• 2022-07
  • 2022-05
  • 2022-04
  • 2021-03
  • 2020-08
  • 2020-07
  • 2020-03
  • 2019-11
  • 2019-10
  • 2019-09
  • 2019-08
  • 2019-07
  • br Corresponding author at Vanderbilt Epidemiology Center Vanderbilt Ingram Cancer


    Corresponding author at: Vanderbilt Epidemiology Center, Vanderbilt-Ingram Cancer Center, Vanderbilt University School of Medicine, MCN B-2104, 1161 21st Avenue South, Nashville, TN, 37232-2400, USA. E-mail address: [email protected] (Q. Cai).
    smoking status can modify this potential association.
    2. Materials and methods
    2.1. Study population
    The present population-based nested case-control study was drawn from the SCCS, a prospective cohort designed to explore the underlying causes of racial disparities in health outcomes. The detailed information regarding the SCCS has been reported elsewhere [18,19]. Briefly, be-tween March 2002 and September 2009, 84,797 adults (40–79 years old) were recruited from 12 southeastern American states (Alabama, Arkansas, Florida, Georgia, Kentucky, Louisiana, Mississippi, North Carolina, South Carolina, Tennessee, Virginia, and West Virginia). Ap-proximately 86% of the participants were recruited from 71 Commu-nity Health Centers (CHCs) where healthcare services are provided that are mainly focused on low-income and uninsured people, and 14% of the participants were recruited through written materials by mailings in these geographic areas. Computer-assisted personal interviews were performed by trained interviewers to collect baseline data regarding the participants’ demographic characteristics, medical history, and poten-tial health risk factors, including dietary habits, physical activity, smoking, alcohol consumption, and anthropometric characteristics. All participants were regularly followed-up, and data linkages to the states’ cancer registries and/or the National Death Index mortality records were used to ascertain incident cancers.
    2.2. Ascertainment of lung cancer
    By the end of 2015, a total of 403 incident lung cancer cases with oral health data available were ascertained. Lung cancer cases were identified by using the tenth revision of the International Statistical Classification of Diseases (ICD-10: C340–C349). All cases were in-dividually matched with four controls on age at enrollment ( ± 2 years), sex, Nicotine (African American, European American, or others), the recruitment source (CHC or general population), and recruitment site (for CHC) of study enrollment. Ultimately, a total of 2015 individuals, including 403 incident lung cancer cases and 1612 matched-controls, were selected for the present study.
    2.3. Assessment of oral health
    At the first follow-up survey for the SCCS (2008–2011), self-re-ported data on tooth loss, tooth decay, and history of periodontal dis-ease were collected by the following questions and response categories: “About how many adult teeth have you lost in your lifetime due to tooth decay or gum disease?” (none, 1–4, 5–10, > 10 but not all of them, and all of them); “How many decayed teeth or cavities do you currently have spermatogenesis have not been treated?” (none, 1–2, 3–5, > 5, and no teeth); “Has a dentist or doctor ever told you that you have gum disease (gingivitis or periodontitis)?” (no or yes); and “At what age were you diagnosed with gum disease (gingivitis or periodontitis)?” (a continuous variable, reported in years).
    2.4. Statistical analysis
    The baseline characteristics of lung cancer risk factors (i.e., age, sex, body mass index [BMI], race, education, household income, marital status, history of chronic obstructive pulmonary disease [COPD], al-cohol drinking, smoking status, and pack-years) were compared be-tween the lung cancer cases and controls using Student’s t-test for continuous variables or chi-square test for categorical variables. Multiple imputation methods were used to impute missing data for tooth loss (n = 18; 0.9% missing), tooth decay (n = 5; 0.3% missing), periodontal disease (n = 35; 1.7% missing), age of periodontal disease (n = 46; 2.3% missing) and other covariates (less than 0.1% missing) 
    by SAS PROC MI. Based on oral health information, tooth loss was grouped into four categories: none, 1–4, 5–10, and more than 10, and tooth decay categorized as none, 1–5, and 6 or more. Age of periodontal disease was divided into two groups as ≥ 40, and < 40 (early onset of periodontal disease). The results, based on the original data and the imputed data, were first compared in crude models. Multivariate lo-gistic regression models were fitted with adjustment for BMI (con-tinuous), education (years of schooling: less than 12 years, 12 years, above 12 years), household income (< $15,000, $15,000-$25,000, > $25,000), history of COPD (yes, no), alcohol drinking (never, ever), smoking status (never, former, current), and race-specific median pack-years (African Americans: ≥ 17.5 pack-years; European Americans: ≥ 33.0 pack-years; other races: ≥ 31.0 pack-years). Linear trend tests for tooth loss and tooth decay were conducted by using the Wald test. All associations were further evaluated in the stratified analyses by race, smoking status, and pack-years; likelihood ratio test tested interactions between oral health status and those factors. All analyses were performed using the SAS software (version 9.4; SAS Institute, Cary, NC, USA).