RKihprebsiTtfii

M

Tg

Fc

3

0d

Effect of Short-Term Rosuvastatin Treatment on EstimatedGlomerular Filtration Rate

Donald G. Vidt, MDa,*, Susan Harris, MDb, Fergus McTaggart, PhDc, Marc Ditmarsch, MDc,Philip T. Sager, MDc, and Jonathan M. Sorof, MDc

To define the effect of short-term rosuvastatin treatment on the estimated glomerularfiltration rate (eGFR), the database of controlled clinical trials in the Rosuvastatin ClinicalDevelopment Program was reviewed. Thirteen studies comprising 3,956 rosuvastatin-treated patients were selected based on a serum creatinine measurement at 6 or 8 weeksafter initiation of rosuvastatin treatment, randomization to approved and marketed rosu-vastatin doses (5 to 40 mg), and unchanged rosuvastatin dose from treatment initiation(baseline) through 6 to 8 weeks of treatment. eGFR was determined with the Modificationof Diet in Renal Disease formula. eGFR significantly increased for each dose of rosuvas-tatin individually and for all doses combined compared with baseline (range �0.9 to �3.2ml/min/1.73 m2). Further analysis of 5 blinded, placebo-controlled trials comprising 525patients showed an increase in eGFR of �0.8 ml/min/1.73 m2 (95% confidence interval�0.1 to �1.5) for all rosuvastatin-treated patients, which was significantly different frombaseline (p <0.04) and from a change of �1.5 ml/min/1.73 m2 in the placebo-treatedpatients (95% confidence interval �2.5 to �0.5, p <0.001). The increase in eGFR forrosuvastatin-treated patients was consistent across all major demographic and clinicalsubgroups of interest, including patients with baseline proteinuria, baseline eGFR <60ml/min/1.73 m2, and in patients with hypertension and/or diabetes. In conclusion, theseresults are consistent with previous rosuvastatin studies that showed an upward trend ineGFR with long-term treatment (>96 weeks) and with the hypothesis that statins may havepleiotropic mechanisms of action that include beneficial renal effects. © 2006 Elsevier

Inc. All rights reserved. (Am J Cardiol 2006;97:1602–1606)

Dactpwc

crcoMmcIsallt�oCb

osuvastatin (Crestor, AstraZeneca, Alderley Park, Unitedingdom) has been shown to be an effective statin for decreas-

ng low-density lipoprotein (LDL) cholesterol and increasingigh-density lipoprotein cholesterol.1 Vidt et al2 recently re-orted that long-term treatment with rosuvastatin (�96 weeks)esults in an increase of approximately 4 ml/min/1.73 m2 in thestimated glomerular filtration rate (eGFR) compared withaseline.2 This same study also found a benefit on eGFR withhort-term treatment, but analysis of the short-term studiesncluded a wide range of treatment duration, i.e., 6 to 52 weeks.o more clearly define the effect of short-term rosuvastatin

reatment on eGFR, an analysis of controlled clinical trialsrom the Rosuvastatin Clinical Development Database, whichncluded a creatinine measurements at 6 to 8 weeks afternitiation of treatment was conducted.

ethods

he trials in the Rosuvastatin Clinical Development Pro-ram were designed and conducted in accordance with the

aThe Department of Nephrology and Hypertension, Cleveland Clinicoundation, Cleveland, Ohio; bAstraZeneca, Wilmington, Delaware; andAstraZeneca, Alderley Park, United Kingdom. Manuscript received June0, 2005; revised manuscript received and accepted December 9, 2005.

* Corresponding author: Tel: 216-444-6765; fax: 216-444-9378.

aE-mail address: vidtd@ccf.org (D.G. Vidt).

002-9149/06/$ – see front matter © 2006 Elsevier Inc. All rights reserved.oi:10.1016/j.amjcard.2005.12.052

eclaration of Helsinki (versions amended October 1996nd October 2000) and in compliance with the ethical prin-iples of good clinical practice. Appropriate ethics commit-ees or institutional review boards approved the researchrotocols, and all patients (or their guardians) gave theirritten, informed consent before initiation of any trial pro-

edure.The Rosuvastatin Clinical Development Program in-

luded a combination of placebo-controlled trials, dose-anging trials, and trials designed to compare the LDLholesterol–lowering efficacy of rosuvastatin with that ofther statins (i.e., atorvastatin, simvastatin, or pravastatin).en or women who had primary hypercholesterolemia (fa-ilial or nonfamilial), primary hypertriglyceridemia, or a

ombined hyperlipidemic state were included in the trials.nclusion criteria varied between studies, including specificelection criteria, such as hypertriglyceridemia, diabetes,nd African-American or South Asian populations. Baselineipid criteria differed from study to study. No upper ageimit was imposed, a broad range of concomitant medica-ions was allowed, and patients with serum creatinine levels

220 �mol/L (�2.5 mg/dl), stable diabetes, hypertension,r cardiovascular disease were included in the program.o-morbid conditions were reported by investigators andased on case report form entries that defined hypertension

s a systolic blood pressure �140 mm Hg and/or a diastolic

www.AJConline.org

bsat

msccre1c(ctwfi

a(tHtrft

cFcefrc1taweiy

rerr

R

Tvtthw

wesa

pCwh

F(t

TBt

AWMWR

BB

B

B

HDC

1603Preventive Cardiology/Renal Effects of Rosuvastatin

lood pressure �90 mm Hg or current use of antihyperten-ion medication. Coronary artery disease was documenteds coronary heart disease, peripheral artery disease, or his-ory of coronary artery disease or atherosclerosis.

To determine the effects of short-term rosuvastatin treat-ent on eGFR, studies from within this trial database were

elected to be pooled for further analyses based on specificriteria. Selection criteria for these trials were (1) serumreatinine measurement at baseline and at 6 or 8 weeks afterosuvastatin initiation to allow calculation of change inGFR; (2) randomization to fixed rosuvastatin doses of 5,0, 20, or 40 mg from baseline through the 6- or 8-weekreatinine measurement (nonresponse-based designs); and3) all necessary demographic and clinical data available foralculation of eGFR. For the fixed-dose trials included inhe analyses, dose decreases were not permitted. Patientsho, in the judgment of the investigator, did not tolerate thexed dose for any reason were withdrawn from the study.

A central laboratory performed all scheduled laboratoryssessments for patients in the rosuvastatin phase II/III trialsMedical Research Laboratories International, with facili-ies in Zaventem, Belgium, for all European centers andighland Heights, Kentucky, for North American and Aus-

ralian centers). The 2 laboratories used identical analyzers,eagents, calibrations, and quality control material, and splitresh samples were analyzed in the laboratories every weeko ensure continued generation of identical results.

Assessment of renal laboratory parameters used methodsonsistent with those described in the National Kidneyoundation’s clinical practice guidelines for the evaluation,lassification, and stratification of patients with kidney dis-ase. Routine urine dipstick assessments were obtained be-ore patients entered a trial. Urine dipstick results wereeported as none, trace, 1�, 2�, 3�, or 4�, whichorresponded to urinary protein concentrations of �10,0 to 20, �30, �100, �300, and �2,000 mg/dl, respec-ively. Serum creatinine levels were measured with thelkaline picrate method of Jaffe.3 eGFR was determinedith the abbreviated Modification of Diet in Renal Dis-

ase (MDRD) formula4 (MDRD � 186 � [serum creat-nine in milligrams per deciliter]�1.154 � [age inears]�0.203 � [0.742 if a woman] � [1.210 if black]).

Descriptive and inferential statistics are provided in thiseport. Paired t tests were performed on the mean changes inGFR for the rosuvastatin versus the placebo group. Theelation between change in eGFR and change in lipid pa-ameters was assessed by Pearson’s correlation coefficients.

esults

he dataset used for the analyses contains 13 pooled, rosu-astatin studies consisting of clinical data from 3,956 pa-ients. From these studies, 2% of rosuvastatin-treated pa-ients and 3% of placebo-treated patients withdrew beforeaving a creatinine measurement at 6 or 8 weeks and therefore

ere not included in the GFR analysis. None of these patients c

ithdrew due to renal adverse events. At the time of adversevent withdrawal, only 1 rosuvastatin-treated patient had aerum creatinine above the upper limit of normal, which hadlso been above the upper limit of normal at study entry.

Baseline demographic and clinical characteristics of theatients included are listed in Table 1. Most patients wereaucasian (79%) and �65 years of age (72%). Men andomen were equally represented. Forty-nine percent hadypertension, and 18% had diabetes. Based on the MDRD

igure 1. Mean changes in eGFR (bars) with 95% confidence intervalswhiskers) for each rosuvastatin dose. Numbers across the bottom representhe patient number for each dose group. *p �0.01 versus baseline.

able 1aseline demographics and clinical characteristics of 3,956 rosuvastatin-

reated patients

ge (yrs) 57 � 12eight (kg) 83 � 18en 2,014 (51%)omen 1,942 (49%)

aceCaucasian 79%Black 12%Hispanic 2%Asian 4%Other 2%aseline eGFR (ml/min/1.73 m2) 69 � 13aseline renal function (ml/min/1.73 m2)�90 6%60–89 71%30–59 23%�30 0%aseline dipstick urine protein�� 69%�� 5%Not done 27%aseline lipids (mg/dl)LDL cholesterol (n � 2,849) 192 � 37Total cholesterol (n � 2,850) 279 � 37High-density lipoprotein cholesterol (n � 2,850) 49 � 12ypertension 49%iabetes mellitus 18%oronary artery disease 30%

Values are means � SDs or numbers of patients (percentages).

alculations, baseline eGFR was categorized as normal

(m5eecdh

aedi4Budbpegb

wglw

ett5rstbowptcdcd1i

c

TBt

V

AWMWR

B

HDCB

e

B

TC

TA

R

G

H

D

B

B

1604 The American Journal of Cardiology (www.AJConline.org)

�90 ml/min/1.73 m2) in 6%, mildly decreased (60 to 89l/min/1.73 m2) in 71%, and moderately decreased (30 to

9 ml/min/1.73 m2) in 23%. No patients had a baselineGFR �30 ml/min/1.73 m2, which is consistent with thexclusion of patients with baseline serum creatinine con-entrations �2.5 mg/dl at study entry. Among patients withata concerning baseline dipstick proteinuria (n � 2,888), 7%ad �1� and 1% had �2� proteinuria at study entry.

Changes in eGFR for rosuvastatin doses of 5 to 40 mgre shown in Figure 1. eGFR significantly increased forach dose of rosuvastatin individually (p �0.01) and for alloses combined compared with baseline (p �0.01). Thencrease in eGFR ranged from �0.9 ml/min/1.73 m2 for the0-mg dose to �3.2 ml/min/1.73 m2 for the 20-mg dose.ecause patients were randomized by dose within individ-al trials but not across the overall clinical program, someemographic and clinical characteristics differed at baselineetween dose groups. In particular, the prevalence of blackatients in the 40-mg group (3%) was approximately sev-nfold lower than that in the combined 5-, 10-, and 20-mgroups (20%). Further, the prevalence of patients with a

able 2aseline demographic and clinical characteristics for placebo-controlled

rials

ariable Rosuvastatin(n � 377)

Placebo(n � 148)

ge (yrs) 58 � 10 59 � 10eight (kg) 81 � 18 82 � 18en 56% 54%omen 44% 46%

aceCaucasian 81% 87%Black 1% 0%Hispanic 2% 1%Asian 17% 12%aseline lipids (mg/dl)LDL cholesterol 155 � 40 148 � 38Total cholesterol 252 � 36 247 � 35High-density lipoprotein cholesterol 44 � 12 42 � 11Triglycerides 276 � 153 315 � 164ypertension 51% 64%iabetes mellitus 40% 61%oronary artery disease 19% 24%aseline dipstick urine protein�� 89% 80%�� 10% 19%��� 96% 92%��� 3% 7%Not done 1% 1%GFR (ml/min/1.73 m2)Baseline 71 � 17 69 � 16Change from baseline �0.8 � 6.9 �1.5 � 6.1aseline renal function (ml/min/1.73 m2)�90 13% 10%60–89 60% 61%30–59 27% 28%�30 0% 1%

Values are means � SDs or percentages of patients.

aseline eGFR �60 ml/min/1.73 m2 in the 40-mg group (

as 20% compared with 26% for combined lower doseroups. Other notable differences in the 40-mg group wereower prevalences of diabetes and hypertension comparedith the lower dose groups.To further define the effect of rosuvastatin on changes in

GFR, additional analyses were performed for the subset ofrials that were designed to be blinded and placebo con-rolled. Five of the 13 trials met these criteria, consisting of25 of the 3,956 total patients. These trials included severalosuvastatin doses for comparison with placebo, which re-ulted in an unequal number of patients receiving rosuvas-atin versus placebo. Table 2 presents the comparison ofaseline demographic and clinical factors for this subgroupf rosuvastatin- and placebo-treated patients, who, overall,ere similar to each other and to the overall population ofatient characteristics listed in Table 1. Placebo-treated pa-ients had higher prevalences of hypertension, diabetes, andoronary artery disease. After pooling all rosuvastatinoses, the increase in eGFR was 0.8 ml/min/1.73 m2 (95%onfidence interval �0.1 to �1.5), which was significantlyifferent from baseline (p �0.04) and from the decrease of.5 ml/min/1.73 m2 in the placebo group (95% confidencenterval �2.5 to �0.5, p �0.001).

Change in eGFR for the 13 trials that met the inclusionriteria was analyzed for specific subgroups of interest

able 3hange in estimated glomerular filtration rate by subgroups

No. ofPatients

Mean � SD 95% CI

otal 3,956 1.8 � 8 1.6–2.0ge (yrs)�65 2,850 2 � 8.2 1.7–2.3�65 1,106 1.3 � 7.4 0.9–1.7aceCaucasian 3,135 1.3 � 7.5 1.0–1.6Black 486 5.5 � 9.7 4.6–6.4Hispanic 93 1.9 � 7.7 0.3–3.5Asian 155 2 � 9.2 0.6–3.4Other 87 �0.4 � 7.3 �1.9–1.1enderMen 2,014 1.9 � 8 1.6–2.2Women 1,942 1.7 � 8 1.3–2.1ypertensive statusPresent 1,936 1.7 � 8 1.3–2.1Absent 2,020 2 � 8 1.7–2.3iabetes statusPresent 700 1.4 � 8 0.8–2.0Absent 3,256 1.9 � 8 1.6–2.2Hypertension � diabetes 517 0.9 � 8 0.2–1.6aseline eGFR (ml/min/1.73 m2)�60 3,036 1.6 � 8.4 1.3–1.9�60 920 2.8 � 6.4 2.4–3.2aseline urinary protein�� 2,727 2.1 � 8.4 1.8–2.4�� 179 1.6 � 9.4 0.2–3.0��� 2,852 2.1 � 8.5 1.8–2.4��� 54 1 � 6.7 �0.8–2.8

CI � confidence interval.

Table 3). Overall, results showed a consistent trend for

ehbisidpFmmm

lrbhgNe

D

Toshdcptaspssvfi(pcfii

ccdTihrspopf

fL

csdbsetNmtepoorhwil

stot1dcodettmicabtcr

useteialpwb

1605Preventive Cardiology/Renal Effects of Rosuvastatin

GFR increases across all subgroups, including age, gender,ypertension status, diabetes status, baseline eGFR, andaseline urinary protein (dipstick assessment). Of particularnterest, patients with baseline evidence of renal diseasehowed the same overall benefit. Specifically, eGFR signif-cantly increased or tended to increase irrespective of baselineipstick proteinuria status (�1� or 2� proteinuria) or im-aired renal function at baseline (eGFR �60 ml/min/1.73 m2).urther, patients with diabetes (n � 700) showed a �1.4l/min/1.73 m2 increase in eGFR. Black patients showed aean increase in eGFR of �5.5 ml/min/1.73 m2, approxi-ately threefold higher than that of the total population.The relation between change in eGFR and change in

ipid parameters was also analyzed. Specifically, for eachosuvastatin dose and for all doses pooled, the correlationetween change in eGFR and change in LDL cholesterol,igh-density lipoprotein cholesterol, total cholesterol, tri-lycerides, and apolipoproteins-A1 and -B was determined.o significant correlations were found between change in

GFR and any component of the lipid profile.

iscussion

he present study, which investigated the short-term effectsf statins on renal function, was conducted using the largesttatin clinical database available to date. The results show aighly significant increase in eGFR from baseline at eachose of rosuvastatin after only 6 to 8 weeks of treatment inonjunction with a decrease in eGFR in patients treated withlacebo. Although retrospective in nature, the strength ofhese analyses include a consistent duration of treatmentcross all studies that were analyzed, randomized and con-istent dose allocation for each patient during the treatmenteriod under study, a single central laboratory using theame method for assaying serum creatinine for every mea-urement, and the use of the most contemporary and bestalidated method for estimating GFR. The validity of thesendings is further supported by a level of significancep �0.01) that is meaningful even with the multiple com-arisons that were performed. Overall, these findings areonsistent with previous studies that have reported a bene-cial effect of statins on renal function that appear to be

ndependent of the extent of lipid lowering.5

Overall, the finding of an increase in eGFR was generallyonsistent across all major subgroups of interest, includingategorization by age, gender, race, hypertension status,iabetes status, baseline eGFR, and baseline proteinuria.his consistency of effect across subgroups is particularly

mportant because many patients with statin indicationsave co-morbid conditions that predispose to progressiveenal disease. Patients with a baseline proteinuria of �2�howed a mean eGFR increase, as did patients with im-aired baseline renal function (�60 ml/min/1.73 m2). Thebservation that mean eGFR also increased in 700 diabeticatients is of particular interest because current guidelines

or the treatment of diabetic patients create a de facto need p

or high-efficacy statin therapy to achieve recommendedDL cholesterol goals.

Although trials were selected for the present study byriteria that included fixed, randomized rosuvastatin doses,pecific design features of the patient trials with respect toosing regimens and target populations resulted in differentaseline characteristics of the dose groups. These differencesubstantially confound a dose-response analysis for changes inGFR. Specifically, race and baseline eGFR values were foundo be important determinants of the extent of eGFR increase.otably, black patients had a greater than threefold higherean eGFR increase compared with the overall study popula-

ion. The 20-mg dose group, which demonstrated the great-st eGFR increase, had the highest prevalence of blackatients (25%), whereas the 40-mg dose group, which dem-nstrated the least eGFR increase, had the lowest prevalencef black patients (3%). In addition, patients with moderateenal impairment at baseline (eGFR �60 ml/min/1.73 m2)ad a 75% higher mean eGFR increase than did patientsith normal or mild renal impairment. Thus, the lesser

ncrease in eGFR in the 40-mg dose group may be due toower prevalences of these high responder subgroups.

The extent of decrease in eGFR in the placebo group isomewhat surprising because of the relatively short duration ofreatment in the present analysis of 6 to 8 weeks. This degreef decrease would extrapolate to a rate of decrease in eGFR inhese patients in the range of 9 to 10 ml/min/1.73 m2 over a-year period. A potential explanation for the extent of eGFRecrease might be a transitory decrease in eGFR that wasaused by withdrawal of a previous statin during the wash-ut phase of the rosuvastatin trials. The analysis databaseid not contain sufficient data to explore this potentialxplanation; however, if true, it would lend further supporto the possibility of a positive statin effect on eGFR. Al-hough the reasons for this finding are not entirely clear, theethod for estimating GFR was uniform for all dose groups,

ncluding placebo. Therefore, comparisons of the relativehange from baseline of the placebo group with the rosuv-statin groups should not be subject to this confoundingias. However, it should be noted that placebo-treated pa-ients had higher prevalences of hypertension, diabetes, andoronary artery disease, which could have contributed to theosuvastatin versus placebo difference in eGFR change.

Although the results from the present rosuvastatin studysing the MDRD GFR equation confirm previous smallertudies, the interpretation is restricted by the limitation ofstimating GFR using measured serum creatinine concen-ration and derived formulas. However, evidence that theGFR reflects a true increase in GFR is found in an exper-mental rat model, in which lovastatin in combination withn angiotensin-converting enzyme inhibitor preserved inu-in-measured GFR and decreased glomerulosclerosis com-ared with placebo.6 Other studies have reported that 3eeks of lovastatin treatment significantly increased renallood flow by 17% and GFR by 49% as mediated through

reglomerular vasodilation.7 In addition, a study of patients

wtaamt

1

2

3

4

5

6

7

8

1606 The American Journal of Cardiology (www.AJConline.org)

ith autosomal dominant polycystic kidney disease showedhat 4 weeks of treatment with 40 mg of simvastatin resulted inn increase in inulin-measured GFR from 124 to 132 ml/minnd in renal blood flow from 806 to 997 ml/min.8 Thus, directeasurement of GFR in animal and human studies corroborate

he findings of an increase in eGFR from large clinical trials.

. Olsson AG, McTaggart F, Raza A. Rosuvastatin: a highly effective newHMG-CoA reductase inhibitor. Cardiovasc Drug Rev 2002;20:303–328.

. Vidt DG, Cressman MD, Harris S, Pears JS, Hutchinson HG. Rosuv-astatin-induced arrest in progression of renal disease. Cardiology 2004;102:52–60.

. Pardue HL, Bacon BL, Nevius MG, Skoug JW. Kinetic study of theJaffe reaction for quantifying creatinine in serum: 1. Alkalinity con-

trolled with NaOH. Clin Chem 1987;33:278–285.

. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A moreaccurate method to estimate glomerular filtration rate from serum cre-atinine: a new prediction equation. Modification of Diet in Renal Dis-ease Study Group. Ann Intern Med 1999;130:461–470.

. Fried LF, Orchard TJ, Kasiske BL. Effect of lipid reduction on theprogression of renal disease: a meta-analysis. Kidney Int 2001;59:260–269.

. Brouhard BH, Takamori H, Satoh S, Inman S, Cressman M, Irwin K,Berkley V, Stowe N. The combination of lovastatin and enalapril in amodel of progressive renal disease. Pediatr Nephrol 1994;8:436–440.

. Stowe NT, Inman SR, Tapolyai M, Brouhard BH, Hodge EE, NovickAC. Lovastatin has direct renal hemodynamic effects in a rodent model.J Urol 1996;156:249–252.

. van Dijk MA, Kamper AM, van Veen S, Souverijn JH, Blauw GJ.Effect of simvastatin on renal function in autosomal dominant polycys-

tic kidney disease. Nephrol Dial Transplant 2001;16:2152–2157.