Journal of Human Hypertension (1998) 12, 761–763  1998 Stockton Press. All rights reserved 0950-9240/98 $12.00 http://www.stockton-press.co.uk/jhh HYPERTENSION ILLUSTRATED The ‘Birmingham Hypertension Square’ for the optimum choice of add-in drugs in the management of resistant hypertension

GYH Lip, M Beevers and DG Beevers University Department of Medicine, City Hospital, Birmingham B18 7QH, UK

Around half of all patients with hypertension require more than one drug in order to bring their blood pressure under control.1–3 There remains, however, some uncertainty on the best combination therapies for these patients. When deciding which drugs to add in, the clinician should first ensure that appropriate non-pharmacological advice has been given and adhered to, and that any underlying cause for the hypertension has been detected. It is also worthwhile to ensure that the patient understands his or her drug regime and is actually taking any treatment as prescribed. The next consideration is whether the choice of first-line drug was logical in the first place. There is reasonable evidence that drugs which work, wholly or in part, by blocking the renin- angiotensin system, tend to be somewhat less effec- tive in patients whose circulating renin and angio- tensin levels are low.4–6 These drugs are the beta- receptor blockers, the angiotensin-converting enzyme (ACE) inhibitors and the angiotensin recep- tor antagonists. ‘Low renin’ states are found in the elderly, black/Afro-Caribbean patients and patients Figure 1 The ‘Birmingham Hypertension Square’ for the optimal with non-insulin-dependent diabetes mellitus choice of add-in drugs in the management of resistant hyperten- (NIDDM).7–10 The more rational first-line agents for sion. such patients are the thiazide diuretics or the cal- cium channel blockers (CCBs). However, there is with the dihydropyridine CCBs and by offsetting now evidence that the ACE inhibitors (and possibly tachycardia and sympathetic stimulation they may the angiotensin II receptor blockers) are specifically actually reduce side effects.15 There is good evi- indicated in patients with diabetes mellitus because dence of synergy between the thiazide diuretics and of their beneficial effects on nephropathy, retino- the beta-blockers, the ACE inhibitors16 as well as the pathy, left ventricular hypertrophy and the possible 17 11–14 angiotensin receptor antagonists. Similarly, the prevention of macro-vacsular end-organ damage. CCBs work well in combination with the classes of These beneficial effects may be independent of the drugs which block the generation or effects of angio- anti-hypertensive properties and indeed blood tensin II.18,19 By contrast there are some drug combi- pressure may be relatively uncontrolled. What then nations which are not particularly effective and are the best add-in drugs? some are unsafe. There is little synergy between thi- There are logical add-on therapies and less logical azide diuretics and the dihydropyridine CCBs.20,21 combinations. For example beta-blockers work well The addition of an ACE inhibitor to a beta-blocker also tends to be less effective because the beta-block- ers will already have blocked renin release.22,23 Correspondence: Professor DG Beevers In order to aid the clinician in sensible prescribing Hypertension illustrated GYH Lip et al 762 we have developed the ‘Birmingham Hypertension ing beta-adrenergic blockade in hypertensive patients. Square’ for add-in drugs in the management of Am Heart J 1976; 91: 634–642. hypertension (Figure 1). Clearly such a simplified 6Bu¨ hlar FR et al. Antihypertensive action of proprano- system cannot cover all eventualities. The clinician lol. Specific antirenin responses in high and normal should opt, however, to approach the square from renin forms of essential, renal, renovascular and malig- nant hypertension. Am J Cardiol 1973; 32: 511–522. any corner, with first-line drugs chosen logically. 7 Nussberger J, Waeber B, Brunner HR. Plasma angioten- The optimal second-line agents are immediately sin II and the antihypertensive action of angiotensin adjacent and indicated by the arrows. There are sev- converting enzymes. Am J Hypertens 1989; 2: 286–393. eral exceptions to the guidelines suggested by the 8 Saunders E. Hypertension in blacks. Med Clin North square. For example it provides guidance only for Am 1987; 71: 1013–1029. the use of the dihydropyridine CCBs. It is generally 9 Hall WD. Pathophysiology of hypertension in blacks. held that the phenylalkylamine CCB, verapamil, can Am J Hypertens 1990; 3: 366S–371S. be hazardous when used in combination with a beta- 10 Ferris JB. The causes of raised blood pressure in insu- blocker because of negative ionotropism and lin-dependent and non-insulin-dependent diabetes. chronotropism. It has, however, been reported, that J Hum Hypertens 1991; 5: 245–254. 11 Lewis EJ, Hunsicker LG, Bain RP, Rohde RD. The effect there is some synergy between verapamil and the 24 of angiotensin-converting-enzyme inhibition on dia- dihydropyridine CCB, nitrendipine. betic nephropathy. N Eng J Med 1993; 329: 1456–1462. The role of the alpha receptor blockers (eg, 12 Chaturvedi N et al. Effect of lisonopril on retinopathy doxazosin) and the centrally acting agents (eg, in normotensive people with type I diabetes. Lancet moxonidine, rilmenidine and methyldopa) in com- 1998; 351: 28–31. bination therapies is less certain and side effects 13 Cruickshank JM, Lewis J, Moore V, Dodd C. Reversi- may limit their use. There is some evidence that the bility of left ventricular hypertrophy by differing types alpha-blockers are synergistic with beta-blockers of antihypertensive therapy. J Hum Hypertens 1992; 6: and the ACE inhibitors although the side effect of 85–90. stress incontinence limits the use of alpha-blockers 14 Estacio RO et al. The effect of nisoldipine as compared 25 with enalapril on cardiovascular events in patients in women. We therefore do not routinely rec- with non-insulin-dependent diabetes and hyperten- ommend the use of these drugs as the first choice of sion. N Engl J Med 1998; 338: 645–652. ‘add-in’ drugs at the present state of knowledge. 15 Nifedipine Atenolol Study Review Committee. Nifedi- The choice of third-line agents for use with the pine and atenolol singly and combined for treatment Birmingham Hypertension Square is more the sub- of essential hypertension, comparative multicentre ject of speculation, anecdote and little reliable trial study in genreal practice in the United Kingdom. Br evidence. At this stage, in resistant hypertension, Med J 1988; 296: 468–472. where the outlook is more closely related to the 16 Zezulka AV et al. Comparison of enalapril and bendro- achieved blood pressure on therapy rather than the fluazide for treatment of systolic hypertension. Am J initial severity of the hypertension,26,27 the Cardiol 1987; 59: 630–633. Birmingham Hypertension Square may provide 17 Soffer BA et al. Effects of Losartan on a background some guidance. of hydrochlorothiazide in patients with hypertension. Hypertension 1995; 26: 112–117. The choice of rational first and second-drugs 18 Singer DRJ, Markandu N, Shore AC, MacGregor GA. using our system, could lead to some financial sav- Captopril and nifedipine in combination for moderate ings for health care resources. For example, in an to severe essential hypertension. Hypertension 1997; elderly Afro-Caribbean patient, the use of enalapril 9: 629–633. with atenolol added in could cost UK £19.91 per 19 Brouwer RML et al. Antihypertensive treatment using month and will be relatively ineffective.28 The sub- calcium antagonists in combination with captopril stitution of bendrofluazide alone at UK £0.15 per rather than diuretics. J Cardiovasc Pharmacol 1990; month may well be more effective, and more accept- 39: 327–332. able. If a thiazide diuretic drug cannot be used then 20 Zezulka AV, Gill JS, Beevers DG. The effects of bendro- a CCB used alone is a reasonable alternative. flumethiazide added to nifedipine in patients with hypertension. J Clin Pharmacol 1987; 27: 41–45. 21 MacGregor GA, Cappuccio FP. Lack of effect of a diuretic added to diltiazem. J Hum Hypertens 1997; References 11: 249. 22 MacGregor GA et al. Captopril in essential hyperten- 1 MRC Working Party. MRC trial of treatment of mild sion: contrasting effects of adding hydrochlorothiazide hypertension. Principal results. Br Med J 1985; 291: or propranolol. Br Med J 1982; 284: 693–696. 97–104. 23 Frishman WH et al. A multi factorial trial design to 2 SHEP Cooperative Research Group. Prevention of assess combination therapy in hypertension. Arch stroke by antihypertensive drug therapy in the persons Intern Med 1994; 154: 1461–1468. with isolated systolic hyperension. JAMA 1991; 265: 24 Nalbantgil I, Onder R, Kiliccioglu B, Turkoglu C. Com- 3255–3264. bination therapy with verapamil and nitrendipine in 3 Avanzini F et al. Antihypertensive efficiency and tol- patients with hypertension. J Hum Hypertens 1993; 7: erability of different drug regimes in isolated systolic 305–308. hypertension in the elderly. Eur Heart J 1994; 14: 25 Marshall HJ, Beevers DG. Alpha-adrenergic blocking 206–212. drugs and female urinary incontinence: prevalence 4 Johnston CI, Burrell LM. Evolution of blockade of the and reversibility. Br J Clin Pharmacol 1996; 42: 507– renin-angiotensin system. J Hum Hypertens 1995; 9: 509. 375–380. 26 Isles CG et al. Mortality in patients of the Glasgow 5 Amery A et al. Mechanism of hypotensive effect dur- Blood Pressure Clinic. J Hypertens 1986; 4: 141–156. Hypertension illustrated GYH Lip et al 763 27 Beevers DG, Fairman MJ, Hamilton M, Harpur JE. The 28 British National Formulary 1998 No. 35. British Medi- influence of antihypertensive treatment on the inci- cal Association and Royal Pharmaceutical Society of dence of cerebral vascular disease. Postgrad Med J Great Britain. 1973; 49: 905–909.