CLINICAL PHARMACOLOGY TRIALS
An ambulatory blood pressure monitoring study of the
comparative antihypertensive efficacy of two angiotensin
II receptor antagonists, irbesartan and valsartan
Giuseppe Manciaa, Krishna Korliparab, Peter van Rossumc,
Giuseppe Villad and Barry Silverte
Background The primary objective of this study was to
compare the change from baseline in mean diastolic
ambulatory blood pressure (ABP) at 24 h post dose
(trough measurement) after 8 weeks of treatment with
irbesartan or valsartan in subjects with mild-to-moderate
hypertension. Secondary objectives included comparing
the mean changes from baseline in systolic ABP at
trough; 24-h ABP; morning and night-time ABP; self-
measured systolic blood pressure (SBP) and diastolic
blood pressure (DBP); and office-measured SBP and
DBP at trough.
Design After a 3-week, single blind, placebo lead-in
period, 426 subjects were randomized to receive either
irbesartan 150 mg or valsartan 80 mg for 8 weeks.
Methods Ambulatory blood pressure measurements
were obtained at baseline and at week 8. Self-measured
morning and evening DBP and SBP readings were
obtained at home over a 7-day period at baseline and at
week 8. Office-measured seated DBP and SBP
measurements were obtained at trough, at baseline,
and at week 8.
Results Irbesartan demonstrated significantly greater
reductions than valsartan for mean change from
baseline in diastolic ABP at trough (�6.73 versus
�4.84 mmHg, respectively; P¼0.035). Irbesartan
produced significantly greater reductions than valsartan
for mean systolic ABP at trough (�11.62 versus
�7.5 mmHg, respectively; Po0.01) and for mean 24-h
diastolic ABP (�6.38 versus �4.82 mmHg,
respectively; P¼ 0.023) and systolic ABP (�10.24
versus �7.76 mmHg; Po0.01). Irbesartan also
produced significantly greater reductions than valsartan
for office-measured seated DBP (�10.46 versus
7.28 mmHg, respectively; Po0.01) and SBP
(�16.23 versus �9.96 mmHg, respectively; Po0.01)
and for self-measured morning DBP (�6.28 versus
�3.75 mmHg, respectively; Po0.01) and SBP
(�10.21 versus �6.97 mmHg, respectively; Po0.01).
Both drugs were well tolerated.
Conclusion Irbesartan was more effective than
valsartan in reducing DBP and SBP at trough and in
providing greater overall 24-h blood pressure-lowering
efficacy. Blood Press Monit 7: 135–142 & 2002
Lippincott Williams & Wilkins.
Blood Pressure Monitoring 2002, 7:135^142
Keywords: ambulatory blood pressure monitoring, angiotensin II
receptor antagonists, irbesartan, valsartan
aOspedale S. Gerardo and Universita` Milano-Bicocca, Milan, Italy; bPike View
Medical Centre, Bolton, UK; cBeatrix Ziekenhuis, Gorinchem, The Nether-
lands; dFondazione S. Maugeri, Pavia, Italy and eStonehill Medical Centre,
Bolton, UK.
Sponsorship: This study was supported by Bristol-Myers Squibb and
Sanofi-Synthelabo
Conflicts of interest: none
Correspondence and requests for reprints to Giuseppe Mancia, MD,
Professor of Medicine, University of Milan-Bicocca, Ospedale S. Gerardo,
Monza, Department of Medicine, Via Donizetti, 106, 20052 Monza, Italy.
Fax: þ39 039 322 274; e-mail: mancia@gcst.inimi.it
Received 25 October 2001 Revised 06 December 2001
Accepted 28 December 2001
Introduction
Angiotensin II receptor antagonists (AIIRAs) are a recently
developed class of antihypertensive agents that have
gained widespread use in clinical practice because of (1)
antihypertensive efficacy similar to that of all agents
classically employed to treat hypertension [i.e., calcium
channel blockers (CCBs), beta-blockers, angiotensin-con-
verting enzyme (ACE) inhibitors, and diuretics], and (2) a
placebo-like tolerability profile, which holds the promise of
favorably impacting long-term compliance with antihyper-
tensive treatment [1].
This placebo-like tolerability profile has been documented
for all AIIRAs and should thus be regarded as a class-related
feature [2]. Evidence has been produced, on the other
hand, that compared with the first available AIIRA,
losartan, subsequent agents of this class including cande-
sartan, irbesartan, and telmisartan, may have greater blood
pressure (BP)-lowering ability, possibly because of differ-
ences in pharmacodynamic or pharmacokinetic properties
[3–8].
Little information, however, is available on the comparative
antihypertensive efficacy among the newer AIIRAs. The
1359-5237 & 2002 Lippincott Williams & Wilkins
I0054851
高亮
I0054851
高亮
I0054851
高亮
present study addressed this question by comparing two
agents for which a greater angiotensin II antagonistic and
BP-lowering effect has been demonstrated compared with
losartan, irbesartan and valsartan [9–13]. Antihypertensive
efficacy in the current analysis was assessed by BP
measurements in the physician’s office, at home, and by
ambulatory monitoring to base results on multiple data
collected in different environmental conditions.
Methods
Subjects
Subjects were males and females aged between 18 and 75
years old with mild-to-moderate essential hypertension
[defined as seated diastolic blood pressure (DBP) between
95 mmHg and 110 mmHg]. Subjects provided informed
written consent and were willing to discontinue previous
antihypertensive medicine if necessary. Subjects were
enrolled from 45 sites in Italy, the UK, and the Netherlands
(Appendix 1).
Study design
This randomized, multicenter, double blind, parallel-group
study compared the antihypertensive efficacy of irbesartan
150 mg with that of valsartan 80 mg, administered once
daily for 8 weeks. Three methods of obtaining BP were
used: 24-h ambulatory blood pressure (ABP) monitoring;
at-home, self-measured BP; and office-measured BP at
247 3 h after the last dose of study medication (i.e., at
trough). The primary objective was to compare the change
from baseline in diastolic ABP at trough after 8 weeks of
treatment. Secondary objectives included comparing the
changes from baseline in systolic ABP at trough; 24-h mean
systolic and diastolic ABP; morning mean and evening
mean systolic and diastolic ABP; self-measured DBP and
systolic blood pressure (SBP); and office-measured seated
DBP and SBP and heart rate at trough. Another secondary
objective was to determine the percentage of subjects
whose office-measured BP at trough was normalized
(DBPo90 mmHg) and who responded to therapy (i.e.,
whose office-measured DBP was normalized or changed
from baseline X 10 mmHg).
A medical history that included a complete physical
examination and an electrocardiogram was obtained at
screening. Brief physical examinations were performed at
each visit, and a complete examination was performed
again at study conclusion. Fasting blood samples were
obtained for complete blood count and serum chemistries
at screening and at study conclusion. All medications
known to affect BP were prohibited. Subjects
were withdrawn from participation in the study if, at
any time, mean office BP measurements over a 3-day
period were SBP4 200 mmHg or DBP4 115 mmHg. Ex-
clusion criteria included known or suspected secondary
hypertension; office seated DBP4 110 mmHg or seated
SBP4 200 mmHg; history of significant cardiac, hepatic, or
renal disease; or uncontrolled diabetes, defined as fasting
blood glucose 4 180 mg/dl (4 10 mmol/l) or non-fasting
blood glucose 4 220 mg/dl (4 11.2 mmol/l), and clinically
abnormal serum creatinine level, defined as X 1.5 mg/dl
(X 132mmol/l).
After a 3-week, single blind, placebo lead-in period,
qualified subjects were randomized on a 1 : 1 ratio and
entered 8 weeks of double-blind treatment. At each study
visit, subjects were monitored for adverse events. Subjects
were instructed to arrive at the clinic in the morning,
before taking their study medication. Subjects were asked
to refrain from consuming alcoholic beverages for 6 h prior
to BP measurements and from smoking for 2 h beforehand.
Measurements
ABP measurements
Twenty-four hour ABP readings were obtained with the
same type of ABP monitor (model 90207; Spacelabs
Medical, Redmond, Washington, USA) for all subjects.
The device was set to obtain readings every 15 min
throughout a 24-h period. ABP readings were conducted
at baseline (at the end of the lead-in period) and at week 8.
During the 24th hour of ABP recording, measurements
were obtained every 3 min, with the patients in controlled
conditions at the clinic (seated, at rest, quiet, and not
smoking).
Office BP and heart rate measurements
Seated office ABP and heart rate measurements were
obtained at baseline and at weeks 4 and 8 at trough.
Subjects were instructed to arrive at the physician’s office
between 0700–1000 h, without having taken that day’s
study medication, eaten breakfast, smoked within 2 h prior
to the BP readings, or consumed alcohol within 6 h prior to
the BP readings. All readings were taken 247 3 h following
the last dose of study medication, in the dominant arm, and
by the same observer. Blood pressure was measured with a
standard, calibrated, mercury sphygmomanometer, follow-
ing 10 min of rest in the seated position. Three consecutive
BP measurements were obtained at least 1 min apart and
the values were averaged, after which heart rate was
measured by pulse count for 30 s and multiplied by two. If
the three consecutive seated DBP readings were not within
8 mmHg of each other, two additional readings were
obtained at least 1 min apart, and heart rate was re-
measured. For safety reasons, if at any time during the
study seated SBP was 4 200 mmHg or seated DBP was
4 115 mmHg based on the mean of three readings, and
these values were confirmed by follow-up within 3 days,
the subject was removed from the study. Standing BP and
heart rate were measured in the same manner, after the
subject had stood in place for 2 min.
136 Blood Pressure Monitoring 2002, Vol 7 No 2
Home BP measurements
All subjects were trained to use a validated, calibrated,
semi-automatic BP self-measurement device (Pressolink-T;
Tam Telesante, Aix-en-Provence, France). The device
includes an Omron HEM-705 BP monitoring device
(Omron Healthcare, Inc, Vernon Hills, Illinois, USA),
software for data collection, and hardware for data
transmission. Self-measured BP was obtained at baseline
and at week 8. Blood pressure values were automatically
recorded and stored by the BP self-measurement device.
Subjects were instructed to take twice-daily (morning and
evening) sequences of three BP readings over seven
consecutive days for each of the two periods while seated.
Three self-measurements were taken twice daily: in the
morning on awakening (before breakfast and study drug
intake) in a seated position, and in the evening at bedtime
in a seated position. Subjects were not instructed to follow
a specific sleep regimen, although awakening and bedtimes
were recorded. Values recorded on the first day of each
7-day period were excluded from analysis. The remaining
values (n¼ 18) were averaged.
Methods of analysis
Patient characteristics
Key baseline variables were age, sex, race, weight, baseline
ABP measurements, baseline self-measured BP, and base-
line office-measured BP at trough.
Sample size
In order to detect a 3 mmHg difference between irbesartan
and valsartan in mean diastolic ABP at trough at week 8
compared with baseline, a two-sided test at a significance
level of a¼ 0.05 with a power of 1�a¼ 0.9, including 155
subjects per group, was required. Total sample size was
n¼ 388 (194 in each group), which was calculated to
anticipate loss to follow-up and unacceptable BP data.
Statistical tests
Descriptive summaries and statistical tests were per-
formed. Subjects maintained a diary of their sleeping
patterns, and these data were used to define daytime hours
(i.e. , hours from awakening until going to bed) and night-
time hours (i.e., hours from going to bed until awakening).
The mean daytime ABP was defined as the arithmetic
mean of the hourly mean values during the hours that the
subject was awake after the morning drug intake, and the
hours from awakening until completion of the 24th hour
post dose the next morning. The mean night-time ABP was
defined as the arithmetic mean of the hourly mean values
while the patient was asleep. The transitional hours, (i.e.,
the time between going to bed and falling asleep, and upon
awakening) were included in the night-time average if the
subject slept at least 30 min during the hour, or in the
daytime average if the patient slept less than 30 min during
the hour. The variables analysed for change from baseline
at week 8 were mean diastolic and systolic ABP at trough;
mean 24-h diastolic and systolic ABP; daytime and
night-time mean diastolic and systolic ABP; morning and
evening mean self-measured DBP and SBP; and seated
and standing mean office-measured DBP and SBP. Group
comparisons were calculated for irbesartan versus valsartan
for the percentage of subjects who were normalized and
subjects who were responders.
The primary efficacy analysis was performed on all
randomized subjects with acceptable ABP data (n¼ 303).
Secondary efficacy analyses were based on data from all
subjects with valid data that could be evaluated. Rigorous
criteria for compliance with each of the efficacy analyses
were employed. Subjects were excluded from individual
efficacy analyses if (1) ABP recording started too early or
too late (7 1 h); (2) the number of valid readings was less
than the percentage of the expected number of readings;
(3) the number of 24-h readings was less than the
percentage of the expected number of readings; and (4)
15 self-measured readings were not available over 5 days.
For the primary efficacy analysis of change in trough ABP at
Week 8, acceptable data was defined as at least 15 valid
readings in the repetitive trough ABP measurement
interval and at least 69 valid readings for the first 23 h
post dose. Of the 426 subjects randomized, 11 discon-
tinued treatment before study completion, and 112 had
data that were not acceptable for statistical analysis
according to the criteria defined earlier. Of the 11 subjects
who discontinued treatment, one subject discontinued
because of hypertension. In total, 303 subjects were
included in the primary efficacy analysis. Table 1 depicts
the number of patients analysed for each study end-point.
For the secondary end-points of self-measured mean
morning ABP, self-measured mean evening ABP, and self-
measured mean overall ABP, different criteria were used to
define the readings as acceptable for statistical analysis. For
Table 1 Number of subjects included in analysis per endpoint
Randomized subjects
per analysis at week 8
Valsartan
80 mg (n)
Irbesartan
150 mg (n)
Total (n)
Subjects included in
analysis of ABPM data
215 211 426
Subjects with valid ABPM data 156 147 303
Subjects included in self-measured
BP data analysis
Morning 140 139 279
Evening 121 117 238
Total 99 98 197
Subjects included in trough office
seated BP data analysis
211 202 413
Subjects included in safety analysis 215 211 426
ABPM, ambulatory blood pressure monitoring; BP, blood pressure.
Efficacy of irbesartan versus valsartan Mancia et al. 137
mean morning measurements, at least 15 valid readings
were required; for mean evening measurements, at least 12
valid readings were required; and for mean overall
measurements, at least 27 valid, paired readings (valid
morning and evening readings on the same day) were
required. In addition, all three end-points had to have at
least 5 days (not necessarily consecutive) of valid readings.
For morning, evening, and overall self-measured ABP, a
total of 279, 238, and 197 subjects were included in the
analyses, respectively.
Summary statistics included mean, standard deviation, and
mean plus standard deviation change from baseline for each
treatment group. For the percentage of subjects who were
normalized and who were responders, within-group 95%
confidence intervals were calculated. The 95% confidence
intervals and P values for the risk ratios of the therapeutic
response variables were calculated post hoc using the
Cochran–Mantel–Haenszel test.
An analysis of covariance (ANCOVA) model was used to
compare the treatment arms with regard to the following
variables: changes in mean diastolic or systolic ABP (at
trough, 24-h, and daytime and night-time); changes in
mean self-measured DBP or SBP; and office-measured
seated and standing DBP and SBP at trough. Treatment
comparisons were carried out using the ANCOVA model
with treatment and country as main factors and baseline
values as a covariate. A second covariate, the ‘number of
awake hours during week 8 of ABP monitoring,’ was also
included in the ANCOVA model for the change from
baseline in daytime and night-time hours.
Safety analysis
All subjects who received at least one dose of study
medication were included in the analysis of safety data. No
statistical tests were performed because of lack of power to
determine statistical significance. Adverse events were
classified as either a concomitant experience (CE)
(unrelated or unlikely to be related to study medication)
or an adverse drug experience (ADE) (related, probably
related, or possibly related to study medication). Extent of
exposure to study drug (i.e., number of days of exposure)
was summarized, and the mean duration of exposure was
calculated for each treatment group.
Results
In this study, 426 subjects were randomized to receive
either irbesartan 150 mg or valsartan 80 mg. Each group was
balanced in demographic characteristics and baseline BP
values (Table 2). The mean age was approximately 55
years, and subjects were predominantly white and male.
Table 3 shows the frequency of common comorbidities
within the two groups.
The values for various BP changes induced by treatment
are shown in Figure 1. While both irbesartan and valsartan
significantly reduced BP from baseline, in the 303 subjects
included in the efficacy analysis for the study’s primary
objective, irbesartan produced a statistically greater reduc-
tion in mean diastolic ABP at trough (�6.73 versus
�4.84 mmHg, respectively; P¼ 0.035). Irbesartan also
produced a statistically greater reduction in mean systolic
ABP at trough (�11.62 versus �7.5 mmHg, respectively;
Po 0.01). In addition, irbesartan caused a significantly
greater reduction compared with valsartan in mean 24-h
diastolic and systolic ABP (P¼ 0.023 and Po 0.01,
respectively); mean daytime diastolic and systolic ABP
(P¼ 0.017 and P¼ 0.02, respectively); and mean morning
self-measured DBP and SBP (Po 0.01 for both).
There was no substantial difference between the two drugs
for mean night-time ABP or for mean evening self-
measured BP change from baseline. Hourly BP profile
was lower with treatment than at baseline in both groups.
During treatment, hourly BP values were usually lower in
the irbesartan than in the valsartan group, the difference
being more evident during the day than during the night
(Fig. 2).
There were 413 subjects included in the efficacy analysis
for office-measured seated and standing DBP and SBP at
Table 2 Baseline demographics and clinical characteristics
Characteristics Valsartan 80 mg Irbesartan 150 mg
n 215 211
Sex (% male) 67 62
Race (% white) 97 96
Mean age (y)/(SD) 55.4 (9.9) 55.1 (9.7)
n 213 209
Mean weight (kg)/(SD) 81.8 (14.5) 79.3 (12.4)
n 179 178
Trough 24th-h ABP
(SBP/DBP; mmHg)
150.0/95.7 148.3/94.2
[SD] [9.9/15.0] [9.2/14.3]
n 175 173
Mean ABP (SBP/DBP; mmHg) 143.8/89.1 142.2/88.2
[SD] [12.0/8.3] [12.7/8.3]
n 214 211
Seated office BP
(SBP/DBP; mmHg)
158/100.8 159.3/100.7
[SD] [14.1/4.6] [13.6/4.2]
n 170 169
Self-measured morning BP
(SBP/DBP; mmHg)
149.2/96.4 149.1/96.6
[SD] [15.0/10.1] [17.8/10.2]
n 159 147
Self-measured evening BP
(SBP/DBP; mmHg)
149.8/95.2 148.8/94.3
[SD] [14.6/10.2] [17.1/9.8]
n 213 210
Seated office HR (bpm) 75.2 74.1
[SD] [10.5] [7.8]
ABP, ambulatory blood pressure; SBP, systolic blood pressure; DBP, diastolic
blood pressure; BP, blood pressure; HR, heart rate. Val