Within-Visit Variability of Blood Pressure and All-Cause and Cardiovascular Mortality Among US Adults

Authors


Paul Muntner, Department of Epidemiology, University of Alabama at Birmingham, 1665 University Boulevard, Suite 230J, Birmingham, AL 35294
E-mail:pmuntner@uab.edu

Abstract

J Clin Hypertens (Greenwich). 2012;14:165–171. ©2012 Wiley Periodicals, Inc.

The association between within-visit variability of systolic blood pressure (SBP) and diastolic blood pressure (DBP) and all-cause and cardiovascular (CVD) mortality was examined using the Third National Health and Nutrition Survey (n=15,317). Three SBP and DBP readings were taken by physicians during a single medical evaluation. Within-visit variability for each participant was defined using the standard deviation of SBP and DBP across these measurements. Mortality was assessed over 14 years (n=3848 and n=1684 deaths from all causes and CVD, respectively). After age, sex, and race-ethnicity adjustment, the hazard ratios (95% confidence intervals) for all-cause mortality associated with the 4 highest quintiles of within-visit standard deviation of SBP (2.00–2.99 mm Hg, 3.00–3.99 mm Hg, 4.00–5.29 mm Hg, and ≥5.30 mm Hg) compared with participants in the lowest quintile of within-visit standard deviation of SBP (<2.0 mm Hg) were 1.04 (0.87–1.26), 1.09 (0.92–1.29), 1.06 (0.88–1.28), and 1.13 (0.95–1.33), respectively (P=.136). The analogous hazard ratios for CVD mortality were 0.95 (0.69–1.32), 0.96 (0.67–1.36), 0.95 (0.74–1.23), and 1.04 (0.80–1.35), respectively (P=.566). No association with mortality was present after further adjustment and when modeling within-visit standard deviation of SBP as a continuous variable. Standard deviation of DBP was not associated with mortality.

The prognostic value of blood pressure (BP) is based on the average of multiple measurements taken during a single or multiple visits.1,2 The American Heart Association recommends that BP be measured ≥2 times during all outpatient encounters and averaged to represent a patient’s BP.3 These guidelines suggest that if the BP values from the first 2 measurements differ by >5 mm Hg, additional measurements should be taken and used to calculate a patient’s BP.

Recent data suggest that BP variability across several visits may be an important risk factor for stroke, coronary heart disease, and all-cause mortality.4–6 Additionally, the diurnal variability of BP during a 24-hour period using ambulatory BP monitoring has been associated with the subsequent occurrence of cardiovascular disease (CVD) events.7–10 However, few data are available on the prognostic importance of variability of BP obtained during a single clinic visit. If associated with outcomes, within-visit variability of systolic BP (SBP) could be a clinically useful measure since it could be assessed at a single visit. Therefore, we sought to determine the association between the variability of BP during a single research study visit, herein referred to as within-visit variability, and all-cause and CVD-related mortality. To do so, we analyzed data on US adults who participated in the Third National Health and Nutrition Examination Survey (NHANES III).

Methods

NHANES III was a stratified, multistage probability survey designed to select a representative sample of the civilian noninstitutionalized US population.11 NHANES III included an in-home interview with 3 BP measurements and a visit to a mobile examination center for a medical evaluation where 3 additional BP measurements were taken. Additionally, approximately 5% of NHANES III participants were selected to attend a second visit to the mobile examination center where BP was measured 3 additional times. Overall, 18,960 adults 20 years and older completed the NHANES III in-home interview and first visit to the mobile examination center between 1988 and 1994. We limited the current analyses to participants with 3 BP measurements during the in-home interview and the first mobile examination center visit (n=15,328). Eleven participants were missing data for mortality follow-up and were excluded. After these exclusions, a total of 15,317 NHANES III participants were included in the current analyses. For a sensitivity analysis on the correlation between within-visit variability and visit-to-visit variability of BP, we limited the analysis to the participants who completed the second visit to the mobile examination center and had 3 BP measurements taken during this visit (n=956). The protocol for NHANES III was approved by the National Center for Health Statistics of the Center for Disease Control and Prevention institutional review board. All participants gave informed consent.

Baseline Data Collection

Demographic and health-related information was collected using a standardized questionnaire during the in-home interview. Use of antihypertensive medications was ascertained via self-report with classes of antihypertensive medications determined through pill bottle review. Antihypertensive medication classes considered for analysis included angiotensin-converting enzyme (ACE) inhibitors, β-blockers, calcium channel blockers, and thiazide-type diuretics. Other classes had too few individuals taking them to provide stable results. During the medical evaluation, height and weight were measured and body mass index was calculated. Diabetes mellitus was defined as a fasting plasma glucose ≥126 mg/dL, a nonfasting plasma glucose ≥200 mg/dL, and/or a self-reported history of diabetes with concurrent use of antidiabetes medication. Serum C-reactive protein (CRP) levels ≥3 mg/L were defined as elevated. Estimated glomerular filtration rate (eGFR) was determined using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation and reduced eGFR was defined as levels <60 mL/min/1.73 m2.12–14 Albuminuria was defined as a urinary albumin to urinary creatinine ratio ≥30 mg/g.14

BP Measurements.  BP was measured manually 3 times during the in-home interview and 3 additional times during each of the 2 visits to the mobile examination center. BP was measured by a trained research assistant during the in-home visit and by a trained clinician during the visits to the mobile examination center. Additional details regarding BP measurement and quality-control procedures are provided in the NHANES III manual of operations (http://www.cdc.gov/nchs/data/nhanes/nhanes3/cdrom/nchs/manuals/bpqc.pdf). The mean and standard deviation (SD) of SBP and diastolic BP (DBP) during each visit were calculated. Using the mean BP at the first visit to the mobile examination center, hypertension was defined as SBP ≥140 mm Hg, DBP ≥90 mm Hg, or antihypertensive medication use. The SD of SBP has been reported to be associated with the mean SBP. Therefore, we also calculated the SD independent of the mean (SDIM) as a second measure of within-visit variability of SBP.15 For the subset of participants who completed the second visit to the mobile examination center, we calculated the visit-to-visit SD of SBP using the means from each of the 3 study visits.

Mortality Follow-Up.  Adult NHANES III participants were followed for mortality through December 31, 2006. Probabilistic matching was used to link NHANES III participants with the National Death Index to ascertain vital status. Matching was based on 12 identifiers for each participant (eg, social security number, sex, date of birth). Identical matching methodology applied to the NHANES I Epidemiological Follow-up Study for validation purposes found that 96.1% of deceased participants and 99.4% of living participants were correctly classified.16 The International Statistical Classification of Diseases and Related Health Problems (ICD)–Ninth Revision (ICD-9) was used for deaths occurring between 1988 and 1998 and ICD-10 for deaths during 1999 and 2006. CVD mortality was defined by any one of ICD-9 codes 390-434 and 436-459 or ICD-10 codes I00-I99.

Statistical Analysis

Variability of SBP has been found to have a stronger association with outcomes than variability of DBP.4,5 Therefore, for the main analyses, we focused on SBP from the first mobile examination center visit. Participant characteristics at baseline were calculated by quintile of within-visit SD of SBP. Tests for linear trend across quintiles were calculated by modeling the median of each quintile as a continuous variable using linear regression for continuous variables or logistic regression for dichotomous variables.

The percentage of participants dying from all causes and CVD were calculated for each quintile of within-visit SD of SBP. Cox proportional hazards models were used to calculate the hazard ratio for all-cause and CVD-related mortality associated with quintile of within-visit SD of SBP with the lowest quintile as the referent category. Initial models included adjustment for age, sex, and race-ethnicity. Subsequent models also included adjustment for physical inactivity, current smoking, body mass index, total cholesterol, diabetes mellitus, eGFR <60 mL/min/1.73 m2, albuminuria ≥30 mg/g, C-reactive protein ≥3 mg/L, history of myocardial infarction, history of stroke, mean SBP, pulse pressure, and use of ACE inhibitors, β-blockers, calcium channel blockers, or thiazide-type diuretics.

The multivariable-adjusted association of within-visit SD of SBP, modeled as a continuous variable, with all-cause and CVD-related mortality was evaluated using Cox proportional hazard models and restricted quadratic splines with knots at the 10th, 50th, and 90th percentiles of the within-visit SD (1.2 mm Hg, 3.1 mm Hg, and 7.0 mm Hg) of the SBP distribution.17 The white-coat effect is positive and larger among individuals with hypertension as compared with those with normotension.18,19 To assess whether hypertension status impacted the association between within-visit variability of BP and mortality, we repeated the above analyses restricted to individuals without hypertension (SBP/DBP <140/90 mm Hg without antihypertensive medication use).

Two sensitivity analyses were conducted. First, the hazard ratios for all-cause and CVD-related mortality associated with quintile of the within-visit SDIM of SBP was calculated. Second, using BP measurements from the in-home study visit, we calculated the hazard ratios for all-cause and CVD-related mortality associated with quintiles of the within-visit SD of SBP. Also, we calculated the hazard ratios for all-cause and CVD-related mortality associated with quintiles of within-visit SD of DBP.

Finally, we assessed the correlation between participants’ within-visit SD of SBP during the in-home visit with their value from the first mobile examination center visit using a Spearman’s correlation coefficient. Additionally, among the participants who attended 3 NHANES III study visits, we calculated the Spearman’s correlation coefficient between within-visit and visit-to-visit SD of SBP.

The proportional hazards assumption of the Cox models was confirmed using Schoenfeld residuals. Data were analyzed using SUDAAN (version 9.0; Research Triangle Institute, Research Triangle Park, NC) to account for the complex sampling design of NHANES III, including unequal probabilities of selection, over-sampling, and nonresponse.

Results

Baseline Characteristics

The range of the within-visit SD of SBP was from 0 mm Hg to 22.5 mm Hg (median, 3.1 mm Hg; 25th–75th percentiles, 2.1–5.0 mm Hg). Participants with higher within-visit SDs of SBP were older and more likely to be non-Hispanic white (Table I). In contrast, with the exception of sex, being physically inactive and having elevated CRP, each of the characteristics studied, was associated with within-visit SD of SBP.

Table I.   Baseline Characteristics by Quintile of Within-Visit SD of SBP From the NHANES III Mobile Examination Center Visit
 Quintile of Within-Visit SD of SBP, mm HgP Value
1 (n=2728)2 (n=3248)3 (n=2920)4 (n=3527)5 (n=2894)
  1. Abbreviations: ACE, angiotensin-converting enzyme; CRP, C-reactive protein; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; MI, myocardial infarction; NHANES III, Third Report of the National Health and Nutrition Examination Survey; PP, pulse pressure; SBP, systolic blood pressure; SD, standard deviation.

Range, mm Hg<2.02.00–2.993.00–3.994.00–5.29≥5.30 
Age, y40.6 (0.7)41.4 (0.5)43.4 (0.4)46.1 (0.6)52.8 (0.8)<.001
Women, %52.256.752.152.353.6.362
Race-ethnicity, %
 Non-Hispanic white74.672.176.677.682.5<.001
 Non-Hispanic black11.712.110.710.08.6<.001
 Mexican American5.75.55.04.63.6<.001
Physically inactive, %22.821.920.920.723.6.563
Current smoker, %28.932.129.726.823.5<.001
Body mass index, kg/m226.4 (0.2)26.0 (0.2)26.6 (0.2)26.6 (0.2)26.8 (0.2).018
Total cholesterol, mg/dL200.2 (1.4)200.1 (1.4)203.5 (1.3)205.6 (1.3)212.5 (1.4)<.001
Diabetes mellitus, %5.05.16.75.58.4<.001
eGFR <60 mL/min/1.73 m2, %3.53.55.46.110.6<.001
Albuminuria ≥30 mg/g, %7.27.38.68.012.4<.001
Elevated CRP, %31.430.532.133.033.2.185
History of MI, %2.12.72.94.16.0<.001
History of stroke, %1.71.41.42.33.0<.001
Mean SBP, mm Hg116.5 (0.6)117.4 (0.5)119.2 (0.5)122.0 (0.6)129.4 (1.2)<.001
Mean DBP, mm Hg72.0 (0.4)72.3 (0.4)72.8 (0.4)73.4 (0.3)74.8 (0.5)<.001
Hypertension, %16.9%17.2%19.4%24.8%40.4%<.001
Mean PP, mm Hg44.5 (0.5)45.1 (0.4)46.4 (0.5)48.7 (0.6)54.6 (1.2)<.001
Antihypertensive medication drug class, %
 ACE inhibitor2.73.13.33.66.5<.001
 β-Blocker4.22.94.45.68.8<.001
 Calcium channel blocker3.13.94.66.18.3<.001
 Thiazide-type diuretic4.54.85.76.812.3<.001

Within-Visit SD of SBP and Mortality

During a median 14 years of follow-up, 3848 participants died. Of these deaths, 1684 were due to CVD. In unadjusted analyses, a graded association was present between higher within-visit SD of SBP and increased all-cause and CVD mortality (Table II). Among participants with a within-visit SD of SBP <2.00 mm Hg, 2.00 to 2.99 mm Hg, 3.00 to 3.99 mm Hg, 4.00 to 5.29 mm Hg, and ≥5.30 mm Hg, 11.3%, 12.6%, 15.3%, 18.3%, and 29.6% died during follow-up, respectively. After adjustment for age, sex, and race-ethnicity, this association was no longer present (P value across quintiles=.136). The results were similar after further adjustment for physical inactivity, current smoking, body mass index, total cholesterol, diabetes mellitus, eGFR <60 mL/min/1.73 m2, albuminuria ≥30 mg/g, CRP ≥3 mg/L, history of myocardial infarction, history of stroke, mean SBP, pulse pressure, and use of ACE inhibitors, β-blockers, calcium channel blockers, or thiazide-type diuretics (P value across quintiles=.690). No association was present between quintile of within-visit SD of SBP and CVD mortality in an age-, sex-, and race-ethnicity–adjusted model (P value across quintiles=.566) or in a fully adjusted model (P value across quintiles=.734).

Table II.   Hazard Ratios for All-Cause and Cardiovascular Mortality by Quintile of Within-Visit SD of SBP From the NHANES III Mobile Examination Center Visit
 Quintile of Within-Visit SD of SBP, mm HgP Value
1 (n=2728)2 (n=3248)3 (n=2920)4 (n=3527)5 (n=2894)
  1. Abbreviations: CVD, cardiovascular disease; SBP, systolic blood pressure; SD, standard deviation. Fully adjusted includes age, sex, race-ethnicity, physical inactivity, current smoking, body mass index, total cholesterol, diabetes mellitus, estimate glomerular filtration rate <60 mL/min/1.73 m2, albuminuria ≥30 mg/g, C-reactive protein ≥3 mg/L, history of myocardial infarction, history of stroke, mean systolic blood pressure, pulse pressure, and use of angiotensin-converting enzyme inhibitors, β-blockers, calcium channel blockers, or thiazide-type diuretics.

Range, mm Hg<2.02.00–2.993.00–3.994.00–5.29≥5.30 
All-cause mortality
 Deaths, No. (%)476 (11.3)608 (12.6)645 (15.3)934 (18.8)1185 (29.6)<.001
 Hazard ratio (95% confidence interval) 
 Age, sex, and race-ethnicity adjusted1 (reference)1.04 (0.87–1.26)1.09 (0.92–1.29)1.06 (0.88–1.28)1.13 (0.95–1.33).136
 Fully adjusted1 (reference)0.99 (0.80–1.24)1.05 (0.86–1.28)1.04 (0.85–1.28)1.03 (0.83–1.26).690
CVD mortality
 Deaths, No. (%)198 (4.8)247 (4.9)277 (5.9)413 (7.6)549 (13.2)<.001
 Hazard ratio (95% confidence interval) 
 Age, sex, and race-ethnicity adjusted1 (reference)0.95 (0.69–1.32)0.96 (0.67–1.36)0.95 (0.74–1.23)1.04 (0.80–1.35).566
 Fully adjusted1 (ref)0.88 (0.62–1.26)0.83 (0.55–1.26)0.91 (0.68–1.21)0.90 (0.67–1.22).734

When within-visit SD of SBP was modeled as a continuous variable using restricted quadratic splines with full multivariable adjustment, no association was present with all-cause or CVD mortality (Figure). No association was present between higher within-visit SD of SBP and all-cause or CVD mortality in the subgroup of individuals without hypertension (data not shown).

Figure FIGURE.

 Multivariable-adjusted association between within-visit variability (standard deviation [SD]) of systolic blood pressure from the Third Report of the National Health and Nutrition Examination Survey (NHANES III) mobile examination center visit and all-cause (top panel) and cardiovascular mortality (bottom panel). Fully adjusted includes age, sex, race-ethnicity, physical inactivity, current smoking, body mass index, total cholesterol, diabetes mellitus, estimate glomerular filtration rate <60 mL/min/1.73 m2, albuminuria ≥30 mg/g, C-reactive protein ≥3 mg/L, history of myocardial infarction, history of stroke, mean systolic blood pressure (SBP), pulse pressure, and use of angiotensin-converting enzyme inhibitors, β-blockers, calcium channel blockers, or thiazide-type diuretics.

The median SDIM was 3.3 mm Hg (range, 0–24.3 mm Hg). In crude analyses, a graded association was present between higher quintile of within-visit SDIM and increased all-cause and CVD mortality (Table SI). These associations were attenuated after age, sex, and race-ethnicity adjustment, and no association was present after full multivariable adjustment (each P value across quintiles >.4).

BP From the In-Home Visit and Visit-to-Visit Variability

The median within-visit SD of SBP from the in-home visit was 2.9 mm Hg (range, 0–29.7 mm Hg). The correlation of within-visit SD of SBP between the in-home visit and the first mobile examination center visit was 0.09. No associations were present between within-visit SD of SBP from the in-home visit and all-cause or CVD mortality after adjustment for age, sex, and race-ethnicity (Table SII). Results were similar after additional multivariable adjustment. The correlation between the within-visit SD of SBP from the first mobile examination center visit and visit-to-visit variability of SBP was 0.12.

Within-Visit SD of DBP and Mortality

The median within-visit SD for DBP was 2.3 mm Hg (range, 0–22.5 mm Hg). Higher quintiles of within-visit SD for DBP were not associated with all-cause or CVD mortality before or after adjustment for age, sex, and race-ethnicity and other potential confounders (Table III).

Table III.   Hazard Ratios for All-Cause and Cardiovascular Mortality by Quintile of Within-Visit SD of DBP From the NHANES III Mobile Examination Center Visit
 Quintile of Within-Visit SD of DBP, mm Hg (range)P Value
1 (<1.15)2 (1.15–2.00)3 (2.01–3.09)4 (3.10–4.24)5 (≥4.25)
  1. Abbreviations: CVD, cardiovascular disease; DBP, diastolic blood pressure; NHANES III, Third Report of the National Health and Nutrition Examination Survey; SD, standard deviation. Fully adjusted includes age, sex, race-ethnicity, physical inactivity, current smoking, body mass index, total cholesterol, diabetes mellitus, estimate glomerular filtration rate <60 mL/min/1.73 m2, albuminuria ≥30 mg/g, C-reactive protein ≥3 mg/L, history of myocardial infarction, history of stroke, mean systolic blood pressure, pulse pressure, and use of angiotensin-converting enzyme inhibitors, β-blockers, calcium channel blockers, or thiazide-type diuretics.

All-cause mortality
 Deaths, No. (%)919 (16.5)639 (17.0)946 (18.1)561 (17.6)783 (18.0).364
 Hazard ratio (95% confidence interval) 
Age, sex, and race-ethnicity adjusted1 (reference)1.02 (0.91–1.15)1.10 (0.97–1.23)1.08 (0.92–1.28)1.03 (0.88–1.19).504
Fully adjusted1 (reference)0.97 (0.83–1.14)1.09 (0.96–1.23)1.03 (0.86–1.24)0.94 (0.82–1.08).699
CVD mortality
 Deaths, No. (%)408 (7.1)268 (7.1)412 (6.9)238 (6.9)358 (8.0).553
 Hazard ratio (95% confidence interval) 
Age, sex, and race-ethnicity adjusted1 (reference)0.96 (0.80–1.16)0.95 (0.78–1.16)0.98 (0.81–1.18)1.01 (0.83–1.23).892
Fully adjusted1 (reference)0.98 (0.79–1.22)0.99 (0.79–1.24)0.92 (0.72–1.18)0.98 (0.80–1.21).735

Discussion

In the current study of US adults, within-visit variability of SBP and DBP were not associated with all-cause or CVD-related mortality. Additionally, the intra-individual correlation between within-visit variability of SBP over time was small. These results suggest that the variability of BP assessed over a very short period (eg, within a clinical encounter) may reflect conditions surrounding the measurement of BP or measurement errors rather than reflecting a true biologic factor with prognostic importance.

The concept that BP variability has prognostic value for all-cause mortality and CVD events is not new.7,10,20,21 Several recent studies have focused on long-term visit-to-visit variability of BP and outcomes.4,5 Strong associations between higher visit-to-visit variability of SBP and outcomes including stroke, coronary heart disease and all-cause mortality have been found in these studies. For example, comparing the highest to lowest decile of the SD of SBP over 7 visits in the United Kingdom Transient Ischaemic Attack (UK-TIA) trial, the hazard ratio for stroke was 6.22 (95% confidence interval [CI], 4.16–9.29) after multivariable adjustment, including mean SBP.4 Additionally, in a recent study of NHANES III, higher levels of visit-to-visit variability of SBP were associated with increased all-cause mortality.5 Specifically, among 956 US adults with BP measured 3 times during 2 months in NHANES III, the multivariable adjusted hazard ratios for all-cause mortality associated with an SD of SBP of 4.80 mm Hg to 8.34 mm Hg and ≥8.35 mm Hg vs <4.80 mm Hg, were 1.57 (95% CI, 1.07–2.18) and 1.50 (95% CI, 1.03–2.18), respectively.

Short-term diurnal BP variability has also been estimated as the SD from noninvasively monitored 24-hour ambulatory BP.22 Studies of patients with hypertension and individuals from the general population have found associations between BP variability on ambulatory monitoring and CVD morbidity and mortality.20,23–25 In a study of 2649 patients with hypertension, each SD of nighttime SBP was associated with a multivariable adjusted hazard ratio for coronary heart disease events of 1.51 (95% CI, 1.06–2.16).26 Additionally, among 1542 Japanese adults, higher daytime SBP variability was associated with a progressively higher hazard ratio for CVD mortality after multivariable adjustment.20 Specifically, the hazard ratios associated with daytime SBP variability of 11.5 mm Hg to 13.9 mm Hg, 13.9 mm Hg to 15.8 mm Hg, 15.8 mm Hg to 18.8 mm Hg, and >18.8 mm Hg vs <11.5 mm Hg were 1.22 (95% CI, 0.44–3.38), 2.08 (0.85–5.07), and 2.69 (95% CI, 1.18–6.13), respectively. However, other studies have not found independent associations between BP variability on ambulatory monitoring and CVD outcomes.23 While the clinical significance of BP variability on ambulatory monitoring is uncertain, variability in ambulatory BP has several features that distinguish it from variability obtained through clinic-measured BP.22 Ambulatory BP is measured in the natural environment, automated measurements are taken, and it provides many more readings to estimate variability.

Few data have been published on the prognostic importance of within-visit BP variability and outcomes. In the Anglo-Scandinavian Cardiac Outcomes Trial-Blood Pressure Lowering Arm (ASCOT-BPLA), higher within-visit SD SBP was associated with an increased stroke risk.4 Specifically, the hazard ratio for incident stroke comparing the highest with the lowest decile of within-visit SD of SBP, adjusted for mean SBP, was 1.52 (95% CI, 1.09–2.13). However, this association was substantially weaker than the association between visit-to-visit variability of SBP and stroke risk. It has long been recognized that the physical conditions under which BP is measured may affect the observed levels.3 These include cuff size relative to arm size, cuff position relative to the heart, posture, physical activity, the state of relaxation of the subject, talking and environmental noise, and ambient temperature. Additionally, inaccuracy of the BP measurement devices may contribute to within-visit variability of BP.27

Recent studies have highlighted the utility of visit-to-visit variability for identifying high-risk patients, conducting risk prediction, and guiding choice of antihypertensive agents. The lack of an association of within-visit BP variability with all-cause and CVD-related mortality has important implications for both patient management and future research on visit-to-visit variability and outcomes. The importance of obtaining several BP measurements in order to accurately capture a patient’s “true” BP at a clinical encounter has been emphasized.28 While BP will vary between these measurements, data from the current study suggest that this variability is not useful for prognosis and that mean BP derived from readings during one visit or multiple visits is more important. Further, in the context of understanding the prognostic importance of visit-to-visit variability of BP, data from the current study suggest that within-visit variability is not clinically significant. Also, within-visit variability is not a good proxy for visit-to-visit variability. Finally, the underlying mechanisms appear to be different for within-visit, diurnal, and visit-to-visit variability of BP. Within-visit variability of BP may be primarily due to conditions surrounding the BP measurement that produce transient and not physiological changes in BP. In contrast, diurnal and visit-to-visit variability may have a biologic underpinning (eg, impaired baroreflex regulation of BP).3

Study Limitations and Strengths

The current study should be interpreted in the context of known and potential limitations. Within-visit variability of BP was calculated based on only 3 BP measurements. It may be possible to derive a more reliable estimate of variability (eg, SD) if more measurements were available. BP was measured manually. While the equipment was routinely calibrated and staff members were re-trained periodically, BP measurements obtained in this manner are subject to substantial error.29 However, this is how BP is often measured in the clinical setting. The investigation of within-visit variability of SBP using automated devices was beyond the scope of the current study. Additionally, nonfatal outcomes were not available for analysis. Thus, we were unable to assess the relationship between within-visit variability of BP and the incidence of nonfatal events. Despite these limitations, the current analysis has many strengths including the large sample size, use of a population-based study, and protocol-driven BP measurements, as well as the long duration of follow-up.

Conclusions

The findings from the present study suggest that short-term within-visit variability of BP is not associated with an increased risk for all-cause or CVD mortality. Additionally, within-visit variability of BP for a given patient was not reproducible. This is in marked contrast to studies of visit-to-visit variability of SBP, wherein strong associations with outcomes and higher levels of reproducibility have been reported.15 The current study suggests that research on BP variability should focus on measurements taken over longer periods.

Ancillary