Maria Czarina Acelajado, MD, Department of Medicine, University of the Philippines-Philippine General Hospital, Taft Avenue, Ermita, Manila 1000, Philippines E-mail:firstname.lastname@example.org
J Clin Hypertens (Greenwich).
Among patients with resistant hypertension (RHTN), there are those whose blood pressure (BP) remains uncontrolled in spite of maximal medical therapy. This retrospective analysis aims to characterize these patients with refractory hypertension. Refractory hypertension was defined as BP that remained uncontrolled after ≥3 visits to a hypertension clinic within a minimum 6-month follow-up period. Of the 304 patients referred for RHTN, 29 (9.5%) remained refractory to treatment. Patients with refractory hypertension and those with controlled RHTN had similar aldosterone levels and plasma renin activity (PRA). Patients with refractory hypertension had higher baseline BP (175±23/97±15 mm Hg vs 158±25/89±15 mm Hg; P=.001/.005) and heart rate, and higher rates of prior stroke and congestive heart failure. During follow-up, the BP of patients with refractory hypertension remained uncontrolled (168.4±14.8/93.8±17.7 mm Hg) in spite of use of an average of 6 antihypertensive medications, while those of patients with controlled RHTN decreased to 129.3±11.2/77.6±10.8 mm Hg. Spironolactone reduced the BP by 12.9±17.8/6.6±13.7 mm Hg in patients with refractory hypertension and by 24.1±16.7/9.2±12.0 mm Hg in patients with controlled RHTN. In patients with RHTN, approximately 10% remain refractory to treatment. Similar aldosterone and PRA levels and a diminished response to spironolactone suggest that aldosterone excess does not explain the treatment failure.
Resistant hypertension (RHTN) is defined as high blood pressure (BP) that requires ≥4 medications for treatment.1 As defined, RHTN includes patients whose BP is controlled or uncontrolled after use of ≥4 medications. While the exact prevalence of RHTN is unknown, cross-sectional studies suggest that it includes approximately 10% to 15% of the general hypertensive population.2 Patients with RHTN are at increased cardiovascular risk compared with patients with more easily controlled hypertension based on an increased likelihood of having other cardiovascular risk factors such as left ventricular hypertrophy, chronic kidney disease (CKD), and obstructive sleep apnea, and on outcome data demonstrating an increased rate of cardiovascular complications.3–5
Patients with RHTN include a wide range of demographic and physiologic characteristics such as age, body weight, race/ethnicity, degree of target organ damage, presence of secondary causes of hypertension, severity and duration of hypertension, and lifestyle factors. These differences undoubtedly reflect multiple underlying etiologies that contribute to the development of treatment resistance. With intensive treatment, particularly if delivered by a hypertension expert, the available data suggest that the majority of patients with RHTN will eventually achieve BP control. Retrospective studies of success rates within hypertension specialty clinics suggest that 50% to 70% of patients referred for RHTN can be controlled with lifestyle changes and use of effective combination therapies.6 More recently, with broader use of mineralocorticoid receptor antagonists, which seem to be specifically beneficial in the treatment of RHTN, control rates are likely even better.
Based on our anecdotal experience, we have come to believe that the small proportion of patients with RHTN who never achieve BP control in spite of maximum medical therapy (ie, refractory hypertension) represent a potentially unique phenotype, distinct from the broader group of patients with RHTN whose BP can be controlled. Identifying and characterizing such patients represents the first step in defining the mechanisms of treatment resistance in this extreme phenotype. Subsequent steps would include targeted studies of the pathophysiology of refractory hypertension, which, in turn, should facilitate development of more effective diagnostic and therapeutic options for this potentially high-risk group.
The purpose of this study was to identify and characterize patients with refractory hypertension that was defined as failure to achieve BP control in spite of maximum medical therapy. In order to estimate prevalence and provide insight into potentially unique causes of their extreme treatment resistance, patients with refractory hypertension were compared with patients with RHTN in whom BP control was achieved.
The study consisted of a retrospective analysis of patients referred to the University of Alabama at Birmingham (UAB) Hypertension Clinic for evaluation and treatment of RHTN. The analysis included an 8-year period (2000–2008). This study was approved by the UAB institutional review board and was conducted according to institutional guidelines.
A patient was defined as having RHTN if BP remained above goal (BP >140/90 mm Hg) in spite of being adherent to an antihypertensive regimen that consisted of ≥3 medications from different classes. A diuretic was required to be part of the regimen unless contraindicated due to prior adverse effects or intolerances and all agents had to be prescribed at effective dosages. Adherence was determined by patient self-report. Analyzed patients were required to have at least 3 clinic visits during a minimum follow-up period of 6 months. A patient was considered to be refractory to treatment if he or she never achieved BP goal. The reference group consisted of patients referred for RHTN whose BP was controlled for 2 consecutive clinic visits during follow-up (ie, controlled RHTN). Patients who underwent procedures for management of secondary causes of hypertension (eg, angioplasty for renal artery stenosis or adrenalectomy for adrenal adenoma) were also considered to have refractory hypertension if the BP remained uncontrolled after the intervention and in spite of maximum medical therapy.
Per our routine evaluation of patients referred for RHTN, assessed variables included medical history (duration of hypertension, comorbidities, number and class of antihypertensive medications, and home or workplace BP), demographics (age, sex, and race), clinical data (clinic BP, heart rate, height, weight, and body mass index [BMI]), and biochemical parameters (serum potassium and creatinine; plasma aldosterone; plasma renin activity [PRA]; and 24-hour urinary sodium, protein, aldosterone, and cortisol excretion). The 24-hour urine collection was performed while the patients were consuming their usual diet. If the sodium excretion was <150 mmol/24 hours, patients underwent oral salt loading for 3 days (either by administration of sodium tablets or by increasing their dietary sodium intake), and the 24-hour urine collection was then repeated. Increased sodium intake was confirmed by a sodium excretion of >200 mmol/24 hours during the repeat 24-hour urine collection.
BP in the clinic was measured by a trained physician after at least 5 minutes of rest, using the auscultatory method, in a seated patient with the arm properly supported and using the correct cuff size. BP was measured in both arms and the average of 2 readings in the arm with the higher BP reading was used. Patients with suspected white coat hypertension (controlled home, work place, or ambulatory BP measurements) were excluded from analysis. Beyond hyperaldosteronism and hypercortisolism, which we routinely test for by measurement of 24-urinary excretion, other secondary causes of hypertension such as renal artery stenosis and pheochromocytoma were screened for if clinically indicated based on medical history, physical findings, and/or laboratory results. Patients who presented with symptoms of obstructive sleep apnea (such as snoring, witnessed apneas, and excessive daytime sleepiness) were referred for polysomnography.
The referred patients were seen by 2 hypertension specialists. During every visit, the patient’s clinic and out-of-office BPs were reviewed, and the treatment regimen was revised if the BP remained above goal. The generalized treatment approach included adding or titrating diuretic therapy, changing the diuretic class to one appropriate for the patient’s kidney function, using medications with complementary mechanisms of action, and adding a mineralocorticoid antagonist to the antihypertensive drug regimen. Thiazide diuretics, usually chlorthalidone at 25 mg/d, were prescribed preferentially for most patients. If the patient’s kidney function precluded the use of a thiazide diuretic, a loop diuretic was prescribed, most commonly furosemide at 20 mg to 40 mg twice daily. Vasodilators, centrally acting antihypertensive agents, and α-adrenergic blockers were added to the regimen after a renin-angiotensin blocker + diuretic + calcium channel blocker + mineralocorticoid receptor antagonist with or without a β-blocker failed to control BP. Adherence was assessed by asking the patient about medication use, perceptions about medication efficacy, and presence of adverse effects, if any. Patients were seen every 4 to 8 weeks, with more frequent visits for patients with uncontrolled BP.
Values are expressed as mean±standard deviation (SD). Baseline variables for patients with refractory hypertension and controlled RHTN were analyzed by 2-sample t test or Wilcoxon signed-rank, where appropriate. A statistically significant level was set at .05.
A total of 387 patients referred for RHTN were screened, of whom 304 met inclusion criteria and were included in the analysis. Eighty-three patients were excluded because of nonadherence, white coat hypertension, or inadequate follow-up. Of the 304 analyzed patients, 29 (9.5%) never achieved BP control and were diagnosed with refractory hypertension. These patients were classified as having refractory hypertension after an average of 6.7 visits during a mean follow-up of 11 months. In contrast, BP goal was achieved in patients with controlled RHTN after an average of 3.8 visits during a mean follow-up of 3.5 months.
Table I, Table II, and Table III list, respectively, the baseline demographic, clinical, and biochemical data of all patients included in the analysis. Patients with refractory hypertension were more likely to be women and African American, but neither difference achieved statistical significance. Patients with refractory hypertension had a shorter duration of hypertension compared with those with controlled RHTN (mean duration 14 vs 18 years). The 2 groups were similar in terms of age, BMI, renal function, plasma and urinary aldosterone levels, PRA, and 24-hour urinary sodium excretion. Patients with refractory hypertension had higher baseline systolic and diastolic BP levels and higher baseline heart rates in spite of taking more antihypertensive medications. Patients with refractory hypertension had higher rates of prior congestive heart failure and stroke compared with those with controlled RHTN, but they were less likely to have been diagnosed with obstructive sleep apnea. A higher percentage of patients with refractory hypertension had CKD (P=.14), but the proportion of patients with stage 4 or 5 CKD in both groups was similar (28% in the refractory group, 31% in the controlled RHTN group). The majority of patients with CKD in both groups had stage 3 disease. Likewise, there was a higher percentage of patients with diabetes in the refractory hypertension group compared with the controlled RHTN group, but the difference was not statistically significant.
Table I. Baseline Demographic Characteristics of Patients With Refractory and Controlled Resistant Hypertension
Refractory Hypertension (n=29)
Controlled RHTN (n=275)
Abbreviations: BMI, body mass index; BP, blood pressure; RHTN, resistant hypertension. Values are expressed as means or number (percentage). P values denote refractory hypertension group compared with the controlled RHTN group.
Systolic , mm Hg
Diastolic BP, mm Hg
Heart rate, beats per min
Antihypertensive drugs, No.
Table II. Baseline Comorbidities in Patients With Refractory and Controlled Resistant Hypertension
Refractory Hypertension (n=29)
Controlled RHTN (n=275)
Abbreviation: RHTN, resistant hypertension. Values are expressed as number (percentage). P values denote refractory hypertension group compared with the controlled RHTN group.
Obstructive sleep apnea
Chronic kidney disease
Coronary artery disease
Congestive heart failure
Table III. Baseline Biochemical Characteristics in Patients With Refractory and Controlled Resistant Hypertension
Refractory Hypertension (n=29)
Controlled RHTN (n=275)
Abbreviations: PRA, plasma renin activity; RHTN, resistant hypertension. Values are expressed as means. P values denote refractory hypertension group compared with the controlled RHTN group.
Plasma aldosterone, pmol/L
24-Hour urine aldosterone, nmol/d
24-Hour urine sodium, mmol/d
At the time of their last visit, the mean systolic and diastolic BP for patients with refractory hypertension were 168.4±14.8 mm Hg and 93.8±17.7 mm Hg, respectively, reflecting a decrease in BP of 6.2/3.2 mm Hg while under our care. The BP remained uncontrolled in spite of taking an average of 6.0±0.8 antihypertensive medications (range, 5–8). In contrast, the patients with controlled RHTN had a mean follow-up BP of 129.3±11.2/77.6±10.8 mm Hg, with an average decrease in BP from baseline of 28.6/10.9 mm Hg. This was accomplished with use of 4.2±1.1 antihypertensive agents (range, 1–7). Resting heart rate remained elevated in patients with refractory hypertension compared with patients with controlled RHTN (81 vs 70 beats per minute, P<.001) in spite of greater use of β-blockers (79.3% vs 60.0%). The intensified antihypertensive regimen caused a rise in serum creatinine levels in some patients, but the elevation did not exceed 20% of the baseline value. Progression to overt kidney failure was not seen in this cohort.
Antihypertensive Drug Use
Patients with refractory hypertension were being treated with more classes of antihypertensive medications on follow-up compared with those with controlled RHTN (6 vs 4). The antihypertensive medications prescribed for both patient groups during follow-up are listed in Table IV. Almost all patients in both groups were receiving a diuretic (most commonly chlorthalidone 25 mg/d or hydrochlorothiazide 25–50 mg/d; and furosemide 20 mg to 40 mg twice daily for patients with estimated glomerular filtration rate [eGFR] by Modification of Diet in Renal Disease [MDRD] formula <30 mL/min). If a diuretic was not being used, it was because of documented adverse effects or because of patient refusal (usually because of perceived intolerances). In the refractory group, all patients, with one exception, were being treated with chlorthalidone or furosemide. The exception was a patient intolerant of both thiazide and loop diuretics. As would be anticipated, patients with refractory hypertension received more of all classes of antihypertensive agents. Twenty-four of the 29 (82.8%) patients with refractory hypertension were treated with spironolactone at a mean dose of 46.4 mg/d, compared with 144 of the 275 (52.4%) of patients with controlled RHTN who received spironolactone at a mean dose of 29.1 mg/d. In all cases, spironolactone was added to the existing thiazide diuretic. Spironolactone was not prescribed to 4 patients with refractory hypertension because of advanced CKD (eGFR by MDRD formula <30 mL/min) and in the remaining patient because of a prior life-threatening episode of diuretic-induced hyponatremia.
Table IV. Antihypertensive Medications Prescribed After ≥6 Months Follow-Up in Patients With Refractory and Controlled Resistant Hypertension
Refractory Hypertension (n=29)
Controlled RHTN (n=275)
Abbreviations: ACE, angiotensin converting enzyme; ARB, angiotensin receptor blocker; RHTN, resistant hypertension. Values are expressed as number (percentage). P values denote refractory hypertension group compared with the controlled RHTN group.
Mean spironolactone dose, mg/d
ACE inhibitor and/or ARB
Calcium channel blocker
Vasodilators and/or centrally acting agents
Response to Spironolactone
In patients with refractory hypertension, spironolactone reduced the mean BP from 178.0±23.2/100.4±12.9 mm Hg to 165.1±12.3/93.8±13.8 mm Hg (P value=.002 and .027, respectively), corresponding to a mean reduction of 12.9/6.6 mm Hg (Figure). In patients with controlled RHTN, spironolactone reduced the mean BP from 153.4±16.9/86.8±13.6 mm Hg to 129.3±11.2/77.6±10.8 mm Hg (P value <.001 and <.001, respectively), resulting in a mean decrease of 24.1/9.2 mm Hg.
We propose to define refractory hypertension as BP that remains uncontrolled in spite of maximal medical therapy. We appreciate that what is regarded as maximal medical therapy will vary on a patient-by-patient basis, depending on drug indications for associated comorbidities, response to prior medications, history of adverse events and intolerances, and risk of interaction with concurrent medications. In the current analysis, a patient’s hypertension was considered refractory if it could not be controlled after being treated in a hypertension specialty clinic for a minimum of 6 months. Based on this definition, the prevalence of refractory hypertension was found in approximately 10% of patients originally referred to our clinic for RHTN. This implies that within this patient population, we were successful at getting the BP to goal in about 90% of patients, representing an improvement in control rates compared with an earlier analysis, which had reported control rates of 50% to 60% in a specialty clinic.6 While there are major methodologic differences between this earlier study and the current analysis, we speculate that the improvement in control rates is related, at least in part, to the increased use of mineralocorticoid receptor antagonists for treatment of RHTN.
Historically, the terms refractory hypertension and RHTN have been used interchangeably to indicate difficult-to-treat hypertension.7,8 The results of the current analysis indicate some potentially important differences and similarities between patients with refractory vs controlled RHTN. The refractory patients presented with more severe hypertension and a history of more frequent cardiovascular complications, including stroke and CHF. The patients with refractory hypertension also had higher resting heart rates that persisted even after titration of treatment, including increased use of β-blockers. This persistently higher heart rate may suggest that greater sympathetic output may be contributing to antihypertensive treatment failure. If confirmed by prospective mechanistic studies, these findings may point to a need for development of more effective methods to block sympathetic outflow in order to achieve better BP control in these patients who are otherwise undergoing failing antihypertensive therapy.
Important similarities between patients with refractory hypertension vs controlled RHTN include similar ages, BMIs, and aldosterone levels. We and others have found that hyperaldosteronism, whether classical primary aldosteronism or relative aldosterone excess, contributes importantly to the development of RHTN.9–13 The current analysis suggests that refractory hypertension is not characterized by more severe aldosterone excess as indicated by normally measured plasma and urinary aldosterone levels. This lack of difference in circulating levels of aldosterone, however, does not exclude greater mineralocorticoid receptor activation due to tissue differences in effective levels of aldosterone or other potential mineralocorticoid receptor activators such as cortisol. Absence of a greater aldosterone effect is, however, also suggested by the diminished response to spironolactone in patients with refractory hypertension. In spite of use of higher doses of spironolactone, the BP response was considerably less in patients with refractory hypertension compared with the controlled RHTN patients, indicating that the refractory patients experience failed therapy with spironolactone, as with other classes of agents. These combined observations suggest that aldosterone excess is not the primary cause of treatment failure in this group of patients. This finding, if confirmed, does not support use of increasingly higher doses of spironolactone in an effort to control BP in these patients.
RHTN is broadly characterized by excess intravascular fluid retention attributable to multiple factors, including obesity, aldosterone excess, high dietary salt intake, and CKD.9,14,15 In the current study, we did not find evidence of greater fluid retention in refractory hypertension vs controlled RHTN patients. Neither aldosterone nor PRA levels were suppressed in patients with refractory hypertension compared with patients with controlled RHTN, as would have been anticipated in the setting of greater fluid retention. In the current analysis, the lack of suppression of PRA argues against greater fluid retention as a cause of treatment failure. If confirmed in prospective analyses, this would have important clinical implications, as it would indicate that intensifying diuretic therapy, as is generally recommended for treatment of RHTN, might not be helpful in controlling BP in patients with refractory hypertension.
A lesser dependence on volume suggests a greater role of increased cardiac output and/or vascular resistance as the cause of treatment failure. This could be due to heightened sympathetic drive and/or increased peripheral resistance secondary to local or circulating pressor agents. In the current analysis, persistently higher heart rates in the refractory patients compared with patients with controlled RHTN provides preliminary support for the former possibility. Better assessment of sympathetic tone is needed to test this potential mechanism of refractory hypertension.
The current study is limited by its retrospective design, presumption of treatment adherence based on patient self-report, and the lack of ambulatory BP monitoring results to systematically exclude patients with white coat RHTN. We routinely assess patient adherence by interview both initially and at follow-up. It is our impression, as has been reported by others,6 that adherence is generally high in patients seen in a referral clinic, such that the confounding effects of nonadherence on the current analysis should have been minimal. Also per routine, we encourage all patients, if they have not already, to purchase a digital monitor to measure home BP. After instructing the patients on good technique, we ask them to keep a home BP diary, which we then rely on to guide therapeutic decisions. Use of these home BP data was used to exclude patients from the current analysis who were suspected of having a pronounced white coat effect. In an ongoing prospective analysis, these deficiencies will be avoided by more rigorous assessments of adherence and systematic screening for white coat hypertension with use of ambulatory BP monitoring.
The current analysis indicated that approximately 10% of patients referred to a hypertension specialty clinic for RHTN had refractory hypertension, defined as hypertension that is uncontrolled in spite of maximum medical therapy. Patients with refractory hypertension presented with more severe BP elevations and had evidence of heightened sympathetic output based on higher heart rates compared with controlled RHTN patients. However, refractory patients and those with controlled RHTN had similar aldosterone/PRA levels and refractory patients had attenuated antihypertensive responses to spironolactone, suggesting that increased sympathetic nervous system activity, rather than aldosterone excess, may be an important contributor to their failure to respond to antihypertensive treatment.
The current analysis suggests that refractory hypertension is not characterized by excess aldosteronism or greater fluid retention compared with controlled RHTN. If confirmed, this would have important implications for clinical management of such patients in that continued titration of diuretic therapy and/or mineralocorticoid receptor antagonists may not be appropriate and might even be counterproductive by triggering reflexive pressor responses and/or increasing likelihood of adverse events.
Disclosure: This study was supported by National Heart, Lung, and Blood Institute grant number RO1-HL79040 (DAC), SCCOR P50 HL077100 (DAC, LJD), T32 HL007457 (RP, SO), and UAB Center for Clinical and Translational Sciences 5UL1 RR025777. The authors have no conflicts of interest to disclose.