0. Composite Endpoints In Clinical Trials
Ajay J. Kirtane, MD, SM
Center for Interventional Vascular Therapy
Columbia University Medical Center /
New York Presbyterian Hospital

1. Ajay J. Kirtane, MD
I have no real or apparent conflicts of interest to report.

2. Conflict of Interest Disclosure
Ajay J. Kirtane
None
Off-label use will be discussed

3. What is a Composite Endpoint?
Endpoint that combines several outcomes
Sub-components
Components are typically directionally related (e.g. death/MI)
But may not be “equal” in severity and definitions can vary across trials
May be related to the final outcome (death) but mechanisms can be different
Death/MI/bleeding
Death/MI/TVR

4. Composite Endpoints We need them
Individual outcomes lack statistical power
But this can be abused
We sometimes dislike them
Components vary in their clinical importance
Treatment effect varies across components
May actually lose power by using a composite endpoint!!!

5. Composite Outcomes in Published CV Trials
304 trials in 14 journals in
73% had composite as primary endpoint, median 3 components
death 98%
myocardial infarction 92%
reintervention 54%
stroke 32%
angina 10%
hospitalization 12%
cardiac failure 9%
Lim et al Ann Int Med 2008

6. Composite Endpoints: Take Care RITA 3 Trial
Patients
895
915
Deaths 26 23
After 4 months MIs 30 34
Refractory angina 39 85
Death, MI or refractory angina (primary endpoint)
86 (9.6%)
133 (14.5%)
Intervention vs Conservative
Overall p=0.001
Fox K et al. Lancet 2002; 360(9335):

7. TYPHOON trial DES vs. BMS in primary PCI
primary endpoint: cardiac death, MI, TVR by 1 year
sirolimus control
(N=355) (N=377)
primary P=.004
cardiac death
myocardial infarction
TVR
c/o S. Pocock

8. 2 Primary Stent Endpoints (at 12 Months)
1) Ischemia-driven TLR*
and
2) Composite Safety MACE = All cause death, reinfarction, stent thrombosis (ARC definite or probable)**, or stroke
Major Secondary Endpoint (at 13 Months)
Binary angiographic restenosis
* Related to randomized stent lesions (whether study or non study stents were implanted);
** In randomized stent lesions with ≥1 stent implanted (whether study or non study stents)

9. Primary Efficacy Endpoint: Ischemic TLR10
Diff [95%CI] =
-3.0% [-5.1, -0.9]
HR [95%CI] =
0.59 [0.43, 0.83]
P=0.002
TAXUS DES (n=2257) 9
EXPRESS BMS (n=749) 8
7.5% 7 6
Ischemic TLR (%) 5
4.5% 4 3 2 1 1 2 3 4 5 6 7 8 9 10 11 12
Time in Months
Number at risk
TAXUS DES
2257
2132
2098
2069
1868
EXPRESS BMS
749
697
675
658
603

10. Primary Safety Endpoint: Safety MACE*10
TAXUS DES (n=2257) 9
EXPRESS BMS (n=749)
8.1% 8
8.0% 7 6
Diff [95%CI] =
0.1% [-2.1, 2.4]
HR [95%CI] =
1.02 [0.76, 1.36]
PNI=0.01
PSup=0.92
Safety MACE (%) 5 4 3 2 1 1 2 3 4 5 6 7 8 9 10 11 12
Time in Months
Number at risk
TAXUS DES
2257
2115
2086
2057
1856
EXPRESS BMS
749
697
683
672
619
* Safety MACE = death, reinfarction, stroke, or stent thrombosis

11. HORIZONS-AMI trial
primary efficacy: target lesion revascularization at 1 year
composite safety: death, reinfarction, stroke, stent thrombosis
TAXUS bare-metal
stent stent
(N=2257) (N=749)
TLR % % P=.002
composite safety % % P=.92
separate re-intervention from major clinical concerns
non-inferiority re safety, components “equally flat”

12. SYNTAX: PCI vs. CABG “SYNTAX fails to show non-inferiority for DES”
1800 patients with left main/3 vessel disease
Primary Endpoint of MACCE:
Composite of death, stroke, MI
repeat revascularisation

13. Cumulative Event Rate (%) Months Since Allocation
MACCE to 12 Months
CABG (N=897)
TAXUS (N=903) 20
P=0.0015*
17.8%
12.1% 10
Cumulative Event Rate (%) 6 12
Months Since Allocation
ITT population
Event Rate ± 1.5 SE. * Fisher’s Exact Test 13

14. Repeat Revascularization to 12 Months
CABG (N=897)
TAXUS (N=903)
P<0.0001* 20
13.7% 10
Cumulative Event Rate (%)
Exhibit 38
5.9% 6 12
Months Since Allocation
ITT population
Event Rate ± 1.5 SE. * Fisher’s Exact Test 14

15. All-Cause Death/CVA/MI to 12 Months
CABG (N=897)
TAXUS (N=903) 20
P=0.98* 10
Cumulative Event Rate (%)
7.7%
7.6% 6 12
Months Since Allocation
ITT population
Event Rate ± 1.5 SE. * Fisher’s Exact Test 15

16. All-Cause Death to 12 Months
CABG (N=897)
TAXUS (N=903) 20
P=0.37* 10
Cumulative Event Rate (%)
4.3%
3.5%
Exhibit 33 6 12
Months Since Allocation
ITT population
Event Rate ± 1.5 SE. * Fisher’s Exact Test 16

17. Cumulative Event Rate (%) Months Since Allocation
CVA to 12 Months
CABG (N=897)
TAXUS (N=903) 20
P=0.003* 10
Cumulative Event Rate (%)
Exhibit 36
2.2%
0.6% 6 12
Months Since Allocation
ITT population
Event Rate ± 1.5 SE. * Fisher’s Exact Test 17

18. Myocardial Infarction to 12 Months
CABG (N=897)
TAXUS (N=903) 20
P=0.11* 10
Cumulative Event Rate (%)
4.8%
3.2%
Exhibit 37 6 12
Months Since Allocation
ITT population
Event Rate ± 1.5 SE. * Fisher’s Exact Test 18

19. SYNTAX Summary
Composite MACCE (death/MI/stroke/revasc) driven by greater repeat revascularization alone
Death/MI/Stroke rates virtually identical
Composite death/MI/stroke had offsetting components
Higher MI with PCI
Higher stroke with CABG
What about other differences not captured in the composite?

20. PARTNER Endpoints
PRIMARY: All-cause mortality over the duration of the study
Superiority test (two-sided), 85% power to detect a difference, α = 0.05, sample size = 350 total patients
CO-PRIMARY: Hierarchical composite of all-cause mortality and repeat hospitalization
Non-parametric method described by Finkelstein and Schoenfeld (multiple pair-wise comparisons)
> 95% power to detect a difference, α = 0.05
This is the Bulleted List slide.
To create this particular slide, click the NEW SLIDE button on your toolbar and choose the BULLETED LIST format. (Top row, second from left)
The Sub-Heading and footnote will not appear when you insert a new slide. If you need either one, copy and paste it from the sample slide.
If you choose not to use a Sub-Heading, let us know when you hand in your presentation for clean-up and we’ll adjust where the bullets begin on your master page.
Also, be sure to insert the presentation title onto the BULLETED LIST MASTER as follows:
Choose View / Master / Slide Master from your menu.
Select the text at the bottom of the slide and type in a short version of your presentation title.
Click the SLIDE VIEW button in the lower left hand part of your screen to return to the slide show. (Small white rectangle) 20

21. Mortality or Repeat Hosp
HR [95% CI] = 0.46 [0.35, 0.59]
P (log rank) <
100
Standard Rx
TAVI 80 60
All-cause mortality or Repeat Hospitalization (%) 40 20
Months
Numbers at Risk
TAVI
179
117
102 56 22
Standard Rx
121 49 23 4

22. Mortality or Repeat Hosp
∆ at 1 yr = 29.1% NNT = 3.4 pts
100
Standard Rx
TAVI
71.6% 80 60
All-cause mortality or Repeat Hospitalization (%) 40
42.5% 20
Months
Numbers at Risk
TAVI
179
117
102 56 22
Standard Rx
121 49 23 4

23. Finklestein & Schoenfeld Analysis (hierarchical multiple pair-wise comparison)
Compare, at random, every TAVI patient with every Standard Rx patient; 179 x 179 (32,041) patient pairs, which did better?
#1, compare “time to death”
72% chance that we know who died first
If so, 63% chance that Standard Rx patient died first and 37% chance that TAVI patient died first
#2, if necessary, compare “time to repeat hospitalization”
17% chance that we know who had repeat hosp first
If so, 75% chance that Standard Rx patient had repeat hosp first and 25% chance that TAVI patient had repeat hosp first
FS Method
Produces a P-value
<
This is the Bulleted List slide.
To create this particular slide, click the NEW SLIDE button on your toolbar and choose the BULLETED LIST format. (Top row, second from left)
The Sub-Heading and footnote will not appear when you insert a new slide. If you need either one, copy and paste it from the sample slide.
If you choose not to use a Sub-Heading, let us know when you hand in your presentation for clean-up and we’ll adjust where the bullets begin on your master page.
Also, be sure to insert the presentation title onto the BULLETED LIST MASTER as follows:
Choose View / Master / Slide Master from your menu.
Select the text at the bottom of the slide and type in a short version of your presentation title.
Click the SLIDE VIEW button in the lower left hand part of your screen to return to the slide show. (Small white rectangle) 23

24. PARTNER: Win Ratio Analysis
Compare every TAVI pt with Standard pt:
Total no. of pairs: 179 x 179 = 32041
Death w TAVI 1st LOSE
Death w standard 1st WIN
Hosp survivor w TAVI 1st LOSE
Hosp survivor w standard 1st WIN
None of the above TIE
Win Ratio = Pairs with TAVI win / Total Number of pairs
Win ratio for composite: 1.87 (95% CI )
Pocock et al Eur Heart J 2011

25. Weighting Components of Composites
Endpoint Weights
Can discount less important outcomes (e.g. a TLR is worth some fraction of a non-fatal NQWMI)
But from whose perspective?
Outside of QOL / Cost-Effectiveness analyses, there is poor guidance on how to weigh endpoints
Issues of interpretability

26. Summary: Composite Endpoints
Advantages
May provide gain in statistical power
Simple summary of several outcomes
Disadvantages
Can be clinically difficult to interpret
May be a mixed bag of “hard” and “softer” outcomes
Combined outcomes of varying importance
Often no clear way to “weigh” these outcomes

27. Summary Composite primary endpoints are of value
When no single component dominates
Statistical power may be increased
Provides a global summary of treatment effect
Composite primary endpoints have problems
What components to include?
Components vary in clinical importance
Treatment effect varies across components
Results often misinterpreted
c/o S. Pocock