Efficacy and toxicity of antihypertensive pharmacotherapy relative to effective dose 50

Abstract

Antihypertensive drugs have usually been approved at doses near the top of their respective dose–response curves. Efficacy plateaus but adverse drug reactions (ADRs), such as falls, cerebral or renal ischaemia, increase as dose is increased, especially in older patients with comorbidities. ADRs reduce adherence and may be difficult to ascertain reliably. Higher doses have generally not been shown to reduce total mortality, which provides a summary of efficacy and safety. Weight loss and other lifestyle measures are essential and may be sufficient treatment in many young and low risk patients. Most antihypertensive drug lower systolic blood pressure by around 10 mmHg, which reduces stroke and heart failure by about a quarter. Clinical trials have not been designed to demonstrate specific blood pressure treatment thresholds and targets, which are mostly extrapolated from epidemiology. Mean population oral effective dose 50 may be the most appropriate dose at which to commence antihypertensive drugs. The dose can then be titrated up if greater efficacy is demonstrated, or lowered if ADRs develop. Lower dose combination therapy may best balance benefit and harms with fewer ADRs and additive, potentially synergistic, efficacy.

1. INTRODUCTION

Systemic hypertension has a lifetime prevalence of >65%.1 However, aspects of antihypertensive pharmacotherapy remain controversial despite extensive research, intensive drug development and multiple guidelines. Problematic areas include patient adherence,2 reliable identification of patients at higher cardiovascular risk, accuracy of blood pressure (BP) measurement, limited agreement on treatment thresholds and targets and disagreement about the merits of different classes of antihypertensive drugs. Adverse drug reactions (ADRs) are common, are generally dose‐ and time‐related,3 are easily overlooked, must be balanced against potential benefits and may reduce patient adherence to treatment.2 Smaller doses are often better tolerated.3 Efficacy at lower than marketed doses may be maintained because dose–response over the approved dose range of many antihypertensive drugs is nearly flat (Figure 1).4 When dose is plotted linearly rather than logarithmically,5 efficacy can be seen to plateau at lower doses than may be appreciated.

Figure 1.

Drug dose response, idealized plots (e.g. metoprolol). Efficacy: continuous line; adverse drug reactions: broken lines. BP, blood pressure; ED50, (estimated) effective dose 50, the mean population oral dose required to produce 50% of the maximal effect

2. RATIONALE FOR TREATMENT OF HYPERTENSION

Systemic hypertension is the most important modifiable risk factor for cardiovascular morbidity and mortality but is multifactorial, including genetic6 and psychosocial7 factors. Lifestyle management, including weight loss, reduced salt intake and increased fitness, can lower BP8 and may be sufficient in younger, lower risk patients for whom annual benefits of antihypertensive drugs are often smaller compared to the effects of ADRs.9 Benefits from antihypertensive drugs are potentially greatest after patients develop symptomatic cardiac10, 11, 12, 13 or cerebrovascular14 disease, diabetes mellitus15 or aneurysm.16 Asymptomatic patients at higher than average risk may be identified by abnormal electrocardiography related to coronary or hypertensive disease, brain imaging evidence of previous stroke, or biochemical evidence of proteinuria or renal failure.

Optimal antihypertensive drug dose is difficult to determine. Drug requirements may be individualized on the basis of improvements in symptoms such as angina, dyspnoea, oedema or headache and findings of heart failure and retinopathy, but must be balanced against adverse effects, including on‐target ADRs such as faintness, fatigue and decline in renal function, to which the elderly may be more prone because of comorbidities.

One of the most consistent benefits of long‐term antihypertensive pharmacotherapy is the lower incidence of stroke (Table 1). However, no additional reduction in stroke has been shown with higher intensity treatment, except in diabetics (Table 2)17 and unduly low BP may increase the risk of stroke, especially in the elderly.18 Although lower BP is associated with less microalbuminuria,19, 20 improved renal outcomes are conspicuously absent in published randomised controlled trials (RCTs).20

Table 1.

Antihypertensive clinical trials by blood pressure reduction. Randomised controlled trials of 1–3 antihypertensive drug/s vs placebo studies (on no concurrent antihypertensive treatment) that recruited >1000 participants

Study (y completed)	n	Durationn (y)	Drug/s	Δ SBP (mean on‐trial) mmHg	LVF	AMI	Stroke	Mortality
								Total	CV
					(active/placebo)
				<10
PATS 199556	5665	1.8	Indapamide	−5.0	‐	25/21a	159/217***	146/158	(87/101)
SYSTCHINA 199757	2394	3.0	Nitrendipine +/− captopril Hydrochlorothiazide	−8.0c	4/8	9/7	45/59*	61/82 **	(33/44*)
ANBPS 197958	3427	4.0	Chlorothiazide +/− methyldopa, Propranolol, Pindolol	~ −8 (DBP ‐5.6)	3/3	33/33	17/31*	25/35	(8/18*)
Totals:					7/11	67/61	221/307	232/275	(128/163)
Mean reduction (weighted)				−6.5	‐b	‐b	−28%	−16%	(−21%)
				>10
SYST‐EUR 199759	4695	2.6	Nitrendipine Enalapril Hydrochlorothiazide	−10.4	37/49	33/45	47/77**	123/137	(59/77)
MRC 198260	17 354	5.0	Bendrofluazide or propranolol +/− methyldopa	−11.5	‐	222/234a	60/109**	157/181	(104/112)
SHEP 199061	4736	4.5	Chlorthalidone Atenolol	−12.0	56/109**	65/100*	106/163***	213/242	(90/112)
MRC‐elderly 199162	4396	5.8	Atenolol or Hydrochorothiazide + Amiloride	−15.0c	‐	128/159a	101/134*	301/315	(161/180)
STOP 199139	1627	2.1	Atenolol Hydrochlorothiazide Amiloride (metoprolol, Pindolol)	−19.0c	19/39	25/28	29/53**	36/63 **	(13/29)
Totals:					112/197	473/566	343/536	830/938	(427/510)
Mean reduction (weighted)				−12	−43%	−16%	−36%	−12%	(−16%)

^aAny coronary events (total).

^binsufficient data to comment.

^cSBP reduction @ end of study.

^* P < .05,

^** P < .01,

^*** P < .001.

AMI, acute myocardial infarction; CV, cardiovascular; LVF, left ventricular failure; SBP, systolic blood pressure.

CV mortality is in parentheses because its definition varied between studies.

Table 2.

Ten trials comparing usual vs high intensity antihypertensive pharmacotherapy (listed by order of extent of SBP reduction)

Trial (n)	Study duration (y)	AMI (%)	Stroke (%)	CV mortality (%)	Total mortality (%)	Renal failure (%)	Syncope/hypotension (%)	End of study SBP (Δ)
CAR‐SIS 2008 (1111)63	2.0
Usual		1.1	1.6	‐	0.9	‐	‐	136
Tight		0.7	0.7		0.7			132
		NS	NS		NS			4
HOT 1997 (18 990)64	3.8
≤ 90		1.3	1.5	1.4	3.0	‐	‐	144
≤ 85		1.0	1.8	1.4	3.1			141
		P = .05	NS	NS	NS			4
≤ 85		1.0	1.8	1.4	3.1			141
≤ 80		1.0	1.4	1.5	3.3			140
		NS	NS	NS	NS			1
VALISH 2008 (3260)65	2.9
Moderate		0.3	1.5	0.7	2.0	0.1	‐	142
Strict		0.3	1.0	0.7	1.6	0.3		137
		NS	NS	NS	NS	NS		5
JATOS 2004 (4418)66	2.0
Mild		0.3	2.4	0.3	2.1	0.5	‐	146
Strict		0.3	2.5	0.4	2.4	0.4		136
		NS	NS	NS	NS	NS		10
UKPDS 1996 (1148)a 15	8.4
Less tight		17.7	9.5	13.8	21.3	2.6	‐	154
Tight		14.1	5.0	9.1	17.7	1.3		144
		NS	P = .01	NS	NS	NS		10
AASK 1993 (1094)67	4.0
Usual		No	No	0.7	1.9	No	5.2	138
Lower		Change	Change	0.6	1.6	Change	6.3	128
				NS	NS		NS	10
BBB 1993 (2127)68	4.9
Unchanged		1.8	1.1	‐	‐	‐	‐	152
Intensified		2.0	0.8					141
		NS	NS					11
SPS3 2012 (3020)69	3.7
Higher target		0.7	2.8	0.7	1.7		0.11	138
Lower target		0.6	2.2	0.6	1.8		0.22	127
		NS	NS	NS	NS		NS	11
ACCORD 2009 (4733)a 41	4.7
Standard		6.2	2.6	2.4	6.1	0.0004	0.2	133
Intensive		5.3	1.5	2.5	6.3	0.002	1.2	119
		NS	P = .01	NS	NS	NS	P < .001	14
SPRINT 2014 (9361)b 17	3.3	(LVF)
Standard		2.5	1.5	1.4	4.5	2.5	4.4	136
Intensive		2.1	1.3	0.8	3.3	4.1	6.9	121
		NS	NS	P = .005	P = .003	P < .001	P < .01	15

^aAll participants had type 2 diabetes mellitus.

^bMore than a third of participants were over 70, over a third obese and over a quarter had renal failure, LVF fell by over a third, from 2.1 to 1.3% (p = 0.002), which appears likely to account for the reduction in CV and total mortality, at the price of about a 70% increase in renal failure and hypotension.

AMI, acute myocardial infarction; CV, cardiovascular; LVF, left ventricular failure; NS, not significant; SBP, systolic blood pressure.

Systemic hypertension is the most prevalent contributing risk factor in left ventricular failure (LVF).21 Several antihypertensive agents—angiotensin converting enzyme inhibitors (ACEIs), β‐adrenoceptor blockers and spironolactone—reduce systolic BP (SBP) by 10 mmHg or less but each reduced total mortality, overall survival, which provides a summary of efficacy and safety, by 11–30%10, 11, 12 in patients with a history of LVF. Total mortality was not reduced on higher, compared to lower, doses.22 The 3 different drug classes when used together may improve survival substantially.

Consideration should be given to the number of patients required to be treated for 1 year to avoid 1 clinical event (number needed to treat to benefit, NNTB) vs drug tolerability, a variety of ADRs and thereby quality of life, which are dose‐related23 and the same whether a drug is being used for treatment or prevention. Antihypertensive drugs are used as treatment in symptomatic LVF but only for prevention in hypertension. The NNTB is around 3 when an ACEI is used as treatment in patients with symptomatic LVF24 but >130 when used in prevention in asymptomatic patients at high risk of cardiovascular disease25 where the higher risk to benefit ratio may predicate lower doses.

3. DIFFERENCES AMONG DRUG CLASSES

In systemic hypertension, there is substantial evidence for reductions in cardiovascular and total mortality with ACEIs,26 thiazide diuretics27 and calcium channel blockers.28, 29 Although often recommended for symptomatic heart disease, β‐blockers appear to have less impact on cardiovascular events in hypertension,30, 31 quite possibly because the slower heart rate at higher doses increases stroke volume and thereby SBP.32 Beta‐blockers have less impact on central aortic pressure.33 No reduction in total mortality has ever been reported in hypertension or LVF, despite their capacity to lower BP, with angiotensin receptor blockers,34, 35 α‐adrenoceptor blockers, furosemide, methyldopa, clonidine, moxonidine, hydralazine or minoxidil.

4. LIMITED EVIDENCE FOR IMPROVED OUTCOMES WITH HIGHER INTENSITY TREATMENT

Meta‐analyses focus mainly on efficacy and include multiple heterogeneous RCTs which employ fixed or highest tolerated dose. At least half of the observed BP reduction appeared to be due to regression to the mean.36 A mean reduction of SBP compared to placebo in meta‐analyses of around 10 mmHg37, 38 was associated with a relative risk reduction of stroke and LVF by about a quarter.

Most RCTs have been very short‐, short‐ or intermediate‐term evaluations of BP lowering with single antihypertensive drugs37, 38 at highest tolerated dose particularly in the interests of marketing. This information is then used in approvals for human clinical use. Large long‐term placebo‐controlled RCTs on 1–3 drugs, in patients on no antihypertensive pharmacotherapy at baseline, demonstrated reductions of total mortality by a weighted mean of up to 16% with a reduction in SBP by a mean of only about 6 mmHg (Table 1). There is a paucity of studies that address outcomes in patients randomized to different doses of an antihypertensive drug. Long‐term antihypertensive drug treatment, even in combinations, has never been shown to lower SBP by more than a mean of 19.5 mmHg, compared to placebo.39 The benefits of pharmacological reduction of BP in RCTs plateau with increased intensity of treatment (Tables 1,2) whilst ADRs increase (Table 2).40

Higher intensity compared to standard antihypertensive pharmacotherapy appears to reduce cardiovascular events only minimally (Tables 1,2), except in diabetes.15, 41 In SPRINT,17 total mortality was significantly lower on more intensive compared to standard antihypertensive pharmacotherapy, mainly attributable to a reduction in LVF, which was prevalent because of the advanced age, high BMI and high prevalence of renal impairment in that study. The NNTB to prevent 1 death was 270, at the cost of increased ADRs, including syncope for which the number needed to treat to cause harm was 100 and decline in renal function with a number needed to treat to cause harm of 56. There was no improvement in total mortality in 9 previous large RCTs which randomized participants to more intensive therapy and lowered SBP 4‐15 mmHg, compared to standard therapy (Table 2).

BP targets in treatment guidelines vary, perhaps because they are partly extrapolated from epidemiological studies, which show, for example that cardiovascular events, on no drug treatment, increase above a SBP of 115 mmHg.42 No published RCTs of drug treatment were designed to establish the optimum target BP. Potential benefits with lower BP on antihypertensive drugs need to be weighed against ADRs, which are dose‐ and time‐related.3 Specific threshold or target BPs may not be appropriate given the problematic nature of BP measurement, which is commonly confounded by anxiety, hydration, posture, temperature, illness43 and regression to the mean. Direct intra‐arterial and indirect sphygmomanometer BP measurement often differ significantly, the latter usually an underestimate, to an extent that varies between patients.44

5. ORAL ED50 CAN GUIDE TREATMENT

Clinicians may favour higher doses of antihypertensive drugs because they are concerned about high cardiovascular risk and persuaded by competitive marketing and aimed at treating to target. Higher doses may add redundancy to cover heavier patients, imperfect compliance and variation in pharmacokinetics. However, smaller doses of effective drugs may often suffice.

The oral effective dose 50 (ED50) can be determined from published drug dose response data. The older definition of ED50 was the dose required to achieve remission of a symptom or disease (potentially difficult to define and standardize) in 50% of treated patients. The definition of ED50 we use here is the mean population dose that causes a reduction in SBP of 50% of the maximal effect. The maximal effect for most antihypertensive drugs is a SBP reduction of 10–15 mmHg45, 46 and so ED50 (Table 3) is usually around 6 mmHg, a seemingly modest reduction but with which in the early RCTs the maximum reductions in total mortality were observed (Table 1). Published RCTs usually employed antihypertensive drug doses 2‐ to 8‐fold higher than ED50. Doses of diuretics and calcium channel blockers have since fallen to near ED50 (Table 3), presumably in part because of ADRs.

Table 3.

Doses of antihypertensive drugs

Daily dose
Drug	Maximum reported (mg)	Approved (mg)	Mean population oral ED50 dose (estimateda; mg)
Diuretics:
Hydrochlorothiazide 46	100	12.5–50	10
Frusemide 70	500	20–40	10
Indapamide 71	5	1.25–1.5	1
Spironolactone 72	200	25–100	40
Eplerenone 73	50	25–50	20
Beta‐blockers:
Metoprolol 74, 75	400	50–100	60
Propranolol 74, 75	320	10–160	80
Atenolol 75	200	25–100	30
Bisoprolol 76	40	1.25–10	5
Calcium channel blockers:
Verapamil 77	480	40–240	180
Diltiazem 78	540	60–360	100
Felodipine 79	20	2.5–10	11
Lercanidipine 80	20	10–20	5
Amlodipine 81	20	5–10	2
Angiotensin converting enzyme inhibitors 45:
Captopril	300	12.5–50	20
Lisinopril	80	2.5–20	10
Enalapril	40	2.5–20	5
Perindopril	32	2–8	4
Ramipril	20	1.25–10	3
Hydralazine 82	400	25–50	60, 130b
For comparison:
Simvastatin 50	160	10–80	15
Rosuvastatin 50	40	5–40	1
Aspirin 83	300	75–100	25
Metformin 84	4000	500–1000	2000
Dapagliflozin 84	20	5–10	4

^aBased on short‐term studies.

^bHigher ED50 in fast‐acetylators, who make up about half of the population.

ED50, (estimated) effective dose 50, the mean population oral dose required to produce 50% of the maximal effect.

Approved doses around ED50 of various drugs (Table 3) have been shown to reduce clinical events and mortality substantially, for example aspirin 30–75 mg47, 48 felodipine 5 mg,49 simvastatin 40 mg50 and metformin 1–4 g, daily.51 A reduction in total mortality of 16% with antihypertensive treatment may seem less than expected when compared to the large reductions in clinical events (Table 1) but not all deaths were cardiovascular and ADRs may contribute to mortality. Higher doses have limited additional impact on efficacy and mainly increase ADRs (Table 2).

6. POLYPILL

Rather than increase dose, an effective strategy may to be to combine different classes of antihypertensive drugs at lower doses4 but RCT data published to date52, 53, 54, 55 are modest and not long term. Various antihypertensive drug combinations of 2 or 3 drugs are already marketed but cardiovascular outcomes on different combinations and doses have not been evaluated, in part because of constraints by drug patents and logistics. A polypill can address different aspects of hypertension pathophysiology. Lower doses in a polypill4 combine the benefit of additive and possible synergistic efficacy with fewer ADRs from each drug. However, synergistic effects between some drugs may increase ADRs, which may be avoided with smaller doses.

7. CONCLUSION

Antihypertensive drugs can improve survival and have been shown to reduce LVF and stroke substantially. Over‐prescribing may compromise safety, tolerability and compliance, with little additional efficacy. Across the approved dose range of most antihypertensive drugs, efficacy plateaus but multiple ADRs continue to increase. Effective lifestyle management along with combined pharmacotherapy may help make higher antihypertensive drug doses unnecessary. Estimated ED50 may be the most appropriate dose at which to commence. The dose can then be increased if greater efficacy proves necessary and is demonstrated, and reduced if ADRs emerge.

7.1. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY,85 and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18.86, 87, 88

COMPETING INTERESTS

There are no competing interests to declare.

Dimmitt SB, Stampfer HG, Martin JH, Ferner RE. Efficacy and toxicity of antihypertensive pharmacotherapy relative to effective dose 50. Br J Clin Pharmacol. 2019; 85: 2218–2227. 10.1111/bcp.14033