Surrogate Endpoints

1
Surrogate Endpoints
2
The Surrogates Story 3 Drug Trials 2 FDA
Policy Issues

• 1. CAST Trial Cardiac Arrhythmias
• 2. Concorde AZT for AIDS
• 3. Erythropoietin for Dialysis Pts
• 1. Accelerated Approval (AA)
• 2. Patient-Reported Outcomes (PRO)

3
PVCs
4
(No Transcript)
5
(No Transcript)
6
(No Transcript)
7
Managing the Wild Card of Cardiovascular
Disease- Sudden Cardiac Death

• Up to half of the ½ million cardiovascular deaths
  are sudden deaths
• Known risk factors for sudden death
• Cardiac ischemia (esp recent)
• Poor ejection fraction fraction
• Ventricular arrhythmias

8
(No Transcript)
9
(No Transcript)
10
Rigorous testing of new anti-arrhythmics
11
Antiarrthymics in the late 80s

• Antiarrthymics for PVCs V-Tach widely
  prescribed
• Between ¼ - ½ million patients/yr were taking
• New antiarrthymic drugs more effective in
  suppression arrthymias
• Scientific EPS selection of best therapies
• Encainide Flecainide Moricizine
• Placebo controlled trials considered unethical
• Even tho no trails showing reduction of
  arrhythmias led to reduction of sudden death

12
Cardiac Arrhythmia Suppression Trial

• CAST-ing doubt on the dominant pradigms

13
High Risk Patients Require Treatment

• Unethical not to treat such high risk patients
• Have to do something
• How would you feel if one of these patients you
  ignored suffered sudden death
• Plus legal ramifications

14
Cardiac Arrhythmia Suppression Trial

• CAST-ing doubt on the dominant pradigms

15
CAST trial 1987

• Randomized, Double-Blind, Placebo Control
• 27 clinical centers, 4,400 patients
• Open-label titration to select drug-responsive
  patients
• 1,727 randomized to Encainide, Flecainide or
  Moricizine
• Trial is discontinued early
• Encainide and Flecainide 1989
• Moricizine 1991

http//clinicaltrials.gov/ct/show/NCT00000526
16
CAST Results
17
Cardiac Arrhythmia Suppression Trial

• CAST-ing doubt on the dominant pradigms

18
Surrogate Endpoints

• Change in a clinical variablenot experienced
  directly by the patient
• Blood pressure
• Serum cholesterol
• Serum Glucose
• PVCs
• Not itself a direct measure of clinical harm or
  benefit
• Patient does not necessarily feel better or worse
  

19
Clinically Relevant Endpoints

• Mortality or survival benefit
• Clinically important change experienced directly
  by the patient
• Reduced pain
• Improved functional status
• Improved quality of life
• Directly measures clinical benefit or harm

20
Hierarchy of endpoints

• Level 1 True clinical-efficacy measure
• Level 2 Validated surrogate endpoint
• Level 3 Non-validated surrogate endpoints
  reasonably likely to predict clinical benefit
• Level 4 A correlate that measures biological
  activity but whose clinical relevance is not well
  established

Fleming, T. Surrogate endpoints and FDAs
accelerated approval process the challenges are
greater than they seem. Health Affairs (2005)
24(1) 67-78
21
Why Use Surrogate Endpoints?

• Reduction in sample size, duration of trial and
  cost
• Easier to show benefits weeks to mos vs. years
• Assess benefits of drug where measurement of
  clinical outcomes would be unethical/invasive
• Because death and other harder outcomes are
  uncommon or delayed well into future

22
Surrogate Reductio ad absurdum

• Elevated WBC count surrogate for severity of
  pneumonia
• So, why not give cytotoxic agents to reduce the
  white count?!

23
Limitations of Surrogate Endpoints

• Surrogates may not be valid predictors of actual
  clinical outcomes
• Rests upon physiologic assumptions
• Persuasiveness of biologic plausibility
• May be statistically significant but not
  clinically significant
• Provide only partial picture of totality of
  drugs effect
• Lend themselves to manipulation by PhARMA
• Many turn out to be misleading red herrings

24
Reasons for failure of surrogate end points
Surrogate is not involved in disease pathway
Prostate Biopsy Finasteride Trials
Disease has multiple pathways and intervention
effects only one pathway mediated through
surrogate
Ventricular Arrythmias Encainide and Flecainide
Trials
Surrogate is not affected by/ insensitive to
interventions effect
CD4 levels HIV drugs
Intervention has mechanisms of action independent
of disease process
Ventricular Arrythmias Encainide and Flecainide
Trials
Fleming, T. R. et. al. Ann Intern Med
1996125605-613
25
Effective antimicrobial treatment or Useless
surrogate?

• 4,000 surgical patients
• Treated patients for nasal staph carriage
• Treated pts 4.6 vs. 21.3 control (plt.001)
• No difference in surgical infections 2.3
  in treated vs. 2.4 in controls

Perl NEJM 2003
26
Zidovudine -- AZT

• Thymidine Analogue
• Synthesized 1964
• Drug in the public domain
• 1984 scientists approach drug companies to
  expedite RD
• Got Burroughs Wellcome to patent (w/ difficulty)
  

Mitsuya, H. et al., 3'-Azido-3'-deoxythymidine
(BW A509U) An antiviral agent that inhibits the
infectivity and cytopathic effect of human
T-lymphotropic virus type III/lymphadenopathy
associated virus in vitro, Proc. Natl. Acad. Sci
USA (82) 1985
27
Pathophysiology of HIV
AZT
http//research.bidmc.harvard.edu/vptutorials/HIV/
home.htm
28
The Evidence it Works
Total viable cells x 10 -5
HIV positive T cells

HIV negative T cells
AZT uM
Mitsuya, H. et al., 3'-Azido-3'-deoxythymidine
(BW A509U) An antiviral agent that inhibits the
infectivity and cytopathic effect of human
T-lymphotropic virus type III/lymphadenopathy
associated virus in vitro, Proc. Natl. Acad. Sci
USA (82) 1985
29
AZT surrogate endpoints the context

• No treatments for AIDS ? desperation
• Bribery to enter trials
• Difficulties of obtaining accurate results
• Drug smuggling rings
• Cook County 100s of deaths per quarter
• Regan administration largely ignores HIV

30
AZT- The Hopes

• New public private partnership
• Burroughs Wellcome submits AZT application in 3
  stages over 7 months
• Drug companies, scientists and regulatory agency
  working together for good of the public
• Sets stage for accelerated approval of other
  drugs

31
AZT The Doubts

• The myth of public health altruism
• Burroughs Wellcome charges 10,000/year
• Largely assumes credit for innovation
• Despite doing relatively little of R D
• Clinical trials on limited population
• White, gay males
• Difficult to adhere to regimen
• Initially 6x/day
• Controversies and questions about role of the
  drug

32
1994 CONCORDE Study
33
(No Transcript)
34
Kaplan-Meier Plot for all causes of death
35

Kaplan-Meier Plot for time to AIDS or Death
36
Kaplan-Meier for time to ARC AIDS or death
37
Kaplan-Meier for time to reduction in CD4 count
less than half of baseline, AIDS or death
38
CONCORDE Conclusions
The small but highly significant and persistent
difference in CD4 count between the groups was
not translated into a significant clinical
benefit. Thus, analyses of the time until certain
concentrations of CD4 were reached revealed
significantly shorter times in the Deferred AZT
treatment group. Had such analyses been regarded
as fundamental, the trial might have been stopped
early with a false-positive result.
39
CONCORDE Conclusions
The small but highly significant and persistent
difference in CD4 count between the groups was
not translated into a significant clinical
benefit. Thus, analyses of the time until certain
concentrations of CD4 were reached revealed
significantly shorter times in the Deferred AZT
treatment group. Had such analyses been regarded
as fundamental, the trial might have been stopped
early with a false-positive result.
40
CONCORDE Conclusions II
This discrepancy in the differences between
Immediate and Deferred AZT treatment groups in
terms of changes in CD4 count and of long-term
clinical response casts doubt on the uncritical
use of CD4 counts as "surrogate endpoints" in
trials, although their value as a prognostic
marker for disease progression in cohorts and
trials is beyond dispute. The reason for this
discrepancy is unclear.
41
Conclusions Continued
This discrepancy in the differences between
Immediate and Deferred AZT treatment groups in
terms of changes in CD4 count and of long-term
clinical response casts doubt on the uncritical
use of CD4 counts as "surrogate endpoints" in
trials, although their value as a prognostic
marker for disease progression in cohorts and
trials is beyond dispute. The reason for this
discrepancy is unclear.
42
(No Transcript)
43
(No Transcript)
44
History of Erythropoietin

• U of C scientist Eugene Goldwasser starts
  research on erythropoietin 1960s
• Purified in 1970s

45
Erythropoietin and Biotech Revolution

• Amgen works with Goldwasser to sequence and clone
  erythropoietin
• 1985 Amgen files patent
• Orphan Drug
• 1989 approved for marketing
• Decade later 5b in profits

46
Erythropoietin Bloods Life-Blood
47
(No Transcript)
48
Policy Context Not that simple

• Natl Kidney Foundation guidelines
• Progressively raised Hb level w/out clinical
  evidence of benefit
• Strong conflicts of interest
• 10/18 panel members w/significant financial
  interest
• 57 of NKF funding from industry
• Medicare reimbursements ? source of profit for
  dialysis centers

49
CHOIR and CREATE Studies

• Researchers in US and France simultaneously
  conduct erythropoietin studies w/ clinical
  endpoints
• Funded by Amgens competitors
• Roche Johnson Johnson
• To get their me-too drugs to market

50
CHOIR Enrollment and Outcomes
Singh A et al. N Engl J Med 20063552085-2098
51
Mean Monthly Hemoglobin Levels
Singh A et al. N Engl J Med 20063552085-2098
52
Kaplan-Meier Estimates of the Probability of the
Primary Composite End Point
Singh A et al. N Engl J Med 20063552085-2098
53
Kaplan-Meier Estimates of Secondary Endpoint of
Death
Singh A et al. N Engl J Med 20063552085-2098
54
CREATE Enrollment, Randomization, and Study
Completion
Drueke T et al. N Engl J Med 20063552071-2084
55
Median Hemoglobin Levels in the
Intention-to-Treat Population during the Study
Drueke T et al. N Engl J Med 20063552071-2084
56
Changes from Baseline to Year 1 in SF-36
Quality-of-Life Scores
Drueke T et al. N Engl J Med 20063552071-2084
57
Changes from baseline to Year 1 in Time to
Dialysis during the Study
Lower Hb Group remains off dialysis longer
Drueke T et al. N Engl J Med 20063552071-2084
58
Current FDA issues with Regulation of Surrogate
Endpoints

• Accelerated approval (1992)
• Formal acceptance of surrogates
• Endpoints reasonably likely to predict clinical
  benefit
• Early marketing approval contingent upon
  post-marketing studies confirming clinical
  benefit
• Patient-Centered Outcomes
• Lancet commentary, controversies

59
(No Transcript)
60
Pressures for Accelerated Approval (AA)

• 1962 Kefauver-Harris Amendments require
  demonstration of efficacy
• Alleged drug lag
• Drug industry and free market economists want
  less regulation
• Patient groups demand more available treatments
  for AIDS and cancer

61
ACT UP Demonstrations
http//aidshistory.nih.gov/search_for_treatments/d
emonstration.html
http//www.actupny.org/documents/FDAhandbook1.html
1990 NIH Headquarters
1988 FDA Headquarters
62
Accelerated Approval --Issues

• More lenient criteria-slipping into surrogates
• Lack of urgency to complete post-marketing study
  commitments
• No teeth
• Unwillingness and lack of power to pull drugs off
  the market

63
Oncology Drug Advisory Committee Report

• Review 8 products approved under AA in first five
  years AA available for oncology rx (3/03)
• Average time for completion of post-approval
  studies 10 years
• Marketing of interventions shown to have little
  benefit, continued

64
Accelerated Approval, Animal Efficacy Rule and
Pediatric Research Equity Act
65
Open Post-Marketing Commitments
Concluded
13
Submitted
Behind Schedule
10
13
64
On schedule
66

• How to handle cases where validation studies
  dont conclusively support the drug?
• What to do if no tangible benefit in face of well
  documented toxicities or safety risks

67

• CAST-ing about for better pradigms

68
Patient-Reported Outcomes (PRO)

• Attempt to weigh patient-centered quality of
  life outcomes
• Step forward care about more than just death
• Want to value and measure quality of life (QOL)
• Respect and weight for how patients are feeling
• True valued outcomes .or back door
  surrogates?
• Susceptible to manipulation
• How to achieve validated measures/scales?

Revicki, Lancet 2/17/2007
69
(No Transcript)
70
Policy analysis found this study related
promotional activities resulted in ? increase
sales worth 375-450 million
71
Take Home Points

• Misconception that if an outcome is a correlate
  it is a valid surrogate end point
• A correlate does not a surrogate make
• Lowering risk marker ? masking or killing the
  messenger
• Multiple other unintended effects to drugs
  besides putative neat mechanisms of action
• Even best surrogates, risk misleading in various
  ways

72
2 Criteria for Valid Surrogate

• Biologic marker must be correlated with the
  clinical endpoint
• Marker must fully capture the net effect of the
  intervention on the clinical-efficacy endpoint

Fleming, Health Affairs 2005
73
(No Transcript)
74
Searching for Surrogates in JAMA Ads

• How many drug ads (denominator)?
• How many ads have efficacy outcome data of any
  sort (numerator 1)?
• How many of the outcomes are actual clinical
  outcomes how many are surrogates?
• Pick out 1 or 2 and analyze them
• As a pharmaceutical researcher design surrogate
  tools and studies to show how well the drug
  works
• As an FDA reviewer critique/discuss real or
  implied surrogate issues related to the drug and
  the ad .