Table 3 Results of Cox proportional hazards model investigating the association between visit-to-visit variability of systolic blood pressure and cardiovascular events in the Japan Epidemiology Collaboration on Occupational Health Study (2008–2019), shown by cardiovascular diseases subtype

From: Visit-to-visit variability of blood pressure and cardiovascular events among the working-age population in Japan: findings from the Japan Epidemiology Collaboration on Occupational Health Study

 

Stroke mortality

CHD mortality

 

Person-years

No of events

HR

95% CI

Person-years

No of events

HR

95% CI

VVV defined by SD

        

  Continuous

411,631

27

1.47

1.29, 1.67

411,631

36

1.28

1.17, 1.41

  Categorical

        

   T1 (lowest)

141,342

3

Ref.

 

141,342

5

Ref.

 

   T2 (middle)

137,424

8

2.63

0.84, 8.28

137,424

10

1.96

0.36, 10.69

   T3 (highest)

132,865

16

4.08

1.49, 11.15

132,865

21

3.23

0.95, 10.97

VVV defined by CV

        

  Continuous

411,631

27

1.53

1.28, 1.83

411,631

36

1.35

1.21, 1.50

  Categorical

        

   T1 (lowest)

139,263

5

Ref.

 

139,263

5

Ref.

 

   T2 (middle)

138,322

7

1.40

0.43, 4.57

138,322

12

2.42

0.47, 12.52

   T3 (highest)

134,047

15

2.87

1.00, 8.21

134,047

19

3.54

1.22, 10.25

VVV defined by VIM

        

  Continuous

411,631

27

1.53

1.27, 1.85

411,631

36

1.35

1.22, 1.51

  Categorical

        

   T1 (lowest)

139,124

5

Ref.

 

139,124

5

Ref.

 

   T2 (middle)

138,375

7

1.40

0.42, 4.63

138,375

12

2.45

0.47, 12.73

   T3 (highest)

134,132

15

2.96

1.05, 8.33

134,132

19

3.66

1.28, 10.45

VVV defined by ASV

        

  Continuous

411,631

27

1.46

1.15, 1.86

411,631

36

1.19

1.07, 1.32

  Categorical

        

   T1 (lowest)

141,388

7

Ref.

 

141,388

4

Ref.

 

   T2 (middle)

140,138

5

0.66

0.31, 1.41

140,138

13

2.88

1.43, 5.78

   T3 (highest)

130,105

15

1.70

0.86, 3.38

130,105

19

3.70

1.39, 9.87

  1. ASV average absolute difference between successive values, CV coefficient of variation, CVD cardiovascular disease, SD standard deviation, VIM variability independent of the mean, VVV visit-to-visit variability
  2. Models were adjusted for baseline information on mean SBP (2008–2011), age, sex, smoking status, body mass index, antihypertensive medication, diabetes and dyslipidemia while sex and dyslipidemia were excluded from the models for CHD mortality using continuous VVV indicators. We modeled age group (in models for CHD mortality using continuous VVV indicators), dyslipidemia and BMI (in models for CHD mortality using categorical VVV indicators) as time-varying covariates. We also accounted for clustering by study site. BMI categories were collapsed to deal with small number of participants included in lean (for stroke mortality) and obesity (for CHD mortality). Models with continuous exposures report HRs per one-SD increase in the VVV indicator.
Back to article page