The data in this study are a subset of the sixth Korean National Health and Nutrition Examination Survey (KNHANES VI) conducted in 2015 by the Korea Centers for Disease Control and Prevention (KCDC). The KNHANES is a nationwide, population-based, cross-sectional survey examining the general health and nutritional status of Koreans. It has been conducted by the Korean Ministry of Health and Welfare and the KCDC since 1998 [15]. The sampling protocol for KNHANES involves a complex, stratified, and multistage probability cluster survey of a representative sample of the noninstitutionalized civilian population of Korea [15].

The number of participants in the KNHANES VI was 6,977 (3,169 males and 3,808 females). The sample size for this study was 4,727, including those in the unweighted study sample who were between 19 and 64 years of age in 2015. The mean age of our sample was 43.39 years (standard deviation, 18.29 years). From the overall KNHANES VI data, we used sociodemographic characteristics (age, sex, household income), oral health-related variables (tooth-brushing frequency, self-oral health cognition, Community Periodontal Index [CPI]), oral and general health status and behavior (smoking status, obesity, diabetes), and YI. The 2015 KNHANES allowed the researchers to obtain national-level estimates with an assigned sampling weight variable that could generate aggregate weighted frequencies [15].

The yogurt type studied was only semi-solid yogurt, not drinkable yogurt. To determine routine YI over the previous week, intake frequency in the last year was classified into 9 categories ranging from “never or almost never” to “3 times per day,” with 1 serving equal to a cup (100 grams). YI was calculated by multiplying the YI frequency over the previous week by the average intake per serving. Participants were divided into 4 categories based on the number of cups of yogurt consumed: 1) <1 cup per week, 2) ≥1 cup per week but <1 cup per day, 3) ≥1 cup but <2 cups per day, and 4) ≥2 cups per day.

Sociodemographic variables and risk factors related to periodontal disease were selected as covariates. The confounders in this study were age, sex, household income, tooth-brushing frequency, self-oral health cognition, smoking status, obesity, and diabetes. Household income was adjusted for the number of household members and categorized into 4 quartiles, ranging from quartile 1 (bottom) to quartile 4 (top). Participants were divided by tooth-brushing frequency into 2 groups: 1) less than twice a day and 2) twice or more a day. Self-oral health cognition was categorized into 3 groups: 1) high, 2) moderate, and 3) low. Participants were classified by smoking status into 3 groups: 1) current smokers (smoking currently and smoked ≥100 cigarettes), 2) ex-smokers (smoked in the past but stopped), and 3) non-smokers (never smoked). We followed the guidelines of the Korean Society for the Study of Obesity and the World Health Organization (WHO) [16] in categorizing the participants into 3 groups by body mass index: 1) underweight (<18.5 kg/m²), 2) normal weight (18.5–24.9 kg/m²), and 3) obese (≥25 kg/m²). Diabetes status was categorized into 3 groups: 1) normal, 2) impaired fasting glucose, and 3) diabetes. Diabetes was defined as a fasting glucose level ≥126 mg/dL, current use of anti-diabetic medications, or a self-reported physician's diagnosis of diabetes mellitus. Impaired fasting glucose and normal glucose tolerance were defined as a fasting glucose level of 100–125 mg/dL and <100 mg/dL, respectively [17].

Periodontal status was evaluated using the CPI developed by the WHO [18]. Ten index teeth—2 molars in each posterior sextant and the upper right and lower left central incisors—were measured for periodontal pocket depth (PD) using a CPI probe that met the WHO guidelines [19] at 6 sites per tooth (mesiobuccal, midbuccal, distobuccal, mesiolingual, midlingual, and distolingual). Probing was carried out by 16 dentists who had received calibration training during a 4-day session for the KNHANES. Five CPI scores were recorded: CPI 0, normal; CPI 1, gingival bleeding; CPI 2, presence of gingival calculus; CPI 3, shallow periodontal pocket (>3.5 mm but ≤5.5 mm); and CPI 4, deep periodontal pocket (>5.5 mm). We defined periodontitis as CPI scores of 3 or 4, which were used to dichotomize participants into non-periodontitis and periodontitis groups. Calibration training to ensure reliability of the oral health survey in the KNHANES was conducted in 2015 and included lectures, web-based photo instruction, training using dental models, a simulated oral health examination with human subjects, and field instruction with a reliability assessment. The inter-Kappa index was 0.827–1.000 at the final assessment.

The results of the oral examinations and questionnaires were analyzed using SPSS version 23.0 (IBM Corp., Armonk, NY, USA). P values less than 0.05 were considered to indicate statistical significance for all analyses. Individually weighted factors were considered, and the complex sampling design of the survey was utilized to obtain the variances. Age and YI were analyzed using the independent t-test for continuous variables, and the other variables were analyzed using the χ² test for non-continuous variables. Multivariate logistic regression analyses were performed to examine the association between YI and periodontal health, adjusting for sociodemographics, oral health-related variables, and health behaviors and status. The complex sampling design was considered in all analyses.

The sample consisted of 4,727 participants with a weighted mean age of 43.4 years (range, 19–64 years). Table 1 shows the distribution of sociodemographic characteristics, oral health-related variables, and health status and behavior by periodontitis status. Among the total sample, 30.8% had periodontitis (not presented in tables). There were significant differences in the distribution of periodontitis with regard to all variables. Individuals with periodontitis were older (54.44±0.61 years) than those without periodontitis (39.25±0.43 years) (P<0.001). More men had periodontitis than women (55.8% vs. 44.2%). On an average, participants with periodontitis brushed their teeth less and had lower self-oral health cognition. The proportion of periodontitis was higher among current smokers than among ex- or non-smokers, and was also higher in obese participants and those with diabetes than in others. Participants without periodontitis showed a significantly higher YI (1.03±0.06 cups/week) than participants with periodontitis (0.77±0.08 cups/week) (P<0.001).

The distribution of subject characteristics by YI is shown in Table 2. Among those who consumed less than 1 cup of yogurt per week, the proportion of men was higher than that of women. However, among those who consumed more than 1 cup of yogurt per week, the proportion of women was higher than that of men. The proportion of families with a high income were higher among those who consumed more than 2 cups per day. People who had never smoked consumed more yogurt than those who were current or ex-smokers. Moreover, people with periodontitis were significantly more common among those who consumed less than 1 cup of yogurt per week than those who consumed more than 2 cups per day.

The results of the multivariate logistic regression analyses are shown in Table 3. With adjustment for age, the risk of periodontitis was 73% lower among individuals in the highest YI group (P<0.001 for trend). In the final multivariate model, the association between YI and periodontitis remained moderately strong and statistically significant (P=0.012 for trend). Individuals who consumed more than 2 cups of yogurt per day were 76% less likely to have periodontitis than those who consumed less than 1 cup per week.