Evidence reviews for the US Preventive Services Task Force (USPSTF) use an analytic framework to visually display the key questions that the review will address to allow the USPSTF to evaluate the effectiveness and safety of a preventive service. The questions are depicted by linkages that relate interventions and outcomes. A dashed line indicates a health outcome that immediately follows an intermediate outcome. BP indicates blood pressure; CVD, cardiovascular disease; ESRD, end-stage renal disease; PAD, peripheral artery disease.
KQ indicates key question; IPD-MA, independent patient–data meta-analysis.
aReasons for exclusion: Aim: Study aim not relevant. Setting: Study not conducted in a relevant primary care or out-of-office setting. Outcomes: Study did not have relevant outcomes or had incomplete outcomes. Population: Highly selected populations who do not represent a primary screening populations and populations treated for hypertension with medication. Intervention: Study used an excluded intervention or screening approach. Study design: Study did not use an included design. Comparator: Study did not use ambulatory blood pressure monitoring reference standard (KQ2, KQ3). Quality: Study did not meet criteria for fair or good quality. Country: Study conducted in a country not identified as “very high” on the 2015 Human Development Index. Publication type: Conference abstract.
eMethods. Literature Search Strategies for Primary Literature
eTable 1. Inclusion and Exclusion Criteria
eTable 2. Study-Design Quality Rating Criteria
eTable 3. Study and Baseline Population Characteristics of Included Studies for KQ2, by Author
eTable 4. Test Characteristics of Included Studies for KQ2, by Author
eTable 5. Test Accuracy Results of Included Studies for KQ2, by Author
eTable 6. Study and Baseline Population Characteristics of Included Studies for KQ3, by Author
eTable 7. OBPM Test Characteristics of Included Studies for KQ3, by Author
eTable 8. OBPM Test Accuracy Results of Included Studies for KQ3, by Author
eTable 9. HBPM Test Characteristics of Included Studies for KQ3, by Author
eTable 10. HBPM Test Accuracy Results of Included Studies for KQ3, by Author
eTable 11. Self-OBPM Test Characteristics of Included Studies for KQ3, by Author
eTable 12. Self-OBPM Test Accuracy Results of Included Studies for KQ3, by Author
eTable 13. Truncated ABPM Test Characteristics of Included Studies for KQ3, by Author
eTable 14. Truncated ABPM Test Accuracy Results of Included Studies for KQ3, by Author
eTable 15. High Level Summary of Included Studies for KQ4, by Outcome
Customize your JAMA Network experience by selecting one or more topics from the list below.
Guirguis-Blake JM, Evans CV, Webber EM, Coppola EL, Perdue LA, Weyrich MS. Screening for Hypertension in Adults: Updated Evidence Report and Systematic Review for the US Preventive Services Task Force. JAMA. 2021;325(16):1657–1669. doi:10.1001/jama.2020.21669
Hypertension is a major risk factor for cardiovascular disease and can be modified through lifestyle and pharmacological interventions to reduce cardiovascular events and mortality.
To systematically review the benefits and harms of screening and confirmatory blood pressure measurements in adults, to inform the US Preventive Services Task Force.
MEDLINE, PubMed, Cochrane Collaboration Central Registry of Controlled Trials, and CINAHL; surveillance through March 26, 2021.
Randomized clinical trials (RCTs) and nonrandomized controlled intervention studies for effectiveness of screening; accuracy studies for screening and confirmatory measurements (ambulatory blood pressure monitoring as the reference standard); RCTs and nonrandomized controlled intervention studies and observational studies for harms of screening and confirmation.
Data Extraction and Synthesis
Independent critical appraisal and data abstraction; meta-analyses and qualitative syntheses.
Main Outcomes and Measures
Mortality; cardiovascular events; quality of life; sensitivity, specificity, positive and negative predictive values; harms of screening.
A total of 52 studies (N = 215 534) were identified in this systematic review. One cluster RCT (n = 140 642) of a multicomponent intervention including hypertension screening reported fewer annual cardiovascular-related hospital admissions for cardiovascular disease in the intervention group compared with the control group (difference, 3.02 per 1000 people; rate ratio, 0.91 [95% CI, 0.86-0.97]). Meta-analysis of 15 studies (n = 11 309) of initial office-based blood pressure screening showed a pooled sensitivity of 0.54 (95% CI, 0.37-0.70) and specificity of 0.90 (95% CI, 0.84-0.95), with considerable clinical and statistical heterogeneity. Eighteen studies (n = 57 128) of various confirmatory blood pressure measurement modalities were heterogeneous. Meta-analysis of 8 office-based confirmation studies (n = 53 183) showed a pooled sensitivity of 0.80 (95% CI, 0.68-0.88) and specificity of 0.55 (95% CI, 0.42-0.66). Meta-analysis of 4 home-based confirmation studies (n = 1001) showed a pooled sensitivity of 0.84 (95% CI, 0.76-0.90) and a specificity of 0.60 (95% CI, 0.48-0.71). Thirteen studies (n = 5150) suggested that screening was associated with no decrement in quality of life or psychological distress; evidence on absenteeism was mixed. Ambulatory blood pressure measurement was associated with temporary sleep disturbance and bruising.
Conclusions and Relevance
Screening using office-based blood pressure measurement had major accuracy limitations, including misdiagnosis; however, direct harms of measurement were minimal. Research is needed to determine optimal screening and confirmatory algorithms for clinical practice.
Hypertension is highly prevalent and one of the most important risk factors for cardiovascular disease (CVD).1-3 Blood pressure can be modified with lifestyle interventions,4-6 and good-quality randomized clinical trials (RCTs) demonstrate the effectiveness of antihypertensive pharmacological treatments to reduce CVD and total mortality.7,8 While office-based screening for hypertension in adults has been standard of care in the US for decades,9 office-based methods may misclassify individuals (white coat or masked hypertension). Contemporary research in blood pressure measurement has considered the potential benefits of out-of-office or novel office-based measurement modalities.
The aim of this updated systematic review was to inform an update of the 2015 US Preventive Services Task Force (USPSTF) recommendation on screening for hypertension in adults (A recommendation).10 This systematic review addressed the benefits and harms of screening for hypertension in adults, test accuracy of initial office-based screening measurements, and methods of confirmatory blood pressure measurement in those who initially screen positive.
This review addressed 4 key questions (KQs) as shown in Figure 1. Methodological details including study selection, a list of excluded studies, additional data analysis methods, and sensitivity analyses are available in the full evidence report.11
MEDLINE, PubMed (publisher supplied records), the Cochrane Central Register of Controlled Trials, and CINAHL were searched through August 17, 2019, to identify literature published after the previous review for the USPSTF12 (eMethods in the Supplement). The scope of this update differs from that of the 2015 review12 in that this review analyzed specificity and sensitivity of hypertension screening and confirmation, required ambulatory blood pressure measurement as the reference standard, included patients with diabetes, and did not address prognosis associated with various blood pressure measurement modalities. All included studies in the prior review and a subset of previously excluded studies were also evaluated, as well as reference lists of other systematic reviews and individual patient–data meta-analyses.13-15 ClinicalTrials.gov and the World Health Organization International Clinical Trials Registry Platform were searched for relevant ongoing trials. Active surveillance was conducted through March 26, 2021, via article alerts and targeted journal searches to identify major studies that might affect the conclusions or understanding of the evidence. No new studies were identified.
Investigators reviewed 21 741 unique citations and 544 full-text articles against a priori eligibility criteria (Figure 2; eTable 1 in the Supplement). All studies were required to enroll untreated adults or stratify results by treatment status and to have been conducted in countries rated as “very high” on the 2015 Human Development Index.16 Eligible populations for KQ2 (initial screening) were unselected based on blood pressure, whereas KQ3 populations (confirmatory screening) were preselected for having at least 1 elevated blood pressure measurement identified by clinic-based screening.
For KQ1 (screening), RCTs and nonrandomized controlled intervention studies were included that reported changes in health outcomes as a result of screening for hypertension compared with no screening. Eligible health outcomes were all-cause and cardiovascular mortality, cardiovascular disease events, symptomatic peripheral artery disease, vascular dementia, end-stage renal disease, and quality of life.
For KQs 2 and 3 (test accuracy), test accuracy studies comparing an initial office blood pressure measurement (OBPM) (KQ2) or confirmatory measurement modality (KQ3) with any ambulatory blood pressure monitoring (ABPM) reference standard were included. Attended and unattended automated office-based blood pressure (AOBP) measurements were eligible OBPM subtypes considered for all questions. The selection of the ABPM reference standard was based on a 2015 systematic review conducted for the USPSTF that concluded that ABPM was associated with cardiovascular events independently of OBPM and thus could serve as a reference standard.12 Other investigators have confirmed this finding.17
Confirmatory methods examined in KQ3 included repeated OBPM, self-OBPM (measurement performed by a patient in the office setting), home blood pressure measurement (HBPM), or kiosk. For KQ2a and KQ3a, included studies reported accuracy of protocol variations compared with an ABPM reference standard (eg, more vs fewer OBPM measures, more vs fewer days of HBPM). Studies needed to report sensitivity and specificity or provide enough data to calculate these values.
For KQ4 (harms), RCTs, nonrandomized controlled intervention studies, and cohort studies were included for the outcomes of quality of life, psychological effects of labeling, and absenteeism. Cross-sectional studies were additionally included for the outcome of ABPM tolerability.
Two reviewers independently assessed the methodological quality of eligible studies. Disagreements were resolved by consensus and, if needed, consultation with a third reviewer. Each study was assigned a quality rating of “good,” “fair,” or “poor,” according to the USPSTF’s study design–specific criteria (eTable 2 in the Supplement).18 Studies rated poor quality because of serious methodological shortcomings were excluded.18 One reviewer abstracted descriptive and outcome data from each included study into standardized evidence tables; a second checked for accuracy and completeness.
Results for KQ1 and KQ4 were analyzed qualitatively because of the small number of included studies reporting individual outcomes.
For test accuracy studies (KQ2 and KQ3), the primary outcomes of interest were sensitivity and specificity. For quantitative pooling, only studies that used both systolic blood pressure (SBP) and diastolic blood pressure (DBP) in their definition of hypertension were included because of relevance to current clinical practice. Because there is a lack of consensus on thresholds recommended by guidelines, thresholds were selected based on values most commonly reported in primary studies: 140/90 mm Hg for OBPM, 135/85 mm Hg for daytime ABPM, 130/80 for 24-hour ABPM, and 135/85 mm Hg for HBPM. Additional results for less commonly reported thresholds are available in the full evidence report.11 In quantitative analysis of KQ2 (initial screening), only studies measuring OBPM at a single visit were included; 2 additional studies measuring blood pressure at multiple visits were included in a sensitivity analysis.19,20 Results for KQ3 (confirmatory measurement) were stratified by the type of confirmatory measure (repeat OBPM, HBPM, self-OBPM, AOBP, and kiosk). Data were sufficient for quantitative syntheses for OBPM and HBPM modalities only; other modalities were qualitatively synthesized. For all pooled analyses, a bivariate model was used to model sensitivity and specificity simultaneously, thus accounting for the correlation between these variables.
Stata version 15.1 (StataCorp) was used for all analyses. All significance testing was 2-sided, and results were considered statistically significant if the P value was .05 or less.
The aggregate strength of evidence was assessed for each KQ using the approach described in the Methods Guide for Effectiveness and Comparative Effectiveness Reviews, based on the number, quality, and size of studies and the consistency and precision of results between studies.21
In total, 52 studies reported in 81 articles were included (Figure 2).19,20,22-100 For all KQs, additional descriptive and outcome data are available in the full report.
Key Question 1. Does screening for hypertension in adults improve health outcomes?
There were no population-based trials comparing hypertension screening with no screening. One good-quality community-based cluster RCT (n = 140 642) conducted in Canada examined the effectiveness of a multicomponent CVD health promotion program on CVD health outcomes when hypertension screening was the primary intervention.47 The community clusters received either the Cardiovascular Health Awareness Program (CHAP) intervention or no intervention. In the CHAP communities, residents 65 years and older were invited to participate in community pharmacy-based blood pressure screenings using an automated instrument and complete a standardized risk profile. Participants received their risk profile, risk-specific educational materials, and local community resource information. At 1-year follow-up, the intervention communities had a reduction in the number of hospital admissions per 1000 for composite events (rate ratio, 0.91 [95% CI, 0.86-0.97]). There were 3.02 fewer annual hospital admissions for CVD per 1000 persons in the intervention group compared with the control group (intervention group, –2.25 per 1000 persons; control group, 0.77 per 1000 persons). There were no statistically significant differences in all-cause mortality among admitted residents (rate ratio, 0.98 [95% CI, 0.92-1.03]; intervention group, –1.47 per 1000 persons; control group, 1.42 per 1000 persons) or in-hospital cardiovascular mortality (rate ratio, 0.86 [95% CI, 0.73-1.01]; intervention group, –0.47 per 1000 persons; control group, 0.2 per 1000 persons).
Key Question 2. What is the accuracy of OBPM during a single encounter as initial screening for hypertension compared with the reference standard (ABPM)?
Twenty fair- to good-quality studies (n = 12 614) examined the test accuracy of OBPM for initial screening for hypertension compared with ABPM (eTable 3 in the Supplement).19,20,23,26,29,32,35-38,42,43,45,46,49,54-56,61-63,67,68,70,75-78,80,86,87,95-97 Participants in the studies were primarily recruited from community-based samples. Only 5 were conducted in the US. Overall, participants represented a wide range of demographic and clinical characteristics, including a large range of blood pressures. The prevalence of hypertension as defined by ABPM in the included studies reflected population heterogeneity and ranged from 12.6%54 to 88.9%.70
Index test measurement protocols were heterogeneous and deviated somewhat from current commonly performed protocols in US practice. Studies mostly used mercury sphygmomanometers with blood pressure measured by the manual auscultatory method, had participants rest for 5 minutes prior to measurement, and used the mean of multiple measurements (up to 9 measurements) at a single sitting (eTable 4 in the Supplement). Most other protocol characteristics were sparsely or variably reported. Studies most commonly used an office blood pressure of greater than 140/90 mm Hg or of 140/90 mm Hg or greater as the diagnostic threshold for the index test (17/20 studies).19,20,23,32,35,38,46,49,54,55,67,70,75,80,87,95,96 Several studies additionally reported accuracy for other thresholds,35,37,54,70,80 and 2 studies used SBP-alone or DBP-alone thresholds.36,78 Only 1 study reported accuracy for an OBPM threshold of 130/80 mm Hg or greater,70 the diagnostic threshold recommended in the 2017 American College of Cardiology/American Heart Association guideline.101 While all but 1 study78 reported that 24-hour ABPM was conducted, most (13) studies used daytime ABPM as a reference standard (eTable 4 in the Supplement). Only 1 study had low risk of bias for all domains and was rated as good quality.55 All other studies were rated fair quality and many had at least medium risk of bias for patient selection, conduct of the index test, and conduct of the reference test.
Meta-analysis of 15 studies using SBP/DBP thresholds and measuring blood pressure at a single visit (n = 11 309) showed a pooled sensitivity of 0.54 (95% CI, 0.37-0.70) and pooled specificity of 0.90 (95% CI, 0.84-0.95) (Figure 3; eTable 5 in the Supplement). Substantial clinical and methodologic heterogeneity among the included studies contributed to considerable statistical heterogeneity not explained by any single participant or test characteristic. Among this set of studies, positive predictive values and negative predictive values ranged widely, from 0.35 to 0.97 and from 0.25 to 0.97, respectively. False-positive and false-negative rates likewise ranged widely (false-positive rate range, 0%-30%; false-negative rate range, 8%-100%). A sensitivity analysis adding 2 studies measuring blood pressure at multiple visits19,20 rendered the same point estimate but with slightly narrower confidence intervals (sensitivity, 0.53 [95% CI, 0.37-0.68]; specificity, 0.91 [95% CI, 0.85-0.95]).
Three additional studies (n = 1268) could not be included in the meta-analysis (eTable 5 in the Supplement). These included 1 study of attended AOBP35 with insufficient reporting for pooling showing sensitivity consistent with the pooled analysis but lower specificity (0.74 [95% CI, 0.66-0.82]) and 2 studies that used SBP-only or DBP-only thresholds.36,78
Four studies (n = 1467) reported results for multiple OBPM thresholds (eTable 5 in the Supplement).37,54,78,80 These studies consistently showed increased sensitivity and decreased specificity as thresholds are lowered. One study reported accuracy for an OPBM threshold of 130/80 mm Hg or greater in addition to 140/90 mm Hg or greater but also lowered the reference standard threshold; therefore, accuracy between the 2 OBPM thresholds cannot be directly compared.70 The resulting sensitivity for the OPBM threshold of 130/80 mm Hg or greater compared with the 130/80 mm Hg or greater daytime ABPM reference standard was 0.56 (95% CI, 0.50-0.61), with specificity of 0.89 (95% CI, 0.83-0.93).
Key Question 2a. What screening protocol characteristics define the best test accuracy?
Substantial clinical and methodological heterogeneity among the 20 included KQ2 studies precluded analysis of protocol differences across studies as explanations for differences in accuracy. Four of the 20 included KQ2 studies reported accuracy for within-study comparisons of protocol characteristics.35,54,78,80 No consistent pattern of test accuracy was identified related to the number of measures and visits used for screening.
Key Question 3. What is the accuracy of confirmatory blood pressure measurement in adults who initially screen positive for hypertension compared with the reference standard (ABPM)?
Eighteen fair- to good-quality studies (n = 57 128) examined the diagnostic accuracy of confirmatory blood pressure measurements compared with an ABPM reference standard in adults with a previously detected elevated OBPM (eTable 6 in the Supplement).25,28,30,33,34,40,44,51,52,57,65,66,69,74,81,88,90,99 The Spanish ABPM Registry included 45 020 untreated individuals and represents much of the included evidence for this question.28 Only 2 studies were conducted in the US.30,44 Participants in the studies included patients referred by primary care physicians to blood pressure clinics because of borderline or elevated blood pressures, consecutive patients referred to ABPM or hypertension clinics, or individuals newly diagnosed as hypertensive by OBPM and not yet treated. Overall, the participants represented a wide range of demographic and clinical characteristics (eTable 6 in the Supplement). The prevalence of hypertension as defined by ABPM among this preselected population ranged from 47%74,99 to 80%.69 Two of the included studies were rated as good quality, with low risk of bias for all domains.65,90 All other studies were rated fair quality.
Four confirmatory blood pressure measurement modalities were examined for this KQ: repeated office blood pressure measurement (repeat OBPM), twice-daily home blood pressure measurement for 3 to 7 days (HBPM), measurement performed by a patient in the office setting (self-OBPM), and a truncated 6-hour ambulatory blood pressure measurement (truncated ABPM).
The majority of evidence (13/18 studies) was for repeat OBPM.27,33,34,40,44,51,52,57,65,81,88,90,99 As in KQ2, most OBPM confirmatory measurements were obtained with the patient seated with at least 5 minutes’ rest, attended by personnel, taken with a mercury sphygmomanometer, used a diagnostic threshold of 140/90 mm Hg or greater, and were conducted at a single visit (eTable 7 in the Supplement). Other protocol details varied widely. Meta-analysis of 8 OBPM confirmation studies (n = 53 183) reporting SBP/DBP thresholds showed a pooled sensitivity of 0.80 (95% CI, 0.68-0.88) and a pooled specificity of 0.55 (95% CI, 0.42-0.66) with high heterogeneity (Figure 4; eTable 8 in the Supplement).27,44,52,57,65,81,88,90 Among these 8 studies, positive predictive values ranged from 0.61 to 0.88 and negative predictive values from 0.30 to 0.82. False-positive rates ranged from 15% to 65% and false-negative rates from 10% to 65%. Five studies did not contribute to the meta-analysis because they used SBP-only or DBP-only index thresholds, reference test thresholds, or both, that are not relevant to current clinical practice or did not provide sufficient data for pooling; these studies similarly reported large variations in accuracy (eTable 8 in the Supplement).33,40,44,51,99 One study reported results for multiple OBPM thresholds with increased sensitivity and decreased specificity as thresholds are lowered.34 No included study reported accuracy for an OBPM threshold of 130/80 mm Hg or greater.
Four studies (n = 1001) examined HBPM as a confirmatory method.25,65,66,69 In these studies, participants were instructed to measure blood pressure for 3 to 7 days in the morning and evening in the seated position after a rest period of usually 5 minutes (eTable 9 in the Supplement). Meta-analysis of these 4 HBPM confirmation studies with a threshold of 135/85 mm Hg or greater (n = 1001) showed a pooled sensitivity of 0.84 (95% CI, 0.76-0.90) and pooled specificity of 0.60 (95% CI, 0.48-0.71) (Figure 4; eTable 10 in the Supplement). Positive predictive values ranged from 0.68 to 0.94 and negative predictive values from 0.46 to 0.86. False-positive rates ranged from 22% to 50% and false-negative rates from 7% to 24%. Two studies reported accuracy for multiple HBPM thresholds.25,69 These studies consistently showed increased sensitivity and decreased specificity as index test thresholds are lowered.
Two studies (n = 698) evaluated an index test in which a participant used an HBPM device to take their own blood pressure in an office setting (self-OBPM) (eTable 11 in the Supplement).66,74 While many fundamental device and protocol characteristics were similar among these studies, thresholds were not comparable, and measurements were unattended by staff in 1 study. Only 1 study used SBP/DBP thresholds relevant to current clinical practice and reported high sensitivity (0.92) and low specificity (0.25) (eTable 12 in the Supplement). The positive predictive value in that study was 0.59 and the negative predictive value was 0.72. The false-positive rate was 75% and the false-negative rate was 8%.
One study (n = 263) reported the accuracy of a truncated (6-hour) ABPM compared with a full 24-hour ABPM test (eTable 13 in the Supplement).30 Sensitivity and specificity were 0.94 and 0.76, respectively, for the subgroup (n = 126) for whom the ABPM indication was borderline hypertension (eTable 14 in the Supplement). Sensitivity and specificity were 0.89 and 0.70, respectively, for the subgroup (n = 137) with suspected white coat hypertension.
Two studies (n = 564) reported the accuracy of multiple confirmation methods against the same ABPM reference standard.65,66 One study (n = 361) reported the accuracy of repeat OPBM and HBPM compared with a daytime ABPM reference standard.65 Sensitivity was high and similar for both index tests (0.85 [95% CI, 0.80-0.88] for OBPM and 0.87 [95% CI, 0.83-0.91] for HBPM). Specificity was low for both modalities (0.43 [95% CI, 0.33-0.54] for OBPM and 0.61 [95% CI, 0.51-0.71] for HBPM). The second study (n = 203) reported the accuracy of HBPM and self-OBPM compared with a daytime ABPM reference standard.66 Sensitivity was high and similar for both index tests (0.93 [95% CI, 0.86-0.97] for HBPM and 0.92 [95% CI, 0.85-0.96] for self-OBPM). Specificity was low for both modalities, with self-OBPM being substantially worse (0.50 [95% CI, 0.40-0.61] for HBPM and 0.25 [95% CI, 0.16-0.35] for self-OBPM).
Key Question 3a. What confirmation protocol characteristics define the best test accuracy?
Five of 18 confirmation studies reported within-study comparisons of protocol characteristics on accuracy.33,44,66,69,74 Evidence on protocol variations for any one confirmation modality was sparse, but very limited evidence from 2 studies (n = 459) may suggest that for HBPM, additional days of measurement beyond 5 do not improve accuracy.66,69
Key Question 4. What are the harms of screening for hypertension in adults?
Thirteen fair- to good-quality studies (n = 5150) examined the harms of screening and diagnosis of hypertension.22,24,39,53,58-60,64,73,82,85,91,93,94 Evidence for KQ4 is derived from heterogeneous populations and studies of limited quality largely performed 2 or more decades ago (eTable 15 in the Supplement). The limited existing evidence suggests that screening is associated with no decrement in quality of life or psychological distress,24,58,82,89,93 and the scant evidence on screening’s effect on absenteeism is mixed.39,73,85 ABPM follow-up testing is associated with minor adverse events including temporary sleep disturbance and bruising.53,60,64,79,89,91,94 Inaccurate diagnoses (false-positive and false-negative results) are also considered harms of screening and confirmation and have been discussed under KQ2 and KQ3 results.
This study reviewed the benefits and harms of screening for hypertension in adults, as well as the accuracy of tests; a summary of the evidence by key question is provided in the Table. The lack of contemporary population-based trials solely evaluating hypertension screening may be expected; such trials would not be considered feasible or ethical given that hypertension screening is standard practice and there is a robust evidence base linking asymptomatic hypertension treatment to improved CVD outcomes.102-107 Thus, the focus of this review was on the accuracy of screening (KQ2) and confirmatory (KQ3) blood pressure measurements, protocol variations that may influence accuracy (KQ2a and KQ3a), and the harms of screening and confirmation of hypertension (KQ4).
To our knowledge, this is the only published systematic review comparing the accuracy of office-based screening with an ABPM gold standard (KQ2). In the context of hypertension confirmation, the results of the present systematic review on the accuracy of confirmation (KQ3) are reasonably consistent with data from the International Database of Ambulatory Blood Pressure in relation to Cardiovascular Outcome (n = 4997) and other systematic reviews of confirmation, even though other reviews have included mixed populations of treated and untreated individuals and populations with and without previous elevated OBPMs.14,15,108-110 The highly variable specificities in these reviews of confirmation likely reflect heterogeneity in populations and measurement protocols.
Any hypertension screening algorithm using measurement modalities other than ABPM alone will incur a considerable number of missed cases of masked hypertension as well as treatment of white coat hypertension. However, the clinical significance of the poor accuracy of OBPM is largely unknown. Subsequent consequences of poor OBPM accuracy could include delays in the identification and treatment of masked hypertension. For white coat hypertension, poor OBPM accuracy could result in unnecessary treatment and exposure to adverse effects or conversely a treatment benefit. Meta-analyses suggest that for untreated individuals generally recruited from population-based cohorts, cardiovascular risk progressively increases in the order of normotension, white coat hypertension, masked hypertension, and sustained hypertension.111-116 There are no clinical effectiveness trials for the treatment of masked hypertension, and subanalyses of trials analyzing the treatment benefit in white coat hypertension have yielded mixed results.117-119 Nonetheless, the robust evidence base supporting hypertension screening and treatment have historically been based solely on OBPM; therefore, participants with white coat hypertension were invariably included in those treatment trials.7,8
Multiple strategies have been suggested to improve accuracy for identifying those with sustained and masked hypertension. AOBP has been suggested as a replacement for traditional office screening and out-of-office confirmation modalities.120 However, there were no included studies of unattended AOBP and only 1 study of attended AOBP reporting test accuracy compared with an ABPM reference standard.35 Other systematic reviews have suggested that, on average, mean AOBP and ABPM values in terms of mm Hg are similar; however, there is substantial heterogeneity and it is unclear if lack of mean mm Hg differences would result in similar diagnostic categorization and treatment decisions.13,121,122 Because higher 10-year CVD risk scores have been associated with an increased prevalence of masked hypertension, CVD risk tools could be useful for identifying specific populations that may benefit from ABPM to identify masked hypertension.123,124 Lowering the OBPM screening threshold is a possible approach to increase test sensitivity for sustained hypertension101 or to additionally identify a population for whom ABPM may be ordered to detect masked hypertension.80,101,125 Despite 2017 guidance from the American College of Cardiology/American Heart Association lowering the OBPM diagnostic threshold to 130/80 mm Hg or greater,101 no studies are available in an untreated population that report accuracy of this threshold compared with 140/90 mm Hg or greater using the same ABPM reference standard threshold. Trials examining the comparative accuracy and feasibility of various blood pressure measurement strategies for diagnostic confirmation of hypertension in primary care are needed; the publication of 1 such trial (BP-CHECK [NCT03130257]) is anticipated in 2021.126
This systematic review has several limitations. First, it excluded accuracy studies in which 20% or more of participants were treated to approximate screening populations. The accuracy of blood pressure measurements may be influenced by blood pressure variability, and variability may be reduced by hypertension medications.127,128 These pooled accuracy estimates therefore may not be applicable to treated populations. Second, for confirmatory test accuracy (KQ3), studies were included that enrolled participants referred for ABPM; while there are indications for ABPM referral outside of diagnostic confirmation, the lack of treatment was considered a proxy for diagnostic confirmation. Third, this review did not include accuracy studies that only reported mm Hg differences between measurement modalities or studies that only included κ values as a measure of agreement because clinical decision-making to initiate pharmacotherapy is based on blood pressures exceeding a defined threshold. Fourth, the reference standard for all accuracy studies was ABPM based on the previous review’s conclusion that there was a robust evidence base that ABPM is predictive of future CVD events129; nonetheless, there is evidence suggesting that HBPM may be an alternative.130 Fifth, foundational evidence supporting screening is derived from treatment trials almost exclusively recruiting patients based on elevated office measurements without out-of-office confirmation.102-107 Sixth, treatment benefits and harms were beyond the scope of this review.
Screening using office-based blood pressure measurement had major accuracy limitations, including misdiagnosis; however, direct harms of measurement were minimal. Research is needed to determine optimal screening and confirmatory algorithms for clinical practice.
Corresponding Author: Janelle M. Guirguis-Blake, MD, Kaiser Permanente Evidence-based Practice Center, Department of Family Medicine, University of Washington, 521 Martin Luther King Jr Way, Tacoma, WA 98405 (firstname.lastname@example.org).
Accepted for Publication: October 15, 2020.
Author Contributions: Dr Guirguis-Blake had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Guirguis-Blake, Evans, Coppola, Weyrich.
Acquisition, analysis, or interpretation of data: Guirguis-Blake, Evans, Webber, Coppola, Perdue, Weyrich.
Drafting of the manuscript: Guirguis-Blake, Evans, Coppola.
Critical revision of the manuscript for important intellectual content: Guirguis-Blake, Evans, Webber, Coppola, Perdue, Weyrich.
Statistical analysis: Perdue.
Administrative, technical, or material support: Evans, Webber, Coppola, Weyrich.
Supervision: Guirguis-Blake, Evans.
Conflict of Interest Disclosures: None reported.
Funding/Support: This research was funded under contract HHSA-290-2015-000017-I-EPC5, Task Order 5, from the Agency for Healthcare Research and Quality (AHRQ), US Department of Health and Human Services, under a contract to support the US Preventive Services Task Force (USPSTF).
Role of the Funder/Sponsor: Investigators worked with USPSTF members and AHRQ staff to develop the scope, analytic framework, and key questions for this review. AHRQ had no role in study selection, quality assessment, or synthesis. AHRQ staff provided project oversight, reviewed the report to ensure that the analysis met methodological standards, and distributed the draft for peer review. Otherwise, AHRQ had no role in the conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript findings.
Disclaimer: The opinions expressed in this document are those of the authors and do not reflect the official position of AHRQ or the US Department of Health and Human Services.
Additional Contributions: The authors gratefully acknowledge the following individuals for their contributions to this project: Tina Fan, MD, MPH, and Brandy Peaker, MD, MPH, at the Agency for Healthcare Research and Quality; current and former members of the US Preventive Services Task Force who contributed to topic deliberations; the National Heart, Lung, and Blood Institute and Centers for Disease Control and Prevention for providing federal partner review of the draft report; Evidence-based Practice Center staff member Jennifer S Lin, MD, MCR, for mentoring and project oversight; and Todd Hannon, MLS, and Jill Pope, BA, for technical and editorial assistance at the Center for Health Research. USPSTF members, peer reviewers, and those commenting on behalf of partner organizations did not receive financial compensation for their contributions.
Additional information: A draft version of this evidence report underwent external peer review from 5 content experts (Beverly Green, MD, MPH, Kaiser Permanente Washington Health Research Institute; Mike LeFevre, MD, MSPH, MU Health Care, Future of Family Medicine; Paul Muntner, PhD, University of Alabama at Birmingham; Daichi Shimbo, MD, Columbia University; and Reem Mustafa, MBBS, PhD, MPH, University of Kansas) and 2 federal partners, the National Heart, Lung, and Blood Institute and Centers for Disease Control and Prevention. Comments were presented to the USPSTF during its deliberation of the evidence and were considered in preparing the final evidence review.
Editorial Disclaimer: This evidence report is presented as a document in support of the accompanying USPSTF Recommendation Statement. It did not undergo additional peer review after submission to JAMA.