*2.7. Sample Size Calculation and Statistical Analysis*

This was a prospective cohort study, and the sample size was calculated based on the data of previous studies that assessed the relationship between dietary intakes and oral functions [22,36]. Since sodium intake was found to be the factor closely related to the BP of older adults in a previous study [5], we found it appropriate to determine the sample size of this study based on the sodium intake. Based on this, the minimum number of participants required in each group was 185, and, hence, the analysis was performed with data, including those of the participants who presented to our study until December 2019 (at this time, the number of participants in each group exceeded the required number as per the sample size calculation).

The normality of the data distribution was verified using the Kolmogorov–Smirnov test (*p* > 0.05); the data were found to be non-normally distributed. To determine the differences in oral health and dietary intake among the participants with different BP groups, Kruskal-Wallis and Chi-square tests were performed, while intergroup comparisons were performed using the Mann-Whitney U-test with Bonferroni correction. The measurement items were specified according to a Kruskal-Wallis or X<sup>2</sup> test (*p* < 0.05, explanatory variable for basic characteristics and oral health items; *p* < 0.1, explanatory variable for dietary intake). We used the Spearman's rank correlation coefficient to examine the relationship between oral health and dietary intake. To determine the impact of hypertension-related factors, we performed binary logistic regression analyses with stepwise methods (input 0.05, remove 0.15). Hypertension, defined as participants classified as Hypertensive or History of hypertension, was considered the objective variable (Normotensive participants = 0, Hypertensive or History of hypertension participants = 1). The explanatory variables were selected from the participants' basic characteristics, oral health, and dietary intake, which were found to be significantly related to hypertension by the Kruskal-Wallis or chi-square test. All statistical analyses were performed using the SPSS Statistics 22.0 (IBM, Tokyo, Japan).
