Metabolic effects of intravenous LCT or MCT/LCT lipid emulsions in preterm infants.

Most lipid emulsions for parenteral feeding of premature infants are based on long-chain triacylglycerols (LCTs), but inclusion of medium-chain triacylglycerols (MCTs) might provide a more readily oxidizable energy source. The influence of these emulsions on fatty acid composition and metabolism was studied in 12 premature neonates, who were randomly assigned to an LCT emulsion (control) or an emulsion with a mixture of MCT and LCT (1:1). On study day 7, all infants received [13C]linoleic (LA) and [13C]α-linolenic acid (ALA) tracers orally. Plasma phospholipid (PL) and triacylglycerol (TG) fatty acid composition and13C enrichments of plasma PL fatty acids were determined on day 8. After 8 days of lipid infusion, plasma TGs in the MCT/LCT group had higher contents of C8:0 (0.50 ± 0.60% vs. 0.10 ± 0.12%; means ± SD) and C10:0 (0.66 ± 0.51% vs. 0.15 ± 0.17%) than controls. LA content of plasma PLs was slightly lower in the MCT/LCT group (16.47 ± 1.16% vs. 18.57 ± 2.09%), whereas long-chain polyunsaturated derivatives (LC-PUFAs) of LA and ALA tended to be higher. The tracer distributions between precursors and products (LC-PUFAs) were not significantly different between groups. Both lipid emulsions achieve similar plasma essential fatty acid (EFA) contents and similar proportional conversion of EFAs to LC-PUFAs. The MCT/LCT emulsion seems to protect EFAs and LC-PUFAs from β-oxidation.

Abstract Most lipid emulsions for parenteral feeding of premature infants are based on long-chain triacylglycerols (LCTs), but inclusion of medium-chain triacylglycerols (MCTs) might provide a more readily oxidizable energy source. The influence of these emulsions on fatty acid composition and metabolism was studied in 12 premature neonates, who were randomly assigned to an LCT emulsion (control) or an emulsion with a mixture of MCT and LCT (1:1). On study day 7, all infants received [ 13 C]linoleic (LA) and [ 13 C]A-linolenic acid (ALA) tracers orally. Plasma phospholipid (PL) and triacylglycerol (TG) fatty acid composition and 13 C enrichments of plasma PL fatty acids were determined on day 8. After 8 days of lipid infusion, plasma TGs in the MCT/LCT group had higher contents of C8:0 (0.50 6 0.60% vs. 0.10 6 0.12%; means 6 SD) and C10:0 (0.66 6 0.51% vs. 0.15 6 0.17%) than controls. LA content of plasma PLs was slightly lower in the MCT/LCT group (16.47 6 1.16% vs. 18.57 6 2.09%), whereas long-chain polyunsaturated derivatives (LC-PUFAs) of LA and ALA tended to be higher. The tracer distributions between precursors and products (LC-PUFAs) were not significantly different between groups. Both lipid emulsions achieve similar plasma essential fatty acid (EFA) contents and similar proportional conversion of EFAs to LC-PUFAs. The MCT/ LCT emulsion seems to protect EFAs and LC-PUFAs from b-oxidation.-Lehner, F., H. Demmelmair Supplementary key words medium-chain triacylglycerols . long-chain triacylglycerols . long-chain polyunsaturated fatty acids . essential fatty acids . stable isotope For many pediatric patients, parenteral nutrition is an essential and often lifesaving therapy (1). Parenteral nutrition of infants and children depends on the use of lipid emulsions that provide a high energy density in an isotonic solution and supply essential fatty acids (EFAs) and fatsoluble vitamins. In premature infants, EFA body stores are very low, whereas their metabolic requirements are high; therefore, EFA supply is of critical importance (2). Conventional fat emulsions for premature neonates are prepared from long-chain triacylglycerols (LCTs), mostly soybean oil. An emulsion based on physical mixtures of LCT and medium-chain triacylglycerols (MCTs) is widely used for the parenteral nutrition of adult patients and provides an energy source that is rapidly oxidized (3). The MCT/LCT emulsion provides less PUFA than an LCT emulsion and thus has been associated with a lower risk of lipid peroxidation and fewer alterations of membrane structures (4). High amounts of linoleic acid (LA; C18:2n-6) and a-linolenic acid (ALA; C18:3n-3) were reported to inhibit D6 desaturation, the initial step in the formation of long-chain polyunsaturated fatty acids (LC-PUFAs) (5). Thus, we hypothesized that a reduced supply of LA and ALA with the MCT/LCT emulsion might enhance LC-PUFA formation.
Because of the fast growth of brain and retina during the perinatal period, an inadequate supply of LC-PUFAs, mainly arachidonic acid (AA; C20:4n-6) and docosahexaenoic acid (DHA; C22:6n-3), may have profound effects on the development of brain and visual function in preterm infants (6,7). Although infants are able to synthesize LC-PUFAs from LA (C18:2n-6) and ALA (C18:3n-3) by desaturation and elongation from the first postnatal week onward, the rate of synthesis is rather low relative to the requirements for tissue incorporation (8,9).
In human adults, infused MCTs are oxidized faster and to a greater extent than LCTs (10), but data on their metabolism in infants are scarce (1,11). We hypothesized that in preterm infants the supply of a MCT/LCT emulsion would result in predominant oxidation of MCT, rather than LCT, as a major energy source; thus, via decreased LCT oxidation, the lower LCT intake might be partly compensated for by a higher availability of LCT for structural functions and for conversion into LC-PUFAs.
In this randomized study in preterm infants who received total parenteral nutrition for 8 days, we compared the effects of a MCT/LCT-based emulsion and of a LCT emulsion on the fatty acid composition of plasma phospholipids (PLs) and triacylglycerols (TGs). For a more detailed description of the metabolism of EFAs and their LC-PUFA derivatives, we applied 13 C-labeled fatty acid tracers, which allow for the comparison of endogenous LC-PUFA synthesis under different conditions (12,13).

SUBJECTS AND METHODS
Premature neonates were recruited in the Division of Neonatology at the University of Pécs. Inclusion criteria for the study were as follows: gestational age between 25 and 37 weeks; birth weight , 3,000 g; the indication for total parenteral feeding (expected enteral feeding , 20% of daily energy intake) for at least 8 days; and the intention to supply a lipid emulsion within 48 h after birth. Infants with known metabolic diseases were excluded.
On day 7 of the study, 10 mg/kg body weight of uniformly 13 Clabeled (98%) LA and 2 mg/kg body weight of uniformly 13 Clabeled (98%) ALA (Martek Bioscience Corp., Columbia, MD) were given orally to the infants, dissolved in a small volume of human milk. Blood samples were obtained on day 1 (before introducing the study emulsions), on day 7 (before tracer application), and on day 8. Biochemical safety parameters were measured on each day of the study (data not shown). The study protocol was approved by the ethical committee of the University of Pécs, and written informed consent was obtained from the parents of each infant before study entry.

Blood samples
Blood samples were obtained by venipuncture and collected in EDTA tubes. Plasma and red blood cells were separated by centrifugation at 1,500 g for 5 min. An aliquot was used to measure plasma lipids (cholesterol, total TG, total PL, and total nonesterified fatty acids) and 3-hydroxybutyrate. The enzymatic analyses were performed in Pécs with standard clinical chemistry methods (Lipid-Kits from Boehringer Mannheim, Mannheim, Germany).
Plasma free carnitine and acylcarnitines were determined by electrospray tandem mass spectrometry (14,15), and aand gtocopherol were determined by high-performance liquid chromatography from 100 Al of plasma (16).
For the analysis of plasma fatty acids, an internal standard was added to the samples (trinonanoin for the quantification of C8:0, C10:0, and C12:0 in TG, dipentadecanoylphosphatidylcholine and tripentadecanoin for the quantification of fatty acids with 14 or more carbon atoms; Sigma, St. Louis, MO). Samples were extracted with hexane-isopropanol, as described previously (17). Plasma PLs and TGs were isolated by thin-layer chromatography (18). For transesterification, TGs were dissolved in a methanolic HCl/hexane mixture, whereas for the transesterification of PL, pure methanolic HCl was used. After cooling of the mixture, water was added to TG samples to achieve phase separation and an aliquot of the organic phase was used for gas chromatography analysis. Evaporation steps were omitted in the preparation of FA methylesters from TG, to avoid losses of volatile medium-chain fatty acids. After cooling of the PL samples, the reactant mixture was neutralized with a dry carbonate buffer. Methylesters were extracted twice into hexane, dried under nitrogen, and taken up in hexane containing 2,6-di-tert-butyl-4-kresol for gas chromatographic analysis (19).

Gas-liquid chromatography and mass spectrometry
An HP 5890 II gas chromatograph (Hewlett-Packard, Waldbronn, Germany), equipped with a split/splitless injector, a flame ionization detector, and a BPX70 column (length, 60 m; inner diameter, 0.32 mm; SGE, Weiterstadt, Germany), was used for quantitative analyses. The temperature program started at 130jC and increased at 3 K per minute up to 210jC. Fatty acid methylesters were identified by comparison of the retention times with those of authentic standard compounds. 13 C contents in plasma PL fatty acid methylesters were determined by gas chromatography-combustion-isotope ratio mass spectrometry using an HP 5890 II gas chromatograph and a Finnigan MAT delta S isotope ratio mass spectrometer (Thermo Finnigan, Bremen, Germany) (9). Samples were analyzed in duplicate, and further calculations were performed with the means of the y 13 C values (%) obtained.
The atom percent 13 C (AP FA ; %) of the fatty acid methylester was calculated as the percentage content of 13 C relative to total C: Absolute tracer concentrations in plasma PL fatty acids were calculated as Amol of 13 C derived from the tracer in each fatty acid per liter of plasma:

Statistical analyses
Statistical analyses were performed using SAS (SAS System for Windows, release 6.12; SAS Institute, Inc., Cary, NC). Results are given as means 6 SD. Differences between the two feeding groups were examined by the U-Mann-Whitney-Wilcoxon test. Significant differences over time within one feeding group were examined by paired t-tests. Statistical significance was assumed at P , 0.05.

RESULTS
Fifteen infants were enrolled and randomized, and 12 (6 per group) completed the study according to the protocol ( Table 2). Three infants, all assigned to the MCT/LCT group, were excluded because of wrong randomization, breaching of the study conditions, and contraindication against the feeding protocol (one infant each). No dropout was related to any adverse effects of the study medication.
In the control group, birth weights were appropriate for gestational age in five infants and small for gestational age in one, whereas all infants in the MCT/LCT group had birth weights appropriate for gestational age. At baseline, the MCT/LCT group tended to have a lower body weight than the control group, which was preserved until day 8 ( Table 2) and day 10 (1,545 6 140 g vs. 1,802 6 282 g). Z-scores for body weight were calculated based on data from the longitudinal study of Kramer et al. (21) and showed no group differences ( Table 2). Average daily nutrient intakes for days 1-9 are shown in Table 3.
TG levels of the two groups were similar ( Table 4) and within the range of reference values (22), and no hypertriacylglycerolemia was observed. Plasma cholesterol concentrations were within the reference range (22). During the study period, cholesterol levels increased significantly to 3.23 6 0.83 mmol/l in the MCT/LCT group and to 3.37 6 0.64 mmol/l in the control group. Nonesterified fatty acid levels were not different between the two groups or between study time points. 3-Hydroxybutyrate concentrations decreased during the study period in the control group. Free carnitine and acylcarnitine concentrations were similar in the two groups. Neither free carnitine nor total carnitine or individual acylcarnitines decreased significantly from day 1 to day 8 in either group.
At study start, the control group had significantly lower levels of g-tocopherol (0.57 Amol/l vs. not detectable, control vs. MCT/LCT group, respectively), whereas there were no differences in a-tocopherol (10.94 vs. 8.65 Amol/l). In both groups aand g-tocopherol levels increased significantly during the study period (at day 8, 54.16 Amol/l a-tocopherol and 17.13 Amol/l g-tocopherol in controls, 49.48 Amol/l a-tocopherol and 9.55 Amol/l g-tocopherol in the MCT/LCT group), with no significant group differences on day 8.
There were no significant differences between groups in percentage values of PL fatty acids at baseline, but LA and g-linolenic acid values on day 8 were significantly higher in the LCT group ( Table 5). Despite a 50% lower LA (C18:2n-6) and ALA (C18:3n-3) supply with the MCT/ LCT emulsion, LA (C18:2n-6) levels increased to a similar 3-fold extent in both groups from day 1 to day 8. ALA (C18:3n-3) levels showed a marked, 13-fold increase in the MCT/LCT group and an 8-fold increase in the LCT group.
On day 8, the MCT/LCT group showed significantly higher levels of medium-chain fatty acids (C8:0 and C10:0). MUFAs and PUFAs in plasma TG were not different on day 1, with the exception of eicosanoic acid (C20:0) ( Table 6). In both groups, LA (C18:2n-6) and ALA (C18:3n-3) in PL and TG increased significantly from baseline to day 8, whereas AA (C20:4n-6) and DHA (C22:6n-3) decreased. In the MCT/ LCT group, there was a trend toward higher LC-PUFA contents in TG and PL, with significantly higher levels of C20: 2n-6 and DHA (C22:6n-3) in TG. At 24 h after tracer administration, infants in the MCT/ LCT group had significantly higher 13 C APE values and 13 C tracer concentrations in plasma PL fatty acids than infants receiving the LCT emulsion (Figs. 1, 2).

DISCUSSION
This study shows that infants receiving the MCT/LCT lipid emulsion, with half the EFA supply than in the LCT emulsion, have similar EFA contents in plasma PL and TG, whereas LC-PUFA levels tend to be higher. Thus, MCTs seem to enhance EFA and LC-PUFA incorporation into circulating lipids.
Both lipid emulsions were well tolerated in the infants that we followed, similar to observations made in previous studies (23). Compared with the control group, the MCT/ LCT group tended toward a lower body weight at baseline (NS), showed a greater weight loss, and regained birth weight later, although the extent of parenteral nutrition and caloric intake was comparable in the two groups. Postnatal weight loss during the first days of life reflects primarily the loss of relatively expanded fetal water com- A PUFA 5 C18:2n-6 + C18:3n-6 + C20:2n-6 + C20:3n-6 + C20:4n-6 + C22:4n-6 + C22:5n-6 + C18:3n-3 + C20:3n-3 + C20:5n-3 + C22:5n-3 + C22:6n-3. Values shown are means (SD) (%, w/w). a P , 0.05 day 8 versus day 1 within one group. b P , 0.01 day 8 versus day 1 within one group. c P , 0.05 MCT versus control. partments and is proportionally greater in infants with lower birth weight (24,25). The postnatal weight loss in the MCT/LCT group (11.4%) is within the expected range of 7.3-14.5% (26). Weight gain can best be compared with Z-scores for weight related to gender and age, which do not differ between the two groups, suggesting that under the conditions of our study the choice of the lipid emulsion did not affect weight gain. The observed plasma TG concentrations are within the accepted reference range, and the results for PL and nonesterified fatty acids are similar to previous reports for premature neonates (27,28). The increase in cholesterol levels during the first days of life has been described in previous studies after parenteral and enteral nutrition (29,30). In agreement with previous studies, we found no appreciable differences in plasma free carnitine and acylcarnitine values between the two groups (28). A significant effect of MCT in the diet or of carnitine-free parenteral nutrition might become apparent after longer periods, but during short periods of parenteral nutrition carnitine supplementation has no demonstrable benefits (31).
The control group showed lower g-tocopherol levels on day 1 but a trend toward higher values on days 7 and 8 (NS), reflecting the higher content of g-tocopherol in the LCT emulsion with its higher content of soybean oil (32). The concentrations of a-tocopherol, the most important lipid-soluble antioxidant, were similar with both emulsions ( Table 2) and well above the 12.4 Amol/l considered a threshold for vitamin E sufficiency in premature neonates (33).
stearic acid (C18:0), oleic acid (C18:1n-9), LA (C18:2n-6), and ALA (C18:3n-3) as the MCT/LCT emulsion, plasma levels of these fatty acids were rather similar in both groups, except for a slightly higher LA (C18:2n-6) value in plasma PL on day 8 in the LCT group (Table 5). This striking effect may result in part from a decreased long-chain fatty acid oxidation in the MCT/LCT group and thus incorporation of a larger proportion of the long-chain fatty acids supplied into plasma lipids. The trend toward slightly higher values of palmitic (C16:0), stearic (C18:0), and oleic (C18:1n-9) acids but slightly lower LA (C18:2n-6) and ALA (C18:3n-3) in the MCT group, compared with the LCT group, might reflect the greater oxidation of polyunsaturated fatty acids than saturated fatty acids (38). Alternatively, long-chain sat-urated fatty acids and oleic acid (C18:1n-9) might have been synthesized in the MCT/LCT group from readily available medium-chain fatty acids (39). Over the course of the postnatal study period, EFAs increased in plasma lipids, whereas LC-PUFAs decreased.
The observed values and their changes were similar to those found in enterally fed preterm neonates and reflect the change from placental fatty acid transfer, with a preferential supply of LC-PUFAs (40), to a feeding regimen supplying predominantly EFAs (9).
Polyunsaturated fatty acid turnover was assessed in this study with 13 C-labeled tracer fatty acids given orally with small volumes of human milk. Plasma APE values of LA (C18:2n-6) and ALA (C18:3n-3) in the MCT/LCT group were significantly higher, presumably because plasma LA (C18:2n-6) and ALA (C18:3n-3) pools were similar in the groups and the tracer intake was higher in this group relative to the infused amounts of LA (C18:2-6) and ALA (C18: 3n-3). Thus, there was less dilution of the tracer before incorporation into PLs or conversion to LC-PUFAs, which would also explain the higher concentrations of 13 C-labeled LA (C18:2n-6), ALA (C18:3n-3), and their labeled derivatives in the MCT/LCT group than in the control group.
Excessive availability of LA (C18:2n-6) or ALA (C18:3n-3) may inhibit desaturase activity, which would influence tracer distribution between precursor and product fatty acids at 24 h after tracer intake, when samples were obtained (41,42). Based on lower DHA (C22:6n-3) incorporation in plasma PLs in preterm infants fed formulae with MCT (43), it has been postulated that MCT might interfere with the conversion of docosapentaenoic acid (DPA; C22:5n-3) to DHA (C22:6n-3), which involves peroxisomal chain shortening of a 24 carbon intermediate (44). However, the observed distribution of the tracer amounts between precursors and products was very similar between groups for the n-6 series [LA, 82. Thus, the relative conversion of EFAs to LC-PUFAs is not influenced by the intake of MCT or the intake of LA (C18:2n-6) and ALA (C18:3n-3) under the conditions of this study. Also, the MCT supply with the lipid emulsion seems not to influence the relative incorporation of the different n-6 and n-3 fatty acids into plasma PLs. For the n-6 series, most of the tracer is found in LA (C18:2n-6), whereas in the n-3 series, the distribution is skewed toward DHA (C22:6n-3), reflecting the fact that LA (C18:2n-6) is the most abundant n-6 fatty acid, whereas DHA shows the highest percentage contribution of the n-3 fatty acids. Of importance, the tracer distribution depends on the relative incorporation of individual fatty acids in PLs but does  not indicate a higher relative LC-PUFA synthesis in the n-3 series. Given no detectable differential effect of the emulsions on the relative conversion of precursors to LC-PUFAs, the trend toward higher LC-PUFA values in the MCT/ LCT group appears to result from reduced LC-PUFA oxidation. In animal studies, dietary LC-PUFAs are oxidized to a limited extent but are preferentially incorporated into structural lipids and oxidized to a lower extent than other fatty acids (saturated, monounsaturated, LA, ALA) (45,46). We presume that the parenteral MCT supply has effectively reduced LC-PUFA oxidation and, thereby, induced the trend toward increased LC-PUFA contents in plasma lipids.
In contrast to LC-PUFAs, the EFA intermediate g-linolenic acid (C18:3n-6) showed significantly higher contents in PLs of the control group, which was also found after longer infusion periods in adults (42) but not after 5 days of lipid infusion in neonates (28). We could not determine tracer contents in g-linolenic acid (C18:3n-6) in the small samples available from these preterm infants and thus can only speculate that a rapid exchange between the relatively large LA (C18:2n-6) pool and the relatively small g-linolenic acid (C18:3n-6) pool might be the underlying metabolic cause.
We conclude that the use of the MCT/LCT emulsion in parenteral nutrition of preterm infants for a period of 8 days is well tolerated and provides equivalent carnitine, vitamin E, and EFA status compared with the LCT emulsion. The concentration of the functionally important n-3 fatty acid DHA (C22:6n-3) was higher in plasma TG of the MCT/LCT group, and there is also a trend toward higher levels of other LC-PUFAs in TG and PL. Because the availability of LC-PUFAs, and particularly of DHA (C22:6n-3), was shown to be of great functional importance in early life for the development of visual acuity (7) and cognitive development (6), the use of the MCT/LCT emulsion might provide important clinical benefits over the use of a standard soybean oil emulsion in these patients.