AT DEPARTMENT OF HEALTH AND HUMAN SERVICES
FOOD AND DRUG ADMINISTRATION
CENTER FOR DRUG EVALUATION AND RESEARCH
ENDOCRINOLOGIC AND METABOLIC DRUGS
Wednesday, July 9, 2003
8120 Wisconsin Avenue
Glenn Braunstein, M.D., Chair
Dornette Spell-LeSane, M.H.A., NP-C, Executive
Dean Follman, Ph.D.
Lynne L. Levitsky, M.D.
Nelson Watts, M.D.
Paul Woolf, M.D.
Special Government Employee Consultants
Thomas O. Carpenter, M.D.
Jeffrey B. Kopp, M.D.
Charles Hennekens, M.D.
Margaret Wierman, M.D.
ACTING INDUSTRY REPRESENTATIVE
John F. Neylan, M.D.
Robert Temple, M.D.
Robert Meyer, M.D.
David Orloff, M.D.
Mary Parks, M.D.
C O N T E N T S
Call to Order and Introductions:
Glenn Braunstein, M.D. 4
Conflict of Interest Statement:
Dornette Spell-LeSane, M.H.A., NP-C 6
Dr. Catherine McComus 10
Welcome and Introductory Comments:
David Orloff, M.D. 12
NDA 21-366 Crestor (rosuvastatin calcium) tablets
agent for iPR Pharmaceuticals, Inc.
Introduction and Regulatory Overview:
Mark S. Eliason, M.Sc. 19
Clinical Development, Efficacy Overview:
James W. Blasetto, M.D., M.P.H. 30
Clinical Development, Safety Overview:
Howard G. Hutchinson, M.D. 43
The Role of Rosuvastatin in the Treatment of
Daniel J. Rader, M.D. 85
Questions from the Committee 100
Joy Mele, M.S. 134
Safety and Dosing:
William Lubas, M.D., Ph.D. 143
Questions from the Committee 160
Open Public Hearing 194
Sidney M. Wolfe, M.D.
Sponsor Comments 206
Charge to the Committee:
David Orloff, M.D. 219
Committee Discussion/Questions 226
Glenn Braunstein, M.D.
P R O C E E D I N G S
Call to Order and Introductions
DR. BRAUNSTEIN: Welcome to the Food and
Drug Administration, Center for Drug Evaluation and
Research, Meeting of the Endocrinologic and
Metabolic Drugs Advisory Committee for July 9,
2003. Today we are going to discuss NDA 21-366,
Crestor, rosuvastatin, calcium tablets from
AstraZeneca Pharmaceuticals, agent for iPR
We will start by going around the table
and introduce ourselves and tell where we are from
and what role we play on the committee. We will
start with Dr. Temple.
DR. TEMPLE: I'm Bob Temple. I am
Director of the Office of Medical Policy at FDA and
I actually direct one of the review divisions, one
of the review offices, although it has nothing to
do with the one that is operating today.
DR. MEYER: I am Bob Meyer. I am Director
of the Office of Drug Evaluation II in CDER.
DR. ORLOFF: David Orloff, Director,
Division of Metabolic and Endocrine Drug Products,
DR. PARKS: Mary Parks, Deputy Division
Director, Metabolic and Endocrine Drug Products,
DR. CARPENTER: Tom Carpenter. I am a
pediatric endocrinologist at Yale University School
of Medicine in New Haven. This is my first meeting
with you all.
MS. SPEEL-LeSANE: Dornette Spell-LeSane,
Executive Secretary for the Committee.
DR. BRAUNSTEIN: Glenn Braunstein,
Chairman of Medicine, Cedars-Sinai Medical Center,
Chair of the Committee.
DR. WOOLF: Paul Woolf, Chairman of
Medicine, Crozer Chester Medical Center,
DR. HENNEKENS: Charlie Hennekens from
Medicine and Epidemiology at the University of
Miami. I am a consultant to the committee for this
DR. FOLLMAN: I am Dean Follman, Assistant
Institute Director for Biostatistics at the
National Institute of Allergy and Infectious
DR. WATTS: Nelson Watts, an
endocrinologist from the University of Cincinnati.
DR. WIERMAN: I am Maggie Wierman, an
endocrinologist from the University of Colorado.
DR. LEVITSKY: I am Lynne Levitsky. I am
Chief of Pediatric Endocrinology at Mass General
Hospital in Boston.
DR. NEYLAN: John Neylan. I am a
nephrologist by training and am Vice President of
Clinical Research and Development at Wyeth
Research. I serve on this committee as the Acting
DR. BRAUNSTEIN: Thank you.
We will now have the conflict-of-interest
Conflict of Interest Statement
MS. SPELL-LeSANE: The following
announcement addresses the issue of conflict of
interest with regard to this meeting and is made a
part of the record to preclude even the appearance
of such at this meeting.
Based on the submitted agenda for the
meeting and all financial interests reported by the
committee participants, it has been determined that
all interests in firms regulated by the Center for
Drug Evaluation and Research which have been
reported by the participants present no potential
for an appearance of a conflict of interest at this
meeting with the following exceptions.
Dr. Glenn Braunstein has been granted a
waiver under 21 U.S.C. 355(n)(4), an amendment of
Section 505 of the Food and Drug Administration
Modernization Act for ownership in stock in a
competitor valued between $5,001 to $25,000.
Because this stock interest falls below the de
minimis exemption allowed under 5 C.F.R
2640.202(a)(2), a waiver under 18 U.S.C. 208 is not
Dr. Thomas Carpenter has been granted a
waiver under 18 U.S.C. 208(b)(3) for his membership
on a competitor's data safety monitoring board on
unrelated matters. He receives less than $10,001
Dr. Charles Hennekens has been granted
waivers under 18 U.S.C. 208(b)(3) and under 21
U.S.C. 355(n)(4), an amendment of Section 505 of
the Food and Drug Administration Modernization Act
for ownership of stock in one of Crestor's
competitors valued between $5,001 to $25,000 for
ownership of a bond in one of Crestor's competitors
valued between $25,001 to $50,000 and for ownership
of stock in another of Crestor's competitors valued
between $5,001 to $25,000. These investments were
made independent of Dr. Hennekens by Sun Trust Bank
which has sole discretionary authority in these
In addition, the 18 U.S.C. 208(b)(3)
waiver is also for Dr. Hennekens' membership on two
data safety monitoring boards for a competitor of
Crestor. He receives less than $10,001 per year
for membership on a competitor's advisory board
where he receives less than $10,001 per year and
for membership on a competitor's data safety
monitoring board. He receives less than $10,000
Finally, the waiver includes consulting
for two of Crestor's competitors. He receives less
than $10,001 per year from each firm.
Dr. Jeffrey Kopp has been granted a waiver
under 18 U.S.C. 208(b)(3) for his consulting for a
competitor on unrelated matters. The less than
$10,001 per year is donated to charity.
Dr. Nelson Watts has been granted a waiver
under 18 U.S.C. 208(b)(3) for his consulting for
two competing firms on unrelated matters. He
receives between $10,001 to $50,000 per year from
Dr. Margaret Wierman has been granted a
waiver under 18 U.S.C. 208(b)(3) for her membership
on a competitor's speakers bureau. She receives
between $10,001 to $50,000 a year annually, also
for her membership on another competitor's speakers
bureau. Less than $5,000 is paid directly to Dr.
Wierman's employer for her research accounts.
Dr. Paul Woolf has been granted waivers
under 18 U.S.C. 208(b)(3) and under 21 U.S.C.
355(n)(4), an amendment of Section 505 of the Food
and Drug Administration Act for ownership of stock
in one of Crestor's competitors valued between
$25,001 and $50,000.
A copy of these waiver statements may be
obtained by submitting a written request to the
agency's Freedom of Information Office, Room 12A30
of the Parklawn Building.
In addition, we would like to disclose
that Dr. John Neylan is participating in this
meeting as an acting industry representative acting
on behalf of regulated industry. In the event that
the discussions involved any other products or
firms not already on the agenda for which an FDA
participant has a financial interest, the
participants are aware of the need to exclude
themselves from such involvement and their
exclusion will be noted for the record.
With respect to all other participants, we
ask, in the interest of fairness, that they address
any current or previous financial involvement with
any firm whose products they wish to comment upon.
DR. BRAUNSTEIN: Thank you.
Dr. Kopp, perhaps you will tell the
audience who you are and what you do.
DR. KOPP: My name is Jeffrey Kopp. I am
a nephrologist with the NIDDK Intramural Research
DR. BRAUNSTEIN: Thank you.
Dr. Catherine McComus has a brief
DR. McCOMUS: Good morning. My name is
Catherine McComus. I am a faculty member at the
University of Maryland. I am here today to ask for
your help on a study that I am conducting with the
FDA on what the public knows and understands about
the conflict-of-interest procedures that the FDA
uses to monitor and manage real or potential
conflicts of interest of its advisory-committee
This is a study that is being conducted
across multiple centers at the FDA. This, I
believe, is the tenth meeting where I have
collected data. I have distributed questionnaires
for members in the audience. I have also
distributed a separate questionnaire for the
advisory-committee members. If you have a chance
to complete it today, there is a box outside this
room where you can deposit it. Otherwise, there is
a business-reply envelope that you can drop it in
and mail it back at your convenience.
I do hope that you will take a few moments
to complete this survey. They are anonymous and
the more responses we get, that better we are able
to represent how people feel about the
conflict-of-interest procedures and to provide
recommendations to the FDA on how we might improve
satisfaction with the procedures.
I will be around today if you have any
questions. There is also my contact information
and a letter that is in the survey research and
please feel free to contact me if you have any
Thank you very much for allowing me to
address the group.
DR. BRAUNSTEIN: Thank you.
Dr. David Orloff will give his
Welcome and Introductory Comments
DR. ORLOFF: Good morning. First, I want
to thank the members of the committee and the
invited consultants for their review of the
materials beforehand, obviously, and for their
agreement to participate in today's meeting.
I don't know if Dr. Braunstein noted it,
but Dr. Kreisberg, Robert Kreisberg, who was
supposed to be attending today as a consultant for
the FDA, was unable to attend due to a last-minute
I also want to thank the FDA reviewers,
primarily Dr. William Lubas and Joy Mele, for their
work not only in reviewing the NDA but in preparing
for today's meeting.
I have some brief introductory remarks
that I will just read, if that is okay with
everyone. Crestor is the seventh HMG CoA-reductase
inhibitor, or statin, to come before the FDA for
review of data addressing safety and efficacy going
back to lovastatin, approved in 1987. Since the
approval of lovastatin, as most in the room
understand, much has been learned about the risks
and benefits of this class of drugs and of
individual members, some, perhaps, more than
With regard to efficacy, HMG CoA-reductase
inhibition, as a pharmacologic approach to lipid
altering, favorably impacts the course of
atherosclerotic cardiovascular disease in a broad
range of populations across ages, genders,
concomitant risk factors, those with diabetes or
without diabetes, in patients with high or low LDL
cholesterol and in those with normal or low HDL
The controlled clinical-trials experience
with this class includes nearly 30,000
statin-treated patients followed in five-year
placebo-controlled trials examining hard
cardiovascular outcomes as well as
noncardiovascular serious morbidity and mortality.
Suffice it to say that lowering LDL
cholesterol with HMG CoA-reductase inhibitors in
at-risk individuals is, I think, irrefutably proven
to reduce all the manifestations of atherosclerotic
cardiovascular disease including cardiovascular
mortality with no evidence from those trials of a
countervailing excess of noncardiovascular deaths.
This, then, is a remarkably effective class of
With regard to specific aspects of the
safety profile of the statins, it has long been
known that statin use is associated with a
dose-related increase incidence of mild to moderate
asymptomatic, often transient and resolving on
therapy, elevations in hepatic transaminases. Rare
cases of serious liver injury have been reported in
association with statin use although causality has
been difficult to establish. I would say that, by
and large, these drugs are safe with regard to the
Also long known, although not well
understood, is a potentially much more serious side
effect of statins, myopathy. This adverse effect
presents across a broad clinical spectrum from
asymptomatic creatine-kinase elevations to marked
creatine-kinase elevations with symptoms to
From clinical trials, we know that marked
creatine-kinase elevations with or without
clinically evident myopathy, which we consider
surrogates for rhabdomyolysis risk, occur with
increasing frequency at increasing doses of drug.
The risk of myopathy in rhabdo appears further
related to a number of different factors, some
better understood than others; for example,
systemic bioavailability of drug, pharmacokinetic
interactions leading to augmented drug exposure,
the "affinity," in quotes, if you will, of drug for
muscle, the potency of the drug as an inhibitor of
HMG CoA-reductase and predisposing factors such as
diabetes, renal failure, hypothyroidism, surgery,
severe acute illness or injury.
Rhabdomyolysis, or fulminant myopathy with
frank necrosis, myoglobinemia and myoglobinuria and
acute pigment-induced renal failure occurs very
rarely in the clinic in, at least retrospectively,
uniquely susceptible individuals in whom it
appears, after the fact, that some threshold muscle
exposure to drug has been exceeded. As above, as I
stated earlier, this is the most serious side
effect of statins, potentially fatal, and the
Finally, in the Crestor Development
Program, a heretofore undescribed renal side effect
of an HMG CoA-reductase inhibitor has been
The original New Drug Application for
Crestor was submitted on June 26, 2001. An
approvable action was taken by the agency on May
31, 2002, based on safety concerns arising out of
the initial review regarding muscle and kidney.
More specifically, several cases of severe myopathy
or rhabdomyolysis occurred in patients treated with
80 milligrams daily, the highest dose initially
There were no cases seen at 40 milligrams,
although patient exposures at 40 milligrams were
far fewer. Based on this primary safety concern
and the marginal incremental LDL lowering seen with
the step from 40 to 80 milligrams, the agency
concluded that 80 milligrams should not be
Because the clinical-trial exposures had
been skewed toward the low and high ends of the
proposed dosage range, further data were deemed
necessary before a decision could be reached on the
20 and 40 milligram doses. The FDA requested that
the sponsor conduct additional trials to augment
the patient exposure at 40 milligrams specifically
as 40 milligram starts, patients de novo treated
with Crestor at a dose of 40 milligrams, in order
to answer this important question, is Crestor more
prone to cause myopathy than currently marketed
statins, or, alternatively, was 80 milligrams
simply too high a dose to be, overall, safe for
This question was particularly important
in light of the experience with Baycol,
cerivastatin, which, as was observed post-approval,
conferred substantial risk of myopathy relative to
other members of the class, a doses effecting
little LDL-cholesterol lowering.
In response to the FDA request, the
sponsor has studied the myopathic risk associated
with Crestor use in a very large premarketing
patient exposure, indeed, by far the largest of any
statin brought before the FDA. The sponsor and the
FDA medical officer, Dr. Lubas, will present data
today that suggests that the risk of myopathy with
Crestor relative to LDL-lowering efficacy is, at
the very least, no greater than that with the other
marketed members of the class. I emphasize the
critical importance of this issue in the evaluation
of the safety of this drug.
In addition, the sponsor was asked to
investigate further the finding of new-onset mild
proteinuria observed mostly in patients taking
Crestor 80 milligrams. Specifically, the sponsor
was charged with investigating the "nature,
magnitude and frequency" of renal adverse events
observed in patients treated with rosuvastatin and
to explore whether these effects were "reversible,
chronic or progressive."
As you will hear presented, the renal
effects occur with very low frequency at doses
below 80 milligrams although in up to 10 percent of
patients taking 80 milligrams. This is not a
finding noted in other statin-development programs
or in long-term trials of statins.
The clinical picture of Crestor-associated
renal effects seems to include variably the
combination of low-grade proteinuria, minor
elevations in creatinine and microscopic hematuria.
This will be discussed by Dr. Lubas and by the
The sponsor, furthermore, will present
information supporting the possibility that these
renal effects represent a mechanism of
action-related class effect of statins on the
proximal statins on the proximal renal tubule.
This requires close attention and discussion in the
evaluation of the safety of this drug.
In addition, the FDA clinical and
statistical reviewers will make further comments on
specific efficacy and safety issues.
I will end my comments there and have a
few more remarks at the time that I charge the
committee later during the proceedings. Thank you
DR. BRAUNSTEIN: Thank you Dr. Orloff.
We will now move on to the sponsor's
NDA 21-366 Crestor (rosuvastatin calcium) tablets
Agent for iPR Pharmaceuticals Incidence.
Introductory and Regulatory Overview
MR. ELIASON: Good morning everyone. My
name is Mark Eliason and I am the US Regulatory
Director for CRESTOR at AstraZeneca.
Mr. Chairman, distinguished members of
this committee, AstraZeneca is pleased to present
information regarding the safety and efficacy of
CRESTOR Tablets, as currently contained in our NDA.
We hope that you will find our presentations this
morning to be helpful in your deliberations later
in the day.
On behalf of AstraZeneca, I wish
acknowledge at this time the multitude of
physicians, and other healthcare professionals who
participated in the very large CRESTOR drug
To begin my presentation, Iūd like to
discuss the development objectives established by
AstraZeneca for a new statin candidate.
From the early information derived from
the molecule, we focused on the development of
rosuvastatin to provide an overall benefit risk
greater beneficial effects on key lipid parameters,
at both the start dose and across the dose range,
when compared to approved drugs in this class; a
similar safety profile in relation to muscle,
liver, and other effects, when compared to approved
drugs in the statin class; and, lastly, a low
potential for significant drug-drug interactions,
especially through the Cytochrome P450 and
P-glycoprotein systems, as plasma levels of other
drugs in this class had been shown to be driven
higher due to drug-drug interactions.
Rosuvastatin is a novel synthetic
inhibitor of HMG-CoA reductase that was discovered
by the Shionogi Company of Japan. In terms of its
structure, at first glance rosuvastatin is a
conventional statin as it resembles other statins
in having the common pharmacophore group, the
group that resembles the HMG substrate.
However, rosuvastatin is distinctive in
its structure as it contains a relatively polar
methane sulfonamide group. This helps to place
rosuvastatin low on the scale of lipophilicity,
near pravastatin, when plotted against the other
statins as shown on the scale on the right of this
This has two consequences for
pharmacology: first, compounds with low
lipophilicity have the potential of being highly
selective for entry into liver cells as compared to
non-hepatic cells. Secondly, compounds low on this
scale are relatively water soluble and therefore
would not require extensive metabolism by the
hepatic CYP P450 system to render them sufficiently
water soluble for excretion.
In essence, preclinically, rosuvastatin
has some of the favorable properties of
pravastatin, namely a high degree of cell
selectivity and a low degree of metabolism by the
cytochrome P450 system.
On this slide, I would now like to briefly
summarize the key pharmacokinetics and disposition
characteristics of rosuvastatin. The absolute
bioavailability of rosuvastatin is approximately 20
percent. The molecule is only moderately bound to
plasma proteins, principally albumin.
Rosuvastatin does not undergo extensive
metabolism in man. Finally, the terminal half-life
of rosuvastatin is approximately 16 to 20 hours.
Moving to our clinical program, our NDA is
supported by a large international clinical
development program. The results of the studies
outlined on this slide will be discussed later in
our presentations. The program included
thirty-three Phase I studies, and twenty-seven
Phase II/III trials. During Phase III, we
evaluated doses from 5 to 80 milligrams.
The safety database from this set of Phase
II/III trials now contains over 12,500 patients
taking rosuvastatin having a total of over 14,000
patient years. As Dr. Orloff had stated earlier
today, this is by far the largest initial approval
NDA database submitted for a statin to date.
The design of the Phase III program trials
included comparative trials to both placebo and key
statin therapies, which included atorvastatin,
simvastatin and pravastatin, as well as to
non-statin therapies, such as niacin and
fenofibrate in hypertriglyceridemic patients. In
addition, we studied rosuvastatin in combination
with niacin and with fenofibrate, as well as
At the completion of the controlled
portion of our Phase III trials, the enrolled
patients were allowed to continue into long-term
rosuvastatin open-labeled extension trials. These
open label extensions are all still active and
continue to add valuable long-term rosuvastatin
safety information to the clinical database.
There were a number of important trial
features in the clinical development program for
rosuvastatin, some of which are presented here on
this slide. For our Phase III program, all
clinical laboratory samples were analyzed at one
central laboratory. This reduced the potential for
As you will see later in our
presentations, we also tried to be as inclusive as
possible in the range of patients enrolled in our
Phase II/III trials. The purpose of this was to
recruit a diverse population of patients, in
various states of health, that would be considered
representative of the general population requiring
To be specific, we had no upper age limit
for our trials so that approximately a third of the
patients participating in our trials were over 65
years of age.
For most of our trials, we allowed patients with
creatinines of up to two and a half milligrams per
deciliter. From this, over 50 percent of the
patients enrolled in our trials had some degree of
Women of childbearing potential were
permitted to enter into most trials, provided that
they were not pregnant and used appropriate
contraception. Finally, we allowed patients into
trials with existing co-morbidities, such as
hypertension, diabetes, and cardiovascular disease,
provided that the patient's condition was stable
prior to randomization.
Now I would like to turn to the Crestor
NDA itself. As Dr. Orloff had previous stated, our
original new drug application for CRESTOR Tablets
was submitted to the FDA in June of 2001. The
initial NDA submission proposed a dose range of 10
to 80 milligrams once daily for rosuvastatin.
As further clinical data became available,
it was evident that the 80-milligram dose provided
additional lipid effects that would be of potential
benefit to those patients with difficult-to-control
However, the emergent profile for the
80-milligram dose did not meet our objectives for
the favorable benefit-risk profile for the general
populations. So, in March of 2002, AstraZeneca and
the Review Division agreed to suspend further
development of the rosuvastatin 80-milligram dose
for the general population, and all patients who
were receiving the 80-milligram daily dose had
their dose reduced to 40-milligram daily.
The NDA action letter was issued in May
2002, noting that the proposed 10, 20 and 40
-milligram doses of rosuvastatin were approvable.
The NDA action letter centered on the request for
additional safety data for patients receiving the
20 and 40-milligram, in order to fully assess the
therapeutic index of rosuvastatin. In addition,
the Division requested additional information
regarding the renal effects observed in the
AstraZeneca and Division representatives
met in July 2002 to outline the data package for
responding to the action letter. At this meeting,
the Review Division requested that a minimum of 600
patients treated with rosuvastatin at the 20
milligram and at the 40-milligram for six months be
included in the response.
From that, an NDA amendment was submitted
in February of this year supporting a proposed 10
to 40-milligram dose range for the general
population. The NDA amendment provided the
requested additional safety information for the 20
and 40-milligram doses, and with the submission of
an interim safety update in June of this year, the
final NDA safety database contains over 12,500
patients treated with rosuvastatin.
The Rosuvastatin Clinical Development
Program supports the proposed CRESTOR Tablet NDA
indications which are fully presented in Section
1.1 of our briefing document. I will, just for
time's sake, go through them here very quickly.
Our first indication involves primary
hypercholesterolemia and mixed dyslipidemia.
A second indication involves patients with
hypertriglyceridemia. Finally, a third indication
involves the genetic familial homozygous
hypercholesterolemic patient population.
The dosing recommendations proposed in the
CRESTOR NDA are outlined on this slide. For
primary hypercholesterolemia, mixed dyslipidemia
and hypertriglyceridemia, the recommended start
dose of CRESTOR is 10 milligrams, once daily, with
a maximum recommended daily dose of 40 milligrams.
A 20-milligram start dose is optional for
patients with LDL-C levels of greater than 190
milligrams per deciliter and aggressive lipid
targets. For the homozygous familial
hypercholesterolemia indication, the recommended
starting dose for CRESTOR is 20 milligrams once
Finally, a 5-milligram dose will be made available
for patients taking cyclosporine.
The rationale regarding these dosing
recommendations will be discussed in our
Regarding the status of the CRESTOR, we
have approval in 24 countries in Europe, Asia and
the Americas, all incorporating the 10-milligram to
40-milligram dose range. In addition to the
described NDA activity, we continue to study
rosuvastatin. Our ongoing trials program,
investigating rosuvastatin in cardiovascular risk
reduction, currently includes approximately 24,000
patients in the U.S. and the rest of the world all
who are taking rosuvastatin.
Also, as part of this program, we have
initiated two clinical-outcomes trials in May of
this year, which will enroll a total of 18,000
patients between them.
With this background in mind, here is the
agenda for remainder of our presentation. Next,
Dr. James Blasetto will present a brief overview of
the key efficacy results from our NDA clinical
After Dr. Blasetto, Dr. Howard Hutchinson
will discuss the safety profile of rosuvastatin
from our NDA clinical program, with a focus on key
safety issues from the statin drug class.
Finally, AstraZeneca has invited Dr.
Daniel Rader, from the University of Pennsylvania,
to present his thoughts as a practicing physician
on the potential role of rosuvastatin in treating
AstraZeneca has also asked the following
individuals to assist in responding to any points
that the advisory committee members may wish to
have addressed during this meeting. In addition to
Dr. Rader, we have Dr. Christie Ballantyne from
Baylor College, Dr. Donald Hunninghake from the
University of Minnesota, Dr. Edmund J. Lewis from
Rush Presbyterian St. Luke's Medical Center, Dr.
Thomas Pearson from the University of Rochester
Medical Center and Dr. Evan Stein from Medical
Research Laboratories International.
Now I would like to introduce Dr. James
Blasetto, Senior Director at AstraZeneca, who will
present the efficacy portion of our presentation.
DR. BLASETTO: Good morning.
I am Dr. James Blasetto, Senior Director,
Clinical Research at AstraZeneca.
Hypercholesterolemia represents a
significant, persistent yet potentially treatable
medical program in the United States. If we look
at the evolution of the Cholesterol Management
Guidelines as proposed by the National Cholesterol
Education Program, we see an ever-increasing need
for more lipid-modifying efficacy.
If we focus in on the most recent
guidelines, the ATP-3 Guidelines launched in 2001,
we see a number of new and important features.
Firstly, identifies the optimal LDL-C level at less
than 100 milligrams per deciliter.
Secondly, the target goal for patients in
the high-risk group has been made more aggressive,
less than 100 milligrams per deciliter, and a
number of patients that qualify for the high-risk
group has been expanded with the introduction of
the CHD risk-equivalent patients.
Thirdly, there is an increased focus on
HDL-C with a secondary target for therapy, the
non-HDL-C goal, for patients with persistent
elevated triglycerides. Thus, with the current
guidelines, it is estimated that over 36 million
patients will require lipid-lowering therapy and
approximately 60 percent of those, or approximately
21 million, will require a treatment LDL-C goal of
less than 100 milligrams per deciliter.
Yet, if we look at recent clinical data,
we see that a treatment gap still exists between
what current therapies can obtain and what is
needed. This is data that was presented by Dr.
Christie Ballantyne in 2001 from the ACCESS Trial,
the Atorvastatin Comparative Cholesterol Efficacy
and Safety Study.
This is a cohort of patients in the CHD
risk category. Patients were treated and titrated
up to achievement of the ATP-2 goal, an LDL-C of
less than or equal to 100 milligrams per deciliter.
If we focus in on the patients that were
treated with up to maximum doses of atorvastatin,
80 milligrams, we see that 28 percent of the
patients did not achieve their LDL-C target goal
and approximately 40 percent of the patients did
not achieve an established non-HDL-C goal.
If we look at the percent of patients that
did not achieve their LDL or non-HDL-C goals with
the other statins at the doses studied, we see the
numbers were even greater. Thus, with the current
guidelines, more patients require more aggressive
treatments yet, with current therapies, a treatment
deficit still exists.
Now, as you heard in the opening remarks,
there were three key objectives that were core to
our Clinical Development Program. My presentation
will focus on efficacy data to support the first
key objective which was to demonstrate greater
beneficial effects on key lipid parameters over
currently marketed statins. In addition, I will
discuss data that addresses efficacy questions
raised to this advisory committee.
Our first LDL efficacy data came from two
Phase II dose-ranging studies. These studies were
prospectively designed to be pooled. The patient
population evaluated were patients with Type IIa
and IIb hypercholesterolemia.
This is the response seen in percent
change from baseline in LDL-C at each of the doses
evaluated. The mean age in the population studied
was 56 years and the mean baseline LDL-C, 190
milligrams per deciliter. Statistically
significant differences compared to placebo at each
of the doses evaluated were seen, a 33 percent
reduction up to a 65 percent reduction in LDL-C.
Now, based on the efficacy that we saw in
these dose-ranging studies, we initially chose to
evaluate two potential starting doses, rosuvastatin
5 milligrams and rosuvastatin 10 milligrams.
Our Phase III data has confirmed the added
benefits on key lipid parameters with the
10-milligram dose compared to the 5-milligram dose
with an indistinguishable safety profile.
This is data from five clinical trials in
our Phase III program which was prospectively
designed to be pooled. The patient populations
studied were patients with Type IIa and IIb
hypercholesterolemia. The mean age in the
population was 58 with a mean baseline LDL-C of 187
milligrams per deciliter.
After twelve weeks of treatment, this is
the response seen in key lipid parameters with
rosuvastatin 10-milligrams and rosuvastatin 5
milligrams. The 10-milligram dose added benefit on
all lipid parameters compared to the 5-milligram
dose, in particular, a 6 percent further LDL-C
reduction and an approximate 5 percent further
Thus, the risk-benefit profile of the
10-milligram dose is better than the 5-milligram
dose and offers a better treatment option as a
starting dose for patients. Thus, our proposed
starting dose for the general population is
rosuvastatin 10 milligrams.
Alternatively, we initially evaluated
doses up to and including the 80-milligram dose.
As you heard in the opening remarks, after an
assessment of the benefit-risk profile of the
80-milligram dose, we elected to back-titrate
patients from 80 milligrams to 40 milligrams and
not to pursue at this time further development of
the 80-milligram dose. Thus, the maximum proposed
dose is rosuvastatin 40 milligrams.
Rosuvastatin 40 milligrams offers benefit
in key lipid parameters compared to the
20-milligram dose for patients requiring more
reductions to achieve their NCEP targets.
This is data from five individual clinical
trials in our development program which looks at
the effects on LDL-C with rosuvastatin 40
milligrams and rosuvastatin 20 milligrams. In each
of these trials, the patient populations studies
were patients with Type IIa and IIb
hypercholesterolemia with a cohort of patients with
heterozygous familiar hypercholesterolemia
evaluated in Trial 30.
In each of these clinical trials, the
40-milligram dose added greater reductions in LDL-C
compared to the 20-milligram dose. In four or five
of the clinical trials, there was a 7 percent or
greater LDL-C reduction seen with the 40-milligram
dose compared to the 20-milligram dose. Thus, for
patients requiring more reductions in LDL-C or
non-HDL-C to achieve their NCEP target goals, the
40-milligram dose offers benefits over the
Thus our proposed dose range is
rosuvastatin 10 to 40 milligrams and, for the
remainder of my presentation, I will focus on the
10 to 40-milligram dose range.
We studied the effects comparatively of
rosuvastatin in several clinical trials.
The largest clinical trial comparatively
done was the STELLAR Trial, Trial 65, as presented
here. This trial included over 2,000 patients.
After a six-week dietary lead-in, patients were
randomized in an open-label fashion to one of the
treatment arms with rosuvastatin, atorvastatin,
simvastatin or pravastatin, as shown, for six weeks
Baseline characteristics in all treatment
arms were well-matched. The mean age in the
population was 57 and the mean baseline LDL-C 189
milligrams per deciliter.
After six weeks of treatment, this is the
response seen in percent change from baseline in
LDL-C. Rosuvastatin, 10 to 40 milligrams on a
milligram-to-milligram basis demonstrated greater
reductions than atorvastatin, simvastatin and
Doubling of the dose of statin therapy
yielded an approximate 4.5 to 5 percent further
LDL-C reduction. If we assess the effects of
patients treated with rosuvastatin 40 milligrams to
those treated with atorvastatin 80 milligrams, we
saw an approximate 4 percent further LDL-C
reduction with rosuvastatin therapy.
If we look at the distribution of LDL-C at
each of the treatment arms, we see that the
distribution of LDL-C was similar in each treatment
arm, the number of outliers was similar and the
median reduction in LDL-C seen with rosuvastatin 40
milligrams was greater than that seen with the
other statin comparators.
The STELLAR Trial was designed to perform
multiple pairwise and dose-to-dose comparisons on
other key lipid parameters. This is the response
in HDL-C after six weeks of treatment in each of
the treatment arms evaluated. Rosuvastatin 20 and
40 milligrams raised the HDL-C approximately 10
Comparatively, the 10-milligram-response
rosuvastatin was statistically greater than the
10-milligram response of pravastatin. The
20-milligram-response rosuvastatin was
statistically greater than the 20 to 80-milligram
response of atorvastatin, the 20 and 40-milligram
response of pravastatin and the 40-milligram
response of simvastatin. The 40-milligram response
of rosuvastatin was greater than the 40 and
80-milligram response of atorvastatin and the
40-milligram response of both simvastatin and
We assessed the results on the important
parameter non-HDL-C goal. Rosuvastatin, at the
40-milligram dose, reduced non-HDL-C by greater
than 50 percent. Comparatively, compared to
similar doses of atorvastatin and similar doses, or
higher doses, of simvastatin and pravastatin,
rosuvastatin reduced non-HDL-C by a greater
Now, to assess the effects on achievement
on NCEP targets at higher doses, we evaluated
rosuvastatin comparative to atorvastatin in a
titration-to-goal study, Study 26.
This is the design of that trial. After a
six-week dietary lead-in, patients were randomized
in a double-blind fashion to one of the treatment
arms with rosuvastatin or a common starting dose,
atorvastatin 10 milligrams, for twelve weeks of
After twelve weeks, patients were then
subsequently titrated to the next highest dose if
they did not achieve their ATP-2 LDL-C targets.
Baseline characteristics in each of the treatment
arms were well matched. In this population of
patients with Type IIa and IIb
hypercholesterolemia, the mean age was 57 and the
mean baseline LDL-C 187 milligrams per deciliter.
After 52 weeks of treatment, this is the
response seen in the percent of patients achieving
target goal. 82 percent of the patients on
rosuvastatin 10 milligrams achieved their target
goal without need for titration compared to 59
percent of the patients on atorvastatin 10
Overall, 96 percent of the patients
achieved target goal with a regimen of rosuvastatin
10 to 40 milligrams compared to 87 percent of the
patients with a regimen of atorvastatin 10 to 80
milligrams. Thus, overall, more patients achieved
their target goal but, in particular, a greater
percentage achieved target goal at the starting
dose without need for titration.
I would like to conclude with an
assessment of rosuvastatin in an important
population of patients, patients with severe
hypercholesterolemia, heterozygous familial
hypercholesterolemia. This represents an important
population of patients because of the severe nature
of their hypercholesterolemia. They are difficult
to treat and have a frequency in the United States
population of approximately 1 in 500.
We assessed the effects of rosuvastatin
comparatively in this population in Trial 30. This
is the design of that trial. It was a large,
multicentered, multinational trial. After a
six-week dietary lead-in, patients were randomized
in a double-blind fashion to rosuvastatin or
atorvastatin 10 milligrams.
In view of the severe hypercholesterolemia
at baseline these patients had, and the increased
efficacy they needed at the start of therapy, we
chose a strategy of starting these patients to
evaluate a 20-milligram starting dose. After six
weeks of treatment, the patients were
force-titrated to the 40-milligram dose and then
ultimately to the 80-milligram dose.
Baseline characteristics were well matched
in both treatment arms. The mean age of the
population was 48, somewhat younger than the data I
previously presented. That is not unexpected with
patients with heterozygous familial
hypercholesterolemia. The baseline LDL-C
demonstrates the severe hypercholesterolemia of
these patients approaching nearly 300 milligrams
The results of the 80-milligram dose will
be presented to show the potential added benefits
of increased efficacy. However, in view of our
proposed dosing recommendations, I will focus my
comments on the 20 and 40-milligram dose response
This is the response in the percent change
from baseline in LDL-C at each of the time points
and doses evaluated. Rosuvastatin 20 milligrams
reduced LDL-C 47 percent and 54 percent reduction
at the 40-milligram dose, statistically greater
than the 20 and 40-milligram dose response seen
If we evaluate the effects on HDL-C, a 12
percent and 10 percent increase in HDL-C seen with
20 and 40 milligrams of rosuvastatin, statistically
greater than the 20 and 40-milligram response seen
The greater LDL-C reduction translated
into more patients achieving their ATP-3 target
goals. 37 percent of the patients with
rosuvastatin 20 milligrams achieved the target goal
and nearly 50 percent with rosuvastatin
40 milligrams, both statistically greater than the
20 and 40-milligram response of atorvastatin.
If we focus in on that high-risk group of
patients requiring a target LDL-C of less than 100
milligrams per deciliter, 17 percent of the
patients achieved that target goal with
rosuvastatin 40 milligrams compared to 3 percent of
the patients with atorvastatin 40 milligrams. This
was statistically different.
So, in summary, data from our Clinical
Development Program has demonstrated rosuvastatin
10 to 40 milligrams reduced LDL-C 50 to 62 percent
as presented in the dose-ranging studies.
Rosuvastatin lowered LDL-C and non-HDL-C more than
atorvastatin, simvastatin and pravastatin across
the dose range. Greater increases in HDL-C were
More patients achieved NCEP goals with a regimen of
rosuvastatin 10 to 40 milligrams than that with
atorvastatin 10 to 80 milligrams, simvastatin 20 to
80 milligrams and pravastatin 20 to 40 milligrams.
I thank you and, at this time, I would
like to introduce Dr. Howard Hutchinson who will
discuss the safety profile of rosuvastatin.
DR. HUTCHINSON: Good Morning.
I am Howard Hutchinson, Vice President for
Clinical Research at AstraZeneca. Today, I am
pleased to be here to present the safety profile
Dr. Blasetto presented the efficacy data
showing the overall benefits of a rosuvastatin
10-milligram to 40-milligram dose range for the
treatment of patients with dyslipidemia. However,
the benefits of a new drug must also be placed in
the context of the potential risks associated with
With this in mind, I will now present data
which addresses the last two objectives of our
development program. This information will show
that the proposed 10-milligram to 40-milligram dose
range for rosuvastatin has a safety profile similar
to other marketed statins, and that rosuvastatin
will have a low potential for significant drug-drug
The safety data I am going to present
today comes from twenty-seven clinical trials
About half the patients were from the United
The overall database is comprised of over 12,500
patients who have had over 14,000 patient years of
treatment with rosuvastatin at doses up to and
including 80 milligrams.
In presenting the safety data, I will
focus on several key areas. First, I will present
the overall demography of our patient population
followed by exposure data, and adverse events. I
will then focus on three areas of interest for
rosuvastatin and statins in general. They are the
liver, skeletal muscle, and renal effects.
I will finish with a brief presentation on
This slide represents the overall
demography for patients in our
all-controlled/uncontrolled plus Real Time
Laboratory Data or RTLD Pool. This pool represents
our largest pool with 12,569 patients and includes
patients exposed to rosuvastatin in both controlled
trials and in open-label extension trials.
As shown, the mean age for subjects in our
program was 58. Approximately one-third of the
patients were 65 years or older, and over 900 were
75 or over. Almost half of the population was
female, and two-thirds of the women were
With regard to ethnicity, most patients
were Caucasian; however, over 1000 patients were of
We set up our development program to be
inclusive. Patients with co-morbid conditions were
permitted to enter studies provided they were
stable at baseline and we allowed patients to enter
most trials with a serum creatinine level up to 2.5
milligrams per deciliter.
As shown, over half of the subjects
enrolled in the program had baseline renal
impairment as determined using the Cockroft-Gault
formula. In addition, over half of the subjects
had baseline hypertension, 36 percent had
documented atherosclerotic cardiovascular disease,
and 16.5 percent had diabetes.
This slide shows the maximum continuous
duration of treatment with the 5-milligram to
80-milligram doses of rosuvastatin from the
clinical trial program. As shown, over 1000
patients were treated with each of these doses.
Importantly, over 7800 patients were treated with
10-milligram proposed starting dose, over 3900
patients were treated with the 20-milligram dose,
and over 4000 were treated with the 40-milligram
dose. Of the 4000 subjects treated with the
40-milligram dose, over 2000 initiated therapy at
Highlighted are the 24-week and 48-week
exposures. Note that over 1300 and 1800 patients
were treated with the 20-milligram and 40-milligram
doses for 24 weeks or longer. 545 and 276 were
treated with these doses for greater than or equal
to 48 weeks. As previously discussed, patients on
80 milligrams were back-titrated to 40 milligrams
during the development program. The 40-milligram
exposures seen in this table represent patients
back-titrated from 80 milligrams and patients never
exposed to 80 milligrams. Importantly, however, all
of the exposures greater than 48 weeks are in
patients who were never exposed to the 80-milligram
dose and over 3700 patients in this pool were never
exposed to the 80-milligram dose.
The last column is the greater than or
equal to 40-milligram treatment group. In this
group, patients treated with 80-milligram dose and
back-titrated to 40 milligrams were considered to
have been treated continuously with rosuvastatin
with at least 40 milligrams of drug.
This group is important because it gives
information regarding the potential for adverse
events to occur very late into therapy. Note that
1165 patients were treated for greater than or
equal to 48 weeks in this group and 874 for greater
than or equal to 96 weeks in this group.
As you will see, the exposures generated
for this analysis are appropriate for evaluating
the overall safety of rosuvastatin at doses up to
and including 80 milligrams.
Today, a detailed review of
patient-reported adverse events will not be
presented so that I can focus on the more critical
issues addressed in the FDA briefing document.
Shown here are the key points summarizing the
adverse event data.
First of all, the data showed that the
frequency and types of adverse events reported for
rosuvastatin were similar to that of the comparator
statins in our program. Second, the frequency and
types of adverse events were similar for the
5-milligram, 10-milligram, 20-milligram and
40-milligram doses of rosuvastatin.
However, at the 80-milligram dose,
increased frequencies of nausea, myalgia, asthenia,
and constipation were observed, in particular,
nausea, myalgia, asthenia and constipation.
Importantly, rosuvastatin was well-tolerated in a
broad spectrum of patients regardless of age, sex,
ethnicity, the presence comorbidities such as
diabetes, hypertension, or renal impairment, and in
patients on medications used to treat comorbid
conditions such as anti-hypertensive agents and
I would now like to turn our attention to
the effects of rosuvastatin on three organs, the
liver, skeletal muscle, and kidneys. I will start
with the liver.
As Dr. Orloff had mentioned earlier, in
general, statins are well tolerated from the
perspective of the liver. Asymptomatic
transaminase elevations are reported for all
statins, and the frequency of the elevations
appears to increase with dose. Importantly, these
elevations have almost never been associated with
liver failure. The effects of rosuvastatin on the
liver are similar to that observed with other
members of the class.
In the rosuvastatin program, liver
function tests were performed at each visit. In
this section, I will present the percentage of
patients with ALT elevations greater than three
times the upper limit of normal on two occasions.
Note that the ALT elevations greater than three
times the upper limit of normal on two occasions is
consistent with the definition of persistent
elevations used in the labels for other marketed
I will not present data on AST elevations.
However, AST elevations in our program mirrored the
We also evaluated patients for ALT
elevations associated with increases in bilirubin.
Importantly, these elevations were rarely observed,
and, in those instances where they were observed,
they were almost always associated with another
illness such as a malignancy or infectious
Shown on this slide is the frequency of
persistent ALT elevations in patients treated with
rosuvastatin from 5 to 80 milligrams in the all
Controlled/uncontrolled plus RTLD Pool. The data
shows that the frequency of ūpersistent ALT
elevationsū ranged from 0.1 percent to 0.5 percent
at rosuvastatin doses from 5 to 40 but increased to
1.4 percent at the 80-milligram dose.
This figure helps to put the overall ALT
results from the rosuvastatin program into context
with that reported in the prescribing information
or summary basis of approval documents for other
marketed statins, specifically fluvastatin, 20, 40,
and 80 milligrams, lovastatin, 20, 40, and 80
milligrams, simvastatin, 40 and 80 milligrams,
atorvastatin, 10, 20, 40, and 80 milligrams and the
data for rosuvastatin.
On the x-axis is plotted the percentage
LDL-C lowering for the various doses of drug which
represents the potential benefits that can be
achieved at a particular dose. On the y-axis, the
frequency of persistent ALT elevations at a given
dose represents a potential risk of the dose. Note
that rosuvastatin at doses from 5 to 40 milligrams
has a low frequency of elevations similar that
observed with other statins. Only at the
80-milligram dose is an increase in frequency of
persistent elevations seen. The increase in
frequency with rosuvastatin at the 80-milligram
dose, however, is in the range observed for
marketed statins. However, the increase observed
with the other marketed statins occurs at lower
levels of LDL-C reduction.
Overall, the data pertaining to possible
liver effects of rosuvastatin obtained from our
development program support its safety with regard
to this organ.
I would now like to turn our attention to
skeletal-muscle findings. Similar to persistent
ALT elevations, adverse skeletal-muscle effects are
a recognized complication of statin therapy.
Adverse effects such as myopathy and rhabdomyolysis
have been reported for all statins. However, the
frequency of such reports is very low within the
recommended dose range.
Similar to the routine evaluation of
liver-function tests in our program, creatine
kinase or CK measurements were performed at each
In this part of my talk, I will present
the following information.
First, I will present data on CK
elevations greater than ten times the upper limit
of normal. This is an objective measure of the
potential of a statin to cause muscle effects.
Next, I will present our cases of
myopathy. In our program, we used a well
established definition of myopathy which is CK
elevations greater than ten times the upper limit
of normal with associated muscle symptoms.
Some of the patients in our program had
rhabdomyolysis at the 80-milligram dose.
Currently, rhabdomyolysis is defined
several different ways in the literature. In the
FDA review, rhabdomyolysis cases are defined as
those patients with myopathy who required
hospitalization to receive intravenous fluids.
Shown on this slide is the frequency of
both symptomatic and asymptomatic CK elevations in
patients treated with rosuvastatin at doses from 5
to 80 milligrams, once again in our largest pool,
the all controlled/uncontrolled plus RTLD Pool. Our
data shows that the frequency of elevations ranged
from 0.2 to 0.4 percent at rosuvastatin doses from
5 to 40 but increased to 1.9 percent at the
If we now look at these cases for patients
with muscle-related symptoms, we have our overall
Shown on this slide are all symptomatic CK
elevations and those with a possible relationship
to treatment. Note that the overall number of
symptomatic CK elevations at doses from 5 to 40
milligrams is low and similar. The overall
frequency increases to 1.0 percent at the
However, many of these patients had
symptomatic elevations related to causes such as
heavy exercise or injury and many resolved on
continued therapy at the same dose of rosuvastatin.
If we exclude those cases with clearly identified
other causes, we have left the cases with a more
likely association to rosuvastatin therapy.
A total of thirteen possibly
treatment-related cases have been identified, one
case each at 20-milligram and 40-milligram doses
and eleven cases at 80-milligram dose. The one case
observed at 20-milligram dose was in a patient who
was also found to have a Coxsackie Type IV viral
infection at the time of the event. Coxsackie Type
IV viral infections have been associated with
The patient at 40 milligrams had a history
of asymptomatic CK elevations as high as 10,000 off
statin therapy who had a CK elevation to 15,000
three days after initiating a weight-lifting
program. Because the patient had associated arm
pain, he was hospitalized to rule out a myocardial
infarction. After ruling out for myocardial
infarction and being discharged, the patient was
restarted on rosuvastatin 40 milligrams and has now
remained on this dose for several months and has
been asymptomatic without CK elevations.
The eleven cases of possibly
treatment-related myopathy at the 80-milligram
gives a frequency of 0.7 percent at this dose.
Importantly, all eleven of these patients recovered
following discontinuation of therapy. Seven
patients were hospitalized to receive intravenous
fluids. During the program, we also had two cases
of myopathy observed in patients on simvastatin 80
milligrams which gave us a frequency of myopathy
for that group of 0.4 percent. One of these
patients was hospitalized to receive intravenous
The eleven 80-milligram myopathy cases do
allow us an opportunity to evaluate the possible
risk factors for myopathy with rosuvastatin. The
three major risk factors that we identified at the
80-milligram dose were age, renal insufficiency,
and hypothyroidism. It is important to note that
these are also identified as risk factors for
myopathy with other marketed statins.
With regard to age, the frequency of
myopathy was 0.2 percent in subjects less than 65
years old and 2.3 percent in subjects 65 years of
age or older. Patients with a creatinine clearance
less than 80 milliliters per minute had a myopathy
frequency of 1.2 percent at the 80-milligram dose
compared to a frequency of 0.2 percent in patients
with a normal renal function or a creatinine
clearance greater than 80 milliliters per minute.
However, whether renal insufficiency is
truly an independent risk factor for myopathy is
difficult to determine from our data since we used
the Cockroft Gault formula and age is a significant
component in the creatinine-clearance calculation.
Although hypothyroidism was an exclusion
criterion in our program, two patients with
myopathy did have an elevated TSH at the time of
With regard to gender, we did not find a
sex-based predisposition to myopathy. However, of
the seven patients hospitalized to receive
intravenous fluids, five were females.
The data from our program show that
rosuvastatin was well tolerated from a
skeletal-muscle perspective. An increased
frequency of adverse skeletal-muscle effects
compared to lower doses of rosuvastatin was
observed at the 80-milligram dose. However, the
vast majority of patients were safely treated even
with the 80-milligram dose.
How do the skeletal-muscle data generated
from this program compare to data for other
statins? To look at this, we looked back at CK
elevations greater than ten times the upper limit
of normal because this provide an objective measure
for evaluating the potential for a dose of a statin
to cause muscle toxicity.
In this slide, we compare the effects of
rosuvastatin on this parameter to results reported
for cerivastatin at 0.2 to 0.8 milligrams,
pravastatin 40 and 80 milligrams, simvastatin, 40
and 80-milligrams, atorvastatin, 10 to 80
milligrams and rosuvastatin.
In this figure, we evaluate the overall
benefits of a dose of a statin with regard to LDL-C
lowering versus the risk of having a CK elevation
greater than ten times the upper limit of normal.
Note that at rosuvastatin doses up to and including
40 milligrams, the frequency of CK elevations is
low and similar to that observed with other
statins. Only at the 80-milligram dose where LDL-C
is reduced 65 percent does the frequency of
elevations increase above that observed for the
highest doses of pravastatin, simvastatin, or
Also observe the marked difference between
rosuvastatin and cerivastatin where at 35 to 40
percent LDL-C lowering, the frequency of CK
elevations is high.
One potential reason that the number of myopathies
with cerivastatin was high is that a much larger
percentage of hypercholesterolemic patients need
LDL-C lowering in the range of 35 to 40 percent.
In order to get this lowering with
cerivastatin, patients needed to be exposed to
doses with a greater likelihood of affecting
Overall, the skeletal-muscle data for
rosuvastatin program show that it was well
tolerated at doses up to and including 40
milligrams. At these doses, the frequency of
adverse effects was similar to that observed for
other marketed statins, but as you have seen in an
earlier presentation, greater lipid modification
can be achieved with rosuvastatin.
At the 80-milligram dose, patients
achieved an additional 2 to 4 percent LDL-C
reduction over the 40-milligram dose. However, the
frequency of adverse skeletal-muscle effects at
this dose increased above that observed for
rosuvastatin 40 milligrams and the highest doses of
other marketed statins.
Although a small number of patients
experienced adverse skeletal-muscle effects at the
80-milligram dose, many patients were safely
treated. 1200 patients under the age of 65 were
treated with the 80-milligram dose and the
frequency of myopathy in this group was 0.2
percent. Importantly, all patients who had a
significant adverse event at this dose recovered.
I would now like to turn our attention to
the effects of rosuvastatin on the kidney.
Adverse statin effects on the kidney are
well documented in terms of renal failure secondary
to myoglobinuria associated with rhabdomyolysis.
However, other potential effects on the kidney are
not well documented. Following the completion of
the initial Phase III studies for rosuvastatin, an
increased frequency of proteinuria was detected
predominantly at the 80-milligram dose.
In response to this finding, additional
investigations were performed to characterize the
frequency, magnitude, and nature of the proteinuria
and to determine the potential for rosuvastatin to
cause acute or progressive injury to the kidney.
In this section, I will present the results of
In the rosuvastatin program, proteinuria
was evaluated primarily using dipstick testing. In
the general population, a prevalence of proteinuria
up to 10 percent on dipstick testing has been
Proteinuria can have an organic etiology,
such as that which occurs in patients with
diabetes, hypertension, and urologic infections or
it can be functional. Functional causes of
proteinuria include exercise, orthostatic
proteinuria, and proteinuria associated with
Proteinuria can occur due to changes in the
glomerulus, the renal tubules or both sections of
the nephron. The types of proteins excreted can
help identify the source of the proteins.
Glomerular proteinuria is due to leakage
of albumin and other larger molecular weight
proteins through the glomerulus and is the type of
proteinuria associated with diabetic kidney disease
Tubular proteinuria, which you will see is the
pattern of proteinuria seen with rosuvastatin, is
due to reduced absorption of normally filtered
low-molecular-weight proteins. The acute and
long-term consequences of this type of proteinuria
are less well defined and must be defined in the
context of the drug or environmental factor causing
Shown here is Table 15 from the FDA
briefing document. For comparative purposes, the
data from the uncontrolled, open-label extension
trials are omitted so that the pool only contains
data from controlled clinical trials.
Presented in this table are the frequency
of developing proteinuria at any time, hematuria at
any time, or the combination of proteinuria and
hematuria at any time for a given dose of statin.
The data in the proteinuria column shows that the
frequency of proteinuria for rosuvastatin at doses
up to and including 40 milligrams is similar to
that observed for comparator statins. However, at
the 80-milligram dose, an increased frequency is
The next column shows the frequency of
hematuria with and without proteinuria. The
frequency of hematuria with rosuvastatin ranged up
to 12 percent compared to a frequency of up to 8
percent on the comparator statins. Other
evaluations, not shown here, have demonstrated that
isolated hematuria is not associated with either
rosuvastatin therapy or therapy with other statins.
The last column shows the frequency of
proteinuria in combination with hematuria from the
program. When comparing the data for rosuvastatin
with the data obtained for other statins, we find
an increased frequency of proteinuria/hematuria at
the 80-milligram dose and possibly a signal at the
40-milligram dose. But note that, at the
40-milligram dose of simvastatin, we also see a
frequency of 0.8 percent.
The observation of an increased frequency
of proteinuria and proteinuria in combination with
hematuria predominantly at the 80-milligram dose
led to a series of investigations to characterize
the magnitude and nature of these findings.
First, we evaluated the patients with the
most significant shifts from baseline in urine
protein levels to determine the amount and types of
proteins excreted. Shown in this table are total
protein and albumin excretion normalized for
urinary creatinine excretion in patients with a
shift from none or trace at baseline to 2-plus or
greater levels of urine protein.
In these patients, the median protein
excretion was only 0.6-milligram protein per
milligram of creatinine. This value correlates to
about 600 milligrams per day. Note that 150
milligrams of protein excretion per day is
Of the total protein excreted, only about
one-third was albumin. In disease states where the
glomerulus is affected, the vast majority of urine
protein excreted is albumin. Thus our, data
suggested that the proteinuria was not glomerular
Electrophoresis results and analyses of
urinary proteins from patients who developed
proteinuria showed that it was primarily tubular in
origin. Our analyses showed that the proteins
excreted were predominantly alpha-1 microglobulin,
beta-2 microglobulin, and retinol-binding protein.
These are proteins typically filtered at the
glomerulus but normally reabsorbed at the level of
Back-titration of patients in our program
from 80 milligrams to 40 milligram allowed us
another opportunity to assess the nature of the
proteins in patients with proteinuria as well as
the reversibility of the proteinuria. The data
showed that at the 80-milligram dose, the greatest
elevation in urine proteins was for
low-molecular-weight proteins and that following
back-titration to 40 milligrams, the greatest
decrease was in these same urine proteins.
Our evaluation of hematuria in patients
with proteinuria revealed that red blood cells were
present on microscopic evaluation. Myoglobin
levels were not elevated in these patients
confirming that the hematuria was not secondary to
muscle breakdown. Importantly, in our
back-titration study, the combination of
proteinuria and hematuria also reversed with
Since the predominant effect observed with
high doses of rosuvastatin was a tubular
proteinuria, we performed a series of preclinical
evaluations to explore a possible mechanism for the
I will start with the Preclinical data.
Preclinical toxicology studies for the various
statins show that all have tubular effects at very
high exposure levels.
However, in almost all of these animal models, the
doses of statin leading to this effect also caused
the animals to be moribund. Therefore, whether the
effects are a primary effect of the statin or due
to other secondary causes cannot be determined.
However, in one animal model, the
cynomolgus monkey, the effect was observed at high
doses of rosuvastatin and pravastatin, but the
doses were not high enough to cause the animal to
become moribund. The fact that the tubular
toxicity was observed in animal models with all
statins, that the types of proteins present in our
clinical studies suggested a tubular proteinuria,
and that these observations appeared to be dose
related, led us to postulate that the proteinuria
was due to an HMG-CoA-reductase inhibitory effect
in proximal tubule cells.
To explore this hypothesis, we evaluated
the effect of statins on albumin uptake in Opossum
kidney tubule cells. This is a well characterized
model for evaluating the potential effects of a
drug on renal tubules.
The results of the studies I am going to
show you were later confirmed in a human
Shown in this figure is the effect of
increasing concentrations of various statins on
albumin uptake in the Opossum kidney cells. The
statins that we are looking at are rosuvastatin,
atorvastatin, simvastatin, pravastatin and
fluvastatin. Note that with all of these statins,
with increasing concentrations, albumin uptake is
The degree of inhibition is closely
related to the degree of cholesterol inhibition in
these cells. Note that once approximately 80 to 90
percent inhibition is observed, the percentage
inhibition in albumin uptake begins to rapidly
To examine whether the observed effects
were due to HMG-CoA reductase inhibition, we also
examined the effects of adding mevalonate, the
down-stream product of HMG-CoA reductase, to the
cells along with the statin.
This is the result of one experiment. The
data show that the effects are consistent with an
HMG-CoA-reductase inhibitory mechanism. The
addition of mevalonate reverses the inhibition
observed with simvastatin and rosuvastatin and this
experiment has been repeated several times with
Having explored a potential mechanism for
the effect, we are still left with an important
question. Why is proteinuria observed following
therapy with high doses of rosuvastatin?
Two major characteristics of rosuvastatin
help to address this issue, First, rosuvastatin is
a highly effective inhibitor of HMG-CoA reductase.
Second, approximately 28 percent of rosuvastatin
systemic clearance is by the kidney, and this
occurs predominantly by tubular secretion.
For other statins, the degree of renal
excretion or the overall effectiveness in
inhibiting HMG-CoA reductase is less than that
observed with rosuvastatin.
Although we have shown that the
proteinuria was predominantly tubular in nature and
probably related to HMG-CoA reductase inhibition,
the next important question to address is whether
treatment with rosuvastatin leads to either short
or long-term renal complications.
To address the issue of short-term or
acute complications, we present here our cases of
acute renal failure from our program. Out of the
12,569 patients treated with rosuvastatin in our
program, eleven patients were reported to have
acute renal failure, one case each at the 5, 10,
and 20-milligram doses, two cases at the
40-milligram dose, and six cases at the
For the five cases at doses below
80-milligram, none were attributed to therapy with
rosuvastatin. Of the six cases at the 80-milligram
dose, four of those were associated with myopathy.
We are left with two cases of acute renal failure
at the 80-milligram dose.
In these two patients on this dose, the
etiology of the renal failure is unclear. Both
patients had symptomatology suggesting a dehydrated
state prior to the onset of renal failure and both
had other comorbidities requiring treatment with
medications which could predispose them to renal
failure independent of therapy with rosuvastatin.
These cases represent two cases out of
264 patients who initiated therapy at the
80-milligram dose and out of a total of 1583
patients treated with this dose. The current
database contains over 4000 patients treated with
rosuvastatin 40 milligrams of whom over 2000
initiated therapy with this dose. No cases of
renal failure have been attributable to therapy
with the 40-milligram dose of rosuvastatin.
Overall, the number of cases of acute
renal failure observed in this program are not
unexpected given the size of the current database
with over 14,000 patient years exposure to
Having shown that rosuvastatin is unlikely
to cause acute or short-term detrimental effects on
renal function at doses up to and including 40
milligrams, we next explored the potential for
long-term treatment in patients with proteinuria
and proteinuria in combination with hematuria to
lead to decrements in renal function.
To do this, we used a creatinine elevation
greater than 30 percent as a marker for a potential
renal effect. This is a sensitive marker and
represents a level of change of about three
standard deviations above the mean change in
creatinine observed in our placebo group for our
In evaluating long-term effects, we once
again to go our all Controlled/uncontrolled and
RTLD data pool. It is, again, our largest pool of
patients and includes patients with the longest
durations of treatment with rosuvastatin.
This analysis includes patients who had a
shift from none or trace proteinuria at baseline to
2-plus or greater proteinuria at the end of
treatment. Using this level of change identifies
subjects with a greater likelihood of developing
treatment-related proteinuria and a level of
proteinuria that should lead to changes in renal
function if an association exists.
Note that similar to the previous
analyses, the frequency of proteinuria was low and
similar at rosuvastatin doses from 5 to 40
milligrams but increased at the 80-milligram dose.
Of the patients who developed proteinuria, no
patient had a 30 percent creatinine elevation at
the end of treatment at the 5-milligram,
10-milligram, or 40-milligram doses of
Two patients had an increase at the
20-milligram dose and eleven patients at the
80-milligram dose did have an elevation. Of these
thirteen patients with elevations, only four
patients had a 30 percent increase above the
highest creatinine value observed during the
pre-randomization period. All four of these
patients were at the 80-milligram dose and two of
the patients had myopathy. For the remaining two
patients, the elevations were less than 0.5
milligrams per deciliter.
If we now look at patients with
proteinuria and hematuria, who represent a subset
of the patients shown on the previous slide, we
find similar results.
The data show that the number and
frequency of patients with this finding is
extremely low at doses up to and including 40
milligrams. An increased frequency is observed at
the 80-milligram dose.
An evaluation of patients treated for 96
weeks or longer gives additional information
regarding the long-term effects of proteinuria. In
this slide is shown information regarding
proteinuria observed at any time, at the last
visit, and the associated creatinine changes
observed at the last visit.
The data show that the frequency of
proteinuria observed at any time is greater than
the frequency observed at the last visit at a given
dose of drug. This suggests that although
proteinuria can occur, in many patients it does
decrease or resolve. This is demonstrated best in
the 80-milligram group where the frequency at any
time is 16.8 percent but decreases to 6.3 percent
at the final visit.
The back-titration data, the greater than
or equal to 40-milligram group, is also helpful
because it contains important information in almost
800 patients, in over 800 patients receiving high
doses of rosuvastatin. Note that the frequency of
proteinuria observed at any time is similar to that
observed at the 80-milligram group. However, at
the last visit, in patients who are now almost
entirely on the 40-milligram dose, the frequency of
proteinuria is similar to that observed with lower
doses of rosuvastatin.
Out of 37 patients with proteinuria at the
80-milligram only eight had proteinuria following
back-titration to 40-milligram demonstrating that
proteinuria was reversible.
The creatinine data is also helpful. Note
that no patients with proteinuria had a creatinine
elevation greater than 30 percent at rosuvastatin
doses up to 40 milligrams. Seven patients on
80-milligram had an elevation. In all seven of
these patients, the elevation resolved on
back-titration to 40 milligrams showing that the
creatinine elevations, like the proteinuria
findings were reversible.
The results for patients with proteinuria
in combination with hematuria, which is again a
subset of the patients in the previous slide,
showed similar results, no evidence for a treatment
effect at rosuvastatin doses up to and including 40
milligrams. At the 80-milligram dose, both
proteinuria and hematuria and the creatinine
elevations were reversible on back-titration.
In the FDA briefing document is a
description of a patient who had an abnormal
urinalysis with a creatinine elevation and a renal
biopsy. The clinical course for this patient has
relevance to the long-term safety of rosuvastatin
and is presented on this slide.
The patient is a 69-year-old African male
with a history of childhood renal disease, stasis
ulcers, and back pain treated with aspirin,
paracetemol, intramuscular penicillin injections,
and topical steroids. At baseline, the subject had
two urinalysis tests. One showed active sediment.
The other showed 1-plus proteinuria without active
After 18 months, the subject had a serum
creatinine measurement of 1.6 milligrams per
deciliter from a baseline of 1.1 milligrams per
deciliter. The urinalysis showed proteinuria and
hematuria. A renal biopsy was performed which
showed acute on chronic tubulointerstitial changes.
The laboratory abnormalities resolved
following discontinuation of rosuvastatin, but
proteinuria recurred upon rechallenges with
rosuvastatin 80 milligrams and atorvastatin 40
milligrams. This case shows that proteinuria can
be observed with another statin if the patient is
Another method for evaluating the
potential adverse effects of a drug on renal
function is to evaluate the long-term effects of
high-dose treatment in patients with baseline renal
laboratory abnormalities since these patients might
be expected to show a greater susceptibility to
adverse renal effects of drugs.
In this slide, we compare the effects of
treatment with at least 40 milligrams of
rosuvastatin for greater than or equal to 96 weeks
in patients with normal and impaired renal
function. Note that, in general, serum creatinine
levels tended to decrease in all groups and the
percentage of outliers was similar in patients with
normal or impaired renal function.
In summary, we have carefully evaluated a
proteinuria and proteinuria/hematuria signal with
regard to frequency, magnitude, nature, and the
potential for rosuvastatin to cause acute or
long-term renal parenchymal damage.
Our data shows that dipstick positive
proteinuria, primarily tubular in origin, was
observed predominately at the 80-milligram dose.
In a small percentage of patients, this finding was
associated with microscopic hematuria. The data
show that the finding was transient in many cases,
reversible and not associated with long-term
detrimental effects on renal function. Although
two cases of renal failure had a temporal
relationship to therapy with the 80-milligram dose,
both of these cases had other identifiable causes.
At doses up to and including 40 milligrams,
rosuvastatin was well-tolerated from the renal
An important question to address is
whether the prescribing information for
rosuvastatin should include renal monitoring. As
shown by the data, routine urinalysis or creatinine
monitoring is not necessary. The data show that
treatment wit rosuvastatin at doses from 5 to
40 milligrams does not result in acute or long-term
adverse effects on renal function. Even at the
80-milligram dose, any changes that were seen were
reversible with back-titration or stopping therapy,
so even at this dose, there is no evidence of a
long-term irreversible effect on renal function.
Having now reviewed our clinical safety
database, I would like to speak to the last
objective that we set for our program, to determine
whether rosuvastatin would have a low potential for
significant drug-drug interactions.
In this regard, I will present the results
of our drug interaction studies in the following
areas; interactions with drugs that are metabolized
through interactions with cytochrome P450
isoenzymes or PgP transporters and interactions
with drugs known to result in an increased
potential for myopathy. in particular, cyclosporine
Our drug interaction studies with the
cytochrome P450 3A4 inhibitors, ketoconazole and
erythromycin, show that rosuvastatin is not
metabolized by this route. No effect on
rosuvastatin AUC was observed with ketoconazole,
and with erythromycin, a clinically insignificant
0.2-fold decrease in AUC was observed.
Interactions with these same two drugs
along with the results of the digoxin-interaction
study also show that rosuvastatin does not interact
with PgP transporters.
Finally, the result of the fluconazole interaction
study shows that rosuvastatin is not metabolized by
cytochrome P450 2C9 or 2C19.
I would now like to address the issue of
interactions with cyclosporine and gemfibrozil.
Our drug-interaction study with cyclosporine
revealed a 7.1-fold increase in rosuvastatin plasma
Shown in this figure are the results for
rosuvastatin compared to data reported for other
statins in the literature. The results for
rosuvastatin are similar to the other statins
except for lovastatin, which appears to have the
Based on the 7.1-fold increase in
rosuvastatin AUC, the dose of rosuvastatin should
be limited to 5 milligrams when used in conjunction
Shown next is our drug interaction study
with gemfibrozil. In this trial, a 1.9-fold
increase in rosuvastatin AUC was observed. This
increase was similar to that reported for
simvastatin, lovastatin, and pravastatin but less
than the interaction observed with cerivastatin.
Once again, based on the level of increase in AUC
and the known risk for myopathy when statins are
co-administered with gemfibrozil, the dose of
rosuvastatin should not exceed 10-milligram in this
Because of the increasing use of other
fibrates, we performed a drug-interaction study
with fenofibrate. As opposed to the 1.9-fold
increase in AUC observed in the gemfibrozil
interaction study, no interaction was observed when
rosuvastatin was co-administered with fenofibrate.
Our drug interactions studies show that
rosuvastatin will have a low potential for
significant drug interactions. However, other
factors besides drug interactions may impact
exposure to rosuvastatin and could therefore impact
on safety. Data from our clinical pharmacology
program revealed that systemic exposure to
rosuvastatin was not affected by age, sex, or the
presence of mild to moderate renal impairment.
In patients with severe renal impairment,
rosuvastatin plasma concentrations increased
approximately 2 to 3 fold. Based on these
findings, we propose that the dose of rosuvastatin
is limited to 10-milligram in this population.
Rosuvastatin plasma concentrations were also
increased in patients with severe hepatic
impairment. Note that, similar to other statins,
rosuvastatin is contraindicated in patients with
active hepatic disease.
Pharmacokinetic evaluations were also
performed to assess effects based on ethnicity. We
did find that exposure to rosuvastatin was
increased approximately 2-fold in Japanese patients
in Japan. However, we do not know whether this was
due to environmental or genetic factors.
Importantly, no differences in exposure were
observed among Caucasians, Black, or Hispanic
This morning, I have reviewed for you the
safety results from our program. In this program,
doses of rosuvastatin up to and including 80
milligrams were thoroughly explored in over 12,500
dyslipidemic patients. This is the largest NDA
ever submitted for a statin.
This program was inclusive. Approximately
one-third of the patients were 65 years or older
and a high percentage of patients had
co-morbidities such as hypertension, diabetes,
renal insufficiency, and atherosclerosis.
The data show that within the proposed
5-milligram to 40-milligram dose range, the safety
profile of rosuvastatin was similar to other
At the 80-milligram dose, the frequency of adverse
skeletal-muscle and renal effects increases above
that observed for currently marketed statins.
However, even at this dose, the majority of
patients were safely treated. Importantly, all
patients with an adverse event at the 80-milligram
We have also demonstrated that
rosuvastatin will have a low potential for
significant drug-drug interactions.
For those patients at risk for significant adverse
events due to drug interactions, our proposed
labeling will reflect the necessary information.
Having now reviewed the overall safety
database, the issue of selecting appropriate doses
of rosuvastatin to market involves weighing the
potentials risks of a dose versus the potential
benefits afforded by its use.
A rosuvastatin 10-milligram to 40-milligram dose
range is appropriate for the general population of
patients with dyslipidemia.
Our data which clearly demonstrate the
excellent lipid modifying benefits of the proposed
10-milligram to 40-milligram dose range at both the
starting dose and across the dose range compared to
other currently marketed statins.
Also, within the proposed dose range, rosuvastatin
brings a high percentage of patients to recommended
NCEP lipid goals.
Why is a 10-milligram start dose
appropriate for the general population of patients
The reason is once again the overall favorable
benefit to risk of this dose.
Our data shows that the 10-milligram dose
provides additional lipid efficacy compared to the
5-milligram dose, without showing a difference in
overall safety. As previously stated, for patients
on cyclosporine, a 5 milligram dose is available.
And last, why is a 40-milligram dose an
appropriate top dose for patients with
First, our data show that the 40-milligram dose of
rosuvastatin provides additional lipid-modifying
benefits compared to the 20-milligram dose.
With regard to safety, our program has
evaluated rosuvastatin at doses up to and including
80 milligrams. Doing this has allowed us the
opportunity to understand our drug and the
potential risks associated with its use.
The 40-milligram dose was studied in over 4000
patients with a demographic similar to that of the
80-milligram group. Over 2000 subjects initiated
therapy at this dose. Our data clearly show that
this dose was well-tolerated.
Adding to the favorable benefit to risk
profile for this dose is the fact that this is not
a recommended starting dose. The 40-milligram dose
is for those patients who do not achieve the
necessary lipid-modifying effects at the
20-milligram dose of rosuvastatin.
So, in summary, using 40-milligram as the
top dose for rosuvastatin will provide an overall
rosuvastatin dose range, which is safe and provides
additional lipid-modifying benefits over current
I would now like to introduce Dr. Daniel
Rader from the University of Pennsylvania who will
briefly discuss the potential role of rosuvastatin
in the treatment of dyslipidemic patients.
The Role of Rosuvastatin
in the Treatment of Dyslipidemia
DR. RADER: Thanks very much.
I am Dan Rader. I direct a preventive
cardiology program at the University of
Pennsylvania in the Lipid Clinic there. I do
research in lipids and atherosclerosis and I see
patients with lipid disorders. I am happy to be
here today to present to you my thoughts, briefly,
on the potential role of rosuvastatin in the
treatment of dyslipidemia.
I would like to start again by reminding
you, and I think you all know at this point, that
we have had a major evolution in the Lipid
Management Guideline from 1988 to the most recent
ATP-3 Guidelines in 2001. These guidelines have
been reflected by increasing aggressiveness of
cholesterol-lowering therapy from initially a focus
on non-statin therapy to, most recently, because of
the more aggressive guidelines, a focus on
high-dose statins and combination therapy in order
to be able to achieve the kinds of aggressive
targets that are recommended in these guidelines.
I would like to point out that Dr. Don
Hunninghake, who is here with us today, has been
part of the NCP from the beginning and, in fact,
chaired the Drug Therapy Section for all three of
the adult treatment panels. So any questions you
have about NCP, we will certainly forward to Don.
What I would like to do so sort of set the
stage and explain to you why I think rosuvastatin
is an important addition to the therapeutic
armamentarium for dyslipidemia is really to point
out that, in fact, we have difficulty achieving
goals in a lot of our patients with dyslipidemia.
To go back to data that is really based on
the ATP-2 Guidelines, this slide reflects four
different studies, all performed in the mid- to
late-90's and published between '99 and 2001 really
asking, in an observational sense, how well were we
doing in terms of getting patients to the ATP-2
I will just point out here that even the
low-risk patients on the left, only about
two-thirds of them were at goal. The medium-risk
patients in the middle, only about a third were at
goal. The high-risk coronary heart-disease
patients who need to be targeted to LDLs less than
or equal to 100 by these guidelines, only about a
fifth to a quarter were at goal. So, clearly, at
that time, many patients were not at goal.
Now, you might ask, maybe patients are not
being treated or maybe they are not being
appropriately titrated and maybe many of them are
just almost at goal but not quite. But, in this
study, one of those four studies, the L-TAP Study
directed by Dr. Tom Pearson, who is also here with
us today, really shows that that is not the case.
In fact, in L-TAP, a lot of the patients who were
not at goal were actually quite far from goal.
Note that on the right a full l6.6 percent
of the patients, nearly as many as were at goal, as
shown at the left, were over 160 milligrams per
deciliter, far from their goal of 100 and 45
percent of the patients in L-TAP who needed to be
targeted to LDLs less than 100 were actually over
130. So I think this demonstrates that it is not
just in terms of getting people to goal, that we
are getting almost there but not quite there.
A lot of people have a long way to go
before they actually get their NCP goals.
This is a study by Ross Simpson and his
colleagues that looked, in a real-world setting, at
following nearly 3,000 patients asking what is
actually happening in these high-risk patients who
need to be targeted to LDLs less than 100. You can
see that, among these patients, when they were
started on a statin, 47 percent, shown on the
right, got to goal at the starting dose. But over
half did not get to goal at starting dose.
I think this is an important point. Many
patients don't get to goal on starting doses of
statins. Of that group of patients, 47 percent
were titrated but more than half were not titrated,
again reflecting an important point. Physicians
often don't appropriately titrate patients to get
them to goals.
Finally, I think perhaps most importantly,
among the patients who were titrated, only
one-third of those patients actually got to goal.
So even among titrated patients, two-thirds of the
patients did not actually get to goal. I think
this illustrates, and is something I am going to
come back to, it is actually difficult to get many
patients to goal even with appropriate titration.
This is recent data. This came out in
Circulation a few months ago from the NHANES Study.
This is data collected between 1999 and 2000 so it
really reflects treatment in the modern era with
all the current statins that we currently have on
There is a lot of data in this report but
I just thought I would focus on one key issue which
is only 47 percent of the hypercholesterolemic
patients who were being actively treated with drug
actually were adequately controlled. So I think,
again, this suggests that yes, failure to treat is
a problem but even among treated patients, failure
to actually get adequate control and treat patients
to goal is a real issue.
Now, maybe it is just that patients are
not being titrated appropriately. Certainly, that
would be a reasonable question to ask. But I want
to bring you back again to this study directed by
Dr. Christie Ballantyne who is also here with us, a
ACCESS Study, which took hypercholesterolemic
patients, randomized them to five different statins
and then titrated as needed to get to goal.
You will see again that, for LDL goals,
even patients randomized to atorvastatin titrated
as needed up to a maximum of 80 milligrams, only a
little over 70 percent of these patients actually
got to goal of LDL less than 100. For HDL
cholesterol, which, in general, is even harder to
reach, only about 60 percent of the patients on the
atorvastatin arm got to goal.
So you can see that even when
appropriately titrated in a controlled setting like
this trial, it is difficult to get many patients to
I have been focusing on our current goals
but I do have to tell you that, in the lipid field,
many of us feel that our current goals may not be
aggressive enough. I am going to show you two
slides that kind of address that issue. One is
this slide that really plots the on-treatment LDL
cholesterol levels on follow up in all the big
statin trials on the x-axis and the percent with
coronary heart-disease events on the y-axis.
You will note that, for both secondary
prevention and primary prevention, there seems to
be a clear linear relationship between the
on-treatment LDL cholesterol level and the percent
with coronary events. This is, admittedly, a crude
way to look at this but I think it gives us some
idea of this relationship.
I also want to point out that there are
two studies on this slide; the Heart Protection
Study, HPS, and the ASCOT Study that came out since
the ATP-3 Guidelines. So we have new data coming
out even since those guidelines that address this
issue of, perhaps, maybe even lower targets would
You will note that, in both of those
studies in the treated groups, the LDL cholesterol
levels in the treated group, the mean level, was
well less than 100.
I wanted to actually explore the Heart
Protection Study in just a little more detail with
this slide. I think this is really quite important
for this concept of should we be treating people
even lower. So the Heart Protection Study enrolled
people almost regardless of their cholesterol
I just thought I would show you this
analysis that the investigators did where they
looked at baseline LDL cholesterol by tertile. You
will note, in the highest tertile group, where the
mean LDL cholesterol was about 140, treatment with
simvastatin lowered LDL to a little over 100 and
lowered cardiovascular events as you can see here.
In the lowest LDL tertile in this group,
the mean LDL was slightly less than 100 at baseline
and you can see that treatment there lowered LDLs
into the 60s and also significant reduced risk. Of
course, we don't really know, if we took everybody
and lowered their LDLs into the 60s, whether we
would see even greater event reductions than we see
in the current statin trials.
But I think, based on data like this, many
of have concluded that the guidelines are very
likely to become more aggressive with regard to the
need to treat LDL. Certainly, speaking for myself,
based on data like this, I treat my high-risk
patients, patients with coronary disease and
diabetes, somewhat more aggressively than just
targeting 100. I think I would really like to see
the LDLs even lower.
I think you can imagine, as our targets
get even lower, as our practice gets even more
aggressive, it is going to be even harder to target
patients appropriately to these goals. So I would
suggest to you that, in fact, despite all the good
drugs that we have on the market, there is still a
medical need in treatment of dyslipidemia. There
is a need for more efficacious therapy to achieve a
few different goals, one of which is greater LDL
and non-HDL cholesterol-lowering at the start dose.
I have already explained to you how many
patients don't get to goal on start dose and,
unfortunately, many physicians don't appropriately
I thought I would show you just one slide
with a little bit of sort of composite data that
really addresses direct head-to-head comparisons of
rosuvastatin at its 10-milligram start dose with
commonly used start doses of other statins. So
these two panels can't be compared with each other.
They are really self-contained but if you look at
the left, these are three different trials, Trials
24 to 26, comparing rosuvastatin 10 milligrams to
atorvastatin 10 milligrams in a head-to-head
What I have selected to show you here is
actually the achievement of both the LDL
cholesterol and the non-HDL cholesterol goals,
really the ultimate goal of the ATP-3 guidelines.
You should be targeting both of these. You can see
that rosuvastatin 10 brought substantially greater
number of patients to this combined goal than
Shown on the right, Trials 27 and 28,
involved direct head-to-head comparisons of
rosuvastatin 10 with simvastatin 20 and pravastatin
20. Again, you see significantly greater bringing
patients to this combined LDL and non-HDL goal with
rosuvastatin 10 compared to the other two statins.
So I think it is safe to say that use of
rosuvastatin 10 milligrams will bring a greater
number of patients to NCP goals and, I would
suggest to you, could have substantial
public-health benefit with regard to that.
Now, I think the second need for more
efficacy therapy in treatment of dyslipidemia is
clearly to achieve greater LDL and non-HDL
cholesterol lowering at maximal dose. We really
need therapies that will get our difficult-to-treat
patients down closer to the goals that we need to
treat these patients to.
Now, to illustrate this point, I would
like to just briefly bring up familial
hypercholesterolemia. The heterozygous form of
this condition is common. There are about 500,000
patients in the U.S. with heterozygous FH for a
frequency of about 1 in 500, more common, I
believe, than Type 1 diabetes, for example.
FH is a serious disease. Even with
treated with our current drugs, the average age of
onset of coronary disease is about 45 to 50 in men
and about 55 to 60 in women and it is difficult to
treat. As I will show you in a second, most FH
patients cannot be adequately treated to NCP goals
using our current therapies.
In this slide, what I decided to do is
show you two different studies. These are two
independent studies both in heterozygous FH
patients, both directed by Dr. Evan Stein, who is
actually here with us today as well. One is a
study that you have already seen from Dr. Blasetto
on the left, but I just kind of encapsulated it
here, looking at rosuvastatin 40 milligrams and
atorvastatin 80 milligrams in these high-risk FH
patients who are being targeted to LDL less than
You can see that the rosuvastatin, as you
saw previously, got substantially more of these
high-risk FH patients to goal.
On the right, for comparison or to flesh
out this concept, I show you another study directed
by Dr. Stein that compared atorvastatin 80
milligrams, so the same comparator, to atorvastatin
40 milligram plus ezetimide, 10 milligrams. You
will note that, although these are different
studies in different populations both involving
over 600 patients, by the way, you will note that
the atorvastatin 80 performed about the same. Only
about 4 percent of these high-risk FH patients got
to goal, and the combination of atorva 40 plus
ezetimide got, again, about 17 percent of the
patients to goal.
So I think the main point here is
rosuvastatin 40 does do better than any other
single monotherapy statin that we have on the
market in terms of treating these
difficult-to-treat patients. But note that still
less than one in five patients are getting to goal.
So I think clearly, with this type of
severe hypercholesterolemic patient, the future is
being able to use rosuvastatin 40--we really need
that dose for these patients--and then adding on
combination therapies including the additional of
ezetimide to the rosuvastatin 40 to try to get more
of these patients to goal.
I would like to turn for a minute to HDL.
HDL is a common condition, low HDL, and represents
an important medical need. It is one of the most
common risk factors in patients with coronary
disease. ATP-3 importantly placed new emphasis on
low HDL as a risk factor and as a potential target
Data are increasingly suggesting that even
modest increases in HDL may translate into
substantial cardiovascular risk reduction. So I
would like to suggest that, in fact, another need
in treatment of dyslipidemia is getting better at
raising HDL cholesterol.
Dr. Blasetto already showed you data from
the STELLAR Trial looking at the comparison with
rosuvastatin with other statins in terms of HDL.
I thought what I would show you here is looking at
the same trial but asking the question what did
rosuvastatin do in terms of raising HDL in a
low-HDL group, people with HDLs less than 40.
You can see here on the left that the HDL
raising in this subgroup with rosuvastatin was
between 12 and 20 percent. So HDL raising
certainly compares favorably to the best
HDL-raising drugs we currently have on the market.
Admittedly, it is difficult to predict
what incremental reductions in LDL and incremental
increases in HDL will do in terms of reduction in
cardiovascular risk. But the NCP and the ATP-3
report did make these following estimates based on
observational studies as well as the randomized
controlled trials that we have available, and that
is that, for every 1 percent decrease in LDL
cholesterol, there would be expected to be a
reduction of coronary heart-disease risk by
approximately 1 percent and that, for every 1
percent increase in HDL cholesterol, there might be
expected to be a reduction in coronary
heart-disease risk by about 3 percent.
So I think you can imagine that if, in
fact, these do hold true, that even incremental
further reductions in LDL, further increases in
HDL, could, in fact, translate into substantial
further risk reduction for the patient.
So, in summary, I suggest to you that
there is a role for rosuvastatin in treatment of
dyslipidemia, that, first of all, the greater LDL
cholesterol and non-HDL cholesterol lowering at the
start dose will, in fact, bring more patients to
goal at start dose and I believe have public-health
benefits as a result.
Second, the greater LDL cholesterol and
non-HDL lowering at the maximal dose of 40
milligrams will make it easier for us to treat our
patients with FH, other forms of severe
hypercholesterolemia, diabetics, many of whom are
also difficult to treat, and I would suggest to you
that we really do need this 40-milligram dose to
more effectively treat these patients.
Finally, the HDL raising of rosuvastatin,
although incremental, certainly would be suggested
to result in increased reduction in cardiovascular
events as well.
So, in summary, I would suggest to you
that, in fact, rosuvastatin does provide an
important and valuable addition to the therapeutic
armamentarium for the treatment of dyslipidemia.
Thank you very much.
DR. BRAUNSTEIN: Thank you for a lovely
We will now take a fifteen-minute break
and reconvene at 10:45 for questions from the
committee to the sponsor.
Questions from the Committee
DR. BRAUNSTEIN: We will open up the
session for questions and answers from the
committee. The committee will also have an
opportunity for questions, both the FDA and the
sponsor, following the FDA's presentation. But now
we will restrict ourselves to sponsor's
Questions? Dr. Hennekens?
DR. HENNEKENS: I was extremely favorably
impressed with the size and scope of this
development program as well as the comprehensive
presentations. Dr. Orloff, in his comments, gave
us some two focused sets of charges that, perhaps,
might merit further consideration. One was he
spoke of perhaps the need for further safety data
directly comparing the 20 and 40 milligrams at the
40-milligram start dose and talked about 600
patients or more. Secondly, further clarification
of the new onset of proteinuria directly at the 20
and 40-milligram doses, Dr. Hutchinson's Slide
CS24, if taken at face value, suggested that those
rates were 0.3 at 20 and 1.3 percent at 40 which,
if real, would be a relative risk of 4.3.
So, perhaps, further clarification of
those two issues might be helpful in our
deliberations, either now or sometime during the
DR. BRAUNSTEIN: Do you want to respond to
DR. HUTCHINSON: Just to clarify your
question, Dr. Hennekens, you are interested in the
DR. HENNEKENS: The second part related to
your presentation was from your Slide CS34.
DR. HUTCHINSON: Yes; the FDA's analysis
of our data.
DR. HENNEKENS: Yes. If you look at the
right-hand column for the 20 versus the
40-milligram dose, it was 0.3 to 1.3, just further
clarification of that would be helpful to me.
DR. HUTCHINSON: If I can show you the
data from our largest pool of patients which will
give you a better feel for the overall frequency of
proteinuria-hematuria in our program, I may be able
to address your specific questions.
This was data that was presented during my
presentation. Now, this takes all patients in our
program that had urinalysis and creatinine
measurements. It looks at what happens in patients
with the most significant degrees of change
regarding proteinuria and hematuria from baseline
and then what happened in those patients at the end
of treatment with regard to creatinine changes.
As you can see, the percentage of patients
that had proteinuria along with some level of
hematuria ranged from 0.10 to 0.2 percent at doses
up to 40 milligrams. We see an increased frequency
of this finding at the 80-milligram dose.
What is critical here is to know whether
or not this finding is associated with any effects
on renal function so we use this sensitive marker,
which is creatinine elevations greater than 30
percent, to evaluate whether or not the proteinuria
and the hematuria that was there had an effect on
As you can see, 0, 0, 1, 0, 8. When we go
back and evaluate these patients because, in our
program, what we use for creatinine baseline was
the value of creatinine closest to Week 0.
However, a number of these patients had multiple
baseline creatinine measurements.
This identified a group of nine patients.
If you go back and evaluate those patients, what
you find is that, in almost all cases, what happens
here is that the patients don't even have an
elevation in creatinine greater than 30 percent of
the maximum value observed the baseline. In the
few numbers of patients, the one or two patients
that do have an elevation, the creatinine elevation
in these patients is less than 0.5 milligrams per
We do have, in a couple of these patients
also follow up after discontinuation of therapy.
Following discontinuation of therapy, what happens
is that creatinine elevation resolved in those
patients where we had follow up.
I think the 96-week data which looks at
proteinuria, hematuria and the creatinine
elevations also gives you some important
information here as well. These are patients that
are going to be exposed for a mean of 2.4 years
with our drug.
Here in these patients we are once again
looking at this combination of proteinuria and
hematuria to determine whether or not it was
associated with any change in serum creatinine,
this 30 percent marker. Just to give you an idea
if somebody had a creatinine change from
0.6 milligrams per deciliter to 0.8 milligrams per
deciliter, they would fulfill this criterion.
But, if we look at this, what we find,
first of all, after 96 weeks, we had one patient in
the 40-milligram dose group that met this
criterion, one patient at the 10-milligram group.
If we look for creatinine increases, we find that
none of these patients had a creatinine increase.
We see that, at 80 milligrams, five patients had an
increase but, in our program, because we
back-titrated patients from 80 to 40 milligrams, we
had the opportunity to follow many of these
80-milligram patients longer term.
What we find is that, in these patients
once they get back-titrated, look at the frequency
of proteinuria and hematuria. It now approximates
what we see at very low doses of rosuvastatin.
These are patients receiving high doses.
Importantly, those five creatinine elevations are
So, from our patient, we have a very large
dataset. We have a very large dataset in general
looking at the 5, 10, 20, 40, and 80-milligram
doses. I think we have provided very good data to
show what the estimates of this finding will be at
the various doses and we have also provided very
substantial data regarding what the short and
long-term consequences of the findings are.
What we have found is that, in general,
transient, reversible, not associated with any
effects on renal function and, at the same time,
that 40-milligram dose is giving patients
additional significant LDL-C reductions which
DR. BRAUNSTEIN: Could you explain what
the difference is between your table that you just
showed and Table 15 from the FDA? I know there are
minor differences in numbers of patients but it was
1.3 versus--you had 0.2 percent up there and they
had 1.3. So why the difference?
DR. HUTCHINSON: The difference is simply
the type of evaluation that was done. In the FDA
evaluation, we are looking here at proteinuria,
hematuria and the combination at any time during
the program. So this takes into account if someone
had proteinuria at Week 2 but didn't have anything
at the end of the day, they would get picked up in
It is a very good analysis if you want to
look for potential signals. But if you want to
evaluate what is happening with regard to renal
function, you need to follow these patients out
long-term and see what occurs. That is the
analysis that I followed up with.
DR. BRAUNSTEIN: Dr. Woolf?
DR. WOOLF: Can you put those back?
DR. HUTCHINSON: Yes.
DR. WOOLF: Sort of following up on the
same issue, what is the time course of the
development of proteinuria and hematuria? Is it
seen within a few weeks? Is it seen in a few
months? You talked about the etiology of
proteinuria but not of the hematuria. Do you have
any idea where that is coming from?
Then I have a final comment about your
suggestion about not really--a recommendation that
we do not need to put I guess the term is a
"warning" in the labeling about monitoring for
DR. HUTCHINSON: Several questions to
DR. WOOLF: Time course, etiology.
DR. HUTCHINSON: Yes; time course, first.
Thank you very much. With regard to time course,
you can see that proteinuria occurs as early as two
weeks following treatment. We observed this
predominantly at the 80-milligram dose. However,
proteinuria can occur later. But the tendency for
the proteinuria, as I showed you with the 96-week
data, is for the proteinuria, should it appear, to
resolve. But it can occur as early as two weeks.
Now, the second question was with regard
to the hematuria. With regard to the hematuria, we
don't have an explanation for the hematuria. If I
can please see the Trial 99 table from the FDA
document, that does address, in some respects, the
In response to our earlier findings from
the program, we went forward and did a prospective
study looking at rosuvastatin 40 milligrams versus
simvastatin 80 milligrams to try to characterize
the frequency of this finding in other statins and
also to understand a little bit about what was
happening with the proteinuria,
This study did not have a placebo lead-in,
a placebo treatment arm, but there was a dietary
lead-in, a six-week dietary lead-in period. During
that time period, patients had one or two
urinalysis samples. They were off statin therapy.
As you can see, during this time period,
we had a 3.4 percent frequency of proteinuria. The
proteinuria greater than 2-plus was 0.6 percent and
hematuria greater than 1-plus was 7.9 percent
during this period.
Following treatment with simvastatin 80
and rosuvastatin 40, we find that hematuria, it was
roughly similar in both of the treatment groups.
We do see a suggestion, however, that there tended
to be slightly more proteinuria with rosuvastatin.
We are not completely clear on where the
hematuria is coming from with regard to the
proteinuria-hematuria potentially seen with
rosuvastatin, particularly at the 80-milligram
dose. What we know about the proteinuria-hematuria
is it seems to follow the same type of course as
the proteinuria does, which is it is transient,
resolves with back-titration from 80 milligrams to
40 milligrams and, once again, not associated with
any acute or long-term effects on the kidney.
DR. WOOLF: Then, in regards to your
suggestion about the labeling, you have roughly 100
patients who have been followed on a 40-milligram
dose for, I think you said two-and-a-half years.
DR. HUTCHINSON: Yes.
DR. WOOLF: One of whom developed
hematuria-proteinuria. We are talking about
patients who are going to be on this essentially
for a lifetime. While two-and-a-half years is
rewarding, a lifetime is, hopefully, a lot longer
If you don't monitor for it, you will
never be able to know that it disappears when
it--to back-titrate. So it is non sequitur. You
have to monitor to able to know that you have to do
something about it. So, to me, it is a disconnect.
DR. HUTCHINSON: That's true if you are
using the 80-milligram dose. However, we are not
suggesting that we are going to be treating
patients with the 80-milligram dose. Now, you say
100 patients, but if I can please see the 96-week
data again, because it is not really just 100
patients that we looked at in this program.
People were not dropping out of our
program because of proteinuria and because of
increased creatinine. So we had the opportunity to
follow these patients long-term.
If we look at the 96-week data, which I
showed earlier, you are talking about 761 patients.
We are also talking about over 1,165 patients in
our program that have been exposed to doses greater
than or equal to 40 milligram for 48 weeks. So,
again, it is not only 100 patients. It is over
DR. WOOLF: At the 40-milligram dose, it
DR. HUTCHINSON: At the 40-milligram dose,
here, that have never been exposed to the
80-milligram dose; correct. It is 100. But, once
again, if this drug was causing significant effects
on the kidney, one would expect that what we are
seeing at 80 milligrams, you would expect to see
the residual of that effect once you drop these
patients back to 40 milligrams.
We don't see it. In fact, the frequency
of the finding approximates the lower dose. So,
with regard to monitoring, you are dealing with
patients with atherosclerosis, diabetes,
hypertensions. These people have fluctuations of
30 percent in creatinine that can occur at almost
It is more likely that they will get a
fluctuation of 30 percent in their serum creatinine
from the other medications that they are on or
their disease than they will due to rosuvastatin or
DR. BRAUNSTEIN: Dr. Follman.
DR. FOLLMAN: I would like to make a
comment about reversibility. I think it will be
easiest to make this comment if you bring up Slide
I think that is not the one I want but I
think I can make the point with this anyway. When
were are looking proteinuria and hematuria and so
on, these are parameters that will wax and wane
with time with biological processes that the
patients are undergoing with measurement error and
who knows what. So, if you look over the course of
the trial and say, "Oh; I have a high rate of
proteinuria," and then you look at the very last
visit and note that it is lower, to what extent is
that evidence of reversibility or to what extent is
that evidence that you have a biological process
that fluctuates some.
So to really sort that out, you would need
a control group in some way. So this relates to
your comments when you say when you back-titrate I
think from 80 milligrams to 40 milligrams amongst
those who had proteinuria, the rate went down.
Once again, I would like a control group
to really feel comfortable that this is evidence
primarily of reversibility rather than just
fluctuations where you happen to catch them when
they had proteinuria and then, when you
subsequently measure it one more time, it is gone.
So we would like to believe that is
evidence of reversibility, I think. But we just
can't really conclude that without a control group.
DR. HUTCHINSON: Let me show you some data
from a substudy that we performed in one of our
open-label extension trials where we took patients
that were on the 80-milligram dose and, when we
were back-titrating these patients, we performed
very careful timed urine measurements as well as
other analyses in these patients.
DR. FOLLMAN: So this is where the group
as a whole is back-titrated at basically a fixed
point in time?
DR. HUTCHINSON: This is within four weeks
of back-titration of patients from 80 to 40
DR. FOLLMAN: What was the reason for
back-titration? Was it based on the patient's
evidence of proteinuria or clinical
DR. HUTCHINSON: Not at all.
DR. FOLLMAN: So it was done to everyone?
DR. HUTCHINSON: This was done to everyone
in the program because we had looked carefully at
our 80-milligram data and it felt, at that time,
that the efficacy that we were getting did not
justify its use in the general population because
of some of the adverse events we were seeing.
However, this is very strong evidence here
that the proteinuria was reversing. These are
patients on rosuvastatin 80 milligrams with
proteinuria. These are patients with elevated
urinary total proteins when they are on the
80-milligram dose and subsequently back-titrated to
40 milligrams. This is four weeks later.
DR. FOLLMAN: This is the whole group?
DR. HUTCHINSON: This is not everyone on
80. This is done in selected sites. The reason it
had to be done that way is because we were doing
careful timed urine collections as part of the
I will show you the whole group in a
DR. FOLLMAN: Okay.
DR. HUTCHINSON: But, in a very careful
evaluation of these patients, you see that going
from 80 milligrams to 40 milligrams, we get a
substantial reversal and decrease in the
proteinuria so, once again, suggesting that the
proteinuria was resolving.
Now, if we take the patients overall, and
there are 752 patients back-titrated here, from 80
milligrams to 40 milligrams, we see that the
frequency of 1-plus or greater proteinuria goes
from 12 percent down to 4.8 percent and greater
than or equal to 2-plus 7.5 percent down to
With regard to proteinuria-hematuria, 21
out of 46 of the patients here at a urine protein
dipstick blood greater than or equal to 1-plus. 20
of the 21 no longer had that combined effect at
four weeks after the back-titration, once again
showing the reversibility, showing this goes away.
DR. FOLLMAN: Did you do the previous
slide in all the patients, the one with the figure
where you showed it went down nicely? It seemed
that that was in the selected group that had high
protein, high urinary protein--
DR. HUTCHINSON: This one was in patients
with elevated urinary total protein.
DR. FOLLMAN: So, once again, this is not
surprising to me that there would be a tendency for
it to go down. Once again, I want to sort out the
reversibility versus just fluctuations going down.
You select them with high values, look at them
again, and they go down.
DR. HUTCHINSON: I believe what we need to
look at is the totality of the data here. This
signal is not seen in a lot of people, first of
all. It is seen predominantly at the 80-milligram
dose. We were not going to be treating patients
any longer with the 80-milligram dose so, in order
to be able to do these types of evaluations, these
patients provided a very nice cohort to study and
we used them to study the reversibility of the
What is very important here is the
consistency of the findings. The key issue here is
if proteinuria or proteinuria-hematuria is
important from the standpoint of causing an effect
on renal function, then, certainly, the patients
that had the greatest levels of proteinuria and
proteinuria-hematuria and have it for the longest
duration, which would potentially be those with it
at the end of the day, would be the most likely
group to have a creatinine elevation if an
But what is amazing here is, out of the
thousands of patients in the program, you evaluate
these people and then you come down with one or two
people at up to 40 milligrams and a handful at 80
milligrams. When you back-titrate the patients on
80 milligrams, the findings seem to reverse.
DR. BRAUNSTEIN: Dr. Carpenter?
DR. CARPENTER: Related to the same issue,
it seems that the concern that the proteinuria is
trying to predict is the concern of progressive
loss of creatinine clearance. We are using
proteinuria here as an overall marker of glomerular
Yet, the studies that you have shown us
that examine the nature of the proteins in the
urine are evidence that this is primarily a tubular
problem. I wondered if you had explored the
tubulopathy any further; that is, maybe some of
this discordance is related to the fact that
glomerular disease is not what is happening but
tubulopathy is what is happening. Have you looked
at other tubular functions such as potassium
wasting, renal-tubular acidosis, things that could
potentially be comorbid events here that the
proteinuria could be marking and that we have not
really seen any data to effect.
DR. HUTCHINSON: Yes. We certainly did
I can show you some data here regarding
serum calcium, phosphorous and potassium in the
patients with or without proteinuria on
rosuvastatin 80 milligrams. You can see that there
are really no differences in the level of serum
creatinine, serum phosphorous, or serum potassium
in patients with or without the proteinuria. So
this seems to be an effect predominantly on tubular
transport within the tubules. We are not getting a
Fanconi's type of picture here with other
abnormalities present as well.
DR. BRAUNSTEIN: Dr. Watts?
DR. WATTS: Just to clarify. To me,
titrate means that you are adjusting the dose based
on some indicator. In the changing from 80
milligrams to 40 milligrams, it seems to me that
back-titrate is not the correct term, that you
simply reduce the dose.
DR. HUTCHINSON: That's fair.
DR. WATTS: I want to explore what Dr.
Woolf raised and what Dr. Follman raised and that
is the time course and is this resolution or is
this variability? The slide you just showed
indicated that 20 percent of patients in the
80-milligram group had proteinuria.
Table 15, and that analysis that you
had--Table 15 of the FDA shows, by my calculations,
there are probably 180 patients in the 40 and
80-milligram dose who had proteinuria and over 300
patients who had hematuria.
It seems to me you can look at the
occurrence of these events by visit. That would be
more convincing to me than what you see at the last
visit represents a resolution rather than
variability because my bet is, if this is sort of
an erratic process, that what you would see at any
visit is what you see at the last visit. It is
only if you look over the totality of the exposure
that you see when it shows up.
Whether or not this is a problem, a
clinically meaningful problem, I don't know but I
share Dr. Carpenter's concern that changes in serum
creatinine may not be the best way to determine
whether or not this is a clinical problem.
DR. HUTCHINSON: Can I please see the data
that looks at our control pool and looks at the
evaluations of proteinuria at various time points,
please. I will try to address your question using
some of our control data.
This slide is a little complicated. I was
hoping to avoid this. But, having said that, what
we are doing here is using the controlled-trial
database. One of the issues within any time
analysis is it can certainly be influenced if one
of the groups has more visits, if the durations of
therapy are longer.
We do know that for the 40-milligram dose
group in our program, we started a large controlled
trial and we had more visits and we were
specifically trying to characterize some of the
findings in our program using that trial. So, in
general, there was a tendency for patients on 40
milligrams to have more visits and we know that,
from our data, if you look at the placebo data, you
can see proteinuria even on placebo.
But here, what we are looking at, is there
are patients in our program that had shifts in
urine protein to 2-plus or greater. This was our
standard definition when we were analyzing our
data. So that is why I am showing you this.
Numbers may change a little bit, if you
are looking at slightly different levels of
proteinuria but I think the trends are roughly the
same. We are looking at Week 4, Week 6, Week 8 and
Week 12. Notice, for some of the doses you see
zeros, and that is because, in the trials that
those patients were involved in, there just wasn't
a visit at that time.
But here, at four weeks, in the
rosuvastatin trials, we can see a signal up to 1.9
and 1.7 at the 40-milligram dose and, at the
80-milligram dose of rosuvastatin, it is 7.3, 8.4
percent at Week 6 ranging anywhere from 1 to 1.5.
If we go out here to Week 8, what we are seeing is
1 percent, 1.2 percent. If we go out here to Week
12, we see 0.8 percent.
Now let's look at our comparators. At
Week 4, we saw 0.3 percent here with simvastatin,
80 milligrams. If we go over to Week 6, we see a
rate as high as 1.6 percent on placebo, 1 percent
on atorvastatin 20. If we go out now to Week 8, we
see 2 percent here in atorvastatin, 22 percent in
simvastatin 20 and, if we go out here to Week 12,
we see rates as high as 1.3 percent.
So, at the end of the day, the proteinuria
can be seen as early as Week 2 but it appears at
various time points. There is no consistency with
regard to, "I can tell you by Week 6 you are going
to see all the proteinuria."
As you can see from this analysis, you can
see rates as high as 2 percent in patients on the
comparators where there is a reasonable number of
patients on the comparators. So what we are seeing
at 40 milligrams does not appear to be
significantly different than what we are seeing
with the comparators.
The fact that we did more measurements at
40 milligrams is probably contributing in part to
the signal that you start to pick up at the
40-milligram dose group when you look at
proteinuria at any time.
I hope that helps.
DR. WATTS: I would like to see that slide
for a little bit longer.
DR. HUTCHINSON: Sure.
DR. WATTS: It could be made less
complicated if, where you have no patients, you
simply put an X and not a 0 because there are lot
of 0s in the incidence column where there are 0 in
the number-of-subjects column.
But, following the 40-milligram dose
across, it looks like there is I don't know whether
to call it an incidence or prevalence as it
continues, because I don't know whether they are
the same patients or new patients, but it is
between 1 to 2 percent. That is not consistently
seen for any of the other groups.
DR. HUTCHINSON: Part of the reason is
because they haven't been measured at some of these
weeks so you are not picking it up. But, at the
end of the day, I think the important point here is
that you can pick up proteinuria with the other
statins. It is there. Whether or not that
represents background or whether or not the statin
is causing an effect, we don't know.
But, if you remember, we presented one of
the cases in a South African patient who had a
creatinine elevation along with proteinuria and
hematuria and, in that particular patient, the
rosuvastatin was stopped. The abnormalities went
away. The patient was rechallenged with
rosuvastatin. The abnormalities, the urinalysis
abnormalities, came back. It was stopped. It went
The patient was then rechallenged with a
lower dose of rosuvastatin, 40 milligrams. And the
urinalysis findings came back. So I think in some
patients there is the potential that this effect
can be seen.
But whether or not the numbers we are
seeing here are background or actually a
statin-related effect, especially for the other
statins, it is difficult to know. I think, with
rosuvastatin at 80 milligrams, we are certainly
seeing a signal and there is potentially a signal
at the 40-milligram dose.
But, once again, the key thing here is
what happens in this patients with small amounts of
proteinuria? Is the proteinuria at the end of the
day resulting in any long-term or short-term
detrimental effects on renal function? This
program is a huge program and we are just not
DR. BRAUNSTEIN: Dr. Kopp?
DR. KOPP: I have a couple of comments.
Maybe I could start with this slide. One of the
problems here is that it only twelve weeks of data.
You could conclude, on the basis of what you said,
that 80 milligrams of rosuvastatin is safe because
there is no proteinuria at Week 8 and Week 12. I
think it simply points out the more valid issue is
what happens after 48 weeks and 96 weeks.
DR. WATTS: There are no patients in the
80-milligram group at Week 8 and Week 12.
DR. KOPP: Oh; is that right? Sorry.
DR. HUTCHINSON: That's right. Exactly.
There are no patients.
DR. WATTS: That is why I am saying it is
an unnecessarily complicated slide because there
are 0s where there are zero potential to have data.
DR. KOPP: Fair enough. Thank you for
clarifying that. There are two issues I would like
to make as comments. The first, one of the reasons
for this variability is that dipstick proteinuria
is not the ideal way to measure it. It may be the
only practical way in a database of 12,000 patients
but I think we need to recognize that urine
concentration has a lot to do with whether the
dipstick is positive or not.
In fact, if you want to be devil's
advocate, you could say, with progressive
tubulointerstitial disease, one of the first
features of renal function to decline is the
ability to concentrate. In a more dilute urine,
you would tend not to see the proteinuria.
I am not necessarily sure that that is
what is going on here, but I think some of this
variability of proteinuria here, say, 4 percent of
the time and then only present in 2 percent of the
patients at the end of the study, may have to do
with the limitations of dipstick proteinuria. So
that is one comment.
The other comment is I think the model
that I am thinking about, and I suspect some of the
other people are, too, is this an agent that causes
tubulopathy that may take a year or two to appear
and cause proteinuria in a small fraction of
patients, maybe 2 percent, maybe 4 percent, of
patients which eventually will damage glomerular
filtration by damaging the effect of glomeruli as
well and lead to a rise in creatinine. But that
may go on at three and four and five and six years.
I think we can't exclude that possibility.
Many tubular toxins, in fact, take many years to
cause their damage. Lithium would be a chronic
class example. So that is two comments.
A couple of specific questions. Could you
put Slide CS25 which was your data about
protein-to-creatinine ration and
albumin-to-creatinine ratio. The point here is
that glomerular proteinuria typically has more than
50 percent albumin; that is, more than 50 percent
of the protein in the urine is albumin.
As you point out, 0.3 is less than 50
percent of 0.8. The probability is that that
represents a mean of many patients. So, do you
have the specific data what fraction of these
roughly 300 patients had glomerular proteinuria?
Was it, in fact, zero or was it a few?
DR. HUTCHINSON: It is not zero. Where we
have SDS page information, it does show that the
predominant pattern that you see is the SDS page,
the tubular pattern.
If I can please see the data from-we
looked at patients in our program that developed
1-plus or greater proteinuria to look at what types
of patterns would be seen on gel electrophoresis.
I want the slides with the patients--
DR. KOPP: While we are looking for that,
the page data are nice, but, in fact, you can get
it from the 300 patients where you measured
albumin, measured protein, measured creatinine and
simply determine. That might be interesting to do.
DR. HUTCHINSON: I can show you some data
in that regard, too, because we did so some of
these measurements as well. After I speak to these
two slides, I think it would be worthwhile with
regard to evaluation of our renal findings, we had
several experts in the field of nephrology look at
our data and advise us on how to appropriately
evaluate our data in this large database.
We have Dr. Ed Lewis with us today. I
think it would be appropriate for Dr. Lewis to make
a couple of comments in this regard as well. But
here we are looking at patients on the 80-milligram
dose in our program. I think that this has--
No; this is not the slide I would like to
see. Can I please have the slide with the patients
who went from 0 to 1-plus proteinuria. That has a
couple of things reversed on it. Give me the data
on the back-titration from 80 to 40 with the
different types of proteins that were measured.
DR. BRAUNSTEIN: While they are looking
for that, perhaps we can take the next question.
DR. KOPP: Can I ask a second question
which changes, now, to the use of the drug in
cyclosporine. Cyclosporine is also a
proximal-tubule nephrotoxin. Do you have any
comment about the occurrence of increased
proteinuria in patients who were on cyclosporine,
rosuvastatin was added, and then the same question
with regard to creatinine elevation. Again,
cyclosporine elevates creatinine by hemodynamic
mechanisms, later by fibrosis. Does rosuvastatin
potentiate those effects?
DR. HUTCHINSON: The studies with
cyclosporine were very short-term. Predominantly,
they were pharmacokinetic studies and we did not
pick up issues with regard to proteinuria or with
creatinine elevations in those patients. But, in
those studies, we were using low doses. I
apologize for the time it took to get this slide.
Hopefully we will find the other one in a second.
These are patients in the substudy who had
timed overnight urine collections, back-titration
from 80 milligrams to 40 milligrams. These are the
various proteins that were looked at along with
n-acetal-glucose aminidase activity. What we see
at the 80-milligram dose is that the proteins that
were most prevalent in the urine were alpha-1
microglobulin, retinal-binding protein. We had
lower levels of beta-2 microglobulin, albumin
transferrin and IgG, but part of this was just due
to stability issues with beta-2 microglobulin.
What is critical here is, once we
back-titrated patients to 40 milligrams, the
largest changes that we were observing were in the
alpha-1 microglobulin and retinal-binding protein
groups. We saw smaller changes with regard to the
Have we found the other slide? We will
have to try to find that over the break.
DR. KOPP: One other question, and I can
yield the floor. How about glycosuria, a follow up
on Dr. Carpenter's question. Any glycosuria in
DR. HUTCHINSON: No.
If the Chairman will allow, I can have Dr.
Lewis come up and comment.
DR. BRAUNSTEIN: I think what we would
like to do is to actually continue this discussion
after the FDA's presentation. But I wanted to give
Dr. Neylan an opportunity to ask his question.
DR. NEYLAN: Thank you, Mr. Chairman. Two
question, both relating to renal effects. The
first, as the sponsor has shown, I think the
tubular-protein composition is certainly
consistent--or, rather, the protein composition is
certainly consistent with a tubular site. I am not
convinced yet that I understand whether this is a
functional or more structural effect, though.
The reason I raise that is that this issue
of hematuria arising in roughly the same incidence
or prevalence as the proteinuria suggests the
possibility that, indeed, there is a structural
element here. As we know, a protein in the urine
can be found in a variety of otherwise normal
states. Hematuria is quite a bit less frequent.
The dipstick is certainly a convenient way
of looking for the presence of hemoglobin but it is
a surrogate for a microscopic examination of urine
sediment. Urine sediment that shows a lot of cells
and casts certainly raises the possibility of an
activity or inflammatory state or even a state of
increased turnover, be it tubular cells or
So my question is when you received the
approvable letter roughly a year ago and went back
to do more detailed analysis of these renal
findings, did you have opportunity to incorporate
some evaluations of the microscopic elements of the
urinalysis, look at sediment beyond just the
dipstick and so could you share those with us?
DR. HUTCHINSON: I don't have a slide to
show that, but we did have urine sediment
evaluations on our patients with proteinuria and it
did not show that these patients were having an
active urine sediment.
DR. NEYLAN: How about in those patients
that had hematuria by dipstick? Were you able to
do any microscopic examinations of those urines?
DR. HUTCHINSON: We know it is red blood
cells. Unfortunately, it is impossible now to go
back at this stage and look at those previous
urines simply because you need to look at fresh
samples for the appearance of the red blood cells.
This is something that we are doing now in our
studies going forward but we don't have the samples
to go back and evaluate them for red-blood-cell
DR. NEYLAN: My second question relates to
cyclosporine. You mentioned that you were able to
do a small study in heart-transplant recipients who
were receiving cyclosporine as presumably one of
the elements of their maintenance immunosuppressive
I am going to guess that, since most
heart-transplant patients are not on cyclosporine
monotherapy, that this was a multidrug regimen.
Were you able to tease out the potential impact or
interaction of cyclosporine from any other elements
in this regimen since there are well-known
interactions with a variety of other
DR. HUTCHINSON: No; we have not done
DR. NEYLAN: I noticed the labeling of
other statins does not necessarily get as specific
as cyclosporine but mentions that, in the face of
immunosuppressants, there can be warnings attached.
DR. HUTCHINSON: One thing that we did do
was go back and look at our database and look at
our hypertensive patients on various types of
antihypertensive treatments because some
antihypertensive treatments certainly can have
effects on the tubules to see if patients having
treatment with those antihypertensive agents
increased the possibility of having proteinuria.
So here we are looking at our highest
proposed dose of rosuvastatin, the 40-milligram
dose. We are looking at the association with
various antihypertensive drugs of proteinuria, so
we are looking at ARBs, ace inhibitors, calcium
channel blockers and diuretics.
As you can see, yes would mean that they
were on the drug. No means not on the drug. This
is the percentage with 2-plus or greater
proteinuria, the percentage with 1-plus or greater
proteinuria. As the data shows, there is no
evidence that patients on these drugs would have an
increased frequency of proteinuria.
So, once again, if there was some
susceptibility there, we would expect to see an
increased frequency and that is not happening.
DR. BRAUNSTEIN: Thank you.
We will now have the FDA's presentation.
Following that, there will be some more questions
from the committee, both to the FDA and to the
DR. BRAUNSTEIN: Ms. Mele will give the
MS. MELE: Good morning.
My name is Joy Mele. I am the FDA
statistical reviewer for the Crestor application.
I will be giving a short presentation on a
few efficacy issues, so we are back to efficacy
now. First, I will show the dose-response effect
on LDL for rosuvastatin in three studies, Studies
8, 33 and 65. Then I will present a detailed
comparison of rosuvastatin to atorvastatin using
data from Studies 33 and 65. Lastly, I will
describe the effect of rosuvastatin on HDL.
To show the dose-response effect on LDL, I
will presenting data from three dose-response
studies in Type IIa and IIb patients, Studies 8, 33
and 65. You have already seen data today from
Studies 8 and 65. 8 was combined with Study 23 in
the sponsor's presentation. I will show you the
data from these two studies and also from Study 33.
Recall, the doses in Study 8 were 1, 2.5,
5, 10, 20 and 40. In Study 33, dosing ranged from
5 milligrams to 80 and, in Study 5, an open-label
study, dosing ranged from 10 milligrams up to 80
Studies 33 and 65 both included
atorvastatin arms. The sample sizes in each
treatment group varied considerably across these
studies with only about 15 in each group in Study 8
to about 40 in Study 33 and about 160 in each
treatment group in Study 65. The baselines were
similar across the studies at about 190 milligram
This is a plot similar to what the sponsor
has already shown you. Here is plotted the mean
LDL percent change from baseline for the full dose
range of rosuvastatin studied in the three trials I
just described. I wanted to show here the
consistency of the results across these individual
Study 8 is shown in blue, Study 33 in
green and Study 65 in red. The Y axis goes from 0
to 70 percent.
Looking at each dose, and taking into
consideration the variability of these estimates, I
would conclude that the responses are very similar
across these studies. A dose response is evident
in each study although, at the high end of the dose
range, the 40 and 80-milligram doses, we see small
differences of about 2 to 3 percent suggesting a
leveling off of effect.
The benefit of doubling 20 milligrams to
40 is evident in Studies 8 and 33, and the sponsor
showed this very nicely on a side earlier, but not
so evident in Study 65, the very large trial. Note
that the 5-milligram dose, which is plotted here,
provides about two-thirds of the lowering seen for
the 40-milligram dose, about 42 percent for 5 and
60 percent for 40. Dr. Lubas will make some
additional comments about the 5-milligram dose in
From the data presented earlier by the
sponsor, it was evident that the rosuvastatin is
more potent than any other marketed statin on a
milligram-per-milligram basis. Looking across the
dose range, though, at what doses are rosuvastatin
and atorvastatin comparable? Is twice the dose of
atorvastatin needed to obtain comparable LDL
lowering? How about four times the dose? I will
address these questions in the next few slide
slides by showing the treatment differences in the
96-percent confidence intervals.
First let's look at a comparison of
rosuvastatin versus two times atorvastatin. Using
the data from 33 and 65, the two largest
dose-response studies in Type IIa and IIb patients,
I have plotted the mean treatment difference in the
90-percent confidence interval for the difference.
The values to the left of the 0 line favor
rosuvastatin while the values to the right favor
At the top of the graph is 5 milligrams
versus atorvastatin 10. Then there is 10 versus
20, 20 versus 40 and 40 versus 80. Study 33 is
plotted above Study 65 for each of the pair of
Focusing first on the blue boxes, the
results look quite consistent, a difference of
about 4 percent in favor of rosuvastatin is seen.
The confidence intervals for Study 65 are tighter
than for Study 33, as would be expected, given the
large sample size, and the differences seen in
Study 65 are statistically significant. This is
the 0 line, and so you can see that these estimates
do not overlap 0.
I just wanted to point out about the
confidence intervals. These confidence intervals
suggest that it is plausible that differences as
small as 1 to 2 percent could be seen, not a
clinically important difference. But they also
suggest that differences as large as 8 percent
could be seen as well, which would be an important
Since 40 is the highest proposed dose for
rosuvastatin and 80 is the highest marketed dose
for atorvastatin, I would like to examine this
Looking first to the graph on the left,
these box plots show that 25th, 50th and 75th
percentiles. The individual observations are
plotted over the boxes. The overlap of the box
plots show that some patients taking atorvastatin
80 can achieve LDL-lowering comparable to changes
seen for 40 milligrams of rosuvastatin although the
relationship of the boxes shows that a higher
percentage of rosuvastatin patients will achieve
significant decreases. The cumulative
distribution plot to the right here, reiterates
this point. The red line is rosuvastatin and the
blue line is atorvastatin 80. The difference
between the lines is illustrated by this vertical
line at 60 percent.
About 23 percent of atorvastatin patients
had a decrease of 60 percent or more while about
twice as many rosuvastatin patients had a decrease
of 60 percent or more.
What about four times the atorvastatin
dose? Notice that all the confidence intervals
overlap 0. Three estimates favor atorvastatin and
two favor rosuvastatin, so there is no consistency
across the estimates although the two estimates
from Study 65--that would be these two
estimates--are close to 0 and suggest
Now let's go on to HDL.
There were four placebo-controlled,
fixed-dose, phase-III, trials in the original
Crestor application. The HDL results for these
trials are listed here. The second column shows
the baseline. The baseline in Studies 8, 23 and
24, all studies in Type IIa/IIb population, is
about 50 milligrams per deciliter. In the Type
IIb/IV population of Study 35, the baseline is
The underlying values indicate those
changes significantly different from placebo. In
general, the results are variable across the
studies in significance and also in magnitude of
effect although some consistency is seen for the
10-milligram dose which would be this column here.
Note that the placebo subtracted estimates
for the last two studies are both 8 percent. The
lack of a dose effect is evident in both Studies 8
and 35 where higher doses show lower mean
responses. You can see that here.
Now we will examine further the
rosuvastatin dose response for HDL using the data
from the large trial, Study 65.
These box plots are of the HDL percent
change from baseline for rosuvastatin in red and
atorvastatin in blue. The grey boxes represent the
confidence intervals about the medians. You can
see a slight shift upwards of the confidence
interval when going from 10 milligrams to
20 milligrams of rosuvastatin. This represents
about a 2 to 3 percent more increase in HDL. Doses
about 20 appear to afford no additional benefit so
there is no clear dose-response relationship.
The results from Study 33, the other trial
I showed you earlier, show a very similar pattern
of research for rosuvastatin that is shown here.
The box plots for atorvastatin are clearly
shifted downward. You can particularly see this if
you focus on the 75th percentile at the top of the
boxes. The atorvastatin response is more variable
compared to the rosuvastatin response. If I showed
you again the results from Study 33, you would see
even more variability among the atorvastatin arms.
So, in summary, rosuvastatin is marginally
more effective than two times the dose of
atorvastatin achieving about a 40 percent more
lowering on LDL. It is clear that some patients
may achieve comparable effects to rosuvastatin 40
with atorvastatin 80. The HDL effects are
variable. There is no clear dose-response
relationship with only a further increase of about
2 to 3 percent seen when doubling the dose from 10
This lack of a dose response is consistent
with what we see in the statin class although the
atorvastatin results suggest more variability at
the higher doses than what was seen for
DR. BRAUNSTEIN: Thank you.
MS. MELE: Dr. Lubas will speak next.
DR. BRAUNSTEIN: We will go on to the
safety and dosing presentation by Dr. Lubas.
Safety and Dosing
DR. LUBAS: Good morning.
My name is William Lubas. I am a medical
officer in the Division of Endocrine and Metabolic
I will be speaking to you today focusing
on the issues of safety and dosing of rosuvastatin.
In the first part of this talk, I will focus on
I will first address the issue of muscle
toxicity associated with the use of statins.
Statin-associated muscle toxicity has included CK
elevations alone, myopathy, which is defined as CK
elevations greater than ten times the upper limit
of normal associated with muscle symptoms, and
rhabdomyolysis, which is a clinical diagnosis which
commonly refers to patients with very high CK
elevations such as greater than 10,000 units per
liter and/or patients requiring hospitalization for
Since safe and effective statins with a
low risk for the development of rhabdomyolysis are
already currently available, any future statins
which would be approved need to have a comparable
or lower risk for this adverse event.
This slide shows the incidence of CK
elevations and myopathy seen with the use of
statins. It summarizes the data from the
clinical-development programs from Baycol,
rosuvastatin, and for the pool of currently
marketed statins. The incidence of myopathy
includes all cases regardless of etiology.
While rosuvastatin doses of 40 milligrams
and lower are within the range seen for other
approved statins, there is a clear break at 80
milligrams. The two highest marketed doses of
Baycol of 0.4 and 0.8 milligrams and the
rosuvastatin dose of 80 milligrams had a similar
frequency of CK elevations greater than ten times
the upper limit of normal and myopathy, as you can
see comparing here to here.
The frequency of CK elevations and
myopathy is still higher for the 80-milligram dose
of rosuvastatin compared to all marketed statins
even if one looks only at treatment-related cases
as reported in the sponsor's presentation earlier
Baycol, at the highest dose, was found to
cause severe myopathy and rhabdomyolysis in
open-market use with a frequency not acceptable for
the benefits of the drug with regard to LDL
cholesterol lowering. Rosuvastatin, at 80
milligrams, is only marginally more effective than
the 40-milligram dose and, relative to currently
marketed statins, was associated with
rhabdomyolysis in phase III of clinical
The expectation of greater risk in the
less-structured and less-monitored setting of
market use led to the conclusion of the
unapprovability of this high dose.
Now I will switch to the discussion of
treatment-emergent renal adverse events now
previously observed with statins which the sponsor
has discussed in detail in their presentation and
which they attribute to the decreased protein
uptake by renal tubular cells due to
statin-mediated inhibition of HMG CoA-reductase in
This slide shows the percentage of
patients in the largest rosuvastatin safety data
pool shown here, including all patients from all
controlled and uncontrolled trials as well as
real-time data with proteinuria by treatment group
at any visit.
Proteinuria is defined as
dipstick-positive urine of plus-plus or greater
with at least one grade increase from baseline
during the trial. The n here refers to the total
number of patients in each group. The simvastatin,
pravastatin and atorvastatin data came from
controlled trials only while the rosuvastatin data
included both controlled trials and open-label
extensions and so had more patients as can be seen
here. It also had longer duration of patient
The rosuvastatin data gave an appearance
of an increase across the range of those who were
studied, but there was a clearly visible transition
at the 80-milligram dose where the peak incidence
was 17 percent compared to all other statins which
had a frequency of less than 4 percent and were
similar to the frequency of 3 percent seen with
This was true for patients on rosuvastatin
in both the controlled and open-label extension
trials which I will show more clearly in a
subsequent slide. Patients who were back-titrated
from the 80-milligram dose to 40 milligrams of
rosuvastatin according to the sponsor, as discussed
already earlier today, had a decrease in the
frequency of proteinuria from about 8 percent at
their last visit on 80 milligrams to about 2
percent at their first follow-up visit on 40
milligrams. This suggests the reversibility of the
proteinuria seen here at 80 milligrams.
This slide shows the percentage of
patients with proteinuria at any visit summarized
by the numbers on top of the bars subgrouped by
dose and categorical increase in creatinine from
baseline, as shown in the box. Proteinuria, again,
refers to dipstick-positive urine of plus-plus or
greater with at least one grade increase from
baseline during the trial.
In this slide, the data are presented for
both the controlled trials, the lighter colors, and
the open-label extension, the darker colors and
labeled OLE. The serum-creatinine data is
superimposed on the bars using tricolors subdivided
by each group, as shown in the insert.
Red corresponds to an increase of greater
than 30 percent from baseline. Green corresponds
to an increase of between 20 and 30 percent from
baseline and blue corresponds to patients with less
than 20 percent increase from baseline.
So, for example, looking at the
80-milligram dose of rosuvastatin in the open-label
extension trials, 17.2 percent of the patients had
proteinuria at any visit. 11 percent of these
patients also had an increase of less than 20
percent; that is, this would also include patients
with creatinine decreases from baseline.
About 2 to 3 percent of these patients had
an increase of 20 to 30 percent represented by the
green bar and 3 to 4 percent had an increase of
greater than 30 percent represented by the red bar.
I should just focus again that this data,
in contrast to what the sponsor has presented, is
data at any visit. The creatinine data is taken at
the exact same time as the proteinuria data.
The higher incidence of proteinuria seen
with the 80-milligram dose is also associated with
higher incidences of serum-creatinine increases of
both greater than 20 percent and greater than 30
percent from baseline. The greater-than-20-percent
increase from baseline increase would correspond to
the red and green bars, and the
greater-than-30-percent increase from baseline
would correspond to the red bars alone.
At doses below 40 milligrams, the
frequency of proteinuria and creatinine increases
from baseline is much lower. So it is hard to draw
clear conclusions about these dose effects. The
fact that the frequency of proteinuria appears to
be higher in the open-label extensions compared to
similar doses in the controlled trials suggests
that the incidence of proteinuria increases over
time. But this can be confounded by the irregular
frequency of sampling of these trials.
In addition to proteinuria, a subset of
these patients had also had microscopic hematuria.
This slide shows the percentage of patients with
combined proteinuria and hematuria at any visit,
subgrouped, again, by dose and categorical increase
in creatinine from baseline. Again, this is at any
visit, not just at the last visit.
Here hematuria represents
dipstick-positive urine of plus or greater with at
least one grade increase from baseline. Over half
of the patients with proteinuria at the
80-milligram dose shown in the previous slide also
had hematuria shown here. So, for example, for the
closed-label trials, 6.1 percent of the patients
out of 11.8 in the previous slide and for the
open-label extensions it was about 10.5 percent out
of 17.2 percent of the patients.
This suggests that these two effects may
be related. About 2 percent of the patients on 80
milligrams had a visit at which they had combined
proteinuria, hematuria and an increase in
creatinine of greater than 30 percent shown by the
red boxes. This was true for both the open-label
extension and the controlled trials at
80 milligrams and suggests an effect on renal
In contrast, only about a third or less of
the cases of proteinuria at doses of 40 milligrams
and lower, seen in the previous slide, also had
hematuria in this slide. The incidence of
hematuria at these doses shown here is below 2
Again, at doses below 40 milligrams of
rosuvastatin, the frequency of combined proteinuria
and hematuria associated with the creatinine
increases from baseline is much lower and so it is
hard to draw any clear conclusions about dose
While statin-mediated inhibition of
protein uptake in renal tubular cells, described by
the sponsor, may partially explain the proteinuria
seen with rosuvastatin, it does not explain the
hematuria or increase in serum creatinine seen
primarily at the 80-milligram dose.
These are cases that the sponsor has
already addressed but I would like to focus on
these a little more. In addition to the
proteinuria and hematuria seen with rosuvastatin,
there were two cases of acute renal failure of
unclear etiology in patients receiving 80
milligrams of rosuvastatin for 15 to 31 days.
One of these patients had acute tubular
necrosis noted on renal biopsy. There was also one
case of chronic tubulo-interstitial nephritis after
18 months of therapy on 80 milligrams of
atorvastatin. The renal biopsy was consistent with
acute and chronic interstitial inflammatory changes
and this patient had a positive rechallenge test
with worsening proteinuria and hematuria with
repeat exposure to rosuvastatin. This patient also
had a positive rechallenge test, as mentioned
before, to another less potent statin suggesting
that this may, in reality, be due to a class
These three cases, while serious,
represent a small number of the patients out of the
total of 12,000 exposed to rosuvastatin or the
1,500 exposed to the 80-milligram dose. It is
important to note that all of these cases were seen
at 80 milligrams and all patients improved after
the drug was discontinued.
There are still several unanswered
questions about the renal findings. First, have
the renal findings been adequately addressed?
Clearly, most of the findings were at the
80-milligram dose which will not be approved. They
were largely reversible on back titration from 80
to 40 milligrams and even patients with serious
adverse events recovered after the drug was
stopped. But the effects at lower doses are less
Next, is some sort of monitoring needed,
possibly at higher doses? Would urinalysis looking
for proteinuria, hematuria and/or serum creatinine
be useful for monitoring? Also, what further
investigations are warranted to better understand
the mechanism and the clinical course of these
effects? Finally, is this a class effect of
In summary, the frequency of CK elevations
and myopathy at doses of 40 milligrams or less is
similar to that seen with other statins. The
frequency of a 30 percent increase in serum
creatinine above baseline in patients with
proteinuria of plus-plus or greater is higher at
doses of 80 milligrams compared to lower doses.
There is a suggestion that there also may
be an increase with 40 milligrams but the overall
incidence of proteinuria is so much lower at 40
that it is hard to draw conclusions from the
current data. Clinical evidence suggests the renal
findings may not be entirely explained by the OK
model of inhibition of protein uptake by renal
The final issue the advisory committee
will be asked to address is dosing. This slide
presents mean LDL cholesterol data from two pooled
trials, 8 and 23, in patients with Type IIa and
IIb, primary hypercholesterolemia and mixed
dyslipidemia with mean baseline LDL cholesterol
levels in the range of 185 to 194.
The sponsor is proposing a start dose of
10 milligrams which would produce a mean LDL change
of minus 50 percent. However, the 5-milligram
dose, which is also available, is very effective at
lowering LDL cholesterol and would produce mean
reductions in LDL of minus 43 percent.
In one study of Type IIa and IIb patients,
the 5-milligram dose resulted in 67 percent of the
cohort reaching ATP-3 goals compared to 80 percent
at the higher dose of 10 milligrams, a difference
of only 14 percent more at the higher dose. It
should be emphasized that, for many patients, the
5-milligram dose may be an adequate start dose
based on baseline LDL levels and targets of
This slide summarizes the recommended
start dose for all currently marketed statins and
the proposed start dose for rosuvastatin. The
sponsor is currently proposing a start dose of 10
milligrams, 20 milligrams for patients with severe
hypercholesterolemia with LDL cholesterol baseline
levels above 190 milligrams per deciliter and 5
milligrams only for patients who are also receiving
The 10-milligram proposed start dose for
rosuvastatin would give mean LDL
cholesterol-lowering greater than seen with all
other currently approved statin start doses, yet
the 5-milligram dose is also very effective.
This slide describes the mean
LDL-cholesterol reduction in statin-therapy
clinical-event trials and it compares them to that