Advisory Committee Briefing Document
Cardiovascular Safety of Rosiglitazone
Endocrinologic and Metabolic Drugs Advisory Committee
Drug Safety and Risk Management Advisory Committee
Meeting on July 13-14, 2010
Prepared by:
GlaxoSmithKline
Philadelphia, PA
(submitted on June 10, 2010)
AVAILABLE FOR PUBLIC DISCLOSURE WITHOUT REDACTION
This Briefing Document is provided by GlaxoSmithKline for confidential use by the US
Food and Drug Administration and its Advisory Committee members in the month
prior to the public meeting. In accordance with FDA’s procedures, this Briefing
Document will be available for public release (without redaction) by FDA within 1-3,
days prior to the public meeting.
EXECUTIVE SUMMARY
Background: The FDA has re-convened a Joint Meeting of the Endocrinologic and
Metabolic Drugs Advisory Committee with the Drug Safety and Risk Management
Advisory Committee to review the totality of new and existing cardiovascular safety data
on AVANDIA® (rosiglitazone maleate) Tablets since these committees met in July 2007.
Following the 2007 Advisory Committee Meeting, additional cardiovascular safety data
on rosiglitazone (RSG) have become available from a number of sources: meta-analyses,
observational studies, and controlled trials.
Meta-analyses: GlaxoSmithKline (GSK) conducted an updated integrated clinical trials
(ICT) analysis incorporating 10 new studies into the prior 42-study ICT. This 52-study
ICT analysis did not show a significant difference between RSG and control for
myocardial ischemic events (hazard ratio 1.098 [95% CI 0.890, 1.354, p=0.38]). This
result is in contrast with the result from the 42-study analysis with a hazard ratio for the
myocardial ischemic events of 1.30 [95% CI 1.004,1.685, p=0.047], available at the time
of the July 2007 Advisory Committee meeting. The change in the hazard ratio between
the 42-study and the 52-study datasets indicates the instability of any of the hazard ratios
determined using this methodology, particularly when the event rate is low and the
confidence intervals are relatively wide. Given these mixed results, randomized clinical
trial data are required to further inform on the cardiovascular safety of RSG.
Observational studies: A number of observational studies have been conducted to
evaluate the occurrence of major cardiovascular events during treatment with a
thiazolidinedione (TZD) (RSG or pioglitazone [PIO]). GSK has comprehensively
reviewed the literature on observational cardiovascular studies that have been published
since June 2007 in which RSG was studied. Twenty-one such studies were identified that
constitute a much larger dataset than the dataset available in 2007, which did not show an
increase in CV risk with RSG. The new observational studies varied by design, length of
time of observation, the duration of diabetes for an individual, the definition of
cardiovascular events, the medications investigated and the comparisons made. In
studies comparing RSG to other anti-glycemic agents, six studies showed no statistically
significant difference in the risk of myocardial infarction for RSG compared to other antiglycemic
agents while three studies showed a statistically significant increased risk of
myocardial infarction for RSG. However, there was no increased risk of cardiovascular
mortality or all cause mortality. In those observational studies that directly compared
rosiglitazone to pioglitazone, the majority (n= 7) of studies showed no statistically
significant difference in the risk of myocardial infarction between rosiglitazone and
pioglitazone; others (n=3) showed that the point estimate favored pioglitazone compared
to rosiglitazone, and one study had mixed results depending on the medication based
strata examined. Only direct head-to-head comparisons in prospective randomizedcontrolled
trials enable widely convincing conclusions about the comparability of RSG
and other anti-glycemic agents.
Controlled CV outcome clinical trials: Since 2007, final data have become available
from several large long-term cardiovascular outcome studies. The RECORD study, an
open label, large, long-term, prospective, randomized, controlled trial in 4,447 type 2
diabetic patients, was designed to evaluate cardiovascular (CV) outcomes for RSG vs
metformin (MET) and sulfonylurea (SU). With a total of 644 primary events, RECORD
achieved its primary endpoint, according to its pre-specified non-inferiority margin of 1.2
[hazard ratio 0.99 (95% CI 0.85-1.16)]. In addition, the result for the secondary endpoint
of major adverse cardiovascular event (MACE) was 0.93 (95% CI 0.74-1.15). This study
demonstrated the non-inferiority of RSG compared to MET and SU on the combined
outcome of CV death and CV hospitalization.
Additional supportive data for cardiovascular safety comes from the National Institutes of
Health (NIH) sponsored BARI 2D study. BARI 2D is a large, long-term, prospective,
randomized, controlled trial in 2,386 patients with type 2 diabetes and stable coronary
artery disease. A post-hoc analysis conducted by NIH of RSG treated patients in the
study did not show an increased risk for all cause mortality, MACE, or myocardial
infarction (MI). These results are consistent with the primary result from RECORD
showing no increase in overall CV risk.
Glycemic efficacy: Rosiglitazone has demonstrated glycemic efficacy, both in the
short-term, with up to 1.5% reduction in glycated hemoglobin A (HbA1c), and in the
long-term, showing more durable glycemic control compared to MET and SU in two
long-term studies (ADOPT and RECORD). Short- and long-term reductions in
microalbuminuria have been demonstrated with RSG treatment. In RECORD there was
no increase in microvascular complications compared to MET and SU, agents that have
demonstrated reductions in microvascular events in the United Kingdom Prospective
Diabetes Study (UKPDS) study.
Safety profile: Rosiglitazone has been extensively studied, both before and following
approval, and its safety profile is well characterized. The association of RSG with fluid
retention and congestive heart failure (CHF), which are recognized TZD class effects,
was well characterized from short-term studies. Another TZD class effect, increased risk
of fracture, only became apparent initially in long-term RSG studies but has now been
included in labeling for both RSG and pioglitazone (PIO).
Comparison of rosiglitazone vs. pioglitazone: There are no completed large, longterm,
head-to-head randomized clinical trials comparing RSG and PIO. A recent Science
Advisory from the American Heart Association and American College of Cardiology
Foundation entitled “Thiazolidinedione Drugs and Cardiovascular Risks” summarized the
available data concerning TZDs and cardiovascular risk, with a focus on ischemic heart
disease (IHD) events. It concluded that there is no reliable evidence to support the choice
between RSG and PIO [Kaul, 2010]. The TIDE study (Thiazolidinedione Intervention
and vitamin D Evaluation), an FDA post-marketing requirement, was initiated after the
2007 Advisory Committee Meeting and will provide a head-to-head randomized
comparison of RSG and PIO. TIDE has been approved by the majority of Institutional
Review Boards (IRBs) and Ethics Committees around the world who have thus far
reviewed the protocol. GSK remains committed to the successful conclusion of the TIDE
study.
Summary and conclusion: Taken together, the totality of the data from ICT,
observational, and large controlled clinical studies continue to support the overall positive
benefit risk profile of RSG as an important medicine for type 2 diabetes mellitus (T2DM)
patients. Better durability of glycemic control with RSG compared to both SU and MET
has been demonstrated in two long-term studies. This durability benefit has the potential,
in real life conditions, to reduce microvascular complications of type 2 diabetes and to
avoid the need for additional therapy, including insulin. Treatment with RSG does not
demonstrate an increased risk in macrovascular complications compared to MET and SU.
Therefore, the overall benefit risk profile for RSG remains positive.