The prevalence of preoperative malnutrition and postoperative nutritional risk in patients undergoing major surgery is estimated to be as high as 70 %(Reference Arends, Baracos and Bertz1,Reference Weimann, Braga and Carli2) . They are associated with an increased risk of infectious complications, prolonged hospital stay and increased mortality(Reference Ziegler3–Reference Adiamah, Skořepa and Weimann5). Major emergency abdominal surgery further increases these risks compared with elective surgery due to the combination of surgical stress and the patient’s physiologic condition at the time of surgery. Complication rates up to 60 % have been reported, and the 30-day mortality rate ranges from 10–40 % depending on the procedure and the patient-related risk factors(Reference Tengberg, Cihoric and Foss6–Reference Wolsted, Møller and Tolstrup8).
Adequate energy and protein supply are essential in critical illness and after emergency surgery to maintain organ function and to counteract the catabolic response inflicted by surgical stress(Reference Weimann, Braga and Carli9). The European Society for Parenteral and Enteral Nutrition and Enhanced Recovery after Surgery societies recommend that postoperative nutrition primarily be provided by the oral or enteral route and preferably be started as early as possible and parenteral nutrition (PN) considered around day seven(Reference Weimann, Braga and Carli9,Reference Burcharth, Falkenberg and Oreskov10) . Early, adequate nutritional support has been associated with superior clinical outcomes and lower complication rates in critically ill patients(Reference Compher, Chittams and Sammarco11–Reference Casaer, Mesotten and Hermans14). However, in many cases enteral nutrition is not feasible in the immediate postoperative period after emergency abdominal surgery due to ileus, nausea, fatigue, and loss of appetite. In these patients, PN can be a short-term viable alternative(Reference McClave, Taylor and Martindale15,Reference Ljungqvist, Scott and Fearon16) .
In elective major abdominal surgery, early supplemental PN has shown to lower postoperative infection rates(Reference Gao, Liu and Zhang17). No high-quality evidence exists on the impact of early supplemental PN after emergency abdominal surgery on postoperative complications. It is, therefore, still unclear if the clinical benefits of early nutrition in critically ill patients and elective surgery also apply to an emergency surgical population.
Postoperative complications after emergency abdominal surgery, especially infectious ones, pose a substantial health burden for society as well as a risk for the individual participant. Ensuring adequate nutritional supplementation is a way to reduce the risk of infections(Reference Zhang, Najarali and Ruo18). We hypothesise that supplying early PN on postoperative day (POD) two will result in a reduced incidence of postoperative infections when compared with postponed PN.
The EATERS trial aims to assess if early v. postponed supplementary parenteral nutrition (SPN) reduces the risk of postoperative nosocomial infections in patients undergoing major emergency abdominal surgery.
Study methods
Trial design
The EATERS trial is an investigator-initiated, multicentre, primary outcome assessor and analyst-blinded, randomised, controlled trial. The EATERS trial is a superiority trial to test if the early PN group is superior to the postponed group. Participants will be recruited from at least eight Danish hospitals.
The trial was prospectively registered (European Union Clinical Trial number: 2023-505378-14-00, ClinicalTrials.gov: NCT06089551) on the 28th of April 2023 and 13th of October 2023, respectively. The trial sponsor is Jakob Burcharth.
Trial conduct
The protocol was prepared according to the ‘Guidelines for the Content of Statistical Analysis Plans in Clinical Trials’(Reference Gamble, Krishan and Stocken19) and the Standard Protocol Items: Recommendations for Interventional Trials guidelines(Reference Hróbjartsson, Boutron and Hopewell20).
The trial will adhere to the protocol, the guidelines for good clinical practice, the Helsinki Declaration in its latest version and the national laws of Denmark.
Randomisation and blinding
Participants are randomised 1:1 in variable block sizes of two, four, six and eight. Each trial site is given a unique ID with its own randomisation table. The allocation sequence was created electronically in R using computer-generated random numbers (Vienna, version 4.3.0) and uploaded to Research Electronic Data Capture, Vanderbilt University, Tennessee. The randomisation sequence will be hidden from trial staff and cannot be accessed. Randomisation will be performed electronically in Research Electronic Data Capture, Vanderbilt University, Tennessee.
The allocation sequence and block sizes are unknown to the investigators and cannot be accessed by trial staff. The participant will provide informed consent on POD2, after which the randomisation will be performed. Informed consent must be obtained prior to randomisation.
Members of the treating staff will be aware of the participant’s randomisation group as the intervention cannot be blinded. The research team will be aware of the allocation while recruitment is ongoing.
The primary outcome will be assessed independently by two sets of board-certified general surgeons. The assessors will be blinded to the allocation, as all information indicative of the allocation will be concealed. Outcome assessors will not assess participants from their own hospital.
The imputation of missing data will be performed prior to exporting it for analysis. A statistician not affiliated with the study will conduct the primary and secondary outcome analyses and will be blinded to the allocation.
Trial population
Screening data, eligibility and recruitment
Screening will be performed by trained research staff, and assessment of eligibility and recruitment will be carried out by medical doctors at each site. Participants will be recruited from the surgical ward, high dependency unit or intensive care unit. Inclusion in the trial and informed consent will be obtained on POD2. POD0 is defined as the date of start of surgery (knife to skin), and POD2 is the second day after start of surgery. Thus, POD2 will not necessarily be 48 h after surgery. Energy intake is assessed as total intake in the previous 24 h. The time period may vary slightly between hospitals depending on meal service, but will consist of lunch and dinner on POD1 and breakfast on POD2 including all meals and snacks in-between. If intake registration is not adequately documented, the participant will not be eligible.
Screening data will be reported in a Consolidated Standards of Reporting Trials flow diagram(Reference Schulz, Altman and Moher21). See Figure 1.
CONSORT flow diagram of participant screening, randomisation and follow-up. Abbreviations: CONSORT: Consolidated Standards of Reporting Trials, NRS-2002: Nutritional Risk Screening 2002, POD: Postoperative day.
Inclusion criteria
Participants must fulfil all inclusion criteria
-
Age ≥ 18 years
-
Primary acute or subacute laparotomy or laparoscopy for perforated hollow viscus, obstruction, ischaemia, bleeding or inflammation
-
NRS-2002 (Nutritional Risk Screening) score < 7
-
No contraindications for oral or enteral feeding
-
Unable to tolerate or take in at least 50 % of caloric target on POD2
-
Capable of providing informed consent at the time of inclusion
Exclusion criteria
Participants will be excluded if they fulfil one of the following exclusion criteria:
-
Laparotomy without closure of the abdominal aponeurosis
-
Intellectual disability or severe psychiatric disorders rendering participation unethical
-
Very limited expected remaining time of living (< 3 months)
-
Pregnant or breast-feeding women
-
Need for PN prior to POD2
-
Laparoscopic appendectomy or laparoscopic cholecystectomy as the sole procedure
-
Refusal to participate in the study
Lastly, patients with NRS-2002 = 7 were excluded, as we were worried about the risk of refeeding syndrome and whether it was ethically justified to potentially postpone PN in a group at extreme nutritional risk.
Participant timeline
Participants are randomised on POD2 to up to five days of supplemental PN initiated on either POD2 or POD5. The caloric target is set at 25 kcal/kg/d. For participants with a BMI ≥ 30, their body weight at BMI 25 will be used. The feeding protocol will be hypoenergetic to reduce the risk of refeeding syndrome and to account for an increased endogenous energy production in the early postoperative period. The aim is to provide 70–80 % of the calculated caloric target, as hypocaloric feeding in critically ill patients is shown to decrease mortality(Reference Zusman, Theilla and Cohen22). The timeline of participants is presented in Figure 2. The intervention is discontinued after five days to avoid extended periods of hypoenergetic feeding, if the participant achieves 70 % of the caloric target orally or enterally, or if the patient is discharged. Participants with a continued need for PN following the intervention will be treated according to regional guidelines. The treating physician may deviate from duration or dose of PN if clinically justified.
Timeline of participants randomised to early or postponed supplemental parenteral nutrition. Abbreviations: EN: Enteral nutrition, E-SPN: Early supplementary parenteral nutrition, L-SPN: Late supplementary parenteral nutrition.
A gradual increase in PN is permissible for participants at high risk of refeeding syndrome. All PN will be off-the-shelf, three-chamber bags of either SmofKabiven® or SmofKabiven Perifer® with the addition of trace elements and vitamins.
The intervention period is from randomisation to POD7 or POD10, depending on the group allocation. If the participant is discharged and readmitted within this time period, the intervention will not be resumed.
Early parenteral group
SPN will be initiated on POD2 and will be adjusted daily according to oral/enteral intake to ensure 70–80 % of the caloric target is achieved. Oral nutritional supplements and tube feeding is permitted and may be prescribed by the treating physician.
Late parenteral group
Participants will receive standard care until POD5. Oral nutritional supplements and tube feeding is permitted and may be prescribed by the treating physician. If oral and enteral intake is < 50 % of the caloric target, SPN will be administered on POD5. Participants achieving an intake > 50 % of the caloric target SPN will not be initiated. Participants subsequently requiring PN will be treated according to regional guidelines.
If PN is initiated prior to POD5, participants will follow the plan suggested by the treating physician.
Outcome definition
The primary outcome measure is the incidence of participants with a postoperative nosocomial infection from randomisation (POD2) until discharge. A nosocomial infection will be defined as one of the following:
-
Bloodstream infection and sepsis
-
Pneumonia
-
Urinary tract infection
-
Surgical site infection
-
Organ or deep-space infection
-
Skin and soft tissue infection
Preoperative or pre-existing infections will not be recorded as a postoperative infection unless it is evident that the infection has been cleared. Definitions for each infection are provided in online Supplementary 1.
Secondary outcomes
-
Number of patients with one or more non-infectious complications with a Clavien-Dindo score above 2.
-
Number and type of infectious complications
-
Days with administration of antibiotics during admission
-
Length of stay
-
Mortality risk at 30 and 90 d
-
Energy and protein intake
-
Post-discharge Simplified Nutrition Assessment Questionnaire (SNAQ) score at 30 and 90 d (± seven days)
-
Readmission risk and readmission diagnosis within 30 and 90 d of surgery of those discharged and alive
-
Serious adverse events
Detailed definitions of the secondary outcome measures are available in online Supplementary 1. Non-infectious complications are defined in online Supplementary 2.
Data collection and registered variables
Data will be collected from electronic patient charts and registered in a Research Electronic Data Capture, Vanderbilt University, Tennessee database designed for this study. Entry of data will be performed by trained staff at each site. Follow-up data via phone calls will be collected by the sponsor site. All participants will be contacted at least twice on separate days. Data entry will be monitored centrally from the sponsor site as well as on-site by the good clinical practice unit.
A trial investigator will record and evaluate adverse events using the Common Terminology Criteria for Adverse Events as well as determine causality between the adverse event and the supplemental PN. All adverse events will be evaluated by a physician.
Baseline variables
-
Age
-
Sex
-
Date of admission to hospital
-
Height
-
Weight
-
NRS-2002 score
-
Clinical Frailty Scale, WHO performance status, and American Society of Anaesthesiologists score
-
Comorbidity and medical history
-
Operative details including operative findings, procedure, duration and blood loss
Daily postoperative variables
-
Weight
-
Enteral energy intake
-
Enteral protein intake
-
Parenteral energy supplement
-
Parenteral protein supplement
Blood samples
-
Leucocyte count, creatinine, blood urea, alanine aminotransferase, alkaline phosphatase, bilirubin and C-reactive protein on POD 3, 7 and at discharge.
Follow-up at 30 d after surgery (± seven days)
-
Death (yes/no).
-
SNAQ score
-
Weight
-
Readmission (yes/no).
Follow-up at 90 d after surgery (± seven days)
-
Death (yes/no).
-
SNAQ score
-
Weight
-
Readmission (yes/no).
Sample size calculation
A combined nosocomial infection risk of 40–60 % after major emergency surgery is reported, which includes local data. We assume a risk reduction from 51 % to 36 % in the intervention group. With α 0·05, β 0·2 and power of 0·8, this results in two groups of 171 participants and a total of 342. As the primary outcome will be assessed using the electronic patient chart which is available for all participants, we do not anticipate missing data for the primary outcome. For this reason, additional participants will not be included in the trial. No sample size calculation was made for the subgroup analyses, and these will therefore be exploratory.
Monitoring
The trial will be monitored by the good clinical practice units in Denmark which act as an independent party. The monitoring includes data quality and protocol compliance. Visits will be at least once a year depending on inclusion rates. In addition, regular monitoring of data entry and protocol deviations will be performed by the trial sponsor or sponsor’s representative.
The decision not to establish a data monitoring committee was made, as the trial is approved as a low-intervention trial.
Timing of outcome assessments and final analysis
The primary outcome analysis will be performed after the last included participant’s discharge or 90 d after inclusion, whichever comes first. Postoperative SNAQ score, weight and mortality will be assessed 30 and 90 d after surgery. Readmissions will be assessed up to 30 and 90 d post-surgery among those alive and discharged. The remaining secondary outcomes will be evaluated with the primary outcome.
Statistical principles
CI and P-values
This study will use a significance level of 5 % and 95 % CI. A P-value < 0·05 will be considered statistically significant. All tests will be two-sided.
No correction for multiplicity will be carried out, but all secondary and exploratory outcomes will be interpreted with this in mind.
Adherence and protocol deviations
The treating physician can deviate from the protocol for all included participants if clinically justified. The reason for deviation will be registered in Research Electronic Data Capture, Vanderbilt University, Tennessee. Protocol deviations and major protocol violations will be registered. Major protocol violations are defined as patients who did not meet the inclusion criteria, fulfilled one or more exclusion criteria or fulfilled one of the following:
-
Did not initiate SPN on POD2
-
Had more than one instance of caloric target outside 70–80 % range
-
Did not initiate SPN on POD5
-
Did not initiate SPN in participants with intake < 50 % on POD5
-
Had more than one instance of caloric target outside 70–80 % range
Withdrawal/follow-up
Only participants who are able to provide informed consent will be included in the study. Participants may, at any time, withdraw their consent. If a participant withdraws the informed consent, data up until that point will be used for analysis.
Analysis
Analysis populations
The analyses will be performed on the intention-to-treat population, which is defined as all randomised participants. The intention-to-treat population will be used for the primary analysis, and the conclusion of this trial will be based on this population.
The per-protocol population is the same as the intention-to-treat population, but excluding participants with one or more major protocol violations.
Analysis methods
Primary analysis
A logistic regression analysis will be adjusted for the trial site on the intention-to-treat population. The primary outcome will be an OR with 95 % CI.
Sensitivity analyses
-
Logistic regression analysis on the per-protocol population adjusted for the trial site.
-
Logistic regression adjusted for the trial site, duration of antibiotic treatment, and intra-abdominal contamination at the time of surgery (Centers for Disease Control and Prevention surgical wound classification).
-
Cox regression analysis using time until the first event.
-
A complete case analysis will be performed.
If a patient withdraws consent during the intervention period, data until withdrawal will be used. If no outcome has occurred in this time period, the patient will be registered as not having had an outcome.
Secondary outcomes
The dichotomous secondary outcome measures will be analysed in the following way. If possible, all secondary outcome measures will be adjusted for the trial site.
-
Non-infectious complications
-
◦ Logistic regression analysis adjusted for the trial site.
-
-
Readmission within 30 and 90 d of surgery
-
◦ Cox regression analysis will be performed with delayed entry at time of discharge adjusted for the trial site.
-
-
Days with administration of antibiotics during admission
-
◦ Linear regression analysis adjusted for trial site. If distributional assumptions are not met, transformations of outcome will be performed.
-
-
Mortality rate
-
◦ Cox regression analysis adjusted for trial site
-
-
Length of stay
-
◦ Linear regression analysis adjusted for trial site. If distributional assumptions are not met, transformations of outcome will be performed.
-
-
Energy and protein intake
-
◦ The two groups will be compared using Wilcoxon rank test on the median intake during the admission. Kaplan–Meier will be used to assess the time to oral intake of at least 70 % of the caloric target and 70 % of protein requirement of 1·3 g/d.
-
-
Post-discharge SNAQ score
-
◦ Linear regression analysis adjusted for trial site. If distributional assumptions are not met, transformations of outcome will be performed.
-
-
Serious adverse events
-
◦ Logistic regression analysis adjusted for trial site
-
For continuous outcome measures, we will assess data distribution. If possible, we will use a generalised linear model. If the distributional conditions for this model are not met, we will perform transformations of outcome or use a non-parametric test adjusted for site.
If possible, all secondary outcome measures will be adjusted for trial site. In case of several sites with very few included participants (< 5), the adjustment will be inclusion at the Herlev site v. other sites.
Pre-Planned subgroup analyses
-
Patients receiving at least one dose of PN
-
NRS-2002 score (≥ 3 or < 3)
-
Surgical approach (laparotomy or laparoscopy)
-
Duration of surgery > 120 min or ≤ 120 min
-
Clinical Frailty Scale score > 3 or ≤ 3
Missing data
The proportion of missing data will be reported. If < 10 % of the data for a given analysis is missing, data will not be imputed.
Model-based multiple imputation will be applied in case of missing data. We will create 20 input datasets. We will use the following variables: Age, sex, American Society of Anesthesiologists (ASA) score, performance status, NRS-2002 score, baseline weight, baseline co-morbidities (CVD, pulmonary disease, diabetes, and previous abdominal surgery, history of malignancy, immunosuppressive treatment), smoking status, type of surgery (laparotomy or laparoscopy), duration of surgery (as continuous), stoma placement and bowel anastomosis. For imputation, we will not use variables that may reveal the treatment allocation. Lau Caspar Thygesen will perform the imputation and will be blinded to the treatment allocation.
Additional analyses
Safety analysis
Occurrence of one or more adverse events during the intervention period in patients who received at least one dose of PN is defined as:
-
Hyperglycaemia ≥ 10 mmol/l
-
Hypoglycaemia < 4·2 mmol/l
-
Hypophosphataemia < 0·76 mmol/l
-
Hyperphosphatemia ≥ 1·41 mmol/l
-
Plasma bilirubin > 25 μmol/l
-
The analysis will be performed using logistic regression analysis adjusted for trial site.
Statistical software
R (R Foundation for Statistical Computing, Vienna, Austria) and SAS (SAS Institute Inc.)
Discussion
The EATERS trial aims to provide novel insights on the timing of supplemental PN in major emergency abdominal surgery, as the question of when to initiate parenteral nutrition in this setting is unanswered. Many of these patients will experience type I intestinal failure often rendering enteral feeding inadequate or not feasible(Reference Klek, Forbes and Gabe23). Supplementation with PN can increase energy provision and reduce the incidence of infections in critically ill adults(Reference Heidegger, Berger and Graf24). In elective gastrointestinal surgery, early supplemental PN has been associated with a lower risk of postoperative infections without increasing costs or number of adverse events(Reference Gao, Liu and Zhang17). However, studies on the effect of supplemental PN in emergency surgery are lacking.
Although the recent Enhanced Recovery After Surgery guideline in emergency laparotomy recommends early PN for patients with contraindications to enteral nutrition, it offers no guidance on when to start early PN or how much PN to provide(Reference Scott, Aggarwal and Aitken25). This highlights a critical gap in the current literature. The EATERS trial will address this gap by providing high-quality data on potential benefits and safety of early supplemental PN. Findings from our study will aid in future guideline development, improve nutritional decision-making and optimise postoperative recovery in the acute surgical setting.
Strengths
The EATERS trial is a multicentre trial which increases the applicability of the findings. The trial will be conducted and monitored according to good clinical practice guidelines. This protocol and statistical analysis plan will be published prior to data analysis and before inclusion of the last participant. Finally, outcome assessors and the trial statistician are blinded to the treatment allocation.
Limitations
This trial is not blinded to clinicians, investigators or participants as this would simply not be feasible. This can result in an increased risk of bias, as it may affect the clinician’s behaviour towards PN.
To reduce the risk of bias, both outcome assessors and the trial statistician will be blinded to the allocation.
Participants may also be subject to protocol violations. We expect protocol violations to be more frequent in the postponed group (late-SPN) with start of PN prior to POD5 being the most common.
Supplementary material
For supplementary material/s referred to in this article, please visit https://doi.org/10.1017/S0007114526106631
Acknowledgements
The EATERS trial is initiated by the EMERGE research group at the Department of Gastrointestinal and Hepatic Disease at Copenhagen University Hospital Herlev. The trial is supported by a grant from the Novo Nordisk Foundation grant, NNF22OC0079459, to Jakob Burcharth. The Novo Nordisk Foundation had no role in the design or writing of this protocol or in the conduct of the trial. Drug expenses during the study are paid by sponsor site, Copenhagen University Hospital Herlev.
Conceptualisation: J. B. H., J. B., D. K., L. C. T. Data curation: J. B. H., J. B., D. K. Investigation: J. B. H., J. B., D. K. Methodology: J. B. H., J. B., D. K., L. C. T. Supervision: J. B., D. K., L. C. T. Writing – original draft: J. B. H. Writing – review and editing: J. B., D. K., L. C. T.
J. B. serves as an advisory board member at Fresenius Kabi and has received grants from the Novo Nordisk Foundation and Independent Research Fund Denmark. L. C. T., D. K. and J. B. H. declare no conflict of interest.
The EATERS trial is registered in the European Clinical Trials Database and has been approved by the Danish Medicines Agency and the Danish Medical Research Ethics Committees (2023-505378-14-00). Only participants capable of providing verbal and written informed consent will be included in this study. Personal information of enrolled participants will be accessible only to staff from the including site. All participants are covered by ‘Danish Patient Compensation’ in case of harm during the trial.
Patients were not involved in the design of this trial. All participants will be offered a report of the study results once the trial is completed.
The trial was approved on the 29th of June 2023. The first participant was included on October 18th 2023, and 280 participants have been included as of 11th of November 2025. The trial is expected to be completed in early 2026.
After completion of the trial, the results will be submitted to a peer-reviewed clinical journal regardless of positive, negative or inconclusive results. A summary of the trial results written in plain language will be made available on ‘https://euclinicaltrials.eu/?lang=en’ after the trial.
The data that support the findings of the trial will be made available on reasonable request to the trial sponsor (JB). The protocol and other trial documents are available at ‘https://euclinicaltrials.eu/?lang=en’
