Intestinal absorption of long-chain polyunsaturated fatty acids in preterm infants fed breast milk or formula 1 3

Transcription

1 Intestinal absorption of long-chain polyunsaturated fatty acids in preterm infants fed breast milk or formula 1 3 Virgilio P Carnielli, Giovanna Verlato, Fabio Pederzini, Ingrid Luijendijk, Anneke Boerlage, Dino Pedrotti, and Pieter JJ Sauer ABSTRACT The importance of long-chain polyunsaturated fatty acids (LCPs) in the development of preterm infants is now well accepted but the source of dietary LCPs to be added to infant formulas remains controversial. We measured dietary intakes, fecal output, and percentages of intestinal absorption of n 6 and n 3 LCPs in healthy preterm infants fed exclusively preterm breast milk (PBM; n = 20), formula without LCPs added (NLCPs; n = 19), formula with LCPs derived from phospholipids (PL-LCPs; n = 19), or formula with LCPs from triacylglycerols (TG-LCPs; n = 19). Intestinal absorption of arachidonic acid was not different in the four groups but docosahexaenoic acid was better absorbed from PL-LCPs than from PBM (88.3 ± 1.8% compared with 78.4 ± 4.0%, P < 0.05) Total absorption of n 6 LCPs was not different between groups but total n 3 LCPs were better absorbed from PL- LCPs than from PBM or TG-LCPs (88.7 ± 1.9%, 79.2 ± 4.4%, and 80.4 ± 2.2%, respectively). In conclusion, docosahexaenoic acid and arachidonic acid were absorbed as efficiently from TG-LCPs formula as from breast milk fat. Absorption of docosahexaenoic acid and n 3 LCPs was greater from PL-LCPs formula than from PBM or TG-LCPs formula. Am J Clin Nutr 1998;67: KEY WORDS Polyunsaturated fatty acids, feeding study, preterm infants, infant formula, breast milk, fat absorption, premature infants, arachidonic acid, docosahexaenoic acid INTRODUCTION Arachidonic acid (AA, 20:4n 6) and docosahexaenoic acid (DHA, 22:6n 3) are the predominant long-chain polyunsaturated fatty acids (LCPs) derived by chain elongation and desaturation of the parent essential fatty acids linoleic acid (18:2n 6) and linolenic acid (18:3n 3), respectively. LCPs may be conditionally essential for premature infants (1 3). AA was associated with growth of fetuses (4) and infants (5). Although it was shown that both preterm (6) and full-term infants (7) are capable of synthesizing AA and DHA from linoleic acid and linolenic acid, infants who consumed formulas containing DHA had better early visual acuity than control infants (1 3, 8). The finding of higher concentrations of LCPs in the brains of infants who died from sudden infant death and were breast-fed than in the brains of bottle-fed control infants also supports the current view of the importance of dietary LCPs during infant development (9, 10). Very little information is available on the amount and ideal source of LCPs to be added to formulas consumed by infants. It is still unclear whether the fat blend in infant formulas should mimic that in breast milk as much as possible or whether dietary intake of LCPs should produce plasma concentrations and biological effects similar to those in the fetus in utero or in breast-fed infants. Breast milk fat consists of 98% triacylglycerols, 1% phospholipids, and 0.5% cholesterol and cholesterol esters (11). Although this composition remains rather constant throughout lactation, fatty acid composition of these lipid classes is affected by diet. DHA, for instance, is found in amounts ranging from 0.05% to 1.4% of total fatty acids (12). LCPs in breast milk are mainly in triacylglycerols, in which they are primarily esterified at the sn-2 and sn-3 positions (13, 14), but they also occur in phospholipids (15). In formulas for infants, LCPs are added to the fat blend by using relatively highly unsaturated lipids, such as fish oil (mainly triacylglycerols), designer oils (mainly triacylglycerols) derived from unicellular organisms, or egg phospholipids. These products differ from breast milk lipids and differences in metabolism might be expected. The differences could result from a different fatty acid composition, from different proportions of triacylglycerols and phospholipids in the fat blend, and from differences in molecular structure. Little is known about the intestinal handling of highly unsaturated molecular species reported to be relatively resistant to the action of pancreatic lipase (16). Hernell at al (17) found relative resistance to hydrolysis of triacylglycerols containing AA and eicosapentaenoic acid (20:5n 3) when triacylglycerols from rat chylomicrons were incubated in vitro with human pancreatic lipase and colipase. Few data are currently available (18) on intestinal absorption of LCPs in preterm infants, who often have impaired fat absorption during the first weeks of life (19 22). We studied dietary intakes, fecal excretion, and intestinal absorption of LCPs in preterm infants fed exclusively preterm breast milk 1 From the Department of Pediatrics, University of Padova, Italy; Ospedale S Chiara, Unità Locale Socio Sanitaria, Valle dell Adige, Trento, Italy; and Sophia Children s Hospital, Erasmus University, Rotterdam, Netherlands. 2 Supported in part by a grant from Nutricia, Zoetermeer, Netherlands. 3 Reprints not available. Address correspondence to VP Carnielli, Servizio di Patologia Neonatale, Dipartimento di Pediatria, Università di Padova, Via Giustiniani 3, Padova, Italy. carnielli@child.pedi.unipd.it. Received September 4, Accepted for publication June 16, Am J Clin Nutr 1998;67: Printed in USA American Society for Clinical Nutrition 97

2 98 CARNIELLI ET AL (PBM), preterm formula without LCPs added (NLCPs), or preterm formulas supplemented with LCPs derived from egg phospholipids (PL-LCPs) or LCPs from triacylglycerols derived from unicellular organisms (TG-LCPs). SUBJECTS AND METHODS Subjects and clinical design Four groups of preterm infants were fed either PBM (n = 20), NLCPs (n = 19), PL-LCPs (n = 19), or TG-LCPs (n = 19) formula. Composition of feedings is shown in Table 1. All infants were fed exclusively with either PBM or one of the study formulas until they were 5-wk old. Infants were free of major diseases and were growing at the time of the study. Clinical characteristics of the infants are shown in Table 2. Participation in the study was voluntary and written informed consent was obtained from the infants parents. The project was approved by the Ethics Committee of Sophia Children s Hospital/Erasmus University Rotterdam, Rotterdam, Netherlands, which agrees with the principles in the Declaration of Helsinki. Subject assignment was done randomly on the basis of unavailability of breast milk or parental decision not to breast-feed. Analyses were conducted by researchers who were blinded to the dietary assignment. The study consisted of 72-h metabolic balance assessments and gas-chromatographic determinations of dietary and fecal fatty TABLE 1 Composition of the study feedings 1 PBM (day 28) NLCPs PL-LCPs TG-LCPs (n = 20) (n = 10) (n = 6) (n = 10) Energy (kj/l) (kcal/l) Protein (g/l) Carbohydrates (g/l) Lactose (g/l) Dextrine maltose (g/l) Fat (g/l) 39 ± Fatty acid (% by wt) 8: : : : : : : :1n :1n :1n :2n :4n :3n :5n :6n Total n 6 LCPs Total n 3 LCPs PBM, preterm breast milk; LCPs, long-chain polyunsaturated fatty acids; NLCPs, no added LCPs; PL-LCPs, LCPs from phospholipids; TG- LCPs, LCPs from triacylglycerols. 2 Values from Anderson et al (23). acids. The balance evaluations were performed in infants 28-d old. Fat and fatty acid composition of PBM was obtained from direct determinations in milk samples representative of the milk fed to each infant during the study. Values for the protein and carbohydrate contents of PBM were obtained from the study by Anderson et al (23). Total lipids were extracted from milk by using the Röse- Gottlieb method (24) and amounts were determined gravimetrically. Plasma concentrations of fatty acids in infants fed PBM were reported recently (25). The NLCPs and TG-LCPs formulas were produced exclusively for this study (Nutricia Research, Zoetermeer, Netherlands) and had similar compositions, including similar fatty acid profiles, except for the addition of LCPs. The PL-LCPs formula used is currently on the market (Aptamil with Milupan; Milupa, Friedrichsdorf, Germany). Balance studies Intakes of formulas or PBM were determined by weighing bottles or, in the case of tube feeding, by recording the volume in syringes. Balance studies with separate collections of urine (24 h) and feces (72 h) were begun when infants were 28-d old. Fecal collections were carried out by bracketing with carmine red. Fifty milligrams of carmine red was dissolved in 3 ml distilled water and given either through a nasogastric tube or orally just before the 1200 feeding. Feces were then collected, including the first red stool and excluding the second first red stool about 3 d later. Collection was done by using plastic sheets placed inside diapers. Corrections were made for accidental losses of feces in diapers (double weighing of the diaper) and for feces sticking to the buttocks of the infants (double weighing of the cleaning swabs). The total amount of feces collected during the 3-d balance period was weighed and homogenized and a small sample of the homogenate was freeze-dried. Fat excretion was determined by using a modification of the method of Jeejeebhoy et al (26), with twice as much hydrochloric acid added. Excretion amounts were calculated by multiplying the volume of feces produced by the concentration of the compound of interest. Intestinal absorption was calculated by dividing the apparent amount absorbed (intake minus excretion) by the intake and multiplying by 100. Determination of fatty acids in formulas and feces Contents of individual fatty acids in the feedings and feces were determined by high-resolution capillary-gas chromatography. Analyses were done in triplicate. Fresh fecal samples (5 10 mg each) were subjected to trans-esterification by hydrochloric acid methanol after the addition of nonanoic and heptadecanoic acids as internal standards. Fatty acid methyl esters were separated and identified with a gas chromatograph (5890 II; Hewlett Packard, Amstelveen, Netherlands) equipped with a fused silica column (Supelcowax 10; 60 m 0.20 mm internal diameter; m film thickness; Supelco, Zwijndrecht, Netherlands), a flame ionization detector (280ºC), and a split-splitless injector used in splitless mode (280 ºC). The gas chromatograph was operated with the following temperature program: 60ºC initially for 5 min, raising of the oven temperature by 20ºC/min until 205ºC was reached, and holding at this temperature for 15 min. The temperature was then increased by 0.20ºC/min until 222ºC was reached. Helium was used as a carrier gas (2 ml/min) and peak areas were calculated with HP- Chem station software (Hewlett Packard) by using nonanoic and

3 LONG-CHAIN PUFA ABSORPTION IN PRETERM INFANTS 99 TABLE 2 Clinical characteristics of the infants studied 1 Characteristic (n = 20) (n = 19) (n = 19) (n = 19) Sex (M/F) 9/11 8/11 8/11 10/9 Birth weight (kg) 1.10 ± ± ± ± 0.25 Gestational age (wk) 29.8 ± ± ± ± 2.0 Weight at balance study (kg) 1.42 ± ± ± ± n = 77; PBM, preterm breast milk; LCPs, long-chain polyunsaturated fatty acids; NLCPs, no added LCPs; PL-LCPs, LCPs from phospholipids; TG- LCPs, LCPs from triacylglycerols. 2 x ± SD; there were no significant differences between groups. heptadecanoic as internal standards. Fatty acids were identified by comparing retention times with known standards (NuChek Prep, Elysian, MN). All reagents were analytic grade. Total n 6 LCPs were calculated as the sum of the following fatty acids: 20:2n 6, 20:3n 6, 20:4n 6, 22:4n 6, and 22:5n 6. Total n 3 LCPs resulted from the sum of 20:3n 3, 20:4n 3, 20:5n 3, 22:5n 3, and 22:6n 3. For the purposes of this study, fecal hydroxy fatty acids were assumed to be represented by the gas-chromatographic peaks that changed retention time after acetylation of fatty acid methyl esters. Acetylation of hydroxy fatty acid methyl esters was done in selected fecal samples with acetyl chloride for 1 h at 100ºC (Merck, Milan, Italy). Under the gas-chromatographic conditions used, no coelution occurred between fecal hydroxy fatty acids and any of the LCPs. Selected fatty acids were identified with use of a quadrupole mass spectrometer (MD 800 GC-MS; Fisons, Milan, Italy). No attempt was made to identify trans fatty acids. Statistical analysis Data are presented as group means ± SEMs unless otherwise stated. Group comparisons were done with analysis of variance; differences among groups were tested with Tukey posttests. All calculations were done with the SYSTAT statistical package (version 5.2; Systat Inc, Evanston, IL). RESULTS Clinical characteristics of the infants studied were not different between groups (Table 2). Dietary intakes of PBM or formula, fecal output, and fat-balance data are given in Table 3. Infants fed PBM had a higher fluid intake, which is often recommended with breast milk, which has a lower energy and protein content than preterm formulas. Fluid intake was also higher in infants fed PL- LCPs formula than in those fed the other two formulas. This made fat intakes comparable among the study groups. The infants fed PBM had the lowest fecal fat content (226.0 ± 69.9 mg/g); the amount was not significantly different from that in the PL-LCPs group (226.4 ± 62.7 mg/g). Infants in the other two formula groups had significantly higher values (305.3 ± 92.6 and ± 59.9 mg/g, respectively, in the NLCPs and TG-LCPs groups; P = 0.002). These differences were also found for percentages of total fat absorption because fecal output was not different among the four groups. Intakes and output of individual fatty acids are given in Tables 4 and 5. Palmitic acid intake in the PBM and PL-LCPs groups was not different but the intake values were significantly higher that those in the NLCPs and TG-LCPs groups. In the PBM group, AA intake was 32.0 ± 2.1 mg kg 1 d 1 but intake of total n 6 LCPs was three times higher (92.9 ± 5.6 mg kg 1 d 1 ). The same difference applied to intakes of DHA and total n 3 LCPs, which were 18.0 ± 1.3 and 37.7 ± 2.9 mg kg 1 d 1, respectively, indicating the presence of other n 3 and n 6 polyunsaturated fatty acids in addition to AA and DHA. Lower amounts of these other fatty acids were found in the PL-LCPs formula but there were virtually none in the TG-LCPs formula. Metabolizable intakes of individual fatty acids in the 77 infants are shown in Table 6. Intestinal absorption of each individual fatty acid for which dietary intake could be calculated is shown in Table 7. Palmitic acid was better absorbed from PBM than from formulas. Linoleic acid was also better absorbed from PBM (88.1 ± 3.0%) and PL-LCPs formula (91.3 ± 1.7%) than from the NLCPs (69.7 ± 3.4%) and TG-LCPs (68.9 ± 4.0%) formulas. AA was absorbed in similar percentages from PBM, PL-LCPs, and TG-LCPs. Percentages of DHA absorption were similar in the PBM and TG-LCPs TABLE 3 Dietary intakes, fecal output, and fat-balance data in preterm infants fed preterm breast milk or formula 1 Variable (n = 20) (n = 19) (n = 19) (n = 19) P value 2 Dietary intake (ml kg 1 d 1 ) 173 ± 14 a 154 ± 5 b 160 ± 5 a 153 ± 11 b Fecal output (g kg 1 d 1 ) 5.3 ± ± ± ± Fecal fat (mg/g) ± 69.9 a ± 92.6 b ± 62.7 a ± 59.9 a,b Fat intake (g kg 1 d 1 ) 7.5 ± 2.1a 6.6 ± 0.4 a,b 6.1 ± 0.2 b 6.6 ± 0.5 b Fat output (g kg 1 d 1 ) 1.3 ± 0.9 a 2.1 ± 0.6 b 1.5 ± 0.4 a,b 1.9 ± 0.7 b Fat absorption (%) 82.3 ± 13.8 a 68.6 ± 9.2 b 75.6 ± 6.9 a,b 69.9 ± 10.9 b x ± SD; n = 77. PBM, preterm breast milk; LCPs, long-chain polyunsaturated fatty acids; NLCPs, no added LCPs; PL-LCPs, LCPs from phospholipids;

4 100 CARNIELLI ET AL groups but lower than that in the PL-LCPs group. DISCUSSION This study showed that in small preterm infants fed PBM, LCPs are not absorbed completely, as has been believed, and that LCPs bound to phospholipids are better absorbed than triacylglycerol LCPs from either formula or breast milk. The study provided further evidence that intestinal absorption of total fat and palmitic acid is better from breast milk than from formula. Reliable calculation of intestinal absorption of individual polyunsaturated fatty acids requires processing of feces immediately after excretion to avoid bacterial and oxidative degradation of the most unsaturated fatty acids. Use of high-resolution capillary-gas chromatography is also necessary to minimize coelution of several unknown peaks (mostly hydroxy fatty acids) that may interfere with determination of the more common dietary fatty acids. We paid attention to these conditions and found that 80% of AA and DHA from breast milk was absorbed. In PBMfed infants, this value approximated the amount of total fat absorption and was similar to the palmitic acid absorption (78.4 ± 4.0%) but lower than that of oleic acid absorption (85.8 ± 3.2%). Absorption of LCPs from breast milk was thus not better than absorption of saturated fat. In breast milk, palmitic acid is preferentially esterified to the center sn-2 position and LCPs are more represented at the sn-2 and sn- 3 positions (14, 15, 27). It is thus conceivable that LCPs at the external position may have a disadvantage compared with fatty acids attached to the center position, which does not have to be cleaved for absorption. Lingual, gastric, and pancreatic lipases preferentially hydrolyze the sn-1 and sn-3 ester bonds of triacylglycerols but their activity is greater toward short- and medium-chain fatty acids than toward long-chain fatty acids, especially LCPs (16, 28, 29). Intestinal absorption of AA and DHA from TG-LCPs formula TABLE 4 Fatty acid intake in preterm infants fed exclusively preterm breast milk or formula at 4 wk of age 1 was similar to that from PBM. This is an important finding because this formula contained highly unsaturated triacylglycerol molecular species, with up to 40% DHA and up to 30% AA. Unlike in breast milk triacylglycerols, in which DHA has a clear positional specificity (14), DHA in algal oils does not have a strong positional specificity; there are similar proportions at the sn-1, sn-2, and sn-3 positions (data not shown). On the basis of work by Bottino et al (16), who described resistance of certain LCPs of fish oils to hydrolysis by pancreatic lipase, and data of Hernell et al (17), relatively low absorption might also have been expected with the highly unsaturated triacylglycerols from algae. However, our data indicated that LCPs from these triacylglycerols were absorbed as efficiently as those from breast milk, even without the lipolytic enzymes that are present in breast milk but not in formulas (17). Differences in the molecular structure of these algal triacylglycerols and fish oils (with which most studies have been performed) could be responsible for the different digestibility and absorption of these two products. DHA from PL-LCPs was absorbed better than DHA from PBM. This finding clearly indicates that intestinal absorption of phospholipids in preterm infants is efficient. Data on intestinal absorption of phospholipids in human infants and adults is lacking and information on this issue is derived mainly from studies in animals (30). It is known, however, that under normal dietary conditions in human adults, dietary phospholipids are a minor portion of the phospholipids presented to the gut for intestinal absorption; the majority are phospholipids of biliary origin (31). Whether this also applies in newborn infants is not known. However, to ensure intake of LCPs from a phospholipid source in the range of intake of breast milk, dietary intake of phospholipids must far exceed the phospholipid content of breast milk (18). Proportions of intestinal absorption representing the sum of n 6 and n 3 LCPs were similar to those of AA and DHA, Fatty acid (n = 20) (n = 19) (n = 19) (n = 19) P value 2 mg kg 1 d 1 8: ± 1.2 a ± 1.7 b 68.0 ± 0.5 c ± 2.5 b : ± 8.5 a ± 1.4 a 77.6 ± 0.6 b ± 2.1 a : ± 51.5 a ± 10.4 b ± 2.3 c ± 16.4 b : ± 55.6 a ± 5.0 b ± 2.4 b ± 7.2 b : ± a ± 6.9 b ± 11.7 a ± 9.7 b : ± 34.1 a ± 3.2 b ± 3.2 a ± 5.0 b : ± 1.0 a 16.1 ± 0.2 b 13.5 ± 0.1 a 17.0 ± 0.3 b :1n ± 16.7 a 12.4 ± 0.2 b 57.5 ± 0.5 c 10.2 ± 0.2 b :1n ± a ± 31.9 b ± 14.2 c ± 47.5 a,b :1n ± 10.1 a 37.7 ± 0.5 b ± 0.9 c 37.4 ± 0.7 b :1n ± 2.1 a 10.2 ± 0.1 b 9.1 ± 0.1 b 10.7 ± 0.2 b :2n ± 64.7 a ± 10.4 b ± 5.2 a ± 16.1 b :4n ± 2.1 a ± 0.2 b 49.7 ± 1.0 c :3n ± 4.1 a 72.9 ± 0.9 b 54.1 ± 0.4 c 71.5 ± 1.3 b :5n ± 0.9 a ± 0.4 b :6n ± 1.3 a ± 0.1 b 37.8 ± 0.7 c Total n 6 LCPs 92.9 ± 5.6 a ± 0.2 b 57.2 ± 1.1 c Total n 3 LCPs 37.7 ± 2.9 a ± 0.5 b 37.9 ± 0.7 a x ± SEM; n = 77. PBM, preterm breast milk; LCPs, long-chain polyunsaturated fatty acids; NLCPs, no added LCPs; PL-LCPs, LCPs from phospholipids; TG-LCPs, LCPs from triacylglycerols. Values with different superscript letters are significantly different, P < 0.05 (Tukey test).

5 LONG-CHAIN PUFA ABSORPTION IN PRETERM INFANTS 101 TABLE 5 Fecal output of fatty acids in preterm infants fed exclusively preterm breast milk or formula at 4 wk of age 1 Fatty acid (n = 20) (n = 19) (n = 19) (n = 19) P value 2 mg kg 1 d 1 8:0 1.2 ± ± ± ± :0 1.9 ± ± ± ± : ± 8.7 a 55.4 ± 7.2 b 16.0 ± 1.0 a 39.9 ± 5.9 a,b : ± ± ± ± : ± 49.0 a ± 16.6 a ± 30.7 b ± 19.4 a : ± 19.8 a ± 10.5 a,b ± 14.0 b ± 14.0 a,b :0 5.4 ± 0.7 a 19.8 ± 1.3 b 11.5 ± 0.7 c 22.7 ± 1.5 b :1n ± 3.6 a 1.3 ± 0.2 b,c 8.3 ± 0.8 ac 0.7 ± 0.2 b :1n ± 64.1 a ± 54.5 b ± 31.4 a ± 64.3 b :1n ± 4.2 a 18.6 ± 2.6 a,b 24.6 ± 2.4 a,b 15.2 ± 2.1 b :1n ± 1.2 a 13.7 ± 1.2 b 3.3 ± 0.3 c 15.2 ± 1.3 b :2n ± 22.5 a ± 28.9 b 57.2 ± 10.8 a ± 31.9 b :4n ± 1.2 a 1.2 ± 0.1 b 3.0 ± 0.3 a,b 9.8 ± 1.2 c :3n ± 1.0 a 8.4 ± 1.1 b 3.0 ± 0.6 a 6.9 ± 1.2 a,b :5n ± 0.4 a 1.7 ± 0.4 a,b 0.7 ± 0.2 b 2.2 ± 0.4 a :6n ± 0.7 a 0.3 ± 0.1 b 1.6 ± 0.2 b 7.4 ± 0.9 c Total n 6 LCPs 19.8 ± 3.4 a 4.6 ± 1.2 b 8.3 ± 0.7 b,c 13.3 ± 1.3 c Total n 3 LCPs 7.2 ± 1.3 a,b 4.2 ± 1.9 b 2.9 ± 0.5 b 9.8 ± 1.1 a x ± SEM; n = 77. PBM, preterm breast milk; LCPs, long-chain polyunsaturated fatty acids; NLCPs, no added LCPs; PL-LCPs, LCPs from phospholipids; respectively. This indicated that other minor LCPs behave in a similar way with respect to absorption. In the PBM group, intakes of n 6 and n 3 LCPs were more than double than those of AA and DHA, respectively. This resulted from the presence of several intermediates of the n 6 and n 3 families in breast milk. Because these fatty acids have biological importance and may be converted into other LCPs, their dietary intakes should be taken into account, especially when breast milk is compared with supplemented formulas. Infants fed PL-LCPs formula had total n 6 and n 3 LCPs intakes that were much higher that those of AA and DHA alone. Eicosapentaenoic acid contributed to 13.5% of total n 3 LCPs in PBM ((5 mg kg 1 d 1 ) and 8.1% in PL-LCPs (2.1 mg kg 1 d 1 ). This difference may deserve attention in future studies. With the TG-LCPs formula, AA and TABLE 6 Metabolizable intake of fatty acids in preterm infants fed exclusively preterm breast milk or formula at 4 wk of age 1 Fatty acid (n = 19) (n = 19) (n = 19) (n = 19) P value 2 mg kg 1 d 1 8: ± 1.2 a ± 1.8 b 66.0 ± 0.6 c ± 2.5 b : ± 8.4 a ± 2.5 a 76.2 ± 0.7 b ± 2.7 a,b : ± 51.1 a ± 12.5 b ± 2.8 c ± 16.9 b : ± 56.3 a ± 7.5 b ± 4.3 b ± 10.4 b : ± a ± 17.7 b ± 33.5 c ± 24.8 b : ± 34.0 a 68.3 ± 11.2 b ± 14.0 c 67.4 ± 17.0 b ± 1.1 a 3.8 ± 1.2 b 2.0 ± 0.7 c 5.7 ± 1.7 b :1n ± 16.1 a 11.0 ± 0.2 b 49.2 ± 1.0 c 9.5 ± 0.3 b :1n ± a,b ± 60.6 a,b ± 37.0 a ± 83.1 b :1n ± 10.8 a 19.0 ± 2.6 b 88.5 ± 2.3 a 22.3 ± 2.2 b :1n ± 2.0 a 3.5 ± 1.2 b 5.7 ± 0.3 c 4.4 ± 1.4 b :2n ± ± ± ± :4n ± 2.1 a NA 16.4 ± 0.4 a 40.0 ± 1.4 b :3n ± 3.8 a 64.5 ± 1.2 b 51.0 ± 0.8 c 64.7 ± 1.8 b :5n ± 1.0 NA 1.4 ± 0.4 NA :6n ± 1.2 a NA 12.0 ± 0.3 a 30.6 ± 1.1 b Total n 6 LCPs 73.0 ± 6.0 a NA 21.6 ± 0.7 c 43.8 ± 1.5 b Total n 3 LCPs 30.5 ± 3.0 a NA 23.0 ± 0.7 b 30.4 ± 1.1 a x ± SEM; n = 77. PBM, preterm breast milk; LCPs, long-chain polyunsaturated fatty acids; NCPs, no added LPCs; PL-LCPs, LCPs from phospholipids,

6 102 CARNIELLI ET AL TABLE 7 Intestinal absorption percentages of fatty acids in preterm infants fed exclusively preterm breast milk or formula at 4 wk of age 1 Fatty acid (n = 20) (n = 19) (n = 19) (n = 19) P value 2 % 8: ± 1.0 a 98.8 ± 0.3 b 97.2 ± 0.3 a 98.6 ± 0.2 b : ± ± ± ± : ± ± ± ± : ± 2.7 a 81.3 ± 1.5 b 84.0 ± 1.1 a,b 84.4 ± 1.7 a,b : ± 4.0 a 51.4 ± 3.0 b 58.7 ± 2.2 b 49.3 ± 4.2 b : ± 5.5 a 26.2 ± 4.2 b 40.2 ± 3.5 b 24.0 ± 6.2 b : ± 6.1 a 23.4 ± 7.9 b 14.8 ± 5.1 c 35.8 ± 10.2 b :1n ± 2.5 a,b 89.6 ± 1.8 a,b 85.5 ± 1.5 b 93.3 ± 2.2 a :1n ± 3.2 a,c 75.0 ± 2.1 b 88.3 ± 1.8 c 77.9 ± 2.5 ab :1n ± 4.0 a 50.2 ± 6.8 b 78.3 ± 2.0 a 59.4 ± 5.6 b :2n ± 3.0 a 69.7 ± 3.4 b 91.3 ± 1.7 a 68.9 ± 4.0 b :4n ± ± ± :3n ± ± ± ± :5n ± 6.9 a ± :6n ± 4.0 a ± 1.8 b 80.6 ± 2.1 a,b Total n 6 LCPs 78.0 ± ± ± Total n 3 LCPs 79.2 ± ± ± 2.2 b x ± SEM; n = 77. PBM, preterm breast milk; LCPs, long-chain polyunsaturated fatty acids; NLCPs, no added LCPs; PL-LCPs, LCPs from phospholipids; DHA intakes were >85% of total n 6 and n 3 LCPs, respectively, and eicosapentaenoic acid was absent. Mean intake of total n 3 LCPs was lower in the PL-LCPs group than in the PBM and TG-LCPs groups (25.9, 37.7, and 37.9 mg kg 1 d 1, respectively, in the three groups). Intakes of LCPs in PBM-fed preterm infants were slightly higher than those reported by Boehm et al (32), who compared a group fed breast milk with groups fed an unsupplemented formula and a supplemented formula similar to the PL-LCPs formula in our study. This difference in intake of LCPs in our study and that in the study by Boehm et al may have resulted from different fluid intakes or from the fact that the infants in our study were fed a PL-LCPs formula with a higher energy concentration than the infants in the study by Boehm et al (332.8 compared with kj/d). Fecal output data indicated that losses of LCPs in preterm infants are not negligible. Infants fed NLCPs formula lost 4.6 ± 1.2 mg n 6 LCPs kg 1 d 1 and 4.2 ± 1.9 mg n 3 LCPs kg 1 d 1. A portion of these fecal LCPs may derive from endogenous sources -bile lipids, enterocytes shed into the gut lumen, and even bacterial walls. It was shown that bile in human adults contains LCPs (33), but no information is available on children or premature infants nor do data exist on the amount of bile produced. It is likely, however, that a large portion of the LCPs of bile lipids is absorbed along the intestinal tract. Although it is theoretically possible that LCPs in feces are not a reflection of gut absorption, circumstantial evidence suggests that this is not the case. First, much higher amounts are found in stools of infants consuming LCPs in their diets. Second, when LCP content of PBM rises, so does the amount excreted in the stool. We also calculated metabolizable intakes (Table 6) because we believe that these values provide better estimates of the amounts of LCPs entering the body than do values derived from diet alone. Furthermore, we believe that fatty acid content, especially LCP content, in infant formulas is better specified in absolute amounts than as a percentage of total fat because fat content of infant formulas may differ markedly. A difference in total fat of 1.2 g/d (for example, an increase from 3.2 to 4.4 g/d) could lead to a difference in estimated LCP intake of 28%. It is also worth emphasizing that in our subjects 20 25% of dietary LCPs were lost in the feces. In conclusion, our study showed that there are substantial amounts of LCPs in the feces of PBM- and formula-fed preterm infants. We also found that DHA from egg phospholipids was absorbed better than DHA from PBM. DHA and AA from singlecell oils were absorbed as efficiently as DHA and AA from breast milk fat. REFERENCES 1. Uauy RD, Birch DG, Birch EE, Tyson JE, Hoffman DR. Effect of dietary omega-3 fatty acids on retinal function of very-low-birthweight neonates. Pediatr Res 1990;28: Birch EE, Birch DG, Hoffman DR, Uauy R. Dietary essential fatty acid supply and visual acuity development. Invest Ophthalmol Vis Sci 1992;33: Carlson SE, Werkman SH, Rhodes PG, Tolley EA. Visual-acuity development in healthy preterm infants: effect of marine-oil supplementation. Am J Clin Nutr 1993;58: Koletzko B, Braunn M. Arachidonic acid and early human growth: is there a relation? Ann Nutr Metab 1991;35: Carlson SE, Werkman SH, Peeples JM, Cooke RJ, Tolley EA. Arachidonic acid status correlates with first year growth in preterm infants. Proc Natl Acad Sci U S A 1993;90: Carnielli VP, Wattimena JDL, Luijendijk IHT, et al. The very-lowbirth-weight premature infant is capable of synthesizing arachidonic and docosahexaenoic acid from linoleic and linolenic acid. Pediatr Res 1996;39: Salem N Jr, Wegher B, Mena P, Uauy R. Arachidonic and docosahexaenoic acid are biosynthesized from their 18-carbon precursors in human infants. Proc Natl Acad Sci U S A 1996;93: Carlson SE, Ford AJ, Werkman SH, Peeples JM, Koo WWK. Visual acuity and fatty acid status of term infants fed human milk and

7 LONG-CHAIN PUFA ABSORPTION IN PRETERM INFANTS 103 formulas with and without docosahexaenoate and arachidonate from egg yolk lecithin. Pediatr Res 1996;39: Farquharson J, Cockburn F, Patrick WA, Jamieson EC, Logan RW. Infant cerebral cortex phospholipid fatty-acid composition and diet. Lancet 1992;340: Makrides M, Neumann MA, Byard RW, Simmer K, Gibson RA. Fatty acid composition of brain, retina, and erythrocytes in breastand formula-fed infants. Am J Clin Nutr 1994;60: Lammi-Keefe CJ, Jensen RG. Lipids in human milk: a review. 2: composition and fat soluble vitamins. J Pediatr Gastroenterol Nutr 1984;3: Innis SM. Human milk and formula fatty acids. J Pediatr 1992;120:S Breckenridge WC. Triglyceride structure of human milk fat. Can J Biochem 1969;47: Martin JC, Bougnoux P, Antoine JM, Lanson M, Couet C. Triacylglycerol structure of human colostrum and mature milk. Lipids 1993;28: Jensen RG. Lipids in human milk-composition and fat-soluble vitamins. In: Lebenthal E, ed. Textbook of gastroenterology and nutrition in infancy. 2nd ed. New York: Raven Press, 1989: Bottino NR, Vandenburg GA, Reiser R. Resistance of certain longchain polyunsaturated fatty acids of marine oils to pancreatic lipase hydrolysis. Lipids 1967;2: Hernell O, Blackberg L, Chen Q, Stenby B, Nilsson A. Does the bile salt-stimulated lipase of human milk have a role in the use of the milk long-chain polyunsaturated fatty acids? J Pediatr Gastroenterol Nutr 1993;16: Morgan C, Spencer SA, Stammers J, Colley J, Hull D. Differences between long-chain polyunsaturated fatty acid absorption in term and preterm infants. Br J Clin Pract 1995;80(suppl): Atkinson SA, Bryan MH, Anderson GH. Human milk feeding in premature infants: protein, fat, and carbohydrate balances in the first two weeks of life. J Pediatr 1981;99: Brooke OG, Wood C, Barley J. Energy balance, nitrogen balance, and growth in preterm infants fed expressed breast milk, a premature infant formula, and two low-solute adapted formulae. Arch Dis Child 1982;57: Jarvenpaa AL. Feeding the low-birth-weight infant IV. Fat absorption as a function of diet and duodenal bile acids. Pediatrics 1983;72: Watkins JB. Lipid digestion and absorption. Pediatrics 1985;75(suppl): Anderson GH, Atkinson SA, Bryan MH. Energy and macronutrient content of human milk during early lactation from mothers giving birth prematurely and at term. Am J Clin Nutr 1981;34: Normalisatie-Instituut. Basismethode voor de bepaling van het vetgehalte van melk volgens Röse-Gottlieb. (Basic method for the determination of the fat content of milk.) s-gravenhage, Netherlands: Normalisatie-Instituut, 1969 (in Dutch). 25. Carnielli VP, Pederzini F, Vittorangeli R, Luijendijk IHT, Pedrotti D, Sauer PJJ. Plasma and red blood cell fatty acid of very-low-birthweight infants fed exclusively with expressed preterm human milk. Pediatr Res 1996;39: Jeejeebhoy KN, Ahmad S, Kozak G. Determination of fecal fats containing both medium and long chain triglycerides and fatty acids. Clin Biochem 1970;3: Breckenridge WC. Stereospecific analysis of triacylglycerols. In: Kuksis A, ed. Handbook of lipid research 1 fatty acids and glycerides. New York: Plenum Press, 1978: Wang CS. Hydrolysis of dietary glycerides and phosphoglycerides: fatty acid and positional specificity of lipases and phospholipases. In: Kuksis A, ed. Fat absorption. Boca Raton, FL: CRC Press, 1986: Yang LY, Kuksis A, Myher JJ. Lipolysis of menhaden oil triacylglycerols and the corresponding fatty acid alkyl esters by pancreatic lipase in vitro: a reexamination. J Lipid Res 1990;31: Borgstrom B. Phospholipid absorption. In: Rommel K, Bohmer R, eds. Lipid absorption: biochemical and clinical aspects. Baltimore: University Park Press, 1976: Akesson B. Content of phospholipids in human diets studied by the duplicate-portion technique. Br J Nutr 1982;47: Boehm G, Borte M, Bohles HJ, Muller H, Kohn G, Moro G. Docosahexaenoic and arachidonic acid content of serum and red blood cell membrane phospholipids of preterm infants fed breast milk, standard formula or formula supplemented with n 3 and n 6 long-chain polyunsaturated fatty acids. Eur J Pediatr 1996;155:410 6.