritional intake using only early oral feeding (food "at will") within an ERAS protocol after pancreatic surgery may be only partially adequate. According to expert opinions (15), artificial nutritional support should be implemented early postoperatively in malnourished patients or those patients at high risk of developing malnutrition, those who develop severe post-operative early after operation, and in well-nourished patients who do not tolerate at least 50% of their caloric and protein requirement by postoperative day 7 for any reason. Instead, other international guidelines for clinical nutrition in critically ill patients (16) state that permissive underfeeding in the first 3 to 7 days after hospital admission is associated with reduced morbidity and mortality because excessive delivery of calories (overfeeding) may cause severe derangement of several organ systems. One recent RCT, showed that in patients submitted to PD and kept "nil by mouth" for 10 days after the operation, immediate parenteral feeding was associated with less complications when compared to tube enteral nutrition (17). One systematic review (18) compared the outcomes of 5 feeding routes after PD and reported no difference in terms of safety and efficacy. In persisting DGE, better outcomes are achieved when artificial nutrition, either parenteral or enteral, is started within 10 days of operation (19).

According to these results, there is an absence of specific evidence concerning the best management of postoperative nutritional support after PD.

Given the lack of strong evidence, we believe that there is a need of a randomised clinical trial specifically design to explore whether given a full nutritional requirement by parenteral feeding in the first 5 days after surgery coupled with oral food "at will" compared to only oral food "at will", within an established ERAS program, could achieve a reduction of the morbidity burden (20).

Study design and management

RASTA will be a randomised controlled multicentre single-blind, parallel arm, superiority, phase-3 trial.

The trial will be conducted in 6 Institutions with proven experience in pancreatic surgery and already applying an ERAS program (list of participating centres and principal investigator of each centre in appendix 1).

Pre-trial training:

Before starting patient enrolment, a serial of meetings will be organised to accomplish:

- Correct definition of eligibility, inclusion and exclusion criteria.

- Accordance on definition of complications.

- Training on randomization process and patient instruction on treatment arms.

- Accordance on ERAS domains.

- The training of outcome assessors will be performed to achieve concordance on the occurrence of the primary and secondary endpoints. Each participating centre will nominate two independent outcome assessors. In case of discordance on the assignment of the endpoint, a third expert will intervene to solve the dispute and classify the patients as complicated or not. Outcome assessors will be blinded to treatments

- Training on how to fill out correctly the case report form.

Patients eligible for participation Adult patients (age ≥ 18 and < 90 years of age) candidate to elective pancreatoduodenectomy for any periampullary or pancreatic cancer.

Exclusion criteria:

- American Society of Anaesthesiologists (ASA) physical status classification > 3

- Weight loss > 15% with respect to usual weight in the last 6 months

- Child-Pugh > A

- Not signed consent

- Palliative surgery

- Placement of a naso-enteric or jejunostomy feeding tube

Screening and randomization processes:

After being screened for inclusion and exclusion criteria, patients or their legal representative will be asked to sign a written informed consent. Than subjects may be enrolled in the study and randomly allocated into two arms. Randomization will be performed the day before operation. All reasons for exclusion after screening will be recorded.

The study needs to be approved by the Ethical Committee of all participating centres.

The local Ethical Committee, as coordinating centre, will provide the "not emendable judgement" according to the Italian legislation.

The trial will be registered at ClinicalTrials.gov. THE RASTA trial will be managed and coordinated by the School on Medicine and Surgery of the Milano-Bicocca University at the Department of Surgery of the San Gerardo Hospital, Monza, Italy. The coordinating centre will be also responsible for treatment allocation and monitoring, and statistical analysis with the support of the Centre of Biostatistics for Clinical Epidemiology of the Milano-Bicocca University.

The expected duration of enrolment is approximately two years.

Procedures (Figure 1)

Study intervention Patients randomised in the treatment arm will be treated with a full ERAS protocol that establishes oral food "at will" plus parenteral nutrition (PN) from postoperative day 1. A 3-bag compartment peripheral parenteral solution (mOsm < 800) containing carbohydrate, lipids and proteins will be infused to deliver 20/25 total Kcal/kg for a total of 5 days after the operation. In case of the occurrence of any complication impairing the full or partial recovery of oral food, the treatment will be continued until clinically indicated.

Control arm Patients randomised in the control arm will be treated with a full ERAS protocol that establishes oral food "at will". In case of the occurrence of any complication impairing the full recovery of oral food within postoperative day 7, patients will receive parenteral nutrition as in the treated arm until clinically indicated.

Procedures common to both arms Patients of both groups will be treated according to the ERAS Society guidelines for perioperative care for PD (6). Attached is the ERAS guidelines for pancreatic surgery. Blood glucose ≥180 mg/dL will be treated with insulin injection (either subcutaneous or by continuous IV infusion)

Randomisation and masking

Patients will be randomly allocated to ERAS or ERAS plus PN at 8:00 PM of the first postoperative days. Randomisation will be performed by a computed-generated permuted-block sequence. A specific code will be generated for each centre to achieve equivalent grouping. The allocation ratio will be 1:1 with a block size of 4. Randomization will be competitive among centers.

Surgeons will not be blinded to treatment arm. Masking to allocation will be impossible to achieve for the nature of the study. Patients will be evaluated for outcomes by assessors not directly involved in patient care and masked to patient allocation.

Outcomes

The primary endpoint of the trial is the complication burden as measured by the Comprehensive Complication Index (CCI) (20) within 30 days after hospital discharge.

Hypothesis tested - H0 hypothesis: The administration of a parenteral nutrition added to the ERAS protocol will not affect the CCI as compared to standard of care (ERAS).

- H1 hypothesis: The administration of a parenteral nutrition added to the ERAS protocol will affect the CCI as compared to standard of care (ERAS).

Post-discharge follow-up is accomplished by weekly outpatient visits. Also telephone interviews will be allowed to monitor patient health state, but in case of warning signs or symptoms of complication, the patients will be asked to refer to the hospital where the operation was performed for further clinical evaluation.

Secondary outcome measures will be:

- Daily calorie delivered by TPN

- Rate of unplanned TPN (control group)

- The rate and severity* of complications at 30 days after discharge.

- Rate of surgical site infections (CDC classification) (21)

- The rate and severity of POPF (22)

- Rate and severity of DGE (23)

- Rate and severity of haemorrhage (24)

- LOS based on predefined criteria, and actual LOS

- the rate of reoperation.

- the rate and duration of intensive care treatment.

- The rate of hyperglycaemia (blood glucose > 180 mg/dL)

- Use of insulin (subcutaneous bolus or continuous infusion)

- Δ plasma prealbumin levels (baseline, POD1 and POD 6)

- Use of morphine

- Body weight at 30 and 90 days

- Readmission rate

- 90-day mortality

- The severity of complications is scored according to the Dindo-Clavien classification (25) (Appendix 3).

Any attending surgeon will decide the day of discharge according to own clinical judgement. However, pre-specified hospital discharge criteria will be set and recorded.

Ethical considerations

Patients in both groups will receive the best possible and most updated perioperative care pathway (ERAS protocol). The experimental group will be treated with supplementary parenteral nutrition to achieve the full nutritional requirement immediately after surgery. Parenteral feeding is already part of the established clinical practice and in use since more than 50 years. The additional risk for the experimental group is to experience more episodes of hyperglycaemia (blood glucose > 180 mg/dL), particularly in subjects with diabetes, which although will be promptly treated with insulin. Similarly, the experimental group may have an additional risk of plasma electrolyte unbalance, which can be easily corrected by standard treatments. In turn, as argued in the introduction section and in the hypothesis statement, the experimental group might experience a reduced morbidity burden.

Safety issue

We considered as adverse events the following:

- the number of patients not reaching tolerance to oral feeding within 7 days after surgery

- the number of patients needing insulin therapy

- the number of patients requiring electrolyte corrections.

- Δ Body weight (difference between baseline and discharge)

Data collection and management

All data will be collected into an electronic database with a double entry to assure consistency of records. In case of missing or implausible data, queries will be mailed to the participating centres to obtain integrations or corrections. Data collectors will be blinded to allocation.

Statistical planning

The sample size of 120 patients per group is necessary to provide a 80% power to detect at least a 30% reduction in the CCI, which is expected to be around 23 (median) (IQR 21-31) or mean 27 (±20 SD) in complicated patients of the control group. The hypothesised reduction of 30% is based on sound clinical relevance. The median CCI of 23 is retrieved from a previous publication (26). The rate of complication in this type of surgery is expected to be approximately of 60%.

A Mann-Whitney test is considered, type I error rate is fixed at 5% (two tails) and an expected drop-out of 10 % is taken into account.

For the binary end-points, the relative risk (RR) with the corresponding 95% confidence interval, comparing the two groups, will be estimated. For the primary end-point, also the risk difference (RD) will be computed. For the numerical end-points the difference in the location parameter (i.e. median pairwise difference) between the two groups with the corresponding 95% confidence interval will be computed. Fisher test and Mann-Whitney test will be adopted to evaluate univariate associations. Incidence of complications over time in the two groups will be described according to the Kaplan-Meier estimator.

A multivariate logistic regression model will be used to identify factors associated with the primary endpoint and to evaluate the effect of treatment adjusting for possible residual confounding. Using logistic regression, the effect of PN over controls on the CCI will be also investigated within pre-specified subgroups to account for possible effect modification. The pre-specified risk factors for this analysis will be:

- NRS-2002 (≥ 3)

- BMI (> 30)

- sex (male)

- age (> 70 years)

- Charlson comorbidity index (>4)

- ASA score (= 3)

- Neoadjuvant treatment

- blood loss (≥ 500 mL)

- Duration of surgery (> 360 min)

- Jaundice

- Biliary stenting

- Diabetes

- PPPD (vs. Whipple)

- PDAC (vs. others)

- Fistula risk score (≥ 7)

- ERAS overall compliance (> 70%) Analyses will be done on the modified intention-to treat analysis principle to represent clinical practice. All the analyses will be performed with the R software.

ETHICAL / LEGAL ASPECTS:

AC / CE approval A copy of the patient's informed consent must be submitted to the CA and CE together with the protocol for written approval. Written approval of the protocol and informed consent must be obtained before starting the recruitment of patients in the study.

The approval of the Competent Authority (AC) / and the favorable opinion of the local Ethics Committee (EC) must be obtained before the start of the experimentation.

The researcher must inform the CA and the CE of any amendments to the protocol, which must be approved by both.

Informed consent The subject is considered enrolled in the study when he voluntarily provides written informed consent. No study procedure will be initiated before the Informed Consent has been signed. Before acquiring informed consent, the investigator explains to the subject what the study consists of, what are the objectives it aims to achieve, what procedures / treatments it will be necessary to carry out on the subject, any benefits for the subject, such as these are the expected adverse reactions, the possibility that unexpected adverse reactions may occur and any other information necessary to understand the study; the explanation must take place in a manner and in a language understandable to the subject and dedicating the time necessary for the subject to fully understand the study and to ask any questions. Once this interview phase is over, the investigator delivers the informed Consent form to the subject who can decide to sign at the moment or to take home, calmly evaluate and deliver at a later time. The delivery of the form will take place at the beginning of the enrollment.

Treatment of personal data The personal data object of the study must be treated in compliance with the European Regulation on the Protection of Personal Data (GDPR), the Legislative Decree 196/2003 and subsequent amendments and additions, and any other Italian law applicable to the protection of personal data (henceforth referred to as the "applicable data protection law").

With regard to personal data made available to the Promoter in pseudo anonymized form, the Institute and the Promoter are considered independent data controllers of the data, each for its own area of competence, and both will act in compliance with the applicable data protection law.

In addition, the Institute and the Promoter will cooperate to take the necessary measures in order to comply with the applicable data protection law, and will also have to implement appropriate technical and organizational measures to guarantee the GDPR compliance requirements.

The Institute is responsible for providing its Principal Investigators and Research Staff with all the necessary information relating to the methods by which the Promoter will eventually collect and manage their personal data before such information is provided.

The Promoter ensures the correctness of the processing of personal data as required by the GDPR. If the user becomes aware of a data breach of personal data, the user himself must promptly notify the other users of this event, and in any case no later than 24 hours from the knowledge of the event. In this case, the user must fully cooperate with the other users to remedy the data breach, fulfilling the mandatory notification within the set times and managing any damage caused.