Fallani M, Amarri S, Uusijarvi A, Adam R, Khanna S, Aguilera M et al.. Determinants of the human infant intestinal microbiota after the introduction of first complementary foods in infant samples from five European centres. Microbiol Read Engl 157: 1385-1392
Although it is well established that early infant feeding has a major influence on the establishment of the gut microbiota, very little is understood about how the introduction of first solid food influences the colonization process. This study aimed to determine the impact of weaning on the faecal microbiota composition of infants from five European countries (Sweden, Scotland, Germany, Italy and Spain) which have different lifestyle characteristics and infant feeding practices. Faecal samples were collected from 605 infants approximately 4 weeks after the introduction of first solid foods and the results were compared with the same infants before weaning (6 weeks of age) to investigate the association with determining factors such as geographical origin, mode of delivery, previous feeding method and age of weaning. Samples were analysed by fluorescence in situ hybridization and flow cytometry using a panel of 10 rRNA targeted group- and species-specific oligonucleotide probes. The genus Bifidobacterium (36.5 % average proportion of total detectable bacteria), Clostridium coccoides group (14 %) and Bacteroides (13.6 %) were predominant after weaning. Similar to pre-weaning, northern European countries were associated with a higher proportion of bifidobacteria in the infant gut microbiota while higher levels of Bacteroides and lactobacilli characterized southern European countries. As before weaning, the initial feeding method influenced the Clostridium leptum group and Clostridium difficile+Clostridium perfringens species, and bifidobacteria still dominated the faeces of initially breast-fed infants. Formula-fed babies presented significantly higher proportions of Bacteroides and the C. coccoides group. The mode of birth influenced changes in the proportions of bacteroides and atopobium. Although there were significant differences in the mean weaning age between countries, this was not related to the populations of bifidobacteria or bacteroides. Thus, although the faecal microbiota of infants after first complementary foods was different to that before weaning commenced, many of the initial influences on microbiota composition were still evident.
Determinants of the human infant intestinal
microbiota after the introduction of first
complementary foods in infant samples from five
3 Sergio Amarri,
Jose M. Vieites,
Jane A. Scott,
Christine A. Edwards
and the INFABIO team
Christine A. Edwards
Received 31 May 2010
Revised 31 January 2011
Accepted 15 February 2011
Institut National de la Recherche Agronomique, Unite
cologie et de Physiologie du Syste
Digestif, Jouy en Josas, France
Department of Paediatrics, Azienda Ospedaliera Santa Maria Nuova, Reggio Emilia, Italy
Department of Paediatrics, University of Modena and Reggio Emilia, Modena, Italy
Department of Clinical Science and Education, Sodersjukhuset, Sachs’ Children’s Hospital,
Karolinska Institutet, Stockholm, Sweden
Department of Paediatric and Adolescent Medicine, University Medical Centre, Mannheim, Germany
Human Nutrition Section, Division of Developmental Medicine, University of Glasgow, Yorkhill
Hospitals, Glasgow, UK
Department of Biochemistry and Molecular Biology, University of Granada, Spain
Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, Stockholm, Sweden
Department of Statistics and Modelling Science, University of Strathclyde, Glasgow, UK
Nutrition and Dietetics, School of Medicine, Flinders University, Adelaide, Australia
Although it is well established that early infant feeding has a major influence on the establishment
of the gut microbiota, very little is understood about how the introduction of first solid food
influences the colonization process. This study aimed to determine the impact of weaning on the
faecal microbiota composition of infants from five European countries (Sweden, Scotland,
Germany, Italy and Spain) which have different lifestyle characteristics and infant feeding
practices. Faecal samples were collected from 605 infants approximately 4 weeks after the
introduction of first solid foods and the results were compared with the same infants before
weaning (6 weeks of age) to investigate the association with determining factors such as
geographical origin, mode of delivery, previous feeding method and age of weaning. Samples
were analysed by fluorescence in situ hybridization and flow cytometry using a panel of 10 rRNA
targeted group- and species-specific oligonucleotide probes. The genus Bifidobacterium (36.5 %
average proportion of total detectable bacteria), Clostridium coccoides group (14 %) and
Bacteroides (13.6 %) were predominant after weaning. Similar to pre-weaning, northern
European countries were associated with a higher proportion of bifidobacteria in the infant gut
microbiota while higher levels of Bacteroides and lactobacilli characterized southern European
countries. As before weaning, the initial feeding method influenced the Clostridium leptum group
and Clostridium difficile+Clostridium perfringens species, and bifidobacteria still dominated the
faeces of initially breast-fed infants. Formula-fed babies presented significantly higher proportions
of Bacteroides and the C. coccoides group. The mode of birth influenced changes in the
Abbreviation: FISH, fluorescence in situ hybridization.
3Present address: Hospitex Diagnostics srl, via P. Lucchese 145, 50019 Sesto Fiorentino, Firenze, Italy.
Microbiology (2011), 157, 1385–1392 DOI 10.1099/mic.0.042143-0
2011 SGM Printed in Great Britain 1385
proportions of bacteroides and atopobium. Although there were significant differences in the
mean weaning age between countries, this was not related to the populations of bifidobacteria or
bacteroides. Thus, although the faecal microbiota of infants after first complementary foods was
different to that before weaning commenced, many of the initial influences on microbiota
composition were still evident.
Early events in the bacterial colonization of the human gut
may have long-term consequences for the development of
chronic diseases, including allergies. We have previously
confirmed, in a cohort of 606 European infants, that the
initial colonization process is heavily influenced by infant
diet (breast milk versus infant formula), mode of birth and
perinatal antibiotics, but that country of birth is also a
major factor (Fallani et al., 2010). This paper describes the
next critical stage in the development of the gut microbiota
which occurs during the process of weaning (complement-
ary feeding). The weaning period starts with the introduc-
tion of solid (non-milk) foods when infants are exposed for
the first time to many different non-digestible carbohy-
drates. A significant proportion of dietary starch also
escapes digestion because of immature pancreatic exocrine
function in young infants (Zoppi et al., 1972; Lebenthal
& Lee 1980; Parrett et al., 2000). These non-digested
carbohydrates enter the colon and influence the microbiota
by providing new substrates that may promote the survival
and dominan ce of species not supported by the limited
number of non-digestible carbohydrates in human milk
and infant formula. As weaning progresses, pancreatic
function, small intestinal absorption and colonic fermenta-
tion capacity also mature (Parrett & Edwards, 1997)
changing the charac teristics of the material reaching the
colon and also conditions in the fermentation chamber.
This is reflected in the microbiota composition and profiles
of short chain fatty acids (Midtvedt & Midtvedt, 1992).
There are, however, very little data on changes in the
composition of the intestinal microbiota during early
weaning (Satokari et al., 2002; Wang et al., 2004; Ahrne
et al., 2005) and how the influence of pre-weaning feeding
methods and perinatal events impact on the microbiota
after the introduction of comp lementary foods.
Therefore, the aim of this work was to characterize the faecal
microbiota of 605 infants from five European countries with
different lifestyle characteristics during early weaning and to
determine which pre-weaning influences were still active. As
531 infants provided samples both pre- and post-weaning,
the study also was able to monitor individual changes in the
microbiota and determine how much these were influenced
by the pre-weaning determinants.
Socio-demographic study. Four weeks after the introduction of
first solid foods (referred to hereafter as post-weaning), mothers were
requested to collect a faecal sample from their infant and complete
a questionnaire about diet and environmental exposures. Within
the European project INFABIO (http://www.gla.ac.uk/departments/
infabio/), post-weaning faecal samples were available from 605 infants
recruited in five different centres, from northern to southern Europe:
109 from Stockholm (Sweden), 149 from Glasgow (UK), 113 from
sseldorf (Germany), 121 from Reggio Emilia (Italy) and 113 from
Granada (Spain). The majority of these infants (531 of 605) had been
similarly studied at 6 weeks old (baseline measurement) and could
therefore provide paired data (Fallani et al., 2010). Of the total 605
infants, 59.l % had been fully breast-fed, 26.8 % fully formula-fed and
14.1 % had been fed a mixture of formula and breast milk at baseline.
The characteristics of the infants from each country that were used in
the analysis in this paper are summarized in Table 1. Ethical
permission was obtained from local ethics committees at each centre
and parents gave written informed consent.
Faecal sample collection, cell fixation and fluorescence in situ
Each faecal specimen was placed in a sterile
plastic box and maintained under anaerobic conditions at 4 uC using
an Anaerocult A (Merck) for a maximum of 4 h before processing for
cell fixation, as described previously (Rochet et al., 2001). Sample
fixation kits were provided at each collection site. Faeces were
homogenized by mechanical mixing for 3 min and aliquots of 1 g
(wet weight) were added to 9 ml anaerobic PBS. The suspension was
mixed to complete homogeneity in a 50 ml stoppered sterile glass jar
with a magnetic bar and aliquots of 0.2 ml suspension were added to
0.6 ml 4 % PFA in PBS. After overnight incubation at 4 uC, fixed
suspensions were stored at –70 uC, and shipped on dry ice for analysis
at a single location.
The FISH method was carried out as described previously (Rigottier-
Gois et al., 2003a, b; Fallani et al., 2010) using a panel of 10 group-
and species-specific probes covalently linked with indodicarbocyanine
(Cy5) at their 59 end to assess the microbiota composition (Fallani
et al., 2006). Enumeration of the different bacterial groups or species
was performed by FISH-FC by combining one specific probe labelled
with Cy5 together with EUB 338 FITC probe in the same tube.
Statistical analysis. Data are expressed as raw data means±SD of
the proportions of cells that hybridized with each of the 10
oligonuclotide probes relative to the total bacteria. A value of zero
was used where a microbial group was undetected below the
threshold of sensitivity of 0.4 % (Fallani et al., 2006). Initially, a
general linear model was used to assess the relationship between the
factors – centre, delivery method, feeding method before weaning,
infant and mother antibiotics, and outcome variables/bacterial groups
– while correcting for feeding method or delivery method.
Comparisons between the main effects were done using the
Bonferroni correction factor to compensate for multiple testing with
a significance level of 5 % to describe the associations. Variables that
were significant at the 5 % significance level in the univariate analyses
were then included in a multivariate analysis to identify the
independent factors associated with the outcome variable of interest.
Multivariate models included interaction terms for country and
delivery method, country and feeding method at baseline, country
and mother antibiotics, and country and antibiotic exposure of
M. Fallani and others
1386 Microbiology 157
infants, which were tested by using PASW (version 18). Non-significant
interactions were not included in the final models.
Similar models were used to assess changes in bacterial populations
between pre- and post-weaning periods. The Bonferroni correction
factor was used to compensate for multiple testing. Comparisons of
the overall changes in bacteria before and after weaning, without
consideration of influencing factors, were performed using paired
t-tests with Minitab (version 14) and a significance level of 5 %.
Microbiota of post-weani ng infants (n5605)
Nearly all post-weaning samples (98 %) could be analysed.
The remaining samples had a yield of RNA that was too
low. The predominant group detected in the post-weaning
infants considered together was Bifidobacterium with a
mean proportion of 36.5 % (±25.8), then C. coccoides
(14.0±15.0 %) and Bacteroides (13.6±15.8 %). Of the sum
of bacterial cells detected, 74.3± 18.9 % were accounted for
with the panel of 10 non-overlapping phylogenetic probes.
In the multivariate analysis, there were very few instances
of interactions between factors tested, e.g. centre of origin
and delivery method, and these have been taken into
account where appropriate.
Impact of centre of origin. As in our previous pre-weaning
study, there was a marked difference in the proportions of
different bacterial groups detected in infants from the
different countries (Table 2), especially for Bacteroides,
bifidobacteria, Atopobium and the Clostridium leptum
group, while less difference was observed for entero-
bacteria, Lactobacillus and Streptococcus groups, and sum of
the proportions. The country of birth did not significantly
affect the C. coccoides group, and C. difficile and C.
perfringens species. Samples from infants in Granada
contained significantly greater proportions of the
Lactobacillus group and significantly lower proportions of
bifidobacteria compared with samples from all others
centres (P,0.003) after correction for the feeding method.
Granada samples also contained greater proportions of
Bacteroides and C. leptum compared with all other centres
(P,0.001) except Du
sseldorf, greater proportions of entero-
bacteria compared with all centres (P,0.031) except Reggio
Emilia, and greater proportions of members of the
Streptococcus group compared with Du
In contrast, samples from Stockholm and Glasgow con-
tained significantly higher proportions of bifidobacteria
compared with those from all other centres (P,0.01).
sseldorf samples had significantly greater proportions of
C. leptum compared with all centres (P,0.048) except
Granada, and significantly greater proportions of Bacteroides
compared with Stockholm (P50.002) and Reggio Emilia
(P,0.001). Reggio Emilia samples contained a lower sum of
the proportions compared with Glasgow and Stockholm
(P,0.001), while samples from Stockholm also had a
significantly greater sum of proportions compared with
those from Du
Impact of pre-weaning feeding method. The pre-weaning
feeding method (breast-fed, formula-fed or mixed-fed)
influenced the relative proportions of Bacteroides, mem-
bers of the genus Bifidobacterium, C. coccoides, C. leptum
groups, and C. difficile + C. perfringens after correction for
centre effect. Infants breast-fed at 6 weeks but sampled post-
weaning presented significantly greater proportions of
bifidobacteria (40.7 % vs 29.2 %, P,0.001), and signifi-
cantly lower proportions of Bacteroides (12.1 % vs 17.1 %,
P50.004) and C. coccoides (11.1 % vs 15.9 %, P50.004) than
formula-fed infants. Infants who had been breast-fed at 6
weeks (before weaning) also presented significantly lower
proportions of C. coccoides post-weaning compared with
those who had been mixed-fed (11.1 % vs 18.0 %, P,0.001).
Impact of the delivery method. Compared with infants
born by Caesarean section, vaginal delivery infants (67 % of the
babies) had higher mean proportions of Bacteroides (16.1 % vs
10.1 %, P,0.001) after correction for feeding method.
Effect of antibiotics. The post-weaning microbiota of
newborns that had received antibiotics perinatally (only
6.5 % of 605 infants investigated) and infants whose
mothers received ant ibiotics were no different in com-
position from those not exposed to antibiotics.
Impact of weaning age. There was a significant difference
in the age of first introduction of solid foods between
countries (P,0.0001), with infants in Reggio Emilia weaned
earliest (16.9±3.6 weeks) and those in Du
sseldorf the latest
(22.4±4.6 weeks; Table 1). However, this did not influence
Table 1. Number and proportion of mothers and infants used in the analysis that belong to each parameter group studied
Not all mothers answered all questions.
Parameter group All centres Glasgow Granada Reggio Emilia Dusseldorf Stockholm
Caesarean section 205 (34.5 %) 43 (30.7 %) 18 (15.9 %) 58 (47.9 %) 46 (41.4 %) 40 (36.7 %)
Mother had antibiotics 175 (29.5 %) 54 (38.3 %) 11 (9.8 %) 44 (36.4 %) 32 (28.8 %) 34 (31.5 %)
Infant had antibiotics 37 (6.53 %) 7 (5.3 %) 6 (5.5 %) 16 (13.4 %) 7 (6.9 %) 1 (0.9 %)
Fully breast-fed at 6 weeks 312 (54.2 %) 58 (43.6 %) 51 (45.1 %) 61 (50.4 %) 58 (56.9 %) 84 (78.5 %)
Formula-fed at 6 weeks 163 (28.3 %) 66 (49.6 %) 41 (36.3 %) 24 (19.8 %) 30 (29.4 %) 2 (1.9 %)
Mixed-fed at 6 weeks 101 (17.5 %) 9 (6.8 %) 21 (18.6 %) 36 (29.8 %) 14 (13.7 %) 21 (19.6 %)
Infant microbiota and weaning
Table 2. Proportions of bacterial groups, genus and species detected in 605 infants’ faecal samples coming from five different European countries, assessed by FISH combined
with flow cytometry with a panel of 10 oligonucleotide probes
Values (raw data, mean±SD) indicate the percentage of cells hybridizing with the corresponding probe versus Eub338 probe. For the sum we included the proportion of cells hybridising with the
combined probe which measured C. perfringens and C. difficile together (done for all 605 samples) rather than the data where the two separate Cdif198 and Cperf191 probes measured the individual
species (done for only 26 samples). Target groups for each probe included (Fallani et al., 2006): Bif164, Bifidobacterium genus; Erec482, species of the genera Clostridium, Eubacterium, Ruminococcus
and Butyribivrio; Clep866, members of the genera Clostridium, Eubacterium, Ruminococcus and Anaerofilum plus Faecalibacterium prausnitzii cluster; Ato291, Atopobium cluster including members
of the genus Coriobacterium; Bac303, members of the genera Prevotella and Bacteroides; Enter1432, Enteric group including E. coli; Lab158, members of the genera Lactobacillus, Enterococcus,
Weissella and Pediococcus; Strc493, Streptococcus group including members of the genus Lactococcus; Cdif198, Clostridium difficile sp.; Cperf191, Clostridium perfringens sp.
City of origin (country) (n) Probe
Bif164 Erec482 Clep866 Ato291 Bac303 Enter1432 Lab158 Strc493 Cdif198* Cperf191* Cperf191+ Cdif198 Sum
Stockholm (SW) (109) 55.7±27.7 9.7±16.0 0.5±2.1 1.3±3.9 8.7±12.5 2.4±5.5 0.2±0.5 1.6±4.4 0.5±0.9 2.7±1.9 0.8±1.6 80.9±17.9
Glasgow (UK) (149) 41.6±24.3 15.0±15.3 0.7±2.5 2.8±6.2 12.8±13.6 2.0±5.7 0.3
±0.8 1.0±1.6 0.3±0.8 3.4±3.1 0.8±1.5 77.0±16.8
sseldorf (GE) (113) 29.9±20.4 15.4±14.2 2.6±5.2 1.4±4.0 17.4±17.2 3.0±4.9 0.6±1.6 0.9±1.8 0.3±0.5 1.8±1.3 0.9±1.9 72.0 ±19.7
Reggio Emilia (IT) (121) 37.2±21.1 12.9±12.9 0.6±2.0 1.3±3.8 8.8±11.0 4.2±5.6 0.7±0.9 1.4±2.8 0.7±0.8 1.4±1.3 0.6±0.9 67.7±18.1
Granada (ES) (113) 17.4±19.2 16.4±15.6 4.2±9.4 2.3±6.0 20.5±20.2 6.0±13.0 3.6±11.6 2.3±6.5 1.3±0.9 2.3±1.2 0.9±1.6 73.7±19.9
Total (605) 36.5±17.8 14.0±15.0 1.6±5.2 1.9±5.0 13.6±15.8 3.4±
7.6 1.0±5.2 1.4±3.8 0.5±0.8 2.4±2.0 0.8±1.5 74.3±18.9
*Cdif198 and Cperf191 were used separately only on 26 samples chosen among the samples with the highest proportions detected with the combined probe, which did not differentiate between the
species, in the different centres (nine in Du
sseldorf, six in Glasgow, three in Granada, three in Reggio Emilia and five in Stockholm).
M. Fallani and others
1388 Microbiology 157
the populations of bifidobacteria or Bacteroides after
correction for pre-weaning feeding method.
Changes in microbiota in infant s who provided
samples both before and after weaning (n5531)
The composition of the faecal microbiota of 531 infants who
gave samples before weaning (at 6 weeks of age) and 4 weeks
after the introduction of first solid foods (Fig. 1) showed that
weaning was accompanied by a significant decrease in
proportions of bifidobacteria (from 40.9 to 37.9 %,
P50.042), enterobacteria (from 7.3 to 3.1 %, P,0.001)
and C. difficile+C. perfringens species (from 3.2 to 0.8 %,
P,0.001), and a significant increase in proportions of the C.
coccoides (from 5.4 to 14 %, P,0.001) and C. leptum (from
0.4 to 1.6 %, P,0.001) groups.
Impact of centre of origin. Changes in bacterial
populations before and after weaning changed dependent
on the centre of origin after correction for feeding method.
This was particularly true for C. coccoides and C. leptum
groups and for enterobacteria, and to a lesser extent for the
Atopobium cluster and Streptococcus group. The country of
birth did not impact on the change in the other bacterial
groups. In Granada, the increase in the proportions of C.
coccoides and C. leptum groups (from 1.6 to 16.4 % and
from 1 to 4.2 %, respectively, P,0.01) and the decrease in
proportions of enterobacteria (from 19.3 to 6 %) after
weaning were significantly more pronounced compared
with the changes in any other centre (P,0.001). Samples
from Granada also had significantly different changes in
the proportion of the Atopobium cluster (from 0.7 to
2.3 %) before and after weaning compared with samples
from Stockholm (from 2.4 to 1.3 %, P50.02) and
sseldorf (from 2.2 to 1.4 %, P50.01 2), and of the
Streptococcus group (from 1.3 to 2.3 %) compared with
Reggio Emilia (from 2.4 to 1.4 %, P50.002) and Du
(from 1.3 to 0.9 %, P50.04 ). Finally, samples from Reggio
Emilia had significantly different changes in the proportion
of enterobacteria (from 8.3 to 4.2 %) before and after
weaning compared with samples from Stockholm (from
1.7 to 2.4 %, P,0.05) and Glasgow (from 3.5 to 2 %,
P,0.05). The centre of origin also affected the change in
sum of proportions (P,0.048).
Impact of the pre-weaning feeding method. Changes in
bacterial populations during weaning were affected by the
pre-weaning feeding method, after correction for country
effect; there was a significant difference (P,0.05) in
proportional changes only for the C. leptum group and
for C. difficile+C. perfringens species after correction for
centre. Infants who were formula-fed before weaning had a
significantly greater increase in the proportion of C. leptum
(from 0.5 to 2.3 %) after weaning compared with pre-
viously breast-fed babies (from 0.4 to 1.2 %, P50.028), and
a decrease in the proportion of C. difficile+C. perfringens
species (from 3.5 to 0.7 %) compared with mixed-fed
babies (from 1.5 to 1.0 %, P50.012).
Impact of the delivery method. The proportion of
Bacteroides before and after weaning (16.1 %) did not
change in babies born by vaginal delivery, while Bacteroides
increased after weaning (from 6.9 to 10.1) in infants born by
Caesarean section; this represents a significant difference
between the two delivery methods (P50.044). Infants born by
vaginal delivery had a decrease in the corresponding pro-
portions of Atopobium cluster before and after weaning while
babies born by Caesarean section had an increase (2.9 to 2.2 %
versus 0.8 to 1.5 %, P50 .044) after correction for feeding
method. The delivery method also affected the total sum of
detected bacterial groups with a more pronounced increase
for infants born by Caesarean section (67.6 to 73.4 % versus
75.5 to 75.9 %, P50.024) after correction for centre of origin.
Effect of antibiotic treatment. The evolution of
proportions of enterobacteria before and after weaning
Fig. 1. Composition of the faecal microbiota of
531 infants before weaning (6 weeks of age)
(black bars) and 4 weeks after the introduction
of first solid foods (grey bars). Values are mean
SD) proportions of the bacterial groups
quantified by FISH-flow cytometry. Asterisks
indicate significant differences between the two
periods (*P,0.05; **P,0.001).
Infant microbiota and weaning
was markedly affected by perinatal antibiotics (P50.003)
after correction for centre, but the greater decrease in
treated infants (16.6 to 3.6 %) than in untre ated ones (from
6.8 to 3.3 %) was associated with higher pre-weaning
proportions of enterobacteria. Proportions of bifido-
bacteria increased in newborns whose mothers received
antibiotics perinatally, while they decreased in infants born
from mothers without treatment (36.9 to 38.9 % versus
39.9 to 33.6 %, respectively, P50.033)
We have previously reported the factors which affect the
infant microbiota at 6 weeks of age (Fallani et al., 2010).
This study assessed the factors affecting the faecal
microbiota composition of European infants, during early
weaning, when more diverse sources of non-digestible
carbohydrates would be expected to reach the colon, by
analysing 605 faecal samples obtained 4 weeks after the
introduction of first solid foods. We also compared the
microbiota of infants with paired data at 6 weeks old and at
4 weeks post-weaning (n5531). At 4 weeks post-weaning,
Bifidobacteria still dominated the intestinal microbiota of
infants, but their proportions had significantly decreased
and the microb iota had diversified. Facultative anaerobes
significantly decreased, while proportions of strictly
anaerobic clostridia increased. Proportions of Bacteroides,
however, did not change and still represented one of the
most predominant groups in the infant gut microbiota
after weaning commenced. This agrees with most previous
studies. Martin et al. (2000) reported a more diverse
microbiota with higher proportions of Bacteroides and
clostridia (C. coccoides and C. leptum) after the introduc-
tion of solid foods. Wang et al. (2004) showed that
Enterobacteriaceae decreased while clostridia increased
after weani ng in two Swedish infants. Using real-time
PCR and Northern hybridization analyses of 40 infants
(0–24 months old), Hopkins et al. (2005) reported that
Bacteroides and Desulfovibrio numbers increased and
Enterococcus faecalis decreased in the 7–12 and 13–24
months age groups, and the C. coccoides group and
Faecalibacterium prausnitzii subgroup increased after 6
et al. (1998) reported that after weaning was
complete, the Bacteroides-Porphyromonas-Prevotella group
rapidly increased to comprise 30 % of 16S rRNA, while
Bifidobacterium rRNA was undetectable. In the present
study, C. difficile and C. perfringens also significantly
decreased after weaning in accordance with previous
observations (Stark & Lee, 1982; Benno et al., 1984;
Fallani et al., 2006).
The present study investigated the impact of geogr aphical
origin on the development of the gut microbiota of babies
born in different European countries before and after
weaning. The few previous studies reported here consid-
ered the microbiota composition for only one or two
countries and the majority investigated a small number of
bacterial groups (George et al., 1996; Gue
rin-Danan et al.,
1997; Sepp et al., 1997). We observed that the strong
association with geographical origin seen at 6 weeks of age
(a gradient of high numbers of bifidobacteria in northern
Europe to higher Bacteroides and lactobacilli in southern
Europe) (Fallani et al., 2010) still persisted 4 weeks
after first introduction of solid foods, particularly for
Bacteroides, bifidobacteria and the C. leptum group, despite
a high variability in the data (Table 2). It was not possible
to consider the impact of each parameter in each country
separately as the numbers in each centre were limited. One
month after weaning started, the faecal microbiota was still
quite different from that of adults (Lay et al., 2005), most
evidently for the Swedish infants where bifidobacteria
represented .50 % of total bacteria and the C. leptum
group was virtually undetected. This was also true for
Spanish infants, although their post-weaning microbiota
appeared clo sest to that of adults (lowest bifidobacteria
and highest Bacteroides,
C. coccoides and C. leptum
proportions). Thus, early diversification of the faecal
microbiota promotes the earlier acquisition of an ‘adult-
The influence of feeding method at 6 weeks persisted after
weaning started. Pre-weaning breastfeeding was still
associated with higher proportions of bifidobacteria, while
formula feeding was characterized by higher proportions of
Bacteroides and C. coccoides in accordance with previous
studies (Stark & Lee, 1982; Hopkins et al., 2005). We also
observed that during weaning, breastfeeding induced a
slower increase of C. leptum and a faster reduct ion of C.
difficile and C. perfringens species compared with formula
or mixed feeding. Satokari et al. (2002) reported similar
changes in the microbiota during weaning between breast-
fed and formula-fed infants. In a group of children at 6
months, where half had started to have the breast milk
supplemented with solid food, George et al. (1996)
observed that the number of enterococci and Bacteroides
increased, whereas enterobacteria and bifidobacteria
The influence of mode of birth also persisted after weaning.
The delay in faecal colonization and the low number of
Bacteroides in the Caesarean group (Fallani et al., 2010) was
still present after weaning. Vaginal delivery was associated
with higher proportions of Bacteroides post-weaning.
Infants born by Caesarean section had a higher increase
in the proportion of Bacteroides, Atopobium and total
nlund et al. (1999) studied 30 infants
delivered by Caesarean section who se mothers had receiv ed
antibiotics and 34 vaginally delivered infants, and observed
a delay in faecal colonization and low Bacteroides in the
Caesarean group during the first 6 months of life, even
after weaning. Bacteroides also showed a higher increase
across the weaning period (2–6 months) in the Caesarean
At 6 weeks there was a marked effect of antibiotic
treatment in infants or their mother perinatally (Fallani
et al., 2010). This effect disappeared after weaning started,
M. Fallani and others
1390 Microbiology 157
suggesting that the effect of early antibiotic administration
will fade rather than persist upon weaning even after the
treatment has stopped, as formerly reported (Mackie et al.,
In conclusion, in this large-scale longitudinal study,
important modifications occurred in the infant gut
microbiota after weaning which still remained ‘infant-like’
1 month after the introduction of the first solid food. The
effects of country of birth, feeding method and mode of
delivery seen at 6 weeks persisted after weaning, whereas
the impact of ant ibiotic treatment disappeared. One of the
most striking observations was that postnatal conditions
and especially early colonization affect the kinetics of
acquisition of a mature microbiota. The longer term
impact on the development of the microbiota, gut
physiology and maturation of the immune system needs
to be explored.
This study was carried out with financial support from the
commission of the European Communities, RTD programme
‘Quality of life and Management of Living Resources’, QLRT-2001
02606, ‘INFABIO’, coordinated by Professor Christine Edwards
(University of Glasgow, UK). It does not necessarily reflect the views
of the commission and in no way anticipates the commission’s future
policy in this area. Other members of the INFABIO team that were
involved in this project are Benoit Theze
Mohammed Khalid Khan (Glasgow University, UK), Francesca
Benatti and Rosa Maria De Mola (both Azienda Ospedaliera Santa
Maria Nuova and University of Modena and Reggio Emilia, Italy),
Horst Schroten (University Children’s Hospital, Du
Germany), and Oscar Thompson, Jose
Maldonado and Marco
Ferrer (all University of Granada, Spain).
, S., Lo
nnermark, E., Wold, A. E., Aberg, N., Hesselmar, B.,
Saalman, R., Strannega
rd, I. L., Molin, G. & Adlerberth, I. (2005).
Lactobacilli in the intestinal microbiota of Swedish infants. Microbes
Infect 7, 1256–1262.
Benno, Y., Sawada, K. & Mitsuoka, T. (1984). The intestinal
microflora of infants: composition of fecal flora in breast-fed and
bottle-fed infants. Microbiol Immunol 28, 975–986.
, J., Sghir, A., Hannequart-Gramet, G., Corthier, G. &
Pochart, P. (1998).
Design and evaluation of a 16S rRNA-targeted
oligonucleotide probe for specific detection and quantit ation
of human faecal Bacteroides populations. Syst Appl Microbiol 21,
Fallani, M., Rigottier-Gois, L., Aguilera, M., Bridonneau, C.,
Collignon, A., Edwards, C. A., Corthier, G. & Dore
, J. (2006).
Clostridium difficile and Clostridium perfringens species detected in
infant faecal microbiota using 16S rRNA targeted probes. J Microbiol
Methods 67, 150–161.
Fallani, M., Young, D., Scott, J., Norin, E., Amarri, S., Adam, R.,
Aguilera, M., Khanna, S., Gil, A. & other authors (2010).
microbiota of 6-week-old infants across Europe: geographic influence
beyond delivery mode, breast-feeding, and antibiotics. J Pediatr
Gastroenterol Nutr 51, 77–84.
George, M., Nord, K. E., Ronquist, G., Hedenstierna, G. & Wiklund, L.
(1996). Faecal microflora and urease activity during the first six
months of infancy. Ups J Med Sci 101, 233–250.
nlund, M. M., Lehtonen, O. P., Eerola, E. & Kero, P. (1999). Fecal
microflora in healthy infants born by different methods of delivery:
permanent changes in intestinal flora after cesarean delivery. J Pediatr
Gastroenterol Nutr 28, 19–25.
rin-Danan, C., Andrieux, C., Popot, F., Charpilienne, A.,
Vaissade, P., Gaudichon, C., Pedone, C., Bouley, C. & Szylit, O.
Pattern of metabolism and composition of the fecal
microflora in infants 10 to 18 months old from day care centers.
J Pediatr Gastroenterol Nutr 25, 281–289.
Hopkins, M. J., Macfarlane, G. T., Furrie, E., Fite, A. & Macfarlane, S.
(2005). Characterisation of intestinal bacteria in infant stools using
real-time PCR and northern hybridisation analyses. FEMS Microbiol
Ecol 54, 77–85.
Lay, C., Sutren, M., Rochet, V., Saunier, K., Dore
, J. & Rigottier-Gois, L.
Design and validation of 16S rRNA probes to enumerate
members of the Clostridium leptum subgroup in human faecal
microbiota. Environ Microbiol 7, 933–946.
Lebenthal, E. & Lee, P. C. (1980). Development of functional
responses in human exocrine pancreas. Pediatrics 66, 556–560.
Mackie, R. I., Sghir, A. & Gaskins, H. R. (1999). Developmental
microbial ecology of the neonatal gastrointestinal tract. Am J Clin
Nutr 69, 1035S–1045S.
Martin, F., Savage, S. A. H., Parrett, A. M., Gramet, G., Dore, J. &
Edwards, C. A. (2000).
Investigation of bacterial colonization of the
colon in breast-fed infants using novel techniques. Proc Nutr Soc 59,
Midtvedt, A. C. & Midtvedt, T. (1992). Production of short chain fatty
acids by the intestinal microflora during the first 2 years of human
life. J Pediatr Gastroenterol Nutr 15 , 395–403.
Parrett, A. M. & Edwards, C. A. (1997). In vitro fermentation of
carbohydrate by breast fed and formula fed infants. Arch Dis Child 76,
Parrett, A. M., Khanna, S. & Edwards, C. A. (2000). Excretion of faecal
starch and fat in breast-fed and formula-fed infants during weaning: a
longitudinal study. Proc Nutr Soc 59, 64A.
Rigottier-Gois, L., Bourhis, A. G., Gramet, G., Rochet, V. & Dore
(2003a). Fluorescent hybridisation combined with flow cytometry and
hybridisation of total RNA to analyse the composition of microbial
communities in human faeces using 16S rRNA probes. FEMS Microbiol
Ecol 43, 237–245.
Rigottier-Gois, L., Rochet, V., Garrec, N., Suau, A. & Dore
, J. (2003b).
Enumeration of Bacteroides species in human faeces by fluorescent
in situ hybridisation combined with flow cytometry using 16S
rRNA probes. Syst Appl Microbiol 26, 110–118.
Rochet, V., Rigottier-Gois, L., Beguet, F. & Dore, J. (2001).
Composition of human intestinal flora analysed by fluorescent in
situ hybridisation using group-specific 16S rRNA-targeted oligonu-
cleotide probes. Genet Sel Evol 33, S339–S352.
Satokari, R. M., Vaughan, E. E., Favier, C. F., Dore, J., Edwards, C.
& Vos, W. M. (2002). Diversity of Bifidobacterium and Lacto-
bacillus spp. in breast-fed and formula-fed infants as a ssessed by
16S rDNA sequence differences. Micro b Ecol Health Dis 14,97–
Sepp, E., Julge, K., Vasar, M., Naaber, P., Bjo
rksten, B. & Mikelsaar, M.
Intestinal microflora of Estonian and Swedish infants. Acta
Paediatr 86, 956–961.
Stark, P. L. & Lee, A. (1982). The microbial ecology of the
large bowel of breast-fed an d f ormula-fed infants during the first
year of life. JMedMicrobiol15, 189–203.
Infant microbiota and weaning
Wang, M., Ahrne
, S., Antonsson, M. & Molin, G. (2004).
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. JMicrobiolMethods59, 53–
Zoppi, G., Andreotti, G., Pajno-Ferrara, F., Njai, D. M. & Gaburro, D.
(1972). Exocrine pancreas function in premature and full term neonates.
Pediatr Res 6, 880–886.
Edited by: H. J. Flint
M. Fallani and others
1392 Microbiology 157