The New England

Journal of Medicine
C o py r ig ht © 2 0 0 1 by t he Ma s s ac h u s e t t s Me d ic a l S o c ie t y

VOLUME 345 S E P T E M B E R 20, 2001 NUMB ER 12

RENOPROTECTIVE EFFECT OF THE ANGIOTENSIN-RECEPTOR ANTAGONIST

IRBESARTAN IN PATIENTS WITH NEPHROPATHY DUE TO TYPE 2 DIABETES

EDMUND J. LEWIS, M.D., LAWRENCE G. HUNSICKER, M.D., WILLIAM R. CLARKE, PH.D., TOMAS BERL, M.D.,
MARC A. POHL, M.D., JULIA B. LEWIS, M.D., EBERHARD RITZ, M.D., ROBERT C. ATKINS, M.D., RICHARD ROHDE, B.S.,
AND ITAMAR RAZ, M.D., FOR THE COLLABORATIVE STUDY GROUP*

D
ABSTRACT IABETES mellitus is increasing in preva-
Background It is unknown whether either the an- lence worldwide and is currently estimat-
giotensin-II–receptor blocker irbesartan or the cal- ed to affect more than 6.5 percent of the
cium-channel blocker amlodipine slows the progres- population of the United States.1 Diabetes
sion of nephropathy in patients with type 2 diabetes is the most common cause of end-stage renal disease
independently of its capacity to lower the systemic in this country, accounting for 40 percent of cases.2
blood pressure. Although the inhibition of the effects of angiotensin
Methods We randomly assigned 1715 hypertensive II has a beneficial effect in patients with nephropathy
patients with nephropathy due to type 2 diabetes to caused by type 1 diabetes,3 no published study with
treatment with irbesartan (300 mg daily), amlodipine
(10 mg daily), or placebo. The target blood pressure
definitive renal outcomes has addressed the issue of
was 135/85 mm Hg or less in all groups. We compared renoprotection in patients with type 2 diabetes — a
the groups with regard to the time to the primary population that differs substantially from patients with
composite end point of a doubling of the base-line se- type 1 diabetes in terms of demographic characteris-
rum creatinine concentration, the development of end- tics, metabolic features, and potential mechanisms of
stage renal disease, or death from any cause. We also glomerular disease.4 Several studies have addressed
compared them with regard to the time to a second- the positive effects of specific antihypertensive agents
ary, cardiovascular composite end point. on cardiovascular morbidity and mortality within this
Results The mean duration of follow-up was 2.6 population.5-8
years. Treatment with irbesartan was associated with We undertook the Irbesartan Diabetic Nephropathy
a risk of the primary composite end point that was 20
Trial to determine whether the use of an angioten-
percent lower than that in the placebo group (P=0.02)
and 23 percent lower than that in the amlodipine sin-II–receptor blocker or a calcium-channel blocker
group (P=0.006). The risk of a doubling of the serum would provide protection against the progression of
creatinine concentration was 33 percent lower in the nephropathy due to type 2 diabetes beyond that at-
irbesartan group than in the placebo group (P=0.003) tributable to the lowering of the blood pressure. We
and 37 percent lower in the irbesartan group than in also compared the groups assigned to different ther-
the amlodipine group (P<0.001). Treatment with irbe- apeutic regimens in terms of overall mortality and the
sartan was associated with a relative risk of end-stage rate of cardiovascular events.
renal disease that was 23 percent lower than that in
both other groups (P=0.07 for both comparisons).
These differences were not explained by differences
in the blood pressures that were achieved. The serum From the Department of Medicine, Rush–Presbyterian–St. Luke’s Med-
ical Center, Chicago (E.J.L., R.R.); the Department of Internal Medicine,
creatinine concentration increased 24 percent more University of Iowa College of Medicine (L.G.H.), and the Department of
slowly in the irbesartan group than in the placebo Biostatistics, University of Iowa College of Public Health (W.R.C.), Iowa
group (P=0.008) and 21 percent more slowly than in City; the Department of Internal Medicine, University of Colorado School
the amlodipine group (P=0.02). There were no signif- of Medicine, Denver (T.B.); the Department of Medicine, Cleveland Clinic
Foundation, Cleveland (M.A.P.); the Department of Medicine, Vanderbilt
icant differences in the rates of death from any cause University School of Medicine, Nashville (J.B.L.); the Department of Med-
or in the cardiovascular composite end point. icine, Ruperto Carola University, Heidelberg, Germany (E.R.); the Depart-
Conclusions The angiotensin-II–receptor blocker ment of Nephrology, Monash Medical Center, Melbourne, Australia
irbesartan is effective in protecting against the pro- (R.C.A.); and the Department of Medicine, Hadassah University, Jerusa-
lem, Israel (I.R.). Address reprint requests to Dr. Edmund J. Lewis at
gression of nephropathy due to type 2 diabetes. This Rush–Presbyterian–St. Luke’s Medical Center, 1650 W. Harrison, Suite
protection is independent of the reduction in blood 515 RA, Chicago, IL 60612.
pressure it causes. (N Engl J Med 2001;345:851-60.) *The participants in the Collaborative Study Group are listed in the Ap-
Copyright © 2001 Massachusetts Medical Society. pendix.

N Engl J Med, Vol. 345, No. 12 · September 20, 2001 · www.nejm.org · 851

Downloaded from www.nejm.org on March 22, 2008 . Copyright © 2001 Massachusetts Medical Society. All rights reserved.
 The Ne w E n g l a nd Jo u r n a l o f Me d ic i ne

METHODS renal disease (as indicated by the initiation of dialysis, renal trans-
We conducted a prospective, randomized, double-blind clinical plantation, or a serum creatinine concentration of at least 6.0 mg
trial in 210 clinical centers. The conduct of the study was man- per deciliter [530 µmol per liter]), or death from any cause. The sec-
aged and monitored by the clinical coordinating center and the var- ondary, cardiovascular end point was the composite of death from
ious committees of the Collaborative Study Group. All decisions cardiovascular causes, nonfatal myocardial infarction, heart failure
regarding the biostatistical protocol, study plans, and analyses, in- resulting in hospitalization, a permanent neurologic deficit caused
cluding the data presented to the data safety and monitoring board, by a cerebrovascular event, or lower limb amputation above the an-
were the product of the biostatistical coordinating center of the kle. The serum creatinine outcomes were confirmed with the use
study group. The biostatistics and data management department of centrally determined concentrations of serum creatinine. All out-
of Bristol-Myers Squibb was responsible for data handling, in- comes were reviewed and classified by an outcome committee.
cluding entry into the master data base, data-base review, and au- Adverse events were recorded at quarterly visits.
dit. The blinded clinical data base was provided to the biostatis-
tical coordinating center of the Collaborative Study Group for Statistical Analysis
the generation of interim reports and for final statistical analyses Comparisons of base-line values and outcomes that were not
for publications. Details of the protocol have been published previ- time dependent among the three treatment groups were made with
ously.9 The institutional review board or ethics committee of each the chi-square test (for categorical data) or analysis of variance (for
center approved the protocol, and all patients gave written informed continuous data). The times to the primary end point and its com-
consent after reviewing a written summary of the study plan. ponents were compared by means of product-limit survival curves
and the log-rank test.10 All analyses were based on the intention-
Study Patients to-treat principle. Comparisons of the survival curves for the sec-
The criteria for eligibility included an age between 30 and 70 ondary, cardiovascular end point were made by means of the
years, a documented diagnosis of type 2 diabetes mellitus, hyperten- Breslow–Gehan test.11 Adjustment for base-line covariates was per-
sion (a systolic blood pressure of more than 135 mm Hg while formed with the use of Cox proportional-hazards models with
sitting, a diastolic blood pressure of more than 85 mm Hg while sit- terms for the treatment assignment and the base-line covariates.12
ting, or documented treatment with antihypertensive agents), and Similar proportional-hazards models were used to adjust for time-
proteinuria, with urinary protein excretion of at least 900 mg per dependent covariates such as the mean arterial pressure. In calcu-
24 hours. The serum creatinine concentration was required to be lating the slopes of the rates of change of the serum creatinine
between 1.0 and 3.0 mg per deciliter (88 and 265 µmol per liter) concentration and creatinine clearance, a mixed model was used in
in women and 1.2 and 3.0 mg per deciliter (106 and 265 µmol per which the data for patients who reached an end point were cen-
liter) in men. sored. Comparisons among subgroups defined according to the fre-
quency of serious adverse events were performed with the use of
Randomization and Treatment Plan methods for comparing Poisson counts.13 All statistical tests were
All angiotensin-converting–enzyme (ACE) inhibitors, angio- two-sided.
tensin-receptor blockers, and calcium-channel blockers were discon- On the basis of the results of our study of type 1 diabetes,3 in
tinued at least 10 days before the screening period, during which which the three-year rate of a doubling of the base-line serum cre-
time blood pressure was controlled with other agents. Eligible pa- atinine concentration, end-stage renal disease, or death was 36 per-
tients were randomly assigned by a central office to one of three cent, we estimated that we would need 550 patients per treatment
treatment regimens: irbesartan (Avapro, Bristol-Myers Squibb, group for an analysis of the primary outcome. The sample size was
Princeton, N.J.), in a dose titrated from 75 to 300 mg per day; selected to achieve 90 percent power to detect a 26 percent dif-
amlodipine (Norvasc, Pfizer, New York), in a dose titrated from ference in the primary end point between the irbesartan group
2.5 to 10 mg per day; or placebo. Antihypertensive agents other and the placebo group at a 5 percent alpha level.
than ACE inhibitors, angiotensin-receptor blockers, and calcium- An independent data and safety monitoring board monitored
channel blockers were used as needed in each group, and the tar- the study. The Lan–DeMets alpha spending-function method13
get blood pressure for all patients was the same (a systolic blood was used to adjust for interim analyses. Four formal interim analy-
pressure of 135 mm Hg or less, or 10 mm Hg lower than the val- ses were performed. The adjusted level of significance for the final
ue at screening if that value was more than 145 mm Hg, and a analysis of the primary outcome was P=0.04.
diastolic blood pressure of 85 mm Hg or less). Survival, end-stage For all other outcomes, a P value of 0.05 or less was considered
renal disease, the cardiovascular end points, the serum creatinine to indicate significance. The protocol defined the comparison of
and potassium concentrations, and the 24-hour urinary protein the irbesartan group with the placebo group as the primary com-
excretion were monitored quarterly. Blood-pressure measurements parison and the comparison of the irbesartan group with the am-
were reviewed by a clinical management committee that made lodipine group as the secondary comparison. Because the protocol
treatment recommendations. The committee also monitored blood identified one primary comparison, the reported P values were
glucose concentrations by measurement of glycosylated hemoglo- not corrected for multiple comparisons.
bin. Adherence to the treatment regimen was monitored by means
of pill counts. Serum creatinine concentrations were determined RESULTS
by a central laboratory. Serum potassium and urinary protein were
measured in four regional laboratories. A protocol was established Between March 21, 1996, and February 25, 1999,
for the management of hyperkalemia and the detection of early 1715 patients underwent randomization. The base-
increases in the creatinine concentration that might be caused by line demographic, clinical, and laboratory character-
renal-artery stenosis.9 We planned to continue recruitment for three istics of the three groups were similar (Table 1), ex-
years and to follow all patients for a minimum of two years after the
end of recruitment; data were censored on December 31, 2000.
cept that a slightly lower proportion of the patients
The mean arterial blood pressure was calculated as the diastolic in the placebo group were female (P=0.02). The out-
blood pressure plus one third of the difference between the systolic comes for these patients are summarized in Table 2.
blood pressure and the diastolic blood pressure. Values for urinary
protein excretion were log-transformed to reduce skewness. Clinical Management
End Points The blood-pressure measurements for the three
The primary end point was the composite of a doubling of the groups are shown in Figure 1. In all three groups, the
base-line serum creatinine concentration, the onset of end-stage proportion of patients in whom the target blood

852 · N Engl J Med, Vol. 345, No. 12 · September 20, 2001 · www.nejm.org

Downloaded from www.nejm.org on March 22, 2008 . Copyright © 2001 Massachusetts Medical Society. All rights reserved.
 E F F EC T OF IRBE SAR TAN ON NEPH R OPATH Y D UE TO T Y PE 2 D IA BETES

TABLE 1. BASE-LINE CHARACTERISTICS OF THE PATIENTS.*

IRBESARTAN AMLODIPINE PLACEBO

GROUP GROUP GROUP P
CHARACTERISTIC (N=579) (N=567) (N=569) VALUE

Age — yr 59.3±7.1 59.1±7.9 58.3±8.2 0.63

Male sex — no. (%) 378 (65) 359 (63) 403 (71) 0.02
Race or ethnic group — no. (%) 0.33
Non-Hispanic white 438 (76) 389 (69) 415 (73)
Non-Hispanic black 63 (11) 87 (15) 78 (14)
Hispanic 28 (5) 29 (5) 26 (5)
Asian or Pacific Islander 24 (4) 34 (6) 27 (5)
Other 26 (4) 28 (5) 23 (4)
Body-mass index† 31.0±5.6 30.9±5.9 30.5±5.9 0.34
Blood pressure — mm Hg
Systolic 160±20 159±19 158±20 0.10
Diastolic 87±11 87±11 87±11 0.97
Insulin use at entry — no. (%) 329 (57) 327 (58) 335 (59) 0.78
History of cardiovascular disease — no. (%) 158 (27) 171 (30) 164 (29) 0.37
Retinopathy — no. (%) 401 (69) 362 (64) 380 (67) 0.19
Serum creatinine — mg/dl‡ 1.67±0.53 1.65±0.61 1.69±0.57 0.50
Urinary protein excretion — g/24 hr
Median 2.9 2.9 2.9 0.40
Interquartile range 1.6–5.4 1.6–5.2 1.8–5.2
Urinary albumin excretion — g/24 hr
Median 1.9 1.9 1.9 0.47
Interquartile range 1.0–3.8 1.0–3.5 1.1–3.5
Glycosylated hemoglobin — % 8.1±1.7 8.2±1.7 8.2±1.7 0.82

*Plus–minus values are means ±SD.

†The body-mass index is the weight in kilograms divided by the square of the height in meters.
‡To convert values to micromoles per liter, multiply by 88.4.

pressure was achieved increased and the mean blood cebo group (P=0.02) and 23 percent lower than that
pressure decreased over the course of the study; the in the amlodipine group (P=0.006) (Table 3). The
mean blood pressure at visits after base line was 140/ relative risk of the primary end point in the placebo
77 mm Hg in the irbesartan group, 141/77 mm Hg and amlodipine groups did not differ significantly. The
in the amlodipine group, and 144/80 mm Hg in the proportions of patients in each group who reached
placebo group. The mean arterial pressure was signif- each of the three components of the primary end
icantly higher (by 3.3 mm Hg) in the placebo group point are shown in Figures 2B, 2C, and 2D. Among
than in the two active-treatment groups (P=0.001 for the patients assigned to irbesartan, the unadjusted
both comparisons), between which it did not differ relative risk of a doubling of the serum creatinine con-
significantly. The distribution of classes of nonstudy centration was 33 percent lower than that among the
drugs used to control blood pressure — primarily di- patients assigned to placebo (P=0.003) (Table 3)
uretics, beta-blockers, peripheral alpha-blockers, and and 37 percent lower than that among the patients
central a2 agonists — was similar in all groups. The assigned to amlodipine (P<0.001). The unadjusted
patients in the placebo group required an average of relative risk of end-stage renal disease (P=0.07) was
3.3 nonstudy drugs for the control of blood pressure, 23 percent lower in the irbesartan group than in ei-
as compared with an average of 3.0 nonstudy drugs ther the amlodipine group or the placebo group (Ta-
among the patients in the irbesartan and amlodipine ble 3). The placebo and amlodipine groups did not
groups. The mean glycosylated hemoglobin values did differ significantly with respect to the relative risk of
not differ significantly among the treatment groups or a doubling of the serum creatinine concentration or
vary significantly over time. of end-stage renal disease. There was no significant dif-
ference among the three groups in the unadjusted risk
Primary, Renal Outcome of death from any cause (Table 3).
The proportions of patients in each group who
reached the primary end point are shown in Figure Secondary, Cardiovascular Outcome
2A. The patients in the irbesartan group had an un- There were no significant differences among the
adjusted relative risk of reaching the primary end treatment groups in the secondary, cardiovascular out-
point that was 20 percent lower than that in the pla- come (Table 3). The patients assigned to receive ir-

N Engl J Med, Vol. 345, No. 12 · September 20, 2001 · www.nejm.org · 853

Downloaded from www.nejm.org on March 22, 2008 . Copyright © 2001 Massachusetts Medical Society. All rights reserved.
 The Ne w E n g l a nd Jo u r n a l o f Me d ic i ne

TABLE 2. OUTCOMES ACCORDING TO STUDY GROUP.*

IRBESARTAN AMLODIPINE PLACEBO ALL

GROUP GROUP GROUP PATIENTS
VARIABLE (N=579) (N=567) (N=569) (N=1715)

Primary composite outcome — 189 (32.6) 233 (41.1) 222 (39.0) 644 (37.6)
no. (%)
Doubling of serum creatinine 98 (16.9) 144 (25.4) 135 (23.7) 377 (22.0)
concentration
End-stage renal disease 82 (14.2) 104 (18.3) 101 (17.8) 287 (16.7)
Death from any cause 87 (15.0) 83 (14.6) 93 (16.3) 263 (15.3)
Secondary composite outcome — 138 (23.8) 128 (22.6) 144 (25.3) 410 (23.9)
no. (%)
Never received study medication 2 (0.3) 8 (1.4) 6 (1.1) 16 (0.9)
— no. (%)†
Lost to follow-up — no. (%)‡ 5 (0.9) 2 (0.4) 4 (0.7) 11 (0.6)
Completed study without primary 385 (66.5) 332 (58.6) 343 (60.3) 1060 (61.8)
outcome — no. (%)
Mean duration of follow-up — 952 924 921 932
days§

*The numbers of patients with the composite end points are lower than the sums of those with
the various components because some patients reached more than one component.
†Patients who never received study medication were included in all analyses according to the in-
tention-to-treat principle.
‡Data are for patients lost to follow-up before reaching the primary outcome or the last scheduled
study visit.
§The duration of follow-up was calculated from the time of randomization to the occurrence of
the first primary outcome or the date of the last scheduled visit when data were censored.

besartan had a rate of congestive heart failure neces- excretion were similar in the three groups at base line
sitating hospitalization that was 23 percent lower (Table 1). The serum creatinine concentration in-
than that among the patients assigned to receive pla- creased 24 percent more slowly in the patients in the
cebo. The patients assigned to receive amlodipine had irbesartan group than in those in the placebo group
a rate of nonfatal myocardial infarction that was 41 (P=0.008) and 21 percent more slowly than in those
percent lower than that among the patients assigned in the amlodipine group (P=0.02), despite the fact
to receive placebo. that larger numbers of patients in the placebo and
amlodipine groups were excluded from further analy-
Effect of Base-Line Covariates and Achieved Mean ses of the serum creatinine concentration when they
Arterial Pressure reached a renal end point. The serum creatinine slopes
Analyses were performed to ensure that the ob- in the placebo and amlodipine groups did not differ.
served differences in outcomes could not be explained The mean (±SE) absolute rates of change in the se-
by imbalances in the distribution of the base-line co- rum creatinine concentration were 0.45±0.04 mg per
variates. The inclusion of these base-line covariates in deciliter per year in the irbesartan group, 0.57±0.04
proportional-hazards analyses did not change the con- mg per deciliter per year in the amlodipine group, and
clusions of the primary analyses. 0.59±0.04 mg per deciliter per year in the placebo
The better renal outcomes in the irbesartan group group. The mean rate of change in creatinine clear-
could not be explained by differences in the mean ance was ¡5.5±0.36 ml per minute per 1.73 m 2 of
arterial blood pressure during follow-up. The mean body-surface area per year in the irbesartan group,
arterial pressure in the irbesartan group was not sig- ¡6.8±0.37 ml per minute per 1.73 m2 per year in the
nificantly different from that in the amlodipine group. amlodipine group, and ¡6.5±0.37 ml per minute per
Furthermore, when we corrected for the mean arte- 1.73 m 2 per year in the placebo group. Proteinuria
rial pressure at each of the quarterly visits in a time- was reduced on average by 33 percent (mean [±SD]
dependent proportional-hazards analysis (Table 3), the decrease in protein concentration, ¡1.1±1.7 g per
results were similar to those of the primary analysis. 24 hours) in the irbesartan group, as compared with
6 percent (¡0.1±2.9 g per 24 hours) in the amlo-
Changes in Renal Function dipine group and 10 percent (¡0.3±4.3 g per 24
The serum creatinine concentration, creatinine hours) in the placebo group. These reductions were
clearance, and levels of urinary protein and albumin maintained throughout the follow-up period.

854 · N Engl J Med, Vol. 345, No. 12 · September 20, 2001 · www.nejm.org

Downloaded from www.nejm.org on March 22, 2008 . Copyright © 2001 Massachusetts Medical Society. All rights reserved.
 E F F EC T OF IRBE SAR TAN ON NEPH ROPATH Y D UE TO T Y PE 2 D IA BETES

IrbesartanH
160 AmlodipineH
Placebo

Systolic
140
Blood Pressure (mm Hg)

120

Mean
100

Diastolic
80

0 6 12 18 24 30 36 42 48 54
Months of Follow-up
Figure 1. Average Systolic, Mean Arterial, and Diastolic Blood Pressures at Randomization (0 Months)
and during Follow-up, According to Treatment Group.
The mean arterial pressure during follow-up was, on average, 3.3 mm Hg lower in the irbesartan and
amlodipine groups than in the placebo group.

Serious Adverse Events DISCUSSION

One episode of an early increase in the serum cre- The angiotensin-II–receptor antagonist irbesartan
atinine concentration suggestive of renal-artery ste- was associated with better renal outcomes than the
nosis necessitated the stopping of the study medica- other agents (amlodipine, placebo, and antihyperten-
tion. Hyperkalemia necessitating a discontinuation of sive agents) we used. A slowing of the rate of pro-
the study medication occurred in 11 of the patients gression of nephropathy was reflected in a significant
in the irbesartan group (1.9 percent), as compared increase in the time to a doubling of the serum cre-
with 3 of those in the amlodipine group (0.5 per- atinine concentration, a measure that approximates a
cent) and 2 of those in the placebo group (0.4 per- halving of the glomerular filtration rate.3 A diminished
cent, P=0.01 for both comparisons). Overall, 23.7 rate of progression of disease was not limited to the
percent of the patients stopped receiving the study patients in whom there was a doubling of the serum
medication without having reached the primary end creatinine concentration during follow-up. The mean
point and before their data were censored. The most increase in the serum creatinine concentration and the
common reason for the discontinuation of the study mean decrease in creatinine clearance were signifi-
medication was the occurrence of a clinical cardio- cantly slower in the entire irbesartan group. The pa-
vascular event. These discontinuations were evenly dis- tients in the amlodipine group had worse renal out-
tributed among the treatment groups. The number of comes than those in the irbesartan group, although
patients who had at least one serious adverse event (61 there was equal control of blood pressure in the am-
percent of the total cohort) reflected the advanced lodipine group. When we adjusted for the disparity
stage of illness and the numerous risk factors in this in blood-pressure control between the irbesartan and
population, but the number did not differ significant- placebo groups, the extent of the estimated renal ben-
ly among groups. The patients in the irbesartan group efit of irbesartan did not decrease significantly. We in-
had a significantly lower rate of adverse events per terpret these results as demonstrating that irbesartan
1000 days of treatment than those in the placebo and was renoprotective in these patients with nephropa-
amlodipine groups (P=0.002). thy due to type 2 diabetes — an effect analogous to

N Engl J Med, Vol. 345, No. 12 · September 20, 2001 · www.nejm.org · 855

Downloaded from www.nejm.org on March 22, 2008 . Copyright © 2001 Massachusetts Medical Society. All rights reserved.
 The Ne w E n g l a nd Jo u r n a l o f Me d ic i ne

A
IrbesartanH
0.6 AmlodipineH
Placebo

Primary End Point

Proportion withH
0.5

0.4

0.3

0.2

0.1

0.0
0 6 12 18 24 30 36 42 48 54
Months of Follow-up
NO. AT RISK
IrbesartanH 579H 555H 528H 496H 400H 304H 216H 146H 65H
AmlodipineH 565H 542H 508H 474H 385H 287H 187H 128H 46H
Placebo 568 551 512 471 401 280 190 122 53

B
Proportion with a DoublingH

0.6
Creatinine Concentration
of Base-Line SerumH

0.5

0.4

0.3

0.2

0.1

0.0
0 6 12 18 24 30 36 42 48 54
Months of Follow-up
NO. AT RISK
IrbesartanH 579H 534H 495H 457H 363H 273H 191H 131H 57H 5H
AmlodipineH 567H 516H 476H 439H 347H 254H 166H 108H 40H 5H
Placebo 569 527 482 436 360 252 173 107 47 2

Figure 2. Cumulative Proportions of Patients with the Primary Composite End Point (Panel A) and Its
Components, a Doubling of the Base-Line Serum Creatinine Concentration (Panel B), End-Stage Renal
Disease (Panel C, facing page), and Death from Any Cause (Panel D, facing page).
The date of onset of end-stage renal disease could not be determined for one patient in the placebo
group and two patients in the amlodipine group. These three patients were excluded from the analyses
shown in Panels A and C.

that of the ACE inhibitor captopril in patients with conversion of angiotensin I to angiotensin II, as well
nephropathy due to type 1 diabetes.3 as for the catalytic degradation of bradykinin.14 The
The renoprotection provided by an angiotensin- renoprotection associated with ACE inhibition has
II–receptor antagonist derives solely from its restric- been shown in rats to be the result of diminished gen-
tion of angiotensin activity. ACE inhibition may be less eration of angiotensin and increased bradykinin con-
specific and less complete. ACE is responsible for the centrations.14 However, studies in humans have sup-

856 · N Engl J Med, Vol. 345, No. 12 · September 20, 2001 · www.nejm.org

Downloaded from www.nejm.org on March 22, 2008 . Copyright © 2001 Massachusetts Medical Society. All rights reserved.
 E F F EC T OF IRBE SAR TAN ON NEPH ROPATH Y D UE TO T Y PE 2 D IA BETES

C
IrbesartanH

Proportion with End-StageH

0.6 AmlodipineH
Placebo

0.5

Renal Disease
0.4

0.3

0.2

0.1

0.0
0 6 12 18 24 30 36 42 48 54
Months of Follow-up
NO. AT RISK
IrbesartanH 579H 549H 523H 501H 418H 327H 234H 162H 78H 7H
AmlodipineH 565H 538H 510H 482H 408H 310H 221H 152H 58H 7H
Placebo 568 542 517 487 418 302 205 141 63 2

D
0.6
Proportion with DeathH

0.5
from Any Cause

0.4

0.3

0.2

0.1

0.0
0 6 12 18 24 30 36 42 48 54
Months of Follow-up
NO. AT RISK
IrbesartanH 579H 563H 550H 530H 452H 355H 264H 196H 99H 10H
AmlodipineH 567H 552H 536H 524H 457H 358H 266H 201H 83H 9H
Placebo 569 553 539 522 465 354 255 185 94 7

ported the notion that ACE inhibitors alter renal teinuria,19 and decreased collagen formation20 that
hemodynamics primarily by diminishing the action may be related to decreased stimulation of transform-
of angiotensin II.15-17 Thus, in both the renoprotec- ing growth factor b by angiotensin II.21
tion demonstrated by ACE inhibition in nephropathy We cannot directly address the issue of whether the
due to type 1 diabetes and that of angiotensin-receptor effects of ACE inhibitors and angiotensin-receptor
blockers in nephropathy due to type 2 diabetes, the blockers would be equivalent in the treatment of pa-
important pharmacologic action appears to be the re- tients with nephropathy due to type 2 diabetes. It may
striction of intrarenal angiotensin activity. The mech- seem reasonable to assume that agents that primarily
anism of renoprotection by agents that block the ac- reduce the generation or effect of angiotensin II
tion of angiotensin II may be complex, involving would have similar clinical results. However, it is im-
hemodynamic factors that lower the intraglomerular portant to caution that ACE inhibitors and angio-
pressure,18 the beneficial effects of diminished pro- tensin-receptor blockers are distinctly different class-

N Engl J Med, Vol. 345, No. 12 · September 20, 2001 · www.nejm.org · 857

Downloaded from www.nejm.org on March 22, 2008 . Copyright © 2001 Massachusetts Medical Society. All rights reserved.
 The Ne w E n g l a nd Jo u r n a l o f Me d ic i ne

TABLE 3. RELATIVE RISKS OF OUTCOMES.*

UNADJUSTED RELATIVE RISK P ADJUSTED RELATIVE RISK P

OUTCOME (95% CI) VALUE (95% CI)† VALUE

Primary composite end point

Irbesartan vs. placebo 0.80 (0.66–0.97) 0.02 0.81 (0.67–0.99) 0.03
Amlodipine vs. placebo 1.04 (0.86–1.25) 0.69 1.07 (0.89–1.29) 0.47
Irbesartan vs. amlodipine 0.77 (0.63–0.93) 0.006 0.76 (0.63–0.92) 0.005
Doubling of serum creatinine
concentration
Irbesartan vs. placebo 0.67 (0.52–0.87) 0.003 0.71 (0.54–0.92) 0.009
Amlodipine vs. placebo 1.06 (0.84–1.35) 0.60 1.15 (0.91–1.46) 0.24
Irbesartan vs. amlodipine 0.63 (0.48–0.81) <0.001 0.61 (0.48–0.79) <0.001
End-stage renal disease
Irbesartan vs. placebo 0.77 (0.57–1.03) 0.07 0.83 (0.62–1.11) 0.19
Amlodipine vs. placebo 1.00 (0.76–1.32) 0.99 1.09 (0.82–1.43) 0.56
Irbesartan vs. amlodipine 0.77 (0.57–1.03) 0.07 0.76 (0.57–1.02) 0.06
Death from any cause
Irbesartan vs. placebo 0.92 (0.69–1.23) 0.57 0.94 (0.70–1.27) 0.69
Amlodipine vs. placebo 0.88 (0.66–1.19) 0.40 0.90 (0.66–1.21) 0.47
Irbesartan vs. amlodipine 1.04 (0.77–1.40) 0.80 1.05 (0.78–1.42) 0.75
Secondary, cardiovascular
composite end point
Irbesartan vs. placebo 0.91 (0.72–1.14) 0.40 0.91 (0.72–1.14) 0.40
Amlodipine vs. placebo 0.88 (0.69–1.12) 0.29 0.88 (0.69–1.11) 0.27
Irbesartan vs. amlodipine 1.03 (0.81–1.31) 0.79 1.03 (0.81–1.32) 0.78

*CI denotes confidence interval.

†The relative risks were adjusted for the mean arterial blood pressure during follow-up.

es of drugs and that one cannot assume equivalence hibition slowed the progression of renal disease.23 In
between them. The pharmacologic activity of these the light of conflicting information from previous
agents is complex. The effect of ACE inhibition on reports, we must limit our recommendations to those
renal hemodynamics could be limited by the non– that can be drawn from the results achieved with the
ACE-dependent generation of angiotensin II that has agent and dosage that we used in this study.
been documented in patients in the hyperglycemic It should be noted that, although the differences
state.17 It is noteworthy that the size-selective dysfunc- were not statistically significant, our irbesartan group
tion of glomerular capillary permeability that charac- had lower rates of death from any cause and of the
terizes diabetic nephropathy is improved by ACE in- secondary, cardiovascular end point than the placebo
hibition in patients with type 1 diabetes but not in group. These differences were slightly smaller than
those with type 2 diabetes.4 The Ramipril Efficacy the absolute differences in the risk of death from any
in Nephropathy study failed to demonstrate renopro- cause and the risk of other cardiovascular end points
tection in patients with nephropathy due to type 2 that were reported in the Heart Outcomes Prevention
diabetes who received ACE inhibitors.22 Patients who Evaluation Study.24 However, the statistically robust
received ramipril lost renal function at a significantly result reported for that study was the product of its
faster rate than those assigned to other antihyperten- statistical power for the detection of differences in
sive agents.22 A subanalysis of patients in the United mortality with a total of 9297 patients, 3577 of whom
Kingdom Prospective Diabetes Study concluded that had diabetes. In our study, in which the sample was
ACE inhibitors and b-adrenergic–blocking agents much smaller, we found no statistically significant dif-
were equally effective in preventing renal damage.8 ferences in the rate of death from any cause or the
Since b-adrenergic–blocking agents were commonly secondary, cardiovascular end point. Our study was
used in our placebo group, one must note the con- not designed to have adequate statistical power for
trast between the results of the United Kingdom Pro- an analysis of these outcomes.
spective Diabetes Study and our results. The popula- The amlodipine group had a significantly higher
tion at risk in that study, however, was small, and the rate of congestive heart failure than the placebo or
comparison between ACE inhibitors and beta-block- irbesartan group. This finding is in keeping with the
ers was not part of its primary design. One study of moderately increased risk of congestive heart failure
the course of proteinuria in patients with microalbu- associated with calcium-channel blockers reported in
minuria and type 2 diabetes indicated that ACE in- a recent meta-analysis that focused on hypertensive

858 · N Engl J Med, Vol. 345, No. 12 · September 20, 2001 · www.nejm.org

Downloaded from www.nejm.org on March 22, 2008 . Copyright © 2001 Massachusetts Medical Society. All rights reserved.
 E F F EC T OF IRBE SAR TAN ON NEPH ROPATH Y D UE TO T Y PE 2 D IA BETES

patients without diabetes.25 Our report of a lower Breen, N. Tolchin, R. Toto, K. Tuttle, J.B. Tyzack, M. Velasquez, H. Ward,
E. Wedell, M. Weir, T. Wiegmann, C. Wilcox, B. Wood, D. Young, M.
risk of coronary events associated with calcium-chan- Ziegler, F.J. Zieve; Canada — G.R. Bailey, R. Bebb, G. Bondy, D. Cattran,
nel–blocker therapy differs from the results of Hans- P. Hamet, S.P. Handa, O. Ojo, G. Pylypchuk, S.A. Ross, P. Tam, H. Til-
son et al. in a larger trial in elderly hypertensive pa- desley, R. Ting, W. Vlahos, B.M. Wolfe; Argentina — L.F. Ferder, F. Margu-
lis, A. Zucchini; Brazil — F.A. de Almeida, J.B. Lopes de Faria, R. Milagres,
tients, which found no difference between therapy W. Oigman, V. Pavan, M.T. Zanella, R. Zatz; Mexico — R. Correa-Rotter,
with a calcium-channel blocker and therapy with an M. Sieiro-Muradas; Puerto Rico — J. Benabe, J.L. Cangiano; Austria — G.
Biesenbach, G. Schernthaner, W. Waldhäausl; Belgium — B. Keymeulen, L.
ACE inhibitor with regard to the risk of coronary Van Gaal, X. Warling; Denmark — H. Perrild; Finland — C. Grönhagen-
events.26 It is possible that calcium-channel blockers Riska; the Netherlands — K. Hoogenberg, P.F.M.J. Spooren, R.P. Ver-
are more efficacious in lowering the rate of coronary hoeven; France — P. Bataille, J. Chanard, F. Chantrel, G. Charpentier, B.
Coevoet, D.J. Cordonnier, P. Drouin, N. El Esper, A. Fournier, T.
events in a population with diabetic nephropathy. Hannedouche, M. Kessler, G. Lambrey, J. Mahoudeau, C. Maynard, F.
Our data reveal that irbesartan is renoprotective in Mignon-Henrion, P. Vanhille, P. Zaoui; Germany — W. Beischer, K. Bergis,
patients with type 2 diabetes and overt nephropathy G.P. Dragoun, H.H. Echterhoff, C. Hammermeister, W. Kleophas, G.
Kraatz, G.A. Müller, E. Standl, G. Stein, C. Wanner, P. Weisweiler; Hun-
and that it significantly slows the progression of glo- gary — G. Bibok, K. Farkas, G. Jermendy, G. Kakuk, L. Kammerer, J.
merulopathy. The beneficial effects of irbesartan were Nagy, G. Tamas; Israel — I. Raz; Italy — V.E. Andreucci, S. Bandinelli, N.
Bellotti, G.A. Cinotti, G. Crepaldi, A. Dal Canton, G. DeFerrari, E. Degli
accompanied by minimal drug-specific serious adverse Esposti, C. Esposito, F. Locatelli, U. Malcangi, R. Mangili, G. Maschio, R.
effects in our patients. Navalesi, G. Norbiato, G. Penno, G. Piccoli, G. Pozza, F. Quarello, A. Ra-
mello, F.P. Schena, A. Sessa, S. Stefoni, C. Zoccali; Poland — S. Czekalski,
J. Drezwoski, H. Fuchs, W. Grzeszczak, I. Kinalska, K. Pynka, J. Sieradzki,
Supported by the Bristol-Myers Squibb Institute for Medical Research
M. Snit, K. Strojek; Portugal — P.L. Neves, C. Pires; Spain — J. Bonet, F.
and Sanofi–Synthelabo.
De Alvaro, E. Esmatjes, R. Marin Iranzo, J.M. Mauri, R. Romero, L. Rui-
Dr. Edmund Lewis has received research grants from Bristol-Myers Squibb
lope, J.A. Vazquez; Sweden — H. Arnqvist, S. Björck, B. Eliasson, L.
and has been a member of speakers’ bureaus sponsored by Bristol-Myers
Hulthen; United Kingdom — R. Bilous, T.L. Dornan, P. Grant, A. Gren-
Squibb.
fell, D. Hopkins, J.N. Harvey, S.R. Heller, J. New, N.W. Oakley, A. Panah-
Dr. Hunsicker has received research grants from Bristol-Myers Squibb and
loo, P.B. Rylance, H. Simpson, H. Tindall, N.J.A. Vaughn, J.P. Vora; Aus-
has been a member of speakers’ bureaus sponsored by Bristol-Myers Squibb.
tralia — P.J. Champion de Crespigny, M.C. D’Emden, R. Gilbert, B.
Dr. Berl has received research grants from Pfizer and has been a member
Jackson, G. Jerums, J. Kelly, P. Kerr, P. O’Connell, P. Phillips, S. Roger, D.
of speakers’ bureaus sponsored by Bristol-Myers Squibb and Pfizer.
Saltissi, J. Whitworth, D. Yue; New Zealand — W. Bagg, J. Baker, G.
Dr. Pohl been a member of speakers’ bureaus sponsored by Bristol-Myers
Braatvedt, P. Drury, P. Manning, D. Simmons; Hong Kong — J. Chan,
Squibb and Pfizer.
T.M.D. Chan; Malaysia — N. Kong; Singapore — A.C. Thai; Taiwan —
Dr. Ritz has been a member of speakers’ bureaus sponsored by Bristol-
W.C. Yang.
Myers Squibb.
Dr. Atkins has received research grants from Bristol-Myers Squibb.
Dr. Rohde has received research grants from Bristol-Myers Squibb.
REFERENCES
1. Mokdad AH, Ford ES, Bowman BA, et al. Diabetes trends in the U.S.:
We are indebted to Deborah Anzalone, M.D., and her staff at 1990-1998. Diabetes Care 2000;23:1278-83.
Bristol-Myers Squibb for their support. 2. US Renal Data System. USRDS 1999 annual data report. Bethesda,
Md.: National Institutes of Health, National Institute of Diabetes and Di-
APPENDIX gestive and Kidney Diseases, 1999:25-38.
3. Lewis EJ, Hunsicker LG, Bain RP, Rohde RD. The effect of angioten-
The following persons participated in the Collaborative Study Group tri- sin-converting–enzyme inhibition on diabetic nephropathy. N Engl J Med
al: Clinical Coordinating Center: E.J. Lewis, T. Berl, R.D. Rohde, E. 1993;329:1456-62.
Muskwe; European Clinical Coordinating Center: E. Ritz, L. Ruilope, P. 4. Ruggenenti P, Mosconi L, Sangalli F, et al. Glomerular size-selective
Klooker, B. Spiller; Pacific Clinical Coordinating Center: R. Atkins, G. dysfunction in NIDDM is not ameliorated by ACE inhibition or by calci-
Jerums, R. Bartholomeusz; European Country Coordinators: R.W. Bilous, um channel blockade. Kidney Int 1999;55:984-94.
L. Hulthen, S. Björck, D.J. Cordonnier, G. DeFerrari, L. Ruilope, G. 5. Curb JD, Pressel SL, Cutler JA, et al. Effect of diuretic-based antihy-
Tamas, L.F. Van Gaal; Biostatistical Coordinating Center: L.G. Hunsicker, pertensive treatment of cardiovascular disease risk in older diabetic patients
W.R. Clarke; Executive Committee: E.J. Lewis, R. Atkins, E. Ritz, T. Berl, with isolated systolic hypertension. JAMA 1996;276:1886-92.
G. Jerums, L. Ruilope, R. Bilous, S. Blumenthal, W. Clarke, D.J. Cordon- 6. Tuomilehto J, Rastenyte D, Birkenhäger WH, et al. Effects of calcium-
nier, D. Hricik, L.G. Hunsicker, P. Klooker, J. Lewis, O. Ojo, M. Pfeffer, channel blockade in older patients with diabetes and systolic hypertension.
M.A. Pohl, J.G. Porush, I. Raz, R.A. Rodby, T.B. Wiegmann; Clinical N Engl J Med 1999;340:677-84.
Management Committee: M.A. Pohl, D.J. Cordonnier, S. Bjorck, S. Blu- 7. Hansson L, Zanchetti A, Carruthers SG, et al. Effects of intensive
menthal, W. Clarke, F. De Alvaro, G. Deferrari, R. Gilbert, L.G. Hunsicker, blood-pressure lowering and low-dose aspirin in patients with hyperten-
P. Klooker, J.B. Lopes de Faria, R. Mangili, E. Reisin, R.A. Rodby, G. sion: principal results of the Hypertension Optimal Treatment (HOT) ran-
Schernthaner, S. Spitalewitz, H. Tindall; Outcomes Confirmation and domised trial. Lancet 1998;351:1755-62.
Classification Committee: T. Berl, P. Drury, E. Esmatjes, D. Hricik, J. 8. UK Prospective Diabetes Study Group. Efficacy of atenolol and capto-
Lewis, F. Locatelli, G. Porush, I. Raz, L. Ruilope, K. Strojek, R. Toto, P. pril in reducing risk of macrovascular and microvascular complications in
Vanhille, T.B. Wiegmann, B.M. Wolfe; Mortality Committee: T. Berl, S.Z. type 2 diabetes: UKPDS 39. BMJ 1998;317:713-20.
Goldhaber, A. Levey, J. Lewis, M. Pfeffer, J.G. Porush, J.L. Rouleau; Data 9. Rodby RA, Rohde RD, Clarke WR, et al. The Irbesartan Type II Dia-
and Safety Monitoring Committee: S. Klahr, L. Fisher, J. Roseman, M.W. betic Nephropathy Trial: study design and baseline patient characteristics.
Steffes, K. Vaux; Collaborating Investigators: United States — R.J. Ander- Nephrol Dial Transplant 2000;15:487-97.
son, S. Anderson, S. Aronoff, S.A. Atlas, P. Barnett, T. Berl, E.M. Ben- 10. Lee ET. Statistical methods for survival analysis. New York: John
jamin, M. Block, S. Blumenthal, B. Bresnahan, G.L. Braden, J.F. Burke, Jr., Wiley, 1992:67-78, 105-7.
G. Collins, J. Cyrus, W. Dachman, P. Dandona, R.A. Dart, J.A. Dicke, T. 11. Idem. Statistical methods for survival analysis. New York: John Wiley,
Dixon, M. Doyle, T. Duncan, L. Dworkin, D. Einhorn, G. Eisner, W.T. El- 1992:250-63.
lison, M. Falkenhain, J. Fialkow, P. Fisher, E. Fishman, R. Garrick, S. Gold- 12. Woolson RF. Statistical methods for the analysis of biomedical data.
man, R. Goldstein, R. Hegeman, R.E. Heinig, W.J. Henry III, D. Hricik, New York: John Wiley, 1987:227-9.
R. Kaplen, J. Kaufman, S. Kirkman, V.A. Klimas, N. Kopyt, N. Kramer, B. 13. Lan KKG, DeMets DL. Discrete sequential boundaries for clinical tri-
Levine, J. Lewis, R. Lipetz, P. Lodewick, J. Lohr, C. Manion, J.H. Mersey, als. Biometrika 1983;70:659-63.
R. Michaels, J. Middleton, J.M. Milas, J. Moore, Jr., M. Moore, J. Navarro, 14. Kon V, Fogo A, Ichikawa I. Bradykinin causes selective efferent arte-
L. Nelson, M.H. O’Shea, C. Ott, M.A. Pohl, J. Porush, D. Price, S.N. Rah- riolar dilation during angiotensin I converting enzyme inhibition. Kidney
man, E. Reisin, R.A. Rodby, F.D. Rogoff, S.J. Rosansky, C. Rosendorff, Int 1993;44:545-50.
M. Ruddy, M. Rudnick, D. Ruff, J. Rydel, G. Schulman, A. Schwartz, S. 15. Price DA, Dé Oliveira JM, Fisher NDL, Hollenberg NK. Renal he-
Schwartz, K. Sethi, K. Shapiro, K. Sharma, G. Shaw, D. Sherrard, D. Sica, modymanic response to an angiotensin II antagonist, eprosartan, in healthy
H.J. Simon, R.T. Simon, R. Smith, B. Snyder, S. Spitalewitz, C. Stehman- men. Hypertension 1997;30:240-6.

N Engl J Med, Vol. 345, No. 12 · September 20, 2001 · www.nejm.org · 859

Downloaded from www.nejm.org on March 22, 2008 . Copyright © 2001 Massachusetts Medical Society. All rights reserved.
 The Ne w E n g l a nd Jo u r n a l o f Me d ic i ne

16. Price DA, Porter LE, Gordon M, et al. The paradox of the low-renin 22. Ruggenenti P, Perna A, Gherardi G, Benini R, Remuzzi G. Chronic
state in diabetic nephropathy. J Am Soc Nephrol 1999;10:2382-91. proteinuric nephropathies: outcomes and response to treatment in a pro-
17. Osei SY, Price DA, Laffel LMB, Lansang MC, Hollenberg NK. Effect spective cohort of 352 patients with different patterns of renal injury. Am
of angiotensin II antagonist eprosartan on hyperglycemia-induced activa- J Kidney Dis 2000;35:1155-65.
tion of intrarenal renin-angiotensin system in healthy humans. Hyperten- 23. Ravid M, Savin H, Jutrin I, Bental T, Katz B, Lishner M. Long-term
sion 2000;36:122-6. stabilizing effect of angiotensin-converting enzyme inhibition on plasma
18. Zatz R , Dunn BR, Meyer TW, Anderson S, Rennke HG, Brenner creatinine and on proteinuria in normotensive type II diabetic patients.
BM. Prevention of diabetic glomerulopathy by pharmacological ameliora- Ann Intern Med 1993;118:577-81.
tion of glomerular capillary hypertension. J Clin Invest 1986;77:1925-30. 24. The Heart Outcomes Prevention Evaluation Study Investigators. Ef-
19. Abbate M, Zoja C, Corna D, Capitanio M, Bertani T, Remuzzi G. In fects of an angiotensin-converting–enzyme inhibitor, ramipril, on cardio-
progressive nephropathies, overload of tubular cells with filtered proteins vascular events in high-risk patients. N Engl J Med 2000;342:145-53.
translates glomerular permeability dysfunction into cellular signals of inter- 25. Neal B, MacMahon S, Chapman N. Effects of ACE inhibitors, calcium
stitial inflammation. J Am Soc Nephrol 1998;9:1213-24. antagonists and other blood pressure lowering drugs: results of prospec-
20. Wolf G, Haberstroh U, Neilson EG. Angiotensin II stimulates the pro- tively designed overviews of randomized trials. Lancet 2000;356:1955-64.
liferation and biosynthesis of type I collagen in cultured murine mesangial 26. Hansson L, Lindholm L, Ekbom T, et al. Randomised trial of old and
cells. Am J Pathol 1992;140:95-107. new antihypertensive drugs in elderly patients: cardiovascular mortality and
21. Kagami S, Border WA, Miller DE, Noble NA. Angiotensin II stimu- morbidity the Swedish Trial in Old Patients with Hypertension-2 study.
lates extracellular matrix protein synthesis through induction of transform- Lancet 1999;354:1751-6.
ing growth factor-beta expression in rat glomerular mesangial cells. J Clin
Invest 1994;93:2431-7. Copyright © 2001 Massachusetts Medical Society.

860 · N Engl J Med, Vol. 345, No. 12 · September 20, 2001 · www.nejm.org

Downloaded from www.nejm.org on March 22, 2008 . Copyright © 2001 Massachusetts Medical Society. All rights reserved.