Share this article on:

Molecular Fingerprints of the Human Fecal Microbiota From 9 to 18 Months Old and the Effect of Fish Oil Supplementation

Andersen, Anders Daniel*; Mølbak, Lars; Michaelsen, Kim Fleischer*; Lauritzen, Lotte*

Journal of Pediatric Gastroenterology and Nutrition: September 2011 - Volume 53 - Issue 3 - p 303–309
doi: 10.1097/MPG.0b013e31821d298f
Original Articles: Hepatology and Nutrition

Objectives: The aim of this study was to monitor changes in the fecal microbiota from 9 to 18 months and to investigate the effect of increasing dietary n-3 polyunsaturated fatty acids on the fecal microbiota.

Patients and Methods: In a double-blind controlled trial with random allocation to daily supplementation with 5 mL of fish oil (FO) or sunflower oil (SO) from 9 to 18 months of age, stool samples were collected from 132 healthy Danish infants. Molecular fingerprints of the bacterial DNA were obtained by terminal restriction fragment length polymorphism (T-RFLP).

Results: The T-RFLP profiles indicated that a few T-RFs became dominant with age (bp100 and 102, both presumed to be Bacteroidetes) concomitantly with an overall increase in the microbial diversity (P = 0.04). Breast-feeding influenced both the T-RFLP profiles at 9 months and the changes from 9 to 18 months, and breast-feeding cessation during the trial modified the response to the dietary oils. In the FO group, the increase in bp102 was significantly reduced among children weaned before compared with those weaned during the trial (P = 0.027), whereas the increase in bp100 was reduced in the preweaned children of the SO group relative to those weaned during the trial (P = 0.004). This was supported by intervention group differences in the changes in bp102 and bp100 among the earlier weaned children (P = 0.06 and P = 0.09, respectively).

Conclusions: Cessation of breast-feeding played a dominant role relative to developmental changes in the fecal microbiota from 9 to 18 months. FO compared with SO supplementation affected changes in large bacterial groups, but only among children who had stopped breast-feeding before 9 months of age.

*Department of Human Nutrition, Faculty of Life Sciences, University of Copenhagen, Frederiksberg C, Denmark

Division of Veterinary Diagnostics and Research, National Veterinary Institute, Technical University of Denmark, Bülowsvej 27, DK-1790 Copenhagen V, Denmark.

Address correspondence and reprint requests to Lotte Lauritzen, Department of Human Nutrition, Faculty of Life Sciences, University of Copenhagen, Rolighedsvej 30, DK-1958 Frederiksberg C, Denmark (e-mail:

Received 9 December, 2010

Accepted 30 March, 2011

The present study was supported by the Danish Council for Strategic Research, Programme Commission for Food and Health; intervention oils were donated by Axellus A/S (Oslo, Norway). registration no. NCT 00631046.

The authors report no conflict of interests.

It is hypothesized that environmental factors such as diet may affect health in later life via effects on the colonization pattern of the developing microbiota in the first years of life. A dynamic colonization of the sterile infant gastrointestinal tract begins immediately after birth and has been shown to be influenced by factors such as mode of delivery (1), infant feeding (2), and other environmental exposures. A stable and highly complex individual-specific “adult-like” microbiota is believed to be established within the first years of life after finishing breast-feeding (3). A healthy microbiota is important for host health, and deviations may affect later health. For instance, aberrations of the early bacterial fecal composition have been associated with atopy development (4,5), differences in systemic immune responses (6), and also risk of obesity in later childhood (7).

Dietary influences on the microbiota early postnatally have been extensively studied. Breast milk contains different nutritional components (eg, galacto-oligosaccharides, long-chain polyunsaturated fatty acids [LCPUFA]), and some of these presumably confer selective advantages to certain bacteria and account for some of the observed differences between exclusively breast-fed and formula-fed infants (2,8,9). In general, however, little is known about the bacterial dynamics in late infancy and early childhood, and the influence of partial breast-feeding has only been sparsely elucidated. To the best of our knowledge, no studies have been conducted to attempt to elucidate this in a prospective manner. Furthermore, in this period of life it is unknown to what extent diet may play a role. One recent intervention study comparing cow's-milk formula and infant formula with or without fish oil (FO) supplementation (rich in n-3 LCPUFA) reported effects of both cow's-milk consumption and FO supplementation among the cow's-milk consumers on the microbiota in 10-month-old infants (10).

Terminal restriction fragment length polymorphism (T-RFLP), a rapid and reproducible (11) DNA fingerprinting technique, has been shown to be efficient when comparing changes in the fecal microbiota of infants (12) and when monitoring the effect of nutritional interventions in adults (13). Using this technique, we have found effects of n-3 LCPUFA on the cecal bacteria in early-weaned piglets (unpublished data). The aim of the present study was to monitor the changes in the fecal bacterial profiles during the complementary feeding period, from 9 to 18 months of age, and in particular to investigate whether FO supplementation compared with sunflower oil (SO) would affect these changes in the gut microbiota differently. Because most children generally cease breast-feeding before 18 months of age, we also explored the influence of partial breast-feeding on the microbiota and whether this had an influence on the effect of the intervention.

Back to Top | Article Outline


Participants and Study Design

One hundred fifty-four healthy 9-month-old infants residing in the capital region of Copenhagen were randomly recruited from the National Danish Civil Registry and enrolled in this double-blinded randomized parallel intervention, which consisted of a daily supplement of either FO or SO for a period of 9 months. Eligible infants were healthy singletons, born ≥37th week of gestation with an appropriate weight for gestational age (14), 5-minute Apgar score >7, no prior dietary FO supplements, and no use of medications that could be expected to influence growth and/or food intake. Parents of all of the participants provided written informed consent and the study was approved by the scientific committees of the capital region of Copenhagen, Denmark (H-A-2007–0088).

Back to Top | Article Outline

Dietary Intervention, Randomization, and Group Allocation

Based on fatty acid analyses the intended dose of 5 mL/day would supply the infants with 1.6 g/day eicosapentaenoic acid (20:5n-3, EPA) + docosahexaenoic acid (22:6n-3, DHA) in the FO group and 3.1 g/day linoleic acid (18:2n-6, LA) in the SO group. The oils were given either with a spoon or mixed with infant foods. Infants had an equal probability of allocation to the 2 groups. The randomization list was developed by personnel not involved in the trial by use of a computer-generated randomization to random permuted blocks (15). The block sizes were 4, 6, 8, and 10 and varied randomly in order. Hence, ID numbers from 1 to 170 were assigned to 1 of the 2 intervention groups. For a given ID number, 10 dark 150-mL plastic bottles containing the intervention oils were subsequently coded and labeled with the ID number and supplement instructions. To minimize oxidation of the fatty acids, parents were instructed to keep all of the unopened bottles in the freezer and refrigerated once opened. The randomization list was kept in a sealed envelope in a safety box at the department. The investigators responsible for the initial contact with the infants and families allocated the next available number in the randomization list on entry into the trial, and 10 bottles of intervention oils (enough for the entire intervention period) were distributed after the baseline examination. Hence, trial investigators were blinded at all times to the participant allocation. Unblinding was performed only after completing all of the data analyses.

Back to Top | Article Outline

Compliance With the Intervention

Compliance was assessed by 2 objective measures. First, parents were requested to return all 10 of the supplied bottles at the 18-month examination and to report any spilling of the oils during the intervention. All of the bottles, including any leftovers, were weighed upon return and the weight was subtracted from the original weight of 10 full bottles. This amount was subsequently divided by the exact number of days a given participant had been in the study. Second, erythrocyte (red blood cell [RBC]) fatty acid composition was determined in the participants before and after the intervention. RBC PUFA composition is a well-known biomarker of long-term PUFA intake (16,17).

Back to Top | Article Outline

Examinations, Sample Collection, and Habitual Diet

All of the participants were examined both before and after the intervention, at 9 months ± 2 weeks and 18 months ± 4 weeks, respectively. Before both visits stool samples were collected and immediately stored in home freezers. On the day of examination the stool sample was transported on ice to the department and immediately transferred to a −80°C freezer until further processing. Parents were asked whether their child had been ill or had received any antibiotic treatment in the weeks before the collection. Stool samples reported to be collected within 2 weeks after use of antibiotics (5 at 9 months and 10 at 18 months) were excluded from the statistical analyses. At each visit, anthropometric measurements, including weight and length, were conducted by 2 trained assessors. Also, at both examinations, a 6-mL blood sample was taken by venipuncture, kept on ice, and separated into plasma, buffy coat, and RBC by centrifugation at 2300g for 10 minutes at 4°C. The fatty acid composition of the RBCs was determined as described previously (18). In brief, lipids were extracted by the Folch procedure (19) and subsequently methylated with BF3. Fatty acid methyl esters were analyzed on an Agilent Technologies (Santa Clara, CA) 6890N gas chromatograph equipped with an automatic sampler and identified by comparing the retention time with standards of known composition (Nu-Chek-Prep, Elysian, MN). Values are expressed as chromatogram area of the specific fatty acid relative to the total chromatogram area (FA%), which is essentially equivalent to weight%. Data on the habitual diet of the participants were obtained through filling of precoded food diaries in 7 consecutive days before both the first and the last examinations. In addition, detailed questionnaires on birth data (eg, mode of delivery) obtained from the hospital journal, infant diet (eg, breast-feeding), infant development, and health were completed at baseline. In between the visits, parents were asked to keep a journal recording infant illness and medication. Questionnaires covering the intervention period (eg, perceived adherence and infant diet and health) were also completed when the infants were 18 months old.

Back to Top | Article Outline

DNA Extraction and Amplification of 16 s rRNA Gene

DNA was extracted from the stool samples using the Maxwell 16 System (Promega, Madison, WI) and Tissue DNA Purification Kit (cat. no. AS1030) according to the manufacturer's instructions. The concentration of DNA in each sample was determined spectrophotometrically and subsequently adjusted to a concentration of 5 μg/mL DNA mL. The bacterial 16S rRNA genes were amplified using the 16S rRNA gene primers Eub-8fm (5′-AGAGTTTGATCMTGGCTCAG-3′) and Eub-926r (5′-CCGTCAATTCCTTTRAGTTT-3′) targeting most bacteria (20). The forward primer was 5′-labeled with 6-carboxyfluorescein (FAM). Each sample was run in duplicates in 2 separate polymerase chain reactions (PCRs). Hence, a total of four 50-μL replicates were made from 2 PCR mixtures containing 5 μL 10×Taq buffer, 0.5 μL of each primer, 4.5 μL MgCl2, 2.0 μL deoxyribonucleotide triphosphate, 0.5 μL Taq GOLD polymerase, and 35.0 μL nuclease-free water. PCR was run in a T3 Thermocycler (Biometra, Göttingen, Germany) under the following conditions: initial denaturation at 94°C for 6 minutes followed by 30 cycles of denaturation at 94°C for 45 seconds, annealing at 56°C for 45 seconds, and extension at 72°C for 2 minutes, with a final extension step at 72°C for 10 minutes. Amplified DNA (10 μL of the 50 μL) was verified by gel electrophoresis on 1.5% agarose gels and subsequently visualized after staining with ethidium bromide. The remaining PCR product from the duplicates within each separate amplification was pooled and purified using the High Pure PCR-Purification Kit (Roche Applied Science, Mannheim, Germany), according to the manufacturer's instructions and eluted in a final volume of 50-μL buffer.

Back to Top | Article Outline

Terminal Restriction Fragment Length Polymorphism

We used T-RFLP to study the fecal microbiota. Approximately 200 ng of each pooled PCR product was digested with 2.0 μL HhaI in the recommended enzyme buffer in a final volume of 20 μL at 37°C for 3 hours. The T-RF analysis was conducted using capillary electrophoresis on a 3130 Genetic Analyser (Applied Biosystems, Carlsbad, CA). Restriction fragments were run in duplicates of 2 μL purified digested PCR product mixed with 0.50 μL MegaBase ET900-R size standard (GE Health Care, Buckinghamshire, UK) and 10 μL formamide. Chromatograms were exported into the BioNumerics software package (Applied Maths, Sint-Martens-Latem, Belgium) and aligned using the internal size standard. Bands were identified using the autosearch function with a minimum profiling limit for each band of 5% of all of the bands within a given sample. Only bands present in both duplicates were accepted as bacterial fragments, and the duplicate with the lowest quality was subsequently discarded, and the other was used for the microbial profiling. The same bands classifier was used for all of the samples during analysis of the T-RFs. Finally, all T-RF lengths and intensities were imported into the statistical software programs Unscrambler version 9.8 (CAMO, Oslo, Norway) and Stata 11.0 (StataCorp, College Station, TX) for further statistical analyses and exploration. Putative identification of the major bacteria characterized by the T-RFs were predicted through in silico digest performed using the online Microbial Community Analysis III (21) with data from the Ribosomal Database Project (RDP [R10, U12] 322.864 Good Quality (>1200) 16SrRNA) and subsequent browsing in the National Center for Biotechnology Information taxonomy browser.

Back to Top | Article Outline


Dichotomous and continuous data on baseline characteristics were analyzed using χ2 tests and unpaired t tests, respectively. Group differences in the RBC fatty acid composition were compared by unpaired t test or Mann-Whitney U test and given in the text as either mean ± SD or median (25th–75th) as appropriate, unless otherwise noted. The assignment of the fragment sizes to length categories (base pairs) was performed by rounding to the nearest integer. In cases in which this led to identical length categories, the resulting bins were combined to a single category. The relative abundance of a T-RF in the samples was calculated as the intensity of a given T-RF divided by the total intensity of all T-RFs in the sample. Only those T-RFs with a relative abundance >1% were used in further analyses. This latter process was performed separately for data from 9 and 18 months, leading to a reduction in the number of T-RFs from 93 to 26 in 9-month samples and 95 to 24 in 18-month samples. Finally, to overcome variations potentially introduced by differences in the amount of DNA injected before the capillary electrophoresis, the method of standardizing the quantity of DNA introduced by Dunbar et al (22) was subsequently applied. This procedure did not affect the number of T-RFs with a relative abundance >1% in the 9-month samples but reduced T-RFs in 18-month samples from 24 to 22. The relative abundance of these most predominant T-RFs was subsequently calculated as described above and the resulting values are used in the statistical analyses. To evaluate changes over time and hence changes potentially affected by the intervention and other dietary influences, changes in the relative abundances of the T-RFs were calculated by subtracting the 9-month value from the value at 18 months. In cases in which a T-RF was not present in, for example, the outcome but in the baseline sample, the former was set to 0 (ie, no abundance). Data are not normally distributed and there are large individual variations in the abundance of the T-RFs. Therefore, all of the group comparisons and changes in the abundance of the T-RFs are statistically evaluated by the Mann-Whitney U test. We have, however, chosen to depict visually the T-RFs as mean values (without standard deviation [SDs]) because this gives the best impression of the bacterial dynamics. The Shannon-Weaver index of diversity (H′) based on all initial T-RFs was used to determine the diversity of the bacterial fragments. The index was calculated by the following equation:



where s is the number of species in the sample, and Pi is the relative abundance of bacterial fragment i in the sample. The index was not normally distributed, and for statistical evaluation, the Mann-Whitney U test was applied. Statistical significance was established at P ≤ 0.05.

We used principal component analysis (PCA) to explore whether there were any group differences in the composition of the overall microbial communities. This was done for both the T-RFs from 18-month samples and for the calculated changes in the relative abundance from 9 to 18 months separately. The T-RFs were standardized (centered and 1/SD) before the modeling phase to ensure that all of the bacterial fragments would equally influence the model. Possible outliers were inspected visually and by Hotelling T2.

Back to Top | Article Outline


The trial was conducted from January 2008 to March 2009. A total of 133 of the initial 154 participants completed the trial. Of these, we obtained stool samples from 132, and only these participants will be presented below. Seven in the SO group and 14 in the FO group left the study during the intervention. Half of the dropouts in the latter were related to the intervention oil. The characteristics of the study population are shown in Table 1. Overall, there were no differences between the intervention groups with respect to the habitual diet of the participants at 9 or 18 months in terms of energy intake and macronutrient composition. However, infants in the FO group had a slightly higher intake of PUFA than infants in the SO group at 9 months, and compared with children in the FO group those in the SO group received a slightly higher part of their daily energy from protein at 18 months of age. Approximately 55% of the infants were no longer breast-fed at the beginning of the intervention with no group difference (P = 0.61).



Based on the weight of the returned intervention oil bottles, the participants had an intake of 3.8 (3.2–4.2) g intervention oil per day (median [25th–75th]), with no group differences (P = 0.17). At 9 months of age, RBC content of n-3 LCPUFA and the ratio of total n-6/n-3 PUFA were similar in the 2 groups (9.7 ± 2.4 FA% and 3.8 ± 1.1 vs 9.7 ± 2.0 FA% and 3.9 ± 1.2 in the children of the FO group and SO group, respectively). Those who had ceased breast-feeding before the intervention had a lower content of n-3 PUFA in their RBC (9.3 ± 2.2 FA%) at baseline compared with those who were still breast-fed at 9 months of age (11.4 ± 1.9 FA%) (P < 0.0001). The intervention oils had a marked effect on changes in both n-3 LCPUFA content of RBC and the ratio of n-6/n-3 PUFA, with larger increases in n-3 LCPUFA in the FO group relative to that in the SO group (12.5 ± 0.7 FA% [mean ± SE] compared with 2.4 ± 0.4 FA%, P < 0.0001) and a larger decrease in the mean (± SE) ratio of n-6/n-3 PUFA (−2.5 ± 0.1 vs −0.8 ± 0.2, P < 0.0001). At the end of the intervention, the mean difference between groups in the RBC content of n-3 LCPUFA was 9.9 (95% CL 8.6–11.3) FA% and the mean difference in the RBC ratio of n-6/n-3 PUFA was 1.8 (1.5–2.0). The estimated oil consumption in the FO group correlated well with the incorporation of n-3 LCPUFA into the RBC (best for EPA: r = 0.6, P < 0.001). The RBC content of n-6 PUFA decreased in the FO group compared with the SO group (P < 0.001), with a total difference of 7.94 (6.1–9.8) FA% between groups.

Figure 1 shows the abundance at 9 and 18 months of the 29 T-RFs that had a relative abundance of >1% in the samples from the entire study population at either both or 1 of these ages. These T-RFs constituted 84% and 83% of the initial total intensity of all T-RFs in the samples at 9 and 18 months, respectively. Nineteen of the T-RFs were present at both sample times, whereas 3 were found only at 18 months and 7 only at 9 months. In general, there were some distinct changes in the predominating microbiota with age. Some of the relatively abundant T-RFs at 9 months disappeared or were markedly reduced (eg, bp371–373 and bp593–597), whereas other T-RFs, especially bp100 and bp102 (both likely to belong to the Bacteroidetes division) seemed to become dominant in the fecal samples at 18 months. When all of the initial T-RFs were included, the bacterial diversity, as estimated by the Shannon-Weaver index, was higher in the 18-month samples than in those from 9 months (2.40 [2.17–2.60] and 2.36 [2.02–2.54], median [25th–75th], P = 0.04, respectively).



Back to Top | Article Outline

Effects of the Intervention and Breast-feeding on the Microbiota

Breast-feeding clearly affected the relative abundances of individual T-RFs. Children who were breast-fed at 9 months had significantly lower abundance at 9 months of bp100 (P = 0.005) and higher levels of bp593 (P < 0.0001), bp595 (P < 0.0006), bp597 (P < 0.0015), and bp373 (P = 0.02) compared with children who were not breast-fed at 9 months. Furthermore, those who ceased breast-feeding during the intervention period had a larger increase in bp100 and larger decrease in both bp593 and bp373 during the intervention than children who had stopped breast-feeding before the initiation of the study (Fig. 1, Inset). At 18 months of age there were no group differences related to breast-feeding in these T-RFs. Breast-feeding practices did not affect the Shannon-Weaver index of bacterial diversity at either 9 or 18 months.

PCA analyses revealed no clear differentiation in the overall microbial community of the samples in relation to the intervention either in the 18-month samples or in the changes of the relative abundance of the T-RFs from 9 to 18 months. Conventional statistics (Mann-Whitney U tests) showed only 1 significant intervention group difference, namely in the change in bp93 during the intervention, which increased by 1.4[0–2.8]% and 0[−2.6–1.6]% (median [25th–75th]) in the FO group and SO group, respectively (Fig. 1, Inset). There were no group differences at 18 months.

Breast-feeding practices during the intervention seemed to modify the effect of the intervention on the changes in the most predominant individual bacterial fragments (Fig. 2). No effect was evident in the children who stopped breast-feeding during the intervention. In those with early breast-feeding cessation, FO tended to result in a larger increase in bp100 and a smaller increase in bp102 compared with the SO group (Fig. 2). These differences are supported further by the differences in the changes of BF versus non-BF within the oil groups—there was a significant difference in bp100 only in the SO group and a significant difference in bp102 only in the FO group. This was also evident among the non-BF group at 18 months, where there was a lower abundance of bp102 in the FO group (P = 0.003) and a higher abundance of bp189 (P = 0.05) compared with the SO group (data not shown).



The bacterial diversity was not affected by intervention, but at 18 months there was a just significant (P = 0.05) lower diversity in samples from FO children who had stopped breast-feeding before the study (2.31 [2.14–2.7]) compared with those who stopped breast-feeding in the course of the intervention (2.48 [2.24–2.62]), whereas there was no such difference between the earlier and later weaned children within the SO group.

Back to Top | Article Outline


The present study presents the first prospective investigation of the fecal microbiota from 9 to 18 months of age, and we show that the fecal microbiota, as assessed by DNA fingerprinting, is still developing in this period of life, presumably under the influence of both diet and age. FO supplementation had a significant influence on the changes in the microbiota compared with SO, but the effect was mainly observed among children who had stopped breast-feeding before the study. Breast-feeding in itself, even in late infancy, also had decisive effects on the fecal bacteria.

There is limited prior evidence of n-3 LCPUFA-mediated effects on bacteria in the gut. Early studies have shown that n-3 LCPUFA in vitro may impair the growth of Bacteroides thetaiotaomicron, but not that of Escherichia coli (23), and that they can modulate the growth and adhesion of different lactobacilli in vitro (24) and in vivo (25–27). In the present study, we could not demonstrate any effect of FO on the overall composition of the infant fecal bacteria based on PCA of the T-RFLP profiles, which is consistent with a recent study investigating the effect of adding n-3 LCPUFA to a standard feed in broiler chickens (28). In contrast to this, inclusion of FO in the diet has in steers been reported to have a major effect on the overall bacterial community and to decrease the number of bacterial bands detected by denaturing gradient gel electrophoresis (29). In accordance with this, we have found that 14 days of FO-rich formula compared with a SO-rich formula given to early-weaned piglets affected the overall bacterial composition in the caeca (unpublished data). Similarly, in 10-month-old infants, our group found that denaturing gradient gel electrophoresis profiles clustered differently in infants who received FO supplementation for 1 month compared with no supplementation among children concomitantly allocated to cow's-milk consumption (10). In the present study, the effect of the FO consumption may be less evident, but we did observe what seemed to be a shift in the age-induced increase in the dominant Bacteroidetes—from bp102 toward bp100—especially among children who were no longer breast-fed at 9 months.

We could speculate on a few reasons why the effect of FO was more pronounced among children who had been weaned before the study. One explanation could be that cessation of breast-feeding makes the microbiota open for new bacteria to establish because of the lack of the continuous supply of prebiotic oligosaccharides from breast milk, and possibly also milk-borne bacteria (30). Hence, the changes with breast-feeding cessation may simply have “diluted” the intervention effects in the children. Alternatively, because breast milk is also a good source of n-3 LCPUFA (31), another explanation could be that the children who had stopped breast-feeding at 9 months had a significantly lower n-3 PUFA status compared with those who were still breast-fed at that age, which should make it easier to detect an effect of the FO supplementation.

Although it is generally assumed that the bacterial community continue to develop and increase in complexity in this period of life (32), changes in the microbial communities from 9 to 18 months have not been investigated. Considering the lack of knowledge on gut bacteria in late infancy and early childhood, the present study adds to the existing knowledge in several ways with respect to both the dynamics of the gut microbiota and the effect of partial breast-feeding. In the present study, we observed a general shift in the T-RFLP profiles, which indicates that the microbiota is still developing from 9 months onward. Our data also seem to confirm a slight increase in bacterial complexity, because the Shannon-Weaver diversity index increased during the intervention period. Although this increase appeared to be almost negligible, it is worth noting that this occurred in spite of major bacterial groupings concomitantly becoming more dominant. This finding is in line with the literature (3,33) and may reflect the transition into the so-called “adult-like” microbiota, in which bacteria such as Bacteroides belonging to the Bacteroidetes division have been shown to become more abundant (3). Breast-feeding supplies the newborn with oligosaccharides on which certain gut bacteria thrive (34,35), and differences in the fecal microbiota of exclusively breast- and bottle-fed infants have previously been demonstrated (9,36). Our data suggest that even partial breast-feeding in late infancy plays a role in the gut bacteria composition. Because the T-RFLP pattern of non–breast-fed infants at 9 months is more similar to that at 18 months than to that of the partial breast-fed 9-month-olds, one could speculate that prolonged breast-feeding temporarily delays the intestinal microbial maturation. Cessation of breast-feeding during the course of the study seemed to induce a catch-up in the general trends of maturation that had already begun to manifest in the children who had stopped breast-feeding before the study. As a result, at 18 months there were no longer any differences between these 2 groups of children, which appears to be in line with results from a previous investigation that reported a more homogenous microbiota postweaning (37).

It should be mentioned that a number of statistical tests were performed and, although we have focused only on the major T-RFs in the profiles, we cannot rule out the possibility of any chance findings. Therefore, our results must be verified in future studies. Furthermore, as in any PCR-based technology, a minor weakness of the DNA fingerprinting method used in the present study is a possible biased representation of the microbial composition because the choice of PCR primer, which may influence the pattern (38). This means that the T-RFs in a given profile are not necessarily the true depiction of the entire bacterial population and their relative abundances in the sample. We did not a priori attempt to address questions related to specific groups of bacteria, but we were interested in monitoring changes in the microbiota over time and as influenced by the dietary intervention for which the method has been applied successfully in previous studies (13). Such relative changes in diversity are generally believed to be reliably reflected by this method (39).

In conclusion, we have shown that the fecal microbiota is still developing from 9 to 18 months of age, and that daily intake of FO compared with SO during this period significantly influenced changes in the abundance of major bacterial groups among children who had stopped breast-feeding. Our results furthermore show that breast-feeding in itself appeared to have a decisive effect on the microbial communities even at this late stage of infancy.

Back to Top | Article Outline


We gratefully acknowledge the contribution of all of the participating children and their families, those who assisted in the conduction of the study, and the biotechnicians for their technical assistance in collecting the data.

Back to Top | Article Outline


1. Gronlund MM, Lehtonen OP, Eerola E, et al. Fecal microflora in healthy infants born by different methods of delivery: permanent changes in intestinal flora after cesarean delivery. J Pediatr Gastroenterol Nutr 1999; 28:19–25.
2. Harmsen HJM, Wildeboer-Veloo ACM, Raangs GC, et al. Analysis of intestinal flora development in breast-fed and formula-fed infants by using molecular identification and detection methods. J Pediatr Gastroenterol Nutr 2000; 30:61–67.
3. Palmer C, Bik EM, DiGiulio DB, et al. Development of the human infant intestinal microbiota. PLoS Biol 2007; 5:1556–1573.
4. Kalliomaki M, Kirjavainen P, Eerola E, et al. Distinct patterns of neonatal gut microflora in infants in whom atopy was and was not developing. J Allergy Clin Immunol 2001; 107:129–134.
5. Wang M, Karlsson C, Olsson C, et al. Reduced diversity in the early fecal microbiota of infants with atopic eczema. J Allergy Clin Immunol 2008; 121:129–134.
6. Sjogren YM, Tomicic S, Lundberg A, et al. Influence of early gut microbiota on the maturation of childhood mucosal and systemic immune responses. Clin Exp Allergy 2009; 39:1842–1851.
7. Kalliomaki M, Collado MC, Salminen S, et al. Early differences in fecal microbiota composition in children may predict overweight. Am J Clin Nutr 2008; 87:534–538.
8. Penders J, Vink C, Driessen C, et al. Quantification of Bifidobacterium spp., Escherichia coli and Clostridium difficile in faecal samples of breast-fed and formula-fed infants by real-time PCR. FEMS Microbiol Lett 2005; 243:141–147.
9. Penders J, Thijs C, Vink C, et al. Factors influencing the composition of the intestinal microbiota in early infancy. Pediatrics 2006; 118:511–521.
10. Nielsen S, Nielsen DS, Lauritzen L, et al. Impact of diet on the intestinal microbiota in 10-month-old infants. J Pediatr Gastroenterol Nutr 2007; 44:613–618.
11. Li F, Hullar MAJ, Lampe JW. Optimization of terminal restriction fragment polymorphism (TRFLP) analysis of human gut microbiota. J Microbiol Methods 2007; 68:303–311.
12. Wang M, Ahrne S, Antonsson M, et al. T-RFLP combined with principal component analysis and 16S rRNA gene sequencing: an effective strategy for comparison of fecal microbiota in infants of different ages. J Microbiol Methods 2004; 59:53–69.
13. Li F, Hullar MAJ, Schwarz Y, et al. Human gut bacterial communities are altered by addition of cruciferous vegetables to a controlled fruit- and vegetable-free diet. J Nutr 2009; 139:1685–1691.
14. Greisen G, Michaelsen KF. Perinatal growth. A practical perinatal growth curve. Ugeskr Laeger 1989; 151:1813–1815.
15. Dallal G. Web site. Published August 2, 2007. Accessed January 22, 2008.
16. Brown AJ, Pang E, Roberts DCK. Persistent changes in the fatty-acid composition of erythrocyte-membranes after moderate intake of n-3 polyunsaturated fatty-acids—study design implications. Am J Clin Nutr 1991; 54:668–673.
17. Kuratko CN, Salem N. Biomarkers of DHA status. Prostaglandins Leukot Essent Fatty Acids 2009; 81:111–118.
18. Lauritzen L, Jorgensen M, Mikkelsen T, et al. Maternal fish oil supplementation in lactation: effect on visual acuity and n-3 fatty acid content of infant erythrocytes. Lipids 2004; 39:195–206.
19. Folch J, Lees M, Stanley GHS. A simple method for the isolation and purification of total lipides from animal tissues. J Biol Chem 1957; 226:497–509.
20. Molbak L, Thomsen LE, Jensen TK, et al. Increased amount of Bifidobacterium thermacidophilum and Megasphaera elsdenii in the colonic microbiota of pigs fed a swine dysentery preventive diet containing chicory roots and sweet lupine. J Appl Microbiol 2007; 103:1853–1867.
21. Shyu C, Soule T, Bent S, et al. MiCA: a Web-based tool for the analysis of microbial communities based on terminal-restriction fragment length polymorphisms of 16S and 18S rRNA genes. Microb Ecol 2007; 53:562–570.
22. Dunbar J, Ticknor LO, Kuske CR. Phylogenetic specificity and reproducibility and new method for analysis of terminal restriction fragment profiles of 16S rRNA genes from bacterial communities. Appl Environ Microbiol 2001; 67:190–197.
23. Thompson L, Spiller RC. Impact of polyunsaturated fatty-acids on human colonic bacterial metabolism—an in-vitro and in-vivo study. Br J Nutr 1995; 74:733–741.
24. Kankaanpaa PE, Salminen SJ, Isolauri E, et al. The influence of polyunsaturated fatty acids on probiotic growth and adhesion. FEMS Microbiol Lett 2001; 194:149–153.
25. Hekmatdoost A, Feizabadi MM, Djazayery A, et al. The effect of dietary oils on cecal microflora in experimental colitis in mice. Indian J Gastroenterol 2008; 27:186–189.
26. Bomba A, Nemcova R, Gancarcikova S, et al. The influence of omega-3 polyunsaturated fatty acids (omega-3 pufa) on lactobacilli adhesion to the intestinal mucosa and on immunity in gnotobiotic piglets. Berl Munch Tierarztl Wochenschr 2003; 116:312–316.
27. Ringo E, Bendiksen HR, Gausen SJ, et al. The effect of dietary fatty acids on lactic acid bacteria associated with the epithelial mucosa and from faecalia of Arctic charr, Salvelinus alpinus (L.). J Appl Microbiol 1998; 85:855–864.
28. Geier MS, Torok VA, Allison GE, et al. Dietary omega-3 polyunsaturated fatty acid does not influence the intestinal microbial communities of broiler chickens. Poult Sci 2009; 88:2399–2405.
29. Kim EJ, Huws SA, Lee MRF, et al. Fish oil increases the duodenal flow of long chain polyunsaturated fatty acids and trans-11 18:1 and decreases 18:0 in steers via changes in the rumen bacterial community. J Nutr 2008; 138:889–896.
30. Martin R, Jimenez E, Heilig H, et al. Isolation of Bifidobacteria from breast milk and assessment of the bifidobacterial population by PCR-denaturing gradient gel electrophoresis and quantitative real-time PCR. Appl Environ Microbiol 2009; 75:965–969.
31. Brenna JT, Varamini B, Jensen RG, et al. Docosahexaenoic and arachidonic acid concentrations in human breast milk worldwide. Am J Clin Nutr 2007; 85:1457–1464.
32. Adlerberth I, Wold AE. Establishment of the gut microbiota in Western infants. Acta Paediatr 2009; 98:229–238.
33. Kurokawa K, Itoh T, Kuwahara T, et al. Comparative metagenomics revealed commonly enriched gene sets in human gut microbiomes. DNA Res 2007; 14:169–181.
34. Coppa GV, Bruni S, Morelli L, et al. The first prebiotics in humans—human milk oligosaccharides. J Clin Gastroenterol 2004; 38:S80–S83.
35. Marcobal A, Barboza M, Froehlich JW, et al. Consumption of human milk oligosaccharides by gut-related microbes. J Agric Food Chem 2010; 58:5334–5340.
36. Hopkins MJ, Macfarlane GT, Furrie E, et al. Characterisation of intestinal bacteria in infant stools using real-time PCR and northern hybridisation analyses. FEMS Microbiol Ecol 2005; 54:77–85.
37. Magne F, Hachelaf W, Suau A, et al. A longitudinal study of infant faecal microbiota during weaning. FEMS Microbiol Ecol 2006; 58:563–571.
38. Schutte UME, Abdo Z, Bent SJ, et al. Advances in the use of terminal restriction fragment length polymorphism (T-RFLP) analysis of 16S rRNA genes to characterize microbial communities. Appl Microbiol Biotechnol 2008; 80:365–380.
39. Hartmann M, Widmer F. Reliability for detecting composition and changes of microbial communities by T-RFLP genetic profiling. FEMS Microbiol Ecol 2008; 63:249–260.

gut bacteria; infancy; n-3 polyunsaturated fatty acids

Copyright 2011 by ESPGHAN and NASPGHAN