Non-compliance

A common cause of resistant hypertension is non-compliance. Patients may not take their medication because of cost, adverse effects or a system of health beliefs, such as that drugs are harmful, or a result of a conspiracy between the pharmaceutical industry and physicians. A particularly difficult problem is the patient who mistakenly attributes any and all symptoms to adverse effects of drugs.7 A blinded N of one crossover study with capsules containing active drug in one bottle and placebo in the other may be helpful in sorting this out, but it may be fruitless in patients whose mind is made up. If asked in a non-judgemental way (eg, how often do you forget to take your medication?), about half of patients will admit to non-compliance.8 Advice about the truly causal adverse effects of medications, as opposed to the long list of incidental adverse effects reported on the long lists in which symptoms are equally common on placebo versus active drug, may help the patient to resume medication. True adverse effects such as cough with ACE inhibitors can be avoided by switching to other medications such as an angiotensin receptor blocker (ARB). Persistence is better with medications that have fewer adverse effects, in particular with ARBs.9

Substances that patients consume that aggravate hypertension

A high salt intake has historically been a particular problem in China. Government efforts and changes in food preservation related to increased availability of refrigeration have made some progress, but survival free of hypertension is still significantly related to the ratio of sodium intake to potassium intake. Most sodium intake is from added condiments and processed foods, and the intake of processed foods has been increasing.10

In a recent US study,11 the combination of a low sodium intake and a diet rich in fruits, vegetables and low-fat dairy products, and reduced in saturated fat and cholesterol (the Dietary Approaches to Stop Hypertension diet), markedly reduced blood pressure, and the effects were greater in patients with higher blood pressures. In patients with systolic pressure >150 mm Hg, the reduction of systolic blood pressure with this combination was 20.7 mm Hg, and with diastolic pressure >90 mm Hg, the reduction was 7.9 mm Hg (figure 2). Diet for stroke prevention is discussed in another article in this issue of the journal; essentially the recommended diet is based on the Cretan Mediterranean diet.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

The combined effects of low sodium and the DASH diet according to baseline blood pressure. The mean end of period SBP measurements (mm Hg) by strata of (A) baseline SBP (<130, 130–139, 140–149, >150 mm Hg) or (B) baseline DBP (<80, 80–84, 85–89, >90 mm Hg). The mean blood pressure values are presented by the high-sodium control diet (circle) or the low-sodium DASH diet (diamond). Differences between diets were determined using linear regression comparing baseline changes in systolic or diastolic blood pressure adjusted for age, female sex, black race and baseline body mass index. Bars indicate 95% CIs. Ctrl, control; DASH, Dietary Approaches to Stop Hypertension trial; DBP, diastolic blood pressure; NA, sodium; SBP, systolic blood pressure. (Reproduced with permission from Elsevier: Juraschek et al.11)

Patients with hypertension should be encouraged to reduce their salt intake towards 2–3 g per day, avoiding added salt and also limiting intake of foods containing high amounts of sodium. Sparing use of light soya sauce, more spices, herbs, ginger, lemon juice, vinegar and hot sauce can help reduce salt intake. Excess alcohol, licorice, birth control pills and non-steroidal anti-inflammatory agents (NSAIDs) may also account for resistant hypertension.

Licorice has mineralocorticoid effects causing hypertension and hypokalaemia,12 and is found in candy, tea and some tonics. Oral contraceptives cause hypertension; there may be several mechanisms. In a rat model combined hormonal therapy induced endothelial dysfunction and upregulation of the ACE-1 gene and angiotensin receptor gene13; however, combined hormonal therapy in postmenopausal women raises the level of aldosterone and lowers the level of plasma renin.14

The only NSAID that does not raise blood pressure is sulindac15; it is highly protein-bound, so very little is filtered into the renal tubule; furthermore, it is a prodrug that is oxidised in the renal tubule back to the inactive form, which does not inhibit formation of vasodilator prostaglandins.15–17

Heavy alcohol use (>3 drinks per day) doubles the risk of hypertension18 and increases the risk of non-lobar intracerebral haemorrhage.19

Therapeutic inertia

Therapeutic inertia has many sources. Busy physicians may focus on other health problems of patients with hypertension, such as diabetes, arthritis, gastrointestinal complaints, headache and so on. When a patient presents with complaints in other areas, the hypertension is not the focus of the clinic visit and may be overlooked. Furthermore, patients and physicians may be reluctant to increase the antihypertensive therapy, in the mistaken belief that white coat hypertension is benign. It is not.20 However, therapeutic inertia can be overcome, essentially by measuring blood pressure at every visit, and a policy that antihypertensive therapy will be intensified whenever target blood pressures are not achieved. In the North American Symptomatic Carotid Endarterectomy Trial,21 we essentially overcame therapeutic inertia by sending a stiff letter to the site investigator every time the medication was not increased, when a patient’s blood pressure exceeded the level specified in the protocol. This resulted in a remarkable reduction of intracranial haemorrhage to only 0.5% of strokes, at a time when 20% of strokes were due to intracerebral haemorrhage.

Diagnostic inertia

When a patient does not respond to usual antihypertensive therapy, it is crucial to consider whether the therapy is appropriate for that patient and to investigate the underlying physiology that is driving the hypertension. Failure to consider reasons why patients’ blood pressure is not controlled may be termed ‘Diagnostic inertia’.22

Among patients with resistant hypertension, an approach that is efficacious is what I call ‘Physiologically Individualized Therapy’ (PhysRx), based on phenotyping by plasma renin activity and aldosterone. This is a form of personalised medicine. An early variant of this was recommended by Laragh in the early 1970s, based on plasma renin activity.23 The results described above (a 50% reduction of stroke in our community)4 were the result of a programme of improved detection and treatment of hypertension mounted by the Department of Family Medicine, in combination with my hypertension clinic, basing therapy on stimulated plasma renin levels. Patients with a high salt intake (very common in China) have low plasma renin levels if unstimulated. By giving a dose of furosemide (0.5 mg/kg) before drawing blood for measurement of plasma renin activity (PRA), the range of plasma levels is spread out, with normal and high PRA levels moving up so that low and suppressed plasma renin can be appreciated.24 I gave oral furosemide to patients living locally 4 hours before blood draws; for those coming from several hours away, we gave intravenous furosemide 30 min before the blood draw. In patients who are already taking a diuretic, ACE inhibitor or ARB, it may not be necessary to give furosemide, as these drugs all stimulate plasma renin activity.o

Egan and colleagues25 carried out a randomised trial of therapy based on plasma renin activity, but did not measure plasma aldosterone. In 77 patients randomised to usual care versus renin-based therapy, they reported lower systolic pressures, a trend to improved blood pressure control and a greater reduction of medication needed among patients with volume hypertension (ie, low renin hypertension). Measuring PRA is useful, but it is important to measure stimulated plasma aldosterone. Once rare causes of hypertension are eliminated (such as pheochromocytoma, adult coarctation of the aorta, and licorice ingestion), there are essentially three types of hypertension, all understood by studying the renin/angiotensin/aldosterone axis (figure 3). (1) Renal hypertension: when the kidney senses low blood pressure it releases renin, activating the formation of angiotensin and in turn release of aldosterone. Patients with renal hypertension, renal cysts, obstruction, stones, renovascular hypertension and parenchymal renal disease including hypertensive nephrosclerosis (which might be thought of as microvascular renovascular hypertension) have a high PRA and secondary hyperaldosteronism. (2) Adrenocortical hypertension: when there is hyperplasia of the adrenal cortex with overproduction of aldosterone, salt and water retention suppresses PRA, so plasma aldosterone is high and PRA is low. (3) Liddle phenotype: when there is a mutation of the renal tubular epithelial sodium channel (ENac, SCNBB1), or a mutation of genes that affect the function of ENaC, there is salt and water retention, which suppresses both PRA and aldosterone. As shown in table 2, the treatment is different for these three types of hypertension. Renal hypertension, with high renin and secondary aldosteronism, is best treated with an angiotensin receptor blocker (ARB). Primary aldosteronism, with low renin and high aldosterone, is best treated with an aldosterone antagonist. Spironolactone can be used for women, but spironolactone in adequate doses for primary hyperaldosteronism usually causes gynecomastia and mastalgia in men, so eplerenone is preferred. Abnormalities of the ENac or variants that affect the function of ENac (low renin/low aldosterone; Liddle phenotype) are specifically treated with amiloride. There are several variants of Liddle syndrome (mutations of ENaC; SCNBB1) and variants of several other genes that affect the function of ENaC (GRK, NEDD4L, CYP4A11, NPPA, UMOD), resulting in Liddle phenotype.26 27

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

The physiology of therapy for resistant hypertension. When the kidney senses low blood pressure, renin is released, leading to production of angiotensin II. In turn, angiotensin II activates production of aldosterone by the adrenal cortex, which leads to salt and water retention. Under healthy conditions, this system is turned off once normal body water levels are restored. Impairment of this system can cause hypertension. Renal hypertension is associated with high renin and secondary hyperaldosteronism; in primary hyperaldosteronism the salt and water retention suppresses renin production, so aldosterone levels are high with low renin levels. Abnormalities of the renal tubular epithelial sodium channel cause salt and water retention, suppressing both renin and aldosterone. To identify the best primary therapy for patients with resistant hypertension, the physiological mechanism underlying the blood pressure elevation must be identified (table 1). (Reproduced with permission from Nature: Spence29).

A clinical trial in Africa provides evidence that physiologically individualised therapy for resistant hypertension as described above markedly improved blood pressure control. Patients in three hypertension clinics, in Nigeria, Kenya and South Africa, whose blood pressure was not controlled to <140 mm Hg systolic or <90 mm Hg diastolic were allocated to usual care versus physiologically individualised therapy based on the algorithm described in table 2. In Kenya this approach had no effect; patients attended clinic less frequently, compliance was less and amiloride was not available. ‘When only the sites in Nigeria and South Africa were considered, systolic control was obtained in 15.0% of UC vs. 78.6% of PhysRx (P<0.0001), diastolic control in 45.0% vs 71.4% (P=0.04), and control of both in 15.0% vs. 66.7% (P=0.0001). If only the Nigerian site (where patients were randomized to the two treatment strategies) is considered, systolic control was obtained in 15% of UC vs 85% of PhysRx (P=0.0001), and diastolic control in 45% vs 75% (P=0.11)’.26 In this trial, the biggest difference in medication between the usual care arm and the PhysRx arm was in the prescription of amiloride: 19%of patients allocated to PhysRx vs 2.8% on usual care (P=0.02). A randomised substudy of this approach is planned in China.