Salivary inflammatory markers and microbiome in normoglycemic lean and obese children compared to obese children with type 2 diabetes

Abstract

Background: There is emerging evidence linking diabetes with periodontal disease. Diabetes is a well-recognized risk factor for periodontal disease. Conversely, pro-inflammatory molecules released by periodontally-diseased tissues may enter the circulation to induce insulin resistance. While this association has been demonstrated in adults, there is little information regarding periodontal status in obese children with and without type 2 diabetes (T2D). We hypothesized that children with T2D have higher rates of gingivitis, elevated salivary inflammatory markers, and an altered salivary microbiome compared to children without T2D.

Methods: Three pediatric cohorts ages 10-19 years were studied: lean (normal weight-C), obese (Ob), and obese with T2D (T2D). Each subject completed an oral health survey, received a clinical oral examination, and provided unstimulated saliva for measurement of inflammatory markers and microbiome analysis.

Results: The diabetes group was less likely to have had a dental visit within the last six months. Body mass index (BMI) Z-scores and waist circumference/height ratios were similar between Ob and T2D cohorts. The number of carious lesions and fillings were similar for all three groups. The gingival index was greater in the T2D group compared to the Ob and C groups. Although salivary microbial diversity was minimal between groups, a few differences in bacterial genus composition were noted.

Conclusions: Obese children with T2D show a trend toward poorer oral health compared to normal weight and obese children without T2D. This study characterizes the salivary microbiome of children with and without obesity and T2D. This study supports a modest link between T2D and periodontal inflammation in the pediatric population.

Fig 1. Alpha diversity as measured using five metrics.

Calculations and plots were generated using the Phyloseq package. Unadjusted p-values using the Kruskal-Wallis test for microbiome alpha diversity measures were as follows: Observed: 0.1594; Chao: 0.2760; Ace: 0.2872; Shannon: 0.1714; and Simpson: 0.0354. No significant differences at p<0.01 were observed between groups.

Fig 2. Bray Curtis based NMDS plots of all samples.

Ellipsoids represent 95% confidence intervals surrounding each cohort. MRPP analysis indicates that control group members are more dissimilar than expected by chance (agreement statistic A = 0.049; p<0.001).

Fig 3. OTUs at genus level that are significantly different across groups (C = control, O = obese, T = Type 2 diabetes).

The relative abundance is shown as log(counts +1) for the significant OTUs with false discovery rate controlled at 0.05 via a Benjamini and Hochberg procedure. The q-values were computed from the p-values from a likelihood ratio test applied to a Negative Binomial generalized linear model fit using the edgeR package. Note, Lautropia, Corynebacterium, and Cardiobacterium were significantly different after adjusting for gingival index.