J Dent Res 90(10)2011 Mode of Delivery and Oral Microbiota 1187 Table 3. Variable Importance (P) for Bacteria Associated with Mode of Delivery from PLS-DA Multivariate Modeling Associated with Vaginal Delivery Associated with Caesarian Section Delivery acterial Group Bacterial grou 1.50 Slackia 3.26 Cardiobacterium hominis 1.44 Veillonella sp. EF509966(intestinal isolate 1.54 Prevotella Cluster P 1.43 Veillonella atypica, Veillonella parvula 1.51 Prevotella loescheii, Prevotella sp HOT 472 1.43 Streptococcus parasanguinis I and ll 1.34 Eubacterium[11][G-7l yurii 1. 43 Veillonella parvula 1.23 Catonella morbi, Catonella sp. HOT 164 43 Gemella haemolysin IM7[G-1]sp HOT 347, TM7[G-2] sp. HOT 350 1.43 Gemella morbillorum 1.26 1.41 Streptococcus australis 1.20 Leptotrichia buccalis, Leptotrichia goodfellowi, Sneath Bifidobacterium animalis nimalis. Bifidobacterium 1.13 lactis Bacteroidetes phylum 1.28 Streptococcus cristatus Kingella oralis, Eikenella sp HOT 0O 1.26 Selenomonas noxia sp HOT 326 Haemophilus sp HOT 035, HOT 036 1.25 Selenomonas sputigena, Selemonas sp HOT 143 1.03 1. 18 Aggregatibacter segnis, Aggregatibacter sp HOT 512 1.0 Prevotella nigrescens 1.05 Campylobacter concisus .03 Corynebacterium matruchotii 05 Streptococcus sp HOT 070, 071 Neisseria elong granulicatella elegans Capnocytophaga sputig streptococcus mutans Prevotella clusterⅣ Prevotella melaninogenica, Prevotella histicola Results are shown for a basic model including all infants with HOMIM microarray data ( n= 63)and (in bold) a second model restricted to breas 0435), length at birth, gestational weeks, and town of residence were associated with mode of birth (P 2 1.0), whereas in the PLS.- A model tricked to breast fed infants and infants born in or after gestational week 3(n=42)(model R2=0.693, Q2=0.496), only length at 3 mos was influential in addition to bacteric 'Unnamed taxa are identified by their Human Oral Taxon(HoT) number from HOMD(Dewhirst et al., 2010) Lactobacillus cluster I (probe w94)targets L. casei, L. paracasei, and L rhamnosus clustering and for identifying the variables characterizing clus- delivery group. S. exigua is a Gram-positive, strictly anaerobic, ers. With x-variables that were scaled to unit variance, and saccharolytic species that has been isolated in root canal infec cross-validation of the explanatory capacity to account for over- tions, periodontitis, extra-oral surgical wounds, and intestinal interpretation, there was good power to discriminate between abscesses(Abiko et al., 2010: Kim et aL., 2010). While it is not modes of delivery. The results were also stable after the model known why S. exigua was detected in the C-section but not in as restricted to infants born in wk 37 or later and those being the vaginal delivery group, it seems possible that the more breast-fed. With PLS-DA, several bacterial species differed diverse microbial biofilm of vaginally delivered infants could between infants based on their delivery method. The biological suppress establishment of this periodontitis-associated species nce of the species evaluated, and dete The association between S. exigua and mode of delivery requires species in the mouth in early infancy of infants delivered by further investigation. C-section, compared with those delivered vaginally, however, is In conclusion, the present study indicates a different coloni- unknown and will require longitudinal evaluation. It is neverthe- zation pattern in the oral cavity between three-month-old infants less notable that a greater microbial diversity in the intestine and delivered vaginally and those delivered by Caesarian section in the mouth has been reported to be associated with health The reasons for the differences are unknown, as is whether (Marsh, 2006: Preza et al., 2008; Sjogren et al, 2009; Turnbaugh these differences have long-term impact on the oral or general et aL, 2009: Luoto et al., 2011) health of the child. Possible reasons for differences will likely a difference based on mode of delivery was detection of include the relative influence of host receptor and mucosal and Slackia exigua in over 76% of the C-section infants(all with saliva immune phenotypes, and interactions with environmental HOMIM scores 22)compared with non-detection in the vaginal exposuresJ Dent Res 90(10) 2011 Mode of Delivery and Oral Microbiota 1187 clustering and for identifying the variables characterizing clus￾ters. With x-variables that were scaled to unit variance, and cross-validation of the explanatory capacity to account for over￾interpretation, there was good power to discriminate between modes of delivery. The results were also stable after the model was restricted to infants born in wk 37 or later and those being breast-fed. With PLS-DA, several bacterial species differed between infants based on their delivery method. The biological significance of the species evaluated, and detection of fewer species in the mouth in early infancy of infants delivered by C-section, compared with those delivered vaginally, however, is unknown and will require longitudinal evaluation. It is neverthe￾less notable that a greater microbial diversity in the intestine and in the mouth has been reported to be associated with health (Marsh, 2006; Preza et al., 2008; Sjögren et al., 2009; Turnbaugh et al., 2009; Luoto et al., 2011). A difference based on mode of delivery was detection of Slackia exigua in over 76% of the C-section infants (all with HOMIM scores ≥ 2) compared with non-detection in the vaginal delivery group. S. exigua is a Gram-positive, strictly anaerobic, asaccharolytic species that has been isolated in root canal infec￾tions, periodontitis, extra-oral surgical wounds, and intestinal abscesses (Abiko et al., 2010; Kim et al., 2010). While it is not known why S. exigua was detected in the C-section but not in the vaginal delivery group, it seems possible that the more diverse microbial biofilm of vaginally delivered infants could suppress establishment of this periodontitis-associated species. The association between S. exigua and mode of delivery requires further investigation. In conclusion, the present study indicates a different coloni￾zation pattern in the oral cavity between three-month-old infants delivered vaginally and those delivered by Caesarian section. The reasons for the differences are unknown, as is whether these differences have long-term impact on the oral or general health of the child. Possible reasons for differences will likely include the relative influence of host receptor and mucosal and saliva immune phenotypes, and interactions with environmental exposures. Table 3. Variable Importance (VIP) for Bacteria Associated with Mode of Delivery from PLS-DA Multivariate Modeling Associated with Vaginal Delivery Associated with Caesarian Section Delivery Bacterial Group1 VIP Bacterial Group1 VIP Streptococcus sanguinis 1.50 Slackia exigua 3.26 Cardiobacterium hominis 1.50 Lactobacillus Cluster I3 1.85 Streptococcus anginosus, Streptococcus gordonii 1.44 Veillonella sp. EF509966 (intestinal isolate) 1.54 Prevotella Cluster I2 1.43 Veillonella atypica, Veillonella parvula 1.51 Prevotella loescheii, Prevotella sp. HOT 472 1.43 Streptococcus parasanguinis I and II 1.34 Eubacterium [11] [G-7] yurii 1.43 Veillonella parvula 1.23 Catonella morbi, Catonella sp. HOT 164 1.43 Gemella haemolysans 1.32 TM7 [G-1] sp. HOT 347, TM7 [G-2] sp. HOT 350 1.43 Gemella morbillorum 1.26 Haemophilus parainfluenzae 1.41 Streptococcus australis 1.20 Leptotrichia buccalis, Leptotrichia goodfellowii, Sneathia sanguinegens 1.31 Bifidobacterium animalis ss. animalis, Bifidobacterium lactis 1.13 Bacteroidetes phylum 1.28 Streptococcus cristatus 1.08 Campylobacter gracilis 1.27 Kingella oralis, Eikenella sp. HOT 009 1.03 Capnocytophaga granulose, Capnocytophaga sp. HOT 326 1.26 Selenomonas noxia 1.03 Haemophilus sp. HOT 035, HOT 036 1.25 Selenomonas sputigena, Selemonas sp. HOT 143 1.03 Kingella oralis 1.18 Aggregatibacter segnis, Aggregatibacter sp. HOT 512 1.03 Prevotella nigrescens 1.05 Campylobacter concisus 1.03 Corynebacterium matruchotii 1.05 Streptococcus sp. HOT 070, 071 1.02 Neisseria elongata 1.05 Granulicatella elegans 1.02 Fusobacterium periodontium 1.05 Capnocytophaga sputigena 1.05 Streptococcus mutans 1.04 Prevotella Cluster IV 1.01 Prevotella melaninogenica, Prevotella histicola 1.01 Results are shown for a basic model including all infants with HOMIM microarray data (n = 63) and (in bold) a second model restricted to breast￾fed infants and those born in gestational week 37 or later (n = 42). In the PLS-DA model including all 63 infants (model R2 = 0.620, Q2 = 0435), length at birth, gestational weeks, and town of residence were associated with mode of birth (VIP ≥ 1.0), whereas in the PLS-DA model restricted to breast-fed infants and infants born in or after gestational week 3 (n = 42) (model R2 = 0.693, Q2 = 0.496), only length at 3 mos was influential in addition to bacteria. 1 Unnamed taxa are identified by their Human Oral Taxon (HOT) number from HOMD (Dewhirst et al., 2010). 2 Prevotella cluster I (probe Y65) targets P. loescheii, Prevotella spp clone GU027, Prevotella spp strain C3MKM081, and Prevotella spp strain TFI B31FD (HOT numbers 317, 472, 658). 3 Lactobacillus cluster I (probe W94) targets L. casei, L. paracasei, and L. rhamnosus