There are direct nutritive and non-nutritive benefits of feeding patients with pediatric acute respiratory distress syndrome (pARDS). Adequate nutrition therapy in support of protein, energy, and micronutrient needs prevents loss of lean body mass, improves protein turnover for production of acute phase and immune proteins, prevents depletion of tissue antioxidant systems which occur with starvation, and is associated with improved 60-day mortality in mechanically ventilated, critically ill children (1–3). Adequate protein delivery prevents loss of respiratory and cardiac muscle function and is associated with increased ventilator-free days and improved mortality in pARDS (4). And, yet, median delivery of enteral nutrition (EN) remains 40–75% of goal over the first week of pediatric intensive care unit (PICU) hospitalization (3, 5–10). In addition, as protein synthesis rates are reliant upon adequate protein delivery, patients may not realize benefits of adjuvant therapies that rely on protein signaling if they fail to meet goal EN.

The gastrointestinal tract is a primary lymphoid organ, housing 70% of all immune cells with the ability to alter systemic inflammatory responses (11–14). Non-nutritive benefits of feeding, therefore, include downstream reduction in pro-inflammatory signaling to the lung (11, 15, 16). The gastrointestinal tract is an important target to improve lung inflammation, during pARDS, and subsequent patient outcomes. The intestinal epithelial barrier is a single-cell monolayer, which must absorb fluids and nutrients, interact with commensal organisms, and prevent entrance of pathogens and their toxic products (17). Intestinal barrier dysfunction has many downstream negative consequences. It is associated with bacterial translocation, endotoxemia, organ failure, immune dysfunction, and lung inflammation (13, 18, 19). The intestinal epithelial barrier, the host microbiome, and intestinal immune system interaction contribute to the pathophysiology of pARDS and present novel therapeutic targets.

Specific nutrients provided at pharmacological doses or nutrition supplemented with immune-modulating factors may directly impact lung pro-inflammatory cytokines and neutrophil accumulation in the setting of pARDS (20–22). Non-nutritive goals of feeding with either a standard enteral formula or an immune-modulating formula include maintenance of intestinal barrier functions, to alter or attenuate the immune or inflammatory responses in pARDS, and to modulate microbiome and/or host–microbe interactions. The host microbiome is emerging as an important mediator in chronic as well as acute inflammatory states. An understanding of the necessary and sufficient nutrition for each of these nutritive and non-nutritive benefits is important to determine the minimum necessary macronutrient and micronutrient delivery during each phase of pARDS to optimize patient outcomes. Optimizing nutritional therapy to delivery the right nutrition to the right patient at the right time is an excellent opportunity to improve outcomes in pARDS.

Currently, knowledge of the normal nutritional needs of children and of the metabolic response to critical illness guides recommendations and goals for nutritional support in pediatric patients (Figure 1). Critical illness and the associated inflammation and tissue injury alter metabolism by inducing a catabolic state, which may exacerbate pre-existing malnutrition (1, 23). Metabolism shifts away from growth to support the production of acute phase proteins, enzymes, and glucose to facilitate recovery (23, 24). This is accomplished through glycolysis, lipolysis, and protein turnover with resulting hyperglycemia, ketone production, and breakdown of skeletal muscle protein (23, 24). This diversion of resources results in cessation of normal growth and development in pediatric patients (23, 24). Prolonged persistence of this catabolic state progressively depletes the body’s nutritional resources, ultimately leading to muscle wasting, impaired cardiopulmonary function, decreased immune response, and poor wound healing (Figure 1). Pediatric patients (especially those with pre-existing nutritional, cardiopulmonary, or muscular disease) are particularly susceptible to these adverse sequelae due to their limited macronutrient and micronutrient reserves. In the setting of pARDS, this loss of lean body mass and diaphragmatic function is likely to amplify the severity of the patient’s respiratory failure and prolong ventilator dependence.

Patients who are undernourished prior to their critical illness experience worse outcomes (Table 1) (26). High quality research to understand how adequacy of nutrition delivery and premorbid nutritional status influences outcomes of pARDS is limited; however, pre-illness nutritional status and adequacy of macronutrient delivery early in critical illness are well established as risk factors for morbidity and mortality of critically ill patients in neonatal and pediatric populations (2, 3, 7, 10, 26–31). Malnutrition is both prevalent and incident in children hospitalized in the PICU; 30% of children admitted to the PICU have pre-existing acute or chronic malnutrition, and up to 58% of patients are discharged from the PICU malnourished (3, 6, 28). Even children who are well nourished prior to their critical illness may develop acute protein–energy malnutrition within 48 h of PICU admission due to limited protein and energy reserves in the setting of severe acute illness (6, 32–35). In a study of 1622 mechanically ventilated children, underweight nutritional status based on admission BMI z-score was associated with an increase in hospital-acquired infections, fewer ventilator-free days, and lower likelihood of hospital discharge (Hazard ratio 0.71, p < 0.001) (Table 1) (26). These high-risk patients may benefit from personalized nutritional strategies.

A growing percentage of children worldwide are now obese (39). While literature on the obese and critically ill pediatric patient is sparse, adult studies demonstrate a complex relationship between obesity and clinical outcomes in critical illness and acute respiratory distress syndrome (ARDS) (40, 41). While overall morbidity, including ventilator days and length of hospitalization, is higher, mortality was lower (40, 41). A mortality benefit may be due to a higher nutritional reserve in the setting of prolonged catabolism (40, 41). Similar to risks in underweight children, obesity based on admission BMI z-score is associated with higher risk of hospital-acquired infections and lower likelihood of hospital discharge (hazard ratio 0.82, p = 0.04) (26). It is important to note, however, that obesity is not exclusive of malnutrition and that the obese child should receive the same consideration for nutritional supplementation (40). Both obese and underweight children are at increased risk compared with normal weight children for adverse outcomes while undergoing mechanical ventilation (26). For this reason, a nutritional support strategy should include early screening and diagnosis of malnutrition (42).

Studies in adults and animals suggest negative impact of excessive nutrition with difficulty with ventilator weaning and either a mortality benefit or non-inferiority with restricted caloric intake (Table 2) (73–79). Several retrospective studies in critically ill children identify overfeeding occurring in PICU patients with negative clinical consequences (53, 80). Overfeeding is associated with delayed ventilator weaning, lipogenesis, hepatic dysfunction, hyperglycemia, increased mortality, and prolonged hospitalization. Complications associated with underfeeding included delayed ventilator weaning, impaired protein synthesis, organ failure, and an increased risk of sepsis (7, 10, 24, 32). Several potential mechanisms exist for the negative impact of overfeeding on patient outcome, such as increase in carbon dioxide production, increased intolerance of EN and parenteral nutrition (PN), refeeding syndrome, azotemia and metabolic acidosis from excess protein administration, hepatic steatosis from excess glucose delivery, hyperglycemia, hypertriglyceridemia, and on the cellular level the suppression of autophagy (53, 62, 81, 82). Administration of PN early in the course of critical illness may suppress autophagy, a necessary form of programmed cell death, which aids in removal of damaged proteins and mitochondria and is thought to play a role in recovery after organ failure (83).

American Society of Parenteral and Enteral Nutrition (ASPEN) Guidelines of the Provision and Assessment of Nutrition Support Therapy in the Adult Critically Ill Patient recommend that patients with ALI/ARDS receive either trophic or full EN over the first week of hospitalization. Even in adults, there is uncertainty regarding timing and dose of nutrition therapy. Evidence exists that hypocaloric nutritional strategies, experienced by the majority of patients, may not benefit specific subsets of the critically ill, such as in the setting of mechanical ventilation, severe burns, and surgery (111–113). Patients at high nutritional risk, defined as either undernourished or overnourished at baseline may benefit from a more aggressive nutritional strategy (114). The intensive nutrition in acute lung injury (INTACT) study was designed to deliver >75% of goal energy and protein needs daily to adult ICU patients via EN from date of diagnosis of ARDS to hospital discharge as compared with standard nutrition support (Table 2) (90). The INTACT study found increased mortality (40 versus 16%, p = 0.02) with aggressive nutritional support (85% of estimated energy expenditure) when compared with standard therapy (55% of estimated energy expenditure) and was stopped early (90). While impossible to apply directly to pediatric patients due to their different metabolic needs and ongoing growth and neurocognitive development, these data reinforce the powerful role nutrition exerts on patient outcome and highlight the need for rigorous research to identify the optimal timing and dose of calories and protein, tailored to narrowly defined patient populations.

Individualized nutritional therapies that are “just right” and meet, but do not exceed, needs are currently recommended by ASPEN guidelines for nutritional prescription in critically ill children (50). “Optimal” energy and protein targets remain elusive, and response to nutrition delivery is likely related to baseline nutritional risk and current diagnosis. Increased utilization of non-invasive ventilation (NIV) in pARDS is associated with worse nutritional adequacy, and nutrition-related outcomes for patients managed with NIV are unknown (115). We do not know if potential benefits to NIV are outweighed by risks related to underfeeding. Underfeeding is harmful, but overfeeding is equally harmful. Multiple clinical trials are criticized for comparing very underfed to underfed patients, or for overfeeding patients. At the heart of these controversies lies the difficultly in determining in real-time calorie and protein needs and response to nutritional intervention. Advances in body composition assessments that are practical during pARDS in ultrasound, bioelectrical impedance, CT imaging, and functional outcomes assessments to monitor lean body mass close monitoring of lean body mass during pARDS a possibility, but require further study (116–119).

There are many research opportunities to improve nutritional supplementation for critically ill patients in general as well as specifically within the framework of pARDS. Emerging research continues to demonstrate the role of gut dysfunction in the development of ARDS (13, 151). Deitch proposed the “gut lymph” hypothesis, whereby, the injured gut allows translocation of bacteria and bacterial products and liberates cytokines and chemokines, which act via the lymphatic duct to activate alveolar macrophages and contribute to pathogenesis of ARDS (152). Clark and Coopersmith proposed the “intestinal crosstalk” theory whereby the intestinal microbiota, immune system, and intestinal epithelial barrier interact and, when dysregulated, worsen systemic inflammation (153–155). Loss of balance in this “crosstalk” may lead to ARDS and is supported by emerging animal and human data examining the microbiome in critical illness. Modulation of intestinal barrier function and microbiome remain novel targets to improve outcomes in pARDS.

Several minimally invasive plasma biomarkers show promise in guiding initiation, titration, and continuation of EN by objectively assessing intestinal epithelial structure and function as well as assessing for the translocation of bacteria and bacterial products (Table 3) (156–158). The intestinal epithelial barrier is a single-cell monolayer tasked with competing roles to absorb nutrients, interact with commensal organisms, and limit entrance of pathogens and their products (Figure 2) (159, 160). It is the largest surface in the body in contact with the external environment. Intestinal barrier dysfunction is associated with the pathogenesis of multiple organ dysfunction, translocation of bacteria and endotoxin, and loss of remote organ immune function (19, 159, 161). Enteral feeding of at least 15% of goal calories restores intestinal barrier function in murine models of critical illness (69). The minimum dose of EN required to maintain intestinal barrier function in children with pARDS is unknown. Biomarkers to monitor intestinal barrier function could guide provider decisions regarding the minimum necessary dose of enteral feeding. In a study of 20 children after cardiopulmonary bypass, biomarker concentrations were associated with symptoms of feeding intolerance (140). Biomarker-based decision rules for initiation, advancement, and maintenance of EN represent an opportunity to reduce practice variation and improve success in achieving and maintaining nutritional goals. The development of biomarkers to accurately and easily measure the catabolic and anabolic balance of the metabolic system is another area of ongoing research which will assist clinicians to more accurately develop nutritional prescriptions (46).

Technological advances in computing power and analytic techniques over the last decade make the study of the human microbiome possible (173). In the human body, bacterial cells exceed human cells in number and perform many essential functions, such as production of short chain fatty acids and vitamins (vitamin K), which aid the human host (174). The developmental maturation of the human microbiome is poorly understood, but it reaches adult patterns by 2–4 years of age (175). Dysbiosis occurs when the symbiotic relationship between human and microbial populations is disturbed and may result in reduced microbiome diversity. Reduced microbiome diversity during infancy, a critical time period during microbiome development, may have lasting consequences on the development of several chronic diseases (176). Murine models where antibiotics are provided during infancy, demonstrate rapid and permanent alteration of metabolic phenotype via transient reduction in intestinal microbiome diversity (176–178). In contrast, an increased microbiome diversity in human infants enhances maturation of the intestinal mucosal immune system, which can influence LPS responsiveness at 1 year of age in humans (179). Thus, normal metabolic and immune imprinting by essential “keystone” microbes, which can be altered by early antibiotic exposures, may have long-term consequences on immune response to critical illness and risk of chronic illness. This emerging field has the potential to shift our targets with regard to control of both acute and chronic inflammation from modulation of human cells to modulation of the bacterial within us. The gastrointestinal microbiota is implicated in control of inflammation in the lung (180). Diet is a potent determinant of intestinal microbiome diversity and, in murine models, trumps genetic background (181). Carmody et al., demonstrated, in a murine model, that rapid changes in diet resulted in rapid shifts in microbial composition (181). As diet is an essential determinant of microbiome diversity and alters lung inflammation, dietary manipulation of the gastrointestinal microbiome may be a new target for treatment of pARDS. We do not know how medical diets alter the intestinal microbiome during pARDS, nor do we understand the short or long-term consequences of nutritional, probiotic, or pre-biotic interventions.

Immunonutrition is well tolerated and results in improved pro-inflammatory cytokine profiles, but is not associated with clear clinical outcome benefits (182). While small single-center studies of isolated pharmaconutrients showed promise, multiple combination nutritional and pharmacologic additives in critically ill adults aimed at modulating the inflammatory and immune response have not shown benefits (Table 2) (183, 184). Specific supplements that have been explored include several antioxidants, arginine, glutamine, metoclopramide, ω-3 fatty acids, zinc, and selenium (103, 185–190). The randomized comparative effectiveness pediatric critical illness stress-induced immune suppression (CRISIS) prevention trial examined the effect in critically ill children of daily enteral zinc, selenium, glutamine, and IV metoclopramide on incidence of nosocomial infections. No difference was seen between the two groups, and the trial was stopped early due to futility (103). Multiple combination nutritional and pharmacologic additives in critically ill adults aimed at modulating the inflammatory and immune response have been trialed, but have generally not shown any benefits (50, 191). The optimal formulation of lipids is an area of active research with continued focus on the role of ω-3 polyunsaturated fatty acids (PUFAs).

The use of ω-3 PUFAs as a nutritional adjuvant have evidence suggesting they may reduce the pro-inflammatory state in ARDS, but a recent systematic review and meta-analysis looking specifically at ARDS patients in the adult ICU found no benefit in clinical outcomes (107, 192–197). The rationale for ω-3 PUFAs is that they may be beneficial in ARDS as they would compete with ω-6 PUFAs, decrease the synthesis of pro-inflammatory eicosanoids, increase production of anti-inflammatory lipid mediators, such as resolvins and protectins, decrease in chemotaxis, decrease reactive oxygen species (ROS) and pro-inflammatory cytokines, and decrease leukocyte binding and activation through decreased expression of adhesion molecules (Figure 3) (192).

Jacobs et al. evaluated the feasibility of ω-3 PUFAs (eicosapentaenoic acid + γ-linolenic acid) in children with ALI and ARDS and examined the impact of ω-3 PUFAs on plasma phospholipid fatty acid concentrations (22). They found that delivery of nutrition enriched with eicosapentaenoic acid + γ-linolenic acid was feasible and resulted in an anti-inflammatory fatty acid profile (22). This study is supportive of future work in this area in children. Whether a change in anti-inflammatory fatty acid profile will improve clinical outcomes is unclear, but promising. Methodological challenges in several adult trials limit conclusions regarding ω-3 PUFAs in adult patients with ARDS (192). Specifically, studies are heterogeneous, and, in some, the control diet was pro-inflammatory.

Several ω-3 PUFA-derived mediators are potential targets to resolve the inflammatory response in the setting of pARDS (Figure 3). Alveolar edema and neutrophil recruitment and activation are early events in ARDS. Resolvins, protectins, and maresins are lipid-derived mediators, and several resolvins are emerging as potential therapeutic targets for resolution of ARDS (Figure 3) (20, 21). These resolvins are derived from the ω-3 PUFAs docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA). Eickmeier et al. found that aspirin-triggered resolvin D1 reduces mucosal inflammation and promotes resolution in a murine model of acute lung injury (21). Aspirin-triggered RV D1 decreased lung inflammation and promoted acute lung injury resolution shortly after injury in an hydrochloric acid model of ARDS by enhancing restitution of barrier integrity, decreasing circulating neutrophil–platelet heterotopic interactions, and regulating inflammatory mediators and nuclear factor-κB (NF-κB) activation (21). Seki et al. found that resolvin E1 protected mice from bacterial pneumonia and acute lung injury with pretreatment by decreased lung neutrophil accumulation, enhanced microbial clearance, decrease in lung pro-inflammatory mediators, and improved survival (20).

Recent evidence suggests an association between vitamin D levels and the risk of ARDS (199–201). Vitamin D deficiency is associated with impaired pulmonary function and increased incidence of viral and bacterial infections and inflammatory disease, including asthma and COPD (202–204). Mechanisms for these associations are unknown, but vitamin D plays a role in macrophage, lymphocyte, and epithelial cell function, critical to ARDS pathophysiology (200, 205).

There are currently limited data to guide nutritional strategies in the critically ill pediatric patient and an absence of data targeted specifically at pARDS (Tables 2 and 3). Current strategies rely on the application of adult data, knowledge of the metabolic derangements caused by critical illness, and knowledge of the nutritional requirements of healthy children. The generally accepted strategy centers on the early identification of need for nutritional supplementation and early initiation of EN targeted toward fulfillment of nutritional goals and implemented by a collaborative, multidisciplinary treatment team. Indirect calorimetry is a recommended adjunct to ensure adequate, but not excessive, nutritional support is supplied. Early prescription of a bowel regimen to limit constipation while avoiding diarrhea is also recommended (50). PN is currently reserved for patients in whom EN is contraindicated or limited, though its routine use as a supplement to EN is an area ripe for research. Before we evaluate the role of immunonutrition to improve outcomes of pARDS, fundamental questions regarding protein and energy requirements during pARDS remain: if delivery should be based on premorbid nutritional risk and if under, normo, or even overfeeding might be beneficial at different stages of illness. Given the mortality benefit of nutritional adequacy on pARDS, it may prove challenging to identify benefits of adjuvant nutritional therapies, unless they are overlaid upon a foundation of adequate macronutrient delivery. Once fundamental questions regarding optimal timing, dose, and route of macronutrient delivery are answered for pARDS, investigations centered on immunonutrition to further improve patient outcomes are appropriate. Systematic investigations to determine the impact of each pharmaconutrient separately and in combination during pARDS are necessary.

In the future, intensivists will likely employ distinct and highly personalized nutritional therapies based on patient premorbid nutritional risk, admission diagnosis, severity of illness, measured energy expenditure, active monitoring of lean body mass, and the composition of patient intestinal microbiome. A select group of high-risk patients are likely to experience benefit from an intensive and personalized, titrated nutritional plan. Long-stay patients, patients with ARDS, sepsis, burns, or pre-existing severe malnutrition are most likely to have improved outcomes if nutrition is closely monitored. Intensive nutritional therapies will likely be titrated daily to meet energy and protein needs. Nutritional therapies, monitoring, and impact on meaningful patient outcomes are likely to change dramatically in the next decade as we understand the role of the microbiome in regulating both intestinal health and also the inflammatory response to pARDS. We may require nutritional or “metabolic” teams to precisely tailor metabolic support to meet, but not exceed, energy, protein, and micronutrient needs, preserve lean body mass, the microbiome, the virome, and intestinal barrier functions. Nutrition is fundamental to provision of critical care services, and, with an understanding of the complex role it plays in illness and health, we will be able to design effective clinical trials of nutritional interventions as primary therapies for pARDS.

Dr. BW and Dr. KT contributed equally to the conception, review of literature, writing, and editing of this manuscript. Both authors approve of the final submitted version of this manuscript.

Dr. KT has received an investigator initiated grant from the Baxter Corporation in support of research to perform a clinical trial of early versus later PN in mechanically ventilated children. Funds provided by Baxter do not support her effort on the project, but provide project funds. BW declares no conflict of interest.