Main

Long chain polyunsaturated fatty acids (LCPUFAs) of the n-6 and n-3 series are metabolically derived from the essential 18 carbon n-6 and n-3 precursors, linoleic acid (LA, 18:2 n-6) and alpha-linolenic acid (ALA, 18:3 n-3), and play major roles in early development, being deeply involved in neurodevelopment and generally in growth processes (1).

Most of the n-3 and n-6 LCPUFA are acquired by the fetus in utero across the placenta (2). The pattern of LCPUFA transferred from the mother to the fetus depends on maternal stores, maternal dietary intakes, and metabolic processes taking place in the placenta (3). By the end of the second trimester of pregnancy, the fetus is capable to directly synthesize LCPUFA from their precursors (4). In the last weeks of pregnancy, a biomagnification process takes places in mammalian species, leading to a preferential accumulation of the LCPUFA derivatives in the fetus (5). Therefore, prematurity directly reduces the LCPUFA bioavailability (6).

Within term and preterm infants, intrauterine growth restriction (IUGR) is a condition associated with failure of the placenta to provide the necessary nutrients required by the fetus to maintain adequate growth. This is shown by the observation that in both the n-3 and n-6 PUFA families, the proportions of the long-chain derivatives, arachidonic acid (AA, 20:4 n-6) and docosahexaenoic acid (DHA, 22:6 n-3), are lower versus an elevation of their precursors, LA and ALA, respectively, in cases of IUGR (7). As result of IUGR due to reduction or an arrest of growth processes at various stages during the fetal life, infants at birth are small for gestational age (SGA), that is, below the 10th percentile for expected weight, compared with infant counterparts with birth weight appropriate for gestational age (AGA).

Studies with stable isotopes have shown that growth retardation slows down or reduces LCPUFA formation (8), the synthesis of n-3 compounds being more extensively affected (9). The evaluation of the LCPUFA status is therefore relevant to appreciate physiologic processes and establish dietary needs of preterm and SGA infants, respectively. Studies so far have compared LCFUFA levels of AGA versus SGA infants in a very wide interval of gestational age in utero (7), or at birth by analyzing umbilical cord plasma phospholipids within a large population inclusive of a small percentage of SGA infants (10). Also, in studies with isotopic tracers, heterogeneous groups have been progressively considered, that is, mildly preterm SGA (around 36 wks) versus preterm AGA (30 wks) versus term AGA (8), and mildly preterm SGA (34 wks) versus mild preterm AGA (34 wks) versus preterm AGA (around 30 wks) (9), based on comparisons of the study group with infants of comparable GA or weight, respectively. No study has also considered associations with maternal dietary intakes throughout pregnancy.

The aim of this investigation was studying four groups, healthy term and mildly preterm infants, further subdivided into AGA and SGA per group, within a matched design by means of the determination of the fatty acid (FA) pattern on whole blood (inclusive of all circulating lipid fractions and cells) collected on a special strip of adsorbent at d 4 on occasion of the Guthrie test, and controlling for maternal dietary habits in pregnancy. Sampling of infant's blood was therefore carried out at a very early age, comparable with sampling of cord blood at birth, but eliminating the possible contamination with maternal blood (11). This minimally invasive method for whole blood FA analysis not only facilitates the collection of samples, but also allows their easy storage and shipment to the analytical laboratories. In most LCPUFA studies, blood samples are separated and plasma and red cell PUFA values reported separately. The whole blood collected with this method represents the available “pool” of LCPUFA, with relevance for human studies on a large-scale basis (12,13), and it is also applicable in rather difficult operating conditions, such as studies in developing countries (14).

METHODS AND SUBJECTS

Term and preterm SGA and AGA infants, born in the neonatal unit of the Pediatric Department of San Paolo Hospital in Milan, were recruited throughout 1 y, starting November 2005, up to October 2006. For any SGA, two AGA, matched for GA ± 0.5 wk, were included. The study was approved by the review board of the Maternal-Infant Department at San Paolo Hospital. Parents gave their informed consent.

Entry criteria were: healthy and clinically stable (Apgar at 5 min >7), and parental informed approval of infant participation obtained after birth. Only preterms (<37th wk GA) clinically stable, admitted (if needed) into the intermediate care unit of our Department, and requiring at most parenteral glucose and amino acids (but not i.v. lipids) after birth were recruited for the study. As a consequence, our preterm infants' sample was constituted by infants with mild prematurity (>32nd wk GA in any case).

Exclusion criteria included habitual maternal alcohol consumption and smoking habits for the possible effects on infants' FA status that seem to be unrelated to being preterm or SGA (15,16). Except cases requiring parenteral nutrient supplementations, all infants were fed their own mother's milk. The choice of a mother to formula feed was an exclusion criterion for the very early consequences of formula feeding on the FA status (17). The FA composition of Italian mothers' milk has been described by our group (18).

Definition and measurements.

Maternal characteristics (age, prepregnancy body weight, weight increase in pregnancy, parity) were drawn from obstetrical charts. Maternal smoking habits and alcohol consumption were investigated through a questionnaire directly administered to mothers after delivery. Gestational age was assessed by the last menstrual period and confirmed by an ultrasound evaluation performed within the 20th wk GA. Maternal dietary habits were analyzed by means of a self-administered food frequency questionnaire (FFQ) consisting of 130 items specific for the examined population (pregnant women) related to the last 6 mos, designed according to Block (19), with a file for the frequency (daily, weekly, monthly) and standard measures indicating the eaten quantity, allowing for a semiquantitative estimation of the nutrient contribution. The FFQ forms were distributed to mothers after delivery, and were collected at discharge. For the data analysis, a software program linking the food frequency data with the nutrient database of food composition of the Italian Institute of Nutrition (20) has been developed according to indications on the correct use of FFQs for different study designs (21).

Infants were weighed naked within 30 min after delivery, and before the first feeding, by means of an electronic integrating scale (Sartorius, AG, Gottingen, Germany; precision ±5.0 g). Crown-to-heel length was measured on a recumbent infant board to the nearest millimeter by a trained operator using a Harpenden (UK) neonatometer. Cranial circumference was measured with a flexible narrow steel tape, which was applied firmly around the head above the supraorbital ridges. Infants were classified as AGA and SGA, respectively, according to Italian curves (22).

Laboratory analysis.

FAs were analyzed in a drop of whole blood absorbed on a strip of chromatography paper by an innovative method validated for reproducibility (23), and already applied to studies in infants (16). Blood collected from a heel prick at the same time of the Guthrie neonatal screening test on d 4 of life was directly subjected to transmethylation for gas chromatography analysis. FA values were then expressed as FA weight percentages (%) of total FA.

Statistical analysis.

Descriptive data are shown as mean and SD values for population characteristics and dietary data. Because FAs have nonparametric within-group distribution, their values are described with median values and interquartile ranges. The SPSS package 14.0 for Windows (SPSS, Chicago, IL) was used for the statistical analyses. Categorical variables were evaluated with the χ2 test. ANOVA with the posthoc Student-Newman-Keuls test were used to evaluate between-group differences among the four study groups in case of parametric distribution. For FA values, with nonparametric distribution, the Kruskal-Wallis test and the posthoc Friedman test were used. Statistical significant level: p < 0.05.

RESULTS

According to the study design, a total of 144 infants, born in the neonatal unit of the Pediatric Department at San Paolo Hospital in Milan and all breastfed, entered the study. According to GA and birth-weight they have been subdivided into 28 term SGA, 56 term AGA, 20 mildly preterm SGA, and 40 mildly preterm AGA. Maternal and anthropometrics characteristics are listed in Tables 1 and 2, respectively. Although no major differences in maternal characteristics were present (except for parity), weight, length, and cranial circumference were higher in term AGA infants versus the other three groups. Term SGA and preterm AGA showed quite similar values for weight, length, and head circumference.

Table 1 Maternal characteristics (mean ± SD, when not differently indicated)
Full size table
Table 2 Gestational age and neonatal characteristics at birth (mean ± SD)
Full size table

Maternal dietary intakes are reported in Table 3. There were no statistically significant differences among the four groups as far as intakes of total energy, proteins, carbohydrates, and total fats were considered. However, a higher saturated fat intake has been found for mothers of term AGA in comparison with mothers of preterm AGA, whereas mothers of both SGA groups did not differ compared with any group.

Table 3 Maternal dietary intakes (mean ± SD)
Full size table

The percentage values of the major classes of FA in whole blood lipids are reported in Table 4. Differences were present for total monounsaturated, lower in the case of term SGA, and total n-3 PUFA, lower in the case of preterms, respectively. As concerns ratios between n-3 and n-6 FA, they were higher in term groups, in parallel with higher total n-3 concentrations.

Table 4 Major fatty acids families (median, interquartile range) as fatty acid % and ratio between total n-3 and n-6 fatty acids in infants' blood
Full size table

Table 5 shows the percentage values of the main PUFA. The precursors of the two polyunsaturated FA series, LA, for the n-6, and ALA, for the n-3 series, respectively, were progressively higher in the sequence from term AGA to preterm groups, whereas the inverse trend occurred for the major LCPUFA, less markedly in the n-6 series, for di-homo-gamma-linolenic acid (20:3 n-6), and more markedly for DHA, the end-product of the n-3 series. Peculiarly, eicosapentaenoic acid (20:5 n-3) reached the highest level in term SGA. Term infants showed a trend toward higher AA levels compared with preterms.

Table 5 Major polyunsaturated fatty acids (median, interquartile range) as fatty acid % in infants' blood
Full size table

Finally, in Table 6 the most relevant product/precursor ratios of both the n-6 and the n-3 series are reported, which are representative of the metabolic steps involved in PUFA biosynthesis, regulated by alternating desaturase and elongase enzymes. In particular, the AA/LA ratio, and then the conversion rate of the essential LA to AA, the main n-6 LCPUFA, showed higher levels in term AGA and lower in term SGA infants and preterm groups, suggesting a reduction of the Δ6 desaturation activity. In the case of the n-3 series the corresponding product/precursor ratio, EPA/ALA, is higher in both term groups, even if for term SGA infants the median value is quite lower, in face of a wide interquartile range in association with maximal EPA concentrations. On the other hand, the index of the cumulative n-3 FA metabolic pathway, expressed by the DHA/ALA ratio, is quite higher in the term AGA group compared with all the other three groups. Therefore, the production of metabolic end products of the n-3 series appears to be interrupted at the level of EPA for term SGA, in which the ratio between the two 20 carbon PUFA, AA and EPA, is significantly lower versus the other three groups.

Table 6 Major product/precursor ratios of n-6 and n-3 series (median, interquartile range) in infants' blood
Full size table

DISCUSSION

We have evaluated the whole blood FA profiles in healthy infants, born at term and preterm, according to a matched study design comparing AGA and SGA infants in both groups, respectively. Because blood sampling was obtained at d 4, even considering the possible differences in the LCPUFA content in milk of mothers with preterm delivery, as found in cases of severe, and not mild, prematurity (24), the amount of milk consumed by the infant during the first days of life would have just a minimal impact on his LCPUFA status, at the point to explain differences in the FA pattern. Moreover, it has already been demonstrated that the relationships between LCPUFA concentrations at birth and postnatal changes during the first 6 wks after birth are just partly dependent on the composition of the supplied milk (25).

To our knowledge this is the first matched study on the FA status of term and preterm infants, subdivided into AGA and SGA, with four defined groups, taking into consideration also the nutrient composition of maternal diet. In addition, the FA analysis has been carried out on whole blood, inclusive of both circulating and membrane cell lipids, and therefore mostly representative of the FA status in the body on a sample out of any possible contamination from placental blood (11), reflecting both maternal passage and fetal/neonatal synthesis.

The analysis of the FA composition shows that essential FA, LA and ALA, are higher, whereas the major metabolic products, AA and DHA, together with the total n-3 PUFA are lower in preterm AGA and SGA, in comparison with the term groups, suggesting that mostly gestational age, i.e. the duration of the intrauterine life, affects blood PUFA levels. On the other side, if we consider the product/precursor ratios, it is evident that all the conversion rates are significantly higher in term AGA infants versus all the other three groups, indicating that optimal biosynthetic performances and/or complete metabolic transfers from the maternal compartment to infants may take place only in case of term delivery and adequate intrauterine growth. These between-neonate differences are not associated with differences in PUFA maternal dietary intakes. Indeed term AGA infants represent the golden standard as far as the FA pattern is considered. Levels of essential FA precursors are lower than those in the other groups. The higher levels of DHA in term AGA indicate that most transfer and incorporation of this FA in cells and tissues takes place during the last weeks of pregnancy in the normally growing infants. On the basis of our findings, however, it is not possible to speculate whether the different distributions of PUFA of the two series are associated to major changes in synthesis within fetal compartments and/or transfer from the maternal FA pool at the placenta level.

Compared with term AGA, term SGA infants have higher levels of 18C carbon precursors but similar concentrations of n-6 LCPUFA and total n-3 FA. As regards EPA levels, they are higher in term SGA than in term AGA. We may speculate that the excess of EPA in term SGA could be representative of a block behind the last metabolic steps in DHA synthesis taking place in the last weeks of term pregnancies. An alternative interpretation is that the lower rate of body growth in SGA versus AGA results in lower incorporation in cells and tissues, especially in phospholipid pools, of DHA, the end product of the n-3 series. Whatever the underlying mechanism for the elevation of EPA, as a consequence, the AA/EPA ratio is quite lower in term SGA, compared in particular with the two AGA groups. To what extent this reduction of AA relative to its analogous 20C of the n-3 series (EPA) could contribute to some unfavorable conditions within this group (26,27) is unknown. The peculiarity of the FA pattern of term SGA infants is further emphasized by the observation of lower levels of total monounsaturated fats. Although the biologic relevance of this observation is unclear, we could speculate that SGA infants near the term could increase the relative use of monoenes for energy production, considering the reduced fat stores in intrauterine growth retardation coupled with a less efficient energy production from glucose (28,29). The high EPA levels within term SGA seem to be apparently in contrast with Rump et al., (10) who found that EPA at birth was not related to weight for gestational age at birth, while a trend toward lower levels of circulating monounsaturated FA in parallel with lower birth weight was present. Although in Rump's study FA were analyzed in cord blood phospholipids, we analyzed whole blood—inclusive of circulating cell membranes, where LCPUFA are more concentrated—collected at d 4 with the Guthrie test.

Our data are consistent with findings from the studies with stable isotopes in IUGR infants showing specific impairments in the final biosynthetic pathways leading to the synthesis of LCPUFA, particularly DHA (8,9). Compared with these studies, the new information from our study includes the lack of associations with major differences in PUFA maternal dietary intakes, the high EPA-low monounsaturated FA levels in term SGA, and that even mild prematurity (that is, an average gestational age of around 35 wks) is associated with lower LCPUFA levels, particularly DHA, compared with term infants. Accordingly, maximal blood accumulation of the n-3 LCPUFA seems to occur in the last weeks of pregnancy, leading to DHA and EPA peaking in term AGA and SGA infants, respectively. Although the DHA peaking in term infants represent the physiologic situation, the EPA peaking in term SGA could be expression of their “unphysiological” metabolic condition. Therefore, although a dietary supply of DHA is advisable, common fish oils, inclusive of both EPA and DHA (and in general with higher EPA concentrations compared with DHA), could not be the best choice for this group. Investigators should elucidate in future studies the correlations between blood and tissue FA profiles (in particular of the CNS) and the associations between biochemical markers and tissue functions (considering the complications of reduced fetal growth at the neurovascular level) (27), to optimize indications of dietary fat requirements in preterm and term infants, based on being either AGA or SGA, respectively.