Non-adherence to Antihypertensive Guidelines in Patients with Asymptomatic Carotid Stenosis

Abstract

Importance:

Hypertension and carotid stenosis are both risk factors for stroke, but the presence of carotid stenosis might dampen enthusiasm for tight control of hypertension because of concerns for hypoperfusion.

Objective:

To determine the extent to which there are opportunities to potentially improve pharmacotherapy for hypertension in patients known to have asymptomatic high-grade carotid stenosis.

Design:

We examined anti-hypertensive medication prescription and adherence to evidence-based hypertension treatment guidelines in a cross-sectional analysis of baseline data of patients enrolled in a clinical trial.

Setting:

The Carotid Revascularization and Medical Management for Asymptomatic Carotid Stenosis Trial (CREST-2) is a multicenter prospective randomized open blinded end-point clinical trial of intensive medical management with or without revascularization by endarterectomy or stenting for asymptomatic high-grade carotid stenosis.

Participants:

1479 participants (38.6% female; mean age 69.8 years) from 132 clinical centers enrolled in the CREST-2 trial as of April 6, 2020 who were taking ≥1 antihypertensive drug at baseline.

Exposures:

Pharmacotherapy for hypertension.

Main Outcome:

Adherence to evidence-based guidelines for treating hypertension.

Results:

Of 1458 participants with complete data, 26% were on one, 31% on 2, and 43% on ≥3 antihypertensive medications at trial entry. Thirty-two percent of participants were prescribed thiazide; 74%, angiotensin converting enzyme inhibitor (ACEI) or angiotensin receptor blocker (ARB); 38%, calcium channel blocker (CCB); 56%, a beta blocker; 11%, loop diuretic; and 27%, other. Of those prescribed a single antihypertensive medication, the proportion prescribed thiazide was 5%; ACEI or ARB, 55%, and CCB, 11%. The prevalence of guideline-adherent regimens was 34% (95% CI, 31–36%).

Conclusions and Relevance:

In a diverse cohort with severe carotid disease and hypertension, non-adherence to hypertension guidelines was common. All preferred classes of antihypertensive drug were under-prescribed. Using staged iterative guideline-based care for hypertension, CREST-2 will characterize drug tolerance and stroke rates under these conditions.

Trial Registration:

ClinicalTrials.gov Number NCT02089217.

Introduction

Recommendations for first-line drugs for treatment of hypertension have been made based in several evidence-based guidelines, most recently JNC-8 in 2014, AHA/ACC/ASH in 2015 and in 2017. ^{1; 2; 3} A cross-sectional analysis of baseline data from the Systolic Blood Pressure Intervention Trial (SPRINT) found suboptimal use of the first-line agent, thiazide, at baseline in patients at higher risk for vascular disease, suggesting that clinical practice significantly diverges from guideline recommendations, possibly compromising best outcomes. ³

The Carotid Revascularization and Medical Management for Asymptomatic Carotid Stenosis Trial (CREST-2) is a set of two parallel randomized multi-center trials comparing outcomes from carotid endarterectomy (CEA) and intensive medical management (IMM) versus IMM alone, and carotid stenting (CAS) and IMM versus IMM alone for asymptomatic carotid stenosis. ⁴ Whether use of current guideline-based therapies and IMM with emphasis on optimal treatment of hypertension will support the previous findings of risk reduction in favor CEA or CAS for asymptomatic severe carotid artery disease is the focus of CREST-2.

We hypothesized that patients with high-grade asymptomatic carotid stenosis and hypertension uncommonly receive antihypertensive drug therapy that adheres to evidence based guidelines. The aim of this study was to assess the validity of this hypothesis by performing a cross-sectional analysis of baseline antihypertensive drug use in CREST-2.

Methods

Study population

CREST-2 patients are screened and enrolled into the trial across 127 centers in US, 4 centers in Canada, and one center in Spain. The primary outcome is the composite of stroke and death within 44 days following randomization and stroke ipsilateral to the target vessel thereafter up to four years. Change in cognition and differences in major and minor stroke are secondary outcomes. Enrollment in each trial is ongoing. Patients at least 35 years of age with high-grade stenosis (≥70%) of the cervical internal carotid artery and no history of stroke or transient ischemic attack ipsilateral to the stenosis within 180 days of randomization were potentially eligible for inclusion. Those with serum creatinine ≥2.5 mg/dl or estimated glomerular filtration rate (eGFR) <30 ml/min/1.73m2 were excluded. Details of eligibility criteria have been published. ⁴

Baseline characteristics and antihypertensive treatment

During the baseline visit, study personnel obtained information on participant characteristics, including co-morbid conditions (e.g. coronary artery disease), demographics, vascular risk factors, height, weight, blood pressure, smoking history, symptoms of stroke using the Questionnaire for Verifying Stroke-free Status (QVSS),⁵ the modified Rankin Scale, the National Institutes of Health Stroke Scale, cognitive testing and selected blood tests. All prescribed baseline antihypertensive medications were documented and were classified in the following categories: thiazide diuretics, angiotensin converting enzyme inhibitors (ACEI), angiotensin receptor blockers (ARB), calcium channel blockers (CCB), loop diuretics, epithelial sodium channel blockers, direct renin inhibitors, beta blockers, alpha blockers, mineralocorticoid receptor antagonists, centrally acting agents, and direct vasodilators.

Assessing adherence to hypertension guidelines

We examined antihypertensive medication prescription at baseline to determine if it met specific criteria for guideline-adherent care, defined according to age, race, and CKD status. ^{1; 2;6} CKD status was determined by inputting baseline clinical and laboratory data into the MDRD equation as follows: eGFR in mL/min/1.73m² = 175 × (serum creatinine in mg/dL)^−1.154 × (age in years)^−0.203 × (0.742 if female) × (1.212 if black). The drug regimen was considered adherent to guidelines in patients <60 years old, non-black, and without chronic kidney disease (CKD) if it consisted of: thiazide or ACEI (or ARB) or CCB if prescribed one drug; or any combination of these classes of medications, if prescribed 2 medications; or all 3 of these classes of medications, if prescribed ≥ 3 medications. The medication regimen was considered adherent to guidelines in patients age ≥60 years or black if it consisted of: thiazide or CCB if prescribed one medication; a combination of thiazide and CCB (add whichever not added as first-line), if prescribed 2 medications; or a combination of thiazide, CCB, and ACEI (or ARB), in addition to whatever else prescribed if prescribed ≥3 medications. The medication regimen was considered adherent to guidelines in patients with CKD if it consisted of: ACEI (or ARB), if on one medication; ACEI or ARB plus thiazide (or loop diuretic) or CCB, if prescribed two medications; or ACEI (or ARB) plus thiazide (or loop diuretic) and CCB, in addition to whatever else prescribed, if prescribed ≥3 medications. The medication regimen was considered adherent to guidelines in patients with a compelling cardiovascular disease indication such as coronary artery disease or heart failure if it consisted of: beta blocker, if prescribed one medication; or beta blocker plus medication classes as described above depending upon age, race, or CKD status, if prescribed ≥2 medications. In terms of guideline adherence, the presence of CKD overrides other considerations including whether black or whether age is greater than 60 years.

Statistical analysis

We examined patterns and prevalence of anti-hypertensive medication prescription in a cross-sectional analysis of baseline data from 1458 participants enrolled in the CREST-2 trial who were taking at least 1 antihypertensive medication at baseline. The main outcome variable was evidence of adherence to hypertension treatment guidelines. We compared socio-demographic variables, history of cardiovascular disease, and kidney function between participants prescribed guideline-adherent regimens with those prescribed guideline-non-adherent regimens.

Results

Guideline adherence

Of the 1648 participants enrolled in CREST-2 as of April 6, 2020, 1479 were prescribed at least one antihypertensive medication at baseline, and 1458 had complete data required for this analysis. This constituted the cohort for this study. The overall prevalence of medication regimens that adhered to hypertension treatment guidelines was 34% (95% CI 31%−36%). In unadjusted analyses, comparing adherent to non-adherent groups, a larger proportion of those < age 60 years, with diabetes, and with CKD were prescribed a guideline-adherent regimen. Those prescribed a guideline-adherent regimen were taking more medications at baseline (2.7 vs. 2.3; p<0.001) and were more likely to be < age 60 years and to have diabetes and CKD (Table 1). Of the 807 patients with a compelling indication for a beta blocker, 74% (n=598) were prescribed a beta blocker. Table 2 shows that patients adherent to guidelines were more likely to be treated with a diuretic or a calcium channel blocker.

Table 1.

Baseline characteristics of CREST-2 participants prescribed at least 1 antihypertensive drug, overall and by adherence to hypertensive guidelines status

	Total (n=1458)	Guideline adherent (n=490)	Guideline non- adherent (n=968)	p-value
Age, mean (SD), years	69. (7.8)	68.8 (8.8)	70.3 (7.2)	0.001
Age ≥60 years (%)	89.5	81.2	93.7	<0.0001
Black (%)	7.2	8.4	6.5	0.19
Female (%)	38.6	36.5	39.7	0.24
Diabetes (%)	39.5	43.4	37.5	0.03
Dyslipidemia (%)	92.8	93.1	92.6	0.76
Hypertension (%)	92.0	91.4	92,4	0.54
SBP, mean (SD), mmHg	139.7 (20.5)	139.5 (20.9)	139.9 (20.4)	0.76
DBP, mean (SD), mmHg	72.5 (10.7)	72.5 (10.7)	72.5 (10.7)	0.99
Prior CVD (%)	58.4	57.5	58.8	0.64
CKD (eGFR <60 ml/min/1.73m2) (%)	32.0	52.5	21.7	<0.001
Current smoker (%)	22.9	20.6	24.1	0.14
Alcohol use (%)	60.5	61.2	60.2	0.73
No. of antihypertensive drugs, mean (SD)	2.4 (1.2)	2.7 (1.5)	2.3 (1.0)	<0.0001

Table 2:

Types of Antihypertensive drugs in CREST-2 participants prescribed at least 1 antihypertensive drug, overall and by adherence to hypertensive guidelines status.

	Total (n=1458)	Guideline adherent (n=490)	Guideline non-adherent (n=968)	p-value
Diuretics (%)	41.4	53.7	35.1	<0.0001
ACEI (%)	47.6	44.7	49.1	0.11
ARB (%)	27.5	31.4	25.5	0.02
CCB (%)	38.3	53.5	30.6	<0.0001
Beta Blocker (%)	56.0	60.4	53.7	0.02
Other HTN drug (%)	27.3	18.2	31.9	<0.0001

Baseline drug prescription patterns

Twenty-six percent (n = 372) of the cohort were prescribed one antihypertensive medication; 31% (n = 453) were prescribed 2 medications; and 43%, ≥3 medications (n = 633). Thirty-two percent (n = 462) of all participants were prescribed a thiazide; 74% (n = 1082), ACEI or ARB; 38% (n = 558), CCB; 56% (n = 816), beta blocker; 11% (n = 162) loop diuretic; and 27% (n = 398) at least one other antihypertensive medication. A combination of ACEI and ARB was prescribed in 13 patients (<1%).

The distribution of antihypertensive medication classes in those prescribed a single medication, 2 medications, or ≥3 medications is shown in the Figure. In the group prescribed a single medication at baseline, 5% were prescribed thiazide, and 70% were prescribed thiazide, ACEI or ARB, or CCB. For those prescribed 2 medications, 28% were prescribed thiazide, and 40% a combination of thiazide, ACEI or ARB, or CCB. For those prescribed ≥3 medications, 50% were prescribed a thiazide, and 26% were prescribed all 3 classes: thiazide, ACEI or ARB, and CCB.

Figure 1.

Distribution (%) of antihypertensive drugs at time of enrollment into the CREST-2 trial by class in participants prescribed: single drug (n = 372); two drugs (n =453); 3 or more drugs (n =633). ACEI denotes Angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; CCB, calcium channel blocker; and Loop, loop diuretic.

Subgroup analyses

Of those participants < 60 years old (153/1458), 23% were prescribed thiazide; 30%, CCB; and 55% neither thiazide nor CCB. Of those ≥60 years old, 33% were prescribed thiazide; 39%, CCB; and 42%, neither thiazide nor CCB. Of those participants ≥60 years old who were prescribed a single medication, 17% were prescribed a thiazide or CCB. Of those participants ≥60 years old who were prescribed 2 medications, 3% were prescribed both thiazide and CCB. Of those participants ≥60 years old who were prescribed ≥3 medications, 27% were prescribed all 3 of these preferred classes (thiazide, CCB, ACEI or ARB).

Among black participants, 50/104 (48%) were prescribed thiazide, 67/104 (64%) CCB, and 19/104 (18%) neither. If prescribed a single medication, 7 of 13 (54%) were prescribed thiazide or CCB; if 2 medications, 2 of 27 (7%) were prescribed both thiazide and CCB, and if ≥3 medications, 24 of 64 (38%) were prescribed all 3 of these classes. Of those participants with CKD, 73% (342/466) were prescribed ACEI or ARB; 54% were prescribed ACEI or ARB plus diuretic (thiazide or loop) or CCB; 9% were prescribed none of the latter classes. If prescribed a single medications, 41 of 91 (45%) were getting ACEI or ARB; if 2 medications, 51 of 118 (43%) were prescribed a combination of ACEI or ARB and a diuretic or CCB, and if ≥3 medications, 98 of 257 (38%) were prescribed all 3 of these classes. Of the 56% of participants prescribed beta blockers, 73% had a compelling CVD indication, however, 209/807 (26%) with compelling indication were not prescribed beta blocker.

Discussion

The main finding of our study was that among patients prescribed antihypertensive medications, a minority of patients entering the CREST-2 trial were prescribed a drug regimen that adhered to current evidence-based guidelines for the management of blood pressure. It seems implausible that this finding could be explained by a general lack of access to healthcare. Every patient had access to healthcare sufficient to undergo imaging to identify carotid stenosis and be considered for a revascularization procedure. Because there is a tendency for trial participants to be more health-conscious than patients with the same condition who do not participate in trials, the problem of lack of adherence to antihypertensive guidelines may be worse for the general patient population with asymptomatic carotid stenosis than what we have observed at enrollment in CREST-2.

Several factors may contribute to the evidence-practice gap regarding treating hypertension in patients with carotid stenosis. For initiation of antihypertensive drug therapy, first-line agents include thiazide diuretics, CCBs, and ACE inhibitors or ARBs.^7;8 The guidelines recommend consideration of co-morbid conditions in choosing which of these should be given first when a single medication is appropriate (for example, CKD: first should be ACEI or ARB). On the other hand, in head to head trials,^9;10 thiazide-type diuretics were superior to CCB and ACEI/ARB in preventing heart failure. In preventing stroke, thiazides were superior to ACEI and CCBs. For black patients and those over age 60, thiazides or CCBs were the best initial choice for single drug therapy. Our choice to examine guideline adherence defined according to age, race, and CKD status follows from these considerations. Otherwise, meta-analyses and the largest individual RCT that compared first-step agents⁹ have found that diuretics, especially thiazide-type, to be an optimal choice. We found thiazides to be under-prescribed; 32% of all participants were prescribed a thiazide at baseline. In the group prescribed a single medication, that proportion was 5%; in those prescribed 2 medications, 28%, and in those prescribed 3 agents, 50%. Fifty-six per cent of patients in our study were prescribed a beta blocker, which was considered first-line when associated with compelling indication, such as in patients with coronary artery disease including stable angina or previous myocardial infarction. Otherwise, beta blockers are not first-line as they have been found to be less effective in preventing cardiovascular events and stroke when compared to thiazides CCBs, and renin angiotensin system blockers.^{11; 12} In our study, of participants prescribed beta blockers, 73% had a compelling indication, while 26% with compelling indication for beta blocker use were not prescribed a beta blocker.

Fear of over treatment could have also played a role in the evidence-practice gap we documented. There are at least two well-described clinical syndromes associated with systemic hypotension and high-grade carotid stenosis: limb shaking transient ischemic attacks and internal watershed infarcts. The etiologic relevance of carotid stenosis in limb shaking transient ischemic attacks is supported by the nearly universal resolution of the attacks with successful revascularization.¹³ Internal watershed infarcts are infarcts located between the deep and superficial perforating arteries of the middle cerebral artery. Carotid steno-occlusive disease is seen more often in patients with internal watershed infarcts than in patients with cortical watershed infarcts.¹⁴ Clinicians may be concerned of the potential to over-treat hypertension and cause transient ischemic attacks or infarcts.

Despite concerns regarding over-treatment, there are several lines of evidence supporting aggressive management of hypertension in patients with carotid atherosclerotic stenosis. In a clinical-pathological study of 1684 patients, preoperative hypertension was associated with carotid plaque macrophages, lipid rich core, and intraplaque hemorrhage, all markers of unstable plaque.¹⁵ There is also compelling evidence from cohort studies that blood pressure control reduces risk of stroke in patients with asymptomatic carotid stenosis.¹⁶ Analysis of pooled data from ECST and NASCET showed the relationship between blood pressure and stroke risk was not affected by the presence of a unilateral carotid occlusion.¹⁷ Similarly, post-hoc analysis of BP control in the COSS study found mean BP < 130 was associated with a lower risk of ipsilateral ischemic stroke among patients with recently symptomatic carotid occlusion and hemodynamic failure.¹⁸

Our study shows that the enrolled population in CREST-2 is rich in opportunities for improving medical care of hypertension. Unlike prior major randomized clinical trials of carotid revascularization with or without medical management, like ACAS and ACST, the CREST-2 trial strives to achieve stringent atherosclerotic risk factor control through the implementation of guideline-based risk factor modification algorithms. Use of guideline-based risk factor modification algorithms have been shown to improve atherosclerotic risk factor control outside of clinical trials.¹⁹ We had previously found in an earlier trial directly comparing endarterectomy to stenting in symptomatic and asymptomatic patients with carotid stenosis (CREST) that merely encouraging site-level adherence to guidelines for management of atherosclerotic risk factors led to significant but woefully inadequate improvements in the risk factor profiles of trial participants.²⁰ Spurred on by these findings, the design of the subsequent CREST-2 trial has incorporated a central core that monitors and advises on management of blood pressure and lipids. ⁴