Oral microbiome in older adults with mild cognitive impairment

Abstract

The association between the oral microbiome and mild cognitive impairment (MCI) remains unclear. This study aimed to investigate such an association among Chinese older adults. Participants without dementia were recruited from the community. A battery of neuropsychological tests was administered to evaluate the cognitive function. The diagnosis of MCI was based on Peterson's criteria. The non-stimulated saliva was collected to extract sequences of the oral microbiome. Forty-seven MCI and 47 cognitively normal participants were included. There was significant difference in alpha diversity and insignificant difference in beta diversity between the two groups of participants. Compared with the cognitively normal group, Gemella haemolysans and Streptococcus gordonii were two significantly decreased species while Veillonella unclassified_Veillonella and Fusobacterium sp._HMT_203 were two significantly increased species in the MCI group. The richness of Gemella haemolysans presented the best discriminate value for MCI with the AUC (Area Under Curve) of 0.707, a cut-off value of 0.008 for relative abundance, the sensitivity of 63.8% and specificity of 70.2%. The dysbiosis of oral microbiome and relative abundance of Gemella haemolysans was significantly associated with MCI. Further studies were needed to develop new treatment strategies targeting the oral microbiome for cognitive impairment.

Keywords: 16S rDNA; Oral microbiome; bioinformatic analysis; mild cognitive impairment; older adults.

Figure 1.

Taxonomic distribution of the top 10 relative abundant phyla in the participants with MCI and normal cognition.

Figure 2.

Taxonomic distribution of the top 10 relative abundant genera in the participants with MCI and normal cognition.

Figure 3.

Principal component analysis showed different beta diversities indicating different microbiome structure between the MCI and cognitively normal groups.

Figure 4.

Principal coordinates analysis showed different beta diversities indicating different microbiome structure between the MCI and cognitively normal groups.

Figure 5.

Upgma(unweighted pair-group method with arithmetic means) showed different beta diversities indicating unsimilar microbiome structure between the MCI and cognitively normal groups.

Figure 6.

Boxplot of the top 10 abundant differentiated genera in the MCI and cognitively normal groups.

Figure 7.

Legends showed composition of different genera of microbiome between the participants with MCI and normal cognition using LEfSe analysis.

Figure 8.

Rocs of Gemella haemolysans (AUC:0.707), Streptococcus gordonii (AUC:0.703), Veillonella unclassified_veillonella (AUC:0.684) and Fusobacterium sp._hmt_203 (AUC:0.620) for the discrimination of MCI.