[Skip to Navigation]
Sign In
The Rational Clinical Examination
July 27, 2021

Does This Adult Patient Have Hypertension?The Rational Clinical Examination Systematic Review

Author Affiliations
  • 1Department of Family Medicine and Community Health, Duke University School of Medicine, Durham, North Carolina
  • 2Center for Novel and Exploratory Clinical Trials, Yokohama City University, Yokohama, Japan
  • 3Department of Biostatistics, University of North Carolina at Chapel Hill, Chapel Hill
  • 4Durham Veterans Affairs Health System and Department of Medicine, Duke University School of Medicine, Durham, North Carolina
  • 5Maine General Internal Medicine, Waterville
  • 6Department of Medicine, Division of General Internal Medicine and Clinical Epidemiology, University of North Carolina at Chapel Hill
  • 7Research Triangle Park, Durham, North Carolina
  • 8North Carolina Translational and Clinical Sciences Institute, University of North Carolina at Chapel Hill
  • 9Department of Medicine, Columbia University Irving Medical Center, New York, New York
  • 10Department of Epidemiology, University of Alabama at Birmingham
  • 11Department of Family Medicine, University of North Carolina at Chapel Hill
  • 12Sheps Center for Health Services Research, University of North Carolina at Chapel Hill
  • 13Department of Medicine, Division of Cardiology, University of North Carolina at Chapel Hill
  • 14now with Department of Internal Medicine, Division of General Internal Medicine, The Ohio State University
JAMA. 2021;326(4):339-347. doi:10.1001/jama.2021.4533
Key Points

Question  How can clinicians best identify adult patients likely to have hypertension?

Findings  Conventional office blood pressure (BP) measurements from a single visit or home BP monitoring performed over a few days each have limited sensitivity and specificity for diagnosing hypertension, especially when the pretest probability is low. The combination of elevated mean office BP plus elevated mean BP on home BP monitoring is most indicative of hypertension, using 24-hour ambulatory BP monitoring as the reference standard.

Meaning  Out-of-office BP measurements should be obtained to complement office BP before diagnosing hypertension in adults.


Importance  Office blood pressure (BP) measurements are not the most accurate method to diagnose hypertension. Home BP monitoring (HBPM) and 24-hour ambulatory BP monitoring (ABPM) are out-of-office alternatives, and ABPM is considered the reference standard for BP assessment.

Objective  To systematically review the accuracy of oscillometric office and home BP measurement methods for correctly classifying adults as having hypertension, defined using ABPM.

Data Sources  PubMed, Cochrane Library, Embase, ClinicalTrials.gov, and DARE databases and the American Heart Association website (from inception to April 2021) were searched, along with reference lists from retrieved articles.

Data Extraction and Synthesis  Two authors independently abstracted raw data and assessed methodological quality. A third author resolved disputes as needed.

Main Outcomes and Measures  Random effects summary sensitivity, specificity, and likelihood ratios (LRs) were calculated for BP measurement methods for the diagnosis of hypertension. ABPM (24-hour mean BP ≥130/80 mm Hg or mean BP while awake ≥135/85 mm Hg) was considered the reference standard.

Results  A total of 12 cross-sectional studies (n = 6877) that compared conventional oscillometric office BP measurements to mean BP during 24-hour ABPM and 6 studies (n = 2049) that compared mean BP on HBPM to mean BP during 24-hour ABPM were included (range, 117-2209 participants per analysis); 2 of these studies (n = 3040) used consecutive samples. The overall prevalence of hypertension identified by 24-hour ABPM was 49% (95% CI, 39%-60%) in the pooled studies that evaluated office measures and 54% (95% CI, 39%-69%) in studies that evaluated HBPM. All included studies assessed sensitivity and specificity at the office BP threshold of 140/90 mm Hg and the home BP threshold of 135/85 mm Hg. Conventional office oscillometric measurement (1-5 measurements in a single visit with BP ≥140/90 mm Hg) had a sensitivity of 51% (95% CI, 36%-67%), specificity of 88% (95% CI, 80%-96%), positive LR of 4.2 (95% CI, 2.5-6.0), and negative LR of 0.56 (95% CI, 0.42-0.69). Mean BP with HBPM (with BP ≥135/85 mm Hg) had a sensitivity of 75% (95% CI, 65%-86%), specificity of 76% (95% CI, 65%-86%), positive LR of 3.1 (95% CI, 2.2-4.0), and negative LR of 0.33 (95% CI, 0.20-0.47). Two studies (1 with a consecutive sample) that compared unattended automated mean office BP (with BP ≥135/85 mm Hg) with 24-hour ABPM had sensitivity ranging from 48% to 51% and specificity ranging from 80% to 91%. One study that compared attended automated mean office BP (with BP ≥140/90 mm Hg) with 24-hour ABPM had a sensitivity of 87.6% (95% CI, 83%-92%) and specificity of 24.1% (95% CI, 16%-32%).

Conclusions and Relevance  Office measurements of BP may not be accurate enough to rule in or rule out hypertension; HBPM may be helpful to confirm a diagnosis. When there is uncertainty around threshold values or when office and HBPM are not in agreement, 24-hour ABPM should be considered to establish the diagnosis.

Add or change institution