Multivariate hazard ratio and 95% confidence interval of association between number of teeth, myocardial infarction and stroke.

§: adjusted for Age (continuous), Sex, BMI (continuous), education (3 categories), occupation (4 categories), Lifestyle (smoking (never, former, current, number of cigarettes per day), alcohol consumption (continuous), physical activity (Cambridge physical activity index), use of vitamin and/or mineral supplements, antibiotics and non–steroidal anti–inflammatory drugs, hormone replacement therapy (women), prevalent diseases and three retained factors from factor analysis of 49 food groups
Model 1: Age (continuous), Sex, BMI (continuous), 1–17 teeth×time interaction.
Model 2: Model 1+ education (3 categories), occupation (4 categories), Lifestyle (smoking (never, former, current, number of cigarettes per day), alcohol consumption (continuous), physical activity (Cambridge physical activity index), use of vitamin and/or mineral supplements, antibiotics, hormone replacement therapy (women) and non–steroidal anti–inflammatory drugs
Model 3: Model 2+ prevalent diseases (MI: prevalent hypertension, angina pectoris, heart failure, transient ischaemic shock, stroke, T2DM and cancer; stroke: prevalent hypertension, angina pectoris, heart failure, MI, transient ischaemic shock, T2DM and cancer).
Model 4: Model 3+ three retained factors from factor analysis of 49 food groups
Model 5: Model 4 + competing risk events of other three incident diseases.
HR: Hazard ratio
* P–value for association;
‡ P–value for linear trend;
† adjusted competing risk events
MI: Group 1–17teeth×time interaction (P = 0.04 for interaction) (model 4). Final model showed no evidence that proportional hazard assumption was violated (P = 0.49).
§Hazard ratios (unadjusted for teeth 1–17 interaction with time): 2.02 (1.26–3.23), 1.59 (1.01–2.5), 1.62 (1.01–2.59), 1.55 (0.96–2.49), 1.42 (0.9–2.52)
Stroke: All variables and final model (model 4) showed no evidence that proportional hazard assumption was violated (P = 0.11).
Multivariate hazard ratio and 95% confidence interval of association between number of teeth, myocardial infarction and stroke.