When More Isn’t Better: Visits for Hypertension

Editor’s Note:

You are reading the 53rd installment of Annals of Emergency Medicine Journal Club. This Journal Club refers to the article by Masood¹ published in the September 2016 edition of Annals. Information about Journal Club can be found at http://www.annemergmed.com/content/journalclub. Readers should recognize that these are suggested answers. We hope they are accurate; we know that they are not comprehensive. There are many other points that could be made about these questions or about the article in general. Questions are rated “novice” (), “intermediate” (), and “advanced” () so that individuals planning a journal club can assign the right question to the right student. The “novice” rating does not imply that a novice should be able to spontaneously answer the question. “Novice” means we expect that someone with little background should be able to do a bit of reading, formulate an answer, and teach the material to others. Intermediate and advanced questions also will likely require some reading and research, and that reading will be sufficiently difficult that some background in clinical epidemiology will be helpful in understanding the reading and concepts. We are interested in receiving feedback about this feature. Please e-mail journalclub@acep.org with your comments.

DISCUSSION POINTS

Masood et al¹ stated that the study goal was to “understand the epidemiology, patient characteristics, and short- and long-term outcomes of emergency department (ED) patients with a primary diagnosis of hypertension.” They conducted a cohort study using administrative data from the National Ambulatory Care Reporting System, which contains deidentified information from all ED visits within the province of Ontario, Canada, beginning in 2002.

1. ED visits for hypertension between April 1, 2002, and March 31, 2012, among adult patients with a valid Ontario health care number were included in the study.

   • 
     How did this study define ED visits for hypertension?
   • 
     What was this definition’s positive predictive value for identifying hypertension-related ED visits according to manual review of medical records? With the information provided in the methods, can you calculate the negative predictive value, sensitivity, or specificity? If not, what information do you need to compute these?
   • 
     Diagnosis codes in Canada do not influence reimbursement for ED visits. Name 2 alternative explanations for why an ED visit might erroneously have hypertension listed as the primary, final diagnosis.
2. These authors report that ED visits for hypertension are increasing yet the “feared complications of hypertension are extremely infrequent.”¹

   • 
     According to this study, emergency physicians should anticipate increasing visits for hypertension. Summarize and discuss the international guideline recommendations about which patients should receive immediate blood pressure reduction in the ED and the evaluation, treatment, and follow-up recommendations for those who do not require immediate blood pressure reduction.
   • 
     Discuss your clinical practice when treating a patient with hypertension who is not currently receiving antihypertensive medications. Do you prescribe a blood pressure–decreasing medication at discharge? If yes, what patient characteristics (age, sex, race, comorbidities, chief complaint, existing antihypertensive prescription, etc) do you consider when you make this decision? If no, explain your rationale for not starting or titrating antihypertensive medications.
   • 
     Recent work by Patel et al² has raised questions about the role of the ED in evaluation or treatment of patients with severely elevated blood pressure (≥180/110 mm Hg) or hypertensive urgency. During 7 years across the Cleveland Clinic health care system, less than 1% of patients with a clinic blood pressure greater than or equal to 180/100 mm Hg were referred to the ED. Among patients who were sent to the ED, only (2%) had pulmonary edema, acute kidney injury, or elevated cardiac biomarker levels, and 80% were discharged home. Discuss the findings of these 2 articles together. How is the role of the ED within the larger health care system evolving in the United States and Canada? What role can the ED play for patients with asymptomatic elevated blood pressure? Is “hypertensive urgency” a useful term?
3. ED visits for hypertension increased from 15,793 in 2002 to 25,950 in 2012. The results included the raw number of visits, crude rate of visits, and the age- and sex-standardized ED visit rates from 2002 to 2012.

   • 
     What is one reason to report age- and sex-standardized ED visit rates for hypertension, in addition to the raw number and crude rates?
   • 
     What reference population did the authors use for the standardized ED visit rates?
4. • 
     Discuss the difference between statistically significant differences and clinically important differences. In a study with a sample size of 206,147 ED visits, how might these 2 concepts influence interpretation of results?
   • 
     Figure 4 in the article by Masood¹ presents the proportion of subsequent hospitalizations for patients who were admitted versus discharged from the ED, according to categories of hypertension complications (stroke, heart failure, acute myocardial infarction, atrial fibrillation, renal failure, aortic dissection, or encephalopathy). The proportions are higher overall for patients who were admitted from the ED. Does this mean that being admitted increases the risk of being hospitalized in the future for complications of hypertension? If not, why not?

ANSWER 1

Masood¹ stated that the study goal was to “understand the epidemiology, patient characteristics, and short- and long-term outcomes of emergency department (ED) patients with a primary diagnosis of hypertension.” They conducted a cohort study using administrative data from the National Ambulatory Care Reporting System, which contains deidentified information from all ED visits within the province of Ontario, Canada, since 2002.

Q1. ED visits for hypertension between April 1, 2002, and March 31, 2012, among adult patients with a valid Ontario health card number were included in the study.

Q1.a How did this study define ED visits for hypertension?

ED visits for hypertension were defined as adult ED visits in the province of Ontario in which the patient had a final, primary ED diagnosis of hypertension as defined by any of the following International Classification of Diseases, 10th Revision (ICD-10) codes: I10, I11, I12, I13, and R030. Patients had to have a valid Ontario health card number to be included in the provincial health administrative database (the Canadian Institutes of Health Information National Ambulatory Care Reporting System). In Ontario, all ED visits have been reported to this database since 2002.

It is not entirely clear whether hospital diagnosis codes were also considered for patients who were hospitalized from the ED.

Q1.b What was the definition’s positive predictive value for identifying hypertension-related ED visits according to manual review of medical records? With the information provided in the methods, can you calculate the negative predictive value, sensitivity, or specificity? If not, what information do you need to compute these?

The positive predictive value (PPV) of using ICD-10 codes I10, I11, I12, and I13 for a primary diagnosis of an ED visit for hypertension was 95.7% (95% confidence interval 94.6% to 96.7%). To obtain the PPV, the investigators divided the number of cases identified by both the ICD-10 definition and chart review (the “reference standard”) by the total number of ED visits for hypertension that were identified by their ICD-10 definition. The PPV is sometimes called the “true positive proportion” or, stated otherwise, the proportion of test positives (ICD-10 diagnosis) that are true positives (chart-review-confirmed cases of hypertension). If we round PPV to 96%, this means that of 100 ED visits identified as being for hypertension by the ICD-10 definition, 4 would have been found on chart review to not actually have been for hypertension (Figure). The reported PPV is derived from a small subset (approximately 0.8%) of all ED visits for hypertension abstracted from the National Ambulatory Care Reporting System and thus “predicts” the relative accuracy of coding in the database, not the clinical diagnosis itself.

Figure.

2×2 Table for calculating hypertension-related ED visits.

To calculate the negative predictive value, we would need to know the number of ED visits in which an ICD-10 code other than I10, I11, I12, or I13 was used for patients with and without a true diagnosis of hypertension. However, the latter cannot be derived from the existing study database because it includes only individuals with specific hypertension-related ICD-10 codes. For the same reason, neither specificity (which requires knowledge of patients with true hypertension who had other diagnosis codes) nor sensitivity (which requires inclusion of patients without true hypertension) can be calculated.

Q1.c Diagnosis codes in Canada do not influence reimbursement for ED visits. Name 2 alternative explanations for why an ED visit might erroneously have hypertension listed as the primary, final diagnosis.

There are several potential alternative explanations for errors in classifying ED visits. Patients who presented to the ED with a complaint for which there was no clear diagnosis but who had an elevated blood pressure (BP) may have had their ED visit coded as being due to hypertension when in fact this finding was incidental and not directly related to the ED visit. Alternatively, for patients with a variety of complaints, clinicians may list the first diagnosis code that comes to mind, rather than the one that may be most relevant to the visit. A diagnosis of hypertension may be given for patients who, despite having a normal BP measurement in the ED, have an underlying diagnosis of chronic hypertension and concurrent, perhaps even long-standing, hypertension-related target-organ damage. Last, it is not uncommon for patients who present to the ED with elevated BP not to have underlying hypertension. Such patients may receive a coded diagnosis of hypertension, particularly if they were referred to the ED for elevated BP from a provider or screening event.

ANSWER 2

Q2. These authors report that ED visits for hypertension are increasing yet the “feared complications of hypertension are extremely infrequent.”¹

Q2.a According to this study, emergency physicians should anticipate increasing visits for hypertension. Summarize and discuss the international guideline recommendations regarding which patients should receive immediate blood pressure lowering in the ED and the evaluation, treatment, and follow-up recommendations for those who do not require immediate blood pressure reduction.

Patients presenting to the ED with elevated BP—for example, when the BP is consistently greater than or equal to 180/120 mm Hg—should be evaluated for evidence of acute target-organ damage. Strong evidence does not currently exist for BP thresholds; for example, “severe [BP] elevation” defined in the Seventh Report of the Joint National Committee as “≥180/120 mm Hg” was based on expert opinion and does not include a reference.

Evaluations for evidence of acute target-organ damage may include a basic metabolic panel to determine renal function and critical electrolyte concentrations, complete blood count to evaluate for microangiopathic hemolytic anemia, urinalysis to look for proteinuria, ECG for evidence of left ventricular hypertrophy, or a chest radiograph to evaluate for cardiomegaly, according to the treating clinician’s discretion.² However, existing data suggest that, absent associated signs or symptoms of acute target-organ damage, identification of clinically important abnormalities with such testing is unlikely and may, as with the ECG, be falsely reassuring.³ The ACEP Clinical Policy for ED patients with asymptomatic elevated BP recommends against routine screening and testing of target organ damage (level C recommendation), except for patients with poor follow-up.⁴

National and international guidelines recommend that patients with signs or symptoms of acute target-organ dysfunction related to hypertension be treated as having true hypertensive emergencies, with admission to a monitored bed and administration of parenteral BP-decreasing medication.^5,6 Generally, a 25% reduction in BP should be targeted in these cases according primarily to expert recommendations, but guidelines for specific diseases, such as intracranial hemorrhage and acute heart failure, may vary.

For patients who have markedly elevated BP without evidence of acute target-organ damage, BP reduction should be pursued during the course of days to weeks with oral medications. The 2006 and 2013 ACEP Clinical Policy publications addressing asymptomatic hypertension in the ED strongly emphasize the importance of outpatient follow-up and provide the following level B and C recommendations (which reflect moderate clinical certainty based on Class II recommendations and consensus opinion, respectively)^4,7:

1. Start long-term oral antihypertensive treatment in the ED primarily for patients without adequate follow-up (B/C).

2. Avoid rapid BP reduction to prevent development of avoidable harm (B).

3. Lower BP gradually for patients whom are started on oral treatment.

4. Avoid normalizing BP during the initial ED visit (B).

More recent work examining asymptomatic primary care patients (n=59,535) with markedly elevated BP (≥180/110 mm Hg) supports this approach with a very low risk of adverse cardiovascular events whether patients were referred to the ED (n=387) or not.⁸ Even 6 weeks of sustained, uncontrolled hypertension are associated with increased risk of worse outcome,⁹ suggesting that a goal of gradual but not delayed BP reduction facilitated by active follow-up of such patients from the ED should be the idealized norm.

Q2.b Discuss your clinical practice when treating a patient with hypertension who is not currently on antihypertensive medications. Do you prescribe a blood pressure-decreasing medication at discharge? If yes, what patient characteristics (age, sex, race, comorbidities, chief complaint, existing antihypertensive prescription, etc) do you consider when you make this decision? If no, explain your rationale for not starting or titrating antihypertensive medications.

There is great practice variability in the management of patients with asymptomatic elevated BP. Evaluation for potential reversible causes such as use of nonsteroidal anti-inflammatory medications or cocaine use is important to help determine whether to initiate treatment based on ED BP measurements. In general, ED patients with BPs that are consistently greater than or equal to 160/100 mm Hg (especially if also elevated during previous ED visits) will, in most cases, have hypertension at follow-up, regardless of pain or anxiety during the ED visit. A reasonable approach for patients with established hypertension (or suspected, undiagnosed hypertension) is to presume inadequate control and consider prescribing a first-line oral antihypertensive. This should only be done after assessment of baseline kidney function and electrolyte balance (because these will influence medication choices) and after a discussion about follow-up and treatment options (eg, lifestyle changes, medication adverse-effect profiles) with the patient.

First-line medications include thiazide (eg, hydrochlorothiazide) or thiazidelike (eg, chlorthalidone) diuretics at 12.5 or 25 mg/day, long-acting calcium-channel blockers (eg, amlodipine at 5 or 10 mg/day), or angiotensin-converting enzyme (ACE) inhibitors or angiotensin-receptor blockers (ARBs) (eg, lisinopril 10 to 20 mg/day, enalapril 2.5 to 5 mg/day, losartan 25 to 50 mg/day). Although existing guidelines are flexible in regard to first-line choice, younger patients may respond better to ACE inhibitors or ARBs. ARBs should be reserved for patients with severe hypertension with ECG evidence of left ventricular hypertrophy¹⁰ or for those who do not tolerate ACE inhibitors. Teratogenicity should be taken into account when prescribing these medications to women with childbearing potential is considered. Black patients and older adults may respond best to thiazide or thiazidelike diuretics and long-acting calcium channel blockers. ACE inhibitor or ARB therapy is associated with better outcomes in patients with a history of coronary artery disease, heart failure, or chronic kidney disease and should strongly be considered.^11–13

The decision about whether and which BP medication to prescribe should be made in conjunction with the patient, taking into consideration the patient’s understanding of hypertension, willingness or ability to tolerate potential adverse effects, dosing schedule (once-daily dosing and combination medications are generally preferred by patients), risk of drug-drug interactions, ability to pay for the medication, and necessary follow-up frequency (eg, more frequent laboratory testing may be needed for patients treated with a diuretic or ACE inhibitor/ARB).

Although many ED clinicians may choose not to prescribe medication on discharge, starting antihypertensives in the ED has been shown to be effective and without evidence of harm in a recent small prospective study.¹⁴

Q2.c Recent work by Patel et al¹⁵ has raised questions about the role of the ED in the evaluation or treatment of patients with severely elevated blood pressure (≥180/110 mm Hg), or “hypertensive urgency.” During 7 years across the Cleveland Clinic Healthcare System, less than 1% of such patients were referred from their clinics to the ED. Among patients who were sent to the ED, only (2%) had pulmonary edema, acute kidney injury, or elevated cardiac biomarker levels, and 80% were discharged home. Discuss the findings of these 2 articles together. How is the role of the ED within the larger health care system evolving in the United States and Canada? What role can the ED play for patients with asymptomatic elevated blood pressure? Is “hypertensive urgency” a useful term?

A very small group of patients with elevated BP will have symptoms suggestive of acute target-organ damage and will require immediate BP reduction with parenteral antihypertensive therapy. However, the majority of patients presenting to the ED for hypertension will either be discharged or will be hospitalized for other reasons. Nonetheless, emergency clinicians can play a key role in identifying undiagnosed or uncontrolled hypertension through screening and provide recommendations for further testing and treatment. Because many ED patients have limited interaction with the health care system, they may go years without a BP measurement. The ED visit could represent an important teachable moment for such patients. Not all patients will need antihypertensive therapy. Counseling patients on the benefits of lifestyle modifications such as salt and alcohol restriction, increased physical activity, tobacco cessation, and weight loss should be discussed, particularly for patients with elevated BP but without an established diagnosis of chronic hypertension.

Hypertensive urgency is a term that has been used historically to describe patients with BP greater than or equal to 180/110 mm Hg and without acute target-organ damage. As described above, there is increasing evidence that BP for these patients should be decreased during the course of days to weeks, and the goal for an ED encounter should not be BP normalization during the index visit. As noted by Gallagher¹⁶ in his pointed 2003 commentary titled “Hypertensive Urgencies—Treating the Mercury?,” the only feature of a putative hypertensive urgency that might conceivably be urgent is not the need for prompt (or even gradual) BP reduction during the ED visit, but rather the need for expeditious outpatient follow-up. Thus, the term hypertensive urgency itself may be a misleading anchor that encourages clinicians to do something immediate for a condition based on an unsubstantiated, and in many ways quixotic, fear of short-term, adverse consequences from perceived inaction.

Despite this understanding, there remains much confusion among patients and non–emergency medicine clinicians about when (if ever) patients should seek acute care for asymptomatic, markedly elevated BP. Masood and Atzema¹ stated that recent Canadian and European guidelines supporting the use of home BP monitoring as a self-management tool for chronic hypertension may have inadvertently contributed to their observed secular trend in ED visits for hypertension. Although that is an intriguing hypothesis, the authors cite limited evidence to support this and correctly encourage efforts to improve patient and professional education as it relates to appropriate ED use in the setting of elevated BP. This is perhaps the most important take-home message from the study because the increase in ED utilization over time did not appear, at least overtly, to be related to issues of changing acuity or clinician practice patterns.

ANSWER 3

Q3. ED visits for hypertension in Ontario, Canada increased from 15,793 in 2002 to 25,950 in 2012. The results included the raw number of visits, crude rate of visits, and the age- and sex-standardized ED visit rates from 2002 to 2012.

Q3.a What is one reason to report age- and sex-standardized ED visit rates for hypertension, in addition to the raw number and crude rates?

In a population with a higher proportion of older patients, crude rates of ED visits for hypertension may be higher simply because hypertension is more common with increasing age. Without standardizing for age, for example, we cannot tell whether an increase in the number of ED visits for hypertension was because the number of older patients increased (for example, because patients with hypertension survived longer) or whether ED visits for hypertension really did increase over time.

This is an “age effect,” which is the consequence of populations growing older rather than a true increase in prevalence over time. Age- and sex-standardized ED visit rates allow us to correct for the age and sex distribution of a population and make a valid comparison over time or across populations with different age or sex distributions.

Similar situations can occur when there are differences by sex or other key factors. Of relevance to this analysis, there was a small but statistically significant decrease in the proportion of women with a primary ED diagnosis of hypertension over time. Given that uncontrolled hypertension is more prevalent in men,¹⁷ sex standardization is critically important to the validity of their results.

Q3.b What reference population did the authors use for the standardized ED visit rates?

The 2002 population of Ontario, analyzed in 5-year groups, was used as the reference population.

ANSWER 4

Q4.a Discuss the difference between statistically significant differences and clinically important differences. In a study with a sample size of 206,147 ED visits, how might these two concepts influence interpretation of results?

Clinical importance is determined by clinical judgment: is the difference between the 2 groups clinically meaningful, and is it likely to influence the risk of disease or outcome?

In contrast, statistical significance is an arbitrary threshold defined by convention during the past century. The threshold of P<.05 is typically used to define statistical significance for historical reasons. Rampant misunderstandings about the meaning of statistical significance and P values prompted a recent statement by the American Statistical Association.¹⁸ A P value says nothing about whether the hypothesis being tested was true or not or about the probability that the “true value” is within the 95% confidence interval. Instead, the P value is the probability, assuming the hypothesis is correct, of obtaining the observed results or a more extreme result because of chance alone. There has been much written about the many, many problems with typical interpretation of P values.¹⁹

Statistical significance is closely tied to sample size. Studies with large sample sizes can produce “statistically significant results” even when between-group differences are quite small. Regardless of the P value, clinicians must decide for themselves whether differences are clinically important.

Q4.b Figure 4 of the article by Masood and Atzema¹ presents the proportion of subsequent hospitalizations for patients who were admitted versus discharged from the ED, according to categories of hypertension complications (stroke, heart failure, acute myocardial infarction, atrial fibrillation, renal failure, aortic dissection, or encephalopathy). The proportions are higher overall for patients who were admitted from the ED. Does this mean that being admitted increases the risk of being hospitalized in the future for complications of hypertension? If not, why not?

Patients admitted from the ED likely represent a higher-risk group because of the presence of other comorbidities and the severity of their presenting symptoms, although a variety of other factors, including socioeconomics and linkage to care, can contribute to patient disposition decisions. Although the latter is less likely to have confounded disposition in this study because it was conducted within the Canadian health care system, in general, healthier, health-literate patients with adequate follow-up and secure home situations are more likely to have better outcomes.