Table 3 Studies reporting protein intake in critically ill patients

Larsson and colleagues [114]

Severely injured patients (burn or fracture of more than two long bones). Randomized during the first week of trauma (n = 39) to five different amounts of N from 0 to 0.3 g/kg/day

Prospective randomized study

Daily and cumulative N balance increased in the groups with a N intake of up to 0.2 g/kg/day versus the no N group (P <0.001)

Ishibashi and colleagues [15]

Immediate post-trauma patients (n = 18) or severely septic patients (n = 5) were divided into three groups (A, B, and C) receiving 1.1, 1.5, and 1.9 g/kg FFMc/day protein respectively

Retrospective study

Average loss of total body protein was 1.2 kg. Loss of body protein was greater in group A compared with groups B (P = 0.013) and C (P = 0.023). Protein loss in group B (1.5 g/kg FFMc/day), was half that of group A (1.1 g/kg FFMc/day). Protein loss in groups B and C was not different. An intake of 1.5 g/kg FFMc/day was equivalent to 1.0 g/day/kg body weight measured at the start of the study. Authors recommend the clinician obtains information on pre-illness bodyweight and prescribes 1.2 g/day/kg

Barr and colleagues [118]

200 ICU patients (npo >48 hours after their admission): 100 before implementation of a nutritional management protocol, 100 afterwards

Prospective evaluation

Risk of death was 56% lower in patients who received EN (HR: 0.44, 95% CI: 0.24, 0.80, P = 0.007)

Martin and colleagues [119]

499 ICU patients with an expected ICU stay of at least 48 hours. Introduction of evidence-based recommendations

Cluster-randomized controlled trial

Implementation of evidence-based recommendations led to more days of EN (P = 0.042), shorter mean hospital stay (P = 0.003) and a trend towards reduced mortality (P =0.058). The mean ICU stay did not differ significantly

Doig and colleagues [117]

1,118 patients in the ICU >2 days. Randomization to guideline or control groups. Guideline ICUs used an evidence-based guideline

Cluster-randomized controlled trial

Guideline ICU patients were fed earlier and reached nutritional goals more often compared with control subjects, but did not show significantly different hospital discharge mortality (P = 0.75), hospital LOS (P = 0.97), or ICU LOS (P = 0.42)

Alberda and colleagues [14]

2,772 mechanically ventilated patients. Prescribed and received energy was reported

Observational cohort study

Patients received only 56 to 64% of the nutritional prescription for energy and 50 to 65% for protein. Increased provision of energy and protein appear to be associated with improved clinical outcomes, particularly when BMI <25 or ≥35 kg/m². A 1,000 kcal increase is associated with improved mortality (P = 0.014) and more ventilation-free days (P = 0.003)

Strack van Schijndel and colleagues [120]

243 sequential mixed medical-surgical patients. Nutrition according to indirect calorimetry and at least 1.2 g protein/kg/day

Prospective observational cohort study

Reaching nutritional goals improves ICU (P = 0.027) and 28-day mortality (P = 0.005) and hospital survival (P = 0.04) in female patients. When only energy goals but not protein goals are met, ICU mortality is not changed. No differences could be observed for male patients

Casaer and colleagues [123]

4,640 ICU patients: 2,312 patients received PN within 48 hours after ICU admission, 2,328 patients received no PN before day 8

Randomized, multicenter trial

Early provision of PN shows a higher complication rate (26.2% vs 22.8% for ICU infections, P = 0.008), longer mechanical ventilation time (9.7% longer, P = 0.006) and renal replacement therapy (3 days' longer, P = 0.008), and a longer mean hospital duration (6.4% higher likelihood to discharge later, P = 0.04), but no significant impact on mortality

Weijs and colleagues [121]

886 mechanically ventilated patients; stratified into three groups: reaching energy and protein target; reaching energy target; and reaching no target

Prospective observational cohort study

Reaching the energy and protein target is associated with a 50% decrease in 28-day mortality. Reaching only the energy target is not associated with an improvement

Arabi and colleagues [124]

240 ICU patients randomly assigned to permissive underfeeding or target feeding

Randomized, controlled trial

Permissive underfeeding may be associated with lower mortality rates. Hospital mortality was lower in the permissive feeding group (30.0% vs 42.5%; relative risk: 0.71; 95% CI: 0.50, 0.99; P = 0.04). However, 28-day all-cause mortality was not significantly different between groups (18.3% vs 23.3%; relative risk: 0.79; 95% CI: 0.48, 1.29; P = 0.34)

Rice and colleagues [125]

200 mechanically ventilated patients with acute respiratory failure, expected to require mechanical ventilation for at least 72 hours randomized to receive initial trophic (10 ml/hour) or full-energy EN for the initial 6 days

Randomized, open-label study

Mortality to hospital discharge was 22.4% for trophic vs 19.6% for full energy (P = 0.62). The trophic group showed a trend for less diarrhea in the first 6 days (19% vs 24% of feeding days; P =0.08) and significantly fewer episodes of elevated gastric residual volumes (2% vs 8% of feeding days; P <0.001)

Singer and colleagues [126]

130 patients expected to stay in ICU >3 days. Randomization to EN with a target determined by indirect calorimetry (study group) or with 25 kcal/kg/day (control group)

Prospective, randomized, controlled trial

Patients in the study group had a higher mean energy (P =0.01) and protein intake (P =0.01) than the control group. They also showed a trend towards reduced mortality (32.3% vs 47.7%, P =0.058), but the number of infectious complications were higher (37 in the study vs 20 in the control group P =0.05)

Allingstrup and colleagues [122]

113 ICU patients. Analyzed according to provided amount of protein and AA

Prospective, observational, cohort study

In the low protein and AA provision group, the Kaplan-Meier survival probability was 49% on day 10, compared with 79% and 88% in the medium and high protein and AA groups on day 10, respectively

Rice and colleagues [127]

1,000 patients with acute lung injury requiring mechanical ventilation. Randomization to trophic or full enteral feeding for the first 6 days

Randomized, open-label, multicenter trial

Initial trophic feeding did not improve 60-day mortality (23.2% vs 22.2%, P =0.77) or infectious complications (P =0.72, P =0.77, and P =0.24 for ventilator-associated pneumonia, Clostridium difficile colitis and bacteremia, respectively) compared with full enteral feeding

Heidegger and colleagues [128]

ICU patients who had received less than 60% of their energy target from EN, were expected to stay >5 days, and to survive >7 days. Randomization to SPN (n =153) or EN (n =152). Protein administration was set to 1.2 g/kg ideal bodyweight/day during the study

Randomized controlled trial

Mean energy delivery between days 4 and 8 was higher for the SPN group (103% vs 77% of energy target). Nosocomial infections, between days 9 and 28, were more frequent in the EN group patients (38% vs 27%, P =0.0248). Overall nosocomial infections were not different