Author + information
- Received May 27, 2015
- Revision received September 4, 2015
- Accepted September 4, 2015
- Published online November 17, 2015.
- Danielle Tientcheu, MD∗,
- Colby Ayers, MS∗,
- Sandeep R. Das, MD, MPH∗,
- Darren K. McGuire, MD, MHSc∗,
- James A. de Lemos, MD∗,
- Amit Khera, MD, MS∗,
- Norman Kaplan, MD∗,
- Ronald Victor, MD† and
- Wanpen Vongpatanasin, MD∗∗ ()
- ∗Cardiology Division, Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, Texas
- †Hypertension Center, Cedars-Sinai Heart Institute, Los Angeles, California
- ↵∗Reprint requests and correspondence:
Dr. Wanpen Vongpatanasin, Hypertension Section, Cardiology Division, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, Texas 75390.
Background Multiple epidemiological studies from Europe and Asia have demonstrated increased cardiovascular risks associated with isolated elevation of home blood pressure (BP) or masked hypertension (MH). Previous studies have not addressed cardiovascular outcomes associated with MH and white-coat hypertension (WCH) in the general population in the United States.
Objectives The goal of this study was to determine hypertensive target organ damage and adverse cardiovascular outcomes associated with WCH (high clinic BP, ≥140/90 mm Hg; normal home BP, <135/85 mm Hg), MH (high home BP, ≥135/85 mm Hg; normal clinic BP, <140/90 mm Hg), and sustained hypertension (high home and clinic BP) in the DHS (Dallas Heart Study), a large, multiethnic, probability-based population cohort.
Methods Associations among WCH, MH, sustained hypertension, and aortic pulsed wave velocity by magnetic resonance imaging; urinary albumin-to-creatinine ratio; and cystatin C were evaluated at study baseline. Then, associations between WCH and MH with incident cardiovascular outcomes (coronary heart disease, stroke, atrial fibrillation, heart failure, and cardiovascular death) over a median follow-up period of 9 years were assessed.
Results The study cohort comprised 3,027 subjects (50% African Americans). The sample-weighted prevalence rates of WCH and MH were 3.3% and 17.8%, respectively. Both WCH and MH were independently associated with increased aortic pulsed wave velocity, cystatin C, and urinary albumin-to-creatinine ratio. Both WCH (adjusted hazard ratio: 2.09; 95% confidence interval: 1.05 to 4.15) and MH (adjusted hazard ratio: 2.03; 95% confidence interval: 1.36 to 3.03) were independently associated with higher cardiovascular events compared with the normotensive group, even after adjustment for traditional cardiovascular risk factors.
Conclusions In a multiethnic U.S. population, both WCH and MH were independently associated with increased aortic stiffness, renal injury, and incident cardiovascular events. Because MH is common and associated with an adverse cardiovascular profile, home BP monitoring should be routinely performed among U.S. adults.
Home blood pressure (BP) monitoring has been endorsed in many hypertension guidelines as part of standard care to guide hypertension treatment, as it has been widely recognized that clinic BP may not accurately reflect out-of-office BP (1–3). The pattern of discordance between home and clinic BP can be divided into 2 major categories: white-coat hypertension (WCH; elevated office BP with normal ambulatory or home BP), or masked hypertension (MH; elevated ambulatory or home BP with normal office BP) (4). The cardiovascular (CV) prognosis of WCH is controversial. Although some studies have shown increased target organ damage and CV complications in patients with WCH (5–7), others have demonstrated similar left ventricular mass (8) and prognosis when patients with WCH were compared with a normotensive population (9). MH was shown to be associated with an increased risk for CV events in multiple populations in Europe and Asia (5,7,10–13). However, these studies included few subjects of African descent, the racial/ethnic group with the greatest burdens of hypertension and hypertensive target organ damage. Furthermore, CV risks associated with WCH and MH differed depending on the presence or absence of antihypertensive treatment (5,6). Among the treated population, MH is proposed to represent inadequately treated hypertension, whereas patients with WCH are at risk for overtreatment because of persistently elevated office BP (14). However, the prognostic significance of treated and untreated MH and WCH has not been evaluated in the general population in the United States.
Accordingly, we determined the extent of target organ complications and CV prognosis associated with MH and WCH in participants of the DHS (Dallas Heart Study), a multiethnic, probability-based population sample of Dallas County, Texas, adults. The presence or absence of WCH and MH was determined by home and clinic BP measurements obtained in the same subjects using the same instruments and protocols. We also determined CV outcomes on the basis of the presence or absence of antihypertensive treatment prescribed to each participant.
The DHS is a multiethnic, probability-based population sample of Dallas County residents ages 18 to 65 years, established in 2000, as previously described (15). This study was designed to oversample African Americans, with a resultant cohort that includes 54% African Americans and 49% women. The median age of our participants was 43 years. All participants in the DHS provided written informed consent, and the University of Texas Southwestern Medical Center Institutional Review Board approved the study. The first DHS data collection included an in-home visit (n = 6,101) to collect medical history, BP, and anthropometric measurements between 2000 and 2002. During the in-home visit, 5 BP measurements were taken in the seated position using an automatic oscillometric device (series 52,000, Welch Allyn, Inc., Arden, North Carolina). The surveyors verified treatment with antihypertensive medications and the type of antihypertensive treatment by drug, dose, and frequency. Of the initial 6,101 participants who were 30 to 65 years of age, 3,557 agreed to second in-home visits with fasting phlebotomy and first-morning void urine samples collected, with serum cystatin C and urinary albumin-to-creatinine ratio (UACR) measured for the present study. Of these 3,557 participants, 3,027 completed third study visits at the University of Texas Southwestern Medical Center, where clinic BP measurement was conducted in the same fashion as during the in-home visits, using the same oscillometric device. During the clinic visit, magnetic resonance imaging was performed to assess aortic pulsed wave velocity (APWV). All participants who completed 3 visits were invited to return for a follow-up study (DHS-2) between 2007 and 2009. Among these participants, 1,623 who were not on antihypertensive treatment at visit 1 returned for repeated BP measurement in DHS-2. During the DHS-2 visit, BP measurement was repeated in the clinic, using the same BP instruments and protocol (16).
Race/ethnicity was self-reported. Detailed methods for magnetic resonance imaging in the DHS have been reported (17). The average of the third to fifth BP values measured at home was used as the home BP, and the average of the third to fifth BP values measured at our medical center was used as the clinic BP. WCH was defined as normal home BP (<135/85 mm Hg) and elevated clinic BP (≥140/90 mm Hg). MH was defined as elevated home BP (≥135/85 mm Hg) and normal clinic BP (<140/90 mm Hg), sustained hypertension (SH) as elevation in both home BP (≥135/85 mm Hg) and clinic BP (≥140/90 mm Hg), and normotension (NT) as normal home and clinic BP. Chronic kidney disease was defined as estimated glomerular filtration rate <60 ml/min/1.73 m2 or elevated UACR of at least 17 mg/g creatinine in women or 25 mg/g creatinine in men, as previously described (18). Each antihypertensive drug reported by DHS participants was categorized into the one of the following classes: angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, beta-blockers, diuretics, aldosterone antagonists, calcium-channel blockers, alpha antagonists, central-acting antagonists, nitrates, and others. The number of classes of antihypertensive drugs is used as the number of drugs being taken by each participant.
Urine albumin and creatinine were measured in the first morning void urine sample, and the UACR was calculated in milligrams per gram for each participant (18). Measurements of cystatin C were completed with a BNII nephelometer (Dade Behring, Inc., Deerfield, Illinois; now Siemens Healthcare Diagnostics, Inc.) with a particle-enhanced immunonephelometric assay (N Latex Cystatin C, Dade Behring, Inc.) (19). APWV was assessed using a breath-hold, velocity-encoded, phase-contrast gradient-echo sequence acquired perpendicular to the course of the ascending aorta, using a 1.5-T whole-body system (Intera, Philips Medical Systems, Best, the Netherlands), as previously described (20). Mortality data were queried from the National Death Index through December 2010. CV death was defined by codes I00 to I99 of the International Statistical Classification of Diseases-10th Revision. Two overlapping approaches were used to capture nonfatal CV events occurring after enrollment, as previously described (21). First, a detailed health survey regarding interval CV events was administered annually to study participants. Second, quarterly tracking was performed for hospital admissions using the Dallas-Fort Worth Hospital Council Data Initiative Database, a consortium of all acute care hospitals in Dallas County. Primary clinical source documents were reviewed for all suspected nonfatal CV events and were independently adjudicated by an endpoint committee blinded to all study data. Adjudicated CV events included unstable angina, myocardial infarction, coronary artery bypass grafting, percutaneous coronary intervention, stroke, transient ischemic attack, cerebrovascular revascularization, hospitalization for atrial fibrillation or heart failure, and CV death. Follow-up data for both fatal and nonfatal events were complete through December 31, 2010.
Continuous variables are reported as medians and interquartile ranges or mean ± SD, as appropriate, and categorical variables are presented as proportions. To account for sampling strategy and nonparticipation, sample weighting was used to determine the prevalence of WCH and MH in Dallas County and in subsets of the population, as previously described (22). For all other analyses evaluating associations within the DHS cohort, no sample weighting was used. Demographic and clinical variables, as well as CV risk factors, were compared using Spearman correlations, with adjustment for race and sex. All variables retained in the final model had p values <0.05. The Kruskal-Wallis test was used to compare differences in APWV, UACR, and cystatin C among the WCH, MH, SH, and NT groups. The Wilcoxon rank sum test was used for pairwise comparisons. Adjustment for multiple testing between WCH, MH, and SH versus control groups was not performed. Patients with missing data were excluded from the analysis.
Associations of WCH, MH, and SH with clinical endpoints were assessed by multivariable Cox proportional hazards regression. Subjects with histories of CV disease at baseline were excluded from the analysis (n = 227). Subjects without clinical endpoints who were lost to follow-up for nonfatal events were censored at the time of last contact (n = 64; median follow-up 5.0 years). Two multivariate models were assembled, with all variables retained in the final multivariate model with p values <0.05, and included the following covariates: model 1: age, race, sex, diabetes, smoking, body mass index, and total cholesterol level; model 2: model 1 plus history of hypertension alone; model 3: model 1 plus treatment of hypertension alone; and model 4: model 1 plus history of hypertension and treatment of hypertension. All p values were 2 sided, and p values <0.05 were considered to indicate statistical significance. Statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, North Carolina) and Prism version 6.0 (GraphPad, La Jolla, California).
Baseline characteristics of all participants are shown in Table 1, stratified by hypertension category. Baseline characteristics of participants stratified by the presence or absence of antihypertensive medication treatment are shown in Online Tables 1 and 2. The relationship between home and clinic systolic BP is shown in Figure 1. The sample weight–adjusted prevalence rates of WCH, MH, and SH were 3.3%, 17.8%, and 12%, respectively. In untreated participants, the prevalence rates were 2.2%, 13.6%, and 7.8%, respectively. Participants with WCH, MH, and SH were more likely to be black, older, and obese, to be on antihypertensive medications, and to have prevalent hypertension, diabetes mellitus, hyperlipidemia, chronic kidney disease, and/or CV diseases (Table 1).
Associations between WCH and MH and subclinical target organ damage
APWV and cystatin C levels were significantly higher in both the WCH and MH groups compared with the NT group (5.53 ± 1.44 m/s and 5.39 ± 1.40 m/s vs. 4.56 ± 1.33 m/s, respectively, and 0.88 ± 0.13 mg/l and 0.87 ± 0.15 mg/l vs. 0.82 ± 0.14 mg/l, respectively; p < 0.01 for both). Median UACR was also higher in the WCH and MH groups compared with the NT group (16 [interquartile range (IQR): −1 to 31] and 13 [IQR: 0 to 34] vs. 2 [IQR: −1 to 15] mg/g creatinine, respectively; p < 0.01). When the analyses were stratified according to the presence or absence of antihypertensive treatment, subjects with untreated WCH and SH had significantly higher APWVs compared with those with NT, whereas the group with untreated MH was not significantly different from the NT group after multivariate adjustment (Figure 2A). In contrast, among the subgroup treated with antihypertensive medications, WCH, MH, and SH were all associated with higher APWVs than observed in the NT group. WCH and MH were associated with higher levels of cystatin C compared with the NT group in the untreated but not in the treated subgroups after multivariate adjustment, whereas SH was associated with increased cystatin C levels, regardless of treatment status (Figure 2B). WCH, MH, and SH were also associated with higher UACR compared with the NT group, regardless of treatment status (Figure 2C). The number of antihypertensive drugs prescribed in the WCH group was similar to numbers prescribed to the MH and NT groups (1.19 ± 0.69 vs. 1.33 ± 0.74 and 1.02 ± 0.84, respectively; p = NS vs. treated NT). The number of antihypertensive drugs prescribed in the SH group (1.37 ± 0.72) was also not different from that in the MH group (1.33 ± 0.74; p = NS) but was higher than in the treated NT group (p < 0.01) (Table 1).
Associations between WCH and MH and CV outcomes
Over a median follow-up period of 9.4 years (IQR: 9.0 to 9.8 years), 47 CV deaths and 194 composite CV events were recorded. Hazard ratios for composite CV events in participants with WCH, MH, and SH are presented in Table 2 and Online Table 3, and Kaplan-Meier survival curves are presented in Figure 3. In analyses adjusting for age, sex, race/ethnicity, body mass index, diabetes mellitus, smoking, and total cholesterol levels (model 1), WCH was independently associated with higher CV risk compared with the NT group (adjusted hazard ratio: 2.09; 95% confidence interval: 1.05 to 4.15). The association of WCH with CV events remained significant in the model after adjustment for both traditional CV risk factors and antihypertensive medication treatment (model 3, Table 2). After adjustment for both self-reported history of hypertension and prevalent antihypertensive treatment in model 4, the association of WCH with composite CV events trended toward significance (p = 0.051).
Similarly, MH was associated with increased CV risk compared with the NT group after adjustment for traditional risk factors (adjusted hazard ratio: 2.03; 95% confidence interval: 1.36 to 3.03) and in all models that accounted for both self-reported history of hypertension and antihypertensive treatment (Table 2). Associations of WCH and MH with composite CV events were consistent in subgroups (all p values for interaction >0.1) (Figure 4).
To determine if the presence of WCH and MH is associated with progression to hypertension, which may influence CV outcomes, we assessed the association of WCH and MH with incident hypertension in DHS-2. Because only clinic BP was obtained in DHS-2, we determined the incidence of hypertension by the presence of antihypertensive drug treatment. We found that higher proportions of participants with WCH and MH developed incident hypertension requiring prescriptions of antihypertensive medications in DHS-2 compared with the NT group after a median follow-up period of 7 years (56% vs. 45% vs. 14%; p < 0.001) (Online Figure 1).
The major findings of our study are 3-fold. First, MH was common in our cohort, occurring in 18% of participants overall and in 14% of those not receiving antihypertensive treatment, comprising >50% of the prevalent hypertension cases in the cohort. Second, both WCH and MH were associated with markers of target organ damage, including increased aortic stiffness and renal damage, as evidenced by albuminuria and higher cystatin C levels. Third, both WCH and MH were associated with increased CV events in a multiethnic, probability-based population sample of Dallas County adults.
Our study demonstrated a strikingly high prevalence of MH, exceeding the combined prevalence of each of the other 2 classifications of WCH and SH. In analyses from the CARDIA (Coronary Artery Risk Development in Young Adults) study, the investigators reported that the prevalence of MH was only 2.1% to 4.4%, whereas the prevalence of WCH was 3.3% to 3.9% (23). However, participants in the CARDIA study were more than a decade younger than our cohort, and the sample size was limited to a subset of 281 subjects. In contrast, a recently published report from the JHS (Jackson Heart Study), which is limited to African Americans, reported 25.9% prevalence of MH and 7.5% prevalence of WCH (24). The difference in ethnic composition and the use of 24-h ambulatory BP monitoring in the JHS may explain the difference in the study results. Nevertheless, our data are quite consistent with those from a large international registry of ambulatory BP monitoring, in which the prevalence of MH among subjects 40 to 50 years of age was between 12% and 28%, and the reported prevalence of WCH was between 3.6% and 4.1% (25).
Previous studies have demonstrated inconsistent results with regard to CV outcomes associated with WCH. Results from a published meta-analysis conducted by Pierdomenico and Cuccurullo (9) showed that the CV prognosis of patients with WCH was not significantly different from that of those with NT among an untreated population. Similarly, results of analyses from IDACO (International Database on Ambulatory Blood Pressure in Relation to Cardiovascular Outcomes) (6) and Syst-Eur (Systolic Hypertension in Europe) (26) showed that WCH was not associated with an increase in CV events in the treated group. In contrast, results from a meta-analysis by Stergiou et al. (5) showed increased risk for CV events in patients with untreated WCH. Explanations underlying the difference in study results are unknown but may be related to the method of BP monitoring or to the nature of the population studied. The analyses from IDACO used ambulatory BP monitoring to define the presence of WCH, and the results were derived from older subjects (>60 years of age) with isolated systolic hypertension, with uncertain generalizability to a broader age range or spectrum of BP abnormalities. In contrast, the meta-analysis from Stergiou et al. (5) was derived from home BP monitoring from 5 populations in Japan, Finland, Greece, and Uruguay. Our study represents the largest U.S. study to date that has evaluated CV outcomes associated with WCH, as determined by home and clinic BP measurements.
Underlying factors that result in differential BP levels between the in-home and the clinical setting may influence outcomes in patients with WCH and MH. WCH is proposed to result from stress-induced activation of the sympathetic nervous system during encounters with health care providers (27), whereas MH is potentially induced by mental stress at home, excessive consumption of alcohol, caffeine consumption, and cigarette smoking, among other potential confounders (14). Furthermore, MH while on drug treatment is proposed to reflect a shift of phenotype from SH to inadequately treated hypertension (14). In contrast, WCH is thought to lead to overtreatment of hypertension because of persistently elevated clinic BP (14). However, in our study, the number of antihypertensive drugs prescribed in the WCH group was similar to that in the group with controlled hypertension, both at home and in the clinic. Because average home BP in the treated WCH group (Online Table 2) was similar to the treated NT group, our study did not suggest overtreatment of hypertension in the WCH group. Similarly, the number of antihypertensive drugs prescribed to patients with MH was not different from the SH group, whereas both clinic and home BPs were lower in the MH than in the SH group (Online Table 2), which argues against undertreatment of hypertension in MH. Nevertheless, our data extend previous observations by demonstrating increased CV risk of MH and WCH, independent of CV risk factors and antihypertensive drug treatment.
Our study was limited by a relatively small number of participants with WCH. Although the CV risk associated with WCH in the nonblack participants appeared to be lower than in blacks (Figure 4), we did not have adequate statistical power to examine an interaction between race and the impact of WCH on CV outcomes, given the relatively small total number of events in nonblacks (63 composite events).
The median time between the home visit and clinic visit in the MH group was 2.30 months (IQR: 1.08 to 7.79 months), and we cannot exclude the possibility that treatment of hypertension between the home and clinic visits may have led to normalization of office BP and overdiagnosis of MH. However, previous studies have shown that very small percentages of participants in community-based BP screening programs received new diagnosis of hypertension or treatment of hypertension after BP screening, unless additional medical or behavioral interventions were implemented (28–31).
The rates of progression to SH among subjects with WCH and MH are also unknown, because home BP was not measured during the follow-up study in DHS-2. However, higher proportions of participants with WCH and MH developed incident hypertension requiring prescription with antihypertensive medications in DHS-2 compared with the NT group, which may explain the increased risk for CV events in WCH and MH.
Home BP was obtained by surveyors, rather than by self-measurement, which may not truly reflect home BP. The use of local lay, ethnically congruent field staff members should minimize the alerting reaction during home BP measurement, which was further reduced by averaging the last 3 of 5 BP measurements.
The doses of antihypertensive drugs in each class are also unknown. The number of BP medications was estimated from the number of drug classes, rather than by direct assessment. However, it is unlikely that more than 1 drug per class was prescribed to each participant.
Last, statistical analysis in this study was not adjusted for multiple testing between WCH, MH, and SH versus control groups. However, our analysis was performed on the basis of a separate, pre-stated hypothesis suggested by prior published evidence. Because we did not stray into non-hypothesis-driven exploration, we believe this is an acceptable approach.
Our study provides the first direct evidence of increased target organ damage and increased long-term risk for CV complications associated with WCH and MH in a multiethnic population in the United States (Central Illustration). Given the high prevalence of MH in our cohort, which could have been missed with office BP monitoring alone, the present study provides support for routine use of home BP monitoring in U.S. adults with and without antihypertensive drug treatment.
COMPETENCY IN MEDICAL KNOWLEDGE: Patients with normal BP readings in the clinic but elevated levels at home (MH) and those whose BP is elevated in the medical office setting but normal at home (WCH) are at elevated risk for target organ damage and adverse CV events in middle age. MH is more common than both sustained and WCH.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: Home BP monitoring should be routinely performed in adults with established or suspected hypertension.
TRANSLATIONAL OUTLOOK: More studies are needed to determine if antihypertensive drug treatment of masked or WCH improves CV outcomes and to identify optimal treatment regimens for patients with each type of hypertension.
For supplemental tables and a figure, please see the online version of this article.
Dr. Vongpatanasin is supported by the University of Texas Southwestern O’Brien Kidney Center and the Kaplan Chair in Hypertension Research. Dr. Victor is supported by the National Center for Advancing Translational Sciences, UCLA Clinical and Translational Science Institute (grant UL1TR000124), the Lincy Foundation, and the Burns and Allen Chair in Cardiology Research. The DHS was funded by the Donald W. Reynolds Foundation and was partially supported by award UL1TR001105 from the National Center for Advancing Translational Sciences of the National Institutes of Health. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- aortic pulsed wave velocity
- blood pressure
- interquartile range
- masked hypertension
- sustained hypertension
- urinary albumin-to-creatinine ratio
- white-coat hypertension
- Received May 27, 2015.
- Revision received September 4, 2015.
- Accepted September 4, 2015.
- American College of Cardiology Foundation
- Aronow W.S.,
- Fleg J.L.,
- Pepine C.J.,
- et al.
- National Clinical Guideline Centre (UK)
- Pickering T.G.,
- Hall J.E.,
- Appel L.J.,
- et al.
- Stergiou G.S.,
- Asayama K.,
- Thijs L.,
- et al.,
- for the International Database on Home Blood Pressure in Relation to Cardiovascular Outcome (IDHOCO) Investigators
- Franklin S.S.,
- Thijs L.,
- Hansen T.W.,
- et al.,
- for the International Database on Ambulatory Blood Pressure in Relation to Cardiovascular Outcomes Investigators
- Björklund K.,
- Lind L.,
- Vessby B.,
- et al.
- Pierdomenico S.D.,
- Cuccurullo F.
- Niiranen T.J.,
- Asayama K.,
- Thijs L.,
- et al.,
- for the IDHOCO Investigators
- Chandra A.,
- Neeland I.J.,
- Berry J.D.,
- et al.
- Kramer H.,
- Toto R.,
- Peshock R.,
- et al.
- Patel P.C.,
- Ayers C.R.,
- Murphy S.A.,
- et al.
- Jain T.,
- Peshock R.,
- McGuire D.K.,
- et al.,
- for the Dallas Heart Study Investigators
- Muntner P.,
- Lewis C.E.,
- Diaz K.M.,
- et al.
- Diaz K.M.,
- Veerabhadrappa P.,
- Brown M.D.,
- et al.
- Fagard R.H.,
- Staessen J.A.,
- Thijs L.,
- et al.,
- for the Systolic Hypertension in Europe (Syst-Eur) Trial Investigators
- Hamilton W.,
- Round A.,
- Goodchild R.,
- et al.
- Fleming S.,
- Atherton H.,
- McCartney D.,
- et al.