Premature neonates have only limited muscle and fat mass and thus decreased hydrolytic capacity of the enzyme lipoprotein lipase.23
As a consequence, they are at higher risk for PN-associated hypertriglyceridemia than term infants.1,2
Tight monitoring of serum triglyceride levels has been recommended in premature neonates to avoid hypertriglyceridemia with the provision of increasing lipid concentrations as recommended by actual PN guidelines.2
Therefore, our primary hypothesis was that, in premature neonates, test treatment would not be inferior to a standard soybean oil emulsion with regard to the serum concentration of triglycerides.
Throughout the study, mean serum triglyceride levels remained below the upper limit of normal range (1.6 mmol/L or 142 mg/dL) in both treatment groups for all time points (baseline, day 5, day 8, post day 9, and last visit). For the present trial, the benchmark of clinical relevance was set at 3.43 mmol/L (304 mg/dL). In the meantime, recommendations of ESPEN/ESPGHAN have become available that consider a concentration of serum triglycerides exceeding 2.82 mmol/L (250 mg/dL) a critical limit calling for a reduction in parenteral lipid dose.2
As a result of the lipid infusion, there was a slight increase in triglycerides from baseline to day 8 in both treatment groups, which was significant in the test group (P
= .03) and tended to be significant in the control group (P
= .07) but was not different between groups (P
= .15). Yet, statistical proof of the noninferiority of the study treatment could not be given in a confirmatory sense. Possibly, the fact that infants in the test group were slightly more immature than control infants might have influenced the tolerance to parenteral lipids, explaining the failure to give confirmatory proof of the equality of treatments administered. Yet, with regard to the critical limits, the slight and comparable increase in triglyceride concentrations as observed with both study treatments (<0.2 mmol/L) may be rated of no clinical relevance, confirming the safety of both treatments administered. Triglyceride concentrations observed in the present trial are also in line with those reported by Tomsits et al24
in a similar study population, supporting their conclusion that a physical mixture of soybean oil, MCTs, olive oil, and fish oil is safe and well tolerated in premature infants requiring PN.
The significant increase in LDL cholesterol in the test group as compared to the controls is not well explained and is in contrast with the results obtained in the study by Tomsits et al.24
In this study, no differences in HDL and LDL cholesterol were observed. In the present study, the values of the LDL/HDL ratio are comparable with breastfed preterm infants.25
For the hematological parameters, changes observed after study treatment in both groups might be ascribed both to the routine blood sampling and to the physiological anemia frequently experienced by preterm babies.26,27
Clinical laboratory parameters showed no clinically relevant changes during the main study phase. Predominantly, values reflect the common postnatal development in this patient population (eg, maturation of kidney function). Consequently, other safety aspects, such as hematological and clinical laboratory parameters, as well as the adverse event profile and evaluation of vital signs indicate that both lipid emulsions were equally safe and well tolerated over 7–14 days.
Premature neonates with low birth weights, receiving PN for longer periods, are at an enhanced risk to develop cholestasis and PN-associated liver disease (PNALD).1,28-31
Monitoring of total and, in particular, direct bilirubin levels during parenteral feeding is thus crucial. In the present trial, mean values for direct bilirubin remained below the critical level of 17.1 µmol/L and <20% of total bilirubin, respectively, before and after study treatment.31
Total bilirubin levels decreased significantly in the neonates receiving the test mixed emulsion after 7–14 days of parenteral feeding but not in controls receiving a soybean oil–based emulsion. Furthermore, although a slight decrease in direct bilirubin could be seen in the test group over the treatment period, this parameter showed a significant increase in controls. Short-term application of the new lipid emulsion was thus associated with beneficial effects on serum bilirubin levels in premature neonates, thereby potentially reducing their risk of developing cholestasis.
The concept that a change in parenteral lipid regimen, from a predominance of n-6 fatty acids to lipid emulsions containing n-3 fatty acids from fish oil, may be effective in the prevention and treatment of PNALD is still under discussion.17,32-34
Beneficial effects on liver function with a mixed emulsion containing soybean oil, MCT, olive oil, and fish oil were observed in a few studies in adult and pediatric patients: in adult intensive care unit patients after major surgery, a lower rise in liver enzymes and in the phospholipids/plasma apolipoprotein A1 ratio (a surrogate marker of liver function) suggested better liver function by PN with the test emulsion than with a soybean oil emulsion.35
In postoperative surgical patients, ALT, aspartate aminotransferase, and α-glutathione S-transferase levels (markers of hepatic integrity) were also significantly lower with the test emulsion as compared to a lipid emulsion based on olive and soybean oils, indicating a better liver tolerability.36
In a study in children receiving long-term PN with the test emulsion, a significant improvement in plasma bilirubin could be observed.37
The present findings must be interpreted with caution, though, regarding possible benefits of the test lipid emulsion on the development of PNALD in premature infants. First, the observed reduction in total bilirubin, to some extent, reflects the common postnatal development in this patient population. Furthermore, patients need to be on PN for a longer time before negative effects on liver function become apparent, whereas most of the neonates in the present trial did not require PN beyond the third week of life. Further trials evaluating the potential hepatoprotective effects of the new lipid emulsion over prolonged periods on a larger number of patients are needed.
Soybean oil–based lipid emulsions contain high amounts of γ-tocopherol but relatively low amounts of α-tocopherol, the main lipophilic antioxidant. In some older studies, infusion of soybean- or safflower–based lipid emulsions has been associated with an increased production of peroxidative intermediates and, therefore, an aggravated risk of oxidative stress.38,39
Oxidative stress, representing a common mediator of the inflammatory process, has been associated with hepatocellular injury in preterm infants on PN.40
In previous studies evaluating the test emulsion in preterm infants and children, parameters of lipid peroxidation did not differ between treatment groups, but vitamin E status and total antioxidant potential were significantly improved with the provision of adequate amounts of vitamin E as compared to controls receiving a soybean oil emulsion.24,37,41
Unfortunately, measurements of vitamin E status and lipid peroxidation were not within the scope of the present investigation. However, with regard to the previous finding, it was assumed also in the present trial that the provision of increased amounts of α-tocopherol with the test emulsion may have contributed to protecting the liver against PN-induced peroxidative damage. Eventually, the partial replacement of soybean oil with fish oil in the new lipid emulsion resulted in a lower phytosterol intake in the infants receiving the test emulsion. It has repeatedly been suggested that phytosterols may represent a further contributing factor to the pathogenesis of PN-related cholestasis.30,42-44
Phytosterol concentrations have been shown to be particularly high in plant oil–based lipid emulsions, whereas fish oil emulsions are free from phytosterols.35
In our study, fatty acid profiles were to be measured in plasma and in RBC phospholipids pre- and posttreatment. Unfortunately, reliable baseline values for RBC fatty acid composition are not available because of difficulties in sample handling that occurred during the isolation of RBCs from the blood samples drawn at baseline. Therefore, only posttreatment results are presented for RBC phospholipid fatty acid composition.
In both plasma and RBCs, values measured for the essential fatty acids LA and α-LNA posttreatment were significantly higher in the group receiving the soybean oil emulsion than in the infants receiving the test emulsion. These findings reflect the different fatty acid composition of the investigational lipid emulsions, thereby confirming compliance to the treatments administered.
Also, the long-chain homologues of α-LNA (ie, EPA and DHA) were significantly increased with the mixed fish oil containing test emulsion as compared to controls’ posttreatment values. Obviously, the increased provision of EPA and DHA with the fish oil component of the tested emulsion more than outweighed the decreased supply of the precursor fatty acid in both compartments. The increased values for EPA and DHA observed in the present study are in line with the findings by Goulet et al37
in pediatric patients on long-term PN showing significantly increased relative amounts of both EPA and DHA in plasma and RBC phospholipids after 28 days. Although Tomsits et al24
also recently reported an increase in RBC-EPA in premature neonates receiving the test emulsion, they did not observe a corresponding increase in RBC-DHA. The reason for this finding could be that the preterm neonates in that study received up to a maximum of 2 g lipids/kg BW/d. In a recent study of preterm infants <1250 g receiving a lipid emulsion containing 10% fish oil, a decrease of DHA in RBC was noticed.45
However, this was less significantly decreased compared to the group receiving a soybean oil–based emulsion.
The significance of these findings has to be related to the intrauterine and postnatal accretion of LC-PUFAs in preterm and term infants. During the third trimester of fetal life, a rapid accretion of DHA, EPA, and AA takes place.46,47
Because of the relatively inefficient endogenous conversion of α-LNA to DHA in the newborn, there seems to be an apparent need for the supplementation of preformed LC-PUFAs in preterm infants.5
It has been calculated that the need for supplementation of DHA in preterm infants is even higher than actually provided by the supplemented preterm formulas.48
It has been suggested that enteral feeding supplemented with DHA is beneficial for the neurodevelopmental outcome of preterm infants.49
Furthermore, it could be hypothesized that measurements of RBC phospholipids allow predictions of this accretion and its potential beneficial effect on visual acuity and mental and psychomotor development in later infancy.50
Indeed, an adequate intake of dietary DHA and AA seems to be required for optimal functional maturation of the retina and visual cortex.11,51-53
It is also generally accepted that LC-PUFAs, particularly the n-3 LC-PUFA from fish oil, possess potent immunomodulatory and anti-inflammatory properties.54
A shift from a predominance of n-6 fatty acids to n-3 LC-PUFAs from fish oil in PN may be effective in modulating the eicosanoid profile toward a reduction in circulating levels of proinflammatory mediators, thereby potentially contributing to decreasing the risk of hepatic injury.30
In the present trial, C-reactive protein levels showed an increase over the main study period in both groups, which was distinctly yet not significantly more pronounced in the infants receiving the soybean oil emulsion.
In plasma and RBC phospholipids, posttreatment values for the long-chain homologue of LA, arachidonic acid (AA, C20:4 n-6), were comparable with both lipid emulsions, indicating that small amounts of AA (0.5%) and LA (18.7%) contained in the test emulsion efficiently compensated for the much higher amount of LA (53%) contained in the soybean oil emulsion. AA has been shown to affect growth and development in early infancy in a positive way.11
Furthermore, the supply of AA in preterm infants has been associated with the weight at birth and growth during the first year of life.55
In the present trial, both body weight and height were significantly increased vs baseline in both treatment groups, independent of lipid source, suggesting that PN with SMOFlipid, providing n-3 fatty acids from fish oil, was equally effective in promoting growth in premature neonates as a conventional soybean lipid emulsion. This finding might be of particular interest, taking into account that 2 trials published in the 1990s have raised concerns that the use of a fish oil–enriched enteral formula might reduce the growth of preterm infants.56,57
It has been shown, however, that the decrease in weight gain was due to a deficiency in AA in the fish oil–supplemented formula and that optimal weight gain requires a well-balanced DHA/AA. This was confirmed in a recent Cochrane Database Systematic Review showing that formula supplemented with DHA and AA, with or without EPA, did not impair growth.58
Consequently, recommendations have been made about the supply of LC-PUFAs with infant formula.59,60
Fish oil–based emulsions for PN naturally contain both EPA and DHA. The ESPGHAN guidelines on neonatal and pediatric PN have no recommendations on the individual LC-PUFAs.2
The increase of EPA observed in this study may be of concern. In a long-term feeding study with a formula high in linolenic acid and marine oil (without additional AA), plasma and RBC values of EPA were far above the values found in the present study.61
The follow-up of those children has been the subject of a number of studies, with the general conclusion that apart from an impaired growth (AA deficiency), the group with high marine oil supplementation had a improved visual and psychomotor development at 12 months compared to the group with standard formulation and comparable with breastfed infants.62-64
A similar increase of EPA in plasma and red blood cells has been found in a recent study in preterm neonates <1250 g receiving a lipid emulsion containing 10% fish oil.45
The n-6/n-3 fatty acid ratio was significantly lower with the test emulsion as compared to controls posttreatment. This reflects the n-6/n-3 fatty acid ratio in the test emulsion corresponding to approximately 2.5:1, which is in accordance with acknowledged recommendations for adults and approximates the ratio in human milk.49,65,66
In contrast, for soybean lipid emulsions, an n-6/n-3 fatty acid ratio of 7:1 has been reported.65
In line with the present data, the authors of a clinical study in surgical patients have reported significantly higher total n-3 and lower total n-6 fatty acids in plasma phospholipids after 5 days of PN, including the mixed fish oil containing test emulsion, resulting in a decreased n-6/n-3 fatty acid ratio as compared to a standard soybean oil emulsion.67
A modification of the RBC phospholipid pattern, as indicated by a significantly lower n-6/n-3 fatty acid ratio, could be seen in another randomized controlled trial in preterm neonates receiving the test emulsion vs a conventional lipid emulsion and in children on long-term PN.24
As for many clinical studies in a sensitive population of preterm infants, this study has a number of shortcomings in addition to the reduced number of patients included. A potential source of bias for the present trial rests in the fact that neonates in both treatment groups received roughly 30% of their total fat intake from enteral supplementation. Yet, the results of the present study show that changes in the fatty acid profiles occurred considerably faster with PN as compared to the findings of investigations with oral supplementation.50
Possibly, the parenteral route might be predominant in comparison to the enteral route in achieving relevant increases in LC-PUFAs within such a short period. Further investigations are requested to investigate how long these effects can be maintained after the switch from PN to EN in this patient setting. A further potential source of imprecision might be ascribed to the fact that the number of study patients decreased rapidly after treatment day 7. Therefore, study variables were evaluated for the main study phase but also until completion of PN, with the means of the last observation carried forward (LOCF), which allowed evaluating the effect of the different lipid emulsions in most patients. Furthermore, evaluation of baseline characteristics revealed that the rate of females in the test group was higher as compared to the control group (69.2% vs 40.7%; P
= .054). According to clinical experience, preterm female neonates generally have a more favorable clinical outcome in neonatal care as compared to preterm males. On the other hand, neonates in the test group were slightly more immature with regard to gestational age, body weight, length, and the age at the start of infusions. Consequently, these factors of baseline inhomogeneity might have outweighed each other with regard to potential bias. Finally, the short study period does not allow drawing conclusions on safety when PN is used for longer periods in preterm infants.
In conclusion, the new lipid emulsion, based on a physical mixture of soybean oil, MCT, olive oil, and fish oil, was well tolerated and safe in terms of clinical laboratory and hematological parameters, AE profile, and vital signs over an infusion period of 7–14 days in preterm infants. This emulsion seems to be equally effective in providing a balanced fatty acid and effective energy supply. The primary variable, triglyceride levels, was considered safety related, and the study revealed no statistical difference between the groups, although noninferiority on the basis of serum triglycerides could not be shown in a confirmatory way. Administration of the test lipid emulsion was associated with significantly reduced total and slightly reduced direct bilirubin levels, indicating a potential beneficial effect of SMOFlipid in terms of cholestasis. Furthermore, infusion of the test lipid emulsion was equally efficient as a standard soybean emulsion in promoting the neonates’ growth while being associated with a modification of the plasma and RBC fatty acid pattern reflecting the fatty acid composition of the novel lipid emulsion. The higher availability of n-3 LC-PUFAs with this emulsion might be considered advantageous with regard to their postulated role in ensuring adequate brain and retinal development in preterm infants and the postulated anti-inflammatory and immunomodulatory role. Further studies on longer use and on a larger population are needed in preterm infants to confirm the safety and efficacy, as well as its effect on mental and visual development of this novel lipid emulsion in neonatal care.