Skip to main content

Table 2 Models used for analysis of the effects of statins on incidence of dementia using the Cox proportional survival method.

From: Simvastatin is associated with a reduced incidence of dementia and Parkinson's disease

 

Model 1

Model 2

Model 3

Variable

HR

p value

95% Confidence limits for HR

HR

p value

95% Confidence limits for HR

HR

p value

95% Confidence limits for HR

Atorvastatin

            

   Versus CV

0.80

0.0005

0.71

0.91

0.90

0.09

0.79

1.02

0.91

0.11

0.80

1.02

   Age

    

1.11

<0.0001

1.10

1.12

1.11

<0.0001

1.10

1.11

   Hypertension

    

1.10

0.0590

1.00

1.20

    

   CVD

    

1.13

0.0005

1.05

1.20

    

   Diabetes

    

1.12

0.0009

1.05

1.20

    

   Charlson Index

        

1.78

<0.0001

1.66

1.91

Lovastatin

            

   Versus CV

0.98

0.70

0.89

1.08

0.95

0.32

0.86

1.05

0.95

0.34

0.86

1.05

   Age

    

1.11

<0.0001

1.10

1.12

1.11

<0.0001

1.10

1.11

   Hypertension

    

1.09

0.07

0.99

1.19

    

   CVD

    

1.12

0.0005

1.05

1.20

    

   Diabetes

    

1.12

0.0009

1.05

1.20

    

   Charlson Index

        

1.74

<0.0001

1.62

1.86

Simvastatin

            

   Versus CV

0.45

<0.0001

0.43

0.47

0.45

<0.0001

0.43

0.48

0.46

<0.0001

0.44

0.48

   Age

    

1.11

<0.0001

1.10

1.11

1.11

<0.0001

1.10

1.11

   Hypertension

    

1.07

0.07

1.00

1.15

    

   CVD

    

1.09

0.0012

1.04

1.15

    

   Diabetes

    

1.08

0.0034

1.03

1.14

    

   Charlson Index

        

1.72

<0.0001

1.63

1.81

  1. HR, hazard ratio, CVD, cardiovacular disease, CV, cardiovascular.
  2. Model 1 included an adjustment for the interaction between dementia and age; model 2 included adjustments for the interaction between dementia and three disorders associated with increased risk of dementia.; model 3 included adjustments for the interaction between dementia and the Charlson Index, which is an index that provides a broad assessment for the amount of chronic disease. For the purpose of analysis of interactions with the Charlson Index, subjects with a Charlson Index of 0 or 1 were compared with those with a Charlson Index of ≥2.