Reduction of Neonatal Parenteral Nutrition Associated Cholestasis Through Lipid Emulsions

BACKGROUND. Parenteral nutrition associated cholestasis (PNAC) is a common complication of prolonged and exclusive parenteral nutrition (PN). PNAC in neonates and infants is multifactorial, including the underlying pathology and the effect of certain PN nutrients. In preterm infants, hepatic immaturity is itself, a predisposing factor. Infants subjected to major surgery are often unable to receive enteral nutrition for a long period of time, during which they require exclusive PN. After that, enteral nutrition is slowly introduced alongside with the reduction of the PN. In major surgery for congenital malformations of the digestive tract, additional risk factors for PNAC are the absence of enteral nutrition, intestinal bacterial translocation and sepsis. Jejunal atresia and gastroschisis are independent risk factors for PNAC. In short bowel syndrome, changes in the bile acids enterohepatic cycle may also contribute to PNAC.

Intravenous lipid emulsions (ILE) used in PN may promote PNAC or protect against it depending on their composition. Phytosterols contained in ILE have been implicated in PNAC in newborns by disrupting bile-acid homeostasis. High intakes of soy-based fatty acids (FA) n-6 from ILE, especially palmitate, may contribute to PNAC since these are precursors of arachidonic acid, a pro-inflammatory mediator.

Medium chain triglycerides (MCT) may have a hepatic protective effect. Hence, ILE containing relatively high amounts of MCT, such as Lipofundin® (B. Braun) theoretically might be advantageous in protecting against PNAC. Long chain triglycerides (LCT) of n-3 family may protect from PNAC thorough its anti-inflammatory activity. The Omegaven® (Fresenius Kabi), exclusively based on LCT n-3 has proved to prevent and reverse PNAC in neonates.

In several new-generation emulsions, including SMOFlipid® (Fresenius Kabi), the reported α-tocopherol content is up to 4- to 5-fold higher than the alpha-tocopherol content of soy-oil emulsions. The α-tocopherol isoform acts as an antioxidant, preventing lipid peroxidation attributable to the high content of long-chain polyunsaturated fatty acid (LC-PUFA).

The ILE Lipofundin® (B Braun), is composed of 50% LCT (soybean oil) and 50% MCT (coconut oil). The other new generation ILE SMOFlipid® (Fresenius Kabi) is composed of 30% LCT n-6 (soybean oil), 30% MCT (coconut oil), 25% monounsaturated fatty acids (olive oil), 15% LCT n-3 (fish oil) and α-tocopherol.

A systematic review found lower bilirubin levels in children with intestinal failure and other conditions receiving parenteral n-3 ILE compared with n-6 ILE. However, it was concluded that current data is insufficient to support the use of parenteral n-3 ILE in children, suggesting further trials examining long-term clinical outcomes and harms.

Among several studies comparing the effect of different ILEs in PNAC in children, only the retrospective study by Pischler et al. (2014) compared SMOFlipid® with Lipofundin®. This study included 127 children aged 0-16 years, including 34 premature infants and 59 children with surgical conditions, including necrotizing enterocolitis.

Until 2011 only Lipofundin® had been used for PN in the the medical-surgical neonatal intensive care unit (NICU) of the Hospital Dona Estefânia. The further availability of SMOFlipid® led the investigators to compare the effect of both ILE on the liver tests associated with PNAC of neonates subjected to corrective surgery for major congenital abnormalities, since to the best of our knowledge no prospective study has made this comparison specifically in this population.

OBJECTIVE. To compare the incidence and severity of PNAC in term and near-term infants subjected to corrective surgery for congenital abnormalities and needing prolonged PN using the ILE SMOFlipid® or Lipofundin®. The investigators hypothesise that the use of SMOFlipid® may be more protective from PNAC than Lipofundin.

METHODS. Design: single-center, randomized, controlled and double-blinded trial: prescribing physicians were unaware of the type of ILE administered and the pharmacist who prepared and randomized the individuals to the interventions was not aware of the liver status of the participants. Simple randomization was performed by the same pharmacist (MLR) using a computer generated random number table.

Were considered eligible every consecutive neonate admitted in the NICU, with gestational age of 34 weeks or over, requiring corrective surgery of congenital anomaly of the digestive tract or indirectly affecting the digestive tract (eg, diaphragmatic hernia). Recruitment occurred in the first 48 hours after birth, if PN with ILE was initiated within the first 48 hours after birth.

Main variables recorded:

• Weekly measurement of serum: total bilirubin, conjugated bilirubin, GGT, alanine -aminotransferase (ALT) and aspartate aminotransferase (AST), alkaline phosphatase and triglycerides.
• Daily parenteral lipid intake (g/Kg)
• Reasons for reducing or stopping ILE

Secondary variables recorded:

• Gestational age, birth weight, prenatal diagnosis of congenital abnormalities
• Main and secondary diagnoses
• Major surgery (date/ day of life)
• Events, especially infectious events (date/ day of life)
• Enteral feeding initiation (date/ day of life)
• Enteral feeding: type of feeding and mode of administration. The percent of enteral intake in relation to daily fluid intake was recorded (date/ day of life): 0-50% and full enteral feeding.
• Daily weight (g)
• Weekly length (cm) and head circumference (cm)
• Weekly or every 2 weeks measurement of serum: total blood count, ionogram, calcium, phosphorus and magnesium

Potential confounders affecting the liver function:

• Sepsis according to described criteria
• Phenobarbital for treatment abstinence syndrome secondary to sedative and analgesic drugs used during the postsurgical period
• Use of ursodeoxycholic acid

Parenteral nutrition protocol based on the National Consensus for Neonatal PN . Whenever possible PN with ILE is initiated within the first 24 hours after birth. As the Pharmacy Service of the Hospital is not available for preparing individualized PN during weekends, infants admitted during this period have initiated a standard solution containing only glucose, calcium and aminoacids. Thereafter, all patients receive a similar aminoacid, glucose, electrolyte and vitamins PN solution plus ILE (SMOFlipid® or Lipofundin®).

ILE is reduced to 0.5-1.5g/kg/d if:

Hypertriglyceridemia (> 250 mg / dL) Hyperglycemia (> 150 mg / dL) Unconjugated bilirubin > 12 mg/dL Acute phase of sepsis Pulmonary hypertension If cholestasis appeared, ILE is restricted to 2-2.5 g/kg/d, the amino acids restricted to 2-2.5 g/kg/d and glucose limited to 12 mg/kg/minute Enteral nutrition protocol is the same in both groups. Minimal enteral feeding is initiated when bowel sounds are audible, and significant abdominal distention and bilious or bloody gastric residuals are absent. Initially, feeds are administered continuously, and changed to bolus feeding as soon as infants can tolerate it. Mother's milk is preferred. However, depending on the patient's condition semi-elemental (Pepti-Junior®, Danone) or elemental formula (Neocate®, Nutricia) may be preferred. Later on, these formulas are replaced with mother's milk as soon as tolerated, or if mother's milk is insufficient or unavailable, preterm formula (Miltina Prem®, Humana GmbH, Germany) may be used in preterms or infant formula (Nan 1®, Nestlé) in full-term infants.

Data collection and storage: Excel® calculation table (Microsoft Office 2007®). Statistical analysis with the support of the Research Unit of Centro Hospitalar de Lisboa Central. Categorical data were presented as frequencies (percentages), and continuous variables as mean and standard deviation (SD) or median and inter-quartile range (25th percentile-75th percentile), as appropriate.

Mixed effects regression models were used to take into account the correlation structure between measures in time. Crude and adjusted odds-ratios with corresponding 95% confidence intervals were calculated.

The level of significance was α = 0.05. Data analysis was performed using the software SPSS 22.0 (SPSS for Windows, Rel. 22.0.1. 2013. SPSS Inc., Chicago, Il, EUA) and Stata (StataCorp. 2013. Stata Statistical Software: Release 13. College Station, TX: StataCorp LP.).

Measures of outcome: Cholestasis and cholestasis severity incidence rates and associated 95% confidence intervals will be accessed for each intervention group.

Cholestasis and severe cholestasis incidence rates will be compared by calculating relative risks and the Number Needed to Harm (NNH) in association with their 95% confidence intervals.

Relative efficacy measures (hazard ratios) and possibly odds ratios will be used if group homogeneity is found.

Identification of confounders using multivariate analysis in logistic regression.