Table 5.

Multiple logistic regression models for the associations of C-Reactive Protein, insulin, and adiponectin with intermediate hyperglycemia and obesity using data from visit 3.

	Intermediate hyperglycemia			Obesity
	OR	P-value	95% CI	OR	P-value	95%CI
Salivary CRP	4·97	0·004*	1·66,14·90	4·53	<0·001*	2·4,8·5
Serum CRP	3·35	0·02*	1·13,9·93	14·75	<0·001*	7·35,29·60
Salivary insulin	2·64	0·03*	1·09,6·38	3·29	<0·001*	1·82,5·97
Serum insulin	5·93	0·003*	1·82,19·34	5·82	<0·001*	3·24,10·45
Salivary adiponectin	0·72	0·5	0·23,2·27	0·54	0·044*	0·3,0·9
Serum adiponectin	0·76	0·5	0·31,1·89	0·43	0·001*	0·26,0·71

N = 323.

The models were adjusted for age, sex, blood pressure, and school (residential areas).

Obesity was defined as BMIz> 2SD. Intermediate hyperglycemia was defined as hemoglobin A1c (HbA1c) between >5.6– <6.4%.

Salivary biomarker variables were stratified by 75^th percentile cutoff. Serum biomarkers variables were stratified by 50^th percentile cutoff (median).

OR, Odds Ratio; CI, 95% Confidence Interval; CRP, C-Reactive Protein.

*, P-Value <0.05.