FDA Perspectives on ErythropoiesisStimulating Agents ESAs for Anemia of Chronic Renal Failure: Hemog

1
FDA Perspectives onErythropoiesis-Stimulating
Agents (ESAs) for Anemia of Chronic Renal
FailureHemoglobin Target and Dose Optimization

• Joint Meeting of the Cardiovascular and Renal
  Drugs Advisory Committee and the Drug Safety and
  Risk Management Advisory Committee
  
• Ellis F. Unger, M.D.
• Deputy Director for Science (Acting)
• Office of Surveillance and Epidemiology (OSE)
• Center for Drug Evaluation and Research (CDER)
• September 11, 2007

2
Outline

• What is the correct hemoglobin target for ESAs
  for anemia of chronic renal failure?
  
• Randomized controlled clinical trials
• Observational data
• ESA responsiveness
• Dose optimization challenges
• Potential path forward

3
Approved Erythropoiesis-Stimulating Agents (ESAs)
for Anemia of Chronic Renal Failure

• Epoetin alfa (Epogen /Procrit) 1989
• Darbepoetin alfa (Aranesp) 2001

4
What is the Correct Hemoglobin Target (or Range)
for ESAs in Anemia of Chronic Renal Failure?
5
Randomized Controlled Clinical Trials of Discrete
Hemoglobin Targets

• Normal Hematocrit
• CHOIR
• CREATE

6
Normal Hematocrit Study

• Goal
• Assess risks and benefits of achieving a normal
  hematocrit in hemodialysis patients with
  clinically evident CHF or ischemic heart disease
  
• Conducted
• 1993 to 1996, with follow-up through 7/1997

7
Normal Hematocrit Study Design

• Open label
• All subjects received Epoetin alfa
• 11 randomization to
• low hematocrit 303 (Hgb 101 g/dL) or
• normal hematocrit 423 (Hgb 141 g/dL)
• At entry, patients
• clinically evident ischemic heart disease or
  CHF
  
• on hemodialysis
• clinically stable on Epoetin alfa
• 1 endpoint Time to death or non-fatal MI

8
Normal Hematocrit Study Results

• Randomization
• n 634 to normal hematocrit (42 3)
• n 631 to low hematocrit (30 3)
• Terminated early
• Our study was halted when differences in
  mortality between the groups were recognized as
  sufficient to make it very unlikely that
  continuation of the study would reveal a benefit
  for the normal-hematocrit group and the results
  were nearing the statistical boundary of a higher
  mortality rate in the normal hematocrit group.
  
• NEJM, 1998

9
Normal Hematocrit Study Results
Hematocrit ()
Time (months)
Mean (95 CI) Hematocrit by Study Month (NEJM,
1998)
10
Normal Hematocrit Study Results
final log rank p 0.01
Probability of death or non-fatal MI ()
Time (months)
Death or Non-Fatal MI by Study Month (NEJM, 1998)
11
Normal Hematocrit Study Results

• Components of Primary Endpoint

12
Normal Hematocrit Study

• Negative Association Between Mean Hemoglobin
  (throughout study) and Mortality

Lower target
Higher target
n
FDA analysis of data collected through 7/5/97
13
Normal Hematocrit Study Summary

• Hemodialysis patients with clinically evident CHF
  or ischemic HD
  
• Targeting a hematocrit of 42 3 versus 30
  3 (Hgb 14 1 versus 10 1 g/dL) associated
  with increased mortality and cardiovascular
  morbidity.
• Somewhat paradoxically, higher mean hemoglobin
  concentrations were associated with survival in
  both treatment arms.

14
CHOIR Study Design

• Open label, Epoetin alfa
• Patients
• no Epoetin alfa in past 3 months
• not on dialysis
• hemoglobin
• 11 randomization to hemoglobin
  11.3 or 13.5 g/dL
  
• Primary endpoint composite of mortality, CHF
  hospitalization, non-fatal stroke, non-fatal MI

15
CHOIR Study Results

• Randomization
• 715 to hemoglobin of 13.5 g/dL
• 717 to hemoglobin of 11.3 g/dL
• Terminated early
• The DSMB recommended that the study be
  terminated in May 2005 at the time of the second
  interim analysisbecause the conditional power
  for demonstrating a benefit for the
  high-hemoglobin group was less than 5 for all
  plausible values of the true effect for the
  remaining data. NEJM, 2006

16
CHOIR Study Results
mean hemoglobin (g/dL)
time (months)
Mean (95 CI) Hemoglobin by Study Month (NEJM,
2006)
17
CHOIR Study Results
Primary Composite Endpoint
High hemoglobin group
Probability of composite event
Low hemoglobin group
Time (months)
18
CHOIR Study Results 1 Endpoint Components
19
CHOIR Study Results

• Negative Association Between Mean Hemoglobin
  (throughout study) and Mortality

Lower target
Higher target
n
FDA exploratory analysis
20
CHOIR Summary

• For pre-dialysis patients, administration of
  Epoetin alfa to target a Hgb of 13.5 versus 11.3
  g/dL is associated with increased mortality and
  CHF hospitalization.
• Paradoxically, higher mean hemoglobin
  concentrations were associated with survival in
  both treatment arms.

21
CREATE Study Design

• Open label, Epoetin beta
• Patients
• mild anemia (hemoglobin 11 to 12.5 g/dL)
• not on dialysis
• no prior ESAs
• 11 randomization to normal hemoglobin (13 15
  g/dL) or subnormal hemoglobin (11 12.5 g/dL)
  
• Epoetin beta begun in subnormal group once
  hemoglobin

22
CREATE Study Design Results

• Primary composite endpoint sudden death, MI,
  acute heart failure, stroke, TIA, angina
  requiring hospitalization, peripheral vascular
  disease complication or cardiac arrhythmia
  requiring hospitalization
• Randomization
• 301 to normal hemoglobin (1315 g/dL)
• 302 to subnormal hemoglobin (1112.5 g/dL)

23
CREATE Study Results
Median (SD) Hemoglobin by Study Month NEJM, 2006
24
CREATE Study Results

• Primary endpoint events (NEJM, 2006)
• - 58/301 in normal hemoglobin group
• - 47/302 in sub-normal hemoglobin group
• HR 0.78 (95 CI 0.53 1.12)
• Few endpoint events despite broad composite
  endpoint
  
• Results directionally support lower hemoglobin
  target

25
Observational Data 58,058 U.S. HD Patients
Database from DaVita, Inc.
Regidor DL, Kopple JD, Kovesdy CP, et al. J Am
Soc Nephrol. 2006171181.
26
NKF K/DOQI Guidelines (2006)

• Cohort-based observational trials and
  cross-sectional analyses of large medical
  databasesconsistently show that higher achieved
  hemoglobin values (including 12 g/dL) are
  associated with improved patient outcomes . The
  failure of observational associations to be
  confirmed by interventional trials renders use of
  observational evidence unsuitable to support the
  development of an intervention guideline
  statement.
• www.kidney.org/professionals/KDOQI/guidelines_anem
  ia/cpr21.htm
  

27
Data to Support Ideal Hemoglobin Target (1)
? morbidity/mortality ? morbidity/mortality
Normal hematocrit (hemodialysis)
10
14
?
? morbidity/mortality ? morbidity/mortality
11.3
CHOIR (pre-dialysis)
13.5
Observational data, by association only
28
Data to Support Ideal Hemoglobin Target (2)

• Observational data from HD patients
• Exploratory analyses of NHCT and CHOIR
• associations between higher mean hemoglobin
  concentration achieved and survival
  
• Association does not prove causality
• Achieved hemoglobin ? hemoglobin target.
• J-shape relation suggests that there is some Hgb
  concentration that is excessive in the CRF
  population.

29
Data to Support Ideal Hemoglobin Target (3)

• Perhaps patients who achieve higher hemoglobin
  concentrations have less advanced renal disease
  and lower CV disease burden ? better outcomes.
  
• We are not aware of a RCT that demonstrates, in a
  convincing way, that a higher hemoglobin target
  is associated with less cardiovascular morbidity
  and mortality than a lower target.

30

• Dose Optimization Challenges
• ESA Responsiveness

31
1. Could we prospectively identify
hypo-responders, at higher risk of cardiovascular
events?2. If hypo-responders could be
identified, how should they be treated?

• Dose Optimization Challenges
• ESA Responsiveness

32
Survival by Hemoglobin (Normal HCT Study)

• Less responsive to ESAs

Lower target
Higher target
n
FDA analysis of data collected through 7/5/97
33
Survival by Mean Weight-Adjusted Epoetin Alfa
Dose (Normal HCT Study)
Dose and responsiveness are inversely rated
Fraction surviving
Highest dose less responsive to ESAs
FDA exploratory analysis
time (months)
34
Prospective Evaluation of ESA- Responsiveness
(Normal HCT Study) (1)

• FDA exploratory analysis
• NHCT study provided unique opportunity to assess
  ESA-responsiveness.
  
• Stable HD patients, maintained on Epoetin alfa
  hematocrit 27 to 33 for 4 weeks
  
• Subjects randomized to normal hemoglobin target
  group had standard protocol-mandated ESA
  challenge
  
• Epoetin alfa dose increased by factor of 1.5 on
  study entry

35
Prospective Evaluation of ESA- Responsiveness
(Normal HCT Study) (2)

• Epoetin alfa-responsiveness calculated for
  patients who received constant weekly Epoetin
  alfa dosing for 2 to 6 weeks following study
  entry.
• Responsiveness º slope of hemoglobin-time
  relation throughout the 2- to 6-week period
  (linear regression).

36
Prospective Evaluation of ESA- Responsiveness
(Normal HCT Study) (3)

• 618 patients randomized to normal hemoglobin
  target
  
• EPO-responsiveness could be calculated for 414
• 117 patients experienced a decrease in
  hemoglobin, despite a 50 increase in Epoetin
  alfa dose
  
• 297 patients experienced no change or an
  increase in hemoglobin divided in quintiles

37
Prospective Evaluation of ESA- Responsiveness
(Normal HCT Study) (4)

• Assessments
• Survival by initial Epoetin alfa-responsiveness
• Overall Epoetin alfa responsiveness (mean
  hemoglobin concentration throughout study) by
  initial Epoetin alfa response

38
Initial Epoetin Alfa Responsiveness Does Not
Predict Subsequent Mortality in the NHCT Study
FDA exploratory analysis
39
Initial Epoetin Alfa Responsiveness Does Not
Predict Subsequent Mortality in the NHCT Study
FDA exploratory analysis
40
Initial Epoetin Alfa Responsiveness Does Not
Predict Overall Hgb Response in the NHCT Study
less initial response more initial response
Mean Hgb throughout study (g/dL)
0
Hgb rate of change with initial 50 increase in
EPO dose (g/dL/week)
FDA exploratory analysis
41
Could we prospectively identify hypo-responders,
at higher risk of cardiovascular events? In the
NHCT Study, where patients had protocol-mandated
50 increase in EPO dose on study entryInitial
Hgb response did not predict subsequent
mortality, and did not predict overall Hgb
response. ESA responsiveness may need to be
assessed on ongoing basis.
42
ESA-Hyporesponsiveness in a Single Patient
43
Conclusions Dose Optimization Challenges ESA
Responsiveness (1)- Prospective
identification of hypo-responders may be
difficult (i.e., erythropoietic response to an
ESA challenge) - Identification of
hypo-responders is feasible in practice
44
Conclusions Dose Optimization Challenges ESA
Responsiveness (2)For hypo-responsive patients,
the labeling suggests a search for causative
factors, but does not explicitly state a maximum
ESA dose, or what constitutes an adequate attempt
to raise hemoglobin.Key Unanswered Question
whether less responsive patients or those with
specific risk factors would experience fewer
cardiovascular events if attempts were not made
to raise their hemoglobin to some ideal target.
45
Dose Optimization Challenges
46
Dose Optimization Challenges

• ESA labeling warns against excessive rate of rise
  of Hgb ( 1 g/dL per 2 weeks)
  
• Is risk related to hemoglobin response?

47
Dose Optimization Challenges Cycling in a
Subject from the Normal Hematocrit Study
48
Dose Optimization Challenges Cycling in a
Subject from the Normal Hematocrit Study
49
Dose Optimization Challenges Cycling in a
Subject from the Normal Hematocrit Study
Week 50, hemoglobin 9.9, rate of change 0.6
g/dL/week
Week 49, hemoglobin 9.3
50
Normal Hematocrit StudyDynamic Analysis of
Relations Between Serious Cardiovascular Events,
Prevailing Hemoglobin, and Preceding Hemoglobin
Rate of Change
serious CV events/ patient-yr
Hemoglobin (g/dL)
Hemoglobin rate of change (g/dL/wk)
FDA Analysis
51
Dose Optimization Challenges

• ESA labeling warns against excessive rate of rise
  of Hgb ( 1 g/dL per 2 weeks)
  
• Hemoglobin oscillations are associated with
  serious cardiovascular events
  
• Due to underlying patient characteristics?
• Worth trying to prevent?

52
Dose Optimization Challenges Development of ESA
Dosing Algorithms

• Limit hemoglobin oscillations
• Prevent excessive hemoglobin rates of change
• Prevent overshoot
• Provide appropriate means to identify and treat
  hypo-responders

53
Summary

• Best RCT data available Ideal hemoglobin
  target is 10 g/dL for HD patients 11.3 g/dL for
  pre-dialysis patients
  
• Data to support a hemoglobin target as high as 12
  g/dL are observational in nature and of limited
  utility
  
• association ? causality
• achieved hemoglobin ? target hemoglobin
• Unknown if ESA-hyporesponsive and/or high-risk
  patients should be treated differently
  
• Little data to show that current labeling
  addresses how best to reduce hemoglobin overshoot
  and cycling

54
Potential Path Forward

• Hemoglobin target conduct prospective,
  randomized, controlled cardiovascular outcome
  study(ies) to determine optimum hemoglobin
  target(s)
• Consider, a priori, disparate targets based on
  risk factor(s)
  
• Develop new dosing paradigm(s)
• Special dosing strategies might be considered for
  hypo-responsive patients and those at higher risk
  of CV events
  
• Strategy could consider futility
• Test prospectively in RCT(s)

Return to Unger