Endocrine Abstracts

Background: With lacking consistent direction of the association of thyroid dysfunction with periodontitis and limitations of previous studies, we aim to look for the probable association between periodontitis and thyroid disorders based on cross sectional and longitudinal results in population-based data of study of health in Pomerania (SHIP).

Methods: The net sample of first SHIP cohort; SHIP-0 (1997-2001), comprised of 4308 individuals aged 20-79 years. A second separate net sample of 4420 adults aged 20-79 years was drawn for SHIP-TREND (2008-2012). For the cross-sectional analysis data from SHIP-0 and SHIP-TREND-0 and for the longitudinal analysis data from SHIP-0 and SHIP-2 (11 years follow-up of SHIP-0) was used. In SHIP-0, SHIP-2 and SHIP-TREND-0, serum TSH, fT3, fT4 levels were measured by an immunochemiluminescent procedure. Periodontal examinations comprised probing depth (PD) and clinical attachment loss (CAL) performed according to the half-mouth method, described in CDC/AAP case definition. In cross sectional analysis thyroid biomarkers were associated with measures of periodontal status by linear, simple and multinomial logistic regression analysis. In longitudinal analysis negative binomial, poisson and linear regression analysis were conducted where necessary.

Results: In cross sectional analysis normal TSH concentration was inversely associated to the number of teeth, and positively associated with measures of periodontitis (PD mean and PD sites) only in SHIP-0. While in longitudinal analysis normal TSH concentration was positively associated with number of teeth lost between SHIP-0 and SHIP-2. High TSH concentration was inversely associated with number of teeth only in SHIP-TREND-0, and positively associated with measures of periodontitis (PD mean and PD sites) only in SHIP-0. Serum fT4 concentrations was inversely associated with CAL site in SHIP-TREND-0 only and positive association with moderate periodontitis only in SHIP-0. In longitudinal analysis serum fT4 was inversely associated with CAL mean and site differences between SHIP-0 and SHIP-2. Serum fT3 concentrations were associated with almost all measures of periodontitis in SHIP-0 and CAL site, moderate and severe periodontitis and loose teeth in SHIP-TREND-0, while in longitudinal analysis serum fT3 was only associated with PD mean differences between SHIP-0 and SHIP-2.

Conclusion: Our findings for association between thyroid function and periodontitis are inconsistent. Irrespective of the differences between the two SHIP cohorts, the overall results are inclined towards association of periodontitis to hypothyroidism. From longitudinal analysis we cannot draw a clear causal association, probably to a lesser than required number of individuals in the analysis.