Abstract
Background
Research has sought to identify optimal protein doses during acute phases of critical illness to optimize outcomes.
Methods
A secondary analysis of the EFFORT Protein trial, which compared high vs usual protein (N = 1301). Only participants with 8 evaluable days of protein intake were included in our analysis. Mean protein intake was categorized as low (<0.8), medium (0.8–1.3), or high (>1.3 g/kg/day). Acute illness phases were define as early (days 1–4) and late (days 5–8). Participants were grouped by protein dose received in each phase. Based on prior evidence, early phase medium protein/late phase high protein served as the referent. The primary outcome was time‐to‐discharge‐alive; secondary outcomes included 60‐day mortality and discharge home.
Results
We identified 819 participants (median [IQR] age 59.0 [46.0, 69.0] years; 60% male). Time‐to‐discharge‐alive did not differ significantly across groups (P = 0.19). The early low/late high‐protein and early high/late high‐protein groups had hazard ratios of 0.63 (95% CI, 0.35–1.11) and 0.70 (95% CI, 0.46–1.06), respectively. Mortality and discharge‐home rates did not differ significantly across protein dose/acute phase groups (P = 0.85 and 0.65, respectively).
Conclusion
We hypothesized that early medium and late high protein would improve outcomes; however, no significant differences between were observed across protein dose/acute phase groups. These findings are hypothesis‐generating and highlight the need for future research to identify biomarkers or scoring tools that better define phase transitions in critical illness, enabling more precise nutrition strategies.
Trial Registration (Primary)
Keywords: acute phase, critically ill, enteral nutrition, mortality, nutrition support, protein
CLINICAL RELEVANCY STATEMENT
Providing adequate protein to critically ill patients is crucial to mitigate muscle wasting, support immune function, and improve outcomes. However, emerging evidence has suggested that early high protein delivery can be harmful, particularly in certain subpopulations. In this analysis, we evaluated combinations of protein provision during the early acute and late acute phases of critical illness and found no significant differences in outcomes based on timing or dose. These findings suggest that current phased‐based definitions of critical illness do not accurately reflect underlying metabolic changes. Biomarkers or scoring tools to identify phase transitions could help to guide tailored protein dosing and provide optimal nutrition therapy in the intensive care unit (ICU).
INTRODUCTION
Critical illness has been partitioned into the early acute, late acute, and late phases; the early acute phase is hallmarked by a hypercatabolic state and metabolic derangements, including proteolysis. 1 , 2 The late acute phase is characterized by severe muscle wasting but stable or normalizing metabolic alterations. 2 It is intuitive to provide protein to patients with critical illness, yet altered mitochondrial function during the early acute phase may compromise the ability to use any exogenous macronutrients being supplied to offset muscle wasting and harm may occur with high‐protein provisions resulting from increased urea cycle activity and toxic metabolite accumulation. 3 , 4
Recent studies have evaluated the association between protein dose during the early and late acute phases of critical illness and outcomes, such as mortality and duration of both intensive care unit (ICU) stay and mechanical ventilation. 5 , 6 , 7 , 8 , 9 , 10 , 11 The NUTRIREA‐3 trial reported a shorter duration of mechanical ventilation in the low‐protein and low‐energy group (0.2–0.4, 6 kcal/kg) given in the first 7 days but found no significant differences in 90‐day all‐cause mortality. 11 Although initially recommended, administration of early high‐protein doses lacks supporting evidence and may be harmful; a recent Bayesian meta‐analysis and systematic review with pooled mean higher and lower protein doses of 1.5 and 0.9 g/kg, respectively, observed a strong probability of increased mortality risk with higher doses. 5 , 6 , 7 , 8 , 9 , 10 , 12 , 13 , 14 Adjusting protein dose based on the phase of critical illness may have the most beneficial effect on outcomes. 8 , 9 , 10
Overall, the literature suggests a time‐dependent association between the dose of protein in the early and late acute phase of critical illness and mortality. 5 , 6 , 7 , 8 , 9 , 10 However, the research in this area has been largely conducted on heterogeneous groups of critically ill adults, whereas the EFFORT Protein trial included participants with nutrition risk, who were presumed to benefit from protein. 5 , 6 , 7 , 8 , 9 , 10 , 12 Therefore, using a nested cohort from a large, international randomized controlled trial (RCT), we aimed to investigate the impact of varying protein doses received (low <0.8, medium 0.8–1.3, or high >1.3 g/kg/day) between the early (days 1–4) and late (days 5–8) acute phases of critical illness on the primary outcome of time‐to‐discharge‐alive and secondary outcomes of 60‐day mortality and proportion discharged home. In participants with nutrition risk, we hypothesized that early phase medium‐protein/late phase high‐protein doses would be associated with decreased time‐to‐discharge‐alive in critically ill adults enrolled in the EFFORT Protein trial. 12
METHODS
Design and study population
We performed a posthoc secondary analysis of The EFFORT Protein trial, 12 a volunteer‐driven, pragmatic RCT in 85 ICUs across 16 countries, which compared the effect of high (≥2.2 g/kg/day) vs usual (≤1.2 g/kg/day) protein dosing on clinical outcomes in critically ill adults. The protocol for the study and the primary results have been previously described. 12 , 15 The Queen's University Research Ethics Committees approved the trial protocol (6021669). A granted waiver of informed consent was provided by the central institutional review board (IRB) at Vanderbilt University for enrolled hospitals that acceded to this central IRB, or local IRB approval was obtained when required. Participants were eligible for inclusion if they were adult patients (≥18 years), were within 96 h of admission to the ICU, had an anticipated duration of mechanical ventilation for ≥48 h after screening, and had at least one nutrition risk factor. 12 , 15 , 16 , 17 Individuals were excluded if they were expected to transition to end‐of‐life care within 7 days of screening, were pregnant, or had >96 h of mechanical ventilation prior to screening or if the clinician felt the participant was not appropriate for high or usual protein intake, protein dose was not achievable, or parenteral nutrition was the sole source of nutrition. 12 , 15
Baseline population characteristics included age, biological sex, weight, height, and body mass index (BMI). Clinical characteristics recorded include the Acute Physiology and Chronic Health Evaluation II score (APACHE II), Sequential Organ Failure Assessment (SOFA), and Charlson Comorbidity Index scores. 12 , 15 , 18 , 19 Nutrition risk scores were calculated using the modified Nutrition Risk in the Critically Ill (mNUTRIC) tool, along with Strength, Assistance in walking, Rising from a chair, Climbing stairs, and Falls (SARC‐F) and frailty scores. 16 , 17 , 20 Primary ICU admitting diagnosis was collected on admission. 12 , 15 The primary outcome was time to discharge alive, which was the time from random assignment until ICU day 60 or hospital discharge. 12 Secondary outcomes were 60‐day mortality and discharge to home vs other facilities. Sixty‐day mortality was defined as the occurrence of death by day 60 postrandomization. Discharge to home was defined as those who were able to be discharged to their homes vs another facility, such as rehab, long‐term care, and so forth. This analysis was completed on a subset of participants in The EFFORT Protein trial who had at least 8‐evaluable days of nutrition intake and remained alive and in hospital on day 9. Participants met the criteria for 8 days of evaluable nutrition intake if nutrition provision data were recorded, including participants who were temporarily ordered nothing by mouth or received their partial nutrition prescription. Participants were excluded if intake was considered “missing” for any days during the study period.
Nutrition intervention and protein dose categories
In the primary study, the intervention was initiated within 96 h of admission and maintained until day 28 or until transition to oral diet or death. Daily nutrition records were kept documenting the actual amount of protein delivered each day for 28 days. Protein intake was dosed using pre‐ICU actual or estimated dry body weight, or an ideal body weight (equal to a BMI of 25 [calculated as weight in kilograms divided by meters squared]) if BMI was ≥30. 12 , 15 This analysis used the amount of protein received, as documented by the study clinicians in both the early and late acute phases, instead of the randomized protein prescription from the primary study.
Protein doses received were then classified into three categories: low: <0.8 g/kg/day, medium: 0.8–1.3 g/kg/day, and high: >1.3 g/kg/day, which were similar to dosing structures previously reported. 8 , 9 , 10 As categorized in other studies, we defined the early acute phase as ICU days 1–4 and the late acute phase as ICU days 5–8. 8 , 9 , 10 Nine protein dose/acute phase groups were assembled using combinations of protein doses received across the two acute phases. For example, a protein dose of <0.8 g/kg/day (low) in the early acute phase followed by a dose of >1.3 g/kg/day (high) in the late acute phase, was labeled as “Early Phase Low Protein, Late Phase High Protein” (Table SA1). Our comparator, or referent, of interest was “Medium‐High” (MH), defined as early phase medium‐protein dose and late phase high‐protein dose, which aligned with previously published literature. 8 , 9 , 10 Our exposures of interest were the remaining eight protein dose/acute phase groups.
Statistical analysis
Characteristics of the participants were compared between the three early dose groups by using the Kruskal‐Wallis test for continuous variables and the chi‐square test for categorical variables. We modeled our primary outcome, time‐to‐discharge‐alive, as the subdistribution hazard ratio of time‐to‐discharge‐alive using a competing risk time‐to‐event analysis in which death was considered a competing risk precluding the possibility of live discharge based on the Fine and Gray extension to the Cox proportional hazard. 21 We included a random shared frailty to account for the effect of heterogeneity across different ICUs, as implemented in the SAS PHREG procedure. We used logistic regression to analyze our secondary outcomes, 60‐day mortality and discharge home vs other facilities. However, we found that the random ICU variance was estimated at zero, which caused some convergence issues. Therefore, we ran the model without ICU as a random effect, but since the covariates included region, there was some control for variability between ICUs. We accounted for a lag effect by separating the exposure period (the first 8 days) and the outcome period by at least 1 day to reduce the risk of reverse causality, for example, low‐protein intake in the last days alive being the consequence of illness severity vs causing it. The modeling controlled for prespecified baseline covariates: age, APACHE II, clinical frailty, mNUTRIC, and SARC‐F scores, ICU admitting diagnosis, and geographical region (adjusted model 1). A second adjusted model (adjusted model 2) included the same covariates with the addition of nonprotein energy received from nutrition support and propofol. All continuous covariates were modeled as linear. Site was modeled as a random effect, but all other effects were fixed. The early phase medium/late phase high‐protein dose used as our referent, aligned with our hypothesis and prior literature suggesting that doses of 0.8–1.2 g/kg in the early phase were associated with reduced mortality and fewer days of mechanical ventilation, and doses >1.2 g/kg did not appear to have clinical benefit until the late acute phase. 8 , 9 , 10 Groups were compared with the referent by subdistribution hazard ratios (time‐to‐discharge‐alive) or odds ratios (binary outcomes) with corresponding 95% CIs. P values testing the equivalence of all nine groups were also provided.
Because of missing data for APACHE II (n = 60), mNUTRIC (n = 60), frailty (n = 61), and SARC‐F (n = 79) scores, we used multiple imputations to generate 50 imputed datasets with the missing values imputed using a fully conditional regression specification based on all the remaining variables in the analysis as implemented by the SAS MI procedure. 22 We ran each analysis on all 50 imputed datasets and then combined results according to Rubin's rules as implemented in the SAS MIANALYZE procedure. 23
Since the primary study results found high‐protein dose was associated with worse outcomes for participants with baseline acute kidney injury, 12 , 24 we performed a sensitivity analysis that removed participants with acute kidney injury at baseline. P < 0.05 was considered statistically significant, and statistical analysis was conducted using SAS version 9.4, except forest plots were generated using R software.
RESULTS
Population and protein dose groups
Of the 1301 participants analyzed in the EFFORT Protein trial, 819 met the inclusion criteria for this analysis (Figure 1). The baseline characteristics of the study population are presented in Table 1. The median age of this cohort was 59.0 years [46.0, 69.0], and 60% of participants were male. The admission category was primarily medical (82.2%), and diagnoses were largely respiratory (43.2%). The baseline demographics of the study cohort were comparable to those of the overall EFFORT trial population, with no major differences observed.
Figure 1.
Patient flow diagram. *Protein dose received at ICU days 1–4. ICU, intensive care unit.
Table 1.
Patient characteristics by early acute delivered protein category.
| Characteristic | Overall (n = 819) | Low protein (n = 269) | Medium protein (n = 302) | High protein (n = 248) | P value |
|---|---|---|---|---|---|
| Age (years) | 59.0 (46.0–69.0) | 58.0 (45.0–70.0) | 59.0 (47.0–69.0) | 59.0 (46.0–69.0) | 0.73 |
| Sexa | 0.20 | ||||
| Male | 491 (60.0%) | 152 (56.5%) | 192 (63.6%) | 147 (59.3%) | |
| Female | 327 (39.9%) | 117 (43.5%) | 109 (36.1%) | 101 (40.7%) | |
| Admission category | 0.09 | ||||
| Medical | 673 (82.2%) | 208 (77.3%) | 253 (83.8%) | 212 (85.5%) | |
| Surgical Elective | 28 (3.4%) | 14 (5.2%) | 7 (2.3%) | 7 (2.8%) | |
| Surgical Emergency | 118 (14.4%) | 47 (17.5%) | 42 (13.9%) | 29 (11.7%) | |
| Primary ICU diagnosis | 0.13 | ||||
| Cardiovascular/Vascular | 58 (7.1%) | 17 (6.3%) | 24 (7.9%) | 17 (6.9%) | |
| Respiratory | 354 (43.2%) | 116 (43.1%) | 136 (45.0%) | 102 (41.1%) | |
| Gastrointestinal | 35 (4.3%) | 20 (7.4%) | 11 (3.6%) | 4 (1.6%) | |
| Neurologic | 140 (17.1%) | 38 (14.1%) | 43 (14.2%) | 59 (23.8%) | |
| Sepsis | 103 (12.6%) | 40 (14.9%) | 35 (11.6%) | 28 (11.3%) | |
| Trauma | 78 (9.5%) | 20 (7.4%) | 34 (11.3%) | 24 (9.7%) | |
| Others* | 51 (6.2%) | 18 (0.07%) | 19 (0.06%) | 14 (0.06%) | |
| BMI, kg/m2 | 26.1 (22.7–32.3) | 26.7 (23.1–32.4) | 26.1 (22.7–32.3) | 25.9 (22.2–31.8) | 0.59 |
| Charlson comorbidity index | 0.9 ± 1.4 (0.0–11.0) | 1.1 ± 1.6 (0.0–8.0) | 0.8 ± 1.2 (0.0–6.0) | 0.8 ± 1.4 (0.0–11.0) | 0.08 |
| SOFA score | 8.7 ± 3.8 (0.0–23.0) | 8.4 ± 4.2 (0.0–22.0) | 8.8 ± 3.6 (0.0–21.0) | 8.8 ± 3.7 (0.0–23.0) | 0.36 |
| 9.0 (6.0–11.0) | 8.0 (5.0–11.0) | 9.0 (6.0–11.0) | 9.0 (6.0–11.0) | ||
| APACHE II scorea | 20.8 ± 7.8 (3.0–51.0) | 21.1 ± 8.4 (3.0–45.0) | 20.9 ± 7.4 (6.0–42.0) | 20.3 ± 7.7 (6.0–51.0) | 0.41 |
| 20.0 (15.0–26.0) | 20.0 (14.0–27.0) | 20.0 (15.0–26.0) | 20.0 (15.0–25.0) | ||
| mNUTRIC scoreb | 4.4 ± 2.0 (0.0–9.0) | 4.4 ± 2.0 (0.0–9.0) | 4.5 ± 2.0 (0.0–9.0) | 4.2 ± 2.0 (0.0–8.0) | 0.42 |
| 5.0 (3.0–6.0) | 5.0 (3.0–6.0) | 5.0 (3.0–6.0) | 4.0 (3.0–6.0) | ||
| Frailtyc | 3.4 ± 1.7 (1.0–8.0) | 3.6 ± 1.8 (1.–8.0) | 3.2 ± 1.7 (1.0–8.0) | 3.3 ± 1.6 (1.0–7.0) | 0.02 |
| 3.0 (2.0–5.0) | 3.0 (2.0–5.0) | 3.0 (2.0–4.0) | 3.0 (2.0–4.0) | ||
| Frailty ≥ 5 | 191 (23.3%) | 76 (28.3%) | 58 (19.2%) | 57 (23.0%) | 0.04 |
| SARC‐F scored | 2.4 ± 2.9 (0.0–10.0) | 2.8 ± 3.0 (0.0–10.0) | 2.1 ± 2.8 (0.0–10.0) | 2.4 ± 2.9 (0.0–10.0) | 0.002 |
| 1.0 (0.0–5.0) | 2.0 (0.0–5.0) | 0.0 (0.0–4.0) | 1.0 (0.0–4.0) | ||
| Acute kidney injury at baseline | |||||
| Yes | 195 (23.8%) | 80 (29.7%) | 64 (21.2%) | 51 (20.6%) | 0.02 |
| No | 624 (76.2%) | 189 (70.3%) | 238 (78.8%) | 197 (79.4%) |
Note : Values reported as n (%), mean ± SD (min–max), or median (IQR). Kruskal‐Wallis test was used for continuous variables and chi‐square test was used for categorical variables. Statistically significant P values < 0.05 are shown in bold.
Abbreviations: APACHE, Acute Physiology and Chronic Health Evaluation; mNUTRIC, modified Nutrition Risk in the Critically Ill; SOFA, Sequential Organ Failure Assessment; SARC‐F, Strength, Assistance in walking, Rise from a chair, Climb stairs, and Falls.
One patient had missing data for biological sex; therefore, the percentage may not add up to the total number.
n = 761 (overall), 235 (low), 289 (medium), 237 (high).
n = 758 (overall), 250 (low), 277 (medium), 231 (high).
n = 740 (overall), 247 (low), 268 (medium), 225 (high).
During the early acute phase (Days 1–4), 32.8% of participants were in the low‐protein group, 36.8% were in the medium‐protein group, and 30.2% were in the high‐protein group. Participant characteristics were mostly similar between the early protein groups, except frailty, SARC‐F scores, and frequency of acute kidney injury were highest in the low‐protein group (Table 1). Median daily intakes of energy and protein during both phases are reported by protein dose combination groups in Supporting Information S1: Table SA1.
Primary and secondary outcomes
Time‐to‐discharge‐alive did not differ significantly among groups (P = 0.19, adjusted model 2). Time‐to‐discharge‐alive hazard ratios compared to the referent are plotted in Figure 2. The referent group had the lowest mortality rate (Figure 3) and the highest rate of discharge home (Figure 4), but the differences were not statistically significant (P = 0.85 and 0.65, respectively, adjusted model 2). All results with and without adjustment are reported in Supporting Information S1: Tables SA2a, SA3a, and SA4a.
Figure 2.
Time‐to‐discharge‐alive between protein dose and acute phase groups. *P value based on adjusted model 2 (covariates + nonprotein energy). Outcome assessed at day 60 (totals may not equal N, as some patients remained hospitalized). HH, early phase high protein/late phase high protein; HL, early phase high protein/late phase low protein; HM, early phase high protein/late phase medium protein; LH, early phase low protein/late phase high protein; LL, early phase low protein/late phase low protein; LM, early phase low protein/late phase medium protein; MH, early phase medium protein/late phase high protein (referent); ML, early phase medium protein/late phase low protein; MM, early phase medium protein/late phase medium protein.
Figure 3.
Sixty‐day mortality between protein dose and acute phase groups. *P value based on adjusted model 2 (covariates + nonprotein energy). HH, early phase high protein/late phase high protein; HL, early phase high protein/late phase low protein; HM, early phase high protein/late phase medium protein; LH, early phase low protein/late phase high protein; LL, early phase low protein/late phase low protein; LM, early phase low protein/late phase medium protein; MH, early phase medium protein/late phase high protein (referent); ML, early phase medium protein/late phase low protein; MM, early phase medium protein/late phase medium protein.
Figure 4.
Discharge to home vs other between protein dose and acute phase groups. *P value based on adjusted model 2 (covariates + nonprotein energy). HH, early phase high protein/late phase high protein; HL, early phase high protein/late phase low protein; HM, early phase high protein/late phase medium protein; LH, early phase low protein/late phase high protein; LL, early phase low protein/late phase low protein; LM, early phase low protein/late phase medium protein; MH, early phase medium protein/late phase high protein (referent); ML, early phase medium protein/late phase low protein; MM, early phase medium protein/late phase medium protein.
Sensitivity analysis
Excluding participants with acute kidney injury did not alter the trend that the referent nutrition group consistently had better outcomes, although there were no statistically significant differences in outcomes (Tables SA2b, SA3b, SA4b).
DISCUSSION
In this post hoc analysis of one of the largest randomized trials examining protein dosing in critically ill adults, we evaluated associations between protein doses received during the early and late acute phases of critical illness and important clinical outcomes, including time‐to‐discharge‐alive and 60‐day mortality. Across all protein dose combinations, we observed no statistically significant associations with any of these outcomes. Although point estimates varied across protein groups, wide and overlapping CIs limited the interpretation of these differences.
In our analysis, participants who received early phase, medium‐protein doses (0.8–1.3 g/kg) followed by increasing to the high dose in the later phase showed a lower risk of 60‐day mortality compared with the other groups, although the finding was not significant. Compared with the referent dose, low doses of protein in both phases seemed to yield worse outcomes. Participants who received the early phase low‐protein/late phase high‐protein dose experienced the slowest time‐to‐discharge‐alive. This finding contrasts with earlier studies suggesting that protein intakes of 0.8–1.2 g/kg in the early phase and >1.2 g/kg in the late phase were most associated with improved outcomes. 8 , 9 , 10 However, in our cohort, the early phase, low‐protein group overall had a higher prevalence of emergency surgeries, acute kidney injury at baseline, and elevated frailty and sarcopenia scores. They also received considerably less energy overall. These factors may reflect greater severity of illness and could have influenced both protein delivery and clinical outcomes. Additionally, the recent PRECISe trial, which implemented a stepped increase in protein delivery, reported a significantly longer time‐to‐discharge‐alive in high protein (2 g/kg) compared with the standard group (1.3 g/kg). 25
The time‐points defining the phases of critical illness are arbitrary. In reality, critically ill adults represent a heterogenous group, and it is unclear when they transition between the proposed phases of critical illness. Objective biomarkers identifying transitions between the phases of critical illness are needed to support a biological rationale to modify protein dose.
The Swiss EFFORT trial measured C‐reactive protein (CRP) levels and their association with outcomes based on nutrition support interventions. 11 , 26 In a secondary analysis of the Swiss EFFORT trial, Merker et al found that the benefits of nutrition delivery were not observed in participants with CRP concentrations of >100 mg/L. 26 Because of the volunteer‐driven design of the EFFORT Protein Trial, obtaining CRP measurements across sites worldwide was not feasible. Whether serum CRP concentrations impacted outcomes based on protein dose is unknown, and future studies are needed to evaluate these associations.
Our post hoc analysis had several limitations. First, although our overall sample size was large, some of the nine protein groups were small. Therefore, there is a risk of type II error and limitations when adjusting for baseline characteristics. The wide CIs for some groups, such as early phase high/late phase low protein (HL), and early phase high/late phase medium protein (HM) also indicate poor precision and limits the ability to detect small or moderate differences from the referent group, early phase medium/late phase high protein (MH). Second, the protein dose groups were based on the dose delivered rather than the assigned dose. Unknown confounders or reverse causality could limit our ability to make causal inferences. By implementing a lag period and including only patients who remained alive and in the ICU for at least 1 day following the exposure period to reduce the risk of reverse causality, we acknowledge that this approach may introduce survivor bias. Additionally, our analysis adjusted for multiple potential confounders, including nonprotein energy intake. However, we acknowledge that adequacy of nutrition support is frequently associated with illness severity, as observed in our low‐protein group. As with all observational studies, residual confounding by illness severity may persist despite statistical adjustment. Third, since ICU populations are heterogeneous, there may be some subgroups that benefit from a different protein dose approach. Because of the inclusion criteria of patients being at nutrition risk, we cannot generalize to those without nutrition risk. Fourth, our study did not account for exercise, which may be important to synergize the effects of protein. In an RCT, de Azevedo et al found that protein doses of 1.48 g/kg and twice‐daily cycle ergometry, compared with a protein dose of 1.21 g/kg protein and standard physiotherapy, had significantly lower rates of ICU‐acquired weakness (P = 0.05). 27 Future research investigating early physiotherapy combined with different protein doses on outcomes is warranted.
Our study had notable strengths. First, this was a secondary analysis of one of the largest randomized trials evaluating protein doses in critically ill adults. Second, our critically ill study population was diverse, including 819 participants from 85 ICUs across 16 countries, enhancing the generalizability of findings. Third, we used a pragmatic approach to define protein dose groups based on actual delivered intake, reflecting real‐world practice. Fourth, our analysis adjusted for multiple potential confounders, including energy intake delivered, which mitigates the risk of confounding by energy dose. Finally, we examined protein dose trajectories across acute phases (early vs late), providing insights into how timing and dose may interact—an area of growing clinical interest.
CONCLUSION
Over the past decade, critical care nutrition research has focused on identifying optimal protein and energy prescriptions. More recently, and under the premise that critical care is partitioned into different phases, research has aimed to identify how nutrition needs may vary across these phases. Based on literature at the time of study initiation, we hypothesized that providing a medium dose of protein (0.8–1.3 g/kg/day) in the early phase, followed by a higher dose (>1.3 g/kg/day) in the later acute phase, would lead to improved clinical outcomes. However, more recent trials published since then have challenged this hypothesis. 11 , 14 , 25 Our post hoc analysis of the EFFORT Protein trial did not demonstrate significant differences in outcomes between the protein dose combinations evaluated across the acute phases of critical illness. Given the inherent limitations discussed, these findings should be considered hypothesis‐generating. Future research is warranted to highlight which patients may benefit most from targeted protein delivery strategies and identify biomarkers and validated scoring tools that define phase transitions in critical illness, enabling more precise and optimal nutrition care.
AUTHOR CONTRIBUTIONS
Lauren E. Tweel: Formal analysis; methodology; writing—original draft; writing—review and editing. Rebecca Brody: Writing—review and editing; supervision; methodology. Hamed Samavat: Methodology; writing—review and editing. Andrew G. Day: Methodology; formal analysis; writing—review and editing. Xuran Jiang: Formal analysis; methodology. Laura Byham‐Gray: Methodology; supervision; formal analysis; writing—review and editing. Jayshil Patel: Writing—review and editing; methodology. Charlene Compher: Conceptualization; methodology; writing—review and editing; investigation. Daren K. Heyland: Conceptualization; writing—review and editing; investigation.
CONFLICT OF INTERESTS STATEMENT
Dr Compher serves in an unpaid leadership role for the American Society for Parenteral and Enteral Nutrition and an unpaid leadership role for the ASPEN Rhoads Research Foundation. Dr Brody serves in an unpaid leadership role for the Academy of Nutrition and Dietetics Ethics Committee.
ETHICS APPROVAL STATEMENT
“The investigator‐initiated trial protocol was approved by the Research Ethics Committees of Queen's University, Canada, and a central institutional review board at Vanderbilt University, TN, USA that granted a waiver of informed consent for sites that acceded to this central institutional review board. Otherwise, where required by local study sites, local ethics approval was obtained, and informed consent was also obtained from designated participant surrogates before randomisation.” Heyland et al. 1
Supporting information
Table SA1. Groups of protein doses combined with acute critical illness phases (N=819). Table SA2a. Hazard ratio of time‐to‐discharge‐alive compared to medium–high‐protein intake group. Table SA2b. Time‐to‐discharge‐alive compared to medium–high‐protein intake group with acute kidney injury patients removed. Table SA3a. Odds ratio of 60‐day mortality compared to medium–high‐protein intake group. Table SA3b. Odds of 60‐day mortality compared to medium–high‐protein intake group with AKI patients removed. Table SA4a. Odds ratio of discharge to home compared to medium–high‐protein intake group. Table SA4b. Odds ratio of discharge to home compared to medium–high‐protein intake group (AKI patients removed). Table SA5. Patient characteristics at baseline—nine groups.
ACKNOWLEDGMENTS
We wish to thank our colleagues at the EFFORT Protein trial sites who aided in data collection and interpretation. A complete list of participating sites can be found in the original publication.
Tweel LE, Brody R, Samavat H, et al. Association between protein dose in the early and late acute phases of critical illness and time‐to‐discharge‐alive: a secondary analysis of a randomized clinical trial. J Parenter Enteral Nutr. 2026;50:183‐191. 10.1002/jpen.70033
REFERENCES
- 1. Sharma K, Mogensen KM, Robinson MK. Pathophysiology of critical illness and role of nutrition. Nutr Clin Pract. 2019;34(1):12‐22. 10.1002/ncp.10232 [DOI] [PubMed] [Google Scholar]
- 2. Singer P, Blaser AR, Berger MM, et al. ESPEN guideline on clinical nutrition in the intensive care unit. Clin Nutr. 2019;38(1):48‐79. 10.1016/j.clnu.2018.08.037 [DOI] [PubMed] [Google Scholar]
- 3. De Waele E, Malbrain MLNG, Spapen H. Nutrition in sepsis: a bench‐to‐bedside review. Nutrients. 2020;12(2):395. 10.3390/nu12020395 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Haines RW, Prowle JR, Day A, Bear DE, Heyland DK, Puthucheary Z. Association between urea trajectory and protein dose in critically ill adults: a secondary exploratory analysis of the effort Protein trial (RE‐EFFORT). Crit Care. 2024;28(1):24. 10.1186/s13054-024-04799-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Allingstrup MJ, Kondrup J, Wiis J, et al. Early goal‐directed nutrition versus standard of care in adult intensive care patients: the single‐centre, randomised, outcome assessor‐blinded EAT‐ICU trial. Intensive Care Med. 2017;43(11):1637‐1647. 10.1007/s00134-017-4880-3 [DOI] [PubMed] [Google Scholar]
- 6. Braunschweig CA, Sheean PM, Peterson SJ, et al. Intensive nutrition in acute lung injury: a clinical trial (INTACT). JPEN J Parenter Enteral Nutr. 2015;39(1):13‐20. 10.1177/0148607114528541 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Braunschweig CL, Freels S, Sheean PM, et al. Role of timing and dose of energy received in patients with acute lung injury on mortality in the Intensive Nutrition in Acute Lung Injury Trial (INTACT): a post hoc analysis. Am J Clin Nutr. 2017;105(2):411‐416. 10.3945/ajcn.116.140764 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Hartl WH, Kopper P, Bender A, et al. Protein intake and outcome of critically ill patients: analysis of a large international database using piece‐wise exponential additive mixed models. Crit Care. 2022;26(1):7. 10.1186/s13054-021-03870-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Koekkoek WAC, van Setten CH, Olthof LE, Kars JCN, van Zanten ARH. Timing of PROTein INtake and clinical outcomes of adult critically ill patients on prolonged mechanical VENTilation: the PROTINVENT retrospective study. Clin Nutr. 2019;38(2):883‐890. 10.1016/j.clnu.2018.02.012 [DOI] [PubMed] [Google Scholar]
- 10. Matejovic M, Huet O, Dams K, et al. Medical nutrition therapy and clinical outcomes in critically ill adults: a European multinational, prospective observational cohort study (EuroPN). Crit Care. 2022;26(1):143. 10.1186/s13054-022-03997-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Reignier J, Plantefeve G, Mira JP, et al. Low versus standard calorie and protein feeding in ventilated adults with shock: a randomised, controlled, multicentre, open‐label, parallel‐group trial (NUTRIREA‐3). Lancet Respir Med. 2023;11(7):602‐612. 10.1016/S2213-2600(23)00092-9 [DOI] [PubMed] [Google Scholar]
- 12. Heyland DK, Patel J, Compher C, et al. The effect of higher protein dosing in critically ill patients with high nutritional risk (EFFORT Protein): an international, multicentre, pragmatic, registry‐based randomised trial. Lancet. 2023;401(10376):568‐576. 10.1016/S0140-6736(22)02469-2 [DOI] [PubMed] [Google Scholar]
- 13. Taylor BE, McClave SA, Martindale RG, et al. Guidelines for the provision and assessment of nutrition support therapy in the adult critically ill patient: Society of Critical Care Medicine (SCCM) and American Society for Parenteral and Enteral Nutrition (A.S.P.E.N.) Crit Care Med. 2016;44(2):390‐438. 10.1097/CCM.0000000000001525 [DOI] [PubMed] [Google Scholar]
- 14. Heuts S, Lee ZY, Lew CCH, et al. Higher versus lower protein delivery in critically ill patients: a systematic review and bayesian meta‐analysis. Crit Care Med. 2025;53(3):e645‐e655. 10.1097/CCM.0000000000006562 [DOI] [PubMed] [Google Scholar]
- 15. Heyland DK, Patel J, Bear D, et al. The effect of higher protein dosing in critically ill patients: a multicenter registry‐based randomized trial: the EFFORT trial. JPEN J Parenter Enteral Nutr. 2019;43(3):326‐334. 10.1002/jpen.1449 [DOI] [PubMed] [Google Scholar]
- 16. Malmstrom TK, Morley JE. SARC‐F: a simple questionnaire to rapidly diagnose sarcopenia. J Am Med Dir Assoc. 2013;14(8):531‐532. 10.1016/j.jamda.2013.05.018 [DOI] [PubMed] [Google Scholar]
- 17. Rockwood K. A global clinical measure of fitness and frailty in elderly people. Can Med Assoc J. 2005;173(5):489‐495. 10.1503/cmaj.050051 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med. 1985;13(10):818‐829. [PubMed] [Google Scholar]
- 19. Vincent JL, Moreno R, Takala J, et al. The SOFA (Sepsis‐related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis‐Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med. 1996;22(7):707‐710. 10.1007/bf01709751 [DOI] [PubMed] [Google Scholar]
- 20. Rahman A, Hasan RM, Agarwala R, Martin C, Day AG, Heyland DK. Identifying critically‐ill patients who will benefit most from nutritional therapy: further validation of the “modified NUTRIC” nutritional risk assessment tool. Clin Nutr. 2016;35(1):158‐162. 10.1016/j.clnu.2015.01.015 [DOI] [PubMed] [Google Scholar]
- 21. Fine JP, Gray RJ. A Proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94(446):496‐509. 10.1080/01621459.1999.10474144 [DOI] [Google Scholar]
- 22. van Buuren S. Multiple imputation of discrete and continuous data by fully conditional specification. Stat Methods Med Res. 2007;16(3):219‐242. 10.1177/0962280206074463 [DOI] [PubMed] [Google Scholar]
- 23. Rubin DB. Multiple Imputation for Nonresponse in Surveys. Wiley Series in Probability and Mathematical Statistics Applied Probability and Statistics. Wiley; 1987:xxix, 258pp.
- 24. Stoppe C, Patel JJ, Zarbock A, et al. The impact of higher protein dosing on outcomes in critically ill patients with acute kidney injury: a post hoc analysis of the EFFORT Protein trial. Crit Care. 2023;27(1):399. 10.1186/s13054-023-04663-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Bels JLM, Thiessen S, van Gassel RJJ, et al. Effect of high versus standard protein provision on functional recovery in people with critical illness (PRECISe): an investigator‐initiated, double‐blinded, multicentre, parallel‐group, randomised controlled trial in Belgium and the Netherlands. Lancet. 2024;404(10453):659‐669. 10.1016/S0140-6736(24)01304-7 [DOI] [PubMed] [Google Scholar]
- 26. Merker M, Felder M, Gueissaz L, et al. Association of baseline inflammation with effectiveness of nutritional support among patients with disease‐related malnutrition: a secondary analysis of a randomized clinical trial. JAMA Network Open. 2020;3(3):e200663. 10.1001/jamanetworkopen.2020.0663 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. de Azevedo JRA, Lima HCM, Frota PHDB, et al. High‐protein intake and early exercise in adult intensive care patients: a prospective, randomized controlled trial to evaluate the impact on functional outcomes. BMC Anesthesiol. 2021;21(1):283. 10.1186/s12871-021-01492-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table SA1. Groups of protein doses combined with acute critical illness phases (N=819). Table SA2a. Hazard ratio of time‐to‐discharge‐alive compared to medium–high‐protein intake group. Table SA2b. Time‐to‐discharge‐alive compared to medium–high‐protein intake group with acute kidney injury patients removed. Table SA3a. Odds ratio of 60‐day mortality compared to medium–high‐protein intake group. Table SA3b. Odds of 60‐day mortality compared to medium–high‐protein intake group with AKI patients removed. Table SA4a. Odds ratio of discharge to home compared to medium–high‐protein intake group. Table SA4b. Odds ratio of discharge to home compared to medium–high‐protein intake group (AKI patients removed). Table SA5. Patient characteristics at baseline—nine groups.