Nutrition in Critically Ill Children with AKI on Continuous RRT: Consensus Recommendations

Abstract

Background

Nutrition plays a vital role in the outcome of critically ill children, particularly those with AKI. Currently, there are no established guidelines for children with AKI treated with continuous RRT (CRRT). A thorough understanding of the metabolic changes and nutritional challenges in AKI and CRRT is required. Our objective was to create clinical practice points for nutritional assessment and management in critically ill children with AKI receiving CRRT.

Methods

PubMed, MEDLINE, Cochrane, and Embase databases were searched for articles related to the topic. Expertise of the authors and a consensus of the workgroup were additional sources of data in the article. Available articles on nutrition therapy in pediatric patients receiving CRRT through January 2023.

Results

On the basis of the literature review, the current evidence base was examined by a panel of experts in pediatric nephrology and nutrition. The panel used the literature review as well as their expertise to formulate clinical practice points. The modified Delphi method was used to identify and refine clinical practice points.

Conclusions

Forty-four clinical practice points are provided on nutrition assessment, determining energy needs, and nutrient intake in children with AKI and on CRRT on the basis of the existing literature and expert opinions of a multidisciplinary panel.

Introduction

Nutrition plays a key role in determining the outcome of critically ill children, such as those with AKI.^1–3 Nutritional support is sometimes deferred until the child is stabilized, and as such, malnutrition and protein–energy wasting are commonly seen in these children.^4,5 In a study by Kyle et al., acute malnutrition was reported in 33% of those with an AKI in the pediatric intensive care unit (PICU), and 23.8% were likely to show chronic malnutrition.⁶ In critically ill pediatric patients, increased metabolic demands, the child's baseline nutritional status, and ongoing growth needs should be considered.^5,7,8 The metabolic changes in critical illness and their pathways are illustrated in Figure 1. In addition, AKI itself contributes to a hypercatabolic state by altering lipid, carbohydrate, and protein metabolism.⁵

Figure 1.

Metabolic changes in critical illness. A schematic diagram represents the metabolic changes in critical illness and their pathways. In critical illness, an increase in catabolic hormones, cytokines, and catecholamines and a decrease in anabolic hormones are observed. This leads to increased net muscle protein degradation and increased movement of free amino acids through the circulation. Amino acids contribute to gluconeogenesis and protein synthesis in the liver. An increase in carbohydrate turnover leads to increased glucose which is, in turn, used as fuel for the brain, RBCs, and kidneys. An increase in lipid turnover leads to an increase in FAs and ketones, which serve as fuel for the brain. Adapted from ref. 9. FA, fatty acid.

Moreover, AKI is associated with volume overload, which leads to difficulties in the provision of adequate nutrition.⁸ Safe and timely nutrition delivery can significantly affect outcomes.^1–3 Determining the nutritional needs of these patients is challenging. Continuous RRT (CRRT) is the preferred treatment modality in critically ill children with severe AKI.^10,11 However, CRRT can increase the risk of new or worsening malnutrition as it may lead to significant macronutrient and micronutrient losses and electrolyte disorders.¹² According to a study by Castillo et al., children with heart disease and chronic renal insufficiency are more likely to be malnourished and malnutrition is associated with increased mortality.¹ Thus, it is crucial to adapt nutrition therapy in children requiring CRRT individually. Mortality and length of stay outcomes associated with malnutrition in children with AKI and/or requiring CRRT are summarized in Table 1. There are limited data on the effect of nutrition on critically ill pediatric patients with AKI treated with CRRT. In clinical practice, nutrition delivery is based on understanding the observed metabolic changes, extrapolating information from adult literature, and existing guidelines for managing critically ill children.

Table 1.

Studies reporting on children with renal failure

Reference (Year)	Type of Study	Sample Size	Study Population	Mean Age, yr	Intervention	Results
Mortality and LOS
Castillo et al. (2012)1	Prospective observational study	174	Critically ill on CRRT	4.5	None	• Mortality in AKI patients with weight <P3 (third percentile) was greater than that of the children with weight >P3 (51% versus 33%; P = 0.037) In the multivariate logistic regression analyses, the only factor associated with mortality was PEW (malnutrition) (OR, 2.11; 95% CI, 1.07 to 4.17; P = 0.032)
Kyle et al. (2017)13	Retrospective chart review	511 • No AKI: 355 • AKI: 156	Critically ill with and without AKI	1.4	Protein feeding	• Patients with AKI versus non-AKI were significantly shorter and weighed less. Weight-for-age and weight-for-height z-scores were not significantly different • However, height-for-age z-scores were significantly lower in AKI versus non-AKI patients, which suggests that they had more height stunting or chronic malnutrition Among non-AKI versus AKI • PICU LOS (days): 5 (4–9) versus 7 (3.5–12); P < 0.002 • Hospital LOS (days): 11 (8–19) versus 15 (10–24); P < 0.003 Mortality: 4.2% versus 8.3%; P = 0.060
Protein and caloric intake
Zappitelli et al. (2009)14	Retrospective database study	15	Critically ill on CRRT	7.7	CRRT (CVVH)	• At day 2 versus baseline: protein intake increased by 3% and caloric intake increased by 59% • At day 5 versus baseline: protein intake decreased by 6.6%, while caloric intake increased by 66.6%
Kyle et al. (2017)13	Retrospective chart review	511 • No AKI: 355 • Resolved AKI: 112 • Persistent AKI: 44	Critically ill with and without AKI	1.3	Protein feeding	Over the first 8 d in PICU, there was an increase in protein intake among patients with • No AKI by 29.7% • Resolved AKI by 40.4% • Persistent AKI by 73.8%
Vega et al. (2018)15		55	PICU and CVICU patients receiving CRRT for >48 h	8.1	Quality improvement program to increase protein prescription and intake Goal protein set at 2.5 g/kg per day	Primary and secondary outcomes of the QI process and interventions demonstrated significant increases in protein provision from 33% to 71% for the first 5 d of CRRT treatment and from 39% to 75% for treatment duration only after PDSA 3 (group 2 versus 3; both P < 0.01)
Folic acid and homocysteine
Schröder et al. (1999)16	Case–control study	21	CKD on dialysis	9.4	2.5 mg folic acid daily	• The median (IQR) plasma homocysteine concentration was 12.0 (9.8–14.3) after 4 wk of folic acid treatment versus 20.0 (13.7–26.0 µmol/L) before treatment (P < 0.0001)
Merouani et al. (2001)17	Case–control study	86 • CRF: 29 • Healthy control: 57	CKD versus age-matched and sex-matched healthy children	12.0	None	• Homocysteine concentrations were higher in patients versus controls (17.3 versus 6.8 µmol/L, P < 0.0001), and hyperhomocysteinemia (>95th percentile for controls: 14.0 µmol/L) was seen in 62.0% of patients versus 5.2% of controls • Folate concentrations were lower in patients versus controls (9.9 versus 13.5 nmol/L; P < 0.01) • Vitamin B12 was similar in patients and controls (322 versus 284 pmol/L; P > 0.05)
Kang et al. (2002)18	Case–control study	14	Predialytic CKD	8.9	Folic acid supplementation (1 mg daily for 4 wk)	• The mean (SD) homocysteine concentration decreased from pretreatment (13.2 [6.6]) to post-treatment (9.3 [4.2] µmol/L) • Folic acid supplementation also lowered the prevalence of hyperhomocysteinemia from 64.3% to 42.9%
Phosphorus, calcium, and magnesium
Behnke et al. (1998)19	Case–control study	90 • Conservative Treatment: 42 • Dialysis: 22 • Transplantation: 26	CKD	11.6	CKD with conservative therapy, dialysis, or transplantation	• The difference in serum phosphorus among patients on the dialysis versus conservative therapy group (34.4%) and with transplantation versus conservative therapy group (−16.6%) • The difference in serum calcium among patients on the dialysis versus conservative therapy group (5.3%) and with transplantation versus conservative therapy group (2.4%)
Baskin et al. (2004)20	Prospective study	24 • Group I: five patients • Group II: ten patients • Group III: nine patients	Chronic dialysis with severe secondary hyperparathyroidism	13.0	Oral or IV calcitriol	• Group I: with normal parathormone levels without calcitriol • Group II: with secondary hyperparathyroidism receiving oral calcitriol • Group III: with secondary hyperparathyroidism receiving IV calcitriol • The change in serum phosphorus at the end versus beginning of the study: group 1 (by −31.6%), group 2 (by −1.3%), group 3 (IV by −6.5%) • The change in serum calcium at the end versus beginning of the study: group 1 (by 3%), group 2 (by 17.3%), group 3 (by 16.9%)
Santiago et al. (2009)21	Prospective observational evaluation study	85	Critically ill children receiving CRRT	4.9	Phosphate added in replacement and/or dialysate fluid	• Hypophosphatemia in patients that received supplement versus those that did not was 85% versus 55% (P < 0.001) • Severe hypophosphatemia in patients that received supplement versus those that did not was 57% versus 16% (P < 0.013)
Ahlenstiel et al. (2010)22	Prospective observational study	16	CKD	12.0	Phosphate education program concept	• The mean (SD) serum phosphate level decreased from 1.94 (0.23) at baseline to 1.78 (0.36)  mmol/L (P = 0.20) in weeks 19–24 • No significant changes in mean serum calcium 2.66 (0.3) at baseline versus 2.66 (0.23) mmol/L (P = 0.21) in weeks 19–24
Aygun et al. (2018)23	Retrospective study	447	PICU patients requiring CRRT	3.77	CRRT	• Patients requiring CRRT are 5.095 times more likely to develop hypomagnesemia and 4.322 times more likely to develop hypocalcemia
Amino acid
Zappitelli et al. (2009)14	Prospective observational study	15	Critically ill on CRRT	7.7	CRRT (CVVH)	• The change in different amino acids at day 2 versus baseline: aspartic acid (25%); alanine (32.8%); cysteine (−37%); isoleucine (38.7%); arginine (72.1%) • The change in different amino acids at day 5 versus baseline: taurine (−58.1%); aspartic acid (225%); glutamic acid (108.8%); proline (46.8%); alanine (45.1%); isoleucine (45.2%); leucine (29.5%); ornithine (42.6%); arginine (30.2%) • Reported only for those with either +25% or −25% change
Maxvold et al. (2000)24	Prospective randomized crossover study	6	Critically ill with AKI on CRRT	10.9	CRRT (CVVH and CVVHD)	• Amino acid clearance was greater in CVVH than in CVVHD, except for glutamic acid • CVVH glutamic acid clearance: 6.7±2.31 ml/min per 1.73 m2; CVVHD glutamic acid clearance: 7.6±2.79 ml/min per 1.73 m2
Kuttnig et al. (1991)25	Prospective observational study	11	Critically ill anuric on CRRT	—	CRRT (CAVH)	• Mean amino acid loss in ultrafiltrate was 0.159±0.008 (SEM) g/kg per day
Phan et al. (2006)26	Retrospective study	7	MSUD on CRRT or IHD	7.5	CRRT or intermittent hemodialysis	• Integrated leucine clearance ranging from 8.1 to 45.5 ml/min per 1.73 m2
Aygun et al. (2018)23	Retrospective record analysis	14	Metabolic disorders on CRRT	0.1	CRRT	• The change in leucine at 4 h versus baseline (by −58.6%) and at 8 h versus baseline (by −89.0%)
Aygun et al. (2019)27	Retrospective record analysis	97	CRRT in children with metabolic disease (versus others)	3.77	CRRT	• The ammonia reduction rate was 3.93±3.68 (per hour) in the MSUD group, 4.94±5.05 (per hour) in the UCD group, and 5.02±4.54 (per hour) in the organic acidemia group. The leucine reduction rate was 3.88±3.65 (% per hour)
Akduman et al. (2020)28	Retrospective chart review	8	CVVHDF in children with inborn metabolic disease	10±8.6 d	CRRT (CVVHDF)	• The mean plasma levels of ammonium were 1120±512.6 and 227.5±141.6 mg/dl before and at the end of the treatment, respectively • Plasma levels of leucine were 2053.5±1282 and 473.5±7.8 μmol/L before and at the end of the treatment, respectively
Lion et al. (2022)29	Prospective observational cohort study	15	CVVHDF in the PICU population	2	CRRT (CVVHDF)	• The median mass removal of amino acids was 299.0 (174.9–452.0) mg/kg per day • Median amino acid clearance was 18.2 (13.5–27.9) ml/min per m2, and median percent loss protein was 14.6% (8.3%–26.7%)
Sgambat and Moudgil (2016)30	Retrospective database review	42	Children with AKI receiving CRRT	7.9	CRRT	• At baseline: 30.7% and 35.7% were TC and FC deficient • At 1 wk: 64.5% (P = 0.03) and 70% (P = 0.03) were TC and FC deficient • At 2 wk: 80% (P = 0.01) and 90% (P = 0.008) were TC and FC deficient • At ≥3 wk: 100% were TC and FC deficient (P = 0.005 and P = 0.01, respectively, versus baseline)
Sgambat et al. (2021)31	Controlled pilot cohort study	48	• Children with AKI on CRRT: 9 • CRRT control: 10 • ICU control: 9 • Healthy control: 20	10	IV carnitine (20 mg/kg per day) added to TPN	• After carnitine supplementation during CRRT, TC, and FC increased from 36.0 to 18 μmol/L to 93.5 and 74.5 μmol/L, respectively • In control CRRT, TC, and FC decreased from 27.2 to 18.6 μmol/L to 12.4 and 6.6 μmol/L, respectively; lower than ICU controls in which TC was 32.0, FC 26.0 μmol/L
Miscellaneous micronutrients
Warady et al. (1994)32	Case series	7	CKD on peritoneal dialysis	0.9	Daily vitamin supplement devoid of vitamin A; dietary vitamin intake was derived from infant formula	• In all cases, the patients' blood concentrations of the water soluble vitamins were equal to or greater than normal infant values • Serum vitamin A levels were elevated despite the lack of supplementation
Pasko et al. (2009)33	Case series	5	Critically ill children with AKI receiving CVVHDF	7.48	Daily standard supplementation of trace elements in pediatric PN	• The median clearance of chromium, copper, manganese, selenium, and zinc during CVVHDF was calculated as 0, 0.59, 2.48, 1.22, and 1.90 ml, respectively, per 1.73 m2 body surface area per minute
Rianthavorn and Boonyapapong (2013)34	Case–control study	20	CKD stages 5	8.0	Oral ergocalciferol (treatment) versus placebo (control)	After 12-wk study • Serum 25D levels in the treatment group were significantly increased from baseline (P = 0.02) and were significantly higher than the serum 25D levels in the controls (P < 0.05)
Escobedo-Monge et al. (2019)35	Randomized trial multicentric study	48	CKD	13.0	Zinc supplementation (15–30 mg/d)	• Despite zinc supplementation, there were no significant changes in zinc levels (pretreatment 73.5 µg/dl versus post-treatment 75 µg/dl; P > 0.05)
Feeding modality
López-Herce et al. (2006)36	Prospective observational study	With AKI: 53 (with CRRT: 38, without CRRT: 15) Other: 473	Critically ill children with AKI (with or without CRRT) versus critically ill children without AKI	Median ages (months) With AKI: 18 AKI with CRRT: 21.5 AKI without CRRT: 7 Other: 5	Transpyloric EN	• Children with AKI more frequently on PN before TEN (56.6%) versus non-AKI (17.5%) • Incidence of shock, hepatic alterations and mortality was significantly higher with AKI • Regarding nutrition, the duration of previous PN was higher in children with CRRT (10 d) than without CRRT (7 d)
Wong Vega et al. (2019)37	Prospective observational study	41	Patients receiving CRRT for >48 h	9	EN and/or PN	• Protein goals met  Only EN: 12.7%±15.5% of the time  Only PN: 35.8%±34.9% of the time  EN+PN: 60.4%±41% of the time • When weaned to full EN support from EN+PN, the average percentage of time meeting protein goals were met decreased from 47.6% to 20.5%±29.9% (n=16) (P < 0.01)

CI, confidence interval; CRRT, continuous RRT; CVVH, continuous venovenous hemofiltration; CVVHD, continuous venovenous hemodialysis; CVVHDF, continuous venovenous hemodiafiltration; EN, enteral nutrition; FC, free carnitine; ICU, intensive care unit; IQR, interquartile range; IV, intravenous; LOS, length of stay; MSUD, maple syrup urine disease; OR, odds ratio; Plan-Do-Study-Act, PDSA; PEW, protein–energy wasting; PICU, pediatric intensive care unit; PN, parenteral nutrition; Quality Improvement TC, total carnitine; TEN, transpyloric enteral nutrition; UCD, urea cycle disorder.

Our workgroup, an international team of pediatric nephrologists and nutrition experts, conducted an integrative literature review to develop clinical practice points for nutritional assessment and management in critically ill children with AKI receiving CRRT. This article presents our methodology, literature review, CRRT clinical practice points, and recommendations for future research.

Methodology

Within the team, comprising 14 physicians and seven experts in pediatric nutrition, groups were assembled as follows: a core working group and a review panel (See Supplemental Table 6). The core working group defined the scope of the project, formulated the clinical questions to be addressed, performed a literature review, and drafted the clinical practice points.

The core working developed clinical questions to assess nutrition therapy needs and describe the associated nutritional challenges specific to children with AKI and those treated receiving CRRT therapy, using the following criteria:

Population

Infants, 37 or more weeks' gestation, to children age 18 years, with AKI receiving CRRT.

Intervention/Input Exposure

Nutrition therapy for the population of interest receiving CRRT.

Comparator

No comparator.

Outcomes

Optimize individual nutritional therapy to meet the unique needs of the population of interest to minimize catabolism.

Study Types

Case–Control Studies, Case series, Case Reports, Cohort studies, Randomized control trials, Systematic reviews, Metanalyses, Literature reviews, Abstracts, Review articles.

The Population, Intervention, Comparison, Outcomes, Study Design (PICOS) table (Supplemental Table 1) and a list of specific PICO questions can be found in the Supplemental Material.

An electronic search using PubMed and an inclusive academic library search (including MEDLINE, Cochrane, and Embase databases) was conducted to find relevant English language articles using different combinations of the following search terms: AKI, AKI, pediatrics, CRRT, renal failure, kidney failure, kidney dysfunction, indirect calorimetry (IC), nutrition, micronutrients, children, nutrition assessment, energy expenditure, calories, calorie needs, energy needs, macronutrients, protein, protein wasting, potassium, carnitine, malnutrition, lipids, vitamins, vitamin B1, thiamine, vitamin B6, pyridoxine, vitamin B9, folic acid, folate, vitamin C, ascorbic acid, trace elements, zinc, copper, ceruloplasmin, selenium, fluid overload, fluid shifts, enteral nutrition (EN), parenteral nutrition (PN), pediatric critical illness, fluid balance, and effluent content. Articles published between 1980 and January 2023 were included. All retrospective, case–control, randomized controlled trials, case series, and case reports were included in this study. Systematic reviews and meta-analyses, literature reviews, and abstracts were not withheld. Review articles were searched to identify relevant articles. All articles were reviewed by at least two independent reviewers. Data were extracted by at least two members, prepared in evidence tables, and reviewed by all members of the group. Supplemental Figure 1 shows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram. Some studies that were outside the purview of the literature review, but contributed important information, have been included in the discussion.

On the basis of this literature review, a summary of existing pertinent literature and 45 clinical practice points were developed. The core working group then met with the review panel, a multidisciplinary, multi-institution panel of pediatric nephrologists and experts in nutrition. The literature summary was distributed to the panel, who met via a series of web conferences every 2–3 weeks to develop the initial practice points. We then used a modified Delphi method to finalize clinical practice points. An electronic survey with 45 recommendations was distributed among the participants in January 2023. Each clinical practice point was individually and anonymously graded by the panelists on a five-point grading system, where 1=absolutely disagree, 2=disagree, 3=agree, 4=more than agree, and 5=absolutely agree. These experts added their comments on existing clinical practice points and proposed new ones. Comments were incorporated, and new clinical practice points added to the list by the core working group and re-presented to the panel. Consensus was reached for the clinical practice points if ≥75% of the respondents chose a score of 3–5, and these were included in the final clinical practice points. Any clinical practice points that did not meet this criterion were reviewed by the Delphi panel (W.V. Shih, V.S. Vitale, A.-M. Brown, and T.E. Bunchman), modified, and then resent back to the participants of the survey. The process was repeated two times. Fourteen members participated in the initial round, and nine members participated in the second round. Any practice points that did not reach consensus criteria after the third round of voting were eliminated (Supplemental Figure 2).

Literature Review

Below, we have highlighted the challenges seen in critically ill children with AKI requiring CRRT. On the basis of this literature review, clinical practice points were developed, as presented in Table 2.

Table 2.

Clinical practice points

Topic	Practice Points	Quality of Evidence	Gradea	Average Score (Out of 5)	% in Agreement with Clinical Practice Point
Nutritional assessment	1. Baseline information from family and chart review, focusing on history of weight loss or poor weight gain velocity before and during hospitalization	Weak	C	4.4	100
	2. NFPE should be conducted	Weak	C	4.5	100
	3. Use anthropometric measurements such as  a. In children <2 yr—admission weight for length z-score  b. In children >2 yr—admission BMI z-score  c. MUAC	Weak	C	3.9	100
Energy needs	4. Determining energy needs  a. IC is the best method to determine caloric dosing in CRRT; the correction factor is not required	Weak	C	3.5	92.3
	b. If IC is not available, the use of equations (Schofield, FAO/WHO/UNU) is suggested in critically ill patients   ▪ <3 yr–use SDI as the equations give extremely low energy goals in the age group; subtract 5–10% for those on a ventilator    i. Calorie estimates should be adjusted based on ICU therapies   ▪ >3 yrs–use the Schofield equation if there is a weight and height; otherwise, use the WHO equation    ii. Use of stress factors should be done cautiously and based on individual patient needs rather than any global recommendations    iii. Goal of meeting 60% energy need within 4–7 d    iv. Recommendation for assessment/consideration of nonintentional calories (propofol, CRRT substrates; solutions containing glucose, citrate, or lactate)	Weak	C	3.5	75
Energy prescription	5. During the early acute phase of illness, energy expenditure is decreased, and not exceeding recommended caloric intake is suggested	Moderate	B	3.7	92.3
	6. When assessing ratio of macronutrients in patients on CRRT, it should be noted that protein needs are higher (related to losses in dialysate); thus, while patients on CRRT may have increased protein requirements, overall energy/calorie intake must be considered to assure adequate carbohydrates and fats are included	Moderate	B	3.9	87.5
Fluid balance	7. Fluid intake (EN, PN, dialysate, and replacement solutions) and output (fluids removed by CRRT, urine, gastrointestinal losses, and insensible losses) should be closely monitored in children with AKI on CRRT	Moderate	C	4.3	100
	8. Fluid status can be monitored in patients with AKI by the following measures: daily weights, intake/output, and physical examination findings	Strong	B	3.9	92.9
	9. Concentrated enteral feeding formulas or maximally concentrated parenteral feeding formulas may be considered before CRRT initiation to minimize fluid delivery and to allow the fluid balance to be managed primarily outside of nutrition support. Once CRRT is initiated, fluid can be liberalized to meet nutritional needs	Weak	C	4.1	100
	10. Enteral and PN should be incorporated into fluid balance, and CRRT should be adapted	Weak	C	4.1	100
Carbohydrates	11. Individualized energy requirement assessment is necessary for appropriate energy prescription	Weak	C	4.1	100
	12. Dialysate glucose concentration of 1–2 g/dl is recommended in critically ill children with AKI requiring CRRT.	Strong	B	3.6	100
	13. CRRT anticoagulant and buffer solutions may contribute hidden calories, and this should be accounted for during treatment if anticoagulation changes as caloric intake would also change	Strong	B	3.7	92.9
Proteins	14. Protein intake in children with AKI should be adjusted to achieve a positive nitrogen balance	Moderate	B	4.1	92.9
	15. We recommend augmenting protein intake in critically ill children on CRRT as follows: 0–2 yr, 3–3.5 g/kg per day; 2–13 yr, 2–2.5 g/kg per day; and 13–18 yr, 2 g/kg per day, which may vary depending on initial and ongoing nutritional assessment	Strong	B	3.7	92.3
	16. Protein intake in children with AKI on CRRT should be appropriate for age with close laboratory monitoring and should not be limited	Strong	B	4.0	92.9
Amino acids	17. Loss of glutamine and carnitine has been reported in pediatric patients on CRRT, but there is inadequate evidence to provide specific replacement guidelines. Monitoring of levels and individual repletion may be considered, especially in patients on CRRT >5–7 d	Moderate	B	4.1	100
Fats	18. TG levels should be monitored daily in pediatric critically ill patients who are on PN until a steady state is achieved and then at least weekly while on PN.	Weak	C	3.9	75
	19. Lipid supplementation on PN: start with 1 g/kg per day and increase to 2–4 g/kg per day over several days. Maximum in infants: 3–4 g/kg per day. Maximum in older children: 2–3 g/kg per day. Laboratory monitoring required to avoid lipid overload	Weak	C	4.1	100
	20. It is recommended that 30%–40% of calories should be derived from fat in these patients, varying depending on the patient and their nutritional needs/initial assessment. Lipid supplementation should start low and be progressively increased as tolerated with laboratory monitoring	Weak	C	4.1	100
	21. Multicomponent lipid emulsions containing soybean oil, MCT, olive oil, and fish oil (SMOFlipid 20%) are preferred in pediatric patients on CRRT due to their positive immunomodulatory and anti-inflammatory effects	Moderate	B	3.6	100
	22. Early enteral feeding and fish oil supplementation have been associated with better outcomes	Weak	C	3.7	84.6
Electrolytes	23. Electrolytes and acid–base status should be monitored and adapted every 6–8 h. Monitoring adjusted as the clinical status and laboratory results change	Weak	C	3.8	92.3
	24. Sodium levels should be corrected gradually; low sodium CRRT solutions may be used to prevent rapid correction of hyponatremia	Strong	B	3.9	100
	25. High concentration citrate solutions used during CRRT are hypertonic in sodium and can lead to hypernatremia. Sodium levels should be monitored and corrected as required	Weak	C	4.3	100
	26. Phosphorus supplementation through replacement or dialysate fluid during CRRT is safe and reduces the need for IV and oral supplementation	Moderate	B	3.9	100
	27. Replacement and dialysate fluids contain potassium, however supplementation on the basis of laboratory findings may be necessary	Weak	C	3.8	100
	28. Low potassium levels should be avoided, and during replenishment, a rapid correction should be avoided to prevent serious adverse events	Strong	B	3.7	100
Vitamins	Vitamin supplementation recommendations 29. Vitamin B1 (thiamine)  i. Addition of 0.83 μg/L effluent plus SDI  i. SDI is as follows   1.<12 mo: 0.3 mg   2. 1–3 yr: 0.5 mg   3. >3 yr: up to 1.2 mg	Weak	C	3.9	83.3
	30. Vitamin B6 (pyridoxine)  i. Supplementation may be required, further studies are required to make recommendations	Weak	C	3.5	87.5
	31. Vitamin B9 (folic acid)  i. Check levels twice weekly  ii. Addition of 4.4 μg/L effluent fluid plus SDI.  iii. SDI is as follows   1. <12 mo: 48 μg   2. 1–3 yr: 90 μg   3. >4 yr: 240 μg	Weak	C	3.4	90.9
	32. Vitamin B12 (cobalamin)  i. No recommendation as it is not lost in effluent	Moderate	B	3.5	91.6
	33. Vitamin C (ascorbic acid)  i. Addition of 1.9 mg/L effluent plus SDI  ii. SDI is as follows   1. <12 mo: 50 mg   2. 1–3 yr: 15 mg   3. >3 yr: 90 mg	Weak	C	3.6	90.9
	34. Fat soluble vitamins  i. No recommendation as they are not lost in effluent	Moderate	B	3.8	100
Trace elements	35. Regular monitoring of trace elements levels should be considered in children receiving CRRT, specifically in prolonged CRRT (beyond 7–10 d)	Weak	C	3.9	100
	36. Loss of selenium has been observed in pediatric patients on CRRT, calculated to be 15% of SDI. If required, selenium should be replaced. A weekly replacement may be necessary	Weak	C	3.5	92.3
	37. Plasma levels of copper and zinc have been slightly decreased in some children with CRRT; no recommendations for supplementation at this time	Weak	C	3.4	100
Route and Timing of nutritional support	38. EN is preferred over PN wherever feasible, with oral intake and early breastfeeding being the preferred method	Moderate	B	4.3	100
	39. EN should be started within 24–48 h of PICU admission as tolerated, and interruptions should be minimized	Moderate	B	4.1	100
	40. Polymeric feeds are the first choice for EN	Moderate	C	3.9	100
	41. Protein and energy dense formulations may be considered in children with AKI requiring fluid restriction	Moderate	B	4.1	100
	42. If nutritional goals are not met with EN after 3–5 d, PN may be used to supplement EN	Weak	C	4.0	92.8
	43. If patients on CRRT cannot initiate EN, PN should be started within 2–3 d for infants/toddlers and 3–5 d for adolescents	Moderate	B	4.0	100
	44. Formulas should be decided on an individual basis. Renal formulas are typically not required for those on CRRT	Weak	C	4.1	100

BMI, body mass index; CRRT, continuous RRT; EN, enteral nutrition; IC, indirect calorimetry; ICU, intensive care unit; IV, intravenous; MCT, medium-chain triglycerides; MUAC, mid–upper arm circumference; NFPE, nutrition-focused physical assessment; PICU, pediatric intensive care unit; PN, parenteral nutrition; SDI, suggested dietary intake; TG, triglyceride; WHO, World Health Organization.

^aGrading: four levels of aggregate evidence quality: A, B, C, and D (see Figure 2).

Assessment of Malnutrition in Children with AKI during CRRT (Clinical Practice Points 1–3)

The nutritional status of a child who is critically ill should ideally be assessed within the first 48 hours of admission.⁷ A significant number of infants and children in the PICU, particularly those with chronic illnesses, exhibit preexisting malnutrition.³⁸ Many screening tools have been developed to identify malnutrition or the risk of malnutrition in pediatric patients.³⁹ A nutrition screening tool is intended to be simple and completed on admission by any health care provider to assess the need for a more detailed nutrition assessment. Presently, there is no consensus on which tool is best for identifying malnutrition or risk of malnutrition in pediatric critical illness.^40,41 It can be assumed that all critically ill patients with AKI who require CRRT are at risk of nutrition deterioration regardless of nutrition status on admission and benefit from a nutrition assessment.

A complete nutrition assessment includes a detailed review of the medical history and dietary intake before admission, a review of anthropometric measurements, a nutrition-focused physical assessment to identify or confirm muscle wasting, subcutaneous fat loss, specific micronutrient-related deficiencies and edema, and a review of laboratory data.^42,43

Anthropometric measurements, including weight, length/height, and head circumference for children younger than 3 years, should be obtained on admission. Weight for length for children younger than 2 years or body mass index for children older than 2 years should be assessed. Measurements are plotted on the World Health Organization (WHO) growth charts for children younger than 2 years and the Centers for Disease Control and Prevention growth chart for older children. If serial measurements are available, the velocity of weight gain should be calculated for those younger than 2 years and compared with standards published by the WHO.⁴⁴ For older children, serial measurements allow for comparison over time. Weight loss or percentage of usual body weight should be noted.

Mid–upper arm circumference (MUAC) is an easily obtained measurement and is useful to diagnose malnutrition and changes in nutritional status over time for children 6 months through 18 years.^45,46 MUAC is sensitive to changes in nutritional status, and unlike daily weights, it is not typically affected by ascites or edema.

Using anthropometric z-scores for weight, weight/length, body mass index, and MUAC will help to determine the degree of malnutrition. The z-score is a value that describes how far a child is away from the mean. A z-score in the range of –1 to –1.99 is mild, and a z-score –2 to –2.99 is moderate, and z-score >3 is classified as severe malnutrition. The Pedi Tools website provides electronic calculators for growth charts and z-scores.⁴⁷

Laboratory values, including prealbumin and albumin, have been used as nutrition markers. The concentration of these proteins decreases during the acute phase response to illness and injury, particularly when inflammation is present. Serum levels are not influenced by nutrition and do not rise until inflammation subsides.⁴⁸ Although albumin is most frequently used, there is inconsistent evidence for an association between albumin and mortality in patients with AKI. Prealbumin has a brief half-life of approximately 2 days allowing levels to reflect recent dietary intake rather than overall nutritional status. Repeated measurements may be used to monitor nutrition over an extended period. Findings indicate a moderate correlation between prealbumin levels and daily caloric intake among PICU patients, even after accounting for daily C-reactive protein levels and illness severity. This relationship is particularly pronounced in children aged older than 1 year.⁴⁹ As prealbumin is renally excreted, serum levels may be affected by kidney disease, thus complicating its use as a biomarker in AKI.⁵⁰ A serum prealbumin level of ≤10 mg/dl was reported to be associated with increased 90-day mortality. Another study observed that serum prealbumin levels of <11 mg/dl were associated with a higher risk of death in adult patients with AKI.⁵¹ In critically ill patients with AKI, serum cholesterol levels <96 mg/dl were associated with greater 28-day mortality. Guimaraes et al. also showed that decreased levels of IGF-1 and cholesterol were associated with increased mortality in intensive care unit patients.⁵² Similarly, serum cholesterol is nonspecific and insensitive in critically ill patients, as it is associated with systemic inflammation, organ dysfunction, and infection.^53,54 There is currently insufficient evidence to recommend specific acute serologic monitoring of these markers for children and adolescents with AKI. Thus, laboratory values, such as albumin, prealbumin, and serum cholesterol, should be interpreted at the discretion of the medical team in combination with the clinical situation.

Alternative tools that have been suggested but require further study include IGF-1, bioimpedance analysis, and ultrasound of quadriceps muscle. Low levels of IGF-1 (<50.6 ng/ml) indicate a catabolic state and have been associated with increased mortality.⁵⁰ Bioimpedance analysis can assess whether a patient is malnourished based on their body composition (fat mass, lean mass, intravascular water, interstitial water, and total body water).^50,53,55,56 However, fluid overload leads to increased estimates of muscle mass, whether measured by bioimpedance, computerized scanning, magnetic resonance imaging, or ultrasound. Many of these tools are not widely available or practical for bedside assessment.

Determining Accurate Caloric Dosing and Challenges with CRRT (Clinical Practice Point 4)

There are several available calorimetry techniques (Supplemental Figure 3). IC is considered the gold standard, measuring the venous O₂ and venous CO₂, which correlate with energy usage in cells, to calculate resting energy expenditure (REE) according to Weir's equation.^57,58 REE is potentially affected by underlying conditions, types of filters, ultrafiltration and dialysate flow rates, medications, and thermal cooling as heat energy is lost to the dialysate and replacement fluids. Traditionally, IC is measured in a stable, immobile patient, as REE will be affected by physical activity.^12,50,59,60 Despite these disadvantages, IC is still considered the method of choice for determining caloric dosing if available.⁵⁸

CRRT creates an additional challenge to IC, as CO₂ is removed from the blood as bicarbonate, underestimating energy needs. CRRT also triggers heat loss, possibly up to 1000 kilocalories per day,⁶¹ called dialysis trauma. Blood in the extracorporeal circuit will lose thermal energy with exposure to cooler dialysate and replacement fluids, which cannot be fully compensated with heating. This cooling effect with CRRT has yet to be studied in nonsedated PICU patients.^12,50,59,61 Caloric-containing components in dialysate exchanged during CRRT, such as citrate, which acts as an anticoagulant for dialysis and a Krebs cycle substrate, should also be factored into caloric dosing using IC.^50,59 Formulas for estimating caloric delivery from nonintentional calories in substitution fluid of continuous venovenous hemofiltration (CVVH) in adults are elaborated in Supplemental Table 2.

The MEtabolic Consequences of Continuous venovenous hemofiltration on Indirect cAlorimetry trial investigated the effect of CO₂ removal and citrate anticoagulation on the measurement of REE. The adjusted REE increased by 2% for low-dose filtration settings with citrate compared with the same CRRT settings with NaCl predilution and by 3% with higher-filtration dose settings and citrate predilution. Accounting for CO₂ losses, there was then no statistical difference in adjusted REE between low-dose and high-dose ultrafiltration with citrate; however, there was still a significant difference without citrate. Any inflammatory effects of the extracorporeal circuit were either minimal or counteracted by thermal cooling effects. This was a small study, and the results need to be interpreted cautiously.^62,63 In a later retrospective analysis, low-dose and high-dose CRRT settings with citrate had no significant difference in an absolute positive balance but did when compared with NaCl predilution.⁶¹ Supplemental Figure 4 presents a schematic representation of the above findings in the retrospective analysis of the MEtabolic Consequences of Continuous venovenous hemofiltration on Indirect cAlorimetry trial.⁶²

Several equations have been developed to estimate caloric dosing, as presented in Supplemental Table 3. Schofield was closest to IC for healthy male children, while Food and Agriculture Organization/WHO/United Nations University was more accurate for female children. None of these equations have been validated for pediatric patients treated with CRRT, as weight and height alone do not accurately represent their complex metabolism. These equations often overestimate REE with <40% agreement, and the Oxford equations have limited validation.^64,65 Usage of the Schofield equations is recommended if IC is unavailable; however, there is a lack of evidence supporting this during CRRT.^12,50 The advantages and disadvantages of the various methods of measuring caloric dosing are presented in Supplemental Table 4.

Energy Prescription (Clinical Practice Points 5–6)

Providing adequate energy intake in pediatric patients with AKI is challenging as nutrient utilization in these patients is often suboptimal, and both underfeeding and overfeeding can lead to complications.⁶⁶ Consideration of age must be taken into account when calculating nutrient utilization as basal metabolism increases rapidly in the first 2 years of life and declines steadily into adolescence.⁶⁷ Some authors have recommended a caloric intake of 20%–30% above requirements estimated using IC or equations to estimate REE. In most children with AKI, this likely provides adequate calories without a risk of significant overfeeding.⁸ During the early acute phase of illness, energy expenditure is decreased, and not exceeding recommended caloric intake could be suggested.⁶⁸ Lipids, proteins, and carbohydrates should be included as energy provisions for these patients. Insulin may be required, and glucose should be carefully monitored as children with AKI are prone to insulin resistance. Hyperglycemia should be avoided as it is associated with poor clinical outcomes.⁵ Protein and carbohydrate metabolism are closely related. Lipogenesis may be induced if there is an overload of carbohydrates. If energy provision is insufficient, amino acids cannot be adequately used irrespective of the supply.^5,69 In critically ill children, such as those with AKI, increased lipid oxidation and reduced glucose oxidation rates may occur. Owing to the increased lipid demands and limited stores, children are potentially susceptible to essential fatty acid (FA) deficiency. The development of hemofilters has led to an increase in the size of solutes that can now be cleared, which means that lipoproteins up to C16 or C17 may be potentially removed. When assessing the ratio of macronutrients in patients on CRRT, it should be noted that protein needs are higher (related to losses in dialysate); thus, while patients on CRRT may have increased protein requirements, overall energy/calorie intake must be considered to assure adequate carbohydrates and fats intake.^5,70

Fluid Balance (Clinical Practice Points 7–10)

The medical community has increasingly recognized the importance of fluid balance, including nutritional fluids and drugs. The four Ds of fluid management should be considered: drug or type of fluid, dose, duration, and de-escalation.⁷¹ Fluid imbalances and hemodynamic instability are often seen in critically ill children.⁷² In oliguric and anuric AKI, there is decreased renal excretion, contributing to fluid retention. Sutherland et al. showed that fluid overload before initiating CRRT in critically ill children is associated with higher mortality. A 1% increase in the severity of fluid overload led to a 3% increase in mortality.⁷³

A meta-analysis, including 44 pediatric studies, confirmed that fluid overload generally is associated with significant morbidity and mortality in critically ill individuals.⁷⁴

The following measures can be used to monitor fluid status: daily weights, intake and output records, physical examination findings, and more invasive measures, including central venous pressure or pulmonary artery occlusion pressures. Frequent weight measurements may not be feasible in the PICU in unstable patients, and potential weight loss during a PICU stay may be masked by fluid overload. Moreover, accurately measuring fluid overload in infants and young children undergoing dialysis presents challenges due to the fact that a portion of the interdialytic weight gain comprises nutritional weight gain, which can equate to a 5%–10% alteration in target weight per week.⁷⁵ Therefore, electrical bioimpedance or point-of-care ultrasound can aid clinical assessment of fluid status in these children.⁷⁴ Parameters used to assess fluid status using point-of-care ultrasound include sonographic lung B-lines, inferior vena cava (IVC) diameter, and IVC collapsibility index and IVC/aorta ratio. During CRRT, fluid intake comprises enteral and intravenous (IV) fluids (including PN, dialysate, and replacement solutions). Fluid output in these patients includes fluids removed by CRRT, urine, blood loss, gastrointestinal losses, and insensible losses.⁷⁶ Fluid administration and removal are carefully regulated to avoid unnecessary positive fluid balance, especially in critically ill children. PN is more prone to contributing to a positive daily fluid balance.¹² Concentrated enteral feeding formulas or maximally concentrated PN formulations are typically used before CRRT initiation to minimize fluid delivery and to allow the fluid balance to be managed primarily outside of nutrition support.⁷⁷ Once CRRT is initiated, fluid can be liberalized to meet nutritional needs. It is important to note that we should not limit nutritional intake simply to conserve volume.

Carbohydrates (Clinical Practice Points 11–13)

Inflammation in patients with in children with AKI requiring CRRT alters carbohydrate metabolism.^78–80 Adequate carbohydrate intake is essential to counteract protein catabolism and lipogenesis.^5,69 Provision of carbohydrates via EN is preferred as it assists in maintaining intestinal mucosa, better regulation of glucose levels, and is associated with improved patient outcomes.⁸¹

In children with AKI, carbohydrates should comprise at least 20%–30% of the total energy intake.^5,12,70 CRRT increases heat loss through extracorporeal circuity and heightens metabolic activation due to immunologic activity.⁵ Both the increased heat loss and metabolism boost affect energy requirements. Additional caloric sources from medication infusions and CRRT solutions containing citrate, glucose, and lactate may be quantitatively relevant and should be considered in calculating delivered energy versus energy requirements.⁷⁸ Conversely, sedation and mechanical ventilation may decrease metabolic needs, underscoring the necessity of individualized patient assessment. Caloric values for citrate-containing fluids are 0.59 kcal/mmol, glucose 0.73 kcal/mmol, and lactate 0.33 kcal/mmol.⁷⁸ CRRT using trisodium citrate solution containing 300–500 mmol of citrate provides an additional 100–200 kcal/d. The use of an acid–citrate dextrose solution with a lactate buffer adds approximately 1200 kcal/d to the daily energy intake.^{12,59,62,63,78,82–86} Lower citrate protocols can reduce these “hidden” calories.⁸⁷ Glucose control is known to be renoprotective, suggesting the target range for serum glucose levels should be between 110 and 150 mg/dl^78,88–90; however, tight glucose control is not recommended in critically ill children with hyperglycemia as younger children are more susceptible to hypoglycemia than adolescents and young adults.^91,92 In addition, glucose can readily traverse the membrane during both convective (CVVH or continuous venovenous hemodiafiltration [CVVHDF]) and/or diffusive (continuous venovenous hemodialysis [CVVHD], CVVHDF) forms of CRRT.⁹³

Protein (Clinical Practice Points 14–16)

Adequate protein intake is imperative in critical illness as it supports wound healing, preservation of skeletal muscle mass, and reduced inflammation.⁹⁴ Patients at intensive care unit are at high risk of muscle atrophy secondary to immobility, sedative medications, and neuromuscular blockade. Protein turnover rates differ across various age groups. Newborns, for instance, exhibit a protein turnover rate of 6.7 g/kg per day, whereas adults maintain a protein turnover rate of approximately 3.5 g/kg per day.⁴³ Acute protein energy malnutrition is more prevalent among children younger than 2 years and has been shown to contribute to increased physiologic instability.⁹⁵ Increasing protein intake has been associated with decreased mortality in critically ill patients.⁹⁶ Children with AKI are at high risk of protein deficits due to their hypercatabolic state, resulting in a high urea nitrogen production rate (180–250 mg/kg per day) and a net negative nitrogen balance.^66,97 Therefore, protein intake should be adjusted to prevent skeletal muscle wasting and to achieve a positive nitrogen balance.^66,98

Several methods have been advocated for evaluating protein status in children, including determining nitrogen balance, urea nitrogen appearance, and normalized protein catabolic rates.⁵⁰ These are elaborated on in Supplemental Figure 5. However, none of these equations have been validated for use in children on CRRT, as protein removal varied on the basis of molecular weights.⁹⁹ Serum amino acid levels do not reflect total body stores or utilization,⁸ and amino acid assays are used for research purposes.^12,100 The Kjeldahl method can be used to measure the removal of nitrogen-containing molecules, but it cannot be used in patients with hyperammonemia and is a hazardous, lengthy, and labor-intensive process rarely used in practice.^12,101

CRRT provides additional challenges in determining adequate protein intake as amino acids, peptides, and proteins are lost through the hemofilter. Protein losses in various modalities are compared in Supplemental Table 5.^102–104 Many factors affect amino acid clearance in CRRT, including the use of convection (CVVH or CVVHDF) and/or diffusion (CVVHD or CVVHDF), blood flow rate, dialysis flow rate, effluent rate, and filter membrane properties.^100,105 Protein intake should not be restricted in patients on CRRT, given the increased losses observed. Bellomo et al. studied nitrogen metabolism in two consecutive cohorts of adults with AKI on CRRT and demonstrated that a high-protein diet could be safely administered to critically ill patients with AKI on CRRT.¹⁰⁶

Critically ill patients on CRRT require increased protein intake to achieve a positive nitrogen balance due to increased protein losses. A prospective, randomized cross-over study by Maxvold et al. demonstrated negative nitrogen balance in children on CRRT despite a protein intake of 1.5 g/kg per day.²⁴ Scheinkestel et al. showed that for adults on CRRT, a protein intake of 2.5 g/kg per day increased the likelihood of achieving a positive nitrogen balance and improving survival.¹⁰⁷ ASPEN guidelines for critically ill adults support protein provision of up to 2.5 g/kg per day for patients on CRRT.⁹⁴ Jonckheer et al. pointed out that compared with the adult ASPEN guidelines for critically ill patients of 1.2–2 g/kg per day, this is a 1.25–2 times increase in protein intake for patients on CRRT.¹² According to American Society of Parenteral and Enteral Nutrition (ASPEN) clinical guidelines for nutritional support of the critically ill child, the suggested protein intake is 0–2 years, 2–3 g/kg per day; 2–13 years, 1.5–2 g/kg per day; and 13–18 years, 1.5 g/kg per day.¹⁰⁸ On the basis of the literature, although limited, and adaptations from adult guidelines, we recommend augmenting protein intake in critically ill children on CRRT as follows: 0–2 years, 3–3.5 g/kg per day; 2–13 years, 2–2.5 g/kg per day; and 13–18 years, 2 g/kg per day.

Amino Acids (Clinical Practice Point 17)

Increased protein catabolism and protein turnover with redistribution of amino acids from skeletal muscle to other tissues is increased in AKI. Amino acids are primarily used for acute phase protein synthesis, gluconeogenesis, and ureagenesis.¹⁰⁹ In AKI, nonessential amino acids, such as tyrosine and arginine, become conditionally essential¹⁰⁵; furthermore, the most prevalent amino acid in skeletal muscle, glutamine, serves as an energy source for enterocytes and is conditionally essential under metabolic stress.¹¹⁰ Glutamine supports immune cells and enterocytes, protects against insulin resistance, and increases the production of antioxidants.¹¹¹ In critically ill adults on CRRT, supplementation of alanyl-glutamine dipeptide 0.3–0.6 g/kg per day is recommended yet; findings regarding supplementation for malnourished children have conflicting results.³⁸ In addition, when comparing modalities of CRRT, clearance of amino acids has been shown to be greater in CVVH than CVVHD.²⁴ Although glutamine losses during CRRT accounted for 25% of all amino acids in children, there are currently no guidelines for glutamine supplementation in children on CRRT.⁷⁸

Carnitine is an essential nutrient that aids in the transport of long-chain FAs into the mitochondrial matrix and is critical to the oxidation of lipids.³¹ It is depleted during CRRT, and its deficiency is associated with increased mortality.^30,112 Carnitine supplementation has been found to replete plasma levels and improve myocardial strain in PICU patients on CRRT.³¹ However, there is inadequate evidence to provide specific carnitine supplementation recommendations as plasma levels may not accurately reflect tissue levels and excretion of free carnitine is dependent on age, with excretion per kg body weight highest in the age group from 3 to 10 years.¹¹³ Similarly, CRRT can intentionally or unintentionally reduce leucine levels, but there is no evidence to guide empiric replacement.²⁷

Fats (Clinical Practice Points 18–22)

Fat homeostasis is believed not to be significantly influenced by CRRT; however, smaller lipoproteins may be lost in the ultrafiltrate. PN with lipid emulsions has been reported to cause vascular access problems, especially when heparin is used as an anticoagulant.^11,12

Critical illness and CRRT induce a proinflammatory state with an activated innate immune system associated with the release of cytokines, endotoxins, and catecholamines, as well as proinflammatory lipid mediators, including leukotrienes and prostaglandins.^{11,76,114,115} Impaired lecithin–cholesterol acyltransferase, involved in lipolysis, is seen in acute renal failure and leads to decreased HDL and LDL.^{30,114,116,117} Changes in fat metabolism are presented in Supplemental Figure 6. Dysregulation of mitochondrial FA β-oxidation due to carnitine deficiency also contributes to hypertriglyceridemia and increased utilization of FA, which is associated with increased inflammation, tissue toxicity.³⁰ Critically ill children are susceptible to FA deficiency due to high demands and limited fat stores if fat free diet is provided.^108,118 Fat metabolism is not significantly influenced by CRRT because of its high molecular weight and insolubility in water.^{12,30,114,116,117}

Hypertriglyceridemia is an increase in plasma fasting triglyceride (TG) concentration above the 95th percentile for age and sex.¹¹⁹ A TG level ≥100 mg/dl (1.13 mmol/L) and a level ≥130 mg/dl (1.47 mmol/L) are considered above the 95th percentile for children of ages 0–9 and 10–19 years, respectively.¹²⁰ Complications associated with hypertriglyceridemia include coagulopathy, thrombocytopenia, hepatomegaly, elevated liver enzymes, hyperbilirubinemia, and respiratory distress.^{11,12,114,116,121} Hypertriglyceridemia is an increase in plasma fasting TG concentration above the 95th percentile for age and sex.¹¹⁹ A TG level ≥100 mg/dl (1.13 mmol/L) and a level ≥130 mg/dl (1.47 mmol/L) are considered above the 95th percentile for children of ages 0–9 and 10–19 years, respectively.¹²⁰ Complications associated with hypertriglyceridemia include coagulopathy, thrombocytopenia, hepatomegaly, elevated liver enzymes, hyperbilirubinemia, and respiratory distress.^{11,12,114,116,121} TG levels should be monitored daily in pediatric critically ill patients who are on PN until goal lipid provision is achieved and then at least weekly while on PN.^5,12

PN with mainly soybean oil-based lipid emulsions is associated with increased inflammation, which may contribute to PN-associated liver disease. By contrast, multicomponent lipid emulsions containing soybean oil, medium-chain triglycerides, olive oil, and fish oil (SMOFlipid 20%) may be less inflammatory in pediatric patients on CRRT. The addition of IV docosahexaenoic acid (DHA) and eicosapentaenoic acid found in fish oil had a favorable effect on the cell membrane and inflammatory process.¹¹⁴

Early enteral feeding and fish oil supplementation have been associated with better outcomes and a reduced length of stay in critically ill patients.^{12,114,121–123}

Electrolytes (Clinical Practice Points 23–28)

Electrolyte abnormalities are common during critical illness, especially in patients with kidney failure. Electrolyte disturbances occur in over 50% of critically children and are often a reason to initiate CRRT.²¹ Electrolyte abnormalities are determined primarily by CRRT solute clearance and the electrolyte composition of the dialysate and replacement solutions.^12,21,77 Determining the need for electrolytes in patients receiving CRRT is more straight forward than determining energy and protein needs since laboratory values can be closely monitored and adjustments made accordingly. Sodium, potassium, magnesium, and phosphorus are lost in the effluent.¹²⁴ Supplementation of electrolytes needs to be adjusted individually depending on plasma levels. Electrolyte abnormalities are common during critical illness, especially in patients with kidney failure. Electrolyte disturbances occur in over 50% of critically children and are often a reason to initiate CRRT.²¹ Electrolyte abnormalities are determined primarily by CRRT solute clearance and the electrolyte composition of the dialysate and replacement solutions.^12,21,77 Determining the need for electrolytes in patients receiving CRRT is more straight forward than determining energy and protein needs since laboratory values can be closely monitored and adjustments made accordingly. Sodium, potassium, magnesium, and phosphorus are lost in the effluent.¹²⁴ Supplementation of electrolytes needs to be adjusted individually depending on plasma levels.

Hypokalemia and hyperkalemia can increase the risk of life-threatening arrhythmias. Hypokalemia is reported in 5%–25% of patients on CRRT.^125–127 Low serum potassium levels of <3 mEq/L should be avoided, and rapid correction may increase mortality.¹²⁸ Potassium is provided in the replacement and dialysate fluids, and additional potassium may be prescribed as needed on the basis of laboratory values. The risk of diarrhea with enteral potassium supplementation must be evaluated.⁷⁷ Studies discussing electrolyte losses in children on CRRT are summarized in Table 1.

Hyponatremia may occur when sodium balance is negative and not adequately compensated by replacement fluids and CRRT.²¹ Hyponatremia associated with AKI is primarily due to a hypervolemic state rather than a sodium deficit. Volume status must be carefully assessed when hyponatremia is present.^77,129 Sodium needs to be gradually corrected to avoid osmotic demyelination syndrome. Using a CRRT solution with a low sodium concentration has been demonstrated to be safe, feasible, and effective in avoiding rapid hyponatremia correction.¹³⁰ Citrate anticoagulation used during CRRT can be associated with hypernatremia, as some citrate solutions have higher sodium content than others. Sodium levels should be monitored. Modifying the flow and composition of solutions given to the patient is useful in preventing and managing hypernatremia.¹³¹ Hyponatremia may occur when sodium balance is negative and not adequately compensated by replacement fluids and CRRT.²¹ Hyponatremia associated with AKI is primarily due to a hypervolemic state rather than a sodium deficit. Volume status must be carefully assessed when hyponatremia is present.^77,129 Sodium needs to be gradually corrected to avoid osmotic demyelination syndrome. Using a CRRT solution with a low sodium concentration has been demonstrated to be safe, feasible, and effective in avoiding rapid hyponatremia correction.¹³⁰ Citrate anticoagulation used during CRRT can be associated with hypernatremia, as some citrate solutions have higher sodium content than others. Sodium levels should be monitored. Modifying the flow and composition of solutions given to the patient is useful in preventing and managing hypernatremia.¹³¹

Many studies report hypophosphatemia in CRRT treated patients, which may negatively affect patient outcomes. CRRT continuously clears the blood of phosphate and, along with intracellular shifts after the initiation of nutritional support, increases the risk of hypophosphatemia, with a reported prevalence of up to 85% of children receiving CRRT.^61,68,69 In addition, research suggests that children younger than 6 years are at a higher risk for developing hypophosphatemia than older children.¹³² Phosphorus levels should be monitored closely and treated accordingly. Supplementing phosphorus in the dialysis fluid is safe and reduces the need for IV and oral supplementation. When phosphate-free fluids are used, supplementation will likely be required. The route of phosphate supplementation should consider the increased risk of diarrhea with enteral administration and the availability of parenteral products. Low-phosphate renal formula feedings are not recommended for patients on CRRT since phosphorus restriction is not often necessary.

Vitamins (Clinical Practice Points 29–34)

Critically ill children receiving CRRT present with altered vitamin requirements due to underlying illness.^12,61,133 In 2019, Jonckheer et al. provided recommendations regarding vitamin supplementation in this population, as summarized in Table 3.¹² The optimal dose of vitamin 1, 25D, and ergocalciferol supplementation has not yet been determined in children with CKD stages 5/5D.³⁴ The results from adult patients demonstrate no discernible difference in vitamin B12 levels after 6 days of RRT.³⁹ Similarly, fat soluble vitamins are not significantly lost in either convective or diffusive RRT modalities.¹³⁵ Despite, not being lost in the effluent, fat soluble vitamin levels have shown to be disrupted in critical illness. These vitamins are affected by many factors, such as inflammation and cholesterol; thus, they must be interpreted with caution.¹³⁶ Low vitamin A levels have been associated with critically ill pediatric patients and may be a potential marker of mortalilty.¹³⁷ Despite the above recommendations, there are currently no data on whether increasing vitamin plasma levels to normal in these patients significantly influences morbidity or mortality 1.¹² Studies on vitamins in children treated with CRRT are presented in Table 1. Changes in vitamins during critical illness are illustrated in Figure 3.

Table 3.

Vitamin supplementation recommendations

Vitamin	Recommended Supplementation		Additional Supplementation
Vitamin B1: based on plasma levels and age	<12 mos	0.3 mg	83 μg/L effluent fluid
	1–3 yr	0.5 mg
	3+ yr	up to 1.2 mg
Vitamin B6	Needed but no recommendations
Vitamin B9 (folic acid): based on SDI, monitor twice weekly	<1 yr	48 µg	4.4 µg/L effluent fluid
	1–3 yr	90 µg
	4+ yr	240 µg
Vitamin B12	Limited loss–none		Limited loss–none
Vitamin C: based on SDI and filtration rate	<1 yr	50 mg	1.9 mg for each liter (1000 ml) of effluent
	1–3 yr	15 mg
	3+ yr	90 mg
Fat soluble vitamins (A, D, E, K)12,134	No supplementation needed		No supplementation needed

SDI, suggested dietary intake.

Figure 3.

Changes in vitamins and trace elements. This diagram illustrates the changes in vitamins and trace elements during critical illness, their cellular metabolic role, contributing factors for their deficiency, and clinical presentation. Adapted from ref. 138. ICU, intensive care unit; PN, parenteral nutrition; GI, gastrointestinal; RER, rough endoplasmic reticulum.

Trace Elements (Clinical Practice Points 35–37)

Trace elements, including copper (Cu), manganese (Mn), selenium (Se), chromium (Cr), and zinc (Zn), are required for a variety of physiologic functions.^{33,112,133,135,139–144} Data from adult studies on supplementation during CRRT are inconclusive, and an overall consensus does not yet exist on determining the loss and replacement.^{12,39,94,112,133,135,144,145} In pediatrics, Pasko et al. measured the levels of Cr, Cu, Mn, Se, and Zn in five children receiving TPN and a standard dose (2000 ml/1.73 m² per hour) of CVVHDF.³³ The extraction coefficient of trace elements was <0.1 in all patients suggesting that very little had crossed the CRRT filter membrane. Factoring for the parenteral supplementation, Pasko et al. concluded that the trace elements loss by CVVHDF is <20% of the corresponding parenteral supplementation.¹⁴⁶ The review by Jonckheer et al.^{12,39,94,112,133,135,144,145} In pediatrics, Pasko et al. measured the levels of Cr, Cu, Mn, Se, and Zn in five children receiving TPN and a standard dose (2000 ml/1.73 m² per hour) of CVVHDF.³³ The extraction coefficient of trace elements was <0.1 in all patients suggesting that very little had crossed the CRRT filter membrane. Factoring for the parenteral supplementation, Pasko et al. concluded that the trace elements loss by CVVHDF is <20% of the corresponding parenteral supplementation.¹⁴⁶ The review by Jonckheer et al.¹² suggested that the only significantly removed trace element is Se and subsequently needs weekly dosing (weak recommendations). Studies on trace elements in children on CRRT are presented in Table 1. Changes in trace elements during critical illness are illustrated in Figure 2.

Figure 2.

Evidence grading and strength of recommendations rubric. RCT, randomized control trial. Adapted from refs. 169 and 170.

Route and Timing of Nutritional Support (Clinical Practice Points 38–44)

In children with AKI treated with CRRT, EN is preferred to PN because of its simplicity, cost-effectiveness, fewer infectious complications, lower mortality, and trophic effects on the gastrointestinal tract.^5,147–149 Barriers to EN include delayed initiation, interruptions due to perceived intolerance, and prolonged fasting around procedures.¹⁵⁰ EN started within 24–48 hours of PICU admission is associated with better outcomes¹⁴⁸; therefore, daily evaluation is recommended. The oral route (breastfeeding in infancy) is the preferred method of EN.¹⁵¹ Other forms include gastric or postpyloric tube feeding. Nasogastric tube feeding is preferred in patients with swallowing difficulties, impaired consciousness, or under deep sedation. Postpyloric feeding may be used in patients who cannot tolerate gastric feeding or are at severe risk of aspiration. Interruptions in EN should be minimized to avoid inadequate and delayed nutrition and reliance on PN.¹⁵² The goal is to reach estimated energy and protein requirements in a stepwise approach¹⁵⁰ per institutional guidelines¹⁵³ and to reach 60%–70% of the energy requirements within the first week.¹⁵⁰ The choice of enteral formulas should be individualized.^154,155 Polymeric formula (regular or peptide-based) is the first choice for EN unless contraindicated.¹⁵⁶ In children requiring fluid restriction, consider protein and energy dense formulations^150,157,158 while monitoring for gastroparesis and osmotic diarrhea.¹⁵⁹ Renal formulas are not required during the acute phase and while on CRRT.^5,147–149 Barriers to EN include delayed initiation, interruptions due to perceived intolerance, and prolonged fasting around procedures.¹⁵⁰ EN started within 24–48 hours of PICU admission is associated with better outcomes¹⁴⁸; therefore, daily evaluation is recommended. The oral route (breastfeeding in infancy) is the preferred method of EN.¹⁵¹ Other forms include gastric or postpyloric tube feeding. Nasogastric tube feeding is preferred in patients with swallowing difficulties, impaired consciousness, or under deep sedation. Postpyloric feeding may be used in patients who cannot tolerate gastric feeding or are at severe risk of aspiration. Interruptions in EN should be minimized to avoid inadequate and delayed nutrition and reliance on PN.¹⁵² The goal is to reach estimated energy and protein requirements in a stepwise approach¹⁵⁰ per institutional guidelines¹⁵³ and to reach 60%–70% of the energy requirements within the first week.¹⁵⁰ The choice of enteral formulas should be individualized.^154,155 Polymeric formula (regular or peptide-based) is the first choice for EN unless contraindicated.¹⁵⁶ In children requiring fluid restriction, consider protein and energy dense formulations^150,157,158 while monitoring for gastroparesis and osmotic diarrhea.¹⁵⁹ Renal formulas are not required during the acute phase and while on CRRT.

Supplementing EN with PN is advantageous if there are GI complications due to EN or if nutritional needs are not met after 3–5 days, particularly in patients on CRRT, as they can have additional nutritional losses due to CRRT, impaired gastrointestinal absorption, and higher rates of malnutrition compared with other critically ill children.^{5,6,66,108,160–163} Supplemental PN can be delayed up to 1 week after PICU admission, provided micronutrients are delivered. Initiating PN in the first 24 hours of PICU admission is associated with worse outcomes, including more prolonged hospitalization and increased infection rates.^150,164 If unable to initiate or advance EN, PN should be started within 2–3 days for infants/toddlers and 3–5 days for children/adolescents. Children on PN met protein goals more often (35.8% time with isolated PN, 60.4% time on supplemental PN, and 12.7% time compared with EN alone).³⁷ Children with AKI were more frequently in receipt of PN before EN (56.6) versus non-AKI (17.5%).³⁶ In addition, shock, abnormal liver biochemistries, and mortality were significantly higher in children with AKI.

Putative Renoprotective Pharmaconutrients

Glutamine has been considered a putative renoprotective pharmaconutrient because of its potential beneficial effects, acting as a modulating factor and cellular substrate.^50,146 Glutamine supports immune cells, enterocytes, and gluconeogenesis; is protective against insulin resistance during stress; and increases the production of antioxidants.^111,165 As previously discussed, there are no recommendations regarding increased glutamine supplementation in critically ill children.⁷⁸

ω-3 FAs have also been considered to have potentially renoprotective pharmaconutrient effects, as these attenuate the body's response to stress, reduce oxidative stress, and positively influence the inflammatory response. These lipid mediators, resolvins (RV) and protectins (PD), which contribute to resolving the inflammatory response, are derived from ω-3 FA.¹⁶⁶ The addition of IV DHA and eicosapentaenoic acid, which are ω-3 polyunsaturated long-chain FA derived from fish oil, have a favorable effect on cell membranes and the inflammatory process. An in vivo study in mice has shown that administration of DHA in an ischemia–reperfusion injury model was associated with decreased polymorphonuclear leukocyte recruitment and with renoprotective responses by the upregulated expression of the renal cytoprotective heme-oxygenase-1 pathway and the high anti-inflammatory mediator PD D1 with overall decreased renal cytokine levels.^78,167 Although there are currently no clinical trial data to support their use in critically ill patients with AKI, the European Society for Clinical Nutrition and Metabolism currently recommends that parenteral lipid emulsions should contain ω-3 FA.¹⁶⁸

Clinical Practice Points

Overall, the panel of experts developed 44 clinical practice points that guide how to assess nutritional status in critically ill children with AKI requiring CRRT and manage nutrient needs in these patients (Table 2).

A multidisciplinary expert group created 44 clinical practice points to help assess and manage nutritional support in children with AKI requiring CRRT. Clinicians should integrate nutritional monitoring into the care of these critically ill pediatric patients, as malnutrition is commonly observed and is a significant prognostic factor for mortality. Unfortunately, there are limited pediatric data on the role of nutrition in critically ill patients with AKI during CRRT. This has made it challenging to establish evidence-based practice guidelines for these patients. However, given that establishing such guidelines would increase the likelihood of achieving nutrition goals in these patients, the current recommendations have been extrapolated from adult data and the existing guidelines for critically ill children. Literature areas with limited or without evidence were also identified. Further research on the nutritional effect of CRRT in critically ill pediatric patients with AKI will allow us to determine dietary needs more accurately and thus potentially reduce morbidity and mortality in critically ill children with AKI.

Disclosures

T.E. Bunchman reports the following: Advisory or Leadership Role: editorial boards of Pediatric Critical Care and Pediatric Nephrology. A. Davenport reports the following: Consultancy: Fresenius Medical Care; Honoraria: Fresenius Medical Company and Nipro Corporation; and Advisory or Leadership Role: Advisory board—Yaqrit; Advisory board—Nipro Corporation; Advisory board—WAK Scientific advisory board; CJASN; Leadership positions; European Dialysis and Transplant Association ERN committee; and President–International Society for Hemodialysis. S.Y. Irving reports the following: Research Funding: Sigma Theta Tau International, Xi Chapter (at University of Pennsylvania School of Nursing); Advisory or Leadership Role: National Board Member, American Society for Parenteral and Enteral Nutrition; and Other Interests or Relationships: RWJF Health Policy Research Scholar Program, Institution Mentor. A. Nada reports the following: Other Interests or Relationships: Member of the American Society of Pediatric Nephrology and Member of the Neonatal Kidney Collaborative. H.K. Yap reports the following: Other Interests or Relationships: National Kidney Foundation, Singapore. All remaining authors have nothing to disclose.

Funding

None.

Author Contributions

Conceptualization: Ann-Marie Brown, Andrew Davenport, Arwa Nada, Rupesh Raina, Weiwen V. Shih, Anvitha Soundararajan, Andrew Suchan.

Data curation: Anvitha Soundararajan, Andrew Suchan.

Investigation: Rupesh Raina.

Methodology: Timothy E. Bunchman, Sharon Y. Irving, Rupesh Raina, Sidharth K. Sethi.

Resources: Rupesh Raina.

Supervision: Rupesh Raina.

Writing – original draft: Khalid Alhasan, Ann-Marie Brown, Aylin S. Crugnale, Andrew Davenport, Sharon Y. Irving, Sai Sudha Mannemuddhu, Arwa Nada, Rupesh Raina, Sidharth K. Sethi, Weiwen V. Shih, Anvitha Soundararajan, Andrew Suchan, Victoria S. Vitale, Jakub Zieg.

Writing – review & editing: Khalid Alhasan, Katarina G. Berry, Ann-Marie Brown, Timothy E. Bunchman, Aylin S. Crugnale, Andrew Davenport, Isabella Guzzo, Sharon Y. Irving, Gerri L. Keller, Natalie H. Lussier, Sai Sudha Mannemuddhu, Archana Myneni, Arwa Nada, Rupesh Raina, Sidharth K. Sethi, Weiwen V. Shih, Anvitha Soundararajan, Andrew Suchan, Victoria S. Vitale, Hui Kim Yap, Jakub Zieg.

Supplemental Material

This article contains the following supplemental material online at http://links.lww.com/KN9/A424.

Supplemental Table 1. PICOS table. Specific PICOS questions. Search criteria.

Supplemental Table 2. Measuring caloric delivery from nonintentional calories in substitution fluid of CVVH in adults.

Supplemental Table 3. Formulae to determine caloric dosing.

Supplemental Table 4. Advantages and disadvantages of methods of measuring caloric dosing.

Supplemental Table 5. Protein losses on RRT: slow IHD, PD, CVVHD.

Supplemental Table 6. Author credentials.

Supplemental Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram.

Supplemental Figure 2. Delphi method.

Supplemental Figure 3. Calorimetry techniques.

Supplemental Figure 4. Bioenergetic balance in CVVH.

Supplemental Figure 5. Important equations for nitrogen balance measures.

Supplemental Figure 6. Changes in fat metabolism during critical illness.