The aim of the present work was to review published studies investigating the dose-related efficacy on blood pressure (BP) of olmesartan and of other commercially available angiotensin II type I receptor blockers (ARBs). Patient population comprises mild to moderate hypertensive adult patients. We selected studies with comparable design and dose ranges. Dose–effect relationship plots were fitted for diastolic (DBP) and systolic (SBP) BP to the simplified Emax model. We also examined controlled studies of olmesartan vs. other individual ARBs. Our overview was based on 7280 patients, of which 5769 received an ARB and 1511 received placebo. Except for losartan, the data fitted correctly to the Emax model, with correlation coefficients ranging from 0.77 to 0.99. BP-lowering efficacy defined as Emax was superior with olmesartan, (DBP/SBP mmHg: −9.0/−12.4) when compared with candesartan (−6.7/−11.3), irbesartan (−6.5/−11.2) and valsartan (−6.3/−8.9). Head-to-head comparisons of olmesartan to each of the other ARBs used at per-label ‘recommended doses’, support the finding of a greater BP-lowering effect of olmesartan. This overview suggests that clinically relevant differences in maximal efficacy, as well as in efficacy of per-label recommended doses can be evidenced among individual ARBs. Olmesartan efficacy was consistently at the highest end of the range of efficacy of ARBs studied.
Antihypertensive therapy is effective in preventing target organ damage and cardiovascular morbidity and mortality. Although still a matter of debate, it is generally agreed that the main mechanism of cardiovascular prevention with antihypertensive agents is related to their blood pressure (BP)-lowering potency [1–3]. Dose–BP-lowering relationship is not well defined for a number of individual antihypertensive agents.
Meta-analysis can be useful in determining dose–response relationships for individual drugs. However, the interpretation of the results of such analyses is hampered by a number of methodological limitations . Randomized, prospective, double blind, controlled studies are the most accurate means of comparing the BP-lowering efficacy between drugs.
The introduction of angiotensin II type I (AT1)-receptor blockers (ARB) was an important breakthrough, offering effective control of BP with placebo-like tolerability, together with effective protection against cardiovascular complications [5,6]. So far, the available evidence suggests that significant clinically relevant differences in BP-lowering efficacy and duration of action are apparent among individual agents from this class of drug [7–14].
The relationship between dose and antihypertensive effect for different ARBs has been reviewed by Hansson . Since then, new studies have been published and new data with recent ARBs have become available .
Olmesartan medoxomil is a highly potent angiotensin II receptor antagonist that selectively and competitively inhibits the angiotensin II type 1 receptor. It has been subjected to a number of comparative studies with other ARBs (valsartan, losartan, candesartan and irbesartan) [17,18].
The aim of the present work was to review published studies investigating the dose-related efficacy of olmesartan and of other commercially available ARBs. We first selected the studies with comparable design, follow-up duration, patient population, BP measurements methods, and dose ranges; we then constructed average dose–effect relationship plots for each individual ARB. Besides, we also examined the published studies with randomized controlled, double-blind head-to-head comparison of olmesartan vs. other individual ARBs.
Among available ARBs, we focused in the present overview only on those ARBs with available published data from head-to-head comparison with olmesartan, namely, candesartan, irbesartan, losartan, and valsartan as well as from placebo-controlled studies with these agents.
Pharmacologically, the dose of a drug required to produce a specified effect in 50% of the population is the ‘median effective dose’ . Thus, regarding antihypertensive agents, it is generally agreed that the dose of a drug required to produce a response rate of 50% of a population with mild to moderate hypertension, defined as a diastolic blood pressure (DBP) drop of 10 mmHg and/or DBP lowering to <90 mmHg, is the ‘median effective dose’.
During drug clinical development, the dose–response relationship has to be established to determine the clinically useful dose-range as well as the ‘optimal dose’ , which could be defined as the lowest maximally effective dose that should provide maximal efficacy without excessive amounts of the agent .
Maximal effective doses for antihypertensive agents are doses close to the top of the dose–response range, beyond which only little additional BP effect can be obtained by further increases in the dose . Additional efficacy, obtained with further dose increments, is usually minor leading to a plateau in the dose–effect relationship curve.
For each individual ARB and based on these definitions, we selected the recommended maintenance dose (RMD). Other studied doses were then expressed as fractions or multiples of the RMD and plotted against their effect on BP so as to obtain dose effect relationship curves. This allowed comparing ‘normalized’ dose–effect relationships. RMDs for the selected ARBs in the present overview were as follows: olmesartan 20 mg, candesartan 8 mg, irbesartan 150 mg, losartan 50 mg and valsartan 80 mg, once daily.
Selection of overviewed studies
We selected studies with comparable study design, study duration, patient population, BP measurements methods, and dose ranges.
Placebo-controlled studies of individual ARBs
All studies were prospective, multicenter, randomized, double-blind, placebo-controlled parallel-group design, and conducted in hypertensive adult patients, over 18 years of age. Hypertension was essential, mild-to-moderate, defined as a mean seated or supine DBP ranging from 95 to 120 mmHg. All studies had an initial single-blind placebo run-in period of 3–5 weeks. The treatment period was of 8 weeks, to ensure that most of the effect of the drug was achieved . In case of longer study duration, only data available at the 8th week visit were selected. Drugs were administered once a day in the morning.
Trough seated or supine BP measurements were obtained 24 h after the last drug intake, after a 5-min rest in the supine or sitting position, by the same clinician, in the same arm, using a same standard, calibrated mercury sphygmomanometer. The mean of three consecutive measurements was calculated. Efficacy assessment was based on the reduction of seated or supine BP in mmHg at trough. Baseline characteristics for each study are presented in Table I.
Table I. Baseline characteristics of patients in placebo-controlled studies.
aWith assessment at 8 weeks (7) or 6 weeks (1).
bWith assessments at 8 weeks.
cInclusion criteria; average values not documented.
The dose–response curve of olmesartan has been defined from an integrated analysis of seven studies carried out in 3055 patients who received either placebo or olmesartan medoxomil 2.5, 5, 10, 20, 40 and 80 mg over 6–52 weeks. Data provided are those reported at 8 weeks for 2511 patients treated with olmesartan medoxomil and 544 with placebo .
Data were extracted from a single study including 365 patients who received either candesartan cilexetil 2, 4, 8, 16 or 32 mg or placebo (64 patients) for 8 weeks .
Data were extracted from a report of an integrated analysis of eight dose-ranging studies of similar design assessing the efficacy of irbesartan over the dose range of 1–900 mg conducted in 2631 hypertensive patients randomized to either irbesartan maleate or placebo treatment during 6–12 weeks . In this report, the analysis was performed on BP data at 8 weeks for seven studies and at 6 weeks for one study (as 8 weeks data were unavailable), in 1954 irbesartan-treated patients and 677 placebo-treated patients.
Results were obtained from a single 8-week study in which 576 hypertensive patients were randomly assigned to losartan potassium 10, 25, 50, 100 or 150 mg, or placebo or enalapril (83 patients not taken into account in this analysis) .
Assessments came from a study comparing valsartan 20, 80, 160 and 320 mg to placebo, in 736 patients .
Fitting of the dose–BP lowering effect relationship
Doses were standardized as fractions of the RMD to allow direct comparisons between ARBs. Placebo-subtracted mean reductions from baseline presented in published studies (Table II) were fitted for DBP and SBP to the simplified Emax model:
Table II. Blood pressure response at 8 weeks in placebo-controlled studies.
Δ: Means change from baseline. Dose in bold: recommended maintenance dose.
Predicted Emax values for Seated SBP (mmHg) at trough
where Ei and Pi are the mean BP reductions from baseline at dose Di for treatment E and placebo P, Emax the maximal effect and ED50 the median effective dose, i.e. the dose required to produce 50% of the maximal effect. This model corresponds to the classical sigmoid log(dose)–response curve (Hill curve) for a receptor-antagonist drug. Reeves et al.  used that model to compute Emax from individual data and did not find any significant difference with the general model, which includes a ‘steepness’ coefficient, the Hill parameter. The maximal effect Emax was extrapolated from the Lineweaver–Burk regression 1/(E − P) vs. 1/D as the inverse of the intercept of the curve:
when ED50/(EmaxDi) = 0, i.e. for an infinite dose D. As the Lineweaver–Burk regression gives the highest weight to the lowest and least precise doses, the four lowest doses (1, 5, 10 and 25 mg, corresponding respectively to 1/150, 1/30, 1/15 and 1/6 RMD) of irbesartan were excluded from the calculations to fit the model correctly. Dose–effect curves were then drawn using the parameters derived from the Lineweaver–Burk regression for each ARB.
Head-to-head comparison of olmesartan vs. other ARBs
Seven studies have been published providing head-to-head comparison of olmesartan vs. other antihypertensive agents. However, only three clinical trials testing the 20 mg RMD of olmesartan to the RMD of other ARBS, in monotherapy, were identified (Table III).
Table III. Blood pressure response in comparative studies of olmesartan vs. other ARBs.
Olmesartan vs. losartan
Olmesartan vs. losartan, valsartan, and irbesartan
The first comparative study focused on the efficacy of losartan 50–100 mg and olmesartan 10–20 mg in a 12-week titration dose approach , in 316 patients with mild to moderate hypertension. The initial dose (50 and 10 mg, respectively) was doubled at week 4 in non-responders.
The second study compared the antihypertensive efficacy of olmesartan (20 mg), losartan (50 mg), irbesartan (150 mg), and valsartan (80 mg) in an 8-week fixed-dose study, in 588 hypertensive patients. Although the study used also ambulatory BP measurements (ABPM), the primary endpoint was casual BP measurements, which were reported in this review .
The third study compared olmesartan 20 mg to candesartan 8 mg in an 8-week ABPM study of 645 patients with mild to moderate hypertension . Our analysis was carried out using conventional casual trough seated blood measurements.
Overall, our overview was based on 7280 patients, of which 5769 subjects received an ARB and 1511 received placebo. Mean reductions in trough diastolic and systolic seated BP, obtained from baseline to 8-week therapy are summarized in Table II. These results were used to establish the dose–response relationship curves (Figures 1 and 2).
Olmesartan produced a true dose-dependent decrease in trough sitting DBP and all doses, i.e. from 2.5 to 80 mg, of olmesartan were significantly more effective than placebo (P ≤ 0.001) . The DBP decrease was greater with the RMD of olmesartan (20 mg: −5.6 mmHg) than with irbesartan (150 mg), losartan (50 mg) and valsartan (80 mg), and similar to that observed with the candesartan RMD of 8 mg (−6.1 mmHg).
Data observed with SBP corroborate the DBP results. The RMD of 20 mg olmesartan produced a greater SBP reduction compared with all other studied ARBs RMD . Dose–response curves drawn for each agent showed a greater efficacy of olmesartan as evidenced by a higher plateau of maximal DBP as well as SBP lowering, except for candesartan which produced similar SBP results.
Maximal BP reductions could be reliably estimated for all agents, except losartan. The data from four ARBs (olmesartan, candesartan, irbesartan and valsartan) were correctly fitted to the Emax model, with correlation coefficients ranging from 0.77 to 0.93 for DBP and 0.84 to 0.99 for SBP (corresponding determination coefficients: 60–97% and 70–98%). This confirms the validity and the reliability of the model we used to draw dose relationship curves. The distribution of the data from losartan was found too atypical to be fitted (DBP/SBP: correlation coefficients 0.39/0.72, determination coefficients 15/52%) and the derived Emax are meaningless. The Emax for DBP/SBP were −9.0/−12.4 for olmesartan, −6.7/−11.3 for candesartan, −6.5/−11.2 for irbesartan and −6.3/−8.9 for valsartan.
Comparative studies of olmesartan vs. other ARBS
Diastolic and systolic BP changes observed from baseline to the end of study were not comparable across studied ARBs (Table III; Figures 3 and 4).
In the first study  comparing losartan 50–100 mg to olmesartan 10–20 mg, the DBP reduction was significantly greater (P < 0.05) in patients who received olmesartan (−10.6 mmHg) compared with those who received losartan (−8.5 mmHg). Further, the SBP reduction was also significantly more pronounced with olmesartan (−14.9 mmHg) than with losartan (−11.6 mmHg) at week 12 (P ≤ 0.05). The mean SBP/DBP difference was of −3.3/−2.1 mmHg. In addition, fewer patients receiving olmesartan required to be titrated to the higher dose (42%) than patients treated with losartan (63%).
A better BP lowering achieved by the olmesartan regimen was also observed in the clinical trial comparing olmesartan with irbesartan, losartan, and valsartan : seated DBP reduction after a 8-week therapy was −11.5 mmHg for olmesartan, while it was −8.2 mmHg for losartan, −7.9 mmHg for valsartan and −9.9 mmHg for irbesartan. The mean reduction in cuff DBP from baseline with the RMD of 20 mg for olmesartan was significantly greater than with the maintenance doses of losartan (50 mg) by 3.3 mmHg (P < 0.0005), of valsartan (80 mg) by 3.6 mmHg (P < 0.0005), and irbesartan (150 mg) by 1.6 mmHg (P < 0.05). Although not statistically significant, the reduction in the mean SBP was greater in patients receiving olmesartan than in those receiving losartan and irbesartan .
A secondary analysis  based on ABPM results showed a greater mean reduction in BP with olmesartan than with valsartan for all ABPM times analyzed (24 h, daytime, night-time and last 2 and 4 h of monitoring). The superiority was also statistically significant vs. losartan at a number of time intervals, and vs. irbesartan for SBP in the last 4 h of monitoring. Finally, olmesartan maintained better BP lowering during the morning period .
In the third head-to-head comparison, olmesartan 20 mg was more effective than candesartan 8 mg in lowering DBP at 8 week (−15.8 mmHg vs. −5.1 mmHg); of note, the difference was significant for the primary criteria, i.e. daytime DBP (by −1.55 mmHg) and for 24-h DBP, as well as for the daytime SBP .
The present overview of placebo-controlled dose–effect studies of ARBs, used at per-label ‘recommended doses’, suggests that the BP-lowering efficacy of olmesartan, defined as the maximal BP decrease observed at the plateau of dose–effect relationships is at the highest end of the range of efficacy compared with that of losartan, irbesartan, valsartan, and candesartan. The overview of head-to-head comparisons of olmesartan to each of the other ARBs respectively, all used at per-label ‘recommended doses’, supports the finding of a greater BP-lowering effect of olmesartan.
Placebo-controlled dose–response studies
Examining dose–response relationship curves confirmed that most of the antihypertensive effects was obtained with the once daily RMD we selected for each ARB, i.e. olmesartan 20 mg, candesartan 8 mg, irbesartan 150 mg, losartan 50 mg and valsartan 80 mg.
The Emax model we used for the calculation of maximal BP lowering effect was applicable to all ARBs, except losartan. Correlation coefficients were very high, ranging from 0.77 to 0.99, which is a robust confirmation of internal validity. It is noteworthy that, using the simplified Lineweaver–Burk regression method, values found for irbesartan (Emax DBP/SBP = −6.5/−11.2 mmHg) are very close to those found by Reeves et al.  (−6.6/−11.3 mmHg) fitting individual data to the complete model (adding the ‘steepness’ Hill coefficient α to the numerator). This supports the external validity of our results.
The superiority of calculated Emax of olmesartan is consistent with the greater efficacy of olmesartan on the observed BP-lowering effect compared with the other studied ARBs at per-label respective ‘recommended doses’. Comparison of these agents was indirect as performed through separate published studies. It should therefore be interpreted with caution in the overall context of the specific approach described in this paper. However, we did restrict our overview to studies with strictly comparable study populations, design and duration, and all results were placebo-corrected.
Head-to-head comparisons studies
Our main finding of greater efficacy of olmesartan in placebo-controlled dose–effect studies is supported by the results of head-to-head comparisons where olmesartan vs. comparator BP-lowering differences ranged from 0.7 to 3.6 mmHg.
Pharmacological differences among ARBs
Several explanations may be the reason for the observed differences in maximal BP-lowering efficacy among the available ARBs. These agents do share a common mechanism of action, namely the antagonism of angiotensin II AT1 receptors. However, their receptor-binding kinetics differs substantially. Olmesartan and candesartan have a higher affinity for the AT1 receptor than all other ARBs . In addition, candesartan and irbesartan are insurmountable antagonists, whereas losartan and valsartan are competitive antagonists. Pharmacokinetic properties of ARBs also differ in terms of oral bioavailability, rate of absorption, metabolism, and route and rate of elimination [31–33]. On the basis of terminal elimination half-lives, losartan, and valsartan, may be classified as shorter acting while candesartan cilexetil, irbesartan and olmesartan as longer acting.
Clinical relevance of BP-lowering efficacy
The greater efficacy in BP lowering may be clinically relevant given that even small BP differences are associated with substantial differences in the incidence of cardiovascular events in hypertension trials. Large hypertension clinical trials pointed out the importance of a prompt and efficient BP control  but also the significant impact of a very few millimeter difference. Two meta-analyses showed consistent results [1,3]: the relationship between BP level and mortality from stroke or myocardial infarction is linear down to low BP levels . In the Hypertension Optimal Treatment study (HOT), there were 28% fewer myocardial infarctions in the treatment group with a target DBP ≤80 mmHg than in the group with a target DBP of ≤90 mmHg, although the difference in the mean DBP achieved by these groups was only 4.1 mmHg . A mean 5-mmHg DBP decrease is associated with reductions of at least 21% in the incidence of coronary heart disease and 34% in the incidence of stroke . More accurately, it has been shown that a mean 7.5 mmHg DBP reduction allowed a risk reduction by −21% in coronary heart disease and by −46% in stroke. The Blood Pressure Lowering Treatment Trialists highlighted that mean −2/−1 mmHg (SBP/DBP) differences observed with ARBs-based regimens vs. control regimens were indeed associated with consistently and significantly lower relative risks of stroke, coronary heart disease, heart failure, major cardiovascular events, cardiovascular death and total mortality .
Dramatic benefit should thus be expected for minor mean DBP reductions such as a 2-mmHg as it would result in a 17% decrease in the prevalence of hypertension, as well as a reduction in the risk of coronary heart disease by −6% and of stroke and transient ischemic attacks by −15% . One of the major findings of the Valsartan Antihypertensive Long-term Use Evaluation (VALUE) trial was that minor mean differences in BP within the range of 1.5 or 2 mmHg over a short time period may be associated with a meaningful lower cardiovascular risk .
In the SHEP study, a mean difference of 3-mmHg in SBP was associated with 10–20% lower rates of heart failure . In the Losartan Intervention For Endpoint reduction (LIFE) trial, BP was reduced to a similar extent in the losartan as well as in the control group . However, at the end of the study, mean sitting SBP was reduced by 1 mmHg more with losartan than with atenolol, and was associated with a significant decrease in relative risk of 13% for the composite endpoint of cardiovascular morbidity, stroke, and myocardial infarction [5,6]. In addition, in the separate analysis of diabetic hypertensive patients, the mean 2-mmHg difference in SBP between the two treatment groups at endpoint might have contributed to marked differences in the reduction in risk. Recently, the Anglo-Scandinavian Cardiac Outcomes Trial-Blood Pressure Lowering Arm (ASCOT-BPLA) study showed a greater efficiency in reducing cardiovascular events observed with a mean difference of 2.7 mmHg in the SBP lowering between the two antihypertensive regimens .
In conclusion, this overview of placebo-controlled dose–effect studies of similar design with individual ARBs suggests that clinically relevant differences in maximal efficacy as well as in efficacy of per-label recommended doses can be evidenced. Among the studied ARBs, olmesartan can reduce DBP and SBP significantly more than other ARBs. This is further supported by the results of published head-to-head comparisons, which have proved olmesartan to have a greater antihypertensive effect than losartan, valsartan, irbesartan and candesartan at per-label recommended doses.
Differences in efficacy between antihypertensive agents do exist and are most likely to be clinically meaningful, considering the relevance of optimal BP control. For a number of antihypertensive agents, assessment of the dose–effect relationship in phase II placebo-controlled studies is not always optimal. Yet, properly performed studies may be critical for the definition of the optimal doses and maximal efficacy of individual agents.
Conflict of interest statement
Faiez Zannad receives consulting and speaking honoraria from several pharmaceutical companies. The present manuscript has been drafted following a report performed by Faiez Zannad as an expert to the Comité de Transparence of the Agence Française de Sécurité Sanitaire et des Produits de Santé (AFSSAPS) and commissionned by Sankyo-Pharma France.
The authors wish to thank Sophie Hermant for her help in the preparation of the manuscript and for editorial support.