Abstract
The association between blood pressure variability (BPV) and the risk of all‐cause mortality and cardiovascular diseases (CVD) is not well understood. The Kailuan study is a prospective longitudinal cohort study on cerebrovascular events and cardiovascular factors. In this study, resting blood pressure was measured at baseline and every 2 years from 2006 to 2007. BPV is mainly defined as the coefficient of variation (CV). Hazard ratio (HR), with 95% confidence intervals (CI) was calculated using Cox regression model. Among 52 387 participants, we identified 1817 who ended up with all‐cause death and 1198 with CVD. Each 4.68% increase in BPV was associated with a 13% increase in the risk of mortality (HR = 1.13, 95% CI = 1.09‐1.18) and a 7% increase in CVD (HR = 1.07, 95% CI = 1.02‐1.13), respectively. After adjustment of confounding factors, the HR of comparing participants in the highest versus lowest quartile of CV of systolic blood pressure (SBP) was 1.37 (1.19, 1.57) for all‐cause death, 1.18 (1.01, 1.39) for CVD. Similar results were also observed when BPV was measured by different parameters. We concluded that visit‐to‐visit BPV was associated with all‐cause death and cardiovascular and cerebrovascular events in Chinese general population.
Keywords: all‐cause mortality, blood pressure variability, cardiovascular diseases, epidemiology
1. INTRODUCTION
Hypertension is the leading preventable risk factor for cardiovascular disease and premature death in China. The prevalence of hypertension is high and increasing in China in recent years.1 Blood pressure variability (BPV) refers to the fluctuation of blood pressure in a certain period of time. Visit‐to‐visit BPV represents episodic hypertension, which is usually untreated because the patient's BP may be within the normal range during the requisite repeated readings.2, 3 Previous studies found that long‐term Systolic Blood Pressure Variability (SBPV) is related to the all‐cause mortality and cardiovascular diseases (CVD) in hypertensive population.4, 5, 6 A meta‐analysis of 77 299 Patients suggested that BPV is a predictor of cardiovascular and all‐cause mortality and stroke.7 However, inconsistent results exist in this association, especially when this long‐term BPV was frequently assessed by different parameters and in various populations. Mancia et al8 hold the opinion that long‐term SBPV is unessential to the occurrence of cardiovascular events and all‐cause mortality in hypertensive patients. Similar results were found in a prospective cohort of 2906 adults.9 Notably, most previous studies focused on patients with chronic diseases, while the correlation of BP variability with all‐cause mortality and CVD in the general population is obscure. Moreover, a study focused on the Asian population was rare. In this study, we performed a prospective cohort study to examine the potential associations between BP variability and the risk of all‐cause mortality and CVD in the Chinese general population.
2. METHODS
2.1. Study design and population
The Kailuan study is a prospective population‐based cohort study conducted in the Kailuan community in Tangshan, a large modern city located southeast of Beijing. The design, method, rationale, and examination details of the Kailuan cohort study have been published previously.10, 11, 12, 13
Between June 2006 and October 2007, a total of 101 510 individuals aged 18 to 98 years were recruited in the study, and 57 927 individuals have since attended reexaminations every 2 years, in 2008‐2009 and 2010‐2011. All participants recruited underwent routine history and physical examination, anthropometry, and laboratory assessment.
Among the 57 927 individuals, 5540 participants were excluded from the analysis due to the following reasons: history of myocardial infarction or stroke at baseline, blood pressure information was available for fewer than 3 visits, death, or CVD during the periods 2006‐2007, 2008‐2009, or 2010‐2011. A flowchart of the present cohort study was shown below (Figure 1).
Figure 1.
Flowchart of the study
All participants provided informed consent and the study was performed according to the guidelines of the Helsinki Declaration and was approved by the Ethics Committee of the Kailuan General Hospital and the Beijing Tiantan Hospital.
2.2. Definition and assessment of blood pressure variability
BP was measured by certified nurses with an appropriately sized cuff and a corrected mercury sphygmomanometer. Each participant was required to refrain from smoking cigarettes, drinking coffee or tea, and exercising for at least 30 minutes before BP measurements. During the measurements, the participants sat with their arms and feet flat, and their upper arms at the height of their heart. SBP and DBP were recorded upon auscultation of the first and fifth Korotkoff sounds, respectively. Each participant was measured 3 times with 30 second intervals. If 2 measurements differed by > 5 mm Hg, BP was remeasured. The final BP was calculated by taking the mean of the 3 measurements.
As previously reported, BP variability was defined as the CV (calculated as standard deviation (SD)/mean BP × 100%) of the BP between visits.5, 14 Variability independent of independent variation of the mean (VIM), maximum‐minimum difference (MMD), and the average real variability (ARV) were also calculated to confirm the credibility of the data by performing different parameters.15
In this study, we focused on the visit‐to‐visit SBPV, but not diastolic BPV, because existing results show more evidence supporting the association between SBPV and future cardiovascular events than of diastolic BPV.5, 16
2.3. Data collection of potential covariates
Data on demographic variables (e.g., age, sex, household income, and education) and past medical history were collected using questionnaires. Blood samples were collected from the antecubital vein after an overnight fast. All blood samples were tested using a Hitachi 747 auto‐analyzer at the central laboratory of the Kailuan General Hospital. Blood glucose, triglycerides (TG), high‐density lipoprotein (HDL), low‐density lipoprotein (LDL), uric acid, and high‐sensitivity C‐reactive protein (hs‐CRP) levels were determined using standard laboratory methods. Height and body weight were measured while the subjects were wearing lightweight clothing and body mass index (BMI) was calculated by bodyweight/height.2
Dyslipidemia was defined as LDL ≥ 3.4 mmol/L, or HDL < 1.0 mmol/L, or TG ≥ 1.7 mmol/L, or current treatment with lipid‐lowering therapy.17
2.4. Follow‐up and primary outcomes
After the 3 visits, all participants were followed up until December 31, 2015 or until the event of CVD or death. All‐cause mortality was observed for 1 more year. Participants were reexamined, including face‐to‐face interviews in 2‐year intervals. The examinations were performed by trained physicians who were masked to the baseline data. The data obtained in the examinations were confirmed by checking the medical reports from the 11 local hospitals and from medical insurance companies. For participants not attending the regular reexaminations, the outcome information was obtained directly by checking death certificates from provincial vital statistics offices, hospital discharge summaries, and medical records. During the follow‐up period, each fatal event was collected through review of death certificates from the provincial vital statistics offices, hospital records, medical insurance data, and interviews with next of kin, relatives, or eyewitnesses, when such undertakings were possible.
The outcome was defined as all‐cause death and CVD. All‐cause mortality was defined as death for any reason. CVD was defined as the first occurrence of CVD and either the first nonfatal CVD event or CVD death without a preceding nonfatal event. CVD includes coronary disease, myocardial infarction, and stroke (including cerebral infarction, cerebral hemorrhage, and subarachnoid hemorrhage). Vital status was determined by the review committee by December 31, 2016.10, 18
2.5. Statistical analysis
The characteristics of the participants included in the current study were analyzed after the participants were stratified into quartiles into the following groups according to the CV of SBP. Continuous variables were reported as the mean ± standard deviation (SD) and categorical data were summarized using frequency and percentage. Analysis of variance test (ANOVA) or the Kruskal‐Wallis test was conducted for continuous variables, and the χ2 test was used for categorical variables for global comparisons between the 4 BPV groups. The Kaplan‐Meier method was used to calculate the cumulative incidence of all‐cause death and CVD in each quartile group of CV of BP. A series of cox regression models were used to calculate multivariable‐adjusted HR (95% CI) for all‐cause death and CVD risk, associated with a quartile of CV of BP to adjust for the effects of confounding covariates, with the lowest quartile serving as the reference. Variables with a P‐value < .2 and the well‐established predictors were selected as adjustment covariates into the multivariable analyses. These variables included age, gender, and body mass index, smoking, drinking, physical activity, education level, income level, hypertension, diabetes mellitus, hyperlipidemia, family history of CVD, and use of antihypertensive medicine. All statistical analyses were conducted using SPSS software (version 22.0).
For further analysis and to explore the nonlinear relationships between BPV as a continuous variable with all‐cause mortality and CVD, we used restricted cubic splines with 3 knots at the 25th, 50th, and 75th percentiles of the BPV distribution.19
In order to exclude the effects of antihypertensive, lipid‐lowering, and hypoglycemic drugs on the results, we conducted a sensitivity analysis by repeating the above statistical analysis on participants who did not take antihypertensive drugs, who did not take lipid‐lowering drugs, and who did not take hypoglycemic drugs, respectively.
All analyses were performed using SPSS version 22, except for the analysis of restricted cubic splines, which was performed with the statistical package R version 3.1.2 (https://www.r-project.org/). All tests were 2‐sided, and a P value of <.05 was considered to be statistically significant.
3. RESULTS
3.1. Baseline characters
Among 101 510 participants in the Kailuan study, a total number of 52 387 participants free of myocardial infarction, stroke, and with completed BPV rate data were included in the analysis. The study population was consisted of 76.6% men and 23.4% women, with an average age of 48.82 ± 11.77 years. The baseline characteristics of the 49 123 subjects with missing BPV data were different from the remaining 52 387 adults. The excluded participants were slightly older and the majority were males. They had higher serum uric acid and hypersensitive C‐reactive protein levels. A higher proportion smoked and used alcohol, and many were hypertensive, diabetic, and/or hyperlipidemia. They also had a less education, income, and physical activity levels (all P < .01) compared to the participants included in this study (Table S1).
The characteristics of the study participants between quartiles of CV of SBP are presented in Table 1. The average CV of BPV was 8.00%, ranging from 0.00‐89.00%. During the 5.91 ± 0.69 years of follow‐up, we identified 1817 incidents of death and 1198 CVD cases were identified in 4.95 ± 0.55 years, respectively. We found that there were significant differences in age, sex, SBP, DBP, BMI, Lg hs‐CRP, habits of drinking and smoking, education and income level, and history of diseases among groups (P < .01). Compared to the lower BPV quintile groups, there were less males, current smokers, current drinkers, lower education and income level, but a larger proportion of participants with elder age, active physical activity, hypertension and diabetes than in the higher BPV quintile groups.
Table 1.
Baseline characteristics of the study population according to the quartiles of coefficient of variation of systolic blood pressure
| Quartiles of CV of SBP | P value | |||||
|---|---|---|---|---|---|---|
| Q1 (<4.64%) | Q2 (4.64% to <7.41%) | Q3 (7.41% to <10.58%) | Q4 (≥10.58%) | Total (N = 52 387) | ||
| (N=13 077) | (N = 13 116) | (N = 13 104) | (N = 13 090) | |||
| Age, years | 47.83 ± 11.54 | 47.52 ± 11.78 | 48.88 ± 11.78 | 51.06 ± 11.66 | 48.82 ± 11.77 | <.001 |
| Heart rate, bpm | 73.46 ± 9.87 | 73.23 ± 9.83 | 73.26 ± 9.79 | 73.39 ± 10.35 | 73.34 ± 9.96 | 0.206 |
| SBP, mmHg | 128.71 ± 15.66 | 127.36 ± 18.19 | 128.84 ± 18.88 | 132.68 ± 24.62 | 129.4 ± 19.71 | <.001 |
| DBP, mmHg | 84.01 ± 9.79 | 83.07 ± 10.61 | 83.76 ± 11.02 | 84.97 ± 13.40 | 83.95 ± 11.30 | <.001 |
| BMI, kg/m2 | 25.18 ± 3.43 | 24.99 ± 3.48 | 25.03 ± 3.46 | 25.01 ± 3.52 | 25.05 ± 3.47 | <.001 |
| Serum uric acid, mmol/L | 287.37 ± 83.01 | 285.11 ± 83.33 | 284.44 ± 83.04 | 285.59 ± 83.82 | 285.63 ± 83.31 | 0.031 |
| Lg hs‐CRP, mg/L | 0.71 (0.28,2.00) | 0.71 (0.29,2.00) | 0.71 (0.28,2.10) | 0.80 (0.30,2.30) | 0.74 (0.29,2.10) | <.001 |
| Male, n (%) | 10 382 (79.4%) | 9930 (75.7%) | 10 004 (76.3%) | 9825 (75.1%) | 40 141 (76.6%) | <.001 |
| Alcohol drinking, n (%) | ||||||
| Never | 7251 (55.4%) | 7518 (57.3%) | 7675 (58.6%) | 7973 (60.9%) | 30 417 (58.1%) | <.001 |
| Past | 357 (2.7%) | 354 (2.7%) | 380 (2.9%) | 413 (3.2%) | 1504 (2.9%) | |
| Current | 5469 (41.8%) | 5244 (40.0%) | 5049 (38.5%) | 4704 (35.9%) | 20 466 (39.1%) | |
| Smoking, n (%) | ||||||
| Never | 7750 (59.3%) | 8008 (61.1%) | 8058 (61.5%) | 8234 (62.9%) | 32 050 (61.2%) | <.001 |
| Past | 703 (5.4%) | 619 (4.7%) | 647 (4.9%) | 657 (5.0%) | 2626 (5.0%) | |
| Current | 4624 (35.4%) | 4489 (34.2%) | 4399 (33.6%) | 4199 (32.1%) | 17 711 (33.8%) | |
| Education level, n (%) | ||||||
| Illiteracy/primary | 10 045 (76.8%) | 9809 (74.8%) | 10 167 (77.6%) | 10 578 (80.6%) | 40 569 (77.4%) | <.001 |
| High school | 2024 (15.5%) | 2126 (16.2%) | 1930 (14.7%) | 1705 (13.0%) | 7785 (14.9%) | |
| College or higher | 1008 (7.7%) | 1181 (9.0%) | 1007 (7.7%) | 837 (6.4%) | 4033 (7.7%) | |
| Income level, n (%) | ||||||
| <800¥ | 11 080 (84.7%) | 11 092 (84.6%) | 11 151 (85.1%) | 11 278 (86.2%) | 44 601 (85.1%) | 0.002 |
| 800‐1000¥ | 1060 (8.1%) | 1076 (8.2%) | 992 (7.6%) | 982 (7.5%) | 4110 (7.8%) | |
| ≥1000¥ | 937 (7.2%) | 948 (7.2%) | 961 (7.3%) | 830 (6.3%) | 3676 (7.0%) | |
| Physical activity, n (%) | ||||||
| Never | 1138 (9.0%) | 1183 (9.3%) | 1144 (9.1%) | 1093 (8.7%) | 4558 (9.0%) | 0.113 |
| Occasionally | 9796 (77.6%) | 9739 (76.7%) | 9720 (76.9%) | 9607 (76.7%) | 38 862 (77.0%) | |
| Frequently | 1685 (13.4%) | 1783 (14.0%) | 1775 (14.0%) | 1827 (14.6%) | 7070 (14.0%) | |
| Hypertension, n (%) | 4956 (37.9%) | 4518 (34.4%) | 4869 (37.2%) | 5590 (42.7%) | 19 933 (38.0%) | <.001 |
| Diabetes mellitus, n (%) | 937 (7.2%) | 889 (6.8%) | 961 (7.3%) | 1185 (9.1%) | 3972 (7.6%) | <.001 |
| Dyslipidemia, n (%) | 8765 (67.0%) | 8751 (66.7%) | 8649 (66.0%) | 8805 (67.3%) | 34 970 (66.8%) | 0.15 |
| Family history of MI or stroke, n (%) | 822 (6.3%) | 882 (6.7%) | 827 (6.3%) | 820 (6.3%) | 3351 (6.4%) | 0.367 |
| Use of antihypertensive, n (%) | 992 (7.6%) | 1007 (7.7%) | 1084 (8.3%) | 1332 (10.2%) | 4415 (8.4%) | <.001 |
Data shown as mean ± SD or frequency (percentage).
BMI, body mass index; DBP, diastolic blood pressure; Lg hs‐CRP, high‐sensitivity C‐reactive protein after logarithmic transformation; MI: myocardial infarction; SBP, systolic blood pressure.
3.2. BPV with all‐cause mortality and CVD
The multiple Cox regression analysis of associations between BPV levels with: all‐cause mortality and CVD is shown in Table 2. The analysis suggested that the increase of BPV is accompanied with higher all‐cause mortality and CVD occurrence. The Cox proportional hazards regression model, after adjustment for confounding factors tested at baseline, showed such associations were maintained. Adjusted for age, sex, body mass index, smoking, drinking, physical activity, education level, income level, hypertension, diabetes mellitus, hyperlipidemia, family history of CVD, and use of antihypertensive medicine, compared with the lowest group, the odds ratios for the association between quartile 4 with all‐cause mortality and CVD were 1.41 (1.23, 1.62) and 1.28 (1.09, 1.51), respectively (Model 2). Model 3 adds mean SBP to adjustment, the highest BPV quartile group for the risk of all‐cause mortality and CVD was 1.37 (1.19, 1.57) and 1.18 (1.01, 1.39). Similar results were also observed when CV of SBP was considered as a continuous exposure variable (per each 4.68% increase). The survival curve was showed in Figure 2.
Table 2.
Hazard ratios for all‐cause mortality, CVD by quartiles of coefficient of variation of systolic blood pressure and per SD increases in the coefficient of variation
| Q1 (<4.64%) | Q2 (4.64% to <7.41%) | Q3 (7.41% to <10.58%) | Q4 (≥10.58%) | Change per SD (+4.68%) | |
|---|---|---|---|---|---|
| All‐cause mortality | |||||
| Case (%) | 361 (2.8) | 384 (2.9) | 440 (3.4) | 632 (4.8) | |
| Model 1a | 1.00 (ref) | 1.10 (0.95, 1.27) | 1.12 (0.97,1.29) | 1.41 (1.24, 1.60) | 1.17 (1.12, 1.22) |
| Model 2b | 1.00 (ref) | 1.11 (0.986, 1.30) | 1.13 (0.97, 1.31) | 1.41 (1.23,1.62) | 1.16 (1.11, 1.21) |
| Model 3c | 1.00 (ref) | 1.11 (0.96, 1.30) | 1.13 (0.98, 1.31) | 1.37 (1.19, 1.57) | 1.13 (1.09, 1.18) |
| CVD | |||||
| Case (%) | 265 (2.0) | 269 (2.1) | 271 (2.1) | 393 (3.0) | |
| Model 1a | 1.00 (ref) | 1.05 (0.89,1.25) | 0.99 (0.83,1.17) | 1.31 (1.12,1.53) | 1.14 (1.08,1.20) |
| Model 2b | 1.00 (ref) | 1.03 (0.87,1.23) | 0.96 (0.80,1.14) | 1.28 (1.09,1.51) | 1.12 (1.06,1.19) |
| Model 3c | 1.00 (ref) | 1.01 (0.85, 1.21) | 0.96 (0.80,1.14) | 1.18 (1.01,1.39) | 1.07 (1.02,1.13) |
Case (%): number of events (cumulative incidence%). Per SD increase in the CV of SBP = 4.68%.
Adjusted for age, gender, and body mass index.
Further adjusted for smoking, drinking, physical activity, education level,income level, hypertension, diabetes mellitus, hyperlipidemia, family history of CVD, use of antihypertensive.
Further adjusted for mean of systolic blood pressure.
Figure 2.
Cumulative mortality rate of (A) all‐cause death and (B) CVD to quartile of BPV estimating by Kaplan‐Meier curve
In sensitivity analyses, the significance still remained in participants excluding those who have taken different medical treatment after fully adjustment (Table 3). In participants not taking hypertensives, hazard ratios for all‐cause mortality and CVD associated with per SD increase in the CV of SBP in participants was 1.12 (1.06, 1.17) and 1.06 (1.00, 1.13), respectively. The result is consistent in participants not taking hypertensinves (N = 43 234), not taking antidiabetics (N = 51 301), and not taking lipid‐lowering drugs (N = 49 854). Moreover, when variability of SBP was calculated using MMD, ARV, and VIM, the relationship remained the same except for ARV and CVD (Table S2). HR with 95% CI for all‐cause mortality and CVD estimated by multivariable Cox regression model, with the restricted cubic spline for BPV, was plotted in Figure 3. HR for stroke morbidity increased with the ascending distributions of BPV, which model the effects of BPV on stroke to assess the dose‐response relation. There was significant linear term of BPV (P < .001).
Table 3.
Hazard ratios for All‐cause mortality, CVD by per SD increased of coefficient of variation of systolic blood pressure in participants under different medical treatment
| All‐cause mortality | CVD | |
|---|---|---|
| Not taking antihypertensives (N = 47 972) | 1.12 (1.06,1.17) | 1.06 (1.00,1.13) |
| Not taking diabetics (N = 49 475) | 1.13 (1.08,1.18) | 1.06 (1.01,1.13) |
| Not taking lipid‐lowering drugs (N = 47 847) | 1.13 (1.08,1.18) | 1.07 (1.02,1.13) |
Data were adjusted for age, gender, body mass index, current smoking, current drinking, education level, income level, physical activity, heart rate, diabetes mellitus, dyslipidemia, serum uric acid, high‐sensitivity, C‐reactive protein after logarithmic transformation, history of MI, family history of stroke, and mean of systolic blood pressure.
Figure 3.
Hazard ratios and 95% confidence intervals for BPV with (A) all‐cause mortality and (B) CVD by using restricted cubic spline regression with three knots with placed at the 25th, 50th, and 75th percentiles of BPV; the red line represented HR and blue lines represented 95% CI
No interactions of BPV with age (P = .36), age > 65 (0.68), sex (P = .36), BMI (P = .16), BMI > 25 kg/m2 (P = .56), antihypertensive medication use (P = .26), or diabetes (P = .43) were found.
4. DISCUSSION
In this large prospective cohort study of 52 387 participants, we demonstrated that visit‐to‐visit BPV is associated with an increased risk of all‐cause mortality and CVD. This result persists after the analysis was adjusted for age, sex, body mass index, smoking, drinking, physical activity, education level, income level, hypertension, diabetes mellitus, hyperlipidemia, family history of CVD, use of antihypertensive medication, and mean SBP.
In present major guidelines for hypertension, average BP is the recommended crucial component in the management of CVD.2 Nevertheless, because BP measurements often fluctuate from one visit to another, average BP can sometimes fail to accurately reflect a patient's actual status. More evidence has shown that this variability in visit‐to‐visit BP is not an accidental phenomenon; it is reproducible and, furthermore, has an independent association with all‐cause mortality and CVD.20
Findings from NHANES III showed that long‐term SBPV elevation is an important predictor of all‐cause mortality and cardiovascular events.21 Data from the ADVANCE trial suggested that BPV was significantly associated with all‐cause mortality, independent of mean BP in patients with type 2 diabetes.22 Muntner et al23 did a follow up; 25 814 patients participated in the Antihypertensive and Lipid‐Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) for 2.7‐2.9 years, and found that SBPV was associated with an increased risk of all‐cause mortality and CVD.
However, the number of visits, time interval between visits, and the BP measurement protocols varied widely across these studies might cause discrepancies, given the uncertainty of the association between visit‐to‐visit BPV with all‐cause mortality and CVD. A recent meta‐analysis, focusing on the association between BPV and CVD, demonstrated that increased long term variability in systolic blood pressure was associated with risk of all‐cause mortality (HR = 1.15, 95% CI = 1.09‐1.22).24 However, most of these studies have focused on susceptible populations, but not on the Asian population.
The result of this study suggested that the biennially BPV is associated with the increase of all‐cause mortality and CVD in the Chinese population. This conclusion is consistent with the previous meta‐analysis, including 37 studies and 41 independent cohorts.25 In the Cox regression analysis, we brought baseline resting heart rate into the model and found that the result is in line with unadjusted model. The conclusion indicated that, after taking baseline SBP into consideration, visit‐to‐visit BPV is still an independent risk marker for stroke. Restricted cubic spline was used to model the effect of BPV to assess the dose‐response relation. Hazard ratio (HR) increased with increasing BPV, which indicated that the higher the level of BPV, the higher the HR for mortality and CVD. Further analysis in participants without taking antihypertensive drugs, without taking lipid‐lowering drugs, and without taking hypoglycemic drugs was in accordance with this result. Different parameters of BPV had consolidated conclusion to calculate the relationship between visit‐to‐visit BPV with all‐cause mortality, which provided more evidence. The only exception was that there was no significant relationship between ARV and CVD. The probable reason was that CV, MMD, VIM, and ARV reflect different determinants. SD and CV put emphasis on the extreme values, whereas ARV provides more weight to the consecutive changes that may have more value for predicting CVD.5 ARV may be less sensitive to the relative low sampling frequency of the ambulatory blood pressure monitoring devices than SD, which is mentioned in the initial study.26 In a meta‐analysis of 77 299 patients, SD in the visit‐to‐visit setting, seems to be superior to others in predicting cardiovascular and all‐cause mortality and stroke, independent of mean SBP, while nonsignificant findings were observed for cardiovascular mortality in the only study investigating the visit‐to‐visit SBPV by VIM or ARV.7 This conclusion is consist with our study. This finding may be attributable to the extensive use of this index and more included studies. In this respect, the prognostic value of ARV in the visit‐to‐visit setting, as well as its comparisons with CV and SD, merits future investigation.
The clinical characteristics associated with higher BPV also caught our interest. The result indicated that education and income level had a protective effect on mortality risk, which was in line with previous studies in other areas.27, 28 These characteristics were closely associated with medication adherence and treatment compliance. Patient education is consistently identified as a marker for improved adherence to recommended drug therapy. This may be an inspiration for physicians and governors to recognize that financial supports and health education could improve adherence to prevention protocols among patients with chronic conditions. In this regard, it is important to consider the impact of media on patient education and mobile health technology to save the cost of treatment.29
The biological mechanisms underpinning BPV remains speculative, but are likely to be driven by arterial stiffness, autonomic nervous system activity, and complex causes. Systolic BPV is closely correlated with other vascular surrogate markers, including endothelial function, arterial stiffness, and systemic atherosclerotic burden.30 Moreover, several studies have been published to prove relationships between visit‐to‐visit BPV and markers of vascular dysfunction, the activity of the sympathetic nerves,31 the effect of different antihypertensive drugs,5 and patients’ adherence to treatment of hypertension.32
Our study has a large sample size and provides information on a broad spectrum of biological and demographic fields. However, as with all observational studies, there are some inherent limitations. First, Our definitions of variability are based upon biennially measurements of BP, more frequent assessments could improve our phenotyping of variability and thus strengthen our associations. Physiologically, BP shows variations over more prolonged periods because of differences among days, months, and seasons,33 with a trend for systolic BP to increase over the years and for diastolic BP to display an age‐related biphasic change.34 Although we recorded BP in the same season every 2 years and brought in CV, MMD, ARV, and VIM to define BPV at the same time, physiological cause and pathological variation cannot be clearly defined. Also, BP in the study was measured manually by a mercury sphygmomanometer, which may allow for more human error than measurements by an automated device. Moreover, there is still no consensus on the best way to define visit‐to‐visit BPV. A more reliable method to calculate BPV could be obtained by excluding the influence of physiological variation.35 Additionally, we have adjusted confounding variables when evaluating the BPV recorded in different years, but there were still some factors we failed to take into consideration. Patients’ emotional state and BP measurement circumference may also influence visit‐to‐visit BPV, but these factors are difficult to adjust for. Lastly, unlike a randomized clinical trial, BPV could be associated with substantial morbidity and mortality in the population we excluded, which might be especially vulnerable to bias due to attrition.
Our findings raise the hypothesis that there are possible benefits from therapies directed specifically towards lowering BPV, which must be scrutinized in prospective randomized clinical trials.
5. CONCLUSIONS
Visit‐to‐visit BPV is significantly associated with the increase of all‐cause mortality and CVD in the general population. The clinical implications to our findings need to be explored further. BPV could be a highly reproducible and routinely accessible measurement that might serve in risk‐stratification and therapeutic target of patients.
CONFLICTS OF INTEREST
There are no conflicts of interest.
Supporting information
ACKNOWLEDGMENTS
We gratefully acknowledge all 11 regional hospitals of the Kailuan Medical Group and the project development and management teams in Beijing Tiantan hospital. We would like to show our sincere gratitude to clinicians, statisticians, and imaging and laboratory technicians who contributed to the study. We also thank Yumeng Liu (an English teacher from Yuquan School attached to Capital Normal University) and Yiqun Zhang (a graduate from Duke University) who helped polish the writing of the article.
Dai L, Song L, Li X, et al. Association of visit‐to‐visit blood pressure variability with the risk of all‐cause mortality and cardiovascular events in general population. J Clin Hypertens. 2018;20:280–288. 10.1111/jch.13192
Contributor Information
Shouling Wu, Email: drwusl@163.com.
Yongjun Wang, Email: yongjunwang111@aliyun.com.
REFERENCES
- 1. He J, Gu D, Chen J, et al. Premature deaths attributable to blood pressure in China: a prospective cohort study. Lancet. 2009;374:1765‐1772. [DOI] [PubMed] [Google Scholar]
- 2. O'Brien E, Asmar R, Beilin L, et al. Practice guidelines of the European Society of Hypertension for clinic, ambulatory and self blood pressure measurement. J Hypertens. 2005;23:697‐701. [DOI] [PubMed] [Google Scholar]
- 3. Parati G, Stergiou GS, Asmar R, et al. European Society of Hypertension guidelines for blood pressure monitoring at home: a summary report of the Second International Consensus Conference on Home Blood Pressure Monitoring. J Hypertens. 2008;26:1505‐1526. [DOI] [PubMed] [Google Scholar]
- 4. Hastie CE, Jeemon P, Coleman H, et al. Long‐term and ultra long‐term blood pressure variability during follow‐up and mortality in 14,522 patients with hypertension. Hypertension. 2013;62:698‐705. [DOI] [PubMed] [Google Scholar]
- 5. Rothwell PM, Howard SC, Dolan E, et al. Prognostic significance of visit‐to‐visit variability, maximum systolic blood pressure, and episodic hypertension. Lancet. 2010;375:895‐905. [DOI] [PubMed] [Google Scholar]
- 6. Chowdhury EK, Owen A, Krum H, et al. Systolic blood pressure variability is an important predictor of cardiovascular outcomes in elderly hypertensive patients. J Hypertens. 2014;32:525‐533. [DOI] [PubMed] [Google Scholar]
- 7. Tai C, Sun Y, Dai N, et al. Prognostic significance of visit‐to‐visit systolic blood pressure variability: a meta‐analysis of 77,299 patients. J Clin Hypertens. 2015;17:107‐115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Mancia G, Facchetti R, Parati G, Zanchetti A. Visit‐to‐visit blood pressure variability, carotid atherosclerosis, and cardiovascular events in the European lacidipine study on atherosclerosis. Circulation. 2012;126:569‐578. [DOI] [PubMed] [Google Scholar]
- 9. Gao S, Hendrie HC, Wang C, et al. Redefined blood pressure variability measure and its association with mortality in elderly primary care patients. Hypertension. 2014;64:45‐52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Zhang Q, Zhou Y, Gao X, et al. Ideal cardiovascular health metrics and the risks of ischemic and intracerebral hemorrhagic stroke. Stroke. 2013;44:2451‐2456. [DOI] [PubMed] [Google Scholar]
- 11. Liu X, Cui L, Wang A, et al. Cumulative exposure to ideal cardiovascular health and incident diabetes in a chinese population: the Kailuan study. J Am Heart Assoc. 2016;5:e004132. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Wu J, Chen S, Zhou Y, et al. Non‐high‐density lipoprotein cholesterol on the risks of stroke: a result from the Kailuan study. PLoS One. 2013;8:e74634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Dai H, Lu Y, Wu S, Yuan H. Os 37‐06 visit‐to‐visit variability of blood pressure and risk of stroke: results of the Kailuan Cohort Study. J Hypertens. 2016;34(Suppl 1):e406. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Lau KK, Wong YK, Teo KC, et al. Long‐term prognostic implications of visit‐to‐visit blood pressure variability in patients with ischemic stroke. Am J Hypertens. 2014;27:1486‐1494. [DOI] [PubMed] [Google Scholar]
- 15. Wei FF, Li Y, Zhang L, et al. Beat‐to‐beat, reading‐to‐reading, and day‐to‐day blood pressure variability in relation to organ damage in untreated Chinese. Hypertension. 2014;63:790‐796. [DOI] [PubMed] [Google Scholar]
- 16. Filomena J, Riba‐Llena I, Vinyoles E, et al. Short‐term blood pressure variability relates to the presence of subclinical brain small vessel disease in primary hypertension. Hypertension. 2015;66:634‐640. [DOI] [PubMed] [Google Scholar]
- 17. Wang A, Chen G, Su Z, et al. Carotid intima‐media thickness and cognitive function in a middle‐aged and older adult community: a cross‐sectional study. J Neurol. 2016;263:2097‐2104. [DOI] [PubMed] [Google Scholar]
- 18. Wang A, Chen S, Wang C, et al. Resting heart rate and risk of cardiovascular diseases and all‐cause death: the Kailuan study. PLoS One. 2014;9:e110985. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Rutherford MJ, Crowther MJ, Lambert PC. The use of restricted cubic splines to approximate complex hazard functions in the analysis of time‐to‐event data: a simulation study. J Stat Comput Sim. 2015;85:777‐793. [Google Scholar]
- 20. Muntner P, Joyce C, Levitan EB, et al. Reproducibility of visit‐to‐visit variability of blood pressure measured as part of routine clinical care. J Hypertens. 2011;29:2332‐2338. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Muntner P, Shimbo D, Tonelli M, Reynolds K, Arnett DK, Oparil S. The relationship between visit‐to‐visit variability in systolic blood pressure and all‐cause mortality in the general population: findings from NHANES III, 1988 to 1994. Hypertension. 2011;57:160‐166. [DOI] [PubMed] [Google Scholar]
- 22. Hata J, Arima H, Rothwell PM, et al. Effects of visit‐to‐visit variability in systolic blood pressure on macrovascular and microvascular complications in patients with type 2 diabetes mellitus: the ADVANCE trial. Circulation. 2013;128:1325‐1334. [DOI] [PubMed] [Google Scholar]
- 23. Muntner P, Whittle J, Lynch AI, et al. Visit‐to‐visit variability of blood pressure and coronary heart disease, stroke, heart failure, and mortality: a cohort study. Ann Intern Med. 2015;163:329‐338. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Stevens SL, Wood S, Koshiaris C, et al. Blood pressure variability and cardiovascular disease: systematic review and meta‐analysis. BMJ. 2016;354:i4098. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Diaz KM, Tanner RM, Falzon L, et al. Visit‐to‐visit variability of blood pressure and cardiovascular disease and all‐cause mortality: a systematic review and meta‐analysis. Hypertension. 2014;64:965‐982. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Mena L, Pintos S, Queipo NV, Aizpurua JA, Maestre G, Sulbaran T. A reliable index for the prognostic significance of blood pressure variability. J Hypertens. 2005;23:505‐511. [DOI] [PubMed] [Google Scholar]
- 27. Li H, Meng Q, Sun X, Salter A, Briggs NE, Hiller JE. Prevalence, awareness, treatment, and control of hypertension in rural China: results from Shandong province. J Hypertens. 2010;28:432‐438. [DOI] [PubMed] [Google Scholar]
- 28. Koch E, Romero T, Romero CX, et al. Impact of education, income and chronic disease risk factors on mortality of adults: does ‘a pauper‐rich paradox’ exist in Latin American societies? Public Health. 2010;124:39‐48. [DOI] [PubMed] [Google Scholar]
- 29. Smith SC Jr. Adherence to medical therapy and the global burden of cardiovascular disease. J Am Coll Cardiol. 2016;67:1516‐1518. [DOI] [PubMed] [Google Scholar]
- 30. Lau KK, Wong YK, Chan YH, et al. Visit‐to‐visit blood pressure variability as a prognostic marker in patients with cardiovascular and cerebrovascular diseases–relationships and comparisons with vascular markers of atherosclerosis. Atherosclerosis. 2014;235:230‐235. [DOI] [PubMed] [Google Scholar]
- 31. Palatini P, Julius S. The role of cardiac autonomic function in hypertension and cardiovascular disease. Curr Hypertens Rep. 2009;11:199‐205. [DOI] [PubMed] [Google Scholar]
- 32. Krakoff LR. Fluctuation: does blood pressure variability matter? Circulation. 2012;126:525‐527. [DOI] [PubMed] [Google Scholar]
- 33. Sega R, Cesana G, Bombelli M, et al. Seasonal variations in home and ambulatory blood pressure in the PAMELA population. Pressione Arteriose Monitorate E Loro Associazioni. J Hypertens. 1998;16:1585‐1592. [DOI] [PubMed] [Google Scholar]
- 34. Wolf‐Maier K, Cooper RS, Banegas JR, et al. Hypertension prevalence and blood pressure levels in 6 European countries, Canada, and the United States. JAMA. 2003;289:2363‐2369. [DOI] [PubMed] [Google Scholar]
- 35. Duan JL, Hao CN, Lu W, et al. A new method for assessing variability of 24 h blood pressure and its first application in 1526 elderly men. Clin Exp Pharmacol Physiol. 2009;36:1093‐1098. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials