Comments
Abstract
Rationale: The optimal nutritional strategy for critically ill adults at high nutritional risk is unclear.
Objectives: To examine the effect of permissive underfeeding with full protein intake compared with standard feeding on 90-day mortality in patients with different baseline nutritional risk.
Methods: This is a post hoc analysis of the PermiT (Permissive Underfeeding versus Target Enteral Feeding in Adult Critically Ill Patients) trial.
Measurements and Main Results: Nutritional risk was categorized by the modified Nutrition Risk in Critically Ill score, with high nutritional risk defined as a score of 5–9 and low nutritional risk as a score of 0–4. Additional analyses were performed by categorizing patients by body mass index, prealbumin, transferrin, phosphate, urinary urea nitrogen, and nitrogen balance. Based on the Nutrition Risk in Critically Ill score, 378 of 894 (42.3%) patients were categorized as high nutritional risk and 516 of 894 (57.7%) as low nutritional risk. There was no association between feeding strategy and mortality in the two categories; adjusted odds ratio (aOR) of 0.84 (95% confidence interval [CI], 0.56–1.27) for high nutritional risk and 1.01 (95% CI, 0.64–1.61) for low nutritional risk (interaction P = 0.53). Findings were similar in analyses using other definitions, with the exception of prealbumin. The association of permissive underfeeding versus standard feeding and 90-day mortality differed when patients were categorized by baseline prealbumin level (≤0.10 g/L: aOR, 0.57 [95% CI, 0.31–1.05]; >0.10 and ≤0.15 g/L: aOR, 0.79 [95% CI, 0.42–1.48]; >0.15 g/L: aOR, 1.55 [95% CI, 0.80, 3.01]; interaction P = 0.009).
Conclusions: Among patients with high and low nutritional risk, permissive underfeeding with full protein intake was associated with similar outcomes as standard feeding.
Should nutritional support be different among critically ill patients based on their baseline nutritional status? There is a dearth of randomized controlled trial data regarding the optimal nutritional strategy in patients at high nutritional risk.
Among patients with high and low nutritional risk, permissive underfeeding with full protein intake was associated with similar 90-day mortality as standard feeding. Neither the Nutrition Risk in Critically Ill score nor other baseline nutritional variables could identify who may benefit from full caloric intake.
Recent randomized controlled trials (RCTs) of restricted caloric intake compared with standard feeding showed no difference in mortality in critically ill patients (1–3). However, it has been suggested that caloric intake should be targeted based on the underlying nutritional status. Observational studies have found that full caloric intake was associated with improved outcomes in critically ill adults at high nutritional risk (4, 5). Conversely, other observational studies have suggested that malnourished patients are at high risk of refeeding syndrome with its adverse consequences, and thus recommend a restricted caloric strategy for these patients (6). There is a dearth of RCT data regarding the optimal nutritional strategy in patients at high nutritional risk. Only one RCT has compared restricted with standard caloric intake in critically ill patients with refeeding syndrome, defined as hypophosphatemia within 72 hours of initiation of nutrition (7). Although the number of days alive after intensive care unit (ICU) discharge did not differ between the two groups, caloric restriction was associated with more patients alive at Day 60 and longer overall survival. However, whether these patients were at high nutritional risk is unclear, because the average body mass index (BMI) in the study population was 28 kg/m2 and only 3% of patients had a BMI less than 18 kg/m2. Therefore, the optimal caloric intake during critical illness in patients at high nutritional risk remains unclear.
A major challenge in tailoring nutrition in critically ill patients is the lack of consensus regarding the definition of nutritional risk (8). The Nutrition Risk in Critically Ill (NUTRIC) score has been proposed to quantify nutritional risk and identify critically ill patients most likely to benefit from aggressive nutrition therapy (4). The score includes age, Acute Physiology and Chronic Health Evaluation II score, Sequential Organ Failure Assessment score, number of comorbidities, days from hospital admission to ICU admission, and IL-6. A modified version of the NUTRIC score which excludes IL-6 has been validated in an observational study; the total score ranges from 0 to 9 with increasing scores indicating higher nutritional risk (5). A strong association was found between nutritional adequacy and 28-day survival in patients with a high NUTRIC score; this association diminished with decreasing NUTRIC score (5). Based on this finding, the NUTRIC score has been proposed to identify critically ill patients most likely to benefit from optimal amounts of macronutrients (5, 8). Other commonly used measures of nutritional risk include BMI, prealbumin (also called transthyretin), transferrin, serum phosphate, urinary urea nitrogen (UUN), and nitrogen balance. With the exception of phosphate (7) the value of these measures in guiding nutritional support has not been examined in an RCT setting (9–11).
The objective of this analysis was to examine the effect of permissive caloric underfeeding compared with standard feeding on 90-day mortality in critically ill adults stratified by baseline nutritional status, using the NUTRIC score and other common measures of nutritional risk.
This was a post hoc analysis of the PermiT (Permissive Underfeeding versus Target Enteral Feeding in Adult Critically Ill Patients) trial, an unblinded pragmatic RCT conducted in seven tertiary-care centers in Saudi Arabia and Canada (ISRCTN Registry: ISRCTN68144998) between November 2009 and September 2014. The institutional review boards of the participating centers approved the study, and informed consent was obtained from surrogate decision makers before enrollment. In this trial, 894 patients were randomized to permissive underfeeding (with a goal of 40–60% caloric requirement) or standard feeding (with a goal of 70–100% caloric requirement) with a goal of similar protein intake (1.2–1.5 g/kg/day) in both groups. Complete details of the study protocol and results have been previously published (2, 12). The trial found no difference in the primary endpoint of 90-day mortality between the permissive and standard feeding groups (relative risk, 0.94; 95% confidence interval [CI], 0.76–1.16; P = 0.58).
We used the validated modified NUTRIC score to categorize patients as high nutritional risk if their score was 5–9, and low nutritional risk if their score was 0–4. We also categorized patients based on BMI (weight in kilograms divided by the square of the height in meters [kg/m2]) measured on the day of ICU admission using the World Health Organization criteria as underweight (<18.5 kg/m2), normal weight (18.5–24.99 kg/m2), overweight (25.0–29.99 kg/m2), obese (30.0–40 kg/m2), or very obese (>40 kg/m2) (13). Patients were categorized based on the lower limit of the laboratory reference value for baseline phosphate less than or equal to 0.70 mmol/L and greater than 0.70 mmol/L, as high and low nutritional risk, respectively. Serum baseline prealbumin less than or equal to 0.10 g/L was considered as an indicator of severe nutritional risk, greater than 0.10 and less than or equal to 0.15 g/L mild to moderate risk, and greater than 0.15 g/L no risk (14, 15). Baseline transferrin values of less than or equal to 1.0 g/L were considered as an indicator of severe nutritional risk, and greater than 1 g/L as no to moderate nutritional risk (15). We categorized patients based on UUN using the median of the cohort as a cutoff value. We calculated nitrogen balance on Day 1 as follows: total protein intake in grams/6.25 – (UUN in mmol/35.7) + 4 g. A negative balance (≤0) is a marker of catabolism, and a positive balance (>0) indicates anabolism (11). Baseline phosphate, prealbumin, and transferrin were obtained at the time of enrollment. Baseline UUN was collected in the first 24 hours after enrollment.
For the high– and low–nutritional-risk groups categorized by NUTRIC score, we compared the permissive underfeeding and the standard feeding groups for the primary endpoint of 90-day mortality. Secondary outcomes included ICU, 28-day, hospital, and 180-day mortality; mechanical ventilation duration; ICU and hospital lengths of stay; hypoglycemia; hypokalemia; hypomagnesemia; hypophosphatemia; packed red blood cells transfusions; feeding intolerance and diarrhea in the first 14 ICU days; and ICU-associated infections during the ICU stay (using published definitions [16]). Between the permissive underfeeding and standard feeding groups, we compared Sequential Organ Failure Assessment score, serum phosphate and potassium, prealbumin, transferrin, UUN, nitrogen balance, bilirubin, PaCO2, hemoglobin, and weight, and examined changes in these variables over time. We conducted additional analyses to examine high and low nutritional-risk categories using other definitions with 90-day mortality as a primary outcome and ICU-associated infections and incident renal-replacement therapy (RRT) as secondary outcomes. Because our analysis showed that prealbumin was a significant effect modifier of the association of the feeding intervention and mortality, we performed supplementary analyses of baseline characteristics, intervention, and outcomes stratified by prealbumin level.
Categorical variables were compared using the chi-square test. Continuous variables were tested for normality; accordingly they were reported as means and SD or medians and quartiles 1 and 3 (Q1–3) and were tested using the Student’s t test or Wilcoxon-Mann-Whitney test. Confounding effect is expected to have been controlled by the randomized design of the original study, which is expected to be true for the current analyses (stratified by variables based on baseline characteristics). Nevertheless, to further control for any residual confounding effect, we examined the association of the intervention and different outcomes after adjustment for a priori selected variables known to be clinically associated with the outcome (Acute Physiology and Chronic Health Evaluation II, baseline creatinine, and sepsis) using logistic and linear regression analyses, as appropriate. Variables were maintained in the model if P value was less than 0.25. Results of the associations were reported as adjusted odds ratio (aOR) or correlation coefficient and 95% CI. We added an interaction term to test for effect modification of the nutritional-risk group on the association of the intervention and outcomes. Kaplan-Meier survival curves were compared using the log-rank test.
For serial measurements, we tested between-group difference, change over time, and between-group difference over time using repeated measures analysis of variance with no imputation for missing values. We performed retrospective power analyses to provide an indication of the smallest difference between the intervention and control groups in the two nutritional-risk categories. In the high–nutritional-risk group, the available sample size (189 in each group) would yield 80% power to detect a 14.4% difference in 90-day mortality. In the low–nutritional-risk group, the available sample size (259 and 257 patients in the two study groups) would yield an 80% power to detect an 8.8% difference in 90-day mortality. Tests were two-sided and statistical significance was determined at P less than 0.025 as per the Bonferroni correction for multiple testing. Analyses were conducted using SAS version 9.2 (SAS Institute, Cary, NC).
Of the 894 patients, 378 (42.3%) were categorized as high nutritional risk (NUTRIC score 5–9) and 516 (57.7%) patients as low nutritional risk (NUTRIC score 0–4). In each nutritional-risk category, demographic, physiologic, and nutritional characteristics of patients randomized to the permissive underfeeding and standard feeding groups were similar (Table 1).
| Variable | High–Nutritional-Risk Group (NUTRIC 5–9) (n = 378) | Low–Nutritional-Risk Group (NUTRIC 0–4) (n = 516) | ||||
|---|---|---|---|---|---|---|
| Permissive Underfeeding (n = 189) | Standard Feeding (n = 189) | P Value | Permissive Underfeeding (n = 259) | Standard Feeding (n = 257) | P Value | |
| Age, yr, mean ± SD | 60 ± 17 | 63 ± 14 | 0.06 | 43 ± 18 | 42 ± 18 | 0.49 |
| Female sex, n (%) | 87 (46.0) | 81 (42.9) | 0.53 | 69 (26.6) | 83 (32.3) | 0.16 |
| Admission category, n (%) | ||||||
| Medical | 175 (92.6) | 170 (90.0) | 0.47 | 161 (62.2) | 165 (64.2) | 0.12 |
| Surgical | 2 (1.1) | 5 (2.7) | 17 (6.6) | 7 (2.7) | ||
| Nonoperative trauma | 12 (6.4) | 14 (7.4) | 81 (31.3) | 85 (33.1) | ||
| Diabetes, n (%) | 104 (55.0) | 103 (54.5) | 0.92 | 55 (21.2) | 50 (19.5) | 0.62 |
| Sepsis, n (%) | 106 (56.1) | 90 (47.6) | 0.10 | 53 (20.5) | 43 (16.7) | 0.28 |
| Traumatic brain injury, n (%) | 7 (3.7) | 8 (4.2) | 0.79 | 48 (18.5) | 55 (21.4) | 0.42 |
| APACHE II score, median (Q1–Q3)* | 26 (22–31) | 27 (22–31) | 0.46 | 17 (13–20) | 17 (12–20) | 0.71 |
| SOFA score on Day 1, mean ± SD | 11.8 ± 3.0 | 11.6 ± 3.1 | 0.49 | 8.5 ± 3.1 | 8.5 ± 3.2 | 0.90 |
| Mechanical ventilation, n (%) | 186 (98.4) | 181 (95.8) | 0.13 | 250 (96.5) | 248 (96.5) | 0.99 |
| Vasopressor, n (%) | 116 (61.4) | 119 (63.0) | 0.75 | 139 (53.7) | 124 (48.3) | 0.22 |
| Renal-replacement therapy, n (%) | 35 (18.5) | 41 (21.7) | 0.44 | 6 (2.3) | 8 (3.1) | 0.58 |
| Inclusion blood glucose, mg/dl, median (Q1–Q3)* | 199 (146–262) | 187 (141–246) | 0.35 | 155 (124–200) | 151 (121–211) | 0.94 |
| Inclusion blood glucose, mmol/L, median (Q1–Q3)* | 11.1 (8.1–14.6) | 10.4 (7.9–13.7) | 0.35 | 8.6 (6.9–11.1) | 8.4 (6.7–11.7) | 0.94 |
| Creatinine, mg/dl, median (Q1–Q3)* | 1.44 (0.85–2.67) | 1.53 (0.94–3.04) | 0.20 | 0.84 (0.69–1.06) | 0.78 (0.67–1.09) | 0.20 |
| Creatinine, μmol/L, median (Q1–Q3)* | 127 (75–236) | 135 (83–269) | 0.20 | 74 (61–94) | 69 (59–96) | 0.20 |
| Bilirubin, mg/dl, median (Q1–Q3)* | 0.87 (0.51–1.80) | 0.91 (0.52–1.71) | 1.0 | 0.70 (0.47–1.25) | 0.76 (0.48–1.24) | 0.61 |
| Bilirubin, μmol/L, median (Q1–Q3)* | 14.9 (8.8–30.9) | 15.6 (8.9–29.3) | 1.0 | 11.9 (8.0–21.4) | 13.0 (8.2–21.2) | 0.61 |
| Platelets, 109/L, median (Q1–Q3)* | 146 (83–240) | 183 (98–278) | 0.01 | 200 (141–262) | 199 (147–252) | 0.80 |
| INR, median (Q1–Q3)* | 1.3 (1.1–1.5) | 1.3 (1.1–1.6) | 0.31 | 1.2 (1.1–1.3) | 1.2 (1.1–1.3) | 0.65 |
| C-reactive protein, mg/L, mean ± SD† | 127 ± 86 | 133 ± 91 | 0.56 | 133 ± 76 | 119 ± 74 | 0.05 |
| Hemoglobin, g/L, mean ± SD | 99 ± 22 | 101 ± 21 | 0.49 | 109 ± 23 | 110 ± 23 | 0.53 |
| BMI, kg/m2, median (Q1–Q3)* | 27.7 (23.5–35.0) | 28.7 (24.6–34.9) | 0.35 | 27.1 (23.1–32.0) | 27.3 (23.5–32.0) | 0.58 |
| NUTRIC score, mean ± SD | 5.88 ± 0.88 | 5.94 ± 0.97 | 0.54 | 2.88 ± 1.09 | 2.86 ± 1.06 | 0.86 |
| Phosphate, mmol/L, mean ± SD† | 1.14 ± 0.57 | 1.12 ± 0.51 | 0.79 | 0.86 ± 0.33 | 0.90 ± 0.35 | 0.14 |
| Albumin, g/L, mean ± SD | 27.6 ± 8.6 | 26.7 ± 5.9 | 0.24 | 28.4 ± 5.8 | 29.3 ± 5.6 | 0.08 |
| Prealbumin, g/L, median (Q1–Q3)*† | 0.11 (0.08–0.16) | 0.10 (0.07–0.14) | 0.04 | 0.13 (0.10–0.16) | 0.13 (0.11–0.17) | 0.29 |
| Transferrin, g/L, mean ± SD† | 1.29 ± 0.53 | 1.22 ± 0.49 | 0.20 | 1.40 ± 0.44 | 1.49 ± 0.48 | 0.05 |
| 24-h urinary urea nitrogen, mmol, median (Q1–Q3)*† | 229 (104–377) | 199 (108–314) | 0.15 | 268 (197–383) | 286 (205–422) | 0.13 |
| Nitrogen balance, g, mean ± SD† | −4.40 ± 6.42 | −3.76 ± 6.02 | 0.44 | −5.85 ± 5.37 | −6.48 ± 7.03 | 0.33 |
| Time from eligibility to randomization, h, median (Q1–Q3)* | 4.0 (0.0–17.0) | 2.3 (0.0–13.1) | 0.16 | 2.0 (0.0–12.0) | 1.0 (0.0–11.7) | 0.68 |
In both NUTRIC nutritional-risk categories, patients in the permissive underfeeding group received fewer calories than the standard feeding group; however, both groups received similar protein throughout the study period (Table 2, Figure 1). Other cointerventions and nutrition-related data are shown in Table 2 and Figure 1.
| Variable | High–Nutritional-Risk Group (n = 378) | Low–Nutritional-Risk Group (n = 516) | ||||
|---|---|---|---|---|---|---|
| Permissive Underfeeding (n = 189) | Standard Feeding (n = 189) | P Value | Permissive Underfeeding (n = 259) | Standard Feeding (n = 257) | P Value | |
| Interventions, mean ± SD | ||||||
| Calculated caloric requirement, kcal/d | 1,747 ± 355 | 1,744 ± 336 | 0.93 | 1,877 ± 384 | 1,916 ± 377 | 0.26 |
| Study caloric target, kcal/d | 987 ± 243 | 1,736 ± 340 | <0.0001 | 1,073 ± 270 | 1,893 ± 386 | <0.0001 |
| Achieved daily caloric intake, kcal | 799 ± 277 | 1,216 ± 420 | <0.0001 | 862 ± 308 | 1,360 ± 491 | <0.0001 |
| % of requirement achieved | 45.8 ± 12.3 | 70.7 ± 22.7 | <0.0001 | 46.4 ± 14.8 | 71.1 ± 21.3 | <0.0001 |
| Caloric source, kcal | ||||||
| Enteral | 719 ± 271 | 1,137 ± 427 | <0.0001 | 756 ± 309 | 1,243 ± 495 | <0.0001 |
| Propofol*† | 45.2 ± 72.7 | 45.1 ± 73.7 | 0.97 | 76.6 ± 96.0 | 78.9 ± 95.6 | 0.82 |
| Dextrose*† | 34.1 ± 50.4 | 35.2 ± 58.2 | 0.83 | 31.0 ± 64.6 | 34.6 ± 61.4 | 0.08 |
| PN*† | 4.8 ± 46.4 | 3.6 ± 41.0 | 0.71 | 1.0 ± 14.7 | 6.5 ± 69.4 | 0.39 |
| Calculated protein requirement, g/d | 82.7 ± 22.0 | 82.2 ± 21.2 | 0.83 | 87.1 ± 20.6 | 91.5 ± 23.1 | 0.02 |
| Achieved protein intake, g/d | 55.6 ± 22.1 | 54.4 ± 22.5 | 0.62 | 58.5 ± 25.1 | 62.8 ± 26.1 | 0.06 |
| % of requirement achieved | 68.4 ± 23.3 | 68.0 ± 26.8 | 0.89 | 68.1 ± 25.2 | 69.1 ± 24.3 | 0.65 |
| Duration of intervention, d | 9.2 ± 4.4 | 9.6 ± 4.3 | 0.36 | 9.0 ± 4.6 | 9.2 ± 4.4 | 0.51 |
| Cointerventions | ||||||
| Received insulin, n (%) | 122 (64.6) | 137 (72.5) | 0.10 | 83 (32.1) | 98 (38.1) | 0.15 |
| Daily insulin dose, units, median (Q1–Q3)* | 8.2 (0.0–36.5) | 17.6 (0.0–39.9) | 0.07 | 0.0 (0.0–6.8) | 0.0 (0.0–10.8) | 0.12 |
| Formulae, n (%) | ||||||
| Disease nonspecific | 72 (38.1) | 63 (33.3) | 0.33 | 198 (76.5) | 180 (70.0) | 0.10 |
| Disease specific | 117 (61.9) | 126 (66.7) | 61 (23.6) | 77 (30.0) | ||
Figure 1. (A–F) Serial measurements of the intervention, cointerventions, and selected outcomes in patients randomized to permissive underfeeding and standard feeding groups with stratification to high (NUTRIC score, 5–9) and low nutritional risk (NUTRIC score, 0–4). Means and 95% confidence intervals are displayed. *Statistical significance for the difference between the two groups on each day. P values for between-group difference, change over time, and between-group difference over time using repeated measures analysis of variance are given for each variable in the high– and the low–nutritional-risk categories. Total scores on the Sequential Organ Failure Assessment range from 0 to 24, with higher scores indicating a greater degree of organ failure. Nitrogen balance was calculated as [total protein intake in grams ÷ 6.25] – [(urinary nitrogen excretion in millimoles ÷ 35.7) + 4 g]. NUTRIC = NUTrition Risk in the Critically Ill; SOFA = Sequential Organ Failure Assessment.
[More] [Minimize]The association between permissive underfeeding versus standard feeding and mortality was nonsignificant in the two categories; aOR of 0.84 (95% CI, 0.56–1.27) for high nutritional risk and aOR of 1.01 (95% CI, 0.64–1.61) for low nutritional risk (interaction P = 0.53) (Table 3). Similarly, there were no differences in ICU, hospital, 28-day, or 180-day mortality between the two groups. Kaplan-Meier survival estimates demonstrated no difference in the probability of survival between the two groups in either nutritional-risk category (Figure 2).
| Outcomes | High–Nutritional-Risk Group (n = 378) | Low–Nutritional-Risk Group (n = 516) | P Value for Interaction | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Permissive Underfeeding (n = 189) | Standard Feeding (n = 189) | Adjusted Odds Ratio or Correlation Coefficient (95% CI) | P Value | Permissive Underfeeding (n = 259) | Standard Feeding(n = 257) | Adjusted Odds Ratio or Correlation Coefficient (95% CI) | P Value | ||
| 28-d mortality, n (%) | 56/189 (29.6) | 59/189 (31.2) | 0.93 (0.60 to 1.44) | 0.74 | 37/258 (14.3) | 38/255 (14.9) | 0.93 (0.56 to 1.53) | 0.76 | 0.93 |
| 90-d mortality, n (%) | 75/189 (39.7) | 83/189 (43.9) | 0.84 (0.56 to 1.27) | 0.40 | 46/256 (18.0) | 44/251 (17.5) | 1.01 (0.64 to 1.61) | 0.96 | 0.53 |
| 180-d mortality, n (%) | 81/188 (43.1) | 90/189 (47.6) | 0.86 (0.57 to 1.29) | 0.45 | 50/250 (20.0) | 50/247 (20.2) | 0.96 (0.61 to 1.50) | 0.85 | 0.60 |
| ICU mortality, n (%) | 43/189 (22.8) | 51/189 (27.0) | 0.80 (0.50 to 1.27) | 0.34 | 29/259 (11.2) | 34/257 (13.2) | 0.82 (0.48 to 1.39) | 0.46 | 0.91 |
| Hospital mortality, n (%) | 67/188 (35.6) | 80/189 (42.3) | 0.73 (0.48 to 1.11) | 0.14 | 41/259 (15.8) | 43/256 (16.8) | 0.90 (0.56 to 1.45) | 0.67 | 0.50 |
| ICU LOS, d, median (Q1 to Q3) | 13.0 (8.0 to 21.0) | 14 (9.0 to 22.0) | −1.4 (−4.0 to 1.3) | 0.30 | 13.0 (7.0 to 20.0) | 13.0 (8.0 to 19.0) | −0.04 (−1.9 to 1.8) | 0.96 | 0.39 |
| Hospital LOS, d, median (Q1 to Q3) | 29.0 (16.0 to 52.5) | 35.0 (17.0 to 62.0) | −8.5 (−21.5 to 4.4) | 0.20 | 27.0 (14.0 to 55.0) | 27.0 (13.0 to 65.0) | −5.2 (−18.2 to 7.9) | 0.44 | 0.72 |
| Ventilation duration, d, median (Q1 to Q3) | 9.0 (6.0 to 16.0) | 10.0 (5.0 to 17.0) | −1.9 (−4.7 to 0.9) | 0.18 | 9.0 (5.0 to 14.0) | 9.0 (5.0 to 15.0) | −2.6 (−5.9 to 0.7) | 0.13 | 0.72 |
| PRBC transfusions, n (%) | 81/189 (42.9) | 82/189 (43.4) | 1.02 (0.67 to 1.53) | 0.94 | 60/259 (23.2) | 60/257 (23.4) | 0.97 (0.64 to 1.47) | 0.89 | 0.90 |
| Cumulative PRBC transfusion over 14 d, units, mean ± SD | 0.17 ± 0.37 | 0.15 ± 0.25 | 0.01 (−0.05 to 0.08) | 0.72 | 0.10 ± 0.20 | 0.10 ± 0.30 | −0.02 (−0.06 to 0.02) | 0.41 | 0.46 |
| Hypoglycemia, n (%) | 6/189 (3.2) | 5/189 (2.7) | 1.22 (0.37 to 4.09) | 0.75 | 0/259 (0.0) | 2/257 (0.80) | NA | NA | NA |
| Hypokalemia, n (%) | 44/189 (23.3) | 50/189 (26.5) | 0.80 (0.49 to 1.28) | 0.34 | 57/259 (22.0) | 41/257(16.0) | 1.48 (0.95 to 2.32) | 0.08 | 0.06 |
| Hypomagnesemia, n (%) | 44/189 (23.3) | 59/189 (31.2) | 0.67 (0.42 to 1.06) | 0.08 | 83/259 (32.1) | 72/257 (28.0) | 1.21 (0.83 to 1.77) | 0.32 | 0.05 |
| Hypophosphatemia, n (%) | 108/189 (57.1) | 99/189 (52.4) | 1.16 (0.76 to 1.75) | 0.49 | 159/259 (61.4) | 162/257 (63.0) | 0.96 (0.67 to 1.37) | 0.81 | 0.46 |
| Incident renal-replacement therapy, n (%) | 22/153 (14.4) | 35/148 (23.7) | 0.45 (0.23 to 0.89) | 0.02 | 7/253 (2.8) | 10/248 (4.0) | 0.55 (0.18 to 1.66) | 0.29 | 0.61 |
| Healthcare-associated infections, n (%) | 66/189 (34.9) | 73/189 (38.6) | 0.85 (0.56 to 1.30) | 0.46 | 95/259 (36.7) | 96/257 (37.4) | 0.99 (0.69 to 1.41) | 0.94 | 0.60 |
| Urinary tract infection, n (%) | 28/189 (14.8) | 25/189 (13.2) | 1.12 (0.63 to 2.02) | 0.70 | 17/259 (6.6) | 23/257 (9.0) | 0.72 (0.37 to 1.37) | 0.31 | 0.36 |
| Catheter-related bloodstream infection, n (%) | 2/189 (1.1) | 8/189 (4.2) | 0.24 (0.05 to 1.16) | 0.08 | 9/259 (3.5) | 11/257 (4.3) | 0.81 (0.33 to 1.98) | 0.64 | 0.19 |
| Ventilator-associated pneumonia, n (%) | 25/189 (13.2) | 33/189 (17.5) | 0.72 (0.41 to 1.27) | 0.25 | 56/259 (21.6) | 57/257 (22.2) | 0.99 (0.65 to 1.51) | 0.95 | 0.41 |
| ICU-associated severe sepsis or septic shock, n (%) | 1/189 (0.5) | 1/189 (0.5) | 1.7 (0.09 to 33.17) | 0.73 | 2/259 (0.8) | 1/257 (0.4) | 1.99 (0.18 to 22.11) | 0.57 | 0.73 |
| Feeding intolerance, n (%) | 36/189 (19.1) | 37/189 (19.6) | 0.97 (0.58 to 1.61) | 0.90 | 31/259 (12.0) | 42/257 (16.3) | 0.71 (0.43 to 1.17) | 0.18 | 0.34 |
| Diarrhea, n (%) | 55/189 (29.1) | 66/189 (34.9) | 0.77 (0.50 to 1.18) | 0.23 | 42/259 (16.2) | 51/257 (19.8) | 0.75 (0.48 to 1.18) | 0.22 | 0.86 |
Figure 2. Kaplan-Meier survival curves for patients in the permissive underfeeding and standard feeding groups stratified by NUTRIC score (high nutritional risk, NUTRIC score 5–9; low nutritional risk, NUTRIC score 0–4) and by prealbumin. NUTRIC = NUTrition Risk in the Critically Ill.
[More] [Minimize]In patients at high nutritional risk, incident RRT was less frequent in the permissive underfeeding group (22 of 153; 14.4%) compared with the standard feeding group (35 of 148; 23.7%; aOR, 0.45; 95% CI, 0.23–0.89; P = 0.02). However, in patients at low nutritional risk, incident RRT was not statistically different in the permissive underfeeding group (7 of 253; 2.8%) compared with the standard feeding group (10 of 248; 4.0%; aOR, 0.55; 95% CI, 0.18–1.66; P = 0.29). Test of interaction showed that the association of the intervention and incident RRT was not modified by the nutritional-risk category (P value for interaction = 0.61). Serial Sequential Organ Failure Assessment scores, and prealbumin, nitrogen balance, phosphate, creatinine, bilirubin, PaCO2, hemoglobin, body weight, potassium, transferrin, and UUN did not differ between the two groups in either nutritional-risk category (Figure 1; see Figure E1 in the online supplement). There were no differences in any other variables, including mechanical ventilation duration and ICU lengths of stay, between the two groups in either nutritional-risk category (Table 3).
The associations of the intervention with 90-day mortality, ICU-associated infections, and incident RRT were not modified by the nutritional-risk category using other nutritional status indicators, including BMI, phosphate, transferrin, 24-hour UUN, and nitrogen balance (Table 4). The only exception was prealbumin; the association of permissive underfeeding versus standard feeding with 90-day mortality differed by baseline prealbumin level (prealbumin ≤0.10 g/L aOR 0.57, 95% CI, 0.31–1.05, P = 0.07; prealbumin >0.10 and ≤0.15 g/L aOR 0.79, 95% CI, 0.42–1.48, P = 0.46; prealbumin >0.15 g/L aOR 1.55, 95% CI, 0.79–3.01, P = 0.20; P value for interaction = 0.009) (Table 4, Figure 2). The association between the intervention and incident RRT and ICU-associated infections did not differ across prealbumin groups. Detailed stratified analysis based on prealbumin categories of baseline characteristics, interventions, and outcomes is provided in the Tables E1–E3.
| Variables | Permissive Underfeeding [n/N (%)] | Standard Feeding [n/N (%)] | Adjusted Odds Ratio (95% CI) | P Value | P Value for Interaction |
|---|---|---|---|---|---|
| 90-d mortality* | |||||
| BMI <18.5 kg/m2 | 7/22 (31.8) | 4/19 (21.1) | 3.95 (0.51–30.47) | 0.19 | 0.36 |
| BMI 18.5–24.9 kg/m2 | 36/128 (28.1) | 39/116 (33.6) | 0.76 (0.43–1.34) | 0.34 | |
| BMI 25.0–29.9 kg/m2 | 36/127 (28.4) | 31/133 (23.3) | 1.27 (0.72–2.26) | 0.41 | |
| BMI 30–39.9 kg/m2 | 33/128 (25.8) | 38/125 (30.4) | 0.88 (0.49–1.56) | 0.66 | |
| BMI ≥40 kg/m2 | 9/40 (22.5) | 14/46 (30.4) | 0.38 (0.12–1.17) | 0.09 | |
| Phosphate ≤0.70 mmol/L | 29/125 (23.2) | 23/110 (20.9) | 1.36 (0.71–2.62) | 0.36 | 0.13 |
| Phosphate >0.70 mmol/L | 84/298 (28.2) | 97/313 (31.0) | 0.77 (0.54–1.11) | 0.16 | |
| Prealbumin ≤0.10 g/L | 28/98 (28.6) | 45/102 (44.1) | 0.57 (0.31–1.05) | 0.07 | 0.009 |
| Prealbumin >0.10 and ≤0.15 g/L | 23/124 (18.6) | 30/128 (23.4) | 0.79 (0.42–1.48) | 0.46 | |
| Prealbumin >0.15 g/L | 30/111 (27.0) | 19/108 (17.6) | 1.55 (0.80–3.01) | 0.20 | |
| Transferrin ≤1.0 g/L | 38/90 (42.2) | 36/89 (40.5) | 0.99 (0.52–1.87) | 0.97 | 0.54 |
| Transferrin >1.0 g/L | 52/268 (19.4) | 61/269 (22.7) | 0.83 (0.54–1.26) | 0.35 | |
| 24-h urinary urea nitrogen ≤262.5 mmol | 49/160 (30.6) | 44/146 (30.1) | 0.92 (0.55–1.53) | 0.74 | 0.99 |
| 24-h urinary urea nitrogen >262.5 mmol | 27/143 (18.9) | 27/143 (18.9) | 0.89 (0.48–1.64) | 0.71 | |
| Negative nitrogen balance | 58/258 (22.5) | 53/235 (22.6) | 0.75 (0.46–1.23) | 0.25 | 0.78 |
| Positive nitrogen balance | 18/44 (40.9) | 18/54 (33.3) | 0.88 (0.33–2.34) | 0.79 | |
| New renal-replacement therapy | |||||
| BMI <18.5 kg/m2 | 3/22 (13.6) | 1/17 (5.9) | NA | NA | 0.74 |
| BMI 18.5–24.9 kg/m2 | 4/119 (3.4) | 9/108 (8.3) | 0.38 (0.10–1.43) | 0.15 | |
| BMI 25.0–29.9 kg/m2 | 2/116 (1.7) | 13/125 (10.4) | 0.05 (0.01–0.43) | 0.006 | |
| BMI 30–39.9 kg/m2 | 12/117 (10.3) | 14/106 (13.2) | 0.82 (0.29–2.29) | 0.70 | |
| BMI ≥40 kg/m2 | 8/32 (25.0) | 7/39 (18.0) | 0.48 (0.10–2.27) | 0.35 | |
| Phosphate ≤0.70 mmol/L | 2/117 (1.7) | 5/105 (4.8) | 0.40 (0.06–2.72) | 0.35 | 0.95 |
| Phosphate >0.70 mmol/L | 27/270 (10.0) | 37/278 (13.3) | 0.51 (0.27–0.95) | 0.03 | |
| Prealbumin ≤0.10 g/L | 9/86 (10.5) | 18/87 (20.7) | 0.48 (1.18–1.28) | 0.14 | 0.08 |
| Prealbumin >0.10 and ≤0.15 g/L | 5/116 (4.3) | 13/117 (11.1) | 0.24 (0.06–0.91) | 0.04 | |
| Prealbumin >0.15 g/L | 9/101 (8.9) | 5/103 (4.8) | 2.24 (0.47–10.62) | 0.31 | |
| Transferrin ≤1.0 g/L | 15/80 (18.8) | 12/72 (16.7) | 1.34 (0.45–4.01) | 0.60 | 0.03 |
| Transferrin >1.0 g/L | 11/249 (4.4) | 25/253 (9.9) | 0.32 (0.14–0.76) | 0.01 | |
| 24-h urinary urea nitrogen ≤262.5 mmol | 13/145 (9.0) | 18/132 (13.6) | 0.56 (0.23–1.38) | 0.21 | 0.77 |
| 24-h urinary urea nitrogen >262.5 mmol | 4/141 (2.8) | 7/142 (4.9) | 0.59 (0.16–2.17) | 0.43 | |
| Negative nitrogen balance | 12/247 (4.9) | 17/229 (7.4) | 0.34 (0.13–0.92) | 0.03 | 0.74 |
| Positive nitrogen balance | 5/38 (13.2) | 8/45 (17.8) | 0.42 (0.05–3.29) | 0.41 | |
| ICU-associated infections | |||||
| BMI <18.5 kg/m2 | 8/22 (36.4) | 7/19 (36.8) | 0.98 (0.27–3.50) | 0.97 | 0.35 |
| BMI 18.5–24.9 kg/m2 | 51/129 (39.5) | 39/120 (32.5) | 1.36 (0.81–2.28) | 0.25 | |
| BMI 25.0–29.9 kg/m2 | 45/127 (35.4) | 57/133 (42.9) | 0.73 (0.44–1.22) | 0.23 | |
| BMI 30–39.9 kg/m2 | 40/130 (30.8) | 47/126 (37.3) | 0.75 (0.45–1.26) | 0.27 | |
| BMI ≥40 kg/m2 | 17/40 (42.5) | 19/47 (40.4) | 0.92 (0.38–2.26) | 0.87 | |
| Phosphate ≤0.70 mmol/L | 57/126 (45.2) | 41/111 (36.9) | 1.41 (0.84–2.38) | 0.20 | 0.10 |
| Phosphate >0.70 mmol/L | 102/300 (34.0) | 120/317 (37.9) | 0.84 (0.61–1.17) | 0.31 | |
| Prealbumin ≤0.10 g/L | 33/98 (33.7) | 44/102 (43.1) | 0.67 (0.38–1.19) | 0.17 | 0.42 |
| Prealbumin >0.10 and ≤0.15 g/L | 62/125 (49.6) | 52/129 (40.3) | 1.46 (0.89–2.40) | 0.14 | |
| Prealbumin >0.15 g/L | 42/111 (37.8) | 43/110 (39.1) | 1.00 (0.57–1.74) | 0.99 | |
| Transferrin ≤1.0 g/L | 34/90 (37.8) | 37/89 (41.6) | 0.85 (0.47–1.55) | 0.60 | 0.48 |
| Transferrin >1.0 g/L | 107/269 (39.8) | 107/272 (39.3) | 1.02 (0.72–1.44) | 0.91 | |
| 24-h urinary urea nitrogen ≤262.5 mmol | 66/160 (41.3) | 59/148 (39.9) | 1.11 (0.70–1.75) | 0.67 | 0.58 |
| 24-h urinary urea nitrogen >262.5 mmol | 61/145 (42.1) | 53/144 (36.8) | 1.25 (0.78–2.00) | 0.36 | |
| Negative nitrogen balance | 112/260 (43.1) | 88/238 (37.0) | 1.14 (0.77–1.68) | 0.52 | 0.23 |
| Positive nitrogen balance | 14/44 (31.8) | 24/54 (44.4) | 0.62 (0.26–1.50) | 0.29 | |
In patients at high and low nutritional risk defined by the NUTRIC score, enteral feeding to deliver moderate nonprotein calories was associated with similar mortality compared with planned delivery of full nonprotein caloric requirements. Our study shows that the most current definitions for nutritional risk, including the NUTRIC score, could not differentiate the risk association between moderate versus full caloric intake and outcomes. However, permissive underfeeding compared with standard feeding had different effects based on baseline prealbumin; patients with low prealbumin levels may have lower 90-day mortality if they received permissive underfeeding compared with standard feeding.
Our study indicates the limitations of available measures proposed for assessing nutritional risk. We found no difference in mortality or any other studied endpoint between high– and low–nutritional-risk groups defined by NUTRIC or most other measures. The NUTRIC score, which is based on nonnutritional data, has been proposed for this purpose based on the finding of improved survival in patients with high NUTRIC scores in a post hoc analysis of a 1,199-patient RCT (5). However, our study does not validate this finding, because we found that the NUTRIC score cannot be used to differentiate between who may or may not benefit from different caloric dosing. Given that nutritional adequacy was defined retrospectively in the observational cohort (5), the influence of residual confounding effects cannot be excluded, which is probably responsible for the discordant results of many observational nutritional studies in critical care. Among all examined indicators in our study, baseline prealbumin differentiated the risk association of permissive underfeeding and mortality; patients with low prealbumin had lower mortality with permissive underfeeding compared with standard feeding. The use of admission prealbumin level to predict response to nutrition has also been demonstrated in a study of hospitalized patients with anorexia nervosa; patients with low prealbumin had a threefold risk of refeeding hypophosphatemia and a twofold risk of hypoglycemia compared with patients with normal prealbumin, independent of BMI (17). However, we did not find differences in the incidence of hypophosphatemia or hypoglycemia in our study based on prealbumin levels.
The risk of inducing refeeding syndrome, defined as hypophosphatemia after reinstituting nutritional support, in malnourished ICU patients has been a concern (18–20). We found no difference in serial phosphate, potassium, and magnesium levels between permissive underfeeding and standard feeding groups in both the high and low nutritional-risk categories defined by NUTRIC score. A recent RCT of restricted versus continued standard caloric intake in patients with refeeding syndrome found that the restricted caloric group had slightly higher phosphate level on Days 1 and 2 (difference of 0.1 mmol/L), although phosphate replacement dose and the number of patients receiving phosphate replacement did not differ between study groups (7). Interestingly, the trial showed that this high-risk group with refeeding syndrome had higher hospital survival with restricted caloric feeding compared with standard feeding (91% vs. 82%; absolute risk difference, 9.2%; 95% CI, 0.7–17.7; P = 0.017), although mortality was a secondary endpoint for the trial (7). These findings are in line with our observation regarding prealbumin, that certain patients with high nutritional risk may have better outcomes with a restricted nutritional strategy.
In both the high– and low–nutritional-risk groups as defined by the NUTRIC scores, we found no differences between permissive underfeeding and standard feeding in indices of protein status including prealbumin, transferrin levels, UUN, and nitrogen balance, keeping in mind the limitations of these parameters in assessing protein metabolism (21).
We also found no differences in ICU-acquired infections with permissive underfeeding versus standard feeding in high or low nutritional-risk categories. In a recent trial of restricted versus continued standard caloric intake in patients with refeeding syndrome, the restricted caloric strategy was associated with a lower incidence of infections than the standard feeding strategy (7). However, this trial did not include patients with no refeeding syndrome, therefore one cannot conclude whether refeeding syndrome defined by post-feeding hypophosphatemia is an effect modifier of the caloric dose on outcomes.
In patients with high nutritional risk (NUTRIC scores 5–9), those allocated to permissive underfeeding had lower use of incident RRT. Indeed, caloric restriction has been shown to be renal-protective in animal models of acute kidney injury (22–24), via several mechanisms including improved insulin sensitivity (24). Our findings suggest a need for further research on the optimal caloric dose in critically ill patients with or at risk for acute kidney injury.
It is important to note that the PermiT trial restricted calories and not protein. In contrast, the EDEN (Early Versus Delayed Enteral Feeding to Treat People With Acute Lung Injury or Acute Respiratory Distress Syndrome) study limited both calories and protein (1). The effects of both protein and caloric restriction and whether the dose of protein affects outcomes in patients with high nutritional risk remains to be studied.
The strengths of this study include the multicenter conduct and the pragmatic inclusion of medical-surgical-trauma critically ill adults, all of whom received early initiation of feeding. Additionally, the baseline characteristics of the groups of permissive underfeeding and standard feeding were balanced in both high and low nutritional risk categories defined by NUTRIC score. A limitation of this study is the post hoc analysis. Given the number of statistical tests, the significant associations observed with prealbumin could be related to multiple testing. Our study did not address whether permissive underfeeding versus standard feeding has a differential effect on long-term outcome based on underlying nutritional status.
Regarding nutrition and long-term outcomes, an observational study of 302 adults receiving prolonged mechanical ventilation (>8 d in the ICU) found that receiving adequate energy in the first 8 days of ICU stay was associated with improved health-related quality of life at 3 months, although this association became nonsignificant by 6 months (25). However, average BMI in the study population was 30.0 ± 8.7 kg/m2 and only 2% had BMI less than 18.5 kg/m2; therefore, the results may not be generalizable to patients with low BMI. In a substudy of the EDEN trial, 174 acute lung injury survivors (mean BMI, 32 kg/m2) underwent long-term follow-up with physical and cognitive evaluation (26). At 6 and 12 months, initial trophic versus full enteral feeding had no effect on either physical performance outcomes (upper arm anthropometrics, muscle strength, pulmonary function, 6-min-walk distance) or cognitive function (26). However, the EDEN study enrolled relatively young and overweight patients, and excluded underweight patients; therefore, the results of this study may not be generalizable to other ICU patients. Additionally, there was no stratification by baseline nutritional status. Therefore, whether moderate versus full caloric intake has different effects in patients with high and low nutritional risk on long-term functional status and cognitive function remains largely unknown and deserves further study.
In conclusion, in patients with high and low nutritional risk alike, enteral feeding to deliver moderate calories with full protein intake was associated with similar mortality compared with standard caloric feeding with full protein requirements. Available nutritional assessment measures do not seem to differentiate the risk association of moderate versus full caloric doses and mortality risk.
| 1. | Rice TW, Wheeler AP, Thompson BT, Steingrub J, Hite RD, Moss M, Morris A, Dong N, Rock P; National Heart, Lung, and Blood Institute Acute Respiratory Distress Syndrome (ARDS) Clinical Trials Network. Initial trophic vs full enteral feeding in patients with acute lung injury: the EDEN randomized trial. JAMA 2012;307:795–803. |
| 2. | Arabi YM, Aldawood AS, Haddad SH, Al-Dorzi HM, Tamim HM, Jones G, Mehta S, McIntyre L, Solaiman O, Sakkijha MH, et al.; PermiT Trial Group. Permissive underfeeding or standard enteral feeding in critically ill adults. N Engl J Med 2015;372:2398–2408. |
| 3. | Peake SL, Davies AR, Deane AM, Lange K, Moran JL, O’Connor SN, Ridley EJ, Williams PJ, Chapman MJ; TARGET investigators and the Australian and New Zealand Intensive Care Society Clinical Trials Group. Use of a concentrated enteral nutrition solution to increase calorie delivery to critically ill patients: a randomized, double-blind, clinical trial. Am J Clin Nutr 2014;100:616–625. |
| 4. | Heyland DK, Dhaliwal R, Jiang X, Day AG. Identifying critically ill patients who benefit the most from nutrition therapy: the development and initial validation of a novel risk assessment tool. Crit Care 2011;15:R268. |
| 5. | 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;35:158–162. |
| 6. | Mehanna HM, Moledina J, Travis J. Refeeding syndrome: what it is, and how to prevent and treat it. BMJ 2008;336:1495–1498. |
| 7. | Doig GS, Simpson F, Heighes PT, Bellomo R, Chesher D, Caterson ID, Reade MC, Harrigan PW; Refeeding Syndrome Trial Investigators Group. Restricted versus continued standard caloric intake during the management of refeeding syndrome in critically ill adults: a randomised, parallel-group, multicentre, single-blind controlled trial. Lancet Respir Med 2015;3:943–952. |
| 8. | Taylor BE, McClave SA, Martindale RG, Warren MM, Johnson DR, Braunschweig C, McCarthy MS, Davanos E, Rice TW, Cresci GA, et al.; Society of Critical Care Medicine; American Society of Parenteral and Enteral Nutrition. 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;44:390–438. |
| 9. | Devoto G, Gallo F, Marchello C, Racchi O, Garbarini R, Bonassi S, Albalustri G, Haupt E. Prealbumin serum concentrations as a useful tool in the assessment of malnutrition in hospitalized patients. Clin Chem 2006;52:2281–2285. |
| 10. | Fuhrman MP, Charney P, Mueller CM. Hepatic proteins and nutrition assessment. J Am Diet Assoc 2004;104:1258–1264. |
| 11. | Bechard LJ, Parrott JS, Mehta NM. Systematic review of the influence of energy and protein intake on protein balance in critically ill children. J Pediatr 2012;161:333–339.e1. |
| 12. | Arabi YM, Haddad SH, Aldawood AS, Al-Dorzi HM, Tamim HM, Sakkijha M, Jones G, McIntyre L, Mehta S, Solaiman O, et al. Permissive underfeeding versus target enteral feeding in adult critically ill patients (PermiT Trial): a study protocol of a multicenter randomized controlled trial. Trials 2012;13:191. |
| 13. | WHO classification of body mass index (BMI) [accessed 2012 Sept 4]. Available from: http://apps.who.int/bmi/index.jsp?introPage=intro_3.html |
| 14. | Shenkin A. Serum prealbumin: Is it a marker of nutritional status or of risk of malnutrition? Clin Chem 2006;52:2177–2179. |
| 15. | Murray MJ, Cousins DB, Pearl RG, Prough DS. Critical care medicine: perioperative management. Philadelphia: Lippincott-Williams and Wilkins; 2002. |
| 16. | CDC/NHSN surveillance definition of healthcare-associated infections [accessed 2015 Mar 28]. Available from: http://www.cdc.gov/nhsn/PDFs/pscManual/17pscNosInfDef_current.pdf |
| 17. | Gaudiani JL, Sabel AL, Mehler PS. Low prealbumin is a significant predictor of medical complications in severe anorexia nervosa. Int J Eat Disord 2014;47:148–156. |
| 18. | Marik PE, Bedigian MK. Refeeding hypophosphatemia in critically ill patients in an intensive care unit: a prospective study. Arch Surg 1996;131:1043–1047. |
| 19. | Marinella MA. Refeeding syndrome and hypophosphatemia. J Intensive Care Med 2005;20:155–159. |
| 20. | Skipper A. Refeeding syndrome or refeeding hypophosphatemia: a systematic review of cases. Nutr Clin Pract 2012;27:34–40. |
| 21. | Weijs PJ, Cynober L, DeLegge M, Kreymann G, Wernerman J, Wolfe RR. Proteins and amino acids are fundamental to optimal nutrition support in critically ill patients. Crit Care 2014;18:591. |
| 22. | Lempiäinen J, Finckenberg P, Mervaala EE, Sankari S, Levijoki J, Mervaala EM. Caloric restriction ameliorates kidney ischaemia/reperfusion injury through PGC-1α-eNOS pathway and enhanced autophagy. Acta Physiol (Oxf) 2013;208:410–421. |
| 23. | Ning YC, Cai GY, Zhuo L, Gao JJ, Dong D, Cui SY, Shi SZ, Feng Z, Zhang L, Sun XF, et al. Beneficial effects of short-term calorie restriction against cisplatin-induced acute renal injury in aged rats. Nephron, Exp Nephrol 2013;124:19–27. |
| 24. | Mitchell JR, Verweij M, Brand K, van de Ven M, Goemaere N, van den Engel S, Chu T, Forrer F, Müller C, de Jong M, et al. Short-term dietary restriction and fasting precondition against ischemia reperfusion injury in mice. Aging Cell 2010;9:40–53. |
| 25. | Wei X, Day AG, Ouellette-Kuntz H, Heyland DK. The association between nutritional adequacy and long-term outcomes in critically ill patients requiring prolonged mechanical ventilation: a multicenter cohort study. Crit Care Med 2015;43:1569–1579. |
| 26. | Needham DM, Dinglas VD, Morris PE, Jackson JC, Hough CL, Mendez-Tellez PA, Wozniak AW, Colantuoni E, Ely EW, Rice TW, et al.; NIH NHLBI ARDS Network. Physical and cognitive performance of patients with acute lung injury 1 year after initial trophic versus full enteral feeding: EDEN trial follow-up. Am J Respir Crit Care Med 2013;188:567–576. |
Supported by King Abdullah International Medical Research Center, Riyadh, Saudi Arabia. S.M.B. is supported by a Canada Research Chair in Critical Care Nephrology.
Author Contributions: Y.M.A., A.S.A., and H.M.A.-D. contributed to conception and design of the study. Y.M.A., A.S.A., H.M.A.-D., H.M.T., S.H.H., G.J., L.M., O.S., M.H.S., M.S., S.M., A.K., S.M.B., and S.M. contributed to acquisition, analysis or interpretation of data, drafting and revising the manuscript for important intellectual content, and approved final version to be published.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
Originally Published in Press as DOI: 10.1164/rccm.201605-1012OC on September 2, 2016
Author disclosures are available with the text of this article at www.atsjournals.org.
