Oral microbiomes of patients with infective endocarditis (IE): a comparative pilot study of IE patients, patients at risk for IE and healthy controls

Figures & data

Table 1. Subjects’ demographics and dental clinical characteristics for patients at moderate risk for IE who had IE or did not, and healthy controls.

Criteria	HC^a	DC^b	IE^c	HC vs. DC	HC vs. IE	DC vs. IE
		data			p-values^d
Subject Count (M/F)	37 (11/26)	28 (14/14)	9 (5/4)	0.1597 USD	0.2852 USD	1.00 USD
Age:				0.007*	0.250*	0.366*
Mean	52.6	62.3	58.4
Median	53	65	57
SD	14.4	15.3	13.2
Range	24–84	25–87	33–77
Ethnicity Count:
M: C/AA/O	7/0/4	10/4/0	5/0/0	0.0975*	0.1814*	0.1814*
F: C/AA/O	21/2/3	8/6/0	4/0/0	0.0591*	0.1003*	0.1814*
Whole mouth average PD (SD)	NC	2.25 (± 0.58)	2.36 (±0.76)			0.449*
Whole mouth average CI (SD)	NC	0.34 (± 0.62)	0.63 (± 0.81)			0.163*
Oral sample combinations:
BST	111	84	27
sub-g&supra-g	NC	54	16
sub-g	NC	28	8

^aHealthy control (HC) group

^bNon-infective endocarditis (IE) disease control (DC) group consisting of individuals at moderate risk for IE per AHA criteria, who did not have IE

^cIndividuals at moderate risk for IE who developed IE

^dp-values of group comparisons were obtained using either Chi-squared test or Wilcoxon rank sum test

M is Male; F is female; C is Caucasian; AA is African American; O is other; PD is pocket depth; CI is calculus index; SD is standard deviation; BST is buccal mucosa, saliva, and tongue; sub-g&supra-g is subgingival and supragingival plaque; sub-g is subgingival plaque alone; NC is not collected.

Statistical analyses were completed in R:

^$Chi-squared test p-value

*Wilcoxon rank sum test p-value

Figure 1. Analytical design.

Table 2. Alpha-diversity analyses show significant differences for HC vs. IE and DC vs. IE comparisons.

Comparison^a	Min^b	Max^c	Avg^d	Stdev^e	p-value^f
HC vs DC BST
HC	12	147	66.5	27.3	0.819
DC	11	181	60.6	32.1
HC vs IE BST
HC	12	147	66.5	27.3	5.73x10^−07
IE	3	82	25.4	20.2
DC vs IE BST
DC	11	181	60.6	32.1	1.76x10^−06
IE	3	82	25.4	20.2
DC vs IE sub-g &supra-g
DC	22	186	81.9	34.4	4.22x10^−05
IE	13	133	44	33.6
DC vs IE sub-g
DC	26	186	87.9	34.3	0.003
IE	22	100	47.5	29.9

^aComparisons consisted of healthy controls (HC), disease controls (DC), and/or infective endocarditis (IE) patient groups with sample data combinations of oral swabs of buccal mucosa (B) and tongue (T) with saliva (S) or data combinations of supra- and subgingival plaque (supra-g&sub-g) or subgingival plaque (sub-g) alone

^bMinimum number of species detected per sample

^cMaximum number of species detected per sample

^dAverage number of species detected per sample

^eStandard deviation of species detected per sample

^fWilcoxon rank sum test significance (α = 0.05)

Table 3. PERMANOVA p-values for healthy controls, non-IE disease controls, and IE patient comparisons show significant beta-diversity changes.

Comparison^a	Disease status p-value^b	Site p-value^c
HC vs DC^d
BST	0.0008	0.0001
HC vs IE^e
BST	0.0001	0.0001
DC vs IE^f
BST	0.0001	0.0001
sub-g&supra-g	0.0001	0.5421
sub-g	0.0002	NA

^aComparisons of healthy controls (HC), non-IE disease controls (DC), and infective endocarditis (IE) patients

^bDisease status Monte-Carlo corrected p-value

^cSite Monte-Carlo corrected p-value

Sample site data combinations included buccal (B), saliva (S), tongue (T), subgingival plaque (sub-g), and/or supragingival plaque (supra-g). PERMANOVA’s were completed using relative abundance square root transformed data and a resemblance S17 Bray-Curtis similarity matrix. PERMANOVA’s were run using type III partial sum of squares, permutation of residuals under a reduced model, with 9,999 permutations, and a Monte-Carlo corrected significance p-value < 0.05. NA: not applicable

Figure 2. Alpha- and beta-diversity results. Alpha-diversity boxplots of IE, HC, and DC groups. Principal Coordinates Analysis (PCoA) of beta-diversity comparisons. a. Boxplots showing the total species (top left), species richness (top right), Shannon diversity indices (bottom left), and Simpson diversity indices (bottom right) of oral samples taken from the buccal mucosa (B), saliva (S), and tongue (T) (BST sample data combination) for patients diagnosed with infective endocarditis (IE; red), non-IE disease controls (DC; orange), and healthy controls (HC; green). b. Principal coordinates analysis (PCoA) of relative abundance data for sample site data combinations of buccal mucosa (B), saliva (S), and tongue (T) combination for a. healthy controls (HC) (green triangles) vs. infective endocarditis (IE) patients (red squares), b. HC vs. non-IE disease controls (DC) (orange circles), and c. DC vs. IE patients. Other comparisons included d. subgingival plaque (sub-g) data combined with supragingival plaque (supra-g) or e. sub-g data alone for DC vs. IE patients.

Figure 3. LEfSe histograms comparing HC, IE, and DC groups. HC vs. IE. HC vs. DC. DC vs. IE. Linear discriminant analysis (LDA) effect size (LEfSe) of healthy controls (HC), non-IE disease controls (DC), and infective endocarditis patients (IE) based on relative abundance data. Comparisons HC vs. IE (a), HC vs. DC (b) and DC vs. IE (c) were performed. Samples sites included buccal mucosa (B), saliva (S), tongue (T), subgingival plague (sub-g), and/or supragingival plaque (supra-g). LEfSe input consisted of ‘Diagnosis’ (disease status HC, DC, and/or IE) as LEfSe ‘Class’ option and ‘Subject’ as LEfSe ‘Subject’ option. Using the all-against-all strategy, the factorial Kruskal-Wallis and pairwise Wilcoxon signed rank test were set at a Monte-Carlo significance of α = 0.05 to calculate LDA scores. The log LDA was set at a threshold of > 4.0 for all comparisons other than HC vs. DC which was set at a threshold > 2.0. An asterisk (*) denotes significance of Mann-Whitney U-test. *< 0.05; **< 0.01; ***< 0.001.

Figure 3. (Continued.)

Figure 4. Receiver Operating Characteristic (ROC) curve analysis. DC vs. IE BST. HC vs. IE BST. DC vs. IE sub-g&supra-g. DC vs. IE sub-g.Receiver Operating Characteristic (ROC) curves based on relative abundance data were determined for each comparison of healthy controls (HC), non-IE disease controls (DC), and/or infective endocarditis patients (IE) groups for the sample data combinations of buccal mucosa (B), saliva (S), tongue (T), subgingival plaque (sub-g) and/or supragingival plaque (supra-g). No species were found to have an area under the curve > 0.6 for the HC vs. DC BST comparison. Mean relative abundance changes of each species probe in the form of fold changes (FC) are shown for each group. The (>) sign is used to depict the group with highest abundance on average. The criterion is the cut-off point selected to discriminate between two populations.