The Modern Treatment of Hypertension

JIS Robertson Elmbank, Manse Road, Bowling, Glasgow G60 5AA, UK

Progress in the treatment of hypertension over the past of the large but seriously ﬂawed American HDFP study 50 years is reviewed. While achievements have been especially have distorted appreciation of therapeutic considerable, they have sometimes been exaggerated beneﬁts. Journal of Human Hypertension (2000) 14, Suppl by uncritical analyses and meta-analyses of trials. Data 1, S51–S62.

Keywords: antihypertensive drugs; treatment trials; arteriosclerosis; stroke; coronary artery disease

Introduction: a case history was noted to have pulsus alternans. Blood pressure readings continued generally high, varying from A 62-year-old male senior executive was referred for 170/88–240/130 mm Hg. Then, just over 1 year from medical examination. A heavy cigarette smoker, he the first consultation, the patient developed sudden was known to have been hypertensive for at least 4 severe headache, rapidly lost consciousness, and years, but remained untreated. His principal com- was found to have a right hemiplegia; he died 2 h plaints were of cough, together with worsening later. His widow refused permission for post-mor- fatigue and exertional dyspnoea over recent weeks. tem examination to be performed. That account1 is On examination he was found to have a blood press- of the last year of the life of an important historical ure of 186/108 mm Hg. Clinically there was gross figure, Franklin D Roosevelt, US President. The year left ventricular enlargement, confirmed both electro- in question, from March 1944 to April 1945, was one cardiographically and radiologically. Arterioscler- of the most fateful in the history of the Western otic changes were seen on examination of the optic world, and the patient was then supposedly in fundi, but there were no retinal haemorrhages or charge of the affairs of the most powerful nation in exudates. There was proteinuria. A diagnosis of the Western alliance. The occasion which engen- hypertensive heart failure was made, and therapy dered such concern from the two independent phys- was begun with cardiac glycosides. Neither a icians2,3 was the Yalta conference, held in February diuretic, nor any other antihypertensive treatment, 1 1945. Mistakes made by the sick Roosevelt at Yalta was given. During the following 3 months, severe have been widely held by historians to have led to hypertension was repeatedly confirmed, values major political problems over the ensuing 50 years.4 ranging around 218–240/118–130 mm Hg. Four The President’s own physician indeed later specu- months after the initial consultation he developed lated on ‘what turn the subsequent course of history retrosternal pain radiating to both shoulders. Myo- might have taken if modern methods for the control cardial infarction was suspected, although serial of hypertension had been available’.1 Such methods electrocardiography over several days disclosed no were of course not at hand in 1945, but have been diagnostic changes. Severe hypertension persisted introduced, with progressive refinement, over the throughout the next 6 months, during which he ensuing five decades. It may therefore be of interest experienced variable but progressively worsening to review the achievements of antihypertensive ther- lassitude and dyspnoea, punctuated intermittently apy during that period. by episodes of nausea and diarrhoea. The patient continued to serve at least nominally as the head of his organization, albeit with manifestly diminishing Complications of hypertension effectiveness, and with frequent periods off work for rest and recuperation. On the occasion of an official The complications of hypertension, which we seek photograph in connection with his occupation he to correct and/or prevent with therapy, are:5 malig- had difficulty in holding up his head unaided, and nant phase hypertension, in which the blood press- another doctor who saw him then was ‘shocked to ure has risen so high, especially if the increase has see a dying man’.2 A further physician, who was to taken place rapidly, that fibrinoid arterial necrosis observe him in more detail at that time, recorded is occurring; hypertensive encephalopathy, nearly that he was ‘very sick . . . with all the symptoms of always, when it occurs, superimposed on the malig- hardening of the arteries of the brain in an advanced nant phase; overt hypertensive heart failure; stroke, stage . . . I doubt, from what I have seen, whether he which can be haemorrhagic or thrombotic; an accel- is fit for his job here’.3 Four days later the patient erated decline of renal function with age; and hypertension-associated coronary artery disease and its sequelae. It will be noted that Roosevelt manifested Correspondence: JIS Robertson, Elmbank, Manse Road, Bowling, several of these complications in the last year of Glasgow G60 5AA, UK his life. The modern treatment of hypertension JIS Robertson S52 Drug treatment of hypertension: a critique of trials In scrutinising the capacity of antihypertensive therapy to limit complications, we are necessarily con- cerned exclusively with drug treatment.5,6 Surgical manoeuvres, including dorsolumbar sympathec- tomy and adrenalectomy, are no longer employed in this context, whilst a range of non-pharmacological approaches, whatever their effect may be on blood pressure, remain unassessed concerning any capacity to limit complications.

Early uncontrolled observations Figure 1 Schematic diagram showing distribution of diastolic With the introduction of the first effective antihyper- blood pressures in a westernised adult population. Indicated are tensive drugs, it soon became apparent that they results from some trials conducted between 1964 and 1985 and could reverse the malignant phase, provided that the levels of presenting diastolic pressure above which a protective effect was shown with antihypertensive drug treatment. renal failure was not too advanced, and also relieve 8 141 5,6 Adapted from Robertson. The first Australian report put that overt hypertensive heart failure. The prognoses, threshold at 100 mm Hg; the later account33 lowered it to 95 untreated, of these complications of severe hyper- mm Hg. The other trials indicated are from refs. 7, 9, 10, 11. Ham- tension were so rapidly and consistently dire, that ilton et al,7 MRC Working Party,9 and Veterans Admin Coop 10,11 controlled trials of therapy were unnecessary for the Study Group. demonstration of benefit, and would, indeed, have been unethical. These positive observations of thera- effect with presenting diastolic pressures as low as peutic benefit have been repeatedly confirmed clini- 90 mm Hg and upwards. cally. The inclusion of trials in Figure 1 should not be taken as suggesting that such studies were above reproach. In that of Hamilton et al,7 for example, Controlled trials while it was the first controlled trial to evince bene- However, the capacity of antihypertensive drug fit, patients were allocated to therapy alternately as treatment to prevent the supervention of compli- they presented at the clinic, rather than at random, cations not yet manifested required the prosecution and for that reason Hamilton’s study has often been of controlled clinical trials. Such studies have been excluded from meta-analyses. conducted with remarkably varying competence from the time of the pioneering work of Hamilton The US Veterans Administration trial and his colleagues in England in 1964.7 Because the evolution of increasingly more effective and accept- The US Veterans Administration Study,10,11 able antihypertensive agents has paralleled although for many years providing much of the case developing expertise in clinical trial methodology, it for prophylactic antihypertensive drug treatment, is perhaps inevitable that the frailties of some initial has been subjected to severe criticism, and some essays in this field have been exposed.6 Even so, it is regard it as now of historic interest only. Concerns disturbing to find several commentators overlooking about the VA Study have included the splitting of often major defects of certain trials, and, conse- the trial into two parts after its commencement; the quently, presenting over-optimistic interpretations loss of many patients to follow-up; that those of the benefits of therapy. included had already very prevalent cardiovascular Figure 1 is taken from a review I published in complications, and hence were unrepresentative of 19868 of some of the early trials of antihypertensive the bulk of patients presenting for therapy; and that drug treatment. The curve illustrates schematically the results were seemingly inspected continually the distribution of diastolic blood pressures to be and the trial discontinued arbitrarily when an appar- found in adult Western or Westernized populations. ently significant result was obtained.6,12 One critic Notably, in 1986 the focus both epidemiologically wrote in 1983: ‘The US Veterans Study broke almost and therapeutically was on the diastolic pressure every rule of trial design and analysis, and if it were value as presaging cardiovascular morbidity. It is offered to a medical journal today it would probably now recognised that systolic pressure also requires not be accepted for publication’.12 to be considered when evaluating the dangers of hypertension, and in analysing trial outcome. Indi- HDFP cated on Figure 1 is the progressive lowering in trials conducted over the 21 years from 1964 of the Most concern then and subsequently however, has threshold of presenting diastolic pressure above centred on the large US Hypertension Detection and which a protective effect could be discerned with Follow-up Program (HDFP),13 of which the defects treatment. The most recent trial to be completed were so serious that I excluded it from Figure 1. then, the first MRC trial9 was, and remains, numeri- Patients qualifying and entering HDFP were ran- cally the largest of such studies; it had shown that domly assigned to receive either special inter- with treatment, there was statistically a protective vention (Stepped Care; SC) or the ostensible control

Journal of Human Hypertension The modern treatment of hypertension JIS Robertson S53 group (Referred Care; RC). SC participants were reduction’.15 In a later, more detailed account19 it given a free preset programme of antihypertensive was concluded ‘results confirm the previous obser- drug therapy in special centres. They were also vation that the reduction in stroke conferred by just advised to avoid a high salt intake. Additionally, SC a few years’ treatment represents much or all of the patients who were overweight, hypercholesterola- full, long-term potential effect of the blood pressure emic, or heavy smokers were offered counselling reduction achieved’. There are few diseases in designed to control those risk factors. Further, the which therapy is completely effective, and prophy- SC patients received substantial financial and other lactic drug use in hypertension appears an material support. Patients allocated to the ‘referred especially improbable candidate as an exception. care’ group (RC), by contrast, were simply directed to their regular physician, to obtain whatever treat- The Gueyffier et al meta-analysis ment he or she might deem to be appropriate. Thus HDFP did not have a strict control group; whilst, For the various reasons given and despite its fre- moreover, cardiovascular risks additional to hyper- quent citation, I shall eschew the Collins et al meta- tension were addressed. Further confusion resulted analysis15 in this review of treatment benefits, and from the adoption of different diagnostic criteria of shall rely more on that by Gueyffier et al.20 The latter coronary artery disease in various subsequent analy- excluded HDFP for the reasons already stated; more- ses.14–17 over, the various treatment trials were there grouped In the event, lower blood pressures were achieved according to the severity of hypertension at entry, on average during the study in the SC as compared the ages of the patients evaluated, and the clinical with the RC patients, and accompanying that, car- circumstances, an approach more discriminating diovascular morbidity was also lower in the SC sub- than that of Collins et al.15 jects, including, in some categories of patient, and Gueyffier et al20 considered altogether 17 trials. varying with the definition employed, less prevalent They examined three studies10,21,22 in marked hyper- coronary artery disease. However, not surprisingly tension (entry diastolic pressure 100–130 mm Hg). given the trial design, mortality from a range of non- Amongst these the principal therapeutic benefit cardiovascular disorders (gastrointestinal, respir- with drug therapy was an 89% reduction in conges- atory, cancerous, and diabetic) was also lower, by tive cardiac failure, P Ͻ 0.001. some 36%, in the SC group,13 almost certainly Two trials were considered in patients hyperten- reflecting the more ready access to diagnostic and sive following stroke. These were distinctly dispar- therapeutic facilities afforded to those patients. ate. A single-centre English study23 entered patients Amongst many critics of HDFP, one may be quoted within 2 weeks of a stroke; amongst these therapy here: ‘The HDFP study ought to be rejected because limited stroke recurrence. A larger multicentre it was not a proper trial of the treatment of hyperten- American trial24 admitted patients up to a year after sion . . . it was large and involved a lot of work and a stroke; in this the only therapeutic benefit was a expense, but its value is very limited . . . [it] seems reduction in congestive heart failure (n = 0 vs 6; P to show that good care is better than bad care, but Ͻ 0.02). it tells us little about hypertension’.12 The most impressive results were found by Gueyffier et al20 in treating hypertensive subjects aged 60 years and more. Seven such trials25–31 were The Collins et al meta-analysis considered, and entry was variously according to Considering these major defects, it is surprising that systolic and/or diastolic hypertension. Overall there HDFP was included in a much cited meta-analysis was a therapeutic reduction by 46% in congestive by Collins and colleagues15 of hypertension treat- heart failure, by 34% in stroke, and by 23% in car- ment trials, and was there said to be ‘unconfounded’ diovascular mortality (all P Ͻ 0.001); by 21% in cor- even though, for the reasons given above, it failed onary events (P Ͻ 0.01); and by 10% in all-cause to meet the authors’ own explicit definition of an mortality (P Ͻ 0.05). unconfounded study. Perhaps even more surpris- Trials in mild-to-moderate hypertension in young ingly, coronary morbidity in HDFP in that meta- or middle-aged subjects yielded much less impress- analysis was computed on the basis of the patients’ ive results. Five such9,32–35 were evaluated (as men- self-assessment, an approach which had previously tioned HDFP was omitted), the entry diastolic press- been rejected by some of those same authors,16,17 as ure ranging from 85–114 mm Hg. The only wholly subjective, and hence prone to bias. significant benefit was a 49% reduction in stroke (P The Collins et al meta-analysis15 did find that anti- Ͻ 0.001); there was no significant therapeutic hypertensive treatment lowered coronary morbidity reduction in coronary morbid events. (not coronary mortality) significantly, by a mean of Gueyffier et al20 noted that in several trials, active 14%. That finding was, however, wholly a conse- drug was started in a proportion of patients allo- quence of the very questionable inclusion, and even cated to placebo, thus eroding statistical power. then employing dubious diagnostic criteria, of When such bias was corrected, the estimate of bene- HDFP.18 The meta-analysis additionally showed a fit in treating hypertension in elderly subjects was highly significant 42% reduction in stroke with improved, the projected number of strokes pre- treatment, a finding which was simultaneously grati- vented per 1000 patient-years of therapy increasing fying and puzzling, since the authors calculated that from five to nine, and of coronary events prevented antihypertensive therapy had thus rapidly achieved from three to four. In mild-to-moderate hypertension ‘virtually all the epidemiologically expected stroke in younger patients a similar correction improved

Journal of Human Hypertension The modern treatment of hypertension JIS Robertson S54 the projected stroke prevention from one to two per than in control subjects drawn from the same popu- 1000 patient-years of therapy, but coronary benefit lation, and this holds even for those complications remained insignificant. such as stroke, known to be particularly responsive to antihypertensive treatment. Importantly, such clinic data have not shown a comparable excess of Recent trials deaths from cancer, and the deficit evidently rep- Trial outcomes published subsequent to the resents an inadequacy of antihypertensive treat- Gueyffier meta-analysis have reinforced those con- ment. The reasons for, and hence the possible recti- clusions, and have not greatly altered them. A Chi- fication of, these therapeutic shortcomings in nese study36 which recruited patients over 60 years hypertension are frequently linked or overlap, and of age, with either systolic or diastolic hypertension, thus can usefully be considered together. found on treatment a significant decrease in strokes and all cardiovascular events; as expected, coronary Organisational aspects artery morbidity was very low in these Chinese 5,46 patients. A European trial37 of isolated systolic First, and most obviously, limited organisational hypertension in elderly subjects also demonstrated skills would have constrained achievements in a therapeutic reduction in stroke and total cardio- medical outcome. For example, the much-quoted 44 vascular morbidity, but not in cardiac failure or cor- report from the Glasgow Blood Pressure Clinic onary morbidity. derive from data obtained over a period of 15 years (1968–1983). Administrative and therapeutic efficiency improved visibly both during and after Drug treatment and renal function that time. On that basis alone, one would now expect much better results than were obtained sev- The effects of therapy in slowing the accelerated eral decades ago. Even so, when those patients with decline of renal function evident in hypertension treated diastolic pressures below 90 mm Hg were have been difficult to evaluate. Uncontrolled studies evaluated separately, morbidity still remained in already mentioned5,6 have demonstrated clear bene- excess of normal; thus the relative therapeutic defi- fits in this regard in malignant phase hypertension, cit was probably not attributable solely to any limi- a situation in which, untreated, rapidly advancing tations in the delivery of therapy. renal impairment is prominent. However, prospective trials in milder hypertension have been too brief, and the indices of renal function employed Inadequate blood pressure reduction probably too crude, for any such benefit to be At least a proportion of the failure of therapy to shown.6,38 Nevertheless, several retrospective analy- eliminate fully the complications of hypertension ses have indicated that effective blood pressure con- seems to be attributable to inadequate blood press- trol can retard a decline in renal function.6,39,40 ure reduction. Surveys have indicated that fewer than 30% of patients have on-treatment pressures Shortcomings of antihypertensive below 140/90 mm Hg,47,48 whilst most benefit is treatment: reasons obtained by lowering the systolic to about 140 mm Hg and the diastolic to about 90 mm Hg, with The trials of antihypertensive drug treatment sum- slight additional advantage at lower values.9,48,49 marised briefly herein have thus revealed differen- Failure to achieve such effective blood pressure tial benefits from such therapy. Those complications reduction has two principal causes: intrinsic inad- to have been most evidently prevented or corrected equacy of some antihypertensive drugs given either are those most obviously a consequence of the high alone or in combination; and the presence of side- blood pressure as such, namely the malignant phase, effects, which will lead often to sub-optimal dosage overt heart failure, and stroke. It has indeed been and poor patient compliance. further held that such therapeutic advantage concerns haemorrhagic rather than thrombotic The J-curve question stroke.41,42 Significant reduction in coronary morbidity has been found only in hypertensive persons A contrary suggestion50,51 is that in hypertensive over the age of 60, even though coronary events out- patients with coronary artery stenosis, therapeutic number strokes in younger patients.20 There is hence lowering of the diastolic pressure to the range 70– undoubtedly therapeutic shortfall concerning those 85 mm Hg might increase the danger of provoking complications related less directly to arterial hyper- myocardial infarction (the ‘J-curve’ concept). The tension, most particularly via atheroma HOT trial48 was designed to illuminate this issue. It (atherosclerosis). comprised three groups of patients, in whom the tar- An additional, rather different, deficiency of anti- get diastolic pressures were respectively р90, р85 hypertensive treatment has emerged, notwithstand- and р80 mm Hg. The results concerning compli- ing the claim made in the Collins et al meta-analy- cations for the three groups showed rather small dif- sis15 that such treatment might eliminate all the ferences; only the trend for the rate of all myocardial epidemiological risk of hypertension-related stroke. infarction to be less the lower the target pressure Several reports43–45 of data from hypertension clin- was of borderline significance. The lowest incidence ics have shown that morbidity and mortality in of major cardiovascular events occurred at a mean treated hypertensive patients have remained higher achieved diastolic pressure of 82.6 mm Hg. While

Journal of Human Hypertension The modern treatment of hypertension JIS Robertson S55 there was a slight increase in morbidity at lower example, in the very large first MRC trial, sudden diastolic values, this was not significant. Surpris- death rates per 1000 patient years in men, in which ingly, the HOT trial publication provided scant sex the great preponderance of such catastrophes information52 on the initially central issue, the occurred, were 2.7 in those patients allocated to important group of 3080 hypertensive patients with receive thiazide treatment, as against 1.1 on beta- a history of coronary artery disease, although blocker, and 1.9 on placebo.63 The difference amongst these an achieved diastolic pressure below between thiazide and beta-blocker in this respect 81 mm Hg did not apparently carry extra risk.53 was highly significant (P = 0.01). Thiazide diuretics Thus although the J-curve controversy continues,54 cause elevation of plasma cholesterol and triglycer- it is not evidently a major practical therapeutic ides.64 The increase is in all fractions and the ratios issue.5 of these therefore remain unchanged. This plasma cholesterol elevation is most marked in the first year of thiazide treatment and abates somewhat there- Adverse effects: drugs used in trials after.65,66 The prognostic significance remains uncer- A further caveat is that many of the drugs employed tain, although undoubtedly troubling. A further hitherto in the treatment of hypertension possess, in aspect of thiazide treatment is that fasting blood addition to their ability to lower blood pressure, a sugar may be elevated and the glucose tolerance test variety of potentially adverse cardiovascular effects, become diabetic.64 Hypokalaemia appears to be which may well have limited their capacity to implicated in this diabetogenic effect, possibly via reduce complications. There are obvious difficulties an action on pancreatic islets. in computing such a hypothecated amalgam of mer- its and demerits, and, not surprisingly, heated Beta-blockers: problems controversies have arisen. In considering these dis- cordant issues, it is first requisite to appraise which Beta-blocking drugs do not change total plasma classes of drug have been employed on those treat- cholesterol, but they alter the ratio of high density ment trials to have shown benefit in hypertension. lipoproteins to that of low density plus very low As I have already discussed herein, there is in turn density lipoproteins (HDL: LDL + VLDL) in a poten- much disagreement on which trials have been suf- tially adverse way.67 Those beta-blockers with high ficiently well designed and executed to warrant intrinsic sympathomimetic activity have least effect inclusion for evaluation. Nevertheless, the oft- whilst the beta-1 selective drugs are intermediate. repeated saw that the only drug classes to have been Beta-blockers have also been noted to lead to hyper- used in trials showing benefit are diuretics and beta- glycaemia when given long term.67 Moreover, non- blockers is clearly false.6,46,54,55 Of 23 trials I sur- selective beta-blockers prolong insulin-induced veyed in 1997,6 diuretics were employed in 22, hypoglycaemia while selective agents probably do albeit in various types and doses, and with widely not. Both non-selective and selective beta-blockers differing attempts at concurrent potassium conser- inhibit the tachycardia and palpitations vation. Centrally-acting drugs were the next most accompanying hypoglycaemia. often used, featuring in 18 trials (methyldopa in 11, rauwolfia in 10, and clonidine in one). Beta-blockers Centrally-acting drugs: problems were employed in fewer than half, nine of the 23 trials. Hydralazine appeared in six, guanethidine in The older types of centrally-acting drugs, rauwolfia five, and a calcium antagonist in four, while a gang- derivatives, methyldopa, and clonidine, employed lion blocker, bethanidine, an alpha-blocker, and an in many of the early treatment trials of hypertension, angiotensin-converting enzyme (ACE) inhibitor, can cause drowsiness and depression, and indeed figured in one trial each. Unspecified supplementary reports of suicide in patients taking reserpine pre- drugs were prescribed in at least six of the 23 trials. vented that drug from becoming popular in Bri- Since that review appeared, four more large studies tain.5,68 have been published, in three36,37,48 of which the basic therapy was a calcium antagonist, and in the Calcium antagonists: problems fourth56 an ACE inhibitor. Furthermore, and most germanely, combined therapy, that is, the require- Dihydropyridine calcium antagonists have been ment to administer two or more drug classes indicated severally as provoking adverse coronary together in order to achieve adequate control of events in susceptible patients,69–74 of causing serious blood pressure, was necessary in from 60–100% of bleeding,75,76 of inducing aplastic anaemia,77 and of patients in various trials.6,57 We may now consider predisposing to cancer78 and suicide.79 These vari- some of the potential demerits of certain types of ous reports have in turn been severely criticised and antihypertensive drug. the issues remain highly controversial.80–84

Diuretics: problems Atherosclerosis and its consequences Diuretics cause hypokalaemia and depletion of body A most obvious therapeutic deﬁciency concerns potassium content, and hence can predispose to car- nevertheless, as described earlier, the comparative diac arrhythmias.58 While any consequent risks have failure to limit the atheromatous (atherosclerotic) been discounted by some commentators,59 others complications of hypertension. In retrospect the have been less complacent.58,60–62 To cite one shortfall is hardly surprising. Early antihypertensive

Journal of Human Hypertension The modern treatment of hypertension JIS Robertson S56 drugs were introduced almost exclusively according the combined beta-blocker/nitric oxide releasing to their capacity to lower high blood pressure, and drug nebivolol. Large arterial compliance is largely thus would be expected to confer less benefit regard- unchanged by propranolol, whereas it is worsened ing those complications only indirectly resulting by hydralazine.92–96 from arterial hypertension.6 Moreover, initial epidemiological and therapeutic attention was directed to the diastolic pressure,6,85 which is particularly asso- Small arterial changes ciated with arteriolar lesions and cardiac failure. In hypertension, the cross-sectional wall:lumen Systolic hypertension, which is of more recent inter- ratio of resistance arteries is increased, a result est, is now seen to be more closely related to arterial mainly of remodelling of medial smooth muscle.97,98 atheroma (atherosclerosis). Four cardiovascular This change both reinforces the progression of accompaniments of hypertension which are hypertension and limits the antihypertensive effec- attracting especial current therapeutic attention are tiveness of agents whose main action is to relax left ventricular hypertrophy, reduced physical com- small arteries. Antihypertensive treatment using a pliance in large arteries, increased media: lumen range of drugs has been shown to diminish the cross-sectional ratio in resistance arteries, and both wall:lumen ratio of resistance vessels.99 Further, in physical and biochemical dysfunction of endo- a direct prospective comparison, an ACE inhibitor thelium. These several abnormalities and their was effective in this regard, whereas a beta-blocker consequences are frequently interlinked. was not.100 ACE inhibitors seem indeed to be especially valuable in this context.101–103 Left ventricular hypertrophy With the development of left ventricular hypertro- Blood flow turbulence phy in hypertension, the left ventricle remains cap- able of emptying well, although it fills poorly. End- Blood flow via a non-branching length of a large diastolic pressure is increased, plasma concen- artery is characterised by a rapidly moving central trations of atrial natriuretic peptide are elevated, axial stream, with a more sluggish peripheral boundary layer, an arrangement which limits endo- coronary arterial filling is impaired, subendocardial 5,42 ischaemia occurs, ventricular ectopic activity is thelial damage from kinetic energy. However, prominent, and prognosis is worsened.5,85 Conse- when a large artery branches, axial stream impinge- quently, recent interest has centred on the capability ment on the vessel wall at the bifurcation causes of different classes of antihypertensive drug to high shear endothelial damage, while opposite to reverse such left ventricular hypertrophy, on the as this low shear boundary layer separation occurs. yet unproven assumption that any such effect would Both are sites where atheromatous (atherosclerotic) be beneficial. Three large but retrospective plaques form. This takes place rapidly, but to a lim- reviews86–88 of this topic have returned rather dif- ited extent, with high shear; more slowly, but fering results, perhaps not surprisingly given the eventually usually more extensively, with low unavoidable deficiencies of analyses of data shear. Flow turbulence exacerbates these trends. obtained originally for other purposes. Classes of Moreover, in the presence of an atheromatous drug which appear effective in correcting left ven- arterial stenosis, turbulence is further worsened, tricular hypertrophy include sympatholytic agents, with more rapid blood flow occurring through the for example methyldopa; angiotensin-converting narrowing, and zones of stasis and recirculation dis- enzyme (ACE) inhibitors; calcium antagonists; and tal to it. beta-blockers. Less clearly of value in this regard are diuretics and alpha-blockers; whilst the so-called New drug developments direct-acting vasodilators hydralazine and minoxi- dil are ineffective. Two reports89,90 indicated that the Two newly-developed types of antihypertensive non-peptide angiotensin II antagonist, losartan, drug may prove to be especially effective in limiting might worsen left ventricular hypertrophy, although the sequelae of disturbances of blood flow and of a later account91 was more reassuring. Undoubtedly, dysfunction of the endothelium. Bosentan104 is an appropriately designed prospective trials employing antihypertensive antagonist of endothelin receptors; well-validated methods will clarify several since increased circulating endothelin has been presently obscure matters. found to be associated with especially extensive arterial atherosclerosis, therein lies the possibility of restricting the formation of hypertension-associated Large arterial changes atheromatous lesions.5,105,106 Nebivolol is a highly In hypertension, important changes take place in the selective beta-1 blocker which has the further walls of large arteries.85,92–95 Physical compliance is capacity to stimulate the endothelial release of nitric diminished, whilst the response to constrictor stim- oxide;107 with this combination it is hoped that uli is enhanced. With these features, systolic press- effective blood pressure lowering will be combined ure is disproportionately increased, an aspect with limitation of both structural and biochemical closely associated with left ventricular hypertrophy, impairment of the endothelium. Likewise the now and blood flow turbulence, which predisposes to the well-established ACE inhibitors, but not nifedipine, formation of atheroma,5,42 is worsened. Large arterial can improve endothelium-dependent vasodilatation compliance is improved by ACE inhibitors, and by in hypertensive patients, although the mechanism of

Journal of Human Hypertension The modern treatment of hypertension JIS Robertson S57 action is not determined.108 These matters are the sis111 is that the human population is unlikely to be subject of ongoing clinical trials. homogeneous in response to variations in salt Other interest has been in antihypertensive drugs intake, a consideration which argues against the also possessed of antiplatelet actions. These include indiscriminate injunction of unsupervised salt some dihydropyridine type calcium antagonists and restriction. Nevertheless, it is the case that many the combined serotonergic and alpha-adrenergic young hypertensive persons (ie, below about the age blocking agent ketanserin.85,109 However, with the of 40) have a significant deficit of body sodium con- reports of various haemorrhagic complications poss- tent, whereas older and more severely hypertensive ibly resulting from the antiplatelet properties of cer- persons frequently have body sodium excess.113 tain calcium antagonists,75,76 enthusiasm for this Thus dietary salt restriction is unlikely to be gener- approach temporarily waned, then revived with ally helpful as a preventive measure, or therapeuti- reports of the beneficial effects found with added cally much useful in young patients, though it might aspirin in the HOT trial.53 reinforce drug treatment in many older and/or more A range of other approaches continues to be severely hypertensive subjects. Meta-analyses of explored.106 While ACE inhibitors and non-peptide controlled randomised treatment trials have broadly antagonists of angiotensin II have proved effective confirmed these precepts,114,115 although at best the in lowering high blood pressure, inhibitors of the antihypertensive effect of salt reduction is mod- enzyme renin itself, potentially of major value, have est.111,112,114,115 Moreover, because salt restriction so far been constrained by low oral bioavailability elevates plasma renin, and raises the plasma and brief duration of action. LDL:HDL ratio,115 both of which changes could pre- Neutral endopeptidase inhibitors block the break- dispose to adverse cardiovascular events, several down of vasodilator natriuretic peptides and so commentators have expressed caution on this issue lower blood pressure. However, as monotherapy unless and until morbidity data are avail- their antihypertensive action is weak. Nevertheless, able.111,112,115 combined inhibition of natriuretic peptide break- down and of ACE could be promising. At least one Potassium, magnesium, calcium such drug, omapatrilat, is currently being evaluated. More speculative is genetic therapy, including the Modest but seemingly real antihypertensive effects possibility of transfection with vectors enhancing have been noted with dietary supplementation of the endogenous production of various antihyperten- either potassium116–119 or magnesium.120,121 Any sive and antihypertrophic substances. antihypertensive effect of dietary calcium reinforce- Again, we may return to some early evidence. As ment has both its advocates122–124 and detract- mentioned above, centrally-acting antihypertensive ors.125,126 drugs featured in a majority of trials which showed clinical benefit with antihypertensive therapy.6 Alcohol However, side-effects limited the popularity of the older agents reserpine, methyldopa and clonidine.5 There is clear evidence of an association between The recent introduction of imidazaline I1-receptor alcohol consumption and raised blood pressure, and agonists such as moxonidine, which are effective in of an antihypertensive effect of alcohol restric- lowering blood pressure yet without the side-effect tion.5,127–129 Further, it has been suggested that alco- burden of the earlier drugs, could open exciting hol intake might be an independent risk factor for therapeutic avenues.68 stroke.130 However, this issue is complicated by repeated observations that moderate alcohol intake 5,131–134 Non-pharmacological treatment limits cardiovascular complications. I now turn to the vexed question of non-pharmaco- Caffeine logical therapy of hypertension. Although several of such non-pharmacological approaches have been Caffeine ingestion has an acute pressor effect,135 widely advocated both as a prelude and as an although current evidence is contrary to habitual accompaniment to drug therapy, none has yet been coffee consumption raising blood pressure long assessed concerning any capacity to affect the com- term,136 or to the dietary elimination of coffee pos- plications of hypertension.5,6 This scientific deficit, sessing a worthwhile antihypertensive effect.5 which might have been expected to engender caution, has sometimes, surprisingly to the present Marine oils writer, contrarily been accompanied by strident advocacy. There are suggestive, but in the view of the present writer,5 as yet inconclusive, data that the consumption of diets rich in marine oils containing polyunsa- Dietary salt turated fatty acids such as eicosapentanoic acid and Perhaps the most prominent amongst several con- docohexanoic acid can lower blood pressure. tentious aspects is whether or not an unnecessarily high dietary intake of sodium chloride (‘salt’) can Weight reduction cause or worsen hypertension. There can currently be few aspects of medicine so beset by zealotry and Reduction of excessive body weight has consistently accusations of bad faith.110–112 Requiring empha- been demonstrated to lower high blood pressure.5 In

Journal of Human Hypertension The modern treatment of hypertension JIS Robertson S58 comparative trials, the effect has exceeded that of malignant phase and overt hypertensive heart fail- salt restriction.111 There is usually no additive blood ure can be readily controlled, and stroke substan- pressure reduction if the two measures are com- tially prevented, much less impact has been made bined.111 on hypertension-associated coronary artery disease and its consequences. I have tried herein to indicate how these problems might be approached. Drug Exercise therapy, as heretofore, must necessarily be central Regular physical exercise, independently of any to our efforts. First, we should identify, and then accompanying weight reduction, can be an effective, address, specific pathophysiological disturbances in and thus far uncontroversial, means of blood press- hypertension. This elusive goal, if attained, would ure reduction.5,137,138 permit more effective, and much better tolerated, treatment. We require therapy attended by fewer side-effects, be they symptomatic, biochemical, or Psychological treatment pathophysiological, a need almost certainly likely to A range of psychological or behavioural approaches, be met if the initial requirement of defining underly- which include yoga, transcendental meditation, ing pathophysiology noted above were to be real- biofeedback procedures, and relaxation training, ized. Therapy should also resolve left ventricular have been reported, at least short-term, and with hypertrophy, improve large arterial compliance, dedicated therapists and subjects, to lower blood diminish small arterial wall:lumen cross-sectional pressure.139 However, the longer-term effects remain ratio, and preserve endothelial structure and func- uncertain, the methods are unsuitable for broad- tion. Antiplatelet properties could be advantageous. scale application, and these approaches have faded Once-daily drug administration is desirable, and from general advocacy.5,46 cheapness comprises an obvious attraction. Despite often fervent advocacy, non-pharmaco- In many aspects, the non-pharmacological approach logical antihypertensive treatment exhibits at to antihypertensive therapy remains very unsatisfac- present widespread inadequacies. More understand- tory. There is obvious appeal to devotees of public ing is needed of the mechanism of nearly all of these ‘healthism’140 and to government administrators or others responsible for providing medical funds. Nevertheless it is disturbing that too often dis- passionate appraisal has been displaced by discor- dant evangelism, exemplified by some contributions to the uncommunicative ‘debates’ concerning dietary salt intake.110 Too often genuine and crucial uncertainties have been clouded by vehement invec- tive. Herein clinical science has been ill served.

Cigarette smoking Less controversial is the influence of cigarette smoking.5 While the acute effects of smoking on arterial pressure are complex, and habitual smokers have lower pressures than do non-smokers, smoking clearly increases the likelihood of a wide range of Figure 2 Photograph of Roosevelt, with Churchill and Stalin, cardiovascular complications, including stroke and taken at Yalta, February 1945. Note Roosevelt’s cigarette. coronary and peripheral arterial disease, and thus greatly compounds the risks of hypertension. Both fibromuscular dysplastic and atheromatous renal artery stenosis are more prevalent in smokers. The incidence of malignant hypertension is higher in smokers, and its prognosis when treated is poorer in those who continue to smoke.141 Furthermore, the capacity of antihypertensive drugs to lower arterial pressure, and to limit a range of cardiovascular complications, is impaired in persistent smokers. Cess- ation of cigarette smoking lessens the vascular risk. Hypertensive patients especially should therefore be firmly enjoined not to smoke.

Conclusions This review has shown that although antihypertensive drug therapy has brought substantial achieve- Figure 3 Speculative impression of how Figure 2 might have ments, these have been distinctly limited, and that appeared had it been possible to treat Roosevelt’s hypertension further improvements must be sought. While the in 1944–1945.

Journal of Human Hypertension The modern treatment of hypertension JIS Robertson S59 approaches. Too often, over-credulous invocation is incidence of myocardial infarction and angina pec- substituted for the critical evaluation of any adverse toris: 5-year findings of the Hypertension Detection and beneficial influences on morbidity and mor- and Follow-up Program. Hypertension 1984; 6 (Suppl tality. A detailed, discriminating scrutiny should be 1): 198–206. an indispensable prelude to the general recommen- 15 Collins R et al. Blood pressure, stroke, and coronary heart disease. Part 2: Short-term reductions in blood dation of such therapy. pressure: overview of randomised drug trials in their I opened this review with a medical account of epidemiological context. Lancet 1990; 335: 827–838. the last year in the life of Franklin Roosevelt. Figure 16 MacMahon SW, Culter JA, Furberg CD, Payne GH. The 2 is the photograph taken at Yalta in February 1945, effects of drug treatment for hypertension on mor- on the occasion when it was noted that he had dif- bidity and mortality from cardiovascular disease: a ficulty even in holding up his head unaided.2 review of randomized controlled trials. Progr Cardiov- Depicted sitting between Churchill and Stalin is the asc Dis 1986; 29 (Suppl 1): 99–118. haggard Roosevelt, still clutching his cigarette, even 17 Cutler JA, MacMahon SW, Furberg CD. Controlled though he was posing for an official photograph. clinical trials of drug treatment for hypertension: a Within a few weeks Roosevelt was to die from the review. Hypertension 1989; 13 (Suppl 1): 36–44. 18 Alderman MH. Meta-analysis of hypertension treat- consequences of untreated, indeed at that time ment trials. Lancet 1990; 335: 1092–1093. untreatable, hypertension. Antihypertensive drugs 19 MacMahon S. Blood pressure and the prevention of were not available in 1945. If such therapy had been stroke. J Hypertens 1996; 14 (Suppl 6): 39–46. possible, history, and the picture (Figure 3), might 20 Gueyffier F, Froment A, Gouton M. New meta-analysis have been very different. of treatment trials of hypertension: improving the estimate of benefit. J Hum Hypertens 1996; 10: 1–8. 21 Wolff FW, Lindeman RD. Effects of treatment in hypertension: results of a controlled study. J Chron Dis 1966; 19: 227–240. References 22 Barraclough MA et al. Control of moderately raised 1 Bruenn HG. Clinical notes on the illness and death of blood pressure: report of a co-operative randomized President Franklin D. Roosevelt. Ann Int Med 1970; 72: controlled trial. Br Med J 1973; 3: 434–436. 579–591. 23 Carter AB. Hypotensive therapy in stroke survivors. 2 Sheehan D. President Roosevelt’s health. Ann Int Med Lancet 1970; i: 485–489. 1970; 73: 344–345. 24 Hypertension-Stroke Cooperative Study Group. Effect 3 Lord Moran. Yalta diary. In: Churchill: The Struggle of antihypertensive treatment on stroke recurrence. J for Survival, 1940–65. Constable: London, 1966, pp Am Med Ass 1974; 229: 409–418. 216–241. 25 Sprackling ME, Mitchell JRA, Short AH, Watt G. Blood 4 L’Etang H. Ailing Leaders in Power, 1914–1994. Royal pressure reduction in the elderly: a randomised con- Society of Medicine Press: London, 1995. trolled trial of methyldopa. Br Med J 1981; 283: 5 Robertson JIS, Ball SG. Hypertension for the Clinician. 1151–1153. W.B. Saunders: London/Philadelphia, 1994. 26 Amery A et al. Mortality and morbidity results from 6 Robertson JIS. Hypertension. In: Pitt B, Julian D, Poc- the European Working Party on High Blood Pressure ock S (eds.) Clinical Trials in Cardiology. W.B. Saund- in the Elderly trial. Lancet 1985; i: 1349–1354. ers: London/Philadelphia, 1997, pp 115–160. 27 Coope J. Warrender TS. Randomised trial of treatment 7 Hamilton M, Thompson EN, Wisniewski TKM. The of hypertension in elderly patients in primary care. Br role of blood pressure control in preventing compli- Med J 1986; 293: 1145–1151. cations of hypertension. Lancet 1964; i: 235–238. 28 Perry HM, Smith WM, McDonald RH et al. Morbidity 8 Robertson JIS. The 1985 Trials of Hypertension Treat- and mortality in the Systolic Hypertension in the Eld- ment. Hypertension Yearbook 1986. Gower Academic erly Program (SHEP) pilot study. Stroke 1989; 20:4– Journals: London/New York/Paris, pp 3–25. 13. 9 Medical Research Council Working Party. MRC trial of 29 SHEP Cooperative Research Group. Prevention of treatment of mild hypertension: principal results. Br stroke by antihypertensive drug treatment in older per- Med J 1985; 291: 97–104. sons with isolated systolic hypertension: final results 10 Veterans Administration Cooperative Study Group. of the Systolic Hypertension in the Elderly Program Effects of treatment on morbidity in hypertension: (SHEP). J Am Med Ass 1991; 265: 3255–3264. results in patients with diastolic pressures averaging 30 Dahlo¨fBet al. Morbidity and mortality in the Swedish 115 through 129 mmHg. J Am Med Ass 1967; 202: Trial in Old Patients with Hypertension (STOP- 1028–1034. Hypertension). Lancet 1991; 338: 1281–1285. 11 Veterans Administration Cooperative Study Group. 31 Medical Research Council Working Party. Medical Effects of treatment on morbidity in hypertension. II. Research Council trial of treatment in older adults: Results in patients with diastolic blood pressures aver- principal results. BMJ 1992; 304: 405–412. aging 90 through 114 mmHg. J Am Med Ass 1970; 213: 32 McFate Smith W. Treatment of mild hypertension: 1143–1152. results of a ten-year intervention trial. Circ Res 1977; 12 Hampton JR. An appraisal of hypertension trials. Med- 40 (Suppl 1): 98–105. icographia 1983; 5 (Suppl 2): 12–15. 33 Australian Mild Hypertension Trial Management Com- 13 Hypertension Detection and Follow-up Program Coop- mittee. Australian therapeutic trial in mild hyperten- erative Group. Five-year findings of the Hypertension sion. Lancet 1980; i: 1261–1267. Detection and Follow-up Program: I. Reduction in 34 Helgeland A. Treatment of mild hypertension: a five mortality in persons with high blood pressure, includ- year controlled drug trial: the Oslo study. Am J Med ing mild hypertension. J Am Med Ass 1979; 242: 1980; 69: 725–732. 2562–2571. 35 Perry HM et al. Evaluation of drug treatment in mild 14 Hypertension Detection and Follow-up Program Coop- hypertension: VA-NHLBI feasibility trial. Ann N Y erative Group. Effect of stepped care treatment on the Acad Sci 1978; 304: 267–288.

Journal of Human Hypertension The modern treatment of hypertension JIS Robertson S60 36 Gong L et al. Shanghai trial of nifedipine in the elderly cardiac arrhythmias, and sudden death. J Hypertens (STONE). J Hypertens 1996; 14: 1237–1245. 1992; 10: 301–316. 37 Staessen JA et al. Randomised double-blind compari- 59 McInnes GT, Yeo WW, Ramsay LE, Moser M. Cardi- son of placebo and active treatment for older patients otoxicity and diuretics: much speculation—little with isolated systolic hypertension. Lancet 1997; 350: substance. J Hypertens 1992; 10: 317–335. 757–764. 60 Siscovick DS et al. Diuretic therapy and the risk of car- 38 Zanchetti A, Sleight P, Birkenha¨ger WH. Evaluation of diac arrest. N Engl J Med 1994; 330: 1852–1857. organ damage in hypertension. J Hypertens 1993; 11: 61 Hoes AW, Grobbee DE, Peet TM, Lubsen J. Do non- 875–882. potassium-sparing diuretics increase the risk of sud- 39 Tomson CRV, Petersen K, Heagerty AM. Does treated den death in hypertensive patients? Recent evidence. essential hypertension result in renal impairment? A Drugs 1993; 47: 711–733. cohort study. J Hum Hypertens 1991; 5: 189–192. 62 Grobbee DE, Hoes AW. Non-potassium-sparing 40 Ruilope L et al. Does an adequate control of blood diuretics and risk of sudden cardiac death. J Hypertens pressure protect the kidney in essential hypetension? 1995; 13: 1539–1545. J Hypertens 1990; 8: 525–531. 63 Medical Research Council Working Party. Coronary 41 Black DG et al. Effects of treatment for hypertension heart disease in the Medical Research Council trial of on cerebral haemorrhage and infarction. Br Med J treatment of mild hypertension. Br Heart J 1988; 59: 1984; 289: 156–159. 364–378. 42 Spence JD. Cerebral consequences of hypertension: 64 Morgan T. The use of diuretic drugs and aldosterone where do they lead? J Hypertens 1996; 14 (Suppl 5): antagonists in hypertension. In: Doyle AE (ed). Clini- 139–145. cal Pharmacology of Antihypertensive Drugs. Hand- 43 Lindholm L, Ejlertsson G, Schersten´ B. High risk of book of Hypertension, Vol. 11, Elsevier: Amsterdam, cerebro-cardiovascular morbidity in well treated male 1988, pp 73–102. hypertensives: a retrospective study of 40–50 year old 65 Medical Research Council Working Party on Mild to hypertensives in a Swedish primary care district. Acta Moderate Hypertension. Adverse reactions to bendro- fluazide and propranolol for the treatment of mild Med Scand 1984; 216: 251–259. hypertension. Lancet 1982; ii: 839–840. 44 Isles CG et al. Mortality in patients of the Glasgow 66 Curb JD et al. Adverse effects of antihypertensive Blood Pressure Clinic. J Hypertens 1986; 4: 141–156. medications in the Hypertension Detection and Fol- 45 Andersson OK et al. Survival in treated hypertension: low-up Program. Progr Cardiovasc Dis 1986; 29 (Suppl follow up study after two decades. BMJ 1998; 317: 1): 73–88. 167–171. 67 Prichard BNC, Owens CWI. Beta-adrenoceptor block- 46 Robertson JIS. Guidelines for the treatment of hyper- ing drugs. In: Doyle AE (ed). Clinical Pharmacology tension: a critical review. Cardiovasc Drugs Ther 1994; of Antihypertensive Drugs. Handbook of Hypertension, 8: 665–672. Vol. 11, Elsevier: Amsterdam, 1988, pp 187–243. 47 Mancia G et al. Blood pressure control in the hyperten- 68 Robertson JIS. Risk factors and drugs in the treatment of sive population. Lancet 1997; 349: 454–457. hypertension. JHypertens1997; 15 (Suppl 1): 43–46. 48 Hansson L et al. Effects of intensive blood-pressure 69 Psaty BN et al. The risk of myocardial infarction asso- lowering and low-dose aspirin in patients with hyper- ciated with antihypertensive drug therapies. J Am Med tension: principal results of the Hypertension Optimal Ass 1995; 274: 620–625. Treatment (HOT) randomised trial. Lancet 1998; 351: 70 Furberg CD, Psaty BM, Meyer JV. Nifedipine: dose- 1755–1762. related increase in mortality in patients with coronary 49 IPPPSH Collaborative Group. Cardiovascular risk and heart disease. Circulation 1995; 92: 1326–1331. risk factors in a randomised trial of treatment based on 71 Furberg CD, Psaty BM. Calcium antagonists: not appro- the beta-blocker oxprenolol: the International Prospec- priate as first line antihypertensive agents. Am J Hy- tive Primary Prevention Study in Hypertension pertens 1996; 9: 122–125. (IPPPSH). J Hypertens 1985; 3: 379–392. 72 Josefson D. Infarction risk found with calcium channel 50 Cruickshank JM. Coronary flow reserve and the J curve blocker. BMJ 1998; 316: 797. relation between diastolic blood pressure and myocar- 73 Cutler JA. Calcium-channel blockers for hypertension: dial infarction. Br Med J 1988; 297: 1227–1230. uncertainty continues. N Engl J Med 1998; 338: 679– 51 Zanchetti A et al. How much should blood pressure 681. be lowered? The problem of the J-shaped curve. J Hy- 74 Pahor M, Psaty BM, Furberg CD. Treatment of hyper- pertens 1989; 7 (Suppl 6): 338–348. tensive patients with diabetes. Lancet 1998; 351: 52 Cruickshank JM. Hypertension Optimal Treatment 689–690. (HOT) trial. Lancet 1998; 352: 573–574. 75 Kario K, Imiya M, Ohta Y, Shimada K. Gastrointestinal 53 Hansson L, Zanchetti A. Hypertension Optimal Treat- bleeding induced by amlodipine. J Hum Hypertens ment (HOT) trial. Lancet 1998; 352: 574–575. 1995; 9: 206–207. 54 Kaplan NM. J-curve not burned off by HOT study. 76 Kario K, Hoshide S, Nakagawa Y, Shimada K. Cerebral Lancet 1998; 351: 1748–1749. haemorrhage triggered by calcium antagonist. J Hum 55 Robertson JIS. The 1993 treatment guidelines: more Hypertens 1998; 12: 143–144. criticisms. Am J Hypertens 1996; 9: 95–97. 77 Laporte J-R et al. Fatal aplastic anaemia associated 56 Hansson L et al. The Captopril Prevention Project with nifedipine. Lancet 1998; 352: 619–620. (CAPPP) in hypertension: final results. J Hypertens 78 Pahor M et al. Calcium channel blockade and inci- 1998; 16 (Suppl 2): 23. dence of cancer in aged populations. Lancet 1996; 348: 57 Hansson L. The benefits of lowering elevated blood 493–497. pressure: a critical review of studies of cardiovascular 79 Lindberg G et al. Use of calcium channel blockers and morbidity and mortality in hypertension. J Hypertens risk of suicide: ecological findings confirmed in popu- 1996; 14: 537–544. lation based cohort study. BMJ 1998; 316: 741–745. 58 Singh BN, Hollenberg NK, Poole-Wilson PA, Robert- 80 Epstein M. Calcium antagonists: still appropriate as son JIS. Diuretic-induced potassium and magnesium first line antihypertensive agents. Am J Hypertens deficiency: relation to drug-induced QT prolongation, 1996; 9: 110–121.

Journal of Human Hypertension The modern treatment of hypertension JIS Robertson S61 81 Mancia G, Van Zwieten PA. How safe are calcium untreated essential hypertension. Hypertension 1995; antagonists in hypertension and coronary heart dis- 25 (Part 1): 474–481. ease? J Hypertens 1996; 14: 13–17. 103 Marceau M, Kouame´ N, Larcourcie`re Y, Cle´roux J. 82 Mackay JA, Sever PS. Calcium channel blockers in Vascular structure in the forearm and calf after 6 the dock: innocent or guilty? J Hum Hypertens 1996; months of angiotensin converting enzyme inhibition 10: 277–280. in elderly hypertensive subjects with left ventricular 83 Hole DJ et al. Cancer risk of hypertensive patients tak- hypertrophy. J Hypertens 1998; 16: 673–679. ing calcium antagonists. J Hypertens 1998; 16: 119– 104 Krum H et al. The effect of an endothelin-receptor 124. antagonist, bosentan, on blood pressure in patients 84 Hamet P. Calcium antagonists and cancer: limitations with essential hypertension. N Engl J Med 1998; 338: of pros and cons. J Hypertens 1998; 16: 1101. 784–790. 85 Robertson JIS. Antihypertensive therapy: achieve- 105 Webb DJ, Strachan FE. Clinical experience with ments, failures, and prospects. J Cardiovasc Pharma- endothelin antagonists. Am J Hypertens 1998; 11 col 1990; 16 (Suppl 7): 102–104. (Suppl): 71–79. 86 Hachamovitch R, Strom JA, Sonnenblick EH, Frish- 106 Bonn D. Extending the targets for antihypertensive man WH. Left ventricular hypertrophy in hyperten- drugs. Lancet 1998; 352: 378. sion and the effects of antihypertensive drug therapy. 107 Van Nueten L et al. Nebivolol versus enalapril in the Curr Problems Cardiol 1988; 13: 369–422. treatment of hypertension: a double-blind random- 87 Cruickshank JM, Lewis J, Moore V, Dodd C. Reversal ised trial. J Hum Hypertens 1997; 11: 813–819. of left ventricular hypertrophy by differing types of 108 Millgard J, Hagg A, Sarabi M, Lind L. Captopril, but antihypertensive therapy. J Hum Hypertens 1992; 6: not nifedipine, improves endothelium-dependent 85–90. vasodilatation in hypertensive patients. J Hum Hyp- 88 Dahlo¨f B, Pennert K, Hansson L. Reversal of left ven- ertens 1998; 12: 511–516. tricular hypertrophy in hypertensive patients: a meta- 109 Doyle AE. Serotonin antagonists and vascular protec- analysis of 109 studies. Am J Hypertens 1992; 5: tion. Cardiovasc Drugs Ther 1990; 4: 13–18. 95–100. 110 Br Med J 1996; 312: May 18th, passim. 89 Himmelmann A, Svensson A, Bergbrant A, Hansson 111 Gibbs CR, Beevers DG versus Robertson JIS. Debating L. Long-term effects of losartan on blood pressure and nutrition: how strong is the evidence that recommen- left ventricular structure in essential hypertension. dations to reduce population average salt intake will J Hum Hypertens 1996; 10: 729–734. reduce mortality from cardiovascular disease? 90 Cheung B. Increased left-ventricular mass after losar- Nutrition Bull 1997; 22: 178–186. tan treatment. Lancet 1997; 349: 1743–1744. 112 Taubes G. The (political) science of salt. Science 91 Tedesco MA et al. Effects of losartan on hypertension 1998; 281: 898–907. and left ventricular mass: a long-term study. J Hum 113 Beretta-Piccoli C et al. Relationship of arterial press- Hypertens 1998; 12: 505–510. ure with body sodium, body potassium, and plasma 92 London GM. Large artery function and alterations in potassium in essential hypertension. Clin Sci 1982; hypertension. J Hypertens 1995; 13 (Suppl 2): 35–38. 63: 257–270. 93 O’Rourke M. Mechanical principles in arterial dis- 114 Midgley JA, Matthews AG, Greenwood CM, Logan A. ease. Hypertension 1995; 26: 2–9. Effect of reduced dietary sodium on blood pressure: 94 Safar ME, Siche J-P, Mallion J-M, London GM. a meta-analysis of randomized controlled trials. JAm Arterial mechanics predict cardiovascular risk in Med Ass 1996; 275: 1590–1597. hypertension. J Hypertens 1997; 15: 1605–1611. 115 Graudal NA, Galløe AM, Garred P. Effects of sodium 95 Safar ME, London GM, Asmar R, Frohlich ED. Recent restriction on blood pressure, renin, aldosterone, cat- advances on large arteries in hypertension. Hyperten- echolamines, cholesterol, and triglyceride: a meta- sion 1998; 32: 156–161. analysis. J Am Med Ass 1998; 279: 1383–1391. 96 Van Bortel LM, Kool M, Struiker-Boudier HM. Effects 116 Siani A et al. Controlled trial of long term oral potass- of antihypertensive agents on local arterial distens- ium supplements in patients with mild hypertension. ibility and compliance. Hypertension 1995; 26: 531– Br Med J 1987; 294: 1453–1456. 534. 117 Grimm RH et al. The inﬂuence of oral potassium 97 Mulvany MJ. Mechanical and other factors involved chloride on blood pressure in hypertensive men on in vascular injury related to hypertension. Blood a low-sodium diet. NEnglJMed1990; 322:569– Press 1994; 3 (Suppl 1): 11–17. 574. 98 Izzard AS, Heagarty AM. Hypertension and the vas- 118 Fotherby MD, Potter JF. Potassium supplementation culature: arterioles and the myogenic response. J reduces clinic and ambulatory blood pressure in eld- Hypertens 1995; 13:1–4. erly hypertensive patients. J Hypertens 1992; 10: 99 Heagerty AM, Bund SJ, Aalkjaer C. Effects of drug 1403–1408. treatment on human resistance arteriole morphology 119 Whelton PK et al. Effects of oral potassium on blood in essential hypertension: direct evidence for struc- pressure: meta-analysis of randomized controlled tural remodelling of resistance vessels. Lancet 1988; trials. J Am Med Ass 1997; 277: 1624–1632. ii: 1209–1212. 120 Mizushima S, Cappucio FP, Nichols R, Elliott P. Diet- 100 Schiffrin EL, Deng LY, La Rochelle P. Effects of a ary magnesium intake and blood pressure: a qualitat- beta-blocker or a converting enzyme inhibitor on ive overview of the observational studies. J Hum Hy- resistance arteries in essential hypertension. Hyper- pertens 1998; 12: 447–453. tension 1994; 23: 83–91. 121 Kawano Y, Matsuoka H, Takishita S, Omae T. Effects 101 Sihm I et al. Normalization of structural cardiovascu- of magnesium supplementation in hypertensive lar changes during antihypertensive treatment with a patients: assessment by ofﬁce, home, and ambulatory regimen based on the ACE inhibitor perindopril. blood pressures. Hypertension 1998; 32: 260–265. Blood Press 1995; 4: 241–248. 122 McCarron DA. Diet and blood pressure – the para- 102 Thybo NK et al. Effect of antihypertensive treatment digm shift. Science 1998; 281: 933–934. on small arteries of patients with previously 123 Birkett NJ. Comments on a meta-analysis of the

Journal of Human Hypertension The modern treatment of hypertension JIS Robertson S62 relation between dietary calcium intake and blood 133 Beilin LJ, Puddey IB, Burke V. Alcohol and hyperten- pressure. Am J Epidemiol 1998; 149: 223–228. sion – kill or cure? J Hum Hypertens 1996; 10 (Suppl 124 Stoto MA. Invited commentary on meta-analysis of 1): 1–5. epidemiologic data: the case of calcium intake and 134 Rimm EB, Klatsky A, Grobbee D, Stampfer MJ. blood pressure. Am J Epidemiol 1998; 148: 229– Review of moderate alcohol consumption and 231. reduced risk of coronary heart disease: is the effect 125 Cappucio FP et al. Epidemiologic association due to beer, wine, or spirits? BMJ 1996; 312: 731–736. between dietary calcium intake and blood pressure: 135 Robertson D et al. Effects of caffeine on plasma renin a meta-analysis of published data. Am J Epidemiol activity, catecholamines, and blood pressure. N Engl 1995; 142: 935–945. J Med 1978; 298: 181–186. 126 Cappucio FP, Elliott P, Follman D, Cutler JA. 136 Salvaggio A, Periti M, Miano L, Zambelli C. Associ- Authors’ response to ‘Comments on a meta-analysis ation between habitual coffee consumption and of the relation between dietary calcium intake and blood pressure levels. J Hypertens 1990; 8: 585–590. blood pressure’. Am J Epidemiol 1998; 148: 232– 137 Jennings G et al. The effects of changes in physical 233. activity on major cardiovascular risk factors, hemo- 127 Saunders JB, Beevers DG, Paton A. Alcohol-induced dynamics, sympathetic function and glucose utiliz- hypertension. Lancet 1981; ii: 653–656. ation in man: a controlled study of four levels of 128 Potter JF, Watson RDS, Skan W, Beevers DG. The activity. Circulation 1986; 73: 30–40. pressor and metabolic effects of alcohol in normo- 138 Ketelhut RG, Franz IM, Scholze J. Efﬁcacy and pos- tensive subjects. Hypertension 1986; 8: 625–631. ition of endurance training as a non-drug therapy in 129 Sleight P. Short term and long term effects of alcohol the treatment of arterial hypertension. J Hum Hyper- on blood pressure, cardiovascular risk, and all cause tens 1997; 11: 651–655. mortality. Blood Press 1996; 5: 201–205. 139 Shapiro AP et al. Non-pharmacologic management of 130 Gill JS et al. Stroke and alcohol consumption. N Engl hypertension. Ann N Y Acad Sci 1978; 304: 222–235. J Med 1986; 315: 1041–1046. 140 Skrabanek P. The Death of Humane Medicine and the 131 St. Leger AS, Cochrane AL, Moore F. Factors associa- Rise of Coercive Healthism. Social Affairs Unit: ted with cardiac mortality in developed countries London, 1994. with particular reference to the consumption of wine. 141 Australian Mild Hypertension Trial Management Lancet 1979; i: 1017–1020. Committee. Australian therapeutic trial in mild 132 Sharpe D. Coronary disease: when wine is red. Lancet hypertension: initial results. Clin Sci 1979; 57 (Suppl 1993; 341: 27–28. 5): 449–452.

Journal of Human Hypertension