Association between tooth loss, chronic conditions, and common risk factors: Results from the 2019 Brazilian Health Survey

Tayse Caroline Cunha de Medeiros,Alessandra Areas E Souza,Rodolfo Coelho Prates,Iain Chapple,Joao Paulo Steffens
DOI: https://doi.org/10.1002/JPER.21-0433
Abstract:Background: The aim of this study was to evaluate the association between systemic non-communicable diseases (NCDs; including lung, kidney, mental and cardiovascular diseases, arthritis, cancer and spinal problems), common risk factors, and tooth loss (TL), as an endpoint of prevalent oral NCDs (periodontitis and caries). Methods: A total of 60,271 noninstitutionalized adults (≥30 years) were evaluated, using data from the 2019 Brazilian National Health Survey. Negative binomial regressions were performed, adjusting for sex, age, ethnicity, educational attainment, oral hygiene, risk factors for periodontitis and caries (diabetes, smoking, and a cariogenic diet). The dependent variable was TL expressed as a numerical value. Results: Diabetes, current smoking and a frequent cariogenic diet were significantly associated with TL-Incidence Rate Ratio (IRR [95% CI]): 1.11 (1.08 to 1.14), 1.28 (1.25 to 1.31), and 0.97 (0.94 to 0.99), respectively. Significant associations were observed for TL and all assessed NCDs, except kidney diseases, cancer and musculoskeletal diseases related to work, with IRR ranging from 1.06 for hypertension and asthma to 1.16 for rheumatoid arthritis. Regular consumption (4 to 7 d/wk) of vegetables, fruits and beef; alcohol up to 8 doses/wk; and physical exercise were associated with a lower IRR for TL (P < 0.05). Obesity, but not overweight, was associated with increased TL (1.05 [1.03 to 1.07]). Smoking, hypertension, arthritis, other mental diseases and spinal problems further increased the IRR for TL in individuals with diabetes (P < 0.05). Conclusion: We conclude that certain chronic systemic conditions are associated with TL in Brazilian adults. This is likely because of shared risk factors; however, causal associations cannot be examined in this cross-sectional dataset.
What problem does this paper attempt to address?