Clinical aspects

Description

Parenteral nutrition implies the administration of a balanced, all-inclusive mixture of nutrients including carbohydrates (glucose), protein (amino acid), lipids, and macro- and micro-minerals. Total parenteral nutrition implies the absence of any other source of nutrition, for example by the enteral route. Partial parenteral nutrition refers to the instance where parenteral nutrition provides a supplement to enteral nutrition.

Carbohydrates. Glucose is an essential source of energy and the most commonly used carbohydrate in total parenteral nutrition. Carbohydrates should provide approximately 35% of daily calorie needs and approximately 60% of nonprotein energy (13). Hyperglycemia and hypoglycemia are a common complications in preterm infants, but the management of blood glucose remains extremely controversial. Although it is clear that low blood glucoses are detrimental in the short and medium term, the exact threshold for harm (or even if there is a threshold) is unknown (109). Conversely, there are multiple theoretical concerns about high blood glucose levels. Once again, however, the threshold for harm is undefined. Some authors have suggested that glucose more than 150 mg/dL should be considered as hyperglycemia (97), but this would likely expose many preterm infants to insulin treatment (and thus the risk of hypoglycemia) for uncertain benefit. Blood glucose levels greater than 360 mg/dL may cause significant osmolar changes resulting in dehydration and electrolyte derangements, and they are statistically associated with an increased risk of cerebral damage, necrotizing enterocolitis, retinopathy of prematurity, and intraventricular hemorrhage (41; 48; 03); however, whether they are simply associations, or whether they are causes or effects of hyperglycemia, is not clear.

In the Swedish EXPRESS study of infants born at least 27 weeks gestational age, the incidence of hyperglycemia (> 180 mg/dL) was approximately 30%, but a 1 g/kg/d increase in glucose infusion rate was only associated with a 1.6% increase in the rate of hypoglycemia (119). There are concerns about the effects of the main tools to reduce hyperglycemia including insulin administration and reduction in glucose intake (and the resulting reduced calorie intake and possible effects on early growth) (86). Use of insulin infusions to prevent or reduce hyperglycemia has been shown to result in modest reduction in hyperglycemia but at the cost of an increased frequency and severity of hypoglycemia (16). In another study, very low birth weight infants who were defined as hyperglycemic (2 measurements more than 4 hours apart > 153 mg/dl) to target a blood sugar between 72 and 108 mg/dL or between 144 and 180 mg/dL showed little differences in outcomes (04). Tighter glucose control was associated with more rapid weight and head circumference gains, but with reduced linear growth and an increased rate of hypoglycemia (04). Pragmatically, the safest approach in treating hyperglycemia is to lower glucose infusion rate rather than add insulin due to the risks associated with it (62).

Guidelines from ESPGHAN suggest that parenteral glucose delivery to preterm infants should begin at 2 to 4 mg/kg/min (2.9 to 5.8 g/kg/d) and increase to 8 to 10 mg/kg/min (11.5 to 14.4 g/kg/d), although many, including the current authors and the National Institute for Clinical Evidence in the United Kingdom, would consider these amounts too low, especially the starting intake where a value of 6 to 9 g/kg/d would seem more appropriate (74). In term infants, ESPGHAN suggests starting at 2.5 to 5.0 mg/kg/min (3.6 to 7.2 g/kg/d) and increase to 5 to 10 mg/kg/min (7.2 to 14.4 g/kg/d) (63). In infants aged greater than 28 days, infusion rates would be lower and would steadily increase from the acute phase of the illness to the stable phase, and then to the recovery phase; they would be higher (as g/kg/d) in smaller infants (63).

In the PICU, blood glucose levels above 145 mg/dL should be avoided (63). ESPGHAN also recommends that blood glucoses above 145 mg/dL be avoided in the NICU, and recurrent levels above 180 mg/dL should be treated with insulin if reductions in glucose infusion have been inadequate to control blood glucose (63). Once again, some, including the current authors, would consider these treatment levels too aggressive.

Unfortunately, despite the importance of blood glucose control, the real definition of what plasma glucose is too low or too high is unknown and may be an inadequate question (perhaps being affected by clinical state, metabolic demands, developmental stage, and the availability of other substrates). Similarly, the balance of risks and benefits of insulin treatment of hyperglycemia in preterm neonates remains unclear (08; 04).

Protein. The major goal in providing amino acids after birth is to prevent negative nitrogen balance, reduce catabolism, and promote protein accretion. Although there is no direct method for measuring protein needs, studies have shown that endogenous loss of protein (in urine, feces, skin cells, and secretions) is about 1 g/kg per day if no supplemental protein is given (29). Estimated daily protein accretion of a fetus at about 28 weeks of gestation is 2 g/kg/day; at least 3 to 3.5 g/kg/day of amino acids is needed to promote protein accretion while allowing for obligatory losses (120). Protein provides 4 kcal/gram and should supply 12% to 17% of total daily calories. In premature infants, the amino acid should start on the first postnatal day with at least 1.5 g/kg/d to achieve an anabolic state; from day 2 onwards, it should be between 2.5 to 3.5 g/kg/d and should be accompanied by nonprotein intakes of more than 65 kcal/kg/d (108).

Studies support that providing 2 to 3.5 g/kg per day of amino acid shortly after birth is well tolerated and improves protein accretion (101; 46; 99; 43). There are now more data to support even higher intakes of up to 4 g/kg/day during the first week (24; 68). Traditionally, infusing amino acids of 4 g/kg/day by day 3 of life in very premature infants has been associated with high urea and ammonia concentrations (10). Contradictory data on developmental outcomes have been reported, Bayley Mental Developmental Index scores were lower with the high amino acid intake, at 12 months, but improved by 24 months (11). Weight, height, and head circumference were lower than in infants with standard amino acid intakes. Most studies have shown no difference in the pH or base deficit between groups taking different AA doses (the initial dose of 0 to 3 g/kg/day with a target range of 2.4 to 4 g/kg/day) (83; 99; 47; 10; 17; 104).

Administering a high dose (> 3 g/kg/day) or an early dose (< 24hrs) of parenteral amino acid is safe and well tolerated, but does not offer significant benefits with regard to growth (57). ESPGHAN guidelines for preterm infants recommend that parenteral amino acid intakes above 3.5 g/kg/d should only be administered as a part of clinical trials (55). In contrast, evidence-based guidelines from the United Kingdom suggest intakes up to 4 g/kg/d are reasonable after the first few days of life in preterm infants (74).

Special attention should be paid to calcium and phosphorus levels in preterm infants during first week of life receiving high amino acid infusion. A study by Bonsante and colleagues highlights the influence of early AA intake on calcium and phosphorus homeostasis (15). Preterm infants (gestational age less than or equal to 33 weeks) were divided into 3 groups according to their mean AA intake during the first week of life. The incidences of hypophosphatemia and hypercalcemia, respectively, were increased in the high (greater than 2 g/kg/day of mean AA intake) AA intake group relative to the moderate (1.5 to 2 g/kg/day) and low (less than 1.5 g/kg/day) intake groups (16).

Emphasis should be placed on providing optimal protein and energy during the first week of life in preterm infants (94), and transitioning to enteral feeds as quickly as safely possible. In extremely low birth weight infants (less than 1000 g), increased intake of protein and energy in the first week was associated with higher Mental Developmental Index Scores (94) and lower likelihood of length growth restriction. In fact, early administration of amino acids improves the weight of preterm infants (106), decreases postnatal growth restriction (24), and improves postnatal head growth (68). Studies have found improved long-term outcomes, such as growth and neurodevelopment. Poindexter and colleagues found significant improvement in growth parameters (weight, length, and head circumference) at 36 weeks postmenstrual age in the infants who received early amino acids, but no difference was found in growth or in neurodevelopmental outcome at 18 months of age (81). Van den Akker and colleagues found no difference in growth, but found a neurodevelopmental advantage at 2 years corrected for boys who received amino acids from the first day of life compared to the infants who received glucose alone (107). Stephens and colleagues reported a retrospective analysis of 150 extremely low birth weight infants and found a positive association between protein intake in the first week of life and scores on the Bayley Mental Developmental Index at 18 months corrected age.

Osborn and colleagues studied higher versus lower amino acid intake on growth and disability-free survival for newborn infants. Higher amino acid intake (2 to 3.5 gm/kg/day) was associated with positive nitrogen balance, earlier regain of birth weight, and increased head circumference growth at hospital discharge. Higher amino acid intake was not associated with effects on days to full enteral feeds, late-onset sepsis, necrotizing enterocolitis, chronic lung disease or severe intraventricular hemorrhage, and mortality before hospital discharge (76). However, a randomized controlled trial by Balakrishnan and colleagues found that high-dose amino acid supplementation (3 to 4 g/kg/d) starting at birth was not associated with improved growth or neurodevelopmental outcomes (07).

Despite the increase in protein nutrition, parenteral nutrition-dependent very premature infants are still at risk for deficiency for the conditionally essential amino acids, particularly tyrosine, glutamine, cysteine, and arginine (100). Even with supplementation to traditional amino acid solutions using a standardized, concentrated, added macronutrients parenteral (SCAMP) nutrition regimen, very premature infants did not achieve conditionally essential amino acid levels that reached reference ranges (67). A Cochrane review of glutamine supplementation, however, did not demonstrate any effect on mortality or major neonatal morbidities including incidence of sepsis, necrotizing enterocolitis, or neurodevelopmental outcomes (65).

Lipids and essential fatty acids. Lipids are iso-osmolar and provide more calories per gram than do protein and carbohydrates. Lipids should provide 25% to 40% of nonprotein parenteral nutrition calories. Lipid emulsion contains triglycerides (soybean oil or safflower oil), egg yolk phospholipid (for emulsification), and glycerol to achieve isotonicity. It provides energy and essential fatty acids, both of which are important for growth and neurodevelopment. Essential fatty acids (linoleic [omega-6] and alpha linolenic acid [omega-3]) are precursors for the synthesis of long chain polyunsaturated fatty acids (PUFA), especially (EPA and DHA), which play a structural role in biological membranes and are involved in retinal and brain development.

In utero, lipids are obtained from the maternal diet and transferred via the placenta to the fetus in the form of PUFA. Of the PUFA delivered, the placenta transfers docosahexaenoic acid (DHA) more so than other fatty acids (linoleic acid, linolenic acid, and arachidonic acid). DHA constitutes 50% of neuronal plasma membranes and is the only n-3 PUFA found in significant amount in the retina and brain.

After delivery, lipids should be started at a rate of 0.5 to 1 g/kg/day to avoid essential fatty acid deficiency (42; 51). Recommendations suggest starting 1 g/kg/day on the first day, incrementally advancing to a goal of 3 to 4 g/kg/day at the end of the first week (53), provided the triglyceride level remains less than 175 to 200 mg/dL (06). There is no evidence that gradual increments in the rate of infusion of lipids improves fat tolerance; however, higher amounts of lipids (2 to 3 g/kg/d) on the first day of life may result in hyperlipidemia (114). In preterm and term infants, parenteral lipid intake should not exceed 4 g/kg/d (55). If the triglyceride level is elevated, lipids need to be decreased according to the serum level of the triglycerides; however, it should be continued at lowest possible rate (0.25 g/kg/d) to prevent essential fatty acid deficiency.

There is evidence that early introduction of parenteral lipids during the first week is safe and well tolerated in very low birth weight infants (113; 55). It is also associated with weight gain and could improve early nutritional support for these preterm neonates (38).

Soybean emulsions contain a high ratio of omega-6 to omega-3 fatty acids, which, in turn, leads to lesser proportions of omega-3 fatty acid derivatives docosahexaenoic acid and eicosapentaenoic acid. Both docosahexaenoic acid and eicosapentaenoic acid are important substances of neural and retinal development with anti-inflammatory properties (93). Soybean oil–based lipid emulsions are rich in proinflammatory w-6 long-chain polyunsaturated fatty acids and phytosterols; both are known to trigger parenteral nutrition–associated liver disease (88). For parenteral nutrition lasting more than 5 days, pure soy emulsions should no longer be used, and composite emulsions with or without fish oil should be the first choice. These composite lipid emulsions contain higher amounts of vitamin E and less phytosterols (103).

Lipids can be infused separately from amino acid/dextrose solution or combined in a 3-in-1 mixture (fats/proteins/dextrose in the same solution). According to a study by Colomb and colleagues, a 3-in-1 total nutrition formula was more manageable and easier to administer (23). It was also shown to be more cost-effective and requires less nursing time than conventional peripheral parenteral nutrition. Despite these benefits, a separate lipid delivery system is preferred due to better tolerance and the ability to cycle lipids off and on without interrupting the amino acid/dextrose infusion. Additionally, the Centers for Disease Control has recommended against the use of the 3-in-1 formula due to issues of precipitation of components, especially in neonates.

In addition to their nutritional role, lipid emulsions can influence numerous pathophysiological processes including oxidative stress, immune response, and inflammation. The benefits of fish oil come from omega-3 PUFA concentrations, which are rich in docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) and have specific anti-inflammatory effects. They are also devoid of phytosterols, which are thought to contribute to parenteral nutrition associated liver disease (PNALD).

In preterm infants receiving soybean oil derived lipid emulsion, DHA and arachidonic acid (ARA) levels (derived from EPA) decline rapidly after birth. Low levels of these fatty acids are associated with morbidity, namely chronic lung disease and late onset sepsis (61; 89). Studies using fish oil lipid emulsions (rich in n-3 PUFA) have suggested that increasing the amount of DHA through the use of fish oil is well tolerated, can decrease the risks of late onset sepsis, hyperbilirubinemia, and cholestasis, and can improve growth in very low birth weight neonates (87; 114). Infants who received lipid minimization with soybean oil emulsions, but not those who received lipid minimization with fish oil emulsions, are at high risk of biochemical essential fatty acid deficiency and slower weight gain (34). Evidence also demonstrates reversal of established hepatic dysfunction (54; 80), but there was no prevention (37). Currently there are 3 lipid emulsion products containing fish oil: Omegaven® (Fresenius Kabi, Germany), with 10% fish oil; Lipoplus®/Lipidem® (B. Braun, Germany), containing 50% medium chain triglycerides (MCT), 40% soybean oil, and 10% fish oil; and SMOFlipid® (Fresenius Kabi, Germany), containing a mixture of 30% soybean oil, 30% MCT, 25% olive oil, and 15% fish oil (52). Newer SMOF lipid mixtures are well tolerated by premature infants and effective in optimizing fatty acid profiles, reducing liver injury and improving direct bilirubin levels in case of parenteral nutrition associated liver disease (87; 80). In a 2014 Australian study comparing SMOFlipid to olive oil lipid emulsion [(ClinOleic® 80% olive oil and 20% soybean oil) (Baxter, Deerfield, IL)], SMOFlipid was well tolerated and showed beneficial effects in terms of reduction of oxidative stress by reducing lipid peroxidation levels (30). Many newborn intensive care units have started replacing soy-based lipid formulations with Smoflipid^® (w-3 enriched lipid emulsion) as the primary component in parenteral nutrition for preterm infants (21).

However, a Cochrane review by Kapoor and colleagues found no benefit of mixed lipid emulsions (LE) (MCT-olive-fish-soy oil-lipid emulsions; MCT-fish-soy oil-lipid emulsions; olive-soy oil-lipid emulsions; and borage-soy oil-lipid emulsions) over pure soy oil based lipid emulsions in terms of death, growth, bronchopulmonary dysplasia (BPD), sepsis, retinopathy of prematurity greater than stage 3, and parenteral nutrition associated liver disease (49).

Macronutrients, vitamins, and micronutrients. Creating parenteral nutrition solutions that mimic enteral nutrition is difficult due to solubility issues, incompatibility issues, photo-degradation of components, and the lack of knowledge of definite levels of nutrients to meet the demands of patients. Essential for complete nutrition are macro minerals (electrolytes), vitamins, and micro minerals (trace elements).

Two parenteral vitamin solutions are available for use in infants in the United States: MVI Pediatrics® (AstraZeneca Pharmaceuticals) and INFUVITE® Pediatric (Boucherville, Quebec, Canada). Both contain water and fat-soluble vitamins. Of particular importance in patients with intestinal failure and those on parenteral nutrition are the fat-soluble vitamins A, D, E, and K (56). The micronutrients such as zinc, copper, selenium, and iodine should be added to parenteral nutrition in very low birth weight infants within a few days after birth. Parenteral supplementation of iron, manganese, chromium, and molybdenum is rarely needed in very low birth weight infants but should be considered in patients with intestinal failure that requires prolonged total or near total parenteral nutrition. Parenteral iron requirements are estimated to be 200 to 250 mcg/kg/d in preterm infants and 50 to 100 mcg/kg/d in term infants. Iron supplementation should be preferably given via enteral route (33).

Trace elements are necessary in many aspects of cell function: enzyme activity, protein and lipid metabolism, immune and inflammatory healing modulation, thyroid and other endocrine functions, and development of the central nervous system.

Table 1. Suggested Parenteral Intakes of Trace Minerals in Infants and Children

Element	Preterm Mcg/kg per day	Term Mcg/kg per day	Children Mcg/kg per day (maximum, Mcg/kg per day)
Zinc* Copper** Selenium*** Chromium*** Manganese** Molybdenum*** Iodide****	400 20 2 0.20 1 0.25 1	250 < 3 mo 20 2 0.20 1 0.25 1	50(5000) 100 ≥ 3 mo 20(300) 2(30) 0.20(5.0) 1(50) 0.25(5.0) 1(1.0)
*Standard trace element preparations do not supply the high requirement of premature infants, so it should be added to parenteral fluid. **Omit in patients with obstructive jaundice, as manganese and copper are excreted primarily in bile. Copper status should be monitored in patients with cholestasis if not provided in total parenteral nutrition. ***Omit in patients with renal dysfunction. Available concentrations of molybdenum and manganese are such that dilution of the manufacture’s product may be necessary. Neotrace® (Lymphomed Co, Rosemont, IL) contains a higher ratio of manganese to zinc than suggested in this table (ie, zinc: 1.5 mg, and manganese: 25 mcg in each mL). Adapted from (51)

Table 2. Daily Electrolyte and Mineral Requirement for Pediatric Patients

Electrolyte	Preterm neonates	Infants/children
Sodium Potassium Calcium Phosphorus Magnesium Acetate and chloride	2-5 mEq/kg 2-4 mEq/kg 2-4 mEq/kg 1-2 mEq/kg 0.3-0.5 mEq/kg As needed to maintain acid-based balance	2-5 mEq/kg 2-4 mEq/kg 0.5-4 mEq/kg 0.5-2 mEq/kg 0.3-0.5 mEq/kg As needed to maintain acid-based balance
Adapted from (64)

In children receiving parenteral nutrition, especially preterm infants, supplementation of electrolytes should be adjusted according to their serum values.

The goal is always to start enteral nutrition, even small trophic feeds or “minimal enteral feeds,” as soon as the gastrointestinal system will allow.

To date, nearly all studies have shown that minimal enteral feeding approaches promote the capacity to feed enterally (42). According to a Cochrane review, infants given trophic feeds had a reduction in days to full feeding, in days that feedings were held, and in length of hospital stay (105). Enteral nutrition has been shown to increase exocrine pancreatic secretion in response to gastrointestinal hormones. Because parenteral nutrition bypasses the gastrointestinal system and is continuous rather than off/on, as in bolus enteral feeding, hormone excretion mimics elevated postprandial levels for extended periods of time. Continuously elevated levels of hormones in patients on prolonged total parenteral nutrition have been suspected to reset the hormone “set point” and lead to BMI increasing year after year in these children (102). Enteral nutrition has also been shown to increase gastric parietal cell mass in patients compared with parenteral nutrition, and a lack of enteral nutrition has been shown to decrease gut associated lymphatic tissue, therefore, increasing the risk for inflammatory disease states (44).

Indications

Parenteral nutrition should be provided as soon as possible after the infant’s birth if enteral nutrition is not possible or will be inadequate. Providing early parenteral nutrition especially in premature infants has shown to improve nitrogen balance and extra uterine growth restriction. Although enteral nutrition is the preferred route, parenteral nutrition is of paramount importance in infants with gastrointestinal disorders such as intestinal atresia, abdominal wall defects (omphalocele, gastroschisis), necrotizing enterocolitis, and short bowel syndrome, and parenteral nutrition is important in the immediate newborn period when enteral nutrition is not feasible because of unstable medical condition.

Contraindications

Parenteral nutrition is contraindicated in patients with functioning bowel who can tolerate enteral nutrition at a level to provide adequate energy and other nutrients for growth.

Outcomes

A coordinated nutritional support team is a key factor to optimize parenteral nutrition, maintain accuracy of the parenteral nutrition delivered, decrease adverse effects, and improve outcome to minimize hospital stay and cost (51). This specialized team includes physicians, dietitians, nursing and ancillary staff, pharmacists, and surgical staff. Strategies to reduce errors, improve quality of care, and provide optimal nutrition have proven to be effective. These processes include offering nutritional education and training for the support team, maintaining an effective parenteral nutrition order form or program, increasing collaboration among team members, creating and adhering to parenteral nutrition guidelines, and having parenteral nutrition rounds to regularly review parenteral nutrition orders. These approaches resulted in improved compliance with safe practice standards, improved percentage of patients with appropriate indications for total parenteral nutrition, adequate glycemic control, increased number of patients receiving parenteral nutrition within 10% of caloric need, and decreased time to full enteral feeds (additional cost savings) (12).

The goal of parenteral nutrition is to supply the optimal energy and nutrients that the patient would be achieving via the enteral route. Benefits of parenteral nutrition is more apparent in premature neonates, especially extremely low birthweight infants and very low birthweight infants. Previous approaches to formulate nutrition plans referenced the growth of similar-aged neonates to each other. Most premature infants experience extra-uterine growth restriction during their hospital course due to the inability to provide optimal nutrients and energy that matches in utero accretion. New strategies are being designed to mimic the growth occurring in the fetus of the same gestational age. Growth restriction is a significant problem as numerous studies have shown definitively that undernutrition, especially of protein, at critical stages of development produces long-term short stature, organ growth failure, neuronal deficits of number and dendritic connections, and later adverse behavioral and cognitive outcomes (42).

Adverse effects

Table 3. Complications of Total Parenteral Nutrition

Acute		Chronic
Metabolic complications		Systemic complications
	• Hypoglycemia • Hyperglycemia • Metabolic acidosis • Hypophosphatemia and other electrolytes imbalance • Hyperlipidemia		• Parenteral nutrition–associated liver disease • Metabolic bone disease
Mechanical complications		Infectious complications
	• Extravasation and tissue necrosis • Infiltration • Thrombosis • Pleural or pericardial effusion • Cardiac arrhythmia from malposition of catheter		• Bacterial infections especially staphylococcal species • Fungal infections: Candida species, Malassezia furfur
(79)

Good evidence-based guidelines are available for the management of parenteral nutrition complications, specifically those related to the need for a central venous line (117).

Glucose intolerance. Hyperglycemia is more common than hypoglycemia in infants receiving total parenteral nutrition. It is more pronounced in extremely low birth weight infants; although glucose infusion rate has a statistically significant but clinically small effect (119). Early provision of well-balanced total parenteral nutrition with amino acid improves glucose tolerance (02). Studies have shown that the early introduction of parenteral amino acids has been shown to decrease the risk of hyperglycemia by stimulating endogenous insulin production (60). Hyperglycemia is often seen in sick patients, resulting from metabolic stress or a surge of counter regulatory hormones such as cortisol and norepinephrine. Steroids, infection, and respiratory distress can increase metabolic demand and create a catabolic state in which higher energy intake is needed, frequently leading to hyperglycemia. Hypoglycemia is not common while on parenteral nutrition, although hypoglycemia is recognized as a risk factor for morbidity, numerous studies have also shown that hyperglycemia in patients on parenteral nutrition correlates with increased hospital complications, as well as higher morbidity and mortality (59; 78).

Parenteral nutrition-associated liver disease (PNALD). Among the complications associated with long-term use of parenteral nutrition, the incidence of parenteral nutrition-associated liver disease (PNALD) ranges from 40% to 85% in infants. Histologic cholestatic changes in the liver can be observed within 2 weeks, and fibrosis is detected within 6 weeks of commencing parenteral nutrition. The use of parenteral nutrition in the neonatal intensive care unit is common with nearly 70% of patients in a 2007 study receiving parenteral nutrition at some point during hospitalization and 21% of patients receiving parenteral nutrition for greater than 21 days or more (22). PNALD may be diagnosed initially by abnormalities in laboratory findings, including increased liver enzymes and/or direct bilirubin concentrations. PNALD is defined as a concentration of direct bilirubin greater than 2 mg/dL on 2 consecutive measurements along with elevations in transaminases not associated with other known causes of cholestasis. Enteral feeding remains the best strategy to reverse and prevent parenteral nutrition-associated liver disease, with even 10% of caloric intake has shown beneficial effects (25). Patients whose direct bilirubin was 10 mg/dL or more had a nearly 40% chance of death or a liver transplant prior to modern practices of lipid reduction strategies and use of fish oil lipid emulsions (118).

The exact etiology of PNALD is not clear, but multiple risk factors have been identified: prematurity, low birth weight, small for gestational age, prolonged course of parenteral nutrition, intestinal starvation, catheter-related sepsis, photo-oxidation of parenteral nutrients, especially amino acids (09; 77), and the presence of surgical issues that preclude enteral nutrition for varied periods of time.

Studies have demonstrated that n-6 PUFA (represents > 60% of fatty acids in soy-based lipid emulsions) may favor the development of PNALD and promote increased pro-inflammatory mediators. In addition, soybean oil emulsions contain high levels of plant-based phytosterols, which correlate with severity of cholestasis (72). Because discontinuing lipid emulsions in neonates and children dependent on parenteral nutrition can result in essential fatty acid deficiency, poor growth, and poor neurodevelopmental outcome, new lipid emulsions have been developed.

Studies have shown that the dose and composition of intravenous fatty acid emulsions may play an important role in the development and progression of PNALD (19). In patients who develop PNALD while on soy or safflower oil emulsions, changing to fish oil emulsion reverses their PNALD (28). Fish oil-based emulsions, with a higher content in n-3 PUFA, have successfully been used in infants with PNALD as treatment and as nutritional support (39; 85). Multiple studies show fish oil-based emulsions are superior to mixed lipid emulsions in treating PNALD (70; 90), and mixed lipid emulsions seem to be useful in preventing PNALD associated with soybean oil lipid emulsions (80). Smoflipid^® has been shown to be safe and is associated with increased levels of docosahexaenoic acid and eicosapentaenoic acid, with conversely lower arachidonic acid levels in both erythrocyte membranes and plasma of premature infants as compared to pure soy emulsions. Smoflipid^® has also been found to have a positive effect in lowering total bilirubin, with the fish oil component associated with the reversal of PNALD in some infants (50; 70; 80; 31). Fish oil emulsion is prescribed at 1 g/kg/d. Studies have shown that when 1 g/kg/d of exclusive fish oil is substituted for soybean-based lipids, direct hyperbilirubinemia is more likely to resolve and incidence of death may be reduced (39; 18; 84).

Restriction of lipid emulsions may prevent the development of PNALD. Cycling parenteral nutrition with an “off” period each day, as opposed to continuous parenteral nutrition, can reduce the incidence of PNALD without associated calcium, phosphorus, magnesium, or vitamin D losses (95). However, cycling parenteral nutrition in infants younger than 6 months has not been well studied and is not recommended. The risk of PNALD can also be reduced by beginning enteral nutrition as soon as possible and by avoiding overfeeding. Decreasing fat emulsion infusion to 1 g/kg/day, based on the evidence that n-6 PUFA and phytosterols are hepatotoxic and pro-inflammatory (72), has proven not to be beneficial (73; 58). However, infants receiving lipid minimization at 1 gram/kg/day in contrast to 3 gram/kg/day had serum direct bilirubin levels rise at a slower rate than controls (19). More studies evaluating the role of lipid restriction are needed. Enteral feeding remains the best strategy to reverse and prevent PNALD, with as little as 10% of caloric intake showing beneficial effects (25; 77).

Metabolic bone disease. Metabolic bone disease, or osteopenia, is defined as a reduction in bone mineral content (osteopenia). It is a multifactorial disorder mainly seen in very low birth weight infants (premature infants less than 32 weeks’ gestation) from lack of fetal mineralization during the last trimester. The incidence is greater in extremely low birth weight infants. The incidence is inversely proportional to birth weight and gestational age of the infant and seen in 16% to 40% of very low birth weight infants and extremely low birth weight infants. The prevalence is higher in breast fed premature infants (40%) compared to preterm formula fed infants (16%) (98; 01). The clinical signs of metabolic bone disease in premature infants appear between 5 and 11 weeks of life and are characterized by an increased work of breathing due to chest wall instability caused by softening of ribs, an enlargement of the cranial sutures, rickets, fractures, and postnatal growth failure (35).

Osteopenia is more likely seen in premature patients not receiving enteral nutrition and most often manifests as hypophosphatemic metabolic bone disease progressing to rickets. Therefore, it is recommended to supply a calcium content of 75 to 90 mg/kg/day, a phosphorus content of 60 to 67 mg/kg/day, and a magnesium content of 7.5 to 10.5 mg/kg/day. This should correspond to a calcium to phosphorus ratio of 1.3:1 by weight and 1:1 by molar ratio in the parenteral solution. Traditionally, parenteral nutrition cannot provide the necessary amounts of calcium and phosphorus due to precipitation in the parenteral nutrition solutions; however, this is no longer a concern in countries that have organic phosphate preparations available (27).

It has been shown that early aggressive parenteral nutrition can impact calcium and phosphorus homeostasis, leading to hypophosphatemia and hypercalcemia. This result is most apparent with higher amino acid intake, namely greater than 2 g/kg/day (15). Because of this, in addition to a preterm infant’s inherent risk of osteopenia, the need for appropriate calcium to phosphorus ratio to combat hypophosphatemia is critical.

There are no specific methods for diagnosing metabolic bone disease. Often serum markers are used, including calcium, phosphate, alkaline phosphatase, parathyroid hormone, and vitamin D. Alkaline phosphatase levels higher than 900 IU/L in preterm infants younger than 33 weeks’ gestational age, associated with serum phosphate levels persistently lower than 5.6 mg/dL (< 1.8 mmol/L), have a diagnostic sensitivity and specificity of 70% and 100%, respectively, for metabolic bone disease of prematurity (35). Commonly used imaging study is the x-ray of long bones and ribs. Osteopenia may be diagnosed on x-ray based on evidence of reduced bone mineral density. Dual energy x-ray absorptiometry is the gold standard technique to assess bone mineral density. Backström and colleagues found that the association of alkaline phosphatase serum levels more than 900 IU/L and phosphate less than 1.8 mmol/L indicates a low bone mineral density with a sensitivity and specificity of 100% and 70%, respectively, compared to dual-energy x-ray absorptiometry measurements in very low birth weight and extremely low birth weight infants younger than 33 weeks’ gestation and a mean birth weight of 1490 g (05).

Although there are no standard recommendations on screening for metabolic bone disease in premature infants, high index of suspicion with abnormal labs warrants further investigations and treatment (40). Hypophosphatemic rickets in these infants respond well to increased phosphate supplement in total parenteral calories and minerals, especially calcium and phosphorus. This form of osteopathy does not respond to vitamin D nutrition or as a dietary supplement. Premature infants receiving breast milk should have appropriate fortification to enhance therapy unless a deficiency is present. Also, ensure appropriate vitamin D store by providing 400 IU of vitamin D. The AAP recommends that all breast-fed, partially breast-fed and non-breast-fed infants consuming less than 1000 IU of vitamin D fortified milk daily should be supplemented daily with a minimum of 400 IU vitamin D (115).

Infectious. Catheter-related blood stream infection is (CLABSI) potentially life threatening and occurs in 2% to 20% of patients receiving parenteral nutrition with highest rate being in extreme premature infant population. CLABSI is defined as a positive blood culture not related to another site of infection in a symptomatic patient with a central venous line in place at the time of or removed within 2 days prior to the onset of infection (20). In the United States, the CLABSI rate in intensive care units is estimated to be 0.8 per 1000 central line days. These complications can be prevented by aseptic line insertion and careful maintenance, such as sterile changing of infusion solutions, minimizing access to the line, and removing the catheters when it is no longer necessary. Adherence to insertion and maintenance bundles and use of checklists have been demonstrated to decrease catheter-related blood stream infection rates in NICUs in multiple studies (91). The Centers for Disease Control and Prevention recommends development and implementation of “bundles” to improve compliance and outcomes (75). Studies have demonstrated that implementing strategies that involve line site care, training for staff and parents, multidisciplinary discharge planning, and monitoring compliance reduce the rates of catheter-related blood stream infections (71). However, strict adherence to unit specific policies for prevention of central line associated blood stream infection is of paramount importance, as high compliance with a checklist for insertion and daily line necessity is significantly associated with lower catheter-related blood stream rate. Long catheter stay is one of the risk factors for CLABSI. The incidence rate of CLABSI increased by 14% per day during the first 18 days after peripherally inserted central catheter (92). A central catheter should be removed as soon as possible once it is no longer necessary.

Current observations have indicated that use of mixed lipid emulsions have been associated with up to 25% decreased incidence of sepsis in very low birth weight infants compared to those receiving soybean lipid emulsions. More randomized controlled trial studies are still needed (113).

In older children, especially those needing prolonged parenteral nutrition and prolonged central venous access, newer modalities that may be helpful in the prevention or treatment of CLABSI include the use of ethanol locks and taurolidine citrate locks (117).

Others. Other complications of parenteral nutrition include hyperlipidemia, hypercalcemia, and electrolyte abnormalities. The risks of these can be reduced by careful prescribing and close monitoring. There is also risk of inadvertent contamination of the parenteral nutrition solutions with exogenous toxic compounds, for example, aluminum (82) and manganese (36; 32). Parenteral nutrition is a common source of medical errors (26), which may be reduced through the use of a limited number of standardized parenteral nutrition formulations (14).

Special considerations

The disease process for which the patient is on parenteral nutrition, the coexisting disease processes, or adverse events of parenteral nutrition may increase the risk of morbidity or mortality. Now that infants are surviving longer on parenteral nutrition, new aspects must be explored concerning optimizing components, maximizing appropriate growth, mediating immune and inflammatory responses, and improving neurodevelopmental outcomes.

Scientific basis

	• Despite being a common cause of medical error, objective evidence in favor of parenteral nutrition (rather than dextrose/ electrolyte solutions) from RCTs is limited.
	• Parenteral nutrition may be required in preterm infants to approximate the in utero accretion rates of nutrients as enteral feeds are advanced, and there is a strong physiological argument in favor of parenteral nutrition, especially its establishment/advancement of enteral feeds is slow or delayed.
	• It is possible that as enteral feeds are advanced more rapidly in preterm infants, the relative benefits of parenteral nutrition may be reduced, especially in larger infants in whom full enteral feeds can be established much more quickly.

More than 750,000 orders for parenteral nutrition are made in U.S. NICU annually, and parenteral nutrition remains a common cause of medical errors (26). In preterm infants, the earlier administration of parenteral nutrition or the addition of intravenous amino acids to dextrose/electrolyte solutions (typically within the first day, rather than at the end of the first week) has been shown to provide short-term growth benefits (69; 96) and improve early nitrogen balance (104). In one observational population-based study, it may also increase survival, albeit at the cost of increased morbidity; however, such studies are subject to the effect of unidentified confounding variables (116).

The evidence-base is less clear for term neonates, especially those requiring relatively brief periods of parenteral nutrition. In one study of term-born infants admitted to the PICU (a subgroup analysis of the PEPaNIC study), later introduction of parenteral nutrition (towards the end of the first week of life) compared to earlier introduction was associated with benefits in terms of reduced infections and earlier PICU discharge (111), but the quality of evidence in this population is generally poor (112; 66).

The main PEPaNIC study compared the late introduction of parenteral nutrition (towards the end of the first week of life) to earlier introduction of parenteral nutrition in infants and children admitted to the PICU (110), including those much older than the previously reported subgroup analysis of term infants (111). Although early weight loss was associated with worse outcomes, this did not differ between the "early" and "late" parenteral nutrition groups (110). Later introduction of parenteral nutrition was associated with significant cost savings, which exceeded the amount expected simply from the nonprescription of parenteral nutrition (111; 110). At the 2-year follow up, there were no differences in survival, growth, neurodevelopmental outcome, or health-related quality of life, but some developmental subdomains (such as executive functioning and working memory) were improved in the group with delayed introduction of parenteral nutrition (112; 45). The subgroup of infants who were undernourished at admission did worse than well-nourished infants, but even in this group delayed introduction of parenteral nutrition had no effect on early weight loss or mortality, but there were, however, significantly fewer line infections and earlier discharge from the PICU (111; 110).