The baseline characteristics of the 4569 participants were shown in Table 1. Within a mean follow-up of 12.1 years, a total of 403 adverse cardiovascular events occurred, including 173 MI and 263 strokes.
Table 1
Baseline clinical characteristics of participants
Characteristics
|
Total
|
Gender
|
Female
|
Male
|
P
|
Age(year)
|
52.00(46.00–61.00)
|
52.00(46.00–61.00)
|
52.00(46.00–61.00)
|
0.620
|
Gender
|
|
|
|
|
female
|
2424(53.05%)
|
NA
|
NA
|
|
male
|
2145(46.95%)
|
NA
|
NA
|
|
BMI(kg/m^2)
|
22.91(20.76–25.17)
|
23.24(21.00–25.49)
|
22.54(20.55–24.77)
|
< .001
|
SBP(mmHg)
|
120.00(110.00–132.00)
|
120.00(110.00–132.00)
|
121.00(112.00–133.00)
|
< .001
|
DBP(mmHg)
|
80.00(71.00–87.00)
|
79.00(70.00–85.00)
|
80.00(73.00–88.00)
|
< .001
|
Antihypertensive drugs
|
|
|
|
|
no
|
4204(93.69%)
|
2216(92.99%)
|
1988(94.49%)
|
0.040
|
yes
|
283(6.31%)
|
167(7.01%)
|
116(5.51%)
|
|
Current smoker
|
|
|
|
|
no
|
3030(67.54%)
|
2249(93.94%)
|
781(37.33%)
|
< .001
|
yes
|
1456(32.46%)
|
145(6.06%)
|
1311(62.67%)
|
|
Current drinker
|
|
|
|
|
no
|
2833(63.79%)
|
2088(89.23%)
|
745(35.46%)
|
< .001
|
yes
|
1608(36.21%)
|
252(10.77%)
|
1356(64.54%)
|
|
Diabetes
|
|
|
|
|
no
|
4267(98.18%)
|
2272(97.80%)
|
1995(98.62%)
|
0.046
|
yes
|
79(1.82%)
|
51(2.20%)
|
28(1.38%)
|
|
Site
|
|
|
|
|
urban
|
1516(33.18%)
|
807(33.29%)
|
709(33.05%)
|
0.864
|
rural
|
3053(66.82%)
|
1617(66.71%)
|
1436(66.95%)
|
|
Composite outcome
|
|
|
|
|
no
|
4166(91.18%)
|
2242(92.49%)
|
1924(89.70%)
|
< .001
|
yes
|
403(8.82%)
|
182(7.51%)
|
221(10.30%)
|
|
2.1. Association of BP level with composite outcome
Table 2 displays the association of BP level with composite outcome. In multivariable adjusted models(model 2), SBP and DBP showed a linear association with composite outcome, respectively. Compared with participants with SBP 0-130 mmHg, the multivariable adjusted HRs[95% CIs] (model 2) were 1.48[1.09-2.00], 2.26[1.69–3.02], 3.15[2.15–4.62], and 3.90[2.25–6.75] for composite outcome in participants with SBP 130–140, 140–160, 160–180, and ≥ 180 mmHg, respectively; a significant positive trend in categorized SBP for incident composite outcome (P for trend < 0.001) was observed; each additional 10 mmHg of SBP was associated with a 23% higher probability of developing composite outcome(1.23[1.16–1.29]). However, in a mutually adjusted model(model 3) including both SBP and DBP, the association with the risk of composite outcome was positive for SBP, but not significant for DBP.
Multivariable adjusted restricted cubic spline analyses showed a linear association of SBP (P for linearity < 0.001, model 2; P for linearity < 0.001, model 1) and DBP (P for linearity < 0.001, model 2; P for linearity < 0.001, model 1) with composite outcome, respectively (Fig. 1; Figure s1 in appendix). Multivariable adjusted restricted cubic spline analyses (model 3) showed a linear association of SBP with composite outcome; no significant linear or nonlinear relationships of DBP with composite outcome were observed (Figure s2 in appendix).
In subgroup analysis, SBP and DBP were more strongly associated with composite outcome in 40–65 age group, non-use of antihypertensive drugs group, and non-history of diabetes group(Fig. 2 and Fig. 3).
Table 2
Associations of BP with composite outcome
|
Model 1
|
Model 2
|
Model 3
|
HR[95%CI]
|
P
|
HR[95%CI]
|
P
|
HR[95%CI]
|
P
|
SBP(mmHg)
|
|
|
|
|
|
|
0-130
|
1.00
|
|
1.00
|
|
1.00
|
|
130–140
|
1.58[1.20–2.09]
|
0.001
|
1.48[1.09-2.00]
|
0.011
|
1.34[0.97–1.84]
|
0.072
|
140–160
|
2.69[2.07–3.48]
|
< .001
|
2.26[1.69–3.02]
|
< .001
|
1.92[1.38–2.67]
|
< .001
|
160–180
|
3.84[2.71–5.43]
|
< .001
|
3.15[2.15–4.62]
|
< .001
|
2.46[1.57–3.87]
|
< .001
|
>=180
|
5.51[3.32–9.15]
|
< .001
|
3.90[2.25–6.75]
|
< .001
|
2.78[1.46–5.28]
|
0.002
|
P for trend
|
|
< .001
|
|
< .001
|
|
< .001
|
10mmHg increase
|
1.28[1.23–1.34]
|
< .001
|
1.23[1.16–1.29]
|
< .001
|
1.18[1.10–1.27]
|
< .001
|
DBP(mmHg)
|
|
|
|
|
|
|
0–80
|
1.00
|
|
1.00
|
|
1.00
|
|
80–90
|
1.89[1.48–2.43]
|
< .001
|
1.71[1.30–2.25]
|
< .001
|
1.40[1.06–1.86]
|
0.019
|
90–100
|
2.57[1.94–3.39]
|
< .001
|
2.24[1.66–3.04]
|
< .001
|
1.53[1.09–2.15]
|
0.014
|
100–110
|
3.59[2.46–5.23]
|
< .001
|
2.58[1.68–3.95]
|
< .001
|
1.36[0.82–2.24]
|
0.230
|
>=110
|
4.80[2.80–8.25]
|
< .001
|
2.97[1.64–5.37]
|
< .001
|
1.17[0.58–2.36]
|
0.667
|
P for trend
|
|
< .001
|
|
< .001
|
|
0.162
|
10mmHg increase
|
1.45[1.34–1.56]
|
< .001
|
1.32[1.20–1.44]
|
< .001
|
1.09[0.96–1.23]
|
0.167
|
Model 1 adjusted for age and gender. Model 2 adjusted for age, gender, BMI, Antihypertensive drugs, smoking status, drinking status, diabetes, and site. Model 3 based on model 2, mutually adjusted for SBP and DBP. |
2.2. Association of BP level with MI
Table 3 displays the association of BP level with MI. In multivariable adjusted models( model 2), SBP and DBP showed a linear association with MI, respectively. Compared with participants with SBP 0-130 mmHg, the multivariable adjusted HRs[95% CIs] were 1.42[0.91–2.21], 1.84[1.18–2.87], 2.75[1.53–4.92], and 2.13[0.81–5.64] for MI in participants with SBP 130–140, 140–160, 160–180, and ≥ 180 mmHg, respectively; a significant positive trend in categorized SBP for incident MI (P for trend < 0.001) was observed; each additional 10 mmHg of SBP was associated with a 17% higher probability of developing MI(1.17[1.07–1.27]). However, in a mutually adjusted model(model 3) including both SBP and DBP, the association with the risk of MI was positive for SBP, but not significant for DBP.
Multivariable adjusted restricted cubic spline analyses showed a linear association of SBP (P for linearity < 0.001, model 2; P for linearity < 0.001, model 1) and DBP (P for linearity 0.001, model 2; P for linearity < 0.001, model 1) with MI, respectively (Fig. 1; Figure s1 in appendix). The risk of MI might be higher in low DBP level group(model 2, Fig. 1). Multivariable adjusted restricted cubic spline analyses (model 3) showed a linear association of SBP with MI; no significant linear or nonlinear relationships of DBP with MI were observed (Figure s2 in appendix).
In subgroup analysis, SBP was more strongly associated with MI in 40–65 age group, non-use of antihypertensive drugs group, non-history of diabetes group, non-current smoker, non-current drinker, and rural site group. DBP was more strongly associated with MI in male group, non-use of antihypertensive drugs group, non-history of diabetes group, non-current smoker, and non-current drinker (Figure s3 and Figure s4).
Table 3
Associations of BP with MI
|
Model 1
|
Model 2
|
Model 3
|
HR[95%CI]
|
P
|
HR[95%CI]
|
P
|
HR[95%CI]
|
P
|
SBP(mmHg)
|
|
|
|
|
|
|
0-130
|
1.00
|
|
1.00
|
|
1.00
|
|
130–140
|
1.50[0.99–2.28]
|
0.056
|
1.42[0.91–2.21]
|
0.122
|
1.27[0.80–2.03]
|
0.310
|
140–160
|
2.21[1.47–3.31]
|
< .001
|
1.84[1.18–2.87]
|
0.007
|
1.54[0.93–2.55]
|
0.093
|
160–180
|
3.55[2.09–6.02]
|
< .001
|
2.75[1.53–4.92]
|
0.001
|
2.09[1.05–4.19]
|
0.037
|
>=180
|
3.41[1.37–8.48]
|
0.008
|
2.13[0.81–5.64]
|
0.126
|
1.47[0.49–4.43]
|
0.490
|
P for trend
|
|
< .001
|
|
< .001
|
|
0.062
|
10mmHg increase
|
1.23[1.15–1.32]
|
< .001
|
1.17[1.07–1.27]
|
< .001
|
1.12[0.99–1.25]
|
0.062
|
DBP(mmHg)
|
|
|
|
|
|
|
0–80
|
1.00
|
|
1.00
|
|
1.00
|
|
80–90
|
1.36[0.94–1.98]
|
0.106
|
1.29[0.86–1.94]
|
0.211
|
1.12[0.73–1.70]
|
0.610
|
90–100
|
2.39[1.60–3.55]
|
< .001
|
2.10[1.36–3.24]
|
0.001
|
1.58[0.97–2.59]
|
0.067
|
100–110
|
3.01[1.70–5.33]
|
< .001
|
2.41[1.29–4.50]
|
0.006
|
1.47[0.70–3.12]
|
0.311
|
>=110
|
1.84[0.57–5.87]
|
0.305
|
1.14[0.34–3.82]
|
0.826
|
0.58[0.15–2.19]
|
0.417
|
P for trend
|
|
< .001
|
|
0.004
|
|
0.468
|
10mmHg increase
|
1.36[1.21–1.53]
|
< .001
|
1.26[1.10–1.44]
|
0.001
|
1.11[0.92–1.34]
|
0.274
|
Model 1 adjusted for age and gender. Model 2 adjusted for age, gender, BMI, Antihypertensive drugs, smoking status, drinking status, diabetes, and site. Model 3 based on model 2, mutually adjusted for SBP and DBP. |
2.3. Association of BP level with stroke
Table 4 displays the association of BP level with stroke. In multivariable adjusted models(model 2), SBP and DBP showed a linear association with stroke, respectively. Compared with participants with SBP 0-130 mmHg, the multivariable adjusted HRs[95% CIs](model 2) were 1.44[0.97–2.14], 2.55[1.78–3.65], 4.21[2.69–6.60], and 6.14[3.34–11.29] for stroke in participants with SBP 130–140, 140–160, 160–180, and ≥ 180 mmHg, respectively; a significant positive trend in categorized SBP for incident stroke (P for trend < 0.001) was observed; each additional 10 mmHg of SBP was associated with a 29% higher probability of developing stroke(1.29[1.21–1.38]). However, in a mutually adjusted model(model 3) including both SBP and DBP, the association with the risk of stroke was positive for SBP, but not significant for DBP.
Multivariable adjusted restricted cubic spline analyses showed a linear association of SBP (P for linearity < 0.001, model 2; P for linearity < 0.001, model 1) and DBP (P for linearity < 0.001, model 2; P for linearity < 0.001, model 1) with stroke, respectively (Fig. 1; Figure s1 in appendix). Multivariable adjusted restricted cubic spline analyses (model 3) showed a linear association of SBP with stroke; no significant linear or nonlinear relationships of DBP with stroke were observed (Figure s2 in appendix).
In subgroup analysis, SBP and DBP were more strongly associated with stroke in 40–65 age group, non-use of antihypertensive drugs group, and non-history of diabetes group(Figure s5 and Figure s6).
Table 4
Associations of BP with stroke
|
Model 1
|
Model 2
|
Model 3
|
HR[95%CI]
|
P
|
HR[95%CI]
|
P
|
HR[95%CI]
|
P
|
SBP(mmHg)
|
|
|
|
|
|
|
0-130
|
1.00
|
|
1.00
|
|
1.00
|
|
130–140
|
1.58[1.10–2.26]
|
0.013
|
1.44[0.97–2.14]
|
0.068
|
1.31[0.87–1.97]
|
0.200
|
140–160
|
3.04[2.20–4.19]
|
< .001
|
2.55[1.78–3.65]
|
< .001
|
2.18[1.45–3.28]
|
< .001
|
160–180
|
5.00[3.34–7.50]
|
< .001
|
4.21[2.69–6.60]
|
< .001
|
3.32[1.94–5.67]
|
< .001
|
>=180
|
8.36[4.80-14.54]
|
< .001
|
6.14[3.34–11.29]
|
< .001
|
4.44[2.14–9.22]
|
< .001
|
P for trend
|
|
< .001
|
|
< .001
|
|
< .001
|
10mmHg increase
|
1.34[1.27–1.42]
|
< .001
|
1.29[1.21–1.38]
|
< .001
|
1.25[1.15–1.37]
|
< .001
|
DBP(mmHg)
|
|
|
|
|
|
|
0–80
|
1.00
|
|
1.00
|
|
1.00
|
|
80–90
|
2.54[1.83–3.52]
|
< .001
|
2.20[1.55–3.13]
|
< .001
|
1.71[1.19–2.46]
|
0.004
|
90–100
|
3.01[2.09–4.34]
|
< .001
|
2.51[1.68–3.74]
|
< .001
|
1.54[0.99–2.39]
|
0.055
|
100–110
|
4.99[3.16–7.88]
|
< .001
|
3.41[2.02–5.73]
|
< .001
|
1.51[0.82–2.77]
|
0.188
|
>=110
|
8.12[4.43–14.88]
|
< .001
|
4.94[2.53–9.65]
|
< .001
|
1.48[0.65–3.35]
|
0.347
|
P for trend
|
|
< .001
|
|
< .001
|
|
0.189
|
10mmHg increase
|
1.55[1.41–1.70]
|
< .001
|
1.39[1.25–1.55]
|
< .001
|
1.08[0.93–1.25]
|
0.333
|
Model 1 adjusted for age and gender. Model 2 adjusted for age, gender, BMI, Antihypertensive drugs, smoking status, drinking status, diabetes, and site. Model 3 based on model 2, mutually adjusted for SBP and DBP. |