EFFORT
The effect of higher protein dosing in critically ill patients with high nutritional risk (EFFORT Protein): an international, multicentre, pragmatic, registry-based randomised trial
D K Heyland. Lancet 2023; Jan 25;S0140-6736(22)02469-2. doi: 10.1016/S0140-6736(22)02469-2.
Clinical Question
- In critically ill patients, does a higher dose of protein compared to usual dose protein, improve time-to-discharge-alive from hospital up to 60 days?
Background
- Critical illness is associated with loss of muscle mass, and the degree of sarcopenia is correlated to clinical outcomes
- Animal studies and physiological principles indicate that protein could be the most clinically important substrate to maintain muscle mass and improve clinical outcomes, and labelled isotope scans in humans have suggested that exogenous amino acids may stimulate an anabolic response
- However, observational studies comparing lower and higher dose protein doses have conflicting results and randomised controlled data is sparse and limited by insufficient separation between groups
- This uncertainty is reflected in international nutrition guidelines, which vary in their recommendations and in registry data, which demonstrates considerable variety in nutrition practices around the world
Design
- Single-blinded, multi-centre, randomised trial
- Central randomisation through a computer-generated schedule that allocated patients in a 1:1 ratio to receive either high-dose or usual dose protein through permuted blocks of 2, 4 or 8
- Due to the COVID-19 pandemic, recruitment stalled and the original primary outcome (60-day mortality) was amended in June 2021
- The blinded data set was accessed at this time to determine pooled event rates, from which estimations of 60-day mortality and median time-to-discharge-alive from hospital were made
- Using these assumptions, the investigators estimated that a sample size of 600 patients per group would achieve 83% power at a two-sided α of 0·05 for a relative risk reduction in hospital mortality of 15% (from 34·6% to 29·5%), combined with a 20% increase in the hazard rate of time-to-discharge-alive among hospital survivors
- Modified intention-to-treat analysis
- Registered at ClinicalTrials.gov and monitored by an independent DSMB
Setting
- 1329 patients recruited in 85 ICUs across 16 countries between January 2018 and December 2021
- 28 did not receive the assigned intervention and were excluded from the analysis, according to the pre-specified modified intention to treat design
Population
- Inclusion: Adults within 96 hours of ICU admission who were expected to remain mechanically ventilated for at least 48 hours, with ≥1 nutritional risk factor:
- Low (≤25 kg/m2) or high (≥35 kg/m2) BMI
- Moderate to severe malnutrition, as defined by local assessments
- Frailty (CFS ≥5)
- Sarcopenia (SARC-F ≥4)
- Exclusions:
- >96 hours of mechanical ventilation prior to screening
- Expected to die or undergo withdrawal of life-sustaining treatments
- Pregnancy
- Where the responsible clinician felt that the patient needed either low or high protein
- Requirement for PN where the recruiting site did not have products to achieve high protein dose targets
- Baseline demographics (high-dose vs normal dose protein):
- Age: 57 vs 57 years
- Male sex: 61% vs 59%
- Medical admission: 85% vs 82%
- Vasopressor: 39% vs 41%
- APACHE-II: 21 vs 21
- Primary diagnosis:
- Respiratory: 43% vs 42%
- Neurological: 16% vs 14%
- Sepsis: 13% vs 14%
- Trauma: 9% vs 10%
- Burns: <1% vs <1%
- BMI: 28 vs 28.6 kg/m2
- Clinical Frailty Score: 3 vs 3
- SARC-F: 1 vs 1
- Unintentional weight loss prior to admission: 15% vs 15%
- Chronic malabsorption prior to admission: 2% vs 2%
- Moderate or severe fat or muscle wasting prior to admission: 18% vs 17%
- AKI at randomisation: 25% vs 23%
- RRT at randomisation: 12% vs 9%
Intervention
- High protein dose
- 2.2 g/kg per day or more, of protein prescribed
Control
- Standard protein dose
- 1.2 g/kg per day or less, of protein prescribed
Management common to both groups
- Interventions carried out for up to 28 days (or until death, or transition to full oral feeding if either occurred prior to 28 days)
- Clinicians encouraged to achieve at least 80% of the protein prescription, which could be achieved through any combination of enteral or parenteral nutrition
- Clinicians were otherwise encouraged to follow published international guidelines and to avoid over-feeding
- Ideal body weight was used for patients with a BMI>30 kg/m2
- Total energy dose was not controlled
Outcome
- Adherence to study protocol: (high-dose vs normal dose protein)
- Time from ICU admit to starting EN: 16.0 vs 16.1 hours
- Type of nutrition:
- EN only: 92.6% vs 91.3%
- PN only: 0.5% vs 2%
- EN + PN: 7.0% vs 6.7%
- Daily protein received: 1.6 vs 0.9 g/kg/day
- Daily energy received: 14.7 vs 13.2 kcal/kg/day
- Patients receiving >110% prescribed energy: 1.6% vs 2.4%
- Primary efficacy endpoint: Incidence of alive hospital discharge by 60 days was not statistically different between groups
- 46.1% vs 50.2% (RR 0.91; 95% CI 0.77 to 1.07)
- Secondary efficacy endpoint: There was no significant difference in 60 day mortality between groups
- 34.6% vs 32.1% (RR 1.08; 95% CI 0.92 to 1.26)
- Tertiary efficacy endpoints: There were no significant differences in hospital mortality, duration of mechanical ventilation, ICU or hospital length of stay between groups
- Hospital mortality: 32.8% vs 28.8%
- Duration of mechanical ventilation: 6.1 vs 6.1 days
- ICU length of stay: 10.0 vs 9.4 days
- Hospital length of stay: 19.3 vs 18.9 days
- Subgroup analysis based on effect on time-to-discharge-alive from hospital: Patients in the high-dose protein group with concomitant AKI or high SOFA score had a worse outcome (HR, where <1.0 favours usual dose):
- AKI at baseline: 0.5 (p=0.001)
- SOFA ≥9 at baseline: 0.8 (p=0.05)
- Subgroup analysis based on effect on 60-day mortality: Patients in the high-dose protein group with concomitant AKI had a worse outcome (HR, where <1.0 favours high-dose):
- AKI at baseline: 1.4 (p=0.02)
Authors’ Conclusions
- “Delivering higher doses of protein to critically ill patients did not improve the time-to-discharge-alive from hospital compared with usual dose protein and might have worsened outcomes for patients with acute kidney injury and greater severity of illness.”
Strengths
- Multi-centre, randomised, controlled trial
- Computer-generated randomisation technique with permuted blocks ensured allocation concealment
- Baseline demographics were well matched between groups, including indices of pre-morbid nutritional status, primary diagnosis and acute physiology scores
- Good adherence to protocol and adequate separation of protein dose between groups enhances internal validity
- Modified intention to treat analysis with few exclusions and minimal attrition enhances internal validity
- Pragmatic design with broad inclusion criteria including patients from 85 ICUs across 16 countries enhances external validity
- Primary outcome was objectively defined, reducing the potential for detection bias in spite of clinician unblinding
- An important addition to our understanding of nutritional management in critically ill patients
Limitations
- Clinicians were not blinded to treatment allocation, which could introduce performance bias particularly with a time-to-discharge-alive endpoint that can be influenced by treating clinicians. The authors speculate, however, that this would be more likely to bias to the intervention arm given prevailing attitudes around high-protein feeding, and this was a negative study
- The study protocol and primary outcome were amended due to slow recruitment during the COVID-19 pandemic. The study was underpowered for the original primary outcome of 60-day mortality
- Certain subgroups of interest were under-represented in this study including trauma, surgery and burns
- Protein delivery was the same throughout the study period, and was not adapted for the theoretical stages of critical illness. The hypothesis that patients in later (and potentially anabolic) stages of critical illness may benefit from higher doses of protein was not tested
- The finding that patients with AKI or high acute physiology scores who receive high protein-dose nutrition have poorer outcomes is from subgroup analysis, and should be considered hypothesis generating
- Mortality, even as a secondary endpoint, may be considered problematic for nutrition trials, and there was no data collected regarding functional outcomes or quality of life in survivors. The hypothesis that high-dose protein may improve the experience of survivorship was not tested
The Bottom Line
- Well designed and executed pragmatic study with high internal and external validity, which reassures me that high-protein delivery for all-comers to ICU with at least one nutritional risk factor is not warranted, and may in fact worsen outcomes in sicker patients or those with AKI
- Future studies may focus on sub-groups that are hypothesised to benefit from high-dose protein such as obesity or trauma, patients in later stages of critical illness, or focus on functional outcomes
- The results support current ESPEN recommendations that 1.3g/kg/day protein can be delivered progressively in critical illness
External Links
Metadata
Summary author: Andrew Achilleos
Summary date: 9th February 2023
Peer-review editor: David Slessor
Picture by: Niklas on Unsplash
