eFigure. Study Flowchart
eTable 1. Comparison of Baseline Characteristics Between Patients Referred to the Hospital and Propensity-Matched Sample From Patients Sent Home
eTable 2. Comparison of Baseline Characteristics Between Patients With and Without Follow-up Between Patients Referred to the Hospital and the Propensity-Matched Sample From Patients Sent Home
eTable 3. Sensitivity Analyses for the Outcomes of Patients With Follow-up Defined as Patients With at Least 1 Repeated Blood Pressure Measurement Within 6 Months of Baseline Visit
eTable 4. Baseline Characteristics of Patients Presenting With Hypertensive Urgency to the Office (Systolic BP ≥200 mm Hg)
eTable 5. Unadjusted Outcomes of Patients With Hypertensive Urgency (Systolic BP ≥200 mm Hg)
eTable 6. Baseline Characteristics of Patients Presenting With Hypertensive Urgency to the Office (Systolic BP ≥220 mm Hg)
eTable 7. Unadjusted Outcomes of Patients Presenting With Hypertensive Urgency to the Office (Systolic BP ≥220 mm Hg)
Patel KK, Young L, Howell EH, Hu B, Rutecki G, Thomas G, Rothberg MB. Characteristics and Outcomes of Patients Presenting With Hypertensive Urgency in the Office Setting. JAMA Intern Med. 2016;176(7):981–988. doi:10.1001/jamainternmed.2016.1509
Copyright 2016 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
The prevalence and short-term outcomes of hypertensive urgency (systolic blood pressure ≥180 mm Hg and/or diastolic blood pressure ≥110 mm Hg) are unknown. Guidelines recommend achieving blood pressure control within 24 to 48 hours. However, some patients are referred to the emergency department (ED) or directly admitted to the hospital, and whether hospital management is associated with better outcomes is unknown.
To describe the prevalence of hypertensive urgency and the characteristics and short-term outcomes of these patients, and to determine whether referral to the hospital is associated with better outcomes than outpatient management.
Design, Setting, and Participants
This retrospective cohort study with propensity matching included all patients presenting with hypertensive urgency to an office in the Cleveland Clinic Healthcare system from January 1, 2008, to December 31, 2013. Pregnant women and patients referred to the hospital for symptoms or treatment of other conditions were excluded. Final follow-up was completed on June 30, 2014, and data were assessed from October 31, 2014, to May 31, 2015.
Hospital vs ambulatory blood pressure management.
Main Outcomes and Measures
Major adverse cardiovascular events (MACE) consisting of acute coronary syndrome and stroke or transient ischemic attack, uncontrolled hypertension (≥140/90 mm Hg), and hospital admissions.
Of 2 199 019 unique patient office visits, 59 836 (4.6%) met the definition of hypertensive urgency. After excluding 851 patients, 58 535 were included. Mean (SD) age was 63.1 (15.4) years; 57.7% were women; and 76.0% were white. Mean (SD) body mass index (calculated as weight in kilograms divided by height in meters squared) was 31.1 (7.6); mean (SD) systolic blood pressure, 182.5 (16.6) mm Hg; and mean (SD) diastolic blood pressure, 96.4 (15.8) mm Hg. In the propensity-matched analysis, the 852 patients sent home were compared with the 426 patients referred to the hospital, with no significant difference in MACE at 7 days (0 vs 2 [0.5%]; P = .11), 8 to 30 days (0 vs 2 [0.5%]; P = .11), or 6 months (8 [0.9%] vs 4 [0.9%]; P > .99). Patients sent home were more likely to have uncontrolled hypertension at 1 month (735 of 852 [86.3%] vs 349 of 426 [81.9%]; P = .04) but not at 6 months (393 of 608 [64.6%] vs 213 of 320 [66.6%]; P = .56). Patients sent home had lower hospital admission rates at 7 days (40 [4.7%] vs 35 [8.2%]; P = .01) and at 8 to 30 days (59 [6.9%] vs 48 [11.3%]; P = .009).
Conclusions and Relevance
Hypertensive urgency is common, but the rate of MACE in asymptomatic patients is very low. Visits to the ED were associated with more hospitalizations, but not improved outcomes. Most patients still had uncontrolled hypertension 6 months later.
Quiz Ref IDApproximately one-third of the US population has hypertension, which is associated with an increased risk for end-organ damage, mainly myocardial infarction, heart failure, stroke, and kidney damage.1- 4 For chronic uncontrolled hypertension, every 20–mm Hg increase in systolic blood pressure (SBP) or 10–mm Hg increase in diastolic blood pressure (DPB) is associated with doubling of vascular mortality.5 Because organ damage results from increased blood pressure over time, physicians may be concerned about the potential for end-organ damage after severely elevated blood pressure, even for short periods. Hypertensive urgency is defined as SBP of at least 180 mm Hg and/or DBP of at least 110 mm Hg, without associated end-organ damage.6- 8 The prevalence of hypertensive urgency in the outpatient setting is unknown, and management is complicated by a lack of observational studies or randomized clinical trials. The seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure recommended gradual blood pressure control for 24 to 48 hours,4 but this was not addressed in the eighth report.9
Many patients with hypertensive urgency undergo evaluation and treatment in the emergency department (ED), where asymptomatic severe hypertension accounts for as many as 27% of medical emergencies10 and 3% of all visits.10,11 A recent American College of Emergency Physicians policy statement on this topic acknowledges the dearth of evidence supporting recommendations for optimal follow-up, checking for target organ damage, and the effectiveness of ED treatment.12 In practice, when patients present to outpatient clinics with severely elevated blood pressure, some are directly admitted to the hospital, some are sent to the ED, and others are sent home with adjustments to antihypertensive therapy. To our knowledge, no retrospective or prospective studies exist to inform management.
To address these concerns, we conducted a retrospective cohort study of all patients presenting with hypertensive urgency in an office setting within a single health care system during a 6-year period. We hypothesized that ambulatory patients with hypertensive urgency would have low rates of cardiovascular events in the short term and that referral to the hospital would not improve outcomes.
Question For ambulatory patients presenting with hypertensive urgency, is referral to the hospital safer than routine outpatient management of blood pressure?
Findings In this retrospective study of 58 535 patients, emergency department visits for hypertensive urgency were associated with an 8.2% rate of hospital admissions at 7 days vs 4.7% for patients sent home, a significant difference. Major cardiac events were rare, and two-thirds of patients still had uncontrolled hypertension at 6 months.
Meaning For patients presenting with asymptomatic hypertensive urgency, efforts should focus on improving follow-up and serial blood pressure control instead of referral to the hospital.
We conducted a retrospective cohort study of all patients presenting with hypertensive urgency to an outpatient office within the Cleveland Clinic Healthcare System from January 1, 2008, to December 31, 2013. The Cleveland Clinic Healthcare System includes outpatient practices in all specialties, consisting of 1 large academic health care center, 10 regional hospitals, 17 family health centers, and more than 75 outpatient locations in northeast Ohio and Florida. This study was approved by the institutional review board of the Cleveland Clinic Healthcare System. Owing to the retrospective nature of the study, informed consent was waived by the institutional review board.
Quiz Ref IDWe defined hypertensive urgency as SBP of at least 180 mm Hg and/or DBP of at least 110 mm Hg.4 We excluded pregnant women and patients referred to the hospital for symptoms or for treatment of conditions other than hypertension. Patients were divided between those sent home from the office and those referred to the hospital (ED or direct inpatient admission) on the same day. To estimate the prevalence, we divided the number of visits for hypertensive urgency by the total number of unique patient visits to outpatient offices of the Cleveland Clinic Healthcare System during the study period. For patients with multiple visits with hypertensive urgency, only the first visit was included.
For each patient, we recorded baseline demographic data, including age, sex, race, blood pressure, and body mass index (calculated as weight in kilograms divided by height in meters squared). We recorded history of atherosclerotic risk factors, including hypertension, diabetes, smoking, coronary artery disease, hyperlipidemia, chronic kidney disease, dialysis, and cerebrovascular events, based on encounter diagnosis codes before the index visit. We also recorded antihypertensives used at the index visit. Data were generated electronically and validated with 100% accuracy through manual medical record review.
All patients were followed up for 6 months and all blood pressure readings were recorded. Patients were considered to have uncontrolled hypertension at 1 month or 6 months if the last blood pressure reading in that period was at least 140/90 mm Hg.1,13 We reviewed all subsequent visits in our health care system and recorded major adverse cardiovascular events (MACE) within 7 days, 1 month, and 6 months of the index visit. MACE included acute coronary syndrome and cerebrovascular events. All outcomes were identified electronically via codes from the International Classification of Diseases, Ninth Revision, and then manually adjudicated through medical record review by 2 independent reviewers (K.K.P. and L.Y.). Acute coronary syndromes (ST and non-ST elevation myocardial infarction and unstable angina) had to have been diagnosed in the medical record by the consulting cardiologist. Similarly, cerebrovascular events (ischemic and hemorrhagic strokes or transient ischemic attacks with attributable territory of origin) had to have been diagnosed by a consulting neurologist. On review, fewer than 10% of the codes represented a new event. We also recorded all-cause hospital admission rates, including ED, observation, and inpatient admissions within 7 days of the baseline visit and from 8 to 30 days (1 month) after to assess for changes in the use of health care resources between the 2 treatment groups. Patients who did not have a subsequent blood pressure reading were considered lost to or unavailable for follow-up and were not included in the analysis of blood pressure control. For patients who were referred to the ED solely for treatment of blood pressure, details of management, including tests and results, were recorded by manual medical record review. Target organ damage was identified by test results.
Patient characteristics were summarized with descriptive statistics. Continuous variables were compared using the unpaired t test and categorical variables were compared using the χ2 test. We created a propensity model with referral to the ED or hospital as the outcome. The model included baseline demographics (age, sex, and race), SBP and DBP values, vascular disease risk factors (hypertension, diabetes, hyperlipidemia, stroke or transient ischemic attack, chronic kidney disease, dialysis, and coronary artery disease), and number of antihypertensives used at baseline. Asymptomatic patients referred to the ED or hospital were then matched with patients sent home in a 1:2 ratio using the propensity-matching algorithm in R software.14 We compared outcomes between patients referred to the ED or hospital and propensity-matched control individuals using Pearson χ2 test or Fisher exact test if the events were rare.15- 17 Baseline characteristics of patients lost to or unavailable for follow-up were compared with those of patients who had follow- up. We then performed a sensitivity analysis in which we limited the data set to patients who had a primary care physician in our health care system and limited the analysis to those patients who had a blood pressure reading within 6 months. We also performed stratified analyses limited to patients with SBPs of at least 200 and at least 220 mm Hg. All analyses were conducted with R studio software (https://www.r-project.org/). Statistical significance was established at a 2-sided P < .05.
Quiz Ref IDDuring the study period, 1 299 019 unique patient office visits occurred during which a blood pressure reading was recorded, and 59 836 of these (4.6%) met the definition for hypertensive urgency. After excluding 851 patients (eFigure in the Supplement), our final sample included 58 535 patients. Characteristics of the population appear in Table 1. Patients had a mean (SD) age of 63.1 (15.4) years; 57.7% were women; and 76.0% were white. The mean (SD) body mass index was 31.1 (7.6). The mean (SD) SBP was 182.5 (16.6) mm Hg, with 10.2% of the patients having an SBP of at least 200 mm Hg. The mean (SD) DBP was 96.4 (15.8) mm Hg with 5.7% of the patients having a DBP of at least 120 mm Hg. We found a documented history of hypertension in 72.9% of patients, and 58.2% of patients were taking 2 or more antihypertensives.
Quiz Ref IDOnly 426 patients (0.7%) were referred to the hospital for blood pressure management. The rest (n = 58 109) were sent home. The mean blood pressure of patients referred to the hospital for management was 16/11 mm Hg higher than that of patients sent home (P < .001). Of the 426 patients who were referred to the hospital, 218 (51.2%) had an SBP of at least 200 mm Hg compared with 5745 (9.9%) of 58 109 patients sent home (P < .001). The distribution of most atherosclerotic risk factors was similar between groups, but patients sent to the hospital were more likely to have a history of hypertension (408 of 424 [96.2%] vs 42 264 of 57 916 [73.0%]; P < .001) and chronic kidney disease (71 of 426 [16.7%] vs 6050 of 58 109 [10.4%]; P < .01). The number and type of antihypertensives taken for both groups are shown in Table 1.
Table 2 shows the unadjusted outcomes for all patients. Overall, the rate of MACE within 7 days, 8 to 30 days (1 month), and 6 months was low (<1%) in both groups. A total of 496 patients experienced MACE within 6 months (205 episodes of acute coronary syndrome and 301 episodes of stroke or transient ischemic attack). Regardless of where patients were treated, the percentage of patients with controlled hypertension at 1 and 6 months was poor. In unadjusted analyses compared with patients sent home, patients referred to the hospital had more MACE at 7 days (2 of 426 [0.5%] vs 61 of 58 109 [0.1%]; P = .02) but no significant difference at 1 and 6 months. They also had a higher 7-day (35 of 426 [8.2%] vs 2311 of 58 109 [4.0%]; P < .001) and 8- to 30-day (48 of 426 [11.3%] vs 3897 of 58 109 [6.7%]; P < .001) hospital admission rates.
Our propensity model had a C statistic of 0.83. For our propensity-matched analysis, we were able to match all 426 patients referred to the hospital in a 1:2 ratio with patients sent home (Table 3). The baseline characteristics of the matched groups appear in eTable 1 in the Supplement. In the propensity-matched analysis, patients sent home (n = 852) compared with patients referred to the hospital (n = 426) had no significant difference in MACE at 7 days (0 vs 2 [0.5%]; P = .11]), 8 to 30 days (0 vs 2 [0.5%]; P = .11), or 6 months (8 [0.9%] vs 4 [0.9%]; P > .99). Patients sent home were more likely to have uncontrolled hypertension at 1 month (735 [86.3%] vs 349 [81.9%]; P = .04) but not at 6 months (393 of 608 [64.6%] vs 213 of 320 [66.6%]; P = .56). They also had lower hospital admission rates at 7 days (40 [4.7%] vs 35 [8.2%]; P = .01) and 8 to 30 days (59 [6.9%] vs 48 [11.3%]; P = .009).
Loss to or unavailability for follow-up occurred in 58 of the 426 patients referred to the hospital (13.6%) and 12 553 of the 58 109 patients sent home (21.6%). Baseline characteristics of patients with and without follow-up did not differ significantly (eTable 2 in the Supplement). Limiting the analysis to patients with follow-up at 6 months did not change the results (eTable 3 in the Supplement). Patients who had a primary care physician in the Cleveland Clinic Healthcare System (44.2% of our sample) were less likely to be lost to or unavailable for follow-up, including 9 of 172 (5.2%) of those referred to the hospital and 3036 of 22 648 (13.4%) of those sent home. Unadjusted and propensity-matched analyses of these patients (eTable 3 in the Supplement) yielded essentially the same results.
When we limited the analysis to patients with higher blood pressure values, 218 of 5963 patients with SBP of at least 200 mm Hg (3.7%) and 81 of 1058 patients with SBP of at least 220 mm Hg (7.7%) at baseline were referred to the hospital (eTables 4 to 7 in the Supplement). The referred patients were younger, were more likely to be African American, and had higher blood pressure values and more comorbidities. As in the overall sample, patients referred to the hospital had less uncontrolled hypertension at 1 month but not at 6 months. MACE outcomes and hospital admission rates did not differ significantly.
Of the 426 patients referred to the hospital, 387 were referred to the ED; the rest were directly admitted to an inpatient unit. On arrival at the ED, mean (SD) SBP was 197.1 (27.8) mm Hg and mean (SD) DBP was 103.1 (18.7) mm Hg. On discharge, mean (SD) SBP had fallen to 166.3 (23.3) mm Hg and DBP to 87.1 (15.4) mm Hg. Eight patients (1.9%) had evidence of target organ damage (4 had pulmonary edema, 2 had acute kidney injury, and 2 had elevated levels of cardiac biomarkers). The cause of hypertensive urgency was unknown in 233 patients (60.2%), attributed to nonadherence to therapy (including running out of medications) in 94 (24.3%), new diagnoses of hypertension in 42 (10.9%), and referral because they were new patients who presented with elevated blood pressure readings in 18 (4.7%). Treatment consisted of a 1-time dose of antihypertensive medication for 224 of 379 patients (59.1%) (with intravenous delivery in 91 and oral medication in 151), whereas the rest received no intervention. Labetalol hydrochloride (42 [10.9%]) was the most commonly used intravenous drug and clonidine (58 [15.0%]) was the most commonly used oral drug. In the ED, 41 of 748 tests ordered (5.5%) and 39 of 387 patients (10.1%) had an abnormal result (Table 4). Of the ED patients, 61 (15.8%) were admitted to an inpatient service and 310 (80.1%) were discharged home. Of the 61 patients admitted from the ED, 9 (14.8%) were admitted for observation, and the mean (SD) length of stay was 3.5 (3.8) days. Among the 310 patients discharged from the ED, a new antihypertensive was added to the home regimen for 82 patients (26.5%), the home dose of antihypertensive therapy was increased for 23 (7.4%) (some patients had both), and no change was made for 257 (82.9%).
In this large retrospective cohort study, almost 1 in 20 patients presenting to the outpatient departments had hypertensive urgency, but fewer than 1 in 100 asymptomatic patients were referred to the hospital or ED for management of blood pressure. The rate of MACE in all patients was low. Quiz Ref IDReferral to the hospital was associated with increased hospitalizations, but not with better outcomes. Disappointingly, most patients with hypertensive urgency still had uncontrolled hypertension 6 months later.
Despite the common nature of the problem, a surprising lack of literature characterizes hypertensive urgency and its outcomes, particularly in the ambulatory setting. Prior studies10,18,19 report the prevalence of hypertensive crisis in the ED to range from 0.5% to 3% of visits, including those for asymptomatic and symptomatic urgency and patients with emergent visits. In comparison, we found that hypertensive urgency was present in nearly 5% of patients in the office setting, although few were referred to the hospital for management.
Little evidence guides the management of hypertensive urgency. Other than treatment of symptoms, no recommendations exist about who should receive acute management in an emergency setting. What benefit patients get from management in the ED is also unclear. We found that patients who were referred for emergency treatment generally had higher blood pressure but were otherwise similar to those treated in the outpatient setting. Testing results were largely unrevealing. In our study, only 5.5% of tests had an abnormal result and only 8 of 387 patients undergoing testing (2.1%) had evidence of target organ injury. Karras et al20 and Nishijima et al21 reported similar rates of meaningful test abnormalities (6% to 7.2%) in patients presenting to the ED with asymptomatic severe hypertension. Moreover, most of these tests could be performed without an ED admission. Interestingly, 142 of our 379 ED patients (37.5%) received no intervention for their blood pressure, 310 of 387 (80.1%) were discharged home, and 257 of 310 (82.9%) did not have a change in their antihypertensive regimen on discharge, which is comparable to previous reports.20,22
Hypertensive urgency is a sign of poorly controlled hypertension and is associated in the longer term with end-organ damage, primarily of the heart, brain, and kidneys.3,5 In the office setting, however, very high blood pressure readings may also be perceived as posing an immediate risk of acute coronary syndrome or stroke, perhaps prompting hospital referral. To date, however, no study has indicated that hypertensive urgency portends acute risk. The few published reports of outcomes in severe hypertension focus on a select subset of symptomatic patients with acute end-organ damage.23,24 These findings cannot be extrapolated to asymptomatic patients presenting to the office or ED. In our cohort, cardiovascular events occurred in less than 1% of patients within 6 months. In our propensity-matched sample, ED referral was not associated with better 7-day, 8- to 30-day, or 6-month cardiovascular outcomes. A larger concern is long-term uncontrolled hypertension, which poses a more serious risk to patients.3,5 In our cohort, 59.7% of patients had uncontrolled hypertension at 6 months, and no difference was found between the group sent to the hospital and those sent home. Although less dramatic, uncontrolled hypertension in these patients is a more likely cause of long-term cardiovascular morbidity and mortality. As such, appropriate follow-up and intensification of antihypertensive therapy should be emphasized.
Our study has several important limitations. The study was conducted in a single health care system in Ohio and Florida and may not be generalizable to other regions. However, our system is large and includes a diverse population of more than 1 million patients seen in inner city, suburban, and rural settings. Given our data source—medical record review—we may not have ascertained all outcomes. Additional events could have occurred in other health care systems and sudden cardiac death at home may not have been captured. However, unless such events were more likely to occur in patients sent home, our findings would be unchanged. Death outside the hospital is not reliably captured in our electronic health record review, so we did not include mortality in our analysis. Given that the major cardiac events were so rare and the referral to the hospital was not associated with a significant difference, mortality was unlikely to be significantly affected. Given the observational nature, our study may be subject to confounding. We were able to successfully match all patients admitted to the hospital with 2 controls, thereby balancing blood pressure, demographics, and important cardiovascular risk factors, but unmeasured confounders may nevertheless have contributed to the observed associations. Last, many of our patients were lost to or unavailable for follow-up. More than 20% of our patients did not have another visit within our health care system within 6 months. These patients in particular may have had important study outcomes that we did not capture. Baseline characteristics of these patients did not differ significantly from those of patients who did return within 6 months, but they may have differed in ways we did not measure. Patients who had a primary care physician designated in the system had better rates of 6-month follow-up (almost 90%), and a sensitivity analysis restricted to these patients did not produce different results.
Our study also has several strengths. To the best of our knowledge, ours is the first study to describe prevalence, characteristics, and outcomes of ambulatory patients presenting with hypertensive urgency across a large health care system. It is also the first study to compare and quantify short-term outcomes depending on the setting of management of hypertensive urgency.
Hypertensive urgency is common in the outpatient setting. In the absence of symptoms of target organ damage, most patients probably can be safely treated in the outpatient setting, because cardiovascular complications are rare in the short term. Furthermore, referral to the ED was associated with increased use of health care resources but not better outcomes. Finally, patients with hypertensive urgency are at high risk for uncontrolled hypertension as long as 6 months after the initial episode. Efforts to improve follow-up and intensify antihypertensive therapy should be pursued.
Corresponding Author: Krishna K. Patel, MD, Department of Internal Medicine, Medicine Institute, Cleveland Clinic Foundation, 9500 Euclid Ave, Desk G10-C, Cleveland, OH 44195 (firstname.lastname@example.org).
Accepted for Publication: February 29, 2016.
Published Online: June 13, 2016. doi:10.1001/jamainternmed.2016.1509.
Author Contributions: Drs Patel and Rothberg had full access to all the data in the study and take full responsibility for the integrity of the data and accuracy of data analysis.
Study concept and design: Patel, Howell, Rothberg.
Acquisition, analysis, or interpretation of data: Patel, Young, Howell, Hu, Rutecki, Thomas, Rothberg.
Drafting of the manuscript: Patel, Hu.
Critical revision of the manuscript for important intellectual content: Young, Thomas, Rutecki, Thomas, Rothberg.
Statistical analysis: Patel, Hu.
Administrative, technical, or material support: Young, Howell, Rothberg.
Study supervision: Rutecki, Thomas, Rothberg.
Conflict of Interest Disclosures: None reported.
Additional Contributions: Peter Imrey, PhD, and Shannon Morrison, MS, Department of Quantitative Health Sciences, Cleveland Clinic, helped with the statistical analysis of the project. No compensation was provided for these contributions.