Share this article on:

Effect of Different Levels of Docosahexaenoic Acid Supply on Fatty Acid Status and Linoleic and α-Linolenic Acid Conversion in Preterm Infants

Sauerwald, Ulrike C.*; Fink, Maria M.*; Demmelmair, Hans*; Schoenaich, Patrick v.; Rauh-Pfeiffer, Astrid A.M.*; Koletzko, Berthold*

Journal of Pediatric Gastroenterology & Nutrition: March 2012 - Volume 54 - Issue 3 - p 353–363
doi: 10.1097/MPG.0b013e31823c3bfd
Original Articles: Hepatology and Nutrition

Objectives: Long-chain polyunsaturated fatty acid (LC-PUFA) enrichment of preterm infant formulas is recommended to meet high demands. Dietary LC-PUFA may inhibit endogenous LC-PUFA synthesis, thus limiting their benefit. We investigated effects of different docosahexaenoic acid (DHA) intakes on plasma and erythrocyte fatty acids and endogenous LC-PUFA synthesis in preterm infants.

Methods: Forty-two preterm infants (birth weight 1000–2200 g) were randomized double-blind to preterm formulas with γ-linolenic acid (0.4%) and arachidonic acid (AA, 0.1%) but different DHA contents (A: 0.04%, B: 0.33%, C: 0.52%); 24 received human milk (HM: 0.51% AA, 0.38% DHA, nonrandomized). Blood was sampled on study days 0, 14, and 28. Uniformly 13C-labeled linoleic acid (2 mg/kg) and α-linolenic acid (1 mg/kg) were applied orally on day 26 and blood samples collected 48 hours later.

Results: On day 28, group A had the lowest and group C the highest plasma phospholipid concentrations of eicosapentaenoic acid and DHA. Erythrocyte phospholipid DHA was lowest in group A, but comparable in groups B, C, and HM. Plasma and erythrocyte AA were lower in formula groups than in HM. DHA intake had no effect on DHA synthesis. LC-PUFA synthesis was lower in HM infants.

Conclusions: DHA supply dose dependently increased plasma DHA. Formula DHA levels of 0.33% matched plasma DHA status of infants fed HM. LC-PUFA synthesis was lower in infants fed HM than formulas with different DHA and low AA contents. With the LC-PUFA supplementation used, DHA in formulas did not inhibit AA or DHA synthesis.

*Kinderklinik and Kinderpoliklinik, Dr. von Hauner Children′s Hospital, University of Munich Medical Centre, München

Department of Paediatrics, Zentralklinikum Augsburg, Germany.

Address correspondence and reprint request to Berthold Koletzko, Professor of Paediatrics, Division of Metabolic and Nutritional Medicine, Kinderklinik and Kinderpoliklinik, Dr von Hauner Children's Hospital, University of Munich Medical Centre, Lindwurmstr 4, 80337 München, Germany (e-mail:

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Website (

Received 2 March, 2011

Accepted 24 August, 2011 registration number: NCT01300130.

The present study was supported in part by the Bundesministerium für Forschung und Bildung, Berlin, Germany (07 ERG 08/1), Deutsche Forschungsgemeinschaft, Bonn, Germany (KO 912/5-1 and 5-2), and Nestlé Nutrition, Vevey, Switzerland. Formulas for this trial were supplied by Nestlé Nutrition, Frankfurt, Germany.

The authors report no conflicts of interest.

Because mammals require the fatty acids linoleic acid (LA, C18:2n-6) and α-linolenic acid (ALA, C18:3n-3) but are not able to synthesize them, these are essential nutrients. From these essential fatty acids, the long-chain polyunsaturated fatty acids (LC-PUFA) of the n-6 and n-3 series are formed by a series of sequential desaturation and elongation steps (1).

Term and especially preterm infants may not be able to synthesize sufficient amounts of LC-PUFA from their precursors LA and ALA relative to their high demands for rapid tissue growth, because LC-PUFA levels decline rapidly after birth in plasma and erythrocytes of formula-fed infants if not supplied exogenously (2). Preterm infants are at an even higher risk of LC-PUFA depletion than term infants, because during the last trimester of pregnancy about 80% of the total amounts of arachidonic acid (AA, C20:4n-6) and docosahexaenoic acid (DHA, C22:6n-3) present in neonatal tissues at term delivery are accumulated (3).

In vitro studies using radiolabeled precursor fatty acids provided the first evidence that endogenous LC-PUFA synthesis occurs in tissues of preterm infants. These studies demonstrated conversion of LA to AA and conversion of LA to γ-linolenic acid (GLA, C18:3n-6) and dihomo-γ-linolenic acid (DGLA, C20:3n-6) in liver microsomes (4,5). The availability of stable isotope-labeled fatty acids and of sensitive analytical methods now allows in vivo investigation of the conversion of fatty acids (6). With this methodology, synthesis of LC-PUFA from LA and ALA was demonstrated in both term and preterm infants (7–9). Although it has not been possible so far to quantify the absolute amounts of LC-PUFA formed endogenously, the results of both isotope and compositional studies in preterm infants suggest that the rate of endogenous synthesis cannot meet their high requirements for LC-PUFA during the early postnatal period (7,10). Because it has been shown that suboptimal LC-PUFA accretion during early life can affect visual and cognitive development (11–13), these fatty acids are regarded as conditionally essential nutrients for preterm infants. Therefore, it is recommended that preterm infant formulas be enriched with n-6 and n-3 LC-PUFA at levels similar to those found in HM (14,15).

Some in vitro and in vivo studies in human adults found that LC-PUFA inhibits LC-PUFA synthesis (16,17). It is not known whether such an inhibitory effect also occurs to an appreciable extent in infants, and if so, which level of LC-PUFA supply may inhibit endogenous synthesis. If product inhibition occurred in preterm infants, then high levels of DHA supplementation would not provide an added advantage, whereas the risk of potential untoward effects of exogenous LC-PUFA supply such as enhanced lipid peroxidation or impaired AA status may increase (18). For these reasons, knowledge about the effects of different levels of LC-PUFA supplementation on both the LC-PUFA status and the conversion rates of essential fatty acids toward n-6 and n-3 LC-PUFA is warranted for optimizing LC-PUFA contents of preterm infant formulas. To our knowledge, these questions have not been addressed simultaneously in a controlled clinical trial in preterm infants.

In the double-blind randomized clinical trial presented here, we studied the effect of 3 preterm infant formulas containing 0.04%, 0.33%, and 0.52% of total fatty acids as DHA on the fatty acid composition of plasma and erythrocyte phospholipids (PLs) and on the endogenous conversion of 13C-labeled LA and ALA into their corresponding LC-PUFA metabolites. A nonrandomized reference group of infants receiving HM was also included.

Back to Top | Article Outline



Sixty-six preterm infants were recruited from the neonatal units at the Zentralklinikum Augsburg, the Josephinum Hospital Augsburg, the University of Munich Medical Centre, the Lachnerklinik Munich, and the Medical Centre of the Technical University of Munich, Germany. Eligible for enrollment were preterm infants with birth weights between 1000 and 2200 g who were exclusively (≥80% of total energy intake) fed formula or human milk (HM). Exclusion criteria were apparent genetic, gastrointestinal or metabolic disorders, artificial ventilation or oxygen supply >30% at the time of enrollment, and administration of parenteral fat emulsion (>1 g triglyceride per kilogram per day for >7 days) before or after study entry.

Back to Top | Article Outline

Experimental Design

Study Entry

Infants were included in the study within the first 14 days of life after parental informed consent was obtained, full enteral feeding (>80 mL · kg−1 · day−1) was established, and at least 80% of the energy intake was derived from either formula or HM. Before the study entry, infants were fed according to hospital routine.

Back to Top | Article Outline


Enrollment into the nonrandomized HM group was based on maternal decision and the capability of providing breast milk. Formula-fed infants were randomized at study entry into 1 of 3 formula groups providing low, medium, or high DHA intake (Table 1). A blockwise randomization schedule for 3 different treatments and 5 study sites was set up using the software Rando (Dr Kühnau, Konstanz, Germany). The formulas were masked to all of the study investigators, hospital staff, and participating families by color coding of the formula labels.

Back to Top | Article Outline


Participating infants received 80% or more of their total daily energy intake as HM or their assigned study formula from study entry until the postconceptional age (PCA) of 48 weeks. HM was fortified with FM 85 (Nestlé Nutrition, Frankfurt, Germany). If formula feeding (<20% of energy intake) became necessary in the HM group, then parents were supplied with the preterm infant formula “Beba FG” (Nestlé Nutrition), which contained LC-PUFA in amounts similar to the high-DHA formula. Study formulas were based on commercially available preterm formulas. The 3 study formulas were similar in their nutrient composition and appearance except for the content of LA and ALA metabolites derived from egg lipid extracts, black currant seed oil, and low eicosapentaenoic acid (EPA, C20:5n-3) fish oil (Table 1). The study formulas were provided by Nestlé Nutrition. The low-DHA formula contained only traces of DHA (0.04% of total fatty acids), the medium-DHA formula contained 0.33% and the high-DHA formula contained 0.52%. Minor differences in AA and EPA contents of the study formulas were because of the fatty acid composition of the egg lipids and fish oil used for the supplementation (Table 1). The liquid formulas were delivered in cans of 390 mL and provided per 100 mL 80 kcal, 4.0 g fat, 2.3 g protein, and 8.6 g carbohydrates.

Back to Top | Article Outline


The following parameters were documented at birth and during the study period: gestational age; Apgar scores at 1, 5, and 10 minutes; duration of artificial ventilation and oxygen supply; special events during pregnancy, birth, and the study period; blood transfusions; occurrence of disease or complications; and corresponding therapy. Type and amount of parenteral infusions and of formula intake were recorded daily from birth until the end of the study. Intake of HM was recorded during the hospital stay by weighing the infant before and after breast-feeding.

Back to Top | Article Outline


Length and head circumference were measured with calibrated measuring equipment at birth, at study entry (study day 0), 14 and 28 days thereafter (study days 14 and 28), and at 48 weeks’ PCA. Weight was recorded daily from birth until study day 28 and at 48 weeks’ PCA. During the hospital stay, measurements were performed by specially trained personnel of the neonatal units. After hospital discharge, the families visited the neonatal units at Augsburg or at the University of Munich Medical Centre for measurements and sample collection, or they were visited at home by a study investigator and an experienced pediatrician. Growth data were expressed as z scores relative for PCA and sex relative to German reference values for preterm infants (19–21).

Back to Top | Article Outline

Sample Collection

In the HM group, 2 to 4 mL of HM were sampled on study days 0, 14, and 28 in dark-brown vials and frozen immediately at −20°C. Blood samples of the participating infants (1.5 mL) were collected in EDTA tubes on study days 0, 14, and 28. Plasma was separated by centrifugation (5 minutes, 1000g, room temperature) and frozen immediately at −20°C. Red blood cells (RBCs) were washed 3 times with NaCl solution (0.9%) and hemolyzed by adding 0.3 mL distilled water. After mixing with 2.0 mL isopropanol containing butylhydroxytoluene (Fluka, Switzerland) as an antioxidant (50 mg/L), the RBC samples were frozen immediately at −20°C. All of the frozen samples were transferred on dry ice to the laboratory within a few days for storage at −80°C until analysis.

Back to Top | Article Outline

Tracer Application

On study day 26, each preterm infant was given orally 2 mg/kg uniformly 13C-labeled LA and 1 mg/kg uniformly 13C-labeled ALA (98% isotopic purity, Martek Biosciences, Columbia, MD). The blood sample on day 28 was obtained exactly 48 hours after tracer application. Tracer enrichments of individual fatty acids were determined from the samples taken on day 28, and baseline content of 13C in the fatty acids from the blood samples taken on study day 14.

Back to Top | Article Outline

Analytical Methods

Plasma PL Fatty Acids

The methods for the determination of the fatty acid composition of plasma PL have been described in detail elsewhere (22). Briefly, plasma lipids were extracted from 0.25 mL plasma with hexane-isopropanol (3:2) after the addition of dipentadecanoylphosphatidylcholine as internal standard (Sigma, Hannover, Germany) (23), and PL were separated by thin-layer chromatography using silica gel plates (Merck, Darmstadt, Germany) according to Carnielli (24). Fatty acid methyl esters were prepared by transmethylation of the PL fatty acids in 1.5 mL methanolic HCl (3 mol/L) (Supelco Inc, Bellefonte, PA) at 85°C for 45 minutes and extracted into hexane.

Back to Top | Article Outline

RBC-Phosphatidylethanolamine and RBC-Phosphatidylcholine Fatty Acids

Six milliliters of isopropanol and 4 mL chloroform were added to the RBC-isopropanol mixture described above. Samples were centrifuged at room temperature at 1400g for 10 minutes and the chloroform phase was passed through a paper filter (Schleicher & Schuell, Dassel, Germany). The remaining isopropanol phase was again extracted with 4 mL chloroform, which was passed through the same filter. The filtrates were combined and the solvent was evaporated using a rotary evaporator (Büchi, Flawil, Switzerland). After redissolving the lipid extract in 400 μL chloroform/methanol (1/1, vol/vol), it was applied on thin-layer chromatography plates (Merck) for lipid class separation. The plates were developed in chloroform/methanol/NH3/Aqua dest. (73/27/2.2/2.8, vol/vol) and sprayed with 2,7-dichlorofluorescein (Merck) in ethanol to visualize the lipid fractions. The phosphatidylethanolamine (PE) and phosphatidylcholine (PC) fractions were transferred separately to dark-brown glass vials and transesterified and extracted as described for plasma PLs. The fatty acid methyl esters derived from RBCs were dissolved in 40 μL hexane containing butylhydroxytoluene (2 g/L) and stored at −80°C until gas chromatographic analysis.

Back to Top | Article Outline

Human Milk

HM lipids were extracted as described by Fidler et al (25). The fat content of the milk sample was determined gravimetrically and the lipid extract was transesterified in 2 mL methanol/HCl (1.5 mol/L) and 0.5 mL hexane at 90°C for 60 minutes. After the addition of water, the hexane phase was transferred to a dark-brown glass vial and stored at −80°C until gas chromatographic analysis.

Back to Top | Article Outline

Gas Chromatography and Gas Chromatography-Combustion-Isotope Ratio Mass Spectrometry

Fatty acid methyl esters were analyzed by gas chromatography using an HP Series II 5890 A gas chromatograph (Hewlett Packard, Böblingen, Germany) with split injection (1:20) and flame ionization detection on a BPX70 (60 m length, 0.32 mm inner diameter; SGE, Weiterstadt, Germany) column. The temperature program started at 130°C, increased by 4°C/minute to 180°C, by 3°C/minute to 200°C, and by 1°C/minute to 210°C, which was held for 20 minutes. Peak identification was verified by comparison with reference standards (Nu-Chek Inc, Elysian, MN).

Fatty acid concentrations in plasma PL (milligrams per liter) were calculated using the internal standard C15:0, which was added to each aliquot of 0.25 mL plasma (see above). Because no defined aliquots of RBCs were used for sample preparation, absolute fatty acid concentrations in RBC-PC and RBC-PE could not be determined. For this reason, relative fatty acid contents (% wt/wt of total fatty acids) were calculated. Calculation of relative fatty acid contents was also performed for plasma PL.

13C contents of fatty acid methyl esters were determined by Gas Chromatography-Combustion-Isotope Ratio Mass Spectrometry (delta S, Finnigan MAT, Bremen, Germany) as described previously (26). All of the samples were analyzed in duplicate and carbon isotope ratios (RFA = 13C/12C) of the fatty acid methyl esters were converted to percentage contribution of 13C to the total C in the fatty acid methyl esters (atom percent, APFA) according to the following equation:

The increase in isotopic enrichment above baseline (atom percent excess, APEFA) was obtained by subtracting the basal APFA values measured on study day 14 from the APFA values measured on day 28.

Absolute tracer concentrations (μmol 13CFA, [μmol/L]) were calculated from the plasma fatty acid concentrations (μmolFA, [μmol/L]) and the APE values by the following calculations:

CFA* is micromoles of carbon in the methylated FA per liter plasma.

The tracer concentrations obtained from Equation 3 express the micromoles of 13C derived from the applied tracer in each fatty acid per liter plasma. Furthermore, the ratios of tracer concentrations between product and precursor fatty acids were calculated as indicators of conversion rates.

Back to Top | Article Outline

Ethical Aspects

The study protocol was approved by the ethics committees of the medical faculties of the Ludwig-Maximilians-University, the Technical University Munich, and the Zentralklinikum Augsburg, Germany. Written informed consent of the parents was obtained after they had been provided with a detailed description of the study. Parents were not approached if hospital records indicated that infants did not receive >80 energy% of either HM or formula.

Back to Top | Article Outline

Statistical Analysis

Data were analyzed with SPSS for Windows 15.0 (SPSS Inc, Chicago, IL). Normally distributed data (growth, food intake, clinical characteristics) were expressed as mean and standard deviations. Differences between groups on study days 0, 14, and 28 were analyzed with analysis of variance and Bonferroni post hoc tests. Changes over time (days 0–28) within groups were analyzed with paired t tests. Fatty acid and isotope data (not normally distributed) were expressed as median and interquartile range. The effect of diet on plasma fatty acid concentration on days 0, 14, and 28 and on endogenous LC-PUFA synthesis was tested according to the Kruskal-Wallis test. If differences between groups were detected, then they were further evaluated with Mann-Whitney U tests. Changes in fatty acid concentrations over time within groups (day 0 to day 14 to day 28) were analyzed using the Friedman test. If changes were found to be significant, then Wilcoxon signed tests were performed. For correlations analysis, the Spearman coefficient of correlation was calculated. The level of statistical significance was set at P < 0.05 for all of the tests.

Back to Top | Article Outline



Of the 66 infants enrolled, 24 received HM and 42 received formula (formula A: 14, formula B: 13, formula C: 15). During the first 28 study days, 9 infants had to be excluded from the HM group because formula feeding exceeded 20 energy%. One formula-fed infant (formula A) was switched to another formula because of diarrhea and was excluded from the study. Four formula-fed infants were excluded because they met exclusion criteria. Of the 52 infants who remained in the protocol until study day 28, 15 were fed HM, 12 formula A, 12 formula B, and 13 formula C. Calculations of tracer enrichments could be made for 41 infants (9 HM, 11 formula A, 12 formula B, 9 formula C) because for 11 infants, blood samples were missing for either day 14 or day 28.

The clinical characteristics of the 52 infants who completed the first 28 days of the study are given in Table 2. No prenatal or postnatal complications that seemed to influence the results of the study were observed in these 52 infants. There were no indications of formula-related adverse events.

Back to Top | Article Outline

Intakes of HM or Formula, AA, EPA, and DHA

Fatty Acid Composition of HM

The fatty acid composition of the HM samples collected on study days 0, 14, and 28 is given in Table 1 (median contents of all of the samples collected).

Back to Top | Article Outline

LC-PUFA Intake Before Study Entry

Before study entry, the mean intakes of AA, EPA, and DHA were not significantly different between groups except for a significantly higher AA intake in group HM (26.2 ± 5.3 mg · kg−1 · day−1) compared with the formula groups (9.5 ± 4.7 mg · kg−1 · day−1).

Back to Top | Article Outline

Intake During the Study Period

There was no difference in the volumes of formula or milk intake between groups. During the study period, the mean AA intake (milligram per kilogram per day) was higher in group HM (day 28: 25.4 ± 7.2) than in the formula groups (day 28: A: 4.6 ± 0.7, B: 6.5 ± 1.1, C: 6.9 ± 0.9). The mean intakes of DHA (milligram per kilogram per day) were lowest in group A (day 28: 1.5 ± 0.2), highest in group C (day 28: 26.1 ± 3.4), and similar in groups B (day 28: 19.5 ± 3.2) and HM (day 28: 15.8 ± 5.5).

Back to Top | Article Outline


z scores for weight, length, and head circumference did not differ between groups at any time point (data not shown).

Back to Top | Article Outline

Plasma PL Fatty Acids

The PUFA concentrations (milligram per liter) in the plasma PL fraction of the 4 groups on study days 0 and 28 are shown in Table 3. On study day 28, groups A and B showed higher LA levels compared with the group HM. ALA increased during the study period in all of the groups and was lower in group HM than in formula groups at all time points. GLA was lower in group HM compared with the formula groups during the whole study period.

AA concentrations were lowest in group A at day 0. On day 28, AA concentrations were significantly lower in all of the formula groups compared with HM. There were no differences in C22:4n-6 and C22:5n-6 between groups at study entry. On study day 28, C22:4n-6 was higher in group A than in group C and higher in group HM than in groups B and C. C22:5n-6 was higher in A than in B and C.

DHA concentrations were similar in all of the groups at study entry. From day 0 to day 28, DHA concentrations rose significantly in groups B and C, but not in groups A and HM. On day 28, both absolute and relative DHA levels were lowest in A, highest in C, and similar in B and HM.

Differences in EPA levels on study day 28 were the same as for DHA. n-3 DPA levels were higher in HM than in the formula groups during the whole study period. The relative fatty acid contents of plasma PL are available in the online-only supplemental data file at

Back to Top | Article Outline

RBC Fatty Acids

The percentages of PUFA in RBC-PC and RBC-PE on study days 0 and 28 are provided in Table 4.

Back to Top | Article Outline


LA, ALA, and GLA proportions were higher in RBC-PC of formula groups compared with HM infants at study entry and on day 28. AA levels decreased in the 4 groups during the study period and were lower in all of the formula groups compared with HM on day 28. Group differences in DHA were less pronounced than in plasma PLs and there were no significant changes over time within groups. On day 28, DHA was lowest in group A and similar in groups B, C, and HM. EPA levels on day 28 were lowest in group A and highest in group C, whereas groups B and HM were similar. DPA was highest in group HM at the end of the study.

Back to Top | Article Outline


Percentages of LA, ALA, and GLA in RBC-PE were higher in all of the formula groups compared with group HM both at study entry and on day 28. As observed in RBC-PC, AA levels decreased in all of the groups from day 0 to 28. On study day 28, AA levels were lower in groups A and B compared with HM. DHA levels did not differ between groups on study day 0 but decreased significantly over time in group A. Group differences at the end of the study were similar to those in RBC-PC. Unlike the changes observed in RBC-PC, EPA increased markedly in C and in B and HM. DPA was lower in A compared with B and HM and lower in B compared with HM on study day 28.

Back to Top | Article Outline

Correlation Between DHA Intakes and DHA Status

DHA intakes (mean of days 26, 27, and 28) correlated significantly with plasma DHA concentration on day 28 (r2 = 0.51, P < 0.001), with no indication of saturation effects (Fig. 1).

Back to Top | Article Outline

Tracer Enrichment

Tracer concentrations of selected PUFA are provided in Table 5 for the 4 study groups. Figures 2 and 3 show the ratios between tracer concentrations in product and precursor fatty acids for the n-6 series (Fig. 2) and the n-3 series (Fig. 3).

Tracer concentrations were 50 times higher in LA than in ALA. Consequently, the tracer concentration ratios were lower for the n-6 series than for the n-3 series. 13CLA was highest in groups A and HM. For n-3 LC-PUFA, 13CEPA was highest in group C and 13CDHA lowest in group HM (Table 5).

The ratios of tracer concentration in products to the tracer concentrations in the precursors LA and ALA were used to compare differences in fatty acid conversion across groups. All of the ratios were consistently lower in group HM compared with the formula groups (Figs. 2 and 3). There was no significant difference in the tracer ratios of n-6 LC-PUFA across the formula groups. Concerning n-3 LC-PUFA, group C showed the highest ratio of 13CEPA/13CALA and group A showed the highest ratio of 13CDPA/13CALA. There was no difference in 13CDHA/13CALA between formula groups.

Back to Top | Article Outline

Relation Between Fatty Acid Intake and Tracer Concentration Ratios

When all 4 groups were combined, intakes of DGLA and AA correlated inversely with 13CDGLA/13CLA (P < 0.05, DGLA: r = −0.52, AA: r = −0.42) and intakes of EPA and DHA correlated inversely with 13CDPA/13CALA (P < 0.05, EPA: r = −0.41, DHA: r = −0.46). Because of the differences in fatty acid composition of the formulas and HM, the relations between intake and tracer concentrations were also analyzed for the 3 formula groups and the HM group separately. Within formula groups, intakes of EPA and DHA correlated inversely with 13CDPA/13CALA (P < 0.05, EPA: r = −0.57, DHA: r = −0.56). In the group HM, intakes of GLA and AA correlated inversely with 13CEPA/13CALA (P < 0.05, GLA: r = −0.77, AA: r = −0.86).

Gestational age at birth correlated significantly with the ratio of 13CAA/13CLA (r = 0.45, P < 0.005). There were no correlations with further clinical parameters.

Back to Top | Article Outline


Fatty Acid Concentrations in Plasma PLs

Plasma DHA concentrations of the preterm infants studied increased linearly with increasing intake of DHA from formula or HM (Fig. 1), which is in agreement with the results of Gibson et al (27) in breast-fed term infants and Smithers et al (28) in preterm infants. After feeding periods of 14 (data not shown) and 28 days, the formula containing 0.33% DHA (formula B) resulted in plasma DHA concentrations that were comparable to those of infants fed HM with a mean DHA level of 0.38%, whereas formula A, containing 0.04% DHA, and formula C with 0.53% DHA led to significantly lower and higher plasma DHA concentrations, respectively (Table 3).

This observation is in line with a number of previous studies showing significantly lower DHA levels in preterm infants not receiving dietary DHA (29–34). In group C, plasma DHA concentration increased by 58% within the first 14 study days (data not shown) and was significantly higher than in infants fed HM, formula A, or formula B from day 14 onward. These observations are in agreement with a previous study by our group in which plasma DHA concentrations in preterm infants supplemented with 0.57% DHA increased by 60% within 14 days and plateaued thereafter (35). In that study, however, infants had been fed formulas devoid of LC-PUFA before study entry and hence showed markedly lower DHA levels (15.0 mg/L) at study entry. As a consequence, the median of the DHA concentration on day 28 (23.9 mg/L) was only half of that measured in group C in the present study on day 28 (46.7 mg/L).

Plasma AA concentrations were similar in all of the formula groups at the end of the study and significantly lower than in the HM group (Table 3). This is likely the result of HM containing 3 to 7 times more AA than the formulas. In addition, the concentration of the AA precursor DGLA was about 20 times higher in HM than in the formulas (Table 1).

To balance the lower AA and DGLA contents of the formulas compared with the HM, the formulas contained 0.4% GLA. Enrichment with this LA metabolite was based on the assumption that Δ6-desaturation of LA to GLA is the rate-limiting step in endogenous AA synthesis (36–38) and GLA intake may therefore stimulate AA synthesis; however, it has been shown that a higher proportion of endogenously synthesized DGLA is converted to AA as compared with exogenous DGLA (39). This may also apply to dietary GLA and explain why supplementation of the formulas with GLA did not appreciably affect plasma PL AA concentrations. At the time the study was designed, guidelines on AA contents of preterm infant formulas were not yet available. Our data support the present recommendation on AA intake of preterm infants in Europe, which is higher than the amount provided with the formulas in the present study (15).

Back to Top | Article Outline

Relative PUFA Contents in RBC-PC and RBC-PE

RBC-PLs are often regarded as long-term markers of fatty acid intake (40), and changes observed in their LC-PUFA contents, therefore, may depend on the duration of the study. Dietary LC-PUFA intake was reported to affect first the plasma PL, then RBC-PC, and last RBC-PE (41). Therefore, the fact that we found smaller differences between groups in RBC-PC and RBC-PE compared with plasma PL may be attributed to the relatively short study period of 28 days. In agreement with Innis (41), some group differences were seen earlier in the PC fraction than in the PE fraction (Table 4). The results of the present study are also in agreement with those of other authors who suggest that PC is a better indicator of AA and PE is a better indicator of DHA status (40,41).

Back to Top | Article Outline

Tracer Concentrations in Plasma PL PUFA

We found appreciable 13C-enrichments over baseline in all of the LA and ALA metabolites, which confirms previous findings that preterm infants synthesize LC-PUFA from LA and ALA (7–9,42). When evaluating the effects of the different diets on the conversion rates of LA and ALA to LC-PUFA, some aspects and limitations related to the study design must be considered, which are discussed below.

Back to Top | Article Outline

Reasoning for Calculations Performed

13C enrichment after tracer application was determined at 1 time point only, and the ratios of tracer concentrations in products and precursors at this time point were used for estimating the conversion intensities. Demmelmair et al (43) demonstrated that these ratios obtained at a single time point correlate with the corresponding ratios of areas under enrichment curves, which are available when multiple blood samples are drawn after tracer application (44,45). Therefore, if tracer application and blood sampling adhere to the planned time schedule, valuable information can be obtained with the present study design. Although no conclusions on the kinetics of endogenous conversion or on absolute conversion rates can be drawn, it allows comparison of the effects of different exogenous fatty acid supplies. This sampling design was chosen for the present study to reduce the number of blood samples and the total amount of blood required, which is crucial for studies in preterm infants.

Some factors influencing tracer enrichments in fatty acids must be taken into consideration. One aspect is the dilution of tracer by PUFA intake. Dietary fatty acid intake diluted the tracer fatty acids and tracer-derived 13C-labeled intermediate and product fatty acids because fatty acid intakes clearly influenced concentrations of corresponding LC-PUFA in plasma PL (Table 3). This dilution effect is taken into account by using tracer concentrations rather than tracer/tracee ratios for describing the appearance of tracer in precursor and product pools. Furthermore, according to Mayes et al (42), absorption efficiency and the proportion of oxidized tracer ALA differ widely in preterm infants. As a consequence, the amount of tracer available in precursor fatty acids (LA and ALA) for conversion into long-chain metabolites varies between infants. The effect of this variation was minimized by the calculation of ratios between tracer concentrations in products and precursors. Plasma tracer concentrations further depend on incorporation of fatty acids into plasma PLs, which is higher for LA compared with other fatty acids, and on fatty acid turnover, which is lower for LA compared with ALA (39,46,47). These effects resulted in higher PL bound tracer concentrations in LA than in ALA observed in the present study.

Back to Top | Article Outline

Comparison of Tracer Ratios Between Formula Groups

As stated above, no conclusions on absolute conversion rates can be drawn with the study design chosen; however, the effects of different exogenous fatty acid supplies on conversion rates can be compared. In the present study, this is true mainly for comparisons between the formula groups because the formulas used differed in their DHA contents only.

Tracer concentration ratios were not different between formula groups for n-6 LC-PUFA and DHA. This suggests that the level of DHA supplementation used did not affect the synthesis of AA from LA and DHA from ALA, respectively. If the same conclusion can also be drawn for infants fed formulas with higher AA contents, then this requires further study.

Concerning the other n-3 LC-PUFA, tracer concentration ratios of 13CDPA over 13CALA were significantly higher in group A compared with the other groups (Fig. 3). DHA intake of the formula-fed infants correlated negatively with 13CDPA/13CALA. Of note, plasma DPA concentrations increased significantly in groups A and B but not in group C during the study period (Table 3). When group A was analyzed separately, 13CDPA/13CALA correlated significantly with the concentration of DPA in plasma PL (r = 0.782, P = 0.008). This indicates that the capability of DPA synthesis varies considerably between infants when n-3 LC-PUFA intake is low and that dietary intake of n-3 LC-PUFA inhibits conversion rates to DPA. In agreement with this finding, increased EPA and DHA consumption was associated with a decrease in the concentrations of labeled EPA and DPA, but not DHA in a study on adult men by Burdge et al (48).

The higher median conversion rate of ALA to n-3 DPA in group A did not result in higher conversion rates to DHA. This observation may be a result of the complex pathway of DHA synthesis from DPA, which is suggested to be a limiting step in DHA synthesis (49).

Both the plasma PL LC-PUFA concentrations (Table 3) and the conversion rates of LA and ALA to LC-PUFA (Figs. 2 and 3) in infants fed formula indicate that in the amounts given here, no relevant inhibition of endogenous LC-PUFA synthesis by dietary DHA occurs.

Back to Top | Article Outline

Comparison of Tracer Ratios Between Formula Groups and the HM Group

As shown in Figures 2 and 3, conversion rates of infants fed formula differed markedly from those fed HM. With regard to n-6 fatty acids, the ratios of 13C20:2n-6, 13CDGLA, and 13CAA to 13CLA were higher in formula-fed than in breast-fed infants (Fig. 2). This may at least partly explain why plasma PL C20:2n-6 concentrations tended to be higher in formula groups than in group HM on day 28, although HM contained 10 times more C20:2n-6. DGLA and AA intakes correlated inversely with the ratios of 13CDGLA and 13CAA to 13CLA. For AA intake this is in agreement with a study of Emken et al (50), in which 2 groups of adults received deuterium-labeled LA after having been on a diet containing 1.7 g AA or 0.21 g AA per day for 50 days. Higher AA intake resulted in 50% lower concentrations of deuterium-labeled DGLA and AA in plasma PL. Concerning n-3 LC-PUFA, all of the tracer concentration ratios were lower in the infants fed HM than formula.

As discussed previously, we must consider that the fatty acid contents of the 3 study formulas were similar except for DHA, whereas there were major differences in the fatty acid compositions of HM and formulas (Table 1). The higher LA and ALA contents of the study formulas and their lower AA contents may have contributed to the higher rates of LC-PUFA synthesis observed in formula-fed infants. Because of the limitations described above, the data obtained in the present study do not allow any conclusions to be reached on the likelihood or on the extent of such an effect.

One must also consider that HM feeding was not randomized, and both maternal factors and components other than LC-PUFA present in HM may have contributed to the observed lower LC-PUFA synthesis.

Back to Top | Article Outline


We found a clear dose-dependent effect of DHA intake on DHA concentrations in plasma PL within a feeding period of 28 days. Similar trends with less pronounced effect sizes were found in RBC-PL. The formula containing 0.33% DHA resulted in DHA levels that matched those in plasma and RBC PL DHA of infants fed HM. The low AA contents of the formulas used resulted in marginally low AA levels in both plasma and RBC-PL of all of the formula groups. Under these conditions and in the amounts given here, LC-PUFA supplementation seems not to inhibit the synthesis of AA from LA, respectively, or DHA from ALA, but conversion rates were markedly higher in infants fed formula compared with infants fed HM. To our knowledge, this is the first study using stable isotope tracers of LA and ALA to investigate the effect of different amounts of DHA supplementation on endogenous LC-PUFA synthesis in preterm infants.

Back to Top | Article Outline


1. Sauerwald TU, Hachey DL, Jensen CL, et al. New insights into the metabolism of long chain polyunsaturated fatty acids during infancy. Eur J Med Res 1997; 2:88–92.
2. Fleith M, Clandinin MT. Dietary PUFA for preterm and term infants: review of clinical studies. Crit Rev Food Sci Nutr 2005; 45:205–229.
3. Clandinin MT, Chapell JE, Leong S, et al. Intrauterine fatty acid accretion rates in human brain: implications for fatty acid requirements. Early Hum Dev 1980; 4:121–129.
4. Chambaz J, Ravel D, Manier MC, et al. Essential fatty acids interconversion in the human fetal liver. Biol Neonate 1985; 47:136–140.
5. Poisson JP, Dupuy RP, Sarda P, et al. Evidence that liver microsomes of human neonates desaturate essential fatty acids. Biochim Biophys Acta 1993; 1167:109–113.
6. Demmelmair H, Sauerwald T, Koletzko B, et al. New insights into lipid and fatty acid metabolism via stable isotopes. Eur J Pediatr 1997; 156:S70–S74.
7. Carnielli VP, Wattimena DJL, Luijendijk IHT, et al. The very low birth weight premature infant is capable of synthesizing arachidonic and docosahexaenoic acids from linoleic and linolenic acids. Pediatr Res 1996; 40:169–174.
8. Salem N, Wegher B, Mena P, et al. Arachidonic and docosahexaenoic acids are biosynthesized from their 18-carbon precursors in human infants. Proc Natl Acad Sci U S A 1996; 93:49–54.
9. Sauerwald TU, Hachey DL, Jensen CL, et al. Effect of dietary a-linolenic acid intake on incorporation of docosahexaenoic and arachidonic acids into plasma phospholipids of term infants. Lipids 1996; 31:131–135.
10. Sauerwald TU, Hachey DL, Jensen CL, et al. Intermediates in endogenous synthesis of C22:6w3 and C20:4w6 by term and preterm infants. Pediatr Res 1997; 41:183–187.
11. Birch EE, Birch DG, Hoffman DR, et al. Dietary essential fatty acid supply and visual acuity development. Invest Ophthalmol Vis Sci 1992; 33:3242–3253.
12. Carlson SE, Werkman SH, Peeples JM, et al. Long-chain fatty acids and early visual and cognitive development of preterm infants. Eur J Clin Nutr 1994; 48:S27–S30.
13. Helland IB, Smith L, Saarem K, et al. Maternal supplementation with very-long-chain n-3 fatty acids during pregnancy and lactation augments children's IQ at 4 years of age. Pediatrics 2003; 111:e39–e44.
14. Koletzko B, Lien E, Agostoni C, et al. The roles of long-chain polyunsaturated fatty acids in pregnancy, lactation and infancy: review of current knowledge and consensus recommendations. J Perinat Med 2008; 36:5–14.
15. Agostoni C, Buonocore G, Carnielli VP, et al. Enteral nutrient supply for preterm infants: commentary from the European Society for Paediatric Gastroenterology, Hepatology, and Nutrition Committee on Nutrition. J Pediatr Gastroenterol Nutr 2010; 50:1–9.
16. Brenner RR. Nutritional and hormonal factors influencing desaturation of essential fatty acids. Prog Lipid Res 1981; 20:41–47.
17. Emken EA, Adlof RO, Rakoff H, et al. Human metabolic studies with deuterated alpha-linolenic acid. Nutrition 1992; 8:213–214.
18. Crozier G. Galli C, Simopoulos AP, Tremoli E. Infant feeding: antioxidant aspects. Fatty Acids and Lipids: Biological Aspects. Basel:Karger; 1994. 96–101.
19. Brandt I. Perzentilkurven für die Gewichtsentwicklung bei Früh- und Reifgeborenen in den ersten fünf Jahren [Percentiles for weight of preterm and term infants during the first five years of life]. der kinderarzt 1979; 5:713–718.
20. Brandt I. Perzentilkurven für das Längenwachstum bei Früh- und Reifgeborenen in den ersten fünf Jahren [Percentiles for length of preterm and term infants during the first five years of life]. der kinderarzt 1980; 1:43–51.
21. Brandt I. Perzentilkurven für das Kopfumfangswachstum bei Früh- und Reifgeborenen in den ersten sechs Jahren [Percentiles for head circumference of preterm and term infants during the first six years of life]. Kinderarzt 1979; 10:185–188.
22. Koletzko B, Schmidt E, Bremer HJ, et al. Effects of dietary long-chain polyunsaturated fatty acids on the essential fatty acid status of premature infants. Eur J Pediatr 1989; 148:669–675.
23. Kolarovic L, Fournier NC. A comparison of extraction methods for the isolation of phospholipids from biological sources. Anal Biochem 1986; 156:244–250.
24. Carnielli VP, Pederzini F, Vittorangeli R, et al. Plasma and red blood cell fatty acid of very low birth weight infants fed exclusively with expressed preterm human milk. Pediatr Res 1996; 39:671–679.
25. Fidler N, Sauerwald T, Pohl A, et al. Docosahexaenoic acid transfer into human milk after dietary supplementation: a randomised trial. J Lipid Res 2000; 41:1376–1383.
26. Demmelmair H, von Schenck U, Behrendt E, et al. Estimation of arachidonic acid synthesis in full term neonates using natural variation of 13C content. J Pediatr Gastroenterol Nutr 1995; 21:31–36.
27. Gibson RA, Neumann MA, Makrides M. Effect of increasing breast milk docosahexaenoic acid on plasma and erythrocyte phospholipid fatty acids and neural indices of exclusively breast fed infants. Eur J Clin Nutr 1997; 51:578–584.
28. Smithers LG, Gibson RA, McPhee A, et al. Effect of two doses of docosahexaenoic acid (DHA) in the diet of preterm infants on infant fatty acid status: results from the DINO trial. Prostaglandins Leukot Essent Fatty Acids 2008; 79:141–146.
29. Clandinin MT, Garg ML, Parrott A, et al. Addition of long-chain polyunsaturated fatty acids to formula for very low birth weight infants. Lipids 1992; 27:896–900.
30. Koletzko B, Edenhofer S, Lipowsky G, et al. Effects of a low birthweight infant formula containing human milk levels of docosahexaenoic and arachidonic acids. J Pediatr Gastroenterol Nutr 1995; 21:200–208.
31. Foreman-van Drongelen MMHP, van Houwelingen AC, Kester AD, et al. Influence of feeding artificial formula milks containing docosahexaenoic and arachidonic acids on the postnatal long-chain polyunsaturated fatty acid status of healthy preterm infants. Br J Nutr 1996; 76:649–667.
32. Boehm G, Borte M, Böhles HJ, et al. 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; 155:410–416.
33. Hoffman DR, Uauy R. Essentiality of dietary omega 3 fatty acids for premature infants: plasma and red blood cell fatty acid composition. Lipids 1992; 27:886–895.
34. O Connor DL, Hall R, Adamkin D, et al. Growth and development in preterm infants fed long-chain polyunsaturated fatty acids: a prospective, randomized controlled trial. Pediatrics 2001; 108:359–371.
35. Koletzko B, Sauerwald U, Keicher U, et al. Fatty acid profiles, antioxidant status, and growth of preterm infants fed diets without or with long-chain polyunsaturated fatty acids. A randomized clinical trial. Eur J Nutr 2003; 42:243–253.
36. Marcel YL, Christiansen K, Holman RT. The preferred metabolic pathway from linoleic acid to arachidonic acid in vitro. Biochim Biophys Acta 1968; 164:25–34.
37. Bernert JT, Sprecher H. Studies to determine the role rates of chain elongation and desaturation play in regulating the unsaturated fatty acid composition of rat liver lipids. Biochim Biophys Acta 1975; 398:354–363.
38. Hassam AG, Sinclair AJ, Crawford MA. The incorporation of orally fed radioactive γ-linolenic acid and linoleic acid into the liver and brain lipids of suckling rats. Lipids 1975; 10:417–420.
39. Pawlosky RJ, Lin YH, Llanos A, et al. Compartmental analyses of plasma 13C- and 2H-labeled n-6 fatty acids arising from oral administrations of 13C-U-18:2n-6 and 2H5-20:3n-6 in newborn infants. Pediatr Res 2006; 60:327–333.
40. Carlson SE, Cooke RJ, Rhodes PG, et al. Effect of vegetable and marine oils in preterm infant formulas on blood arachidonic and docosahexaenoic acids. J Pediatr 1992; 120:S159–S167.
41. Innis SM. n-3 fatty acid requirements of the newborn. Lipids 1992; 27:879–885.
42. Mayes C, Burdge GC, Bingham A, et al. Variation in [U-13C] alpha linolenic acid absorption, beta-oxidation and conversion to docosahexaenoic acid in the preterm infant fed a DHA-enriched formula. Pediatr Res 2006; 59:271–275.
43. Demmelmair H, Iser B, Rauh-Pfeiffer A, et al. Comparison of bolus versus fractionated oral applications of 13C linoleic acid in humans. Eur J Clin Invest 1999; 29:603–609.
44. Emken EA, Adlof RO, Gulley RM. Dietary linoleic acid influences desaturation and acylation of deuterium-labeled linoleic and linolenic acids in young adult males. Biochim Biophys Acta 1994; 1213:277–288.
45. Rohwedder WK, Duval SM, Wolf DJ, et al. Measurement of the metabolic interconversion of deuterium-labeled fatty acids by gas chromatography/mass spectrometry. Lipids 1990; 25:401–405.
46. Hodson L, McQuaid SE, Karpe F, et al. Differences in partitioning of meal fatty acids into blood lipid fractions: a comparison of linoleate, oleate, and palmitate. Am J Physiol Endocrinol Metab 2009; 296:E64–E71.
47. Pawlosky RJ, Hibbeln JR, Novotny JA, et al. Physiological compartmental analysis of alpha-linolenic acid metabolism in adult humans. J Lipid Res 2001; 42:1257–1265.
48. Burdge GC, Finnegan YE, Minihane AM, et al. Effect of altered dietary n-3 fatty acid intake upon plasma lipid fatty acid composition, conversion of [13C] alpha-linolenic acid to longer-chain fatty acids and partitioning towards β-oxidation in older men. Br J Nutr 2003; 90:311–321.
49. Llanos A, Li Y, Mena P, et al. Infants with intrauterine growth restriction have impaired formation of docosahexaenoic acid in early neonatal life: a stable isotope study. Pediatr Res 2005; 58:735–740.
50. Emken EA, Adlof RO, Gulley RM. Effect of dietary arachidonic acid on metabolism of deuterated linoleic acid by adult male subjects. Lipids 1998; 33:471–480.

human milk; long-chain polyunsaturated fatty acids; premature infants; preterm infant formula; stable isotopes

Supplemental Digital Content

Back to Top | Article Outline
Copyright 2012 by ESPGHAN and NASPGHAN