Introduction

Patients with resistant hypertension (RH) are at risk of poor outcomes. Adherence to medication schedules and treatment intensification improve blood pressure (BP) control; however, little is known about their effects on RH or their associations with outcomes in this patient group. A retrospective cohort study of two integrated health plans conducted between 2002 and 2006 reported that 1 in 50 patients with incidental hypertension developed RH.1 In addition, the patients who developed RH were found to be at increased risk of cardiovascular events; thus, greater effort should be made to improve hypertension outcomes in this population.1 The misdiagnosis of RH during hypertension treatment, for example, it being diagnosed as white coat hypertension or non-adherence to anti-hypertensive medication, sometimes occurs in primary care. Thus, we should refocus on what RH is and how it should be treated. Below, we review the recent findings regarding the pathogenesis and treatment of RH.

Avoidance of hasty diagnosis of RH

RH is sometimes diagnosed in patients with spurious resistant HT whose BP was inaccurately measured and/or evaluated. The protocol for measuring BP should always be strictly adhered to. It is essential to confirm whether or not the cuff size is appropriate, the patient is in a proper position, the arm is at a proper height and the patient has had sufficient rest prior to the measurement. Pseudohypertension, a BP reading that erroneously appears highly elevated, is incorrectly diagnosed as RH, especially in elderly patients, because progressive arteriosclerosis, common among the elderly, results in an increase in arterial stiffness, which makes it difficult to compress the arteries without great pressure during the measurement of BP. Thus, in order to make an accurate diagnosis of true RH, it is necessary to carefully consider all of the clinical information about each patient, including the physical characteristics and laboratory findings, as well as the methods used to measure BP.

It is well known that RH is defined as resistance to treatment when administered three or more different kinds of anti-hypertensive medications, including diuretics, to decrease BP. There is a white coat effect that could result in an overestimation of actual BP during treatment. It is generally accepted that the long-term prognosis of hypertensive patients showing a white coat effect may not be as severe, and with fewer critical cardiovascular complications, as that of patients who have true RH. Ambulatory BP monitoring (ABPM) is useful for evaluating diurnal changes of BP, such as normal or elevated asleep mean. We can differentiate white coat hypertension from persistent hypertension by ABPM, as described in detail below. We should measure BP by using ABPM for an accurate diagnosis of true RH.

It is also well known that uncontrolled hypertension, which is sometimes misdiagnosed as RH, exists in such cases showing non-adherence to lifestyle modification and anti-hypertensive therapy. Furthermore, behavioral medicine has a crucial role in the hypertension management of patients with spurious resistant HT; before a diagnosis of RH is made, individual patient education and empowerment, and non-adherence to medications and lifestyle modification should be improved. Another problem is ‘patient–physician adherence’. We need to focus on the various cognitive and behavioral factors or characteristics of each hypertensive patient that may have an influence on communication and treatment. We should always consider the possibility of spurious resistant HT among the patients with persistent and uncontrolled hypertension before making a definitive diagnosis of RH. A correct diagnosis of true RH can only be made after ensuring that BP was accurately measured, proper adherence to the patient–physician relationship and administration of anti-hypertensive medications, and differentiating white coat hypertension from severe persistent hypertension among patients by ABPM.

Definition and prevalence

We often encounter patients who display persistently high BP values despite taking anti-hypertensive drugs. Hypertensive patients whose BP does not return to normal after lifestyle modification and treatment with at least three different kinds of anti-hypertensive drugs, including diuretics, are diagnosed with RH and/or refractory hypertension. Comparative studies have suggested that uncontrolled RH should be referred to as refractory hypertension, and controlled RH should be termed resistant or difficult-to-control hypertension,2 although RH is essentially considered to be the same condition as refractory hypertension. The prevalence of RH was reported to be 13% in the Japanese J-HOME study and ranged from 30–50% in the following Western studies: ALLHAT,3 CONVINCE,4 LIFE,5 INSIGHT6 and VALUE.7 We need to establish a definitive definition of RH that is based on the findings of home BP studies and/or measurements obtained with ABPM and to clarify the physiological relevance and causative mechanisms of poor BP control under anti-hypertensive drug administration. Recent analyses of the data obtained during the National Health and Nutrition Examination Survey (NHANES), which involved subjects drawn from representative samples of the adult, non-institutionalized, civilian population, estimated the prevalence of RH to be 8.9±0.6% of the US hypertensive population in 2003–2008.8 A time-sequence comparison of the NHANES data from 1998 through 2008 suggested that, unlike hypertension, RH is becoming more prevalent (for example, it displayed an incidence of 20.7% from 2005 to 2008) due to the aging of the general population and the increasing frequency of obesity.8

Pathogenesis of RH

We sometimes encounter RH patients who are complicated with obesity, sleep apnea syndrome, primary aldosteronism (PA) or white coat syndrome as well as RH patients who have been prescribed inappropriate anti-hypertensive drugs or are not following the prescribed treatment schedule. In such cases, we ask the patients to modify their lifestyles, for example, to adopt a low-salt diet or exercise more, and attempt to improve their adherence to the prescribed treatment schedule. It is essential to refer such patients to specialist hypertension clinics, as RH patients are often at high risk of cardiovascular events and/or secondary hypertension.

RH can occur for the following reasons: (1) inadequate measurement of BP due to white coat syndrome, the use of an inappropriately sized cuff for determining BP or pseudohypertension; (2) insufficient treatment including inadequate adherence to the prescribed medication, insufficient lifestyle modification or the inappropriate use of drugs; (3) the existence of poorly controlled conditions including hypervolemia, obesity, sleep apnea syndrome, excess drinking or the consumption of BP-elevating chemicals or foods and (4) the misdiagnosis of secondary hypertension. It is particularly important to focus on the presence of hypervolemia, which is usually caused by high salt intake in the absence of diuretic treatment and/or the impairment of renal function.

A previous retrospective study recruited patients from two health systems, who had been diagnosed with incidental hypertension from 2002 to 2006, and then selected patients who developed RH or uncontrolled BP despite adherence to ⩾3 anti-hypertensive medications.9 They then examined the associations between medication adherence or treatment intensification and 1-year BP control in order to elucidate which patient characteristics affect BP control. As a result, they found that 49% of the RH patients were male (mean age: 60 years old). They also reported that at 1 year after diagnosis fewer of the RH patients were taking diuretics, β-blockers or angiotensin-converting enzyme inhibitors (ACE-I)/angiotensin receptor blockers (ARB) than at the baseline.9 Thus, we need to clarify the reason why patients with uncontrolled BP do not undergo treatment intensification.9 Moreover, in a study analyzing urine for various anti-hypertensive drugs and their metabolites, it was reported that poor adherence was the most common cause of poor BP control in patients with apparent RH, being twice as common as secondary hypertension.10 We therefore should be aware that poorly controlled hypertension under multidrug anti-hypertensive treatment does not always represent true resistance.

ABPM and arterial stiffness

Most cases of hypertension involve a white coat effect, which leads to the overestimation of BP. The prevalence and clinical characteristics of white coat or isolated office RH are usually determined by comparing clinic BP values with either daytime home BP measurements or awake BP mean values obtained from ABPM.11 Thus, so-called RH might include patients with normal or elevated asleep BP mean values.11 The authors of the latter study indicated that RH should be classified into isolated-office RH, masked RH and true RH, and also emphasized that ABPM should be regarded as a clinical requirement for the diagnosis of RH.11 Therefore, when assessing patients with suspected RH, we should measure BP using ABPM, rather than with the conventional cuff methods used to determine clinic BP in order to differentiate it from white coat, isolated office and masked hypertension. A previous study investigated the influence of the timing of hypertensive treatment on circadian BP patterns and the degree of BP control in order to clarify the relevant clinical and laboratory parameters of RH patients using 48-hour ABPM.12 These findings demonstrated that bedtime-based hypertension medication regimens are recommended for the treatment of patients who have been mistakenly judged to be resistant to therapy by conventional cuff methods or after treatment with a morning-based treatment regimen.12 It is suggested that we need to consider when patients should be administered anti-hypertensive drugs in order to achieve well-controlled BP levels. Such decisions should be taken on the basis of ABPM measurements of the diurnal changes in BP, which should be taken into account before definitely diagnosing RH.

It was reported that the variations in systolic BP, as estimated by ABPM, were positively correlated with carotid intima-media thickness, pulse wave velocity and the ambulatory arterial stiffness index (AASI), suggesting that ABPM, which is easier to perform at clinics than pulse wave velocity, and assessments of the variability in 24-hour BP might be useful for evaluating arterial stiffness.13 On the other hand, it was proposed that home BP variability on a single occasion was correlated with pulse wave velocity, independent of other known risk factors, in Japanese patients with type 2 diabetes.14 Another study investigated the correlation between circadian variability in BP/AASI and renal function.15 They demonstrated that hypertensive patients who displayed greater circadian variability in systolic BP as determined by ABPM exhibited significantly higher resistive index values. In addition, they found that daytime variability in systolic BP was more strongly correlated with resistive index than nighttime BP variability,15 suggesting that monitoring both daytime and nighttime BP variability might be useful for assessing pathological conditions in hypertensive patients and preventing cardiovascular disease in RH patients.

According to The Spanish Ambulatory Blood Pressure Monitoring Registry, 14 461 patients fulfilled the criteria for RH, whereas the BP of 13 436 hypertensive patients was successfully controlled using three or fewer anti-hypertensive drugs.16 The individuals with RH were older, more obese and had suffered from hypertension for a longer period than the controlled patients. They were also more likely to suffer from diabetes, dyslipidemia, reduced renal function, microalbuminuria or left ventricular hypertrophy and to display a history of cardiovascular events. In multivariate analysis, the duration of hypertension, obesity, abdominal obesity, left ventricular hypertrophy, a reduced estimated glomerular filtration rate and microalbuminuria were found to be independently associated with RH. In addition, ABPM revealed that white coat hypertension is common among RH patients, as is masked hypertension among patients with apparently controlled BP.16 Only the ABPM-based diagnosis of RH was reported to be an independent predictor of all-cause mortality.17 Higher ambulatory BP was reported to be a predictor of cardiovascular morbidity and mortality in RH patients, whereas office BP had no prognostic value.17 In another study, the prognostic value of nocturnal BP reduction in RH was examined in order to evaluate its importance as a predictor of cardiovascular morbidity and mortality.18 As a result, it was clearly demonstrated that the non-dipping pattern was an independent predictor of the composite end point (hazard ratio, 1.74; 95% confidence interval, 1.12–2.71) and cardiovascular mortality (hazard ratio, 2.31; 95% confidence interval, 1.09–4.92).18 It was also reported that ABPM provides a more reliable assessment of BP in patients with chronic kidney disease (CKD) and RH.19 Moreover, it is known that RH and the non-dipping pattern are valuable predictors of cardiovascular outcomes and that several electrocardiographic findings, such as prolonged ventricular repolarization, serial changes in the strain pattern and left ventricular hypertrophy, renal symptoms of albuminuria, and a reduced glomerular filtration rate are powerful cardiovascular risk markers in RH.20

CKD and RH

The Olivetti Heart Study (OHS), which was performed in southern Italy, aimed to clarify the predictors of RH in the general population21 and found that the incidence of RH during the 8-year follow-up period was 10.1% among hypertensive participants. They concluded that the urinary albumin/creatinine ratio, an elevated fractional excretion of sodium (FENa) value, increased cholesterol levels and a higher basal BP level were independent predictors of RH.21 Thus, it is suggested that microalbuminuria as an indicator of CKD and the degree of sodium retention, as measured by the FENa, might be strong predictors of future RH. It is strongly recommended that dietary sodium restriction be included in hypertension management strategies.22 In another study examining the relationship between obstructive sleep apnea (OSA) and RH in CKD and end-stage renal disease, it was reported that the mean systolic BP (SBP) values of the CKD and end-stage renal disease groups were significantly higher than that of the non-CKD group despite the fact that the former groups were administered more anti-hypertensive medications.23 In addition, the CKD and end-stage renal disease groups exhibited higher incidences of RH and more severe sleep apnea than the non-CKD group.23 Thus, we need to check for the presence of sleep apnea in RH.

Lifestyle modification for RH

It is speculated that unhealthy lifestyle factors might contribute to apparent treatment-RH.24 The latter study analyzed lifestyle factors, including obesity, physical inactivity, current smoking, heavy alcohol consumption, a low Dietary Approaches to Stop Hypertension (DASH) diet score and high sodium-to-potassium (Na/K) intake, in patients with treatment-RH and concluded that unhealthy lifestyle factors were not independently associated with apparent treatment-RH.24 On the other hand, obese patients often display treatment-resistant arterial hypertension, possibly due to neurohumoral activation.25 Moreover, it remains unclear how weight loss strategies and anti-hypertensive treatments should be combined to achieve optimal clinical outcomes.25 It is also known that visceral obesity increases the risk of arterial hypertension.26 The latter study analyzed 5065 hypertensive patients with visceral obesity and found that, after excluding undertreated patients, 13.9% had RH. Moreover, RH was more common in obese individuals with body mass indices of ⩾35 and morbidly obese individuals.26 These findings suggest that under treatment, the underuse of diuretics in multidrug regimens, and the white coat effect are the most common reasons for the overdiagnosis of RH in patients with visceral obesity and that obesity is an independent risk factor for RH.26 It was also reported that higher aortic stiffness is associated with reduced pulse pressure amplification with increasing adiposity in hypertensive African-American adolescents.27 Thus, we should pay attention to the existence of obesity when diagnosing RH.

The effect of physical exercise on hypertension, especially RH, is disputed. In a previous study, 50 subjects with RH were randomly assigned to participate or not participate in an 8–12-week treadmill exercise program, in which the target lactate level was 2.0±0.5 mmol l−1.28 As a result, it was found that physical exercise decreases BP, even in subjects who exhibit poor responses to medical treatment. Thus, physical exercise as well as measures aimed at reducing patients’ body weights should be included in therapeutic programs for RH.28

Mineralocorticoid receptor antagonist treatment for RH

Hypertension is a major risk factor for CKD, but although mineralocorticoid receptor antagonists (MRA) have been demonstrated to be effective at managing RH, they are not always used to treat CKD patients because they induce hyperkalemia. A retrospective study reported that 32 patients with RH associated with stage 3 CKD were treated with spironolactone and 4 were treated with eplerenone during a median follow-up period of 312 days.29 The MRA induced significant decreases in the patients’ SBP and diastolic BP, increased their serum potassium levels from 4.0±0.5 to 4.4±0.5 mEq l−1 (P=0.0001), with the highest value being 5.8 mEq l−1 and decreased their estimated glomerular filtration rates.29 MRA are quite effective at reducing BP in subjects with RH associated with stage 3 CKD, although close biochemical monitoring is recommended because of the increased risk of hyperkalemia and worsening renal function.29

It was previously reported that aldosterone levels are correlated with the severity of OSA in RH30, and that the blockade of aldosterone reduces the severity of OSA. Overall, these findings are consistent with excess aldosterone levels contributing to the worsening of underlying OSA. It was suggested that an excess of aldosterone worsens OSA by inducing fluid accumulation within the neck, which can cause upper airway resistance.30 They also reported that marked anti-hypertensive effects are observed when MRA are added to the treatment regimens of patients with RH, further suggesting that excess aldosterone levels are an important cause of RH. Furthermore, it has been suggested that a close association exists between hyperaldosteronism, RH and OSA.31

In a previous study, compared with the control group, the RH patients displayed significantly higher aldosterone-to-renin ratios and plasma aldosterone, 24-hour urinary aldosterone (UAldo), brain-type natriuretic peptide and atrial natriuretic peptide levels as well as significantly lower plasma renin activity and serum potassium levels.32 The detection of a significant correlation between 24-hour UAldo levels and cortisol excretion suggested that a common stimulus, such as corticotropin, might underlie the excess aldosterone levels observed in patients with RH.32 Thus, we need to consider the involvement of excess aldosterone and cortisol levels in the pathogenesis of RH. Excess aldosterone is also considered to have detrimental metabolic effects that contribute to metabolic syndrome and endothelial dysfunction, which in turn contribute to the development of RH as well as cardiovascular disease and CKD.33 Moreover, elevated aldosterone levels could have a crucial role in promoting insulin resistance and RH.34 Aldosterone enhances the tissue generation of oxygen-free radicals and systemic inflammation, resulting in impaired insulin metabolic signaling, reduced endothelial-mediated vasorelaxation, and associated cardiovascular and renal structural and functional abnormalities.34 It was also reported that the cardiovascular and renal abnormalities associated with insulin resistance are mediated, in part, by aldosterone-based non-genomic and genomic signaling through the mineralocorticoid receptor (MR).34 Furthermore, primary aldosteronism (PA), a clinical condition in which aldosterone is produced in excess and renin production is suppressed, is common in patients with RH, and recent studies have suggested that excess aldosterone might have a pathophysiological role in the relationship between hypertension and OSA. Obesity is common in patients that display OSA, excess aldosterone production and/or RH, and might link these conditions.35 On the other hand, the National Health and Nutrition Examination Survey (NHANES), which was conducted from 2005 to 2008, examined self-reported and study-measured health and sleep characteristics to test the associations between self-reported sleep symptoms and severe/RH in a nationally representative sample of adults.36 As a result, it was found that there was no significant association between self-reported snoring/snorting and RH,36 suggesting that sleep disturbance is not correlated with RH.

In a previous study, we found that 61 (6%) of 1020 hypertensive patients were diagnosed with PA from 1995 until 1999.37, 38 In addition, only 18% of the patients exhibited serum K levels of ⩽3.3 mEq l−1. Thus, many clinicians seem to misdiagnose PA as essential hypertension because of the absence of hypokalemia.37, 38 In a subsequent study, we reported that the incidence of PA is unexpectedly high among patients with hypertension and that aldosterone-producing adenoma, which seems to be a main cause of PA, can be treated with unilateral adrenalectomy.39 Among the 61 patients with definitely diagnosed PA in the above-mentioned study, 39% exhibited BP values of >180/110 mm Hg at outpatient clinics,37, 38 although their BP was not analyzed by ABPM. It was also reported that PA displays an incidence of 17–23% among patients with RH,40 suggesting that we should consider the possibility of PA and screen patients with RH for it in order to protect them from fatal cardiovascular events. Screening for PA is strongly recommended for patients with RH.41 Moreover, a study examining the prevalence of PA among RH patients reported that PA displayed a markedly higher prevalence among subjects with RH or stage III hypertension than among the general hypertensive population.42 Another report investigating the prevalence of PA among RH patients demonstrated that it was substantially lower than previously reported, although the RH patients did display a high prevalence of PA-based hyperaldosteronism.43

We prospectively examined the prevalence of endocrine hypertension among 1435 untreated hypertensive patients who visited Yokohama Rosai Hospital, Yokohama, and Social Insurance Central General Hospital, Tokyo. The prevalences of PA, Cushing’s syndrome including subclinical Cushing’s syndrome and Cushing’s disease, pheochromocytoma and renovascular hypertension were 7.2, 1.9, 0.3 and 0.5%, respectively. Unilateral adrenalectomy was performed in the 75 patients with unilateral hyperaldosteronism among the 104 patients with PA, 6 patients with Cushing’s syndrome, 5 patients with subclinical Cushing’s syndrome and 6 patients with pheochromocytoma. Five patients with Cushing’s disease were treated with transsphenoidal adenomectomy, and seven patients with renovascular hypertension were treated with percutaneous transluminal angioplasty. One year after the surgery/percutaneous transluminal angioplasty, the hypertension of 74 patients (71%) had been completely cured, and that of 30 patients (29%) had been improved under treatment with low doses of anti-hypertensive agents. The prevalence of endocrine hypertension, such as adrenal and renovascular hypertension, was 10.0% among the 1435 untreated hypertensive patients, and the hypertension cure rate was 71% among the 104 patients who underwent surgery or percutaneous transluminal angioplasty. These findings clearly demonstrated a high prevalence rate of endocrine hypertension among hypertensive patients, suggesting that we should appropriately screen hypertensive patients in order to differentiate endocrine hypertension from essential hypertension. Secondary endocrine hypertension is strongly associated with RH and induces cardiovascular events. We should avoid misdiagnosing secondary endocrine hypertension to prevent RH.

Spironolactone has been demonstrated to have an anti-hypertensive effect in patients with ABPM-diagnosed RH.44 In the latter study, 175 patients were subjected to clinical and complementary examinations at the baseline, before being given spironolactone at doses of 25–100 mg per day. As a result, mean reductions of 16 and 9 mm Hg, respectively, were observed in the patients’ 24-hour SBP and diastolic BP (DBP). Their office SBP and DBP also decreased (by 14 and 7 mm Hg, respectively).44 Accordingly, the authors concluded that the administration of spironolactone to patients with true RH is safe and effective at decreasing their BP, especially in patients with abdominal obesity and lower arterial stiffness.44 Spironolactone is also used for cases of PA and RH that do not involve secondary etiologies. It was reported that the dose ranges for these two conditions overlap significantly.45 Thus, we cannot differentiate PA from essential RH according to the BP-lowering effect of spironolactone.

It was also reported that the pathogenesis of MR-associated hypertension in patients with normal plasma aldosterone levels is mediated by MR activation via pathways other than high aldosterone levels, such as increased MR levels, increased MR sensitivity and MR overstimulation by factors such as Rac1.46 For RH patients with high plasma aldosterone levels, MRA should be given as a first-line therapy, whereas for those with normal aldosterone levels, ARB or ACE-I should be given as a first-line therapy, and MRA can be given as an additional treatment.46 In Japan, both the selective MRA eplerenone and the non-selective MR antagonist spironolactone are generally used for the treatment of hypertension.47 Sato suggested that the roles of spironolactone and eplerenone in the treatment of hypertension should be carefully differentiated, and we agree with this point.47

In a study investigating the role of cortisol in the pathogenesis of RH, patients with RH were screened for the presence of subclinical hypercortisolism using the 1 mg dexamethasone overnight suppression test, and hypercortisolism was confirmed when two saliva samples collected at 2300 h displayed cortisol levels of at least 3.6 nmol l−1. One hundred and twelve patients (prevalence: 26.5%) tested positive for hypercortisolism without overt Cushing’s syndrome. The patients who produced positive screening results were older, more likely to be males and displayed higher prevalence of diabetes and target organ damage and higher nighttime SBP values than the patients who produced negative screening results.48 As a result, the authors concluded that patients with RH display a relatively high prevalence of subclinical hypercortisolism, which is associated with several markers of worse cardiovascular prognosis.48 In addition, increased circulating levels of glucocorticoids are observed in patients with lifestyle-related disorders, which can also activate MR signaling in cardiovascular, adipose, skeletal muscle, neuronal and liver tissue.34

Renal artery denervation as a treatment for RH

Treatments for RH involving carotid baroreceptor activation or renal sympathetic denervation have recently been developed. On the other hand, it was reported that while these techniques resulted in marked falls in office BP, the associated ambulatory BP reductions were rather modest. The disparity between the office and ambulatory BP reductions might be explained by sympathetic overactivity.49 Renal artery denervation50, 51 was also reported to reduce SBP and DBP in patients with drug-RH and to reduce atrial fibrillation recurrence when combined with pulmonary vein isolation.52

Conclusion

Clinicians should be aware that RH patients are at risk of poor outcomes. We do not know the exact prevalence of RH among patients with treated hypertension, although it was reported that the prevalence of RH was ∼13% in the Japanese J-HOME study and ranged from 30–50% in Western countries. The latter figure appears to be too high and might have been affected by the inclusion of patients with uncontrolled hypertension or clinically suspected secondary hypertension. We need to perform further clinical trials for clarifying the exact prevalence of RH, because Carey53 has recently described that the prevalence of RH is still a matter of debate. Endocrine hypertension is sometimes present among RH patients. We rule out secondary hypertension during hypertensive patients’ first visits, that is, before administering any medications, because secondary hypertension including endocrine hypertension can easily evolve into RH. RH patients often exhibit salt retention, which can be caused by excess aldosterone production or increased renin–angiotensin system activity due to overactivity of the sympathetic nervous system. We should clarify the pathophysiological mechanisms of the salt retention and sympathetic nervous system activation observed in RH. Furthermore, RH can be treated using aldosterone antagonists or by blockading the renin–angiotensin system, as well as device-based therapy involving renal denervation and carotid baroreceptor activation. We summarize the optimal method for diagnosing and treating RH in Tables 1 and 2. The speed at which Japan’s population, which shows a preference for eating salty foods, is aging is accelerating. Thus, we need to prevent lifestyle-related diseases, including hypertension, especially among aged individuals, as the prevalence of RH-associated complications remains high.

Table 1 Diagnosis of resistant hypertension and its complications
Full size table
Table 2 Treatment of resistant hypertension
Full size table

Finally, we need to focus on the existence of RH among treated hypertensive patients in order to avoid the early onset of fatal cardiovascular events and decrease medical costs.