Acute Myocardial Infarction and Periodontal Disease

Acute Myocardial Infarction and Periodontal Disease Research Findings The study examined the association between acute myocardial infarction and periodontal diseases using cross sectional design. The analysis was carried out using the SPSS/PC Windows version 21.0 software package (IBM, Inc.). The sample size taken for the study was 80 (Cases=40, Control=40). The bivariate association between the studied variables, acute MI and periodontitis (dichotomized) was analyzed with the appropriate test. A significance level of p≠¤0.05 was considered significant and the odds ratios with 95% confidence intervals were calculated. Further, conditional logistic regression analysis/cox regression analysis (1:1 matched pairs) was used to assess the independent contribution of periodontal diseases to the risk of acute myocardial infarction and also to find the relationship between AMI and other possible explanatory variables. The risk factors such as tobacco habit, smoking, dietary habits, family history of diabetes, were forced into the model. The following secti on presents the results. Descriptive Statistics and Preliminary Analyses Association between acute myocardial infarction and study variables. The table below presents the association between Acute Myocardial Infarction and study variables. The results showed that odds of outcome (AMI) were significantly higher in subjects with periodontitis, smoking habits, hypertension and mixed dietary habits. Out of the total 80 patients, the majority of the periodontitis patients (82.5%) were present within the case group (AMI patients) (p=0.026). Similarly, the prevalence of smoking (52.5% vs. 27.5%, p=0.031) and hypertension (52.5% vs. 47.9%, p Table 6:  Association between Acute Myocardial Infarction and study variables *p Association between periodontitis and study variables. The table below presents the association between Periodontitis and study variables. The results showed that odds of outcome (periodontitis) were significantly higher in subjects with smoking habits, hypertension and alcohol drinking (p Table 7:  Association between Periodontitis and study variables Conditional Logistic Regression Analysis Using Cox Proportional Hazard Model Following the preliminary analysis’ cox regression analyses were used to assess the independent contribution of periodontal diseases to the risk of acute myocardial infarction and also to find the relationship between an AMI event and possible explanatory variables. To control the effects of multiple potential confounders, multivariate model were also fitted by modeling periodontitis as a time varying covariant in a model. Cox proportional hazard analysis allowed the researcher to include the predictor variables (covariates) one by one into the subsequent models. This provided estimated coefficients for each of the covariates and allowed the researcher to assess the impact of multiple covariates in the same model. We can also use Cox regression to examine the effect of continuous covariates such as BMI. The following recoding was done to examine the association between AMI and periodontitis. Socio economic status=0 (Reference category): Lower; 1=Upper Lower; 2=Lower middle; 3=Upper middle; 4=Upper: Family history=0 (Reference category): No; 1=Yes; Exercise=0 (Reference category): Yes; 1=No ; Hyper tension=0 (Reference category): No; 1=Yes: Diabetes=0 (Reference category): No; 1=Yes: Dietary habit=0 (Reference category): Vegetarian; 1=Mixed: Smoking habit=0 (Reference category): Non-smoker; 1= Former smoker; 2=Smoker: Smokeless tobacco habit=0 (Reference category): Non-users; 1= Former user; 2=Current u ser: Alcohol drinking=0 (Reference category): Non-drinker; 1= Current drinker; 2=Irregular abstainer: Marital status=0 (Reference category): Unmarried; 1= Married; 2=Divorced. The conditional logistic regression estimates the odds ratio, and an exact 95% confidence interval. Table 3, below presents the association between AMI and Periodontitis using Cox regression. Table 8:  Cox Regression Dependent variable: Acute Myocardial Infarction Conditional logistic regression analysis outcomes indicated the presence of a significant association between AMI and periodontitis (Beta=1.358, p= .039 In all the stages, for regular exercise the beta value is negative which means it is a protective factor or is inversely related as acute myocardial events, however this association is not statistically significant to report. Statistical Analysis of Other Clinical Parameters (DMFT, CPI and LOA Scores) Test for normality. To test the assumption of normality, the study used the Kolmogorov-Smirnov and Shapiro-Wilks test. From this test, the Sig. (p) value was compared to the priori alpha level (level of significance for the statistic) – and a determination was made as to reject (p ÃŽ ±) the null hypothesis. The Table 1 below shows that where ÃŽ ± = 0.001, given that p Table 9:  Test for Normality Test for homogeneity of variance (equality of variances). Further, to test the assumption of homogeneity of variance, where the null hypothesis assumes no difference between the two group’s variances (H0: 2 ÏÆ' 1 = 2 ÏÆ' 2), a non-parametric Levene’s test for equality of variances is the most commonly used statistic to verify the equality of variances in the samples (homogeneity of variance) especially for non-normally distributed data. Therefore, Kruskal Wallis one-way analysis Leven’s test was applied. The Levene’s test uses the level of significance set a priori for the t test analysis (e.g., ÃŽ ± = .05) to test the assumption of homogeneity of variance. However, in SPSS it’s challenging to execute Leven’s test for non-normally distributed data in one step. Hence steps were applied to create three new variables such as ranked data, group mean ranks and deviation from mean ranks. Finally, the differences were computed using ANOVA and the p value was found to be Table 10:  Test Statistics Mann-Whitney U and Wilcoxon W test : comparing medians. As the data is non-homogenous and non-normally distributed, Mann-Whitney U and Wilcoxon W tests were used to compare the median scores of DMFT, CPI and LOA scores, and also to check the significance of differences. Null Hypothesis: Median score of DMFT, CPI and LOA is same for both case and control. Alternative hypothesis: Median score of DMFT, CPI and LOA differs between case and control. Table 11:  Test Statistics DMFT Score CPI Score LOA Score Mann-Whitney U 403.500 340.500 374.500 Wilcoxon W 1223.500 1160.500 1194