Abstract
BACKGROUND
The prognostic value of the short-term blood pressure variability (BPV) from the 24-hour ambulatory blood pressure monitoring (ABPM) remains controversial. The present study aimed to investigate the long-term prognostic value of a high BPV in normotensive and hypertensive subjects from a community-based population.
METHODS
A cohort of 624 normotensive and 633 untreated hypertensive Taiwanese participants (overall 669 men, aged 30–79 years) with baseline ABPM and 20-year all-cause and cardiovascular mortality data was drawn from a community-based survey. BPV was assessed by the read-to-read average real variability of the 24-hour diastolic and systolic blood pressure (SBP) (ARVd and ARVs, respectively).
RESULTS
In Cox proportional hazards analysis, ARVd predicted cardiovascular mortality independently of office SBP (hazard ratios (HRs) and 95% confidence intervals (CIs) per 1 SD: 1.31 (1.10–1.55), respectively, bivariate analysis), 24-hour SBP (HR: 1.19, 95% CI: 1.00–1.43), and conventional risk factors (age, sex, smoking, total cholesterol, high-density lipoprotein cholesterol, and fasting blood glucose, HR: 1.40, 95% CI: 1.18–1.67). In subjects with hypertension, a high vs. low ARVd (median: 8.8mm Hg) significantly predicted cardiovascular mortality (HR: 2.11, 95% CI: 1.23–3.62 and HR: 2.04, 95% CI: 1.19–3.51, respectively), when the conventional risk factors plus office SBP or 24-hour SBP were accounted for, respectively. Similar but less significant results were obtained with ARVs. A high ARVd or ARVs did not significantly predict cardiovascular mortality in the normotensive subjects.
CONCLUSIONS
A high short-term BPV is significantly predictive of long-term cardiovascular mortality in untreated hypertensive but not normotensive community-based subjects, independently of office or 24-hour SBP.
Keywords: ambulatory blood pressure, blood pressure variability, cardiovascular mortality, hypertension.
Longstanding recommended targets for hypertension control have been brachial systolic blood pressure (SBP) and diastolic blood pressure (DBP) values taken from a mercury or an oscillometric cuff-based blood pressure monitor in a clinic setting, assuming that the single-visit SBP and DBP (or even the average of 3 readings) may represent the theoretical true underlying levels of blood pressure.1 The “usual blood pressure” hopefully reflects the average blood pressure over time, i.e., a day, several days, weeks, or months, because the impact of chronic increases in the average blood pressure was initially considered to be more substantial than the short-term blood pressure fluctuation.1 However, blood pressure fluctuates moment-by-moment and may change remarkably within periods of a day, month, and year. In fact, the blood pressure variability (BPV) occurring over 24 hours,2 and the long-term visit-to-visit BPV3 have been associated with prognostic relevance independent of the average blood pressure. Additionally, the visit-to-visit BPV could entirely explain the benefit of amlodipine vs. atenolol in the Anglo-Scandinavian Cardiac Outcomes Trial Blood Pressure Lowering Arm (ASCOT-BPLA).4 In particular, the short-term BPV from the 24-hour ambulatory blood pressure monitoring (ABPM) has been considered as an important contributory factor for cardiac and vascular alterations, subclinical organ damage, progression of microalbuminuria/proteinuria and renal dysfunction, and cardiovascular events and mortality.5 In a recent meta-analysis of 8,938 subjects from 11 random general population samples with known cardiovascular outcomes over a median of 11.3 years, higher BPV assessed by the average real variability (ARV) in ABPM predicted cardiovascular events while accounting for the average 24-hour SBP.6 However, the contribution from ARV to risk stratification over and above 24-hour SBP is considered to be trivial6 and further prospective studies with hard end points to define its potential application are required.7
Because BPV varies with blood pressure levels,8,9 the predictive value of BPV may differ between the general population and subjects with high blood pressure.10 Therefore, the present study aimed to investigate the long-term prognostic value of a high BPV in normotensive and hypertensive subjects from a community-based population.
METHODS
Study population
The study cohort of 1,257 normotensive (n = 624) and untreated hypertensive (office SBP ≥ 140mm Hg or DBP ≥ 90mm Hg, n = 633) Taiwanese participants (669 men, mean age 53.2±13.1 years, range 30–79 years) was drawn from a community-based cardiovascular survey conducted in 1992–1993 and has been reported previously.11 In brief, about equal numbers of hypertensive and normotensive subjects in each of the 5 age strata: 30–39 years, 40–49 years, 50–59 years, 60–69 years, and over 70 years, from representative urban and rural regions in Taiwan and in the Kinmen island, were invited to participate in the survey.12 Every participant received a baseline comprehensive cardiovascular evaluation, including complete medical history and physical examination, arterial tonometry, nondirectional Doppler flow velocimetry, and echocardiography.11,13 Overnight fasting serum and plasma samples were drawn for blood chemistry analysis. None of the participants had a previous history of diabetes mellitus, angina pectoris, or peripheral vascular disease, and none had clinical or echocardiographic evidence of other significant cardiac diseases. None of them had received any vasoactive drugs. All participants gave informed consent, and the study was approved by the institutional review board at Johns Hopkins University.
Office and 24-hour ambulatory blood pressures and BPV
Office SBP and DBP were measured manually using a mercury sphygmomanometer by experienced cardiologists. Two or more measurements separated by at least 5 minutes were taken from the right arm of participants after they had been seated for at least 5 minutes. Reported blood pressure values represent the average of at least 2 consecutive measurements. Pulse pressure was calculated as (SBP − DBP).
All study subjects underwent 24-hour ABPM (Model 90217, Spacelabs, Redmond, WA) on a usual working day and they were instructed to work normally and to rest or sleep before midnight and to rise after morning.12 The recorder was programmed to measure blood pressure at 20-minute intervals during the daytime and at 60-minute intervals during the nighttime. Nighttime was defined as 11:00 pm to 6:00 am, and daytime was defined as 7:00 am to 10:00 pm.12 The 24-hour readings were not edited manually, and only records with 80% successful measurements were included in the analysis. BPV of SBP and DBP over 24 hours14,15 was evaluated by the read-to-read ARV.16 ARV was calculated as16:
Where BP = blood pressure, k ranges from 1 to N, N is the number of blood pressure readings in 24 hours. ARVd and ARVs denote ARV of 24-hour DBP and 24-hour SBP, respectively.
Follow-up
The causes and dates of death for the participants who had deceased before 31 December 2013 were obtained by linking our database with the National Death Registry through a unique, personal identification number given to every Taiwan citizen. The National Death Registry database registers valid information based on the certified death certificates coded according to the International Classification of Disease, Ninth Revision (ICD-9). The ICD-9 codes representing cardiovascular deaths were 390–459. The accuracy of cause-of-death coding in Taiwan’s National Death Registry database has been validated.17
Statistical analysis
Statistical analysis was performed using SPSS software (version 17.0, SPSS, Chicago, IL). All data were expressed as proportions or means and SD. Associations of ARVd and ARVs with all-cause and cardiovascular mortalities over a median follow-up of 20 years were examined by Cox proportional hazard regression analysis. Crude and adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated for each SD increment. To investigate whether BPV improves the prediction of cardiovascular mortality in subjects with normotension and hypertension, we dichotomized ARVd and ARVs by median values, in addition to the analysis of the continuous variables. A global comparison of the survival curves obtained with the Kaplan–Meier method using the log-rank test was also included for normotensive and hypertensive subjects with high vs. low BPV. The statistical significance was established at 2-tailed P <0.05.
RESULTS
The characteristics of hypertensive and normotensive subjects are summarized in Table 1. Hypertensive subjects were older and had higher waist circumference, body mass index, total cholesterol, fasting blood glucose, and lower high-density lipoprotein cholesterol levels and estimated glomerular filtration rate, as compared with normotensive subjects.
Table 1.
Demographic, biochemical, office blood pressure, and ambulatory blood pressure and variability variables of the study population stratified by office hypertension status
| Variable | All (n = 1,257) | Hypertension (n = 633) | Normotension (n = 624) | P |
|---|---|---|---|---|
| Age, years | 53.1±13.1 | 55.9±12.4 | 50.2±13.2 | <0.001 |
| Sex (male), n (%) | 669 (53.2%) | 349 (27.8%) | 320 (25.5%) | 0.100a |
| Waist circumference, cm | 85.3±9.2 | 88.7±8.7 | 81.8±8.2 | <0.001 |
| Height, cm | 159.3±8.7 | 159.2±8.8 | 159.5±8.6 | 0.56 |
| Body mass index, kg/m2 | 24.7±3.7 | 25.9±3.8 | 23.5±3.1 | <0.001 |
| Cholesterol, mg/dl | 198.7±38.7 | 203.3±38.1 | 193.8±38.8 | <0.001 |
| HDL-C, mg/dl | 50.8±12.8 | 48.5±12.0 | 53.2±13.2 | <0.001 |
| Fasting glucose, mg/dl | 101.1±27.2 | 105.5±33.8 | 96.3±16.2 | <0.001 |
| Fasting glucose ≥ 126mg/dl, n (%) | 67 (5.3 %) | 57 (9.0 %) | 10 (1.6 %) | <0.001a |
| eGFR, ml/min/1.732 | 92.0±23.2 | 85.5±20.0 | 99.8±24.5 | <0.001 |
| eGFR < 60ml/min/1.732, n (%) | 115 (9.1 %) | 81 (12.8 %) | 34 (5.4 %) | <0.001a |
| Smoking (yes), n (%) | 317 (25.2%) | 150 (23.7%) | 167 (26.8%) | 0.12a |
| Office blood pressure | ||||
| SBP, mm Hg | 136.1±24.4 | 154.8±18.7 | 117.2±11.2 | <0.001 |
| DBP, mm Hg | 83.6±13.9 | 93.3±11.6 | 73.9±8.0 | <0.001 |
| PP, mm Hg | 52.5±17.4 | 61.5±18.6 | 43.3±9.7 | <0.001 |
| Ambulatory blood pressure | ||||
| 24-hour SBP, mm Hg | 126.9±17.5 | 138.5±15.6 | 115.1±9.8 | <0.001 |
| 24-hour DBP, mm Hg | 80.9±11.7 | 88.3±10.7 | 73.3±6.9 | <0.001 |
| 24-hour PP, mm Hg | 46.0±9.4 | 50.2±10.2 | 41.8±6.0 | <0.001 |
| 24-hour HR, bpm | 77.4±8.7 | 77.7±9.1 | 77.2±8.2 | 0.33 |
| Daytime SBP, mm Hg | 128.6±17.9 | 140.4±16.0 | 116.6±10.0 | <0.001 |
| Daytime DBP, mm Hg | 82.4±12.0 | 90.0±11.0 | 74.7±7.1 | <0.001 |
| Daytime PP, mm Hg | 46.3±9.6 | 50.5±10.4 | 41.9±6.1 | <0.001 |
| Daytime HR, bpm | 80.3±9.3 | 80.4±9.8 | 80.2±8.8 | 0.68 |
| Nighttime SBP, mm Hg | 119.8±17.9 | 130.6±16.7 | 108.9±11.1 | <0.001 |
| Nighttime DBP, mm Hg | 74.6±11.8 | 81.3±11.1 | 67.8±7.9 | <0.001 |
| Nighttime PP, mm Hg | 45.2±9.9 | 49.3±10.7 | 41.1±6.8 | <0.001 |
| Nighttime HR, bpm | 67.1±8.4 | 67.8±8.6 | 66.4±8.3 | 0.004 |
| Blood pressure variability | ||||
| ARVd, mm Hg | 8.8±2.8 | 9.7±3.0 | 8.0±2.2 | <0.001 |
| ARVs, mm Hg | 9.7±2.7 | 10.7±2.9 | 8.6±2.1 | <0.001 |
Abbreviations: ARVd, average real variability of 24-hour DBP; ARVs, average real variability of 24-hour SBP; DBP, diastolic blood pressure; HDL-C, high-density lipoprotein cholesterol; eGFR, estimated glomerular filtration rate, HR, heart rate; PP, pulse pressure; SBP, systolic blood pressure.
aBy chi-square test.
As expected, hypertensive subjects had significantly higher office and ambulatory SBP, DBP, and pulse pressure than normotensives. Consequentially, hypertensives also had significantly higher ARVd and ARVs.
During a median follow-up period of 20 years, 355 subjects died, 90 from a cardiovascular cause. Table 2 summarizes the predictive values of the individual risk factors based on univariable and multivariable analyses. Office SBP, 24-hour SBP, ARVd, and ARVs significantly predicted all-cause and cardiovascular mortalities in univariate Cox proportional hazard analysis (Table 2, (A)) and in Cox models adjusting for the conventional cardiovascular risk factors including age, sex, smoking, total cholesterol, high-density lipoprotein cholesterol, and fasting blood glucose (Table 2, (B)). In bivariable Cox models, both ARVd and ARVs significantly predicted mortalities independently of either office SBP (Table 2, (C)) or 24-hour SBP (Table 2, (D)). In Cox models adjusting for the conventional cardiovascular risk factors plus office SBP (Table 2, (E)), only ARVd was predictive of all-cause and cardiovascular mortalities. In Cox models adjusting for the conventional cardiovascular risk factors plus 24-hour SBP (Table 2, (F)), neither ARVd nor ARVs was predictive of mortalities.
Table 2.
Continuous data analysis: hazard ratios and 95% confidence intervals per 1 SD increment of each variable for 20-year all-cause and cardiovascular mortalities (n = 1,257) by unadjusted and adjusted Cox proportional hazards regression models
| Variable | All-cause mortality (n = 355) | Cardiovascular mortality (n = 90) |
|---|---|---|
| A. Univariable analysis | ||
| Age (13.2 years) | 2.92 (2.61–3.26) | 2.84 (2.28–3.53) |
| Sex (male) | 1.28 (1.04–1.58) | 1.14 (0.76–1.71) |
| Smoking (yes) | 2.02 (1.64–2.49) | 2.27 (1.51–3.43) |
| Body mass index (3.7kg/m2) | 0.96 (0.87–1.07) | 1.02 (0.83–1.25) |
| Cholesterol (38.7mg/dl) | 1.08 (0.97–1.20) | 1.27 (1.05–1.55) |
| HDL-C (12.8mg/dl) | 0.92 (0.82–1.02) | 0.81 (0.65–1.02) |
| Fasting glucose (27.2mg/dl) | 1.18 (1.10–1.27) | 1.26 (1.13–1.40) |
| Office SBP (24.4mm Hg) | 1.53 (1.39–1.69) | 2.26 (1.88–2.71) |
| 24-hour SBP (17.54mm Hg) | 1.49 (1.36–1.64) | 2.10 (1.79–2.45) |
| ARVd (2.76mm Hg) | 1.43 (1.32–1.56) | 1.67 (1.43–1.94) |
| ARVs (2.73mm Hg) | 1.52 (1.40–1.65) | 1.81 (1.55–2.10) |
| B. Adjusted for conventional cardiovascular risk factorsa | ||
| Office SBP (24.4mm Hg) | 1.21 (1.09–1.35) | 1.88 (1.52–2.32) |
| 24-hour SBP (17.54mm Hg) | 1.20 (1.08–1.33) | 1.77 (1.48–2.12) |
| ARVd (2.76mm Hg) | 1.20 (1.09–1.32) | 1.40 (1.18–1.67) |
| ARVs (2.73mm Hg) | 1.15 (1.04–1.28) | 1.47 (1.21–1.78) |
| C. Adjusted for office SBP | ||
| 24-hour SBP (17.54mm Hg) | 1.21 (1.04–1.41) | 1.51 (1.16–1.96) |
| ARVd (2.76mm Hg) | 1.28 (1.16–1.40) | 1.31 (1.10–1.55) |
| ARVs (2.73mm Hg) | 1.35 (1.23–1.49) | 1.38 (1.15–1.66) |
| D. Adjusted for 24-h SBP | ||
| Office SBP (24.4mm Hg) | 1.32 (1.13–1.54) | 1.64 (1.23–2.18) |
| ARVd (2.76mm Hg) | 1.26 (1.14–1.39) | 1.19 (1.00–1.43) |
| ARVs (2.73mm Hg) | 1.36 (1.23–1.50) | 1.35 (1.12–1.63) |
| E. Adjusted for conventional cardiovascular risk factors and office SBP | ||
| 24-hour SBP (17.54mm Hg) | 1.11 (0.96–1.28) | 1.44 (1.13–1.85) |
| ARVd (2.76mm Hg) | 1.15 (1.04–1.27) | 1.20 (1.00–1.45) |
| ARVs (2.73mm Hg) | 1.08 (0.97–1.22) | 1.20 (0.97–1.49) |
| F. Adjusted for conventional cardiovascular risk factors and 24-hour SBP | ||
| ARVd (2.76mm Hg) | 1.14 (1.03–1.27) | 1.11 (0.92–1.34) |
| ARVs (2.74mm Hg) | 1.07 (0.95–1.21) | 1.13 (0.90–1.42) |
Parentheses indicate SDs or 95% confidence intervals. All bold values indicate significant P value < 0.05.
Abbreviations: ARVd, average real variability of 24-hour DBP; ARVs, average real variability of 24-hour SBP; DBP, diastolic blood pressure; HDL-C, high-density lipoprotein cholesterol; SBP, systolic blood pressure.
aConventional cardiovascular risk factors include age, sex, smoking, total cholesterol, HDL-C, and fasting blood glucose.
In categorical analysis, high office SBP (≥140mm Hg, according to hypertension guidelines18), high 24-hour SBP (≥130mm Hg, according to hypertension guidelines18), high ARVd (median cutoff: 8.8mm Hg), and high ARVs (median cutoff: 9.3mm Hg) were significantly associated with cardiovascular mortality, without or with adjustment for the conventional cardiovascular risk factors (Table 3).
Table 3.
Categorical data analysis: HR and 95% CIs of high vs. low office or ambulatory 24-hour systolic blood pressure or short-term blood pressure variability for 20-year cardiovascular mortality (n = 1,257) by unadjusted and adjusted Cox proportional hazards regression models
| Unadjusted HR (95% CI) | Adjusted HR (95% CI) | |
|---|---|---|
| Office SBP, ≥ vs. <140mm Hg | 4.80 (2.98–7.75) | 3.11 (1.85–5.23) |
| 24-hour SBP, ≥ vs. <130mm Hg | 4.70 (2.99–7.39) | 2.92 (1.81–4.72) |
| ARVd, ≥ vs. <8.8mm Hg | 4.40 (2.77–7.00) | 2.78 (1.72–4.48) |
| ARVs, ≥ vs. <9.3mm Hg | 4.22 (2.60–6.87) | 2.07 (1.24–3.46) |
Adjusted independent variables: age, sex, smoking, total cholesterol, high-density lipoprotein cholesterol, and fasting blood glucose; cut points: office SBP ≥ 140mm Hg (by consensus); 24-hour SBP ≥ 130mm Hg (by consensus); ARVd ≥ 8.8mm Hg (median) ARVs ≥ 9.3mm Hg (median).
Abbreviations: ARVd, average real variability of 24-hour DBP; ARVs, average real variability of 24-hour SBP; CIs, confidence intervals; DBP, diastolic blood pressure; HRs, hazard ratios; SBP, systolic blood pressure.
When the whole study population was divided into 4 subgroups according to hypertension status defined by office blood pressure (≥140/90mm Hg) and high vs. low ARVd, with the subgroup of normotension and low ARVd as the reference, the subgroup of “hypertension and high ARVd” had significantly greater risks for cardiovascular mortality, without or with adjustment for the conventional cardiovascular risk factors plus office SBP (Table 4, Model 1) or 24-hour SBP (Table 4, Model 2), respectively. When using the subgroup of “hypertension and low ARVd” as the reference, the subgroup of “hypertension and high ARVd” had a significantly greater risk for cardiovascular mortality while accounting for the conventional risk factors plus office SBP (HR: 2.11, 95% CI: 1.23–3.62; Table 4, Model 1) or 24-hour SBP (HR: 2.04, 95% CI: 1.19–3.51; Table 4, Model 2).
Table 4.
Categorical data analysis: HRs and 95% CIs of high vs. low short-term blood pressure variability for 20-year cardiovascular mortality (n = 1,257) in categories of normotension and hypertension by unadjusted and adjusted Cox proportional hazards regression models
| Unadjusted | Model 1 | Model 2 | |
|---|---|---|---|
| HR (95% CI) | HR (95% CI) | HR (95% CI) | |
| Normotension and low ARVd | 1 (reference) | 1 (reference) | 1 (reference) |
| Normotension and high ARVd | 3.86 (1.37–10.84) | 2.63 (0.91–7.59) | 2.47 (0.85–7.14) |
| Hypertension and low ARVd | 4.96 (1.97–12.50) | 1.93 (0.70–5.38) | 2.12 (0.80–5.67) |
| Hypertension and high ARVd | 15.12 (6.53–35.00) | 3.91 (1.47–10.42) | 4.19 (1.65–10.64) |
| Hypertension and low ARVd | 1 (reference) | 1 (reference) | 1 (reference) |
| Hypertension and high ARVd | 3.05 (1.80–5.15) | 2.11 (1.23–3.62) | 2.04 (1.19–3.51) |
| Normotension and low ARVs | 1 (reference) | 1 (reference) | 1 (reference) |
| Normotension and high ARVs | 2.56 (1.11–5.93) | 1.21 (0.50–2.93) | 1.04 (0.43–2.53) |
| Hypertension and low ARVs | 3.10 (1.32–7.30) | 1.03 (0.39–2.74) | 1.25 (0.50–3.11) |
| Hypertension and high ARVs | 9.47 (4.85–18.5) | 1.66 (0.68–4.07) | 1.88 (0.82–4.27) |
| Hypertension and low ARVs | 1 (reference) | 1 (reference) | 1 (reference) |
| Hypertension and high ARVs | 3.05 (1.61–5.80) | 1.77 (0.90–3.45) | 1.70 (0.87–3.34) |
Model 1: adjusted independent variables: age, sex, smoking, total cholesterol, high-density lipoprotein cholesterol, fasting blood glucose and office SBP; Model 2: adjusted independent variables: age, sex, smoking, total cholesterol, high-density lipoprotein cholesterol, fasting blood glucose and 24-hour SBP; cut points: ARVd ≥ 8.8mm Hg (median) ARVs ≥ 9.3mm Hg (median). All bold values indicate significant P value < 0.05.
Abbreviations: ARVd, average real variability of 24-hour DBP; ARVs, average real variability of 24-hour SBP; CIs, confidence intervals; DBP, diastolic blood pressure; HRs, hazard ratios; SBP, systolic blood pressure.
Categorical analyses with high and low ARVs instead of ARVd provided similar but less significant results (Table 4). Results of the categorical analyses for all-cause mortality were similar but less significant and are shown in the Supplementary Tables S1 and S2.
The differential predictive values of a high BPV in subjects with normotension and hypertension defined by office blood pressure are illustrated by Kaplan–Meier survival analysis in Figure 1. Normotensive and hypertensive subjects with a high BPV indicated by a high ARVd or ARVs had a significantly higher risk for cardiovascular mortality than those with a low BPV (Figure 1). The predictive value of a high BPV was much greater in subjects with hypertension than in subjects with normotension.
Figure 1.
Kaplan–Meier survival analyses of cardiovascular (CV) mortality in normotensive and hypertensive subjects with high or low blood pressure variability. (a) High vs. low read-to-read average real variability (ARVd) of the ambulatory 24-hour diastolic blood pressure in normotensive subjects; (b) high vs. low ARVd in hypertensive subjects; (c) high vs. low read-to-read average real variability (ARVs) of the ambulatory 24-hour systolic blood pressure in normotensive subjects; (d) high vs. low ARVs in hypertensive subjects. Hypertension was defined by office blood pressure ≥140/90mm Hg.
DISCUSSION
In this community-based normotensive and untreated hypertensive cohort with a 20-year follow-up, we found higher ABPM derived short-term BPV variables (ARVd and ARVs) were predictive of more cardiovascular mortality independent of either office SBP or 24-hour SBP in the bivariable analysis, and independent of the conventional risk factors in the multivariable analysis. Higher ARVd was predictive of cardiovascular mortality independent of conventional risk factors plus office SBP. Both ARVd and ARVs were no longer significantly and independently associated with cardiovascular mortality when 24-hour SBP was added to the multivariable models of the conventional risk factors. In categorical data analysis, a high vs. low BPV predicted more than 2-folds of cardiovascular mortality independently of the conventional risk factors. In subgroup analysis, a high ARVd significantly and predicted cardiovascular mortality independently of office SBP or 24-hour SBP in addition to the conventional risk factors in subjects with hypertension. Thus, the short-term BPV may be useful in long-term risk stratification in the untreated hypertensive subjects, beyond level of blood pressure.
Although the short-term BPV derived from ABPM has been associated with target organ damage and cardiovascular events,19 its clinical relevance remains to be established.20 In a recent review addressing the predictive value of BPV, over and beyond level of pressure, in randomly selected population samples, the authors found that none of the variability indices substantially refined risk profiling over and beyond the blood pressure level.7 The conclusion was mainly based on the key finding from a meta-analysis of individual data including >8,000 people with an average 11-year follow-up that ARVd (and also ARVs) was a significant and independent predictor of mortality and of cardiovascular and stroke events but the proportion of the risk explained by ARVd was very low.6 Our results from the continuous data analysis were generally consistent with the meta-analysis but with less statistical significance probably due to insufficient power. However, our categorical data analysis revealed that a high BPV may be useful in identifying high risk hypertensives subjects, beyond the level of blood pressure. In effect, the meta-analysis also demonstrated that the association between ARVd and 10-year risk of composite cardiovascular event was stronger with higher levels of 24-hour SBP.6 Therefore, the reported risk stratification power of ARV in the meta-analysis cohort may have been substantially diluted because of the inclusion of a large proportion of normotensive subjects.6
In a hospital-based study involving 550 sequential untreated hypertensive patients with a mean follow-up of around 5 years, high vs. low ARVs of daytime SBP was associated with more cardiovascular events, including fatal and nonfatal myocardial infarction, coronary or peripheral revascularization, heart failure requiring hospitalization, fatal and nonfatal stroke, and renal failure requiring dialysis, (HR: 2.29, 95% CI: 1.06–4.94).10 The daytime ARVs was a nonsignificant predictor of risk when included in the model as a continuous variable (HR: 1.29, 95% CI: 0.98–1.69).10 Our community-based study provides similar results and has the strength to extend that ARVd is better than ARVs as an independent predictor of 20-year cardiovascular mortality in the untreated hypertensives, beyond level of blood pressure. The superiority of ARVd over ARVs is consistent with previous studies.6 In patients with arterial stiffness, the amplitude of fluctuation with blood pressure levels is likely greater for SBP than for DBP so ARVd may be less dependent on 24-hour SBP than ARVs (Supplementary Table S3).
Office blood pressure measurement has been used for the detection and management of hypertension in clinical practice for a very long time. On the other hand, mounting evidence suggests that blood pressure from ABPM is more closely related to hypertension related end organ damage21,22 and has stronger relationship to morbidity and mortality than office blood pressure.23,24 However, in the present study, we found office SBP was very powerful in predicting long-term cardiovascular mortality (Tables 2 and 3). Although 24-hour SBP was significantly predictive of cardiovascular mortality independent of office SBP and conventional risk factors, office SBP in combination with other conventional risk factors was already a powerful risk prediction model to which addition of novel risk factors would hardly improve the model further.21 Moreover, BPV and 24-hour SBP are usually significantly correlated (Supplementary Table S3).10 These may partly explain why BPV as a continuous variable was significantly predictive of cardiovascular mortality independently of office SBP or 24-hour SBP in the bivariable models but was less significant when conventional risk factors were also accounted for.16
Limitations of the present study
We did not collect nonfatal cardiovascular events for the study population. Therefore, our multivariate analyses might suffer from over-adjustment (too many covariates in the presence of a limited event numbers) and insufficient power so the independent predictive value of BPV could not be clearly demonstrated in all analyses. In the present study, considering a 2-sided level of 0.05, we had 66% power to detect the difference of cardiovascular mortality between hypertensive subjects with high vs. low ARVd in the log-transformed HR (2.04), but only had 8.3% power to detect the association (HR: 1.11) between per 1 SD change of ARVd and risk of cardiovascular mortality in the whole population. In addition, the predictive value of BPV for all cardiovascular events cannot be simply inferred from the data for cardiovascular mortality as reported in the present study. History of drug therapy during the long-term follow-up period was not available. Thus, impact of antihypertensive medications on the prognostic values of office SBP, 24-hour SBP, and BPV could not be assessed in the present study.
In conclusion, a high short-term BPV is significantly predictive of long-term cardiovascular mortality in untreated hypertensive but not normotensive community-based subjects, independently of office or 24-hour SBP.
SUPPLEMENTARY MATERIAL
Supplementary materials are available at American Journal of Hypertension (http://ajh.oxfordjournals.org).
DISCLOSURE
The authors declare no conflict of interest.
Supplementary Material
ACKNOWLEDGMENTS
This work was supported in part by grants from the National Science Council (NSC 99-2314-B-010-034 -MY3 and MOST 103-2314-B-010-050-MY2), an intramural grant from the Taipei Veterans General Hospital (grant V104C-140), Research and Development contract NO1-AG-1-2118, grants from the Ministry of Health and Welfare (MOHW104-TDU-B-211-113-003), and the Intramural Research Program of the National Institute on Aging, National Institutes of Health.
REFERENCES
- 1. Rothwell PM. Limitations of the usual blood-pressure hypothesis and importance of variability, instability, and episodic hypertension. Lancet 2010; 375:938–948. [DOI] [PubMed] [Google Scholar]
- 2. Mancia G, Bombelli M, Facchetti R, Madotto F, Corrao G, Trevano FQ, Grassi G, Sega R. Long-term prognostic value of blood pressure variability in the general population: results of the Pressioni Arteriose Monitorate e Loro Associazioni Study. Hypertension 2007; 49:1265–1270. [DOI] [PubMed] [Google Scholar]
- 3. 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]
- 4. Rothwell PM, Howard SC, Dolan E, O’Brien E, Dobson JE, Dahlöf B, Poulter NR, Sever PS; MRC Trial Investigators Effects of beta blockers and calcium-channel blockers on within-individual variability in blood pressure and risk of stroke. Lancet Neurol 2010; 9:469–480. [DOI] [PubMed] [Google Scholar]
- 5. Parati G, Ochoa JE, Lombardi C, Bilo G. Assessment and management of blood-pressure variability. Nat Rev Cardiol 2013; 10:143–155. [DOI] [PubMed] [Google Scholar]
- 6. Hansen TW, Thijs L, Li Y, Boggia J, Kikuya M, Björklund-Bodegård K, Richart T, Ohkubo T, Jeppesen J, Torp-Pedersen C, Dolan E, Kuznetsova T, Stolarz-Skrzypek K, Tikhonoff V, Malyutina S, Casiglia E, Nikitin Y, Lind L, Sandoya E, Kawecka-Jaszcz K, Imai Y, Wang J, Ibsen H, O’Brien E, Staessen JA. Prognostic value of reading-to-reading blood pressure variability over 24 hours in 8938 subjects from 11 populations. Hypertension 2010; 55:1049–1057. [DOI] [PubMed] [Google Scholar]
- 7. Asayama K, Wei FF, Liu YP, Hara A, Gu YM, Schutte R, Li Y, Thijs L, Staessen JA. Does blood pressure variability contribute to risk stratification? Methodological issues and a review of outcome studies based on home blood pressure. Hypertens Res 2015; 38:97–101. [DOI] [PubMed] [Google Scholar]
- 8. Li Z, Snieder H, Su S, Harshfield GA, Treiber FA, Wang X. A longitudinal study of blood pressure variability in African-American and European American youth. J Hypertens 2010; 28:715–722. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Mancia G, Ferrari A, Gregorini L, Parati G, Pomidossi G, Bertinieri G, Grassi G, di Rienzo M, Pedotti A, Zanchetti A. Blood pressure and heart rate variabilities in normotensive and hypertensive human beings. Circ Res 1983; 53:96–104. [DOI] [PubMed] [Google Scholar]
- 10. Pierdomenico SD, Di Nicola M, Esposito AL, Di Mascio R, Ballone E, Lapenna D, Cuccurullo F. Prognostic value of different indices of blood pressure variability in hypertensive patients. Am J Hypertens 2009; 22:842–847. [DOI] [PubMed] [Google Scholar]
- 11. Sung SH, Cheng HM, Wang KL, Yu WC, Chuang SY, Ting CT, Lakatta EG, Yin FC, Chou P, Chen CH. White coat hypertension is more risky than prehypertension: important role of arterial wave reflections. Hypertension 2013; 61:1346–1353. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Chen CH, Ting CT, Lin SJ, Hsu TL, Chou P, Kuo HS, Wang SP, Yin FC, Chang MS. Relation between diurnal variation of blood pressure and left ventricular mass in a Chinese population. Am J Cardiol 1995; 75:1239–1243. [PubMed] [Google Scholar]
- 13. Chen CH, Ting CT, Lin SJ, Hsu TL, Ho SJ, Chou P, Chang MS, O’Connor F, Spurgeon H, Lakatta E, Yin FC. Which arterial and cardiac parameters best predict left ventricular mass? Circulation 1998; 98:422–428. [DOI] [PubMed] [Google Scholar]
- 14. Bilo G, Giglio A, Styczkiewicz K, Caldara G, Maronati A, Kawecka-Jaszcz K, Mancia G, Parati G. A new method for assessing 24-h blood pressure variability after excluding the contribution of nocturnal blood pressure fall. J Hypertens 2007; 25:2058–2066. [DOI] [PubMed] [Google Scholar]
- 15. Parati G, Bilo G, Vettorello M, Groppelli A, Maronati A, Tortorici E, Caldara G, Mancia G. Assessment of overall blood pressure variability and its different components. Blood Press Monit 2003; 8:155–159. [DOI] [PubMed] [Google Scholar]
- 16. Mena L, Pintos S, Queipo NV, Aizpúrua JA, Maestre G, Sulbarán T. A reliable index for the prognostic significance of blood pressure variability. J Hypertens 2005; 23:505–511. [DOI] [PubMed] [Google Scholar]
- 17. Lu TH, Lee MC, Chou MC. Accuracy of cause-of-death coding in Taiwan: types of miscoding and effects on mortality statistics. Int J Epidemiol 2000; 29:336–343. [DOI] [PubMed] [Google Scholar]
- 18. Chiang CE, Wang TD, Ueng KC, Lin TH, Yeh HI, Chen CY, Wu YJ, Tsai WC, Chao TH, Chen CH, Chu PH, Chao CL, Liu PY, Sung SH, Cheng HM, Wang KL, Li YH, Chiang FT, Chen JH, Chen WJ, Yeh SJ, Lin SJ. 2015 guidelines of the Taiwan Society of Cardiology and the Taiwan Hypertension Society for the management of hypertension. J Chin Med Assoc 2015; 78:1–47. [DOI] [PubMed] [Google Scholar]
- 19. Parati G, Pomidossi G, Albini F, Malaspina D, Mancia G. Relationship of 24-hour blood pressure mean and variability to severity of target-organ damage in hypertension. J Hypertens 1987; 5:93–98. [DOI] [PubMed] [Google Scholar]
- 20. Taylor KS, Heneghan CJ, Stevens RJ, Adams EC, Nunan D, Ward A. Heterogeneity of prognostic studies of 24-hour blood pressure variability: systematic review and meta-analysis. PLoS One 2015; 10:e0126375. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Wang TJ, Gona P, Larson MG, Tofler GH, Levy D, Newton-Cheh C, Jacques PF, Rifai N, Selhub J, Robins SJ, Benjamin EJ, D’Agostino RB, Vasan RS. Multiple biomarkers for the prediction of first major cardiovascular events and death. N Engl J Med 2006; 355:2631–2639. [DOI] [PubMed] [Google Scholar]
- 22. Bliziotis IA, Destounis A, Stergiou GS. Home versus ambulatory and office blood pressure in predicting target organ damage in hypertension: a systematic review and meta-analysis. J Hypertens 2012; 30:1289–1299. [DOI] [PubMed] [Google Scholar]
- 23. Staessen JA, Thijs L, Fagard R, O’Brien ET, Clement D, de Leeuw PW, Mancia G, Nachev C, Palatini P, Parati G, Tuomilehto J, Webster J. Predicting cardiovascular risk using conventional vs ambulatory blood pressure in older patients with systolic hypertension. Systolic Hypertension in Europe Trial Investigators. JAMA 1999; 282:539–546. [DOI] [PubMed] [Google Scholar]
- 24. Sega R, Facchetti R, Bombelli M, Cesana G, Corrao G, Grassi G, Mancia G. Prognostic value of ambulatory and home blood pressures compared with office blood pressure in the general population: follow-up results from the Pressioni Arteriose Monitorate e Loro Associazioni (PAMELA) study. Circulation 2005; 111:1777–1783. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.