Resistant hypertension-defining the scope of the problem

Abstract

The updated scientific statement by the American Heart Association has defined resistant hypertension (HTN; RH) as uncontrolled blood pressure (BP) ≥ 130/80 mmHg, despite concurrent use of 3 anti-HTN drug classes comprising a calcium channel blocker, a blocker of renin-angiotensin system, and a thiazide diuretic, preferably chlorthalidone. Using the updated BP criteria, the prevalence of RH in the United States is found to be modestly increased by approximately 3–4% among treated population. Meta-analysis of observational studies have demonstrated that pseudo-RH from white coat HTN or medication nonadherence is as much common as the truly RH. Thus, screening for pseudo-resistance in the evaluation of all apparent RH is of utmost importance as diagnosis of white-coat HTN requires no treatment, while medication nonadherence would benefit from identifying and targeting barriers to adherence.

Introduction

Elevated blood pressure (BP) is the largest single risk factor contributing to the global deaths of at least 9.4 million per year.¹ A recent estimate indicated that >870 million adult population globally have systolic BP (SBP) above 140 mmHg.^1,2 Disparities in hypertension (HTN) prevalence and control rates continues to widen.² From 2000 to 2010, prevalence of controlled HTN increased by from 18 to 28% in high-income countries but paradoxically reduced from 8.4 to 7.7% in low- and middle-income countries.² In the United States (US), prevalence of HTN has recently increased from 32% to 46% as the result of lower BP threshold proposed by the American College of Cardiology (ACC)/American Heart Association (AHA) High BP guideline from 140/90 to 130/80 mmHg.³ Although this threshold is not adopted uniformly in all countries, many guidelines begin to reduce SBP target to <130 mmHg at least in the high-risk population.⁴

Definition and epidemiology

Following the 2017 ACC/AHA high BP guideline, to the AHA scientific statement addressing resistant HTN (RH) has been updated in 2018.⁵ RH is defined as SBP at least 130 mmHg or diastolic BP (DBP) at least 80 mmHg despite the concurrent use of 3 anti-HTN drug classes at the maximal or near maximal doses. The combination of drugs comprising a long-acting calcium channel blocker (CCB), a blocker of the renin-angiotensin system (RAS), such as angiotensin-converting enzyme inhibitor (ACEI) or angiotensin receptor blocker (ARB), and a diuretic is preferred before RH diagnosis can be ascertained.⁵ RH also included treatment with ≥4 classes of anti-HTN medication regardless of BP.⁵ This definition differs from the previous definition in 2008 in several ways. First, the target BP was lower than the prior definition in 2008, which proposed a SBP of at least 140 mmHg and DBP of at least 90 mmHg for most HTN patients and at least 130/80 mmHg only for those with diabetes mellitus (DM) or chronic kidney disease (CKD). Third, the use of CCB in combination with RAS blockade and a diuretic is now specified, whereas the 2008 definition requires only diuretic as a part of regimen.⁶ The recommendation is logical based on efficacy and safety in preventing cardiovascular disease (CVD) events in HTN patients of these 3 drug classes, which are now considered the first-line therapy for HTN.^7,8 In addition, combination therapy using these agents was shown to have additive effect on BP in a wide range of population of different race/ethnicity and socioeconomic status trials.^9–11 Similar to the previous recommendation, chlorthalidone is still the preferred diuretic over hydrochorothiazide for HTN treatment based on safety and efficacy.¹²

The global prevalence of RH is estimated to be approximately 14.7% among treated population.¹³ There was no significant difference in the RH prevalence among low vs. high income countries or regions of the world.¹³ Factors predisposing to RH include advanced age, obesity, DM, and CKD.¹⁴ Using the new BP targets, the prevalence of RH is modestly increased based on office BP criteria (Table 1). Patel et al. analyzed data from the nationally representative National Health and Nutrition Examination Survey (NHANES) between period of 2007 to 2014 and report prevalence of RH to be 15.95% in US adults with treated HTN which is increased from 12.0% based on the old definition by the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure guideline.¹⁵ Carey et al. performed analysis from NHANES from the period of 2009 to 2014 and found a similar result with the absolute net increase in the prevalence of RH of 2% (from 17.7 to 19.7% among participants taking anti-HTN medications).¹⁴ Prevalence of RH in the Action to Control Cardiovascular Risk in Diabetes (ACCORD)-BP trial, which is limited to DM population with HTN, was shown to be 14.4% according to the current definition as compared to prevalence of 10.2% using 140/90 mmHg as the target.¹⁶ Similarly, prevalence of RH is slightly higher (21.5% by new definition vs. 17.6% by prior definition) in the Systolic Blood Pressure Intervention Trial (SPRINT) trial which enrolled high risk population with and without CKD.¹⁷ The incidence of RH in the SPRINT and ACCORD by the updated definition in patients treated intensively, however, was considerably higher than the incidence in the standard arm that aimed to achieve SBP of <140 mmHg (30 cases vs. 10 cases per 100 patient-years).¹⁶ After enrolling in the study for 1 year, 22% of SPRINT participants and 36% ACCORD participants met the definition of RH.¹⁶ In contrast, the incidence of RH among UK population in the primary care clinic registry is substantially lower of 0.93 per 100 person years in 1996 and 2.07 cases per 100 person years in 2004. Since then, incidence is declining such that it is reported to be only 0.43 per 100 person years in 2015 for the reasons not yet entirely clear.¹⁸ Observational studies from the US cohorts have reported RH incidence (using prior definition of 140/90 mmHg) of 0.7 cases per 100 person-years of follow-up among the Kaiser Permanente health systems.¹⁹ The difference in the incidence and prevalence of RH observed in clinical trials vs. general population is likely to reflect the high CVD risks of patients participated in the SPRINT and ACCORD trials as well as lower target SBP goal of <120 mmHg. In addition, most of these epidemiological studies are limited by lack of data regarding the prevalence or incidence of RH in the context of treatment with specific combination drug therapy preferred by the guideline.

Table 1.

Prevalence of apparent RH in adults with treated hypertension according to the current vs. prior definitions.

Population	Time period	n	Prevalence (%)
			By 2008 Guidelines	By 2018 Guidelines
Observational
NHANES57	1988–1994	2755	9.4
NHANES57	1999–2004	3031	11.7
NHANES58	2003–2008	3710	12.8
NHANES57	2005–2008	2586	14.5
NHANES15	2007–2014	5512	12.0	15.9
NHANES14	2009–2014	4158	17.7	19.7
EURIKA59	2009–2010	5220	14.3
REGARDS60	2003–2007	14,731	14.1
De Nicola et al. (CKD)61	2003–2005	436	30.0	30.0
RIACE study (Diabetes)62	2007–2008	15,933	18.7	18.7
Yokohama et al. (Diabetes)63	2012–2013	1737	17.0	17.0
Jackson Heart Study64	2000–2013	524	8.2
Clinical trials
MASTERPLAN Study (CKD)65	2004–2010	788	34.1	34.1
SPRINT17	2015	9361	17.6	21.5
ACCORD-BP16	2010	4733		14.4
ALLHAT66	1994–2002	14,684	12.7
INVEST67	1997–2003	17,190	37.8

Nevertheless, a recent analysis from NHANES showed that the most common 3-drug regimen included ACEI, β-blocker, and thiazide diuretic.¹⁴ Only <20% of HTN patients taking ≥3 anti-HTN drugs were prescribed with the regimen of CCB plus ACEI (or ARB) plus thiazide diuretics.¹⁴ More importantly, only 3% of RH patients were prescribed chlorthalidone.¹⁴ Thus, prevalence of RH may be substantially lower if analysis is restricted to individuals who are prescribed only with the optimal therapy.

Refractory HTN (RFH) is a subset of RH that is defined as failure to achieve target BP despite 5 or more anti-HTN drug regimen which include a thiazide diuretic.²⁰ This subtype of HTN is rare with prevalence of approximately 10% of patients with RH referred to tertiary care center.²⁰ A recent analysis showed that prevalence of RFH was greater among ACCORD-BP than SPRINT participants. After enrolling in the trial for 1 year, 2.2% of SPRINT participants and 5.3% of ACCORD-BP participants met RFH.¹⁶

Pseudo-RH

The new AHA definition of RH also proposed that individuals with pseudo-resistance should not be considered as RH.⁵ Pseudo-RH could be related to white coat effect or nonadherence to medication.²¹ Meta-analysis of 91 epidemiological studies including >3.2 million HTN patients globally showed that pseudo-resistant form of HTN is as common as true-RH, each with the prevalence of approximately 10% among treated HTN population.¹³ Multiple epidemiological studies have demonstrated that white coat HTN (WCH), defined as isolated HTN in office, is associated with increased all-cause and CVD mortality in the population without anti-HTN drug treatment.^22,23 In contrast, uncontrolled WCH, defined as isolated elevation of BP in the office among the treated individuals, is not associated with increased CVD risk in multiple population studies.^22–25 Since RH by design is only limited to the treated population, WCH phenomenon is likely a benign phenotype in this population, though it has not been specifically studied. Based on the 2017 ACC/AHA guideline and the cohort studies, the cutoff for normal 24-hour ambulatory BP is now lower from 130/80 to 125/75 mmHg while the cutoff for normal daytime and nighttime ambulatory BP or home BP is reduced from 135/85 to 130/80 mmHg and 120/70 to 110/65 mmHg, respectively.^3,26,27 A large observational study from the Spanish Ambulatory Blood Pressure Monitoring Registry including >8000 patients with apparent RH showed that 37.5% of patients had WCH based on prior definition of elevated office BP and normal 24-hour ambulatory BP of at least 130/80 mmHg.²⁸ Only 62.5% of the patients had true RH. However, prevalence of true RH is higher if nighttime BP is used to determine presence of RH up to 75% using the prior cutoff of 120/70 mmHg.^29,30 Thus, out-of-office BP monitoring is essential in verifying presence of RH. Many experts and guidelines have proposed that automated office BP (AOBP) measurements without healthcare professionals present in the room, is an alternative to out-of-office BP monitoring by avoiding white coat effect and alerting reaction. AOBP, which is obtained using oscillometric BP devices that are programmed to have a delayed start, was shown to yield 5–20 mmHg lower BP readings than office BP obtained during manual auscultation by the healthcare personnel, approaching the levels achieved by ambulatory BP monitoring.³¹ However, more recent studies have not shown significant difference between unattended vs. attended AOBP by the healthcare professionals, particularly if automated devices are used instead of manual auscultation after a resting period before BP measurement.^32,33 Furthermore, nighttime BP, which is a stronger predictor of CVD outcome than daytime ambulatory BP and office BP,³⁴ is not captured by AOBP.

Nonadherence to medications in RH

Nonadherence to medications is another major cause of pseudo-RH which has been reported in 20–60% of apparent RH patients. Risk factors associated with nonadherence to anti-HTN medications include younger age, female sex, as well as elevated resting heart rate.^35–39 The number of anti-HTN medications prescribed is shown to predict medication nonadherence in some³⁵ but not all studies.^36,40 Numerous methods have been developed to assess medication nonadherence including detailed questionnaire, pill count, prescription fill rate, electronic pill-boxes, or direct observation. It is well known that patient’s self-report adherence is not reliable, as patients tend to overestimate their adherence to antihypertensive medications.⁴¹ Physicians’ ability to predict patients’ adherence to anti-HTN medication is also often poor.⁴² Direct observation is the most direct and accurate way in determining adherence but is not feasible in the outpatient setting. Electronic pillbox systems, such as the Medical Event Monitoring System, are considered to be highly accurate in assessing medication adherence by capturing both the frequency of medication usage via tracking of bottle cap openings.⁴³ However, the high cost associated with the bottles and the electronic software as well as lack of coverage by most healthcare systems limits its use in the clinical setting.⁴⁴ Patients may not bring the bottles to the clinic to allow data capture or actually take the medications after opening the bottles.⁴⁵ Pill count is laborious for the healthcare providers and mostly used in the research setting. It is found to have only modest accuracy between 50 and 70% in determining medication adherence when compared to electronic pill bottles^46,47 or measurement of drug levels.⁴⁸ More detailed assessment of self-report adherence via questionnaire such as the Morisky medication adherence scale and the Adherence to Refills and Medications Scale have not been shown to consistently separate out adherent from nonadherent patients.^40,49 Pharmacy refill data are another indicator of medication adherence but the data may not be accurate if the patients are not in the integrated health care system or do not take the dispensed medication.

More recently, increasing number of studies have indicated that biochemical measurement of drug levels in the serum or urine samples, using highly sensitive high-performance liquid chromatography-tandem mass spectrometry technique are highly reliable in detecting medication nonadherence in HTN.^{36,37,50–52} Testing for both urine and serum antihypertensive drug levels is available for clinical use and covered by most health insurance plans because serum/plasma levels of most anti-HTN medications is available in the US and Germany.^52,53 The advantage of this technique is the ease of use without interrupting clinic workflow as most patients are required to visit laboratory for other biochemical testing as part of routine care. Observational studies from hypertension referral practices and clinical trials showed that 30–60% of patients with presumed drug resistance were nonadherent to treatment to at least one drug^36,40,52 despite verbal report of medication adherence at the time of testing.

Furthermore, observational studies have shown that information from the measurement may help clinicians to uncover to underlying cause of nonadherence to specific drugs that were undetected in the plasma or urine. When the nonadherent patients were given therapeutic drug monitoring (TDM)-guided feedback and counseling to overcome barrier to adherence to the drugs that were undetected, BP improved substantially at subsequent visits despite no significant change in the anti-HTN regimen.^40,52 Although TDM is inevitably associated with extra cost of testing, TDM was shown to be cost effective in several independent analyses from Germany, England, and the US.^53–55

Conclusions

Patients with RH are known to have higher risk of HTN target organ damage and CVD events.^19,56 It is unknown if the BP level itself or other comorbidities as well as the underlying pathophysiology are responsible for the adverse CVD outcomes. A recent analysis from the SPRINT trial showed that patients with RH who were randomized to intensive BP lowering have better CVD outcomes than those randomized to standard arm, suggesting that uncontrolled HTN plays a large role.¹⁷ Screening for pseudo-resistance is the important first step in the evaluation of all patients with apparent RH as patients with WCH syndrome require no further intervention while those with medication nonadherence may benefit from TDM in facilitating BP control via identifying and targeting barriers to adherence.

Abbreviations and acronyms

ACC

American College of Cardiology

ACCORD

Action to Control Cardiovascular Risk in Diabetes

ACEI

angiotensin converting enzyme inhibitor

AHA

Ammerican Heart Association

AOBP

ambulatory out-of-office blood pressure

ARB

angiotensin receptor blocker

BP

blood pressure

CKD

chronic kidney disease

CVD

cardiovascular disease

DBP

diastolic blood pressure

DM

diabetes mellitus

HTN

hypertension

NHANES

National Health and Nutrition Examination Survey

RAS

Renin Angiotensin System

RFH

refractory hypertension

RH

resistant hypertension

SBP

systolic blood pressure

SPRINT

Systolic Blood Pressure Intervention Trial

TDM

therapeutic drug monitoring

US

United States

WCH

white coat hypertension