The fact that elevated blood pressure carries an increased risk of ischemic heart disease, congestive heart failure, stroke, chronic renal failure and death has been appreciated for decades.^1,2 Development of thiazide diuretics as effective drugs for treatment of hypertension in the 1950s and 1960s and the results of clinical trials, done initially in patients with moderate and severe hypertension in the 1960s and 1970s, led to a rapid increase in the treatment of hypertension and the development and testing of new classes of drugs. An estimated 50 million Americans have hypertension.

Proper technique in the measurement of BP cannot be overemphasized, since the studies on which management recommendations are based all use standardized methods. Aneroid sphygmomanometers are less accurate and more difficult to calibrate than mercury sphygmomanometers and the latter should be used when available. Some causes of office blood pressure (BP) values that do not represent the patient’s typical range include smoking, "white coat hypertension," and oral agonists including NSAIDS, recreational drugs, herbals that include ephedra (e.g. MaHuang), high doses of alcohol, and over-the-counter or self-prescribed beta adrenergic medicines (especially cold and asthma medicines and diet pills).

Common errors in blood pressure measurement include use of a cuff that is too small for the patient's arm circumference, failure to support the patient’s arm, failure to relax the patient, failure to ask about caffeine ingestion or tobacco use within the previous 30 minutes, failure to seat the patient, provide support for the patient's back, or bare the patient’s arm, and failure to give the patient five minutes of quiet rest prior to taking the BP.^3, All of these errors will give BP measurements that are higher than the true BP. Sitting with legs crossed can raise the BP by 5 mmHg. Cuff size artifacts may be as great as 15 mmHg. The bladder should nearly (> 80 percent) or completely encircle the arm. Other common errors are digit preference (i.e. 150 vs. 148 or 152) and rounding down. Two or more readings separated by two minutes should be averaged. Additional readings should be taken if the first two differ by more than 5 mmHg. Hypertension should not be diagnosed on the basis of a single measurement. Initial elevated readings should be confirmed on at least two separate visits over several weeks (unless SBP > 210 mmHg and/or DBP > 120 mmHg).

Issues of particular relevance in the history include history and symptoms of cardiovascular disease (CVD), smoking, alcohol intake, and major stress at home or at work. Physical examination should focus on evidence of hypertensive retinopathy, cardiac hypertrophy, arteriosclerosis (pulse diminution), thyroid nodules or enlargement, renal or other arterial bruits, edema, and focal neurological signs. Hypertensive retinopathy or evidence of left ventricular hypertrophy (LVH) indicate that hypertensive damage is present in the vascular system and that the patient is at increased risk for additional target organ damage and morbid events. These patients should be treated.

Tests that are indicated as part of the baseline assessment of the hypertensive patient include CBC (to rule out erythrocytosis), urinalysis (UA), chem 8 (to examine sodium, potassium, creatinine, glucose, and calcium), fasting lipid profile, and EKG. If cells or casts are present on UA, genitourinary disease should be excluded: these findings cannot be explained solely by hypertension. Significant proteinuria, while possibly a consequence of hypertensive renal damage, is an indication for evaluation for glomerulonephritis. Chest x-ray and echocardiography are not routinely indicated, despite the greater sensitivity of echocardiography to detect LVH compared to the surface EKG. Routine twenty-four hour BP monitoring is not recommended despite greater ability than routine BP to predict target organ damage.⁶ Investigation for pheochromocytoma is indicated if the history suggests episodic hypertension with headache, tachycardia or palpitations, or sweating.

Investigation for renovascular hypertension is indicated in the presence of a renal bruit and in selected patients with resistent hypertension, as defined below. The gold standard test for anatomic renal artery stenosis (RAS) is renal angiography. There is controversy regarding the best non-invasive imaging strategy.⁷ The best test for physiological renovascular hypertension is the captopril renal scan.⁸ In unilateral RAS, a single dose of captopril with subsequent ACE inhibition leads to impairment of renal function on the affected side, which is dependent on angiotensin II to maintain filtration pressure, with consequent unilateral late excretion of tracer. Patients must not be taking ACE inhibitors or A_II blockers at the time of the test. The main anticipated benefits of successful diagnosis and definitive surgical treatment of RVH, usually by angioplasty and stenting, are preservation of remaining function in the affected kidney; protection of the contralateral kidney and the vascular system from increased BP and from the possible effects of elevated renin; and elimination or reduction in drugs needed for BP control. These benefits have not been quantified in controlled trials and remain controversial.⁹

Lifestyle modifications have been shown in clinical trials to lower BP. Diet modification to increase consumption of fruits and vegetables and reduce intake of total and saturated fat (even in the absence of reduction in salt intake),¹⁰ reduction in salt intake,¹¹ regular aerobic exercise,¹² weight loss, and reduction of excessive alcohol intake (to two ounces a day or less) have all been shown to be effective. In the DASH-Sodium trial, sodium reduction decreased blood pressure by 7.0/3.8 mmHg in those older than 45 years of age and by 3.7/1.5 mmHg in those < 45.¹¹ Smoking raises BP transiently (15-30 minutes). Smoking cessation may be associated with weight gain and thereby have adverse effects on BP, but should be strongly encouraged because of positive effects on overall cardiovascular risk. There is a habituation response to caffeine, so that reduction or elimination of coffee has little effect on hypertension. Potassium¹³ and calcium supplementation may each slightly reduce blood pressure.

Drug treatment. There are presently six classes of antihypertensive drugs available in the US: thiazide diuretics, β-blockers, α-blockers, ACE inhibitors (ACEIs), A_II blockers (ARBs), and calcium channel blockers. The Treatment of Mild Hypertension Study (TOMHS)¹⁴ compared lifestyle modification alone (plus placebo) vs. lifestyle modification plus monotherapy with one of five different medications (acebutol, amlodipine, chlorthalidone, doxasine and enalapril) in patients with baseline DBP 90-99 mmHg. Mean BP reductions were significantly greater with all five medications plus lifestyle modification compared with lifestyle modification alone, and the BP reductions at 48 months did not differ significantly among the five medication regimens. All drugs were well tolerated. Quality of life was improved in all drug treatment groups, suggesting that even though the symptomatology is diffuse, hypertensives are often symptomatic. Consensus recommendations published in 1997 and due for revision are to use a thiazide diuretic or β-blocker for initial monotherapy.³ A thiazide is preferred over a β-blocker in the elderly (see below). Exceptions include diabetics and other subgoups discussed below. At least three lines of evidence support the recommendation of thiazides or β-blockers: clinical trial data documenting effects of these treatments on cardiovascular endpoints, head-to-head comparisons showing that their effects on blood pressure are approximately equivalent to that of other drug classes, and lower cost. In the ALLHAT Trial, which compared treatment strategies based on four medication classes (diuretic, β-blocker, calcium channel blocker, and the α-blocker doxasin), the doxasin arm was terminated because of a 25 percent excess in CVD events in the doxasin arm compared to the thiazide diuretic arm.¹⁵ Alpha blockers are not recommended as first-line treatment.

More recently published trials support the use of drugs that block or inhibit the renin-angiotensin system. The HOPE Trial randomized high risk patients, defined as > 55 years of age and having a history of cardiovascular disease or diabetes and at least one other major cardiovascular risk factor (but without heart failure) to the ACEI ramipril or placebo.¹⁶ All of the major endpoints were reduced in the treatment group and the authors concluded that "Ramipril significantly reduces the rates of death, myocardial infarction, and stroke in a broad range of high-risk patients who are not known to have low ejection fraction or heart failure.” The LIFE Study compared treatment based on the ARB losartan to treatment based on the β-blocker atenolol. Hypertensives with EKG-diagnosed LVH received escalating doses to 100 mg/day of losartan or atenolol, with addition of a thiazide diuretic and other medications as needed to achieve treatment target. Virtually identical mean blood pressures were obtained in the two groups, but the losartan group had significantly greater reduction over four years in the primary endpoint of death, MI, or stroke (23.8 vs. 27.9 per 1,000 patient-years; RR 0.87; 95% CI 0.77-0.98; p = 0.02).¹⁷ Losartan-treated patients also had greater reduction in LVH, lower incidence of diabetes (25%), and lower rate of adverse events. These findings suggest that this class of drug reduces cardiovascular risk through mechanisms in addition to blood pressure reduction, and that the total cardiovascular risk reduction is greater than that obtained by treatment with β-blockers. A subgroup analysis from the LIFE Study focused on hypertensives with LVH who also had diabetes found a similar benefit of ARBs.¹⁸ These findings support the use of ARBs and possibly ACEIs as initial monotherapy.

Diuretics are effective and safest at low doses (hydrochlorothiazide 12.5 to 25 mg QD or 50 mg QOD or the equivalent), reflecting a therapeutic dose-response curve with a low onset. Toxicity, primarily hypokalemia and hypomagnesemia, is also dose-related, but generally occurs at higher doses. While these features of diuretic use have been appreciated for many years, a well-conducted case-control study found an increased risk of sudden death among hypertensives treated with higher doses, with an odds ratio of 1.7 (95% CI 0.7-4.5) for hydrochlorothiazide 50 mg/day and an odds ratio of 3.6 (95% CI 1.2-10.8) for 100 mg/day, compared to 25 mg/day.¹⁹ Use of a combination of low dose diuretic and potassium-sparing drug (e.g. dyazide) was associated with a reduced risk. This study also found that the risk of sudden death was lower for patients treated with low dose thiazide diuretics compared to β-blockers, consistent with the finding in the Medical Research Council study²⁰ that coronary heart disease mortality was lower in hypertensives treated with a combination thiazide and potassium-sparing drug compared to those treated with a β-blocker. Thus, thiazides properly used can be effective and safe. For patients who become potassium-depleted, a combination drug with a potassium-sparing agent is preferable to potassium supplements for reasons of adherence. Some experts recommend using potassium-sparing agents in all patients not on other drugs, and who do not have other diseases (e.g., diabetes or renal failure), that inhibit potassium excretion. Many experts prefer spironolactone (aldactone) as the diuretic of choice. Low doses (25mg/day) are effective but may take six to eight weeks for maximum effect. Gynecomastia is rare at doses below 50 mg/day.

Calcium channel blockers have been associated in several case control studies with increased CVD morbidity and mortality.²¹ Short acting calcium channel blockers, which were never approved by the FDA for hypertension treatment, may have deleterious effects related to rapid onset and trough level offset resulting in blood pressure cycling. Close reading of the epidemiological data suggests that the apparently increased risk of CVD outcomes in patients treated with long-acting calcium channel blockers may be explained by "confounding by indication," that is, patients selected for treatment with calcium channel blockers may have been at higher risk because of relative contraindications to other classes of antihypertensives. In particular, diabetic hypertensives in these studies were less likely to be treated with thiazide diuretics or β-blockers and more likely to be treated with calcium channel blockers than nondiabetic hypertensives.²² A recent trial of a long-acting calcium channel blocker in hypertensives age > 60 years with SBP 160-219 and DBP < 95 mmHg found significant reductions at median two years of follow-up in all-cause mortality, CVD mortality, CVD events, and stroke, compared to a placebo control group. Reductions were present in the those both with and without diabetes analyzed as subgroups.²³ This study did not address the relative effectiveness or safety of this class of drugs compared to other classes. Two recent meta-analyses came to differing conclusions about the effects on cardiovascular risk of calcium channel blockers compared to other classes of drugs.^24,25 Current consensus³ (as well as local opinion) is that long-acting calcium channel blockers are effective, useful, and acceptable in terms of safety, but ought to be used after efforts at blood pressure control with other classes of drugs. In most patients the dose response curve for hypertension is such that maximum benefit is achieved at moderate doses (Procardia XL [nifedipine] 60 mg/day or Norvasc [amlodipine] 10 mg/day). Published data do not support use of higher doses.

Combination Drug Treatment. Data from several national surveys show that overall BP control rates are quite low and that most controlled hypertensives are on more than one drug. This is especially the case for patients with proteinuric renal failure, obesity, diabetes, and significant systolic hypertension. Combination treatment has two important pharmacodynamic advantages. First, many combinations have synergistic effects (e.g., diuretics with β-blockers, ACE inhibitors, or ARBs). Secondly, combination therapy allows use of moderate doses of each drug with fewer side effects.

The goal of treatment is to lower the risk of morbid sequelae (CVD, renal failure and CHF) at the lowest financial cost and with the least inconvenience and side effects for the patient. The question of how far to lower the BP has been controversial. The so-called J-curve describes a putative relationship between BP and mortality in which those with the lowest BP have increased mortality compared to those with ideal blood pressure at the bottom of the downslope. Two reviews have concluded that while the J-shaped BP-mortality relationship exists, it does not reflect a cause-and-effect relation between BP reduction and mortality but an artifact arising from the fall in BP due to CVD or CHF in hypertensives.^26,27 The Hypertension Optimal Treatment (HOT) trial randomized 18,790 patients from 26 countries to target DBP of 90 vs. 85 vs. 80 mmHg with average follow-up time of 3.8 years. Rates of CVD events, CVD mortality, and all-cause mortality were lowest in the groups assigned to target DBP of 80 mmHg. The nadir of risk for CVD mortality was at a mean achieved DBP of 86.5 mmHg, and for major CVD events at 82.6 mmHg. The study conclusions were that the data demonstrate the benefits of lowering blood pressure to 140 mmHg systolic and 85 mmHg diastolic or lower. Efforts to lower blood pressure to 120 mmHg systolic and 70 mmHg diastolic, conferred little further benefit but did not cause any significant risk.²⁸

A recent synthesis²⁹ recommended a treatment goal of less than 140/90 mmHg, preferably lowering DBP to 85 mmHg. A treatment goal of 130/80 mmHg is recommended in patients with diabetes^30,31 and possibly in some other high risk groups.²⁹ We recommend agreeing with the patient on a specific blood pressure goal, documenting the goal in the chart, and considering home monitoring by the patient.

Treatment monitoring. Patients should be seen within one to two months of initiating or modifying treatment to assess adequacy of BP control, degree of adherence, and presence of side effects.³ Once stabilized, follow-up at three to six month intervals is appropriate. Potassium and magnesium levels should be checked periodically in patients on diuretics; potassium levels should also be monitored in patients on ACE inhibitors. A 24 hour urine collection may be useful in assessing sodium and potassium intake.

Many experienced clinicians encourage hypertensive patients to buy a home BP monitoring device, measure and record their own BP on a regular basis, and, in some patients, self-regulate their regimen to maintain control. This approach is appealing for several reasons: it is patient-centered, patient-empowering, and may enhance adherence; it uses more data points than can be obtained in the doctor’s office; and it avoids "white coat hypertension" and consequent overtreatment. Home BP devices use aneroid (rather than mercury column) sphygmomanometers and consequently are more difficult to calibrate. Patients can learn to auscultate their own BP using a stethoscope, but most home BP devices use an automated sensor. These devices vary in reliability and accuracy³² and are not subject to any FDA standard. One recent review recommends several Omron devices including the Omron HEM 737 Intellisense.³² Patients who use these devices require training and calibration of their technique to an accurate mercury sphygmomanometer. Coin-operated BP devices measure a single BP in an atypical situation and are highly inaccurate and unreliable. Devices that measure BP in the index finger are inaccurate. The Visiting Nurse Service can also perform home BP monitoring.

We and others^3,32 believe that home BP monitoring has a useful role in the treatment of hypertension. It is possible that BP readings obtained at home under resting conditions more closely approximate BP readings obtained in clinical studies under protocol conditions than do casual office BP.

Mild hypertension is a term previously used for DBP 90-99 and SBP < 160 mmHg, corresponding to Stage 1 hypertension in current terminology (Table 1). The condition is clinically important because of the much greater prevalence of Stage 1 than Stage 2 or Stage 3 hypertension. The British Medical Research Council (MRC) treatment trial of mild hypertension randomized 17,354 hypertensives aged 35 to 60 with initial DBP 90-109 and SBP < 200 mmHg, almost all of whom were followed in general practices, to a thiazide, propranolol or placebo. Stroke was significantly reduced by treatment (1.4 vs 2.6 per 100 patient-years), and significantly more by thiazide than propranolol. No differences were found in incidence of coronary artery disease events or overall mortality. Based on these data, treatment of 850 mild hypertensives for one year would prevent one stroke.²⁰ The true NNT may be smaller if treatment is directed to mild hypertensives who have had multiple repeat elevated BP measurements despite attempts at lifestyle modification. A meta-analysis of 14 randomized placebo-controlled clinical trials of antihypertensive treatment included analysis of the five studies (including the MRC study, which was larger than the other four combined) with entry DBP < 110 mmHg. In pooled analysis of these five trials of mild hypertension, in which the usual aim was to reduce DBP to 90 mmHg or less, stroke was significantly reduced, while CHD and vascular mortality were reduced with marginal statistical significance. Reductions were statistically significant when studies of moderate or severe hypertensives were included.³³ These findings support recommendations to initiate treatment in Stage 1.

Systolic pressure, diastolic pressure, and pulse pressure. For 40 years, treatment trials and treatment goals were based on diastolic blood pressure. Data from the Framingham Study showing that systolic blood pressure was a stronger predictor of CVD events than diastolic blood pressure were known, but clinical trial data to support action directed at SBP reduction were not available until relatively recently. The Systolic Hypertension in the Elderly (SHEP) Study,³⁴ discussed below, was the first major trial to address this issue directly. Recently the pulse pressure (difference between systolic and diastolic pressure) has attracted attention as a marker of cardiovascular risk, particularly in older patients in whom most CVD events occur. Studies have found an association of increased pulse pressure with myocardial infarction³⁵ and congestive heart failure³⁶ as well as other CVD endpoints. Pulse pressure increases with age due to progressive arterial stiffening and loss of elastic capacitance/diastolic recoil in the aorta, producing both higher systolic and lower diastolic pressures. Deleterious hemodynamic effects of higher systolic pressure include increased afterload, myocardial wall stress, and oxygen demand, and subendocardial iscehmia, impaired relaxation, and hypertrophy. Lower diastolic pressure may impair myocardial perfusion and contribute to subendocardial ischemia. Antihypertensive drugs that increase the pulse pressure may therefore have undesirable pathophysiological properties in older patients.

Hypertension in the elderly. An overview summarizing data in 16,164 patients over the age of 60 years in ten randomized trials found that diuretics were 50 percent more effective than β-blockers in lowering blood pressure.³⁷ Moreover, all-cause mortality was reduced by 14 percent and myocardial infarction by 26 percent in the diuretic-treated hypertensives, compared to no reductions in those treated with β-blockers. JNC VI no longer recommends β-blockers as first-line treatment in older people. The SHEP Study also supports the choice of diuretics rather than β-blockers in this group.³⁴ Several pathophysiological mechanisms have been invoked to explain these observations. “The hemodynamic profile in the elderly is characterized by low cardiac output, low heart rate, and high peripheral resistance, with all [these] conditions being exacerbated by β-blockers. Also because β-adrenergic responsiveness is diminished in the elderly, higher doses of the β-blocker would be required."³⁸ A synthesis of data from 1,670 hypertensive patients over age 80 in eight randomized trials that included patients in this age group showed benefit from blood pressure reduction.³⁹

Isolated systolic hypertension (ISH) refers to a SBP > 160 mmHg and DBP < 90 mmHg. ISH is quite prevalent, affecting as many as 9 percent of people aged 65 and older.⁴⁰ The Systolic Hypertension in the Elderly Program (SHEP)³⁴ randomized 4,736 people aged 60 and above (mean age 72) with SBP 160 to 219 (mean 170) mmHg and DBP < 90 mmHg to treatment with chlorthalidone 12.5-25 mg/day (step 1) and atenolol 25-50 mg/day vs. placebo control. This study documented a 36 percent reduction in stroke incidence (5.2 vs. 8.2 per 100 people at 4.5 years of follow-up; p <0.001) and a 32 percent reduction in incidence of total major fatal and nonfatal cardiovascular events (12.2 vs. 17.5 per 100 people at 4.5 years of follow-up). These data support drug treatment of patients with ISH, with an expected individual benefit of 5.3 cardiovascular events including 3.0 strokes prevented over 4.5 years for every 100 people treated. Subgroup analysis supported benefit for both those with and without baseline EKG abnormalities. These data support treatment of elderly patients with ISH. In treating elderly patients with ISH, care should be taken not to reduce the diastolic pressure below 65 mmHg.²⁸

Borderline isolated systolic hypertension (ISH) refers to a SBP 140 to 159 mmHg and DBP < 90 mmHg. After age 50, diastolic BP tends to fall even though systolic BP continues to rise with age. By age 60, most clinical decision-making in hypertension management is based on the systolic BP. The decision whether to use drug treatment for patients with borderline ISH still not clearly defined. Observational data in subjects with borderline ISH show increased CVD incidence and CVD mortality.⁴⁰ Patients with borderline ISH should have BP followed yearly and may be counseled regarding lifestyle changes that may lower BP. Trial data support treatment of patients with combined diastolic and borderline systolic hypertension but do not directly address treatment of borderline ISH. Many experts recommend treatment of borderline ISH in special circumstances including diabetes, proteinuric renal disease, congestive heart failure, or aortic aneurysm.

Hypertension in patients with diabetes should be tightly controlled. The benefit (absolute risk reduction or number of CVD events prevented) obtained by treating hypertension is greater in patients at higher baseline risk, and diabetes approximately doubles the baseline risk of CVD. The UK Prospective Diabetes Study had a substudy of hypertensives with type 2 diabetes (N=1,148). This trial provides strong data showing reductions in both macrovascular and microvascular events in the group randomized to a policy of tight blood pressure control (BP goal < 150/85 mmHg) compared to the group assigned to less tight control (BP goal <180/105 mmHg). For example, there were comparative reductions of 18 percent in all-cause mortality and 44 percent in stroke at median 8.4 years of follow-up.⁴¹ This study design also provided a randomized comparison of the efficacy of an ACE inhibitor (captopril) vs. β-blocker (atenelol). Captopril and atenolol were equally effective in reducing blood pressure and in reducing macrovascular and microvascular endpoints. For example, 31 percent in the captopril goup and 37 percent in the atenolol group developed clinical grade albuminuria (>300 mg/L), and the proportions with hypoglycemic attacks were also not different.⁴² This study provided no evidence favoring differential benefits of ACE inhibitors in hypertensives with diabetes. Other data have led expert opinion in the U.S. to favor ACE inhibitors and possibly A_II blockers in patients with diabetes and micro-albuminuria.⁴³

Each drug class poses special issues when used in patients with diabetes. Hyperkalemia is a concern when using ACE inhibitors in diabetics with nephropathy because of reduced renin secretion, consequent reduction in aldosterone levels, reduced ability to excrete potassium, and a failure of the increased potassium level to properly stimulate aldosterone production ("low-renin low-aldo state"). β-blockers blunt the adrenergic-mediated responses to hypoglycemia. Central adrenergic inhibitors and peripheral alpha blockers may potentiate orthostatic hypotension and sexual dysfunction. Thiazide diuretics may increase glucose intolerance and slightly increase serum lipid levels. Diuretics are often necessary, especially in obese and poorly controlled diabetics, who tend to retain salt and water (insulin resistance and hyperinsulinemia promote sodium retention),⁴⁴ since hypertension is generally difficult to control in salt and fluid overloaded patients. The insulin resistance and hyperinsulinemia that characterize obese patients with type 2 diabetes also increase sympathetic tone, thereby contributing to hypertension.⁴⁴ Thus, weight reduction, as difficult as it may be to achieve, may be of substantial benefit to such patients.

Hypertension in women has been the subject of controversy.⁴⁵ In the MRC study²⁰ 48 percent of subjects were women, and in the SHEP study³⁴ 70 percent were women, so data are not lacking. The controversy arises because treatment appears to lower CVD risk by a smaller amount in women than in men, and in the MRC study treated women had an increase in all-cause mortality. Subgroup analysis produces instability of estimates of risk reduction due to smaller numbers of subjects in the groups; women have fewer endpoints than men, further diminishing statistical precision. On the whole, the available data and JNC recommendations support treating women using the same classification scheme as men. Individualization of treatment implies awareness of gender-specific characteristics that influence treatment decisions. Many women using oral contraceptives experience a small but detectable rise in BP, although most of these data come from women on higher doses of estrogen and progesterone than are presently used. If hypertension develops, the oral contraceptive may be stopped or treatment initiated for the hypertension depending largely on patient preference. Postmenopausal estrogen replacement may safely be prescribed for hypertensive women and may even lower BP. A few women experience a rise in BP, and BP monitoring is recommended for women on either oral contraceptives or estrogen replacement.

Resistant hypertension, defined as hypertension that does not respond to at least three drugs, one of which is a diuretic, has a number of causes. In this situation, it is good practice to assess the patient for secondary causes (e.g., RVH or endocrine hypertension) or nonadherence, rather than simply adding another drug. The most common cause is nonadherence. Volume overload, due to high salt intake, lack of concomitant diuretic therapy, or renal insufficiency should be considered. True resistant or refractory hypertension, not explained by causes such as these, is an indication for evaluation for renovascular hypertension.

Adherence is a central issue in the management of hypertension.⁴⁶ This is because hypertension is asymptomatic in most patients and remains so for 10 to 20 years until end organ damage becomes clinically significant, while treatment is a daily task, the drugs are often costly, and side effects are sometimes a problem. Labeling patients as hypertensive often has negative psychological effects, and anger, denial or both may be used as defenses. Nonadherence follows naturally in this psychological milieu. It is common to think of nonadherence in the context of resistant hypertension, but it is better to think of promoting compliance as essential to the task of managing hypertension from the day of first diagnosis. Creating a positive, nonjudgmental psychological milieu and partnership approach to management that nurtures the patient's own motivations and goal-oriented behavior will improve the chances of achieving treatment goals and satisfaction.

33. Collins R, Peto R, MacMahon S, et al. Blood pressure, stroke and coronary heart disease. Part 2, short-term reduction in blood pressure: overview of randomized drug trials in their epidemiological context. Lancet 1990;335:827-38.
34. SHEP Cooperative Research Group. Prevention of stroke by antihypertensive drug treatment in older persons with isolated systolic hypertension: final results of the Systolic Hypertension in the Elderly Program (SHEP). JAMA 1991;265:3255-64.
35. Madhavan S, Ooi WL, Cohen H, Alderman MH. Relation of pulse pressure and blood pressure reduction to the incidence of myocardial infarction. Hypertension 1994;23:395-401.
36. Chae CU, Pfeffer MA, Glynn RJ, Mitchell GF, Taylor JO, Hennekens CH. Increased pulse pressure and risk of heart failure in the elderly. JAMA 1999;281:634-639.
37. Messerli FH, Grossman E, Goldbouri U. Are β-blockers efficacious as first-line therapy for hypertension in the elderly? A systematic review. JAMA 1998;279:1903-1907.
38. Messerli FH. β-blockers as first-line therapy for hypertension in the elderly. Am J Managed Care 1998;4(12 Suppl):S757-S764.
39. Gueyffier F, Bulpitt C, Boissel JP, et al. Antihypertensive drugs in very old people: a subgroup meta-analysis of randomized trials. Lancet 1999;353:793-796.
40. Sagie A, Larson MG, Levy D. The natural history of borderline isolated systolic hypertension. New Engl J Med 1993;329:1912-17.
41. UK Prospective Diabetes Study Group. Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKDPS 38. BMJ 1998;317:703-713.
42. UK Prospective Diabetes Study Group. Efficacy of atenolol and captopril in reducing risk of macrovascular and microvascular complications in type 2 diabetes: UKDPS 39. BMJ 1998;317:713-720.