Stacey Sheridan, M.D., M.P.H.;a Michael Pignone, M.D., M.P.H.;b Katrina Donahue, M.D., M.P.H.c
The authors of this article are responsible for its contents, including any clinical or treatment recommendations. No statement in this article should be construed as an official position from the Agency for Healthcare Research and Quality, or the U.S. Department of Health and Human Services.
Address correspondence to: Stacey L. Sheridan, M.D., M.P.H.; Division of General Medicine and Clinical Epidemiology; University of North Carolina at Chapel Hill; 5039 Old Clinic Building, CB 7110; Chapel Hill, NC 27599. E-mail: ssherida@email.unc.edu.
Select for copyright and reprint information.
Structured Abstract
Introduction
Methods
Results
Discussion
Acknowledgments
References
Notes
In 1996, the U.S. Preventive Services Task Force (USPSTF) recommended screening adults for hypertension. Since that time the published literature on hypertension has expanded rapidly, necessitating new examination of the evidence supporting screening.
To update the 1996 evidence review on screening for hypertension.
We used the 1996 Guide to Clinical Preventive Services, recent systematic reviews, and focused searches of MEDLINE® to identify new evidence relevant to detecting and treating hypertension.
When a good-quality, recent systematic review was available, we used it to summarize previous research; we searched MEDLINE® only for more recent articles. Two authors reviewed abstracts (and if necessary full texts) of potentially relevant articles to determine if they should be included.
One author extracted data from included studies into evidence tables.
Hypertension can be effectively detected through office measurement of blood pressure. Treatment of elevated blood pressure in adults can reduce cardiovascular events. The magnitude of risk reduction depends on the degree of hypertension and the presence of other cardiovascular risk factors. Available studies have found no important adverse effects on psychological well-being and mixed effects on the absenteeism rates of adults who are screened and labeled as being hypertensive.
Substantial indirect evidence supports the effectiveness of screening adults to detect hypertension and treating them to reduce cardiovascular disease.
In the United States, hypertension is responsible for 35 percent of all cardiovascular events (myocardial infarction and stroke), 49 percent of all episodes of heart failure, and 24 percent of all premature deaths.1 Patients with hypertension have 2 to 4 times more risk for stroke, myocardial infarction, heart failure, and peripheral vascular disease than patients without hypertension.2 Additionally, they have an increased risk for end-stage renal disease, retinopathy, and aortic aneurysm.1,3,4 This substantial burden of suffering from hypertension, in combination with a feasible and accurate means of detection and a clear benefit from treatment,5 have led to a widespread recommendation for screening for hypertension.
In 1996, the U.S. Preventive Services Task Force (USPSTF) reviewed the evidence regarding screening for hypertension.5 Based on its review, the USPSTF strongly recommended screening adults 21 and older using standard office sphygmomanometry. Although they did not recommend a specific interval for screening, they noted that measurement every 2 years for patients with previously normal blood pressures and every year in persons with borderline levels may be prudent.
In this report, we systematically examine newer evidence relevant to screening for hypertension in adults to assist the U.S. Preventive Services Task Force in updating its recommendations and the Guide to Clinical Preventive Services.5
To examine the role of outpatient clinical screening for hypertension in adults, we first developed an analytic framework depicting key questions of interest to the USPSTF (Figure 1, 16 KB). The arrows in the analytic framework represent steps in the chain of logic connecting screening with defined outcomes: cardiovascular disease (CVD), renal failure, and reduction of mortality.
Four key questions guided our literature searches and synthesis of the evidence:
We used the 1996 Guide to Clinical Preventive Services5 and searches of MEDLINE®, DARE, or the COCHRANE Collaboration Library for systematic reviews published since 1995 and focused searches of MEDLINE® to identify new evidence on the benefits of detecting and treating hypertension. When a good quality, recent systematic review was available, we used it to summarize previous research and searched MEDLINE® only for more recent articles. When such a review was not available, we identified English-language articles from comprehensive searches of the MEDLINE® database between January 1966 and January 2002 and used manual searches of relevant articles and personal libraries, as well as peer review, to ensure that we included all appropriate articles.
We included recent systematic reviews and individual observational studies or randomized controlled trials (RCTs) that examined the following topics:
Two authors reviewed abstracts of potentially relevant articles to determine if they should be included. When the reviewers disagreed, we obtained the full articles and resolved the disagreements by consensus.
For each topic for which we performed a comprehensive review (e.g., the prognostic ability of home blood pressure monitoring, the prognostic ability of ambulatory blood pressure monitoring, and the adverse effects of screening for hypertension), a single reviewer extracted data from the included studies and created evidence tables. Using the guidelines developed for the USPSTF reviews, we evaluated the internal and external validity of each study.6 Experts and USPSTF members reviewed our work for accuracy and completeness at meetings in 2001 and early 2002.
RCTs of screening (versus no screening) would provide the best evidence about the effects of screening for hypertension on CVD and mortality. We identified no such studies. Many trials of hypertension treatment that compared pharmacologic and behavioral intervention to usual care, however, showed a beneficial effect of treatment in patients who were enrolled on the basis of elevated blood pressures detected on screening examinations. These findings suggest that screening may be beneficial, but examination of additional evidence is necessary to determine the value of screening. Screening would be beneficial if:
Hypertension, a clinically significant elevation in blood pressure, is usually defined in adults as a diastolic blood pressure (DBP) of 90 mm Hg or higher or a systolic blood pressure (SPB) of 140 mm Hg or higher.7 The Joint National Committee on Hypertension defined stage 1 hypertension as blood pressures ranging from 140 to 159 systolic or 90 to 99 diastolic, stage 2 as pressures ranging from 160 to 179 systolic or 100 to 109 diastolic, and stage 3 as those equal to or greater than 180 systolic or 110 diastolic.7
The relative risk for adverse events associated with hypertension is continuous and graded;2,8 systolic blood pressure predicts cardiovascular risk better than diastolic pressure.9 Even modest elevations of blood pressure in young adulthood are associated with increased risk for cardiovascular events in middle age.10 The absolute risk for cardiovascular events, however, varies depending on the presence of other cardiovascular risk factors, including smoking, diabetes, and abnormal blood lipid levels, as well as the duration of blood pressure elevation.11 To determine how we can most effectively identify patients with hypertension for whom treatment will be beneficial, we consider both which screening method for detection of an elevated blood pressure is most accurate and reliable and whether screening a subset of individuals with other cardiovascular risk factors is more effective than screening everyone in targeting treatment to those who will benefit most.
The accuracy and reliability of blood pressure measurement in adults depends on the method of screening. Characteristics of several common methods of screening are described below. We are unable to provide information on test accuracy owing to limitations of existing evidence, particularly the absence of a "gold standard" against which to judge screening.
Office Blood Pressure Measurement. Office blood pressure measurement using an appropriate upper arm cuff with mercury or aneroid sphygmomanometry is the standard screening test for hypertension. If performed correctly, it provides a measure of blood pressure that is highly correlated with intra-arterial measurement (correlation coefficients 0.94-0.98) and is highly predictive of cardiovascular risk.12
Although office blood pressure measurement is highly predictive when performed correctly, measurement errors occur in clinical practice.13 Errors in measuring blood pressure may result from instrument, observer, or patient factors. As noted by the USPSTF in 1996:5, p.40 "Examples of instrument error include manometer dysfunction, pressure leaks, stethoscope defects, and cuffs of incorrect width or length for the patient's arm size. The observer can introduce errors due to sensory impairment (difficulty hearing Korotkoff sounds or reading the manometer), inattention, inconsistency in recording Korotkoff sounds (e.g., Phase IV vs. Phase V), and subconscious bias (e.g., digit preference for numbers ending with zero or preconceived notions of "normal" pressures)."
Overly rapid release of air from the blood pressure cuff may also be a problem, as may knowledge of the previous readings. Moreover, "[t]he patient can be the source of misleading readings due to posture and biologic factors. Posture (i.e., lying, standing, sitting) and arm position in relation to the heart can affect results by as much as 10 mm Hg. Biologic factors include anxiety, meals, tobacco, alcohol, temperature changes, exertion, and pain."5 In a recent systematic review of factors affecting office blood pressure measurement, McAlister and Straus documented many of the same sources of errors.14
Because of these limitations in the precision of blood pressure measurement, experts commonly recommend that clinicians diagnose hypertension only after they obtain 2 or more elevated blood pressure readings at each of 2 or more separate visits over a period of 1 to several weeks.7 This recommendation follows the pattern of blood pressure measurement in the randomized trials that established the benefits of antihypertensive therapy and represents a compromise between reliable detection of elevated pressures and clinical practicality. One investigation using statistical modeling suggested that patients with blood pressures above 95 mm Hg or below 90 mm Hg could be accurately designated (80% confidence) as having or not having elevated blood pressure after 3 visits, whereas those with borderline elevated blood pressures required 5 visits.15 As expected, increasing precision required an increased number of measurements, particularly for those with borderline hypertension.
Home Blood Pressure Monitoring. Because of the greater number of measurements that can be taken at home, home blood pressure monitoring may provide a better assessment of average blood pressure than periodic office measurement. However, home blood pressure monitoring is subject to the same sources of measurement error as office blood pressure measurement. Additionally, we found only preliminary data to suggest that home monitoring may correlate better with the risk for clinical cardiovascular endpoints than office measurement;16 no definitive study is available to clarify such claims.14
Ambulatory Blood Pressure Measurement. Ambulatory blood pressure monitoring provides a measure of average blood pressure over 24 hours rather than isolated single values obtained in office checks or at home. Fewer data are available about ambulatory blood pressure monitoring than about office-based blood pressure measurement.14,17,18 However, multiple fair quality prospective cohort studies have found that ambulatory blood pressure measurement may be a slightly better predictor of clinical cardiovascular outcomes than clinic-based approaches.19-25 Lack of good quality studies and higher monetary costs have limited the use of ambulatory blood pressure monitoring as a screening test. Ambulatory blood pressure monitoring has, however, been useful to stratify individuals who have normal blood pressures at home but elevated blood pressures in the office, so-called white coat hypertension. A recent study by Little and colleagues examined the potential usefulness of ambulatory blood pressure monitoring in this situation.26 The study found that many patients with elevated clinic blood pressures had normal ambulatory blood pressure (<135/85).
As previously noted, the risk for cardiovascular events and the potential benefit from screening and subsequent treatment of hypertension depend on both the degree and duration of blood pressure elevation and the presence of other cardiovascular risk factors, such as age, sex, lipid disorders, smoking, and diabetes.11,27,28 Because the degree and duration of blood pressure elevation are unknown before screening, selective screening to identify individuals who would benefit most from detection and treatment of hypertension would need to target individuals with other cardiovascular risk factors. We found no studies that examined the relative effectiveness, cost-effectiveness, or harms of targeting screening for hypertension only to those patients with other cardiovascular risk factors instead of to all patients who present at a physician's office. We additionally found no studies that examined the optimal frequency of screening based on a patient's prior blood pressure levels or other cardiovascular risk factors.
For patients who are screened, estimates of the potential benefit of treatment can be improved both by carefully measuring the degree of blood pressure elevation and by assessing the contribution of other risk factors to global cardiovascular risk.11,27,28 The feasibility of routine global risk assessment, however, has not been well studied.
Treatment for hypertension has traditionally been defined as pharmacological or nonpharmacological therapy to reduce blood pressure. Recent trials have shown, however, that the ability of a pharmacological treatment to reduce undesirable cardiovascular outcomes may not correspond directly to its ability to lower blood pressure.29,30 We therefore separate our reviews of pharmacological and nonpharmacological therapies and focus on the efficacy of these therapies in reducing CVD events (e.g., heart attacks, strokes, heart failure, and sudden death). When CVD outcomes were unavailable, we considered the effect of the intervention on blood pressure levels. Here we focus on the results of several recent high quality systematic reviews that summarize the evidence on therapy for hypertension.
Stage 3 Hypertension. A 1996 meta-analysis of 3 small trials conducted in the 1960s in patients with diastolic blood pressure (DBP) greater than 110 mm Hg found that treatment with reserpine or alpha-methyl dopa reduced the odds of congestive heart failure by 86 percent (odds ratio [OR] 0.14; 95% confidence interval [CI] 0.05 to 0.41).31 Because other events were too infrequent over the 1- to 2-year durations of these trials, the authors could not determine the effects of treatment on stroke, major coronary events, CVD mortality, or total mortality. The numbers needed to treat (NNT) over 5 years ranged from 200 to 400 to prevent 1 congestive heart failure event.
Stage 1 and Stage 2 Hypertension. Although no studies have examined the effects of treatment only for persons with DBPs of 90 mm Hg to 99 mm Hg, persons with DBPs of 90 to 109 mm Hg benefit from treatment. The Gueyffier et al. systematic review examined 5 trials (1.4 to 7 years in duration) in individuals with DBPs of 90 mm Hg to 109 mm Hg and found that treatment of hypertension in adults younger than 60 reduced stroke (OR 0.51; 95% CI 0.39 to 0.69) but had no effect on coronary heart disease events, CVD deaths, or total mortality.31 In patients older than 60, the majority of whom had blood pressures in this range, treatment reduced total mortality (OR 0.90; 95% CI 0.81 to 1.00), CVD death (OR 0.77; 95% CI 0.67 to 0.89), stroke (OR 0.66; 95% CI 0.56 to 0.77), coronary heart disease events (OR 0.79; 95% CI 0.68 to 0.92), and congestive heart failure (OR 0.54; 95% CI 0.43 to 0.68).31 The numbers needed to treat over 5 years ranged from approximately 900 to prevent 1 stroke to 10,000 to prevent 1 case of coronary heart disease event or death.
Isolated Systolic Hypertension in the Elderly. Staessen et al. recently performed a systematic review of 8 trials, including two published since 1996, comparing treatment of isolated systolic hypertension (systolic blood pressure [SBP] greater than 160 mm Hg but DBP less than 95 mm Hg) in patients older than 60 with pharmacological therapy (e.g., diuretics, beta-blockers, or calcium-channel blockers) or with placebo treatment.28 The authors found that active treatment reduced stroke (RR 0.70; 95% CI 0.59 to 0.82), coronary heart disease events (RR 0.70; 95% CI 0.66 to 0.90), CVD mortality (RR 0.82; 95% CI 0.71 to 0.96), and total mortality (RR 0.87; 95% CI 0.78 to 0.98). The number needed to treat over 5 years to prevent 1 cardiovascular event was 18 (95% CI 17 to 19) in men and 38 (95% CI 36 to 40) in women.
RCTs of nonpharmacological therapies have examined CVD events as outcomes. Fair to good evidence supports the effectiveness of several nonpharmacological interventions for reducing blood pressure in patients with hypertension over periods up to 1 to 2 years (e.g., weight reduction in overweight patients,32-34 increased physical activity,35 sodium reduction,36-43 potassium supplementation,44-47 decreased alcohol intake,48 and stress management.49,50,51 The magnitude of blood pressure reduction differs by intervention, ranging from 2 to 15 mm Hg for studied interventions: 4-15 mm Hg reduction in SBP for 2-10 kg weight reduction;32 5-7 mm Hg reduction in SBP with moderate to vigorous exercise;35 5.8 mm Hg reduction in SBP with 100 mmol/L (1 teaspoon) reduction in salt intake;36 3.1 mm Hg reduction in SBP with 60 mmol (= 60 meq) of potassium supplementation;44 3.3 mm Hg reduction in SBP with 50 percent reduction in alcohol use in persons drinking 20-40 drinks/week;48 and 9-10 mm Hg reduction in SBP for persons receiving single or multi-component stress management interventions.49 Importantly, some of these interventions (e.g., weight reduction, physical activity) have other beneficial effects that may not be mediated through changes in blood pressure or changes in the incidence of cardiovascular events.
Nonpharmacological therapy also appears to be effective for the primary prevention of hypertension. This topic is outside the scope of our review; we refer the reader to a recent review by the National High Blood Pressure Education Program52 for more information.
We identified 10 cohort studies examining the adverse effects of screening for hypertension and subsequently labeling a person as "hypertensive." These studies were reported in 14 papers and used multiple study designs.53-66 Some studies compared the consequences of screening in hypertensive persons who previously were either aware or unaware of their hypertension; these studies addressed the differential effects of screening and diagnosis versus monitoring of hypertension. Others compared the unintended consequences of labeling an individual as "hypertensive" vs "normotensive." We identified no studies that compared the effect of labeling or not labeling all individuals who were identified to have high blood pressure.
We identified 6 cohort studies examining the psychological effects of screening and labeling individuals with hypertension.53-59 Five compared the unintended psychological effects of labeling an individual as "hypertensive" or "normotensive";53-58 1 compared the psychological effects of screening and labeling hypertensive individuals who previously were either aware or unaware of their hypertensive diagnosis.59 These studies did not find any evidence for adverse psychological effects. However, 3 of these studies had high attrition rates;53,54,56,57 1 reported no statistical comparisons;59,67 and 1 reported only within-group changes in absenteeism among those who were newly diagnosed with hypertension.55
One retrospective and 4 prospective cohort studies showed an increase in absenteeism of 2.2 days to 5.7 days per year in newly screened and diagnosed individuals.59-65 One study, however, reported no statistical comparisons of absenteeism among groups;59,67 2 reported only within group statistical comparisons of absenteeism among those who were previously unaware of their hypertensive diagnosis, making no comparisons with the control group of aware hypertensives;60,61,65 and 1 showed only a trend toward a statistical increase in absenteeism among patients who were previously unaware (compared with aware) of their diagnosis of hypertension.62 Two other cohort studies did not find an increase in absenteeism,56-58 although the study by Rudd and colleagues was again noted to have high attrition rates, which limited conclusions. Differences in the way absenteeism has been defined may explain these discrepancies. In many cases, the reasons for work absenteeism were not directly measured, and absenteeism may be explained by visits to a medical provider for the treatment of hypertension. Additionally, many studies did not control for the effects of comorbid illness or the complications of hypertension, nor did they report how patients were informed of their blood pressure status.
In summary, we found fair quality evidence suggesting that screening and labeling adults with hypertension produces no adverse effects on psychological well-being and mixed effects on absenteeism rates for jobs.
Serious or life-threatening adverse drug reactions have been rare in clinical trials of drugs for hypertension, particularly in those that have used low-dose regimens. Less serious adverse effects have been noted to occur with most drugs.68 A full review of these effects is beyond the scope of this report, but it is clear that clinicians should take adverse effects into account when deciding whether to treat and which treatment to use because the majority of patients with hypertension are asymptomatic.
Strong indirect evidence supports screening adults for hypertension. Hypertension is an important contributor to CVD morbidity and mortality. It is predictive of CHD events and is reliably detected through screening blood pressure measurements using a standard arm blood pressure cuff and sphygmomanometer. Additionally, treatment of adult hypertensive patients with drug therapy and possibly nonpharmacological interventions can reduce blood pressure and the incidence of cardiovascular events, including myocardial infarction, heart failure, and stroke. The degree of risk reduction depends on patients' levels and possibly duration of blood pressure elevation, their other risk factors for CVD, and the choice of antihypertensive treatment.
This review extends our knowledge about several specific questions germane to hypertension screening. Recent prospective studies suggest that ambulatory blood pressure monitoring may be a better predictor of cardiovascular risk than clinic blood pressure measurement, although its utility for screening remains uncertain. Recent treatment trials have confirmed large potential benefit in detecting and treating isolated systolic hypertension in the elderly, highlighting the importance of screening and treating older adults. A detailed analysis of studies addressing the potential harms of hypertension screening found no evidence of adverse effects on psychological well-being and evidence of mixed effects on absenteeism rates in those who are screened and labeled.
Although substantial indirect evidence supports screening, many important issues about screening remain unanswered. Among the high priority questions are the following:
Additionally, despite relatively clear evidence supporting screening and the widespread use of clinical blood pressure measurement, identification and treatment of hypertension remains suboptimal for the U.S. population as a whole. A recent population-based study using National Health and Nutrition Exam Survey (NHANES III) data reported that 31 percent of hypertensive Americans are unaware that they have hypertension, 17 percent are aware of their diagnosis but are not being treated, and 29 percent are being treated but have not controlled their blood pressure.69 Healthy People 2010 aims to reduce all of these numbers to 5 percent.70 Substantial progress in organization of care and access to care will be required to approach the Healthy People 2010 goals.