Design Issues in ABS Trials: Surrogates Endpoints & Non-Inferiority Trials October 29, 2003

Design Issues in ABS Trials: Surrogates Endpoints &

Non-Inferiority Trials

October 29, 2003

Thomas R. Fleming, Ph.D.Professor and Chair of Biostatistics

University of Washington

FDA Anti-Infective Drugs Advisory Committee

Design Issues in ABS Trials

• Criteria for Study Endpoints

• Use of Surrogate Endpoints

• Non-Inferiority Design Issues

• Choice of the NI Margin• Bio-creep with Repeated NI Trials

• Active vs. Placebo controlled trials

• Time to Event Analyses in ABS Trials

Design Issues in ABS Trials

• Criteria for Study Endpoints

• Use of Surrogate Endpoints

• Non-Inferiority Design Issues

• Choice of the NI Margin• Bio-creep with Repeated NI Trials

• Active vs. Placebo controlled trials

• Time to Event Analyses in ABS Trials

Criteria for Study Endpointsin ABS Clinical Trials

• Measurable/Interpretable

• Sensitive

• Clinically relevant ~ Resolution/Improvement of ABS Symptoms ~ Reducing the time to Resolution

Design Issues in ABS Trials

• Criteria for Study Endpoints

• Use of Surrogate Endpoints

• Non-Inferiority Design Issues

• Choice of the NI Margin• Bio-creep with Repeated NI Trials

• Active vs. Placebo controlled trials

• Time to Event Analyses in ABS Trials

Use of Surrogate Endpoints

Treatment Effects on Surrogate Endpoints eg: ~ Radiological Resolution of Disease ~ Microbiological Outcomes

• Establishes Biological Activity

• But Not Necessarily Clinical Efficacy

Surrogate True ClinicalEndpoint OutcomeDisease

Intervention

Time

Illustration: AIDS Patients with MAI Bacteremia

Clarithromycin Dose (mg bid) 500 1000 2000

Bacterial Load 145 34 25

12 wk Mortality 5.7% 25.5% 28.0%

Chaisson et al, 1994

Bacterial 12 week Load Mortality Disease

Intervention

Time

Validation of Surrogate Endpoints

Property of a Valid Surrogate

Effect of the Intervention on the Clinical Endpoint

is reliably predicted by the

Effect of the Intervention on the Surrogate Endpoint

Prentice’s Sufficient Conditions

1. The surrogate endpointmust be correlated with

the clinical outcome

2. The surrogate endpointmust fully capture

the net effect of treatmenton the clinical outcome

Validation of Surrogate Endpoints

Statistical Meta-analyses of clinical trials data

Clinical Comprehensive understanding of the

~ Causal pathways of the disease process ~ Intervention’s intended and unintended

mechanisms of action

Clinical Endpoints & Surrogates in ABS Clinical Trials

• Clinical Endpoints:

~ Resolution/Improvement of ABS Symptoms

• Surrogate Endpoints:

~ Radiological Resolution of Disease

~ Microbiological Outcomes

Validation of Surrogate Endpoints requires

more than correlation with Clinical Endpoints

Design Issues in ABS Trials

• Criteria for Study Endpoints

• Use of Surrogate Endpoints

• Non-Inferiority Design Issues

• Choice of the NI Margin• Bio-creep with Repeated NI Trials

• Active vs. Placebo controlled trials

• Time to Event Analyses in ABS Trials

Dual Goals of Non-Inferiority Trials

• To enable a direct evaluation of the clinical efficacy

of EXP relative to Active Control

• To contribute evidence to the evaluation of efficacy

of EXP relative to Placebo

Non-Inferiority Trials… Some Requirements

• Active Control Effect

ICH E9: “A suitable active comparator… could be a widely used therapy

whose efficacy in the relevant indicationhas been clearly established & quantified

in well-designed & well documented superiority trials

& which can be reliably expected to have similar efficacy in the contemplated AC trial.”

Non-Inferiority Trials… Some Requirements

STD should have clinical efficacy

• that is of substantial magnitude

• that is precisely estimated

• with estimates that are relevant to the setting in which the non-inferiority trial

is being conducted

Factors Influencing Choice of Margin

PLA – AC Cure Rate

• Active Control Effect

~ magnitude of Active Control effect Eg: = 45% – 80% = –35%

~ precision of estimate Eg: 2 s.e. = 10% ( 175/arm )

~ estimates relevant to setting of NI trial • Population • Supportive care • Endpoint assessment

*

( ) –45% –35% –25% –12.5% 0%

Factors Influencing the Choice of Marginand Interpretation of NI Trial Results

• Clinical Relevance of Changes in: Benefits, Risks/Tolerance,

Convenience, Resistance, etc.

• Active Control Effect

Factors Influencing Choice of Margin

• Clinical Relevance of Changes in: Benefits, Risks/Tolerance,

Convenience, Resistance, etc.

Clinical importance of:

- reduction in efficacy by - altered safety/tolerance profile

- altered convenience of administration

- altered resistance or drug/drug interactions

ICH E10: “The determination of the margin in a non-inferiority trial is based on

both statistical reasoning & clinical judgment,

and should reflect uncertainties in the evidence on which the choice is based,

and should be suitably conservative.”

The Choice of the Margin in an NI Trial

“Bio-creep” with Repeated NI Trials

Eg: Anti-viral Drugs Advisory Comm (10/4/01)

Empiric Anti-fungal therapy of febrile neutropenic patients

• Amphotericin B Deoxycholate

• Ambisome vs Amphotericin B49.9% v 49.1% Mycosis Study Gp #32

• Voriconazole vs Ambisome23.7% v 30.1% 95% CI: (– 12, – 0.1)

Overview of Placebo-controlled ABS Trials

• Fourteen Placebo-controlled Trials of Antimicrobials conducted in ’69 –’03

(with nine conducted since ’96)

• Outcome: Antimicrobial effect onResolution or improvement of ABS symptoms, assessed at fixed time between 7-14 days

-70

-50

-30

-10

10

30

50

70A

nti

mic

rob

ial

Eff

icac

y (

95%

CI)

11

33

44

22

55 6677

88 99

1010

1111

1212 1133

1144

1155

1166

StudyStudy

• ICH E10: “The determination of the margin in a non-inferiority trial is based on

both statistical reasoning & clinical judgment,

should reflect uncertainties in the evidence on which the choice is based,

and should be suitably conservative.”

• When one cannot justify a non-trivial margin, placebo controlled trials provide an ethically and scientifically reliable approach

to assessing the benefit-to-risk profile

NI Trials vs Placebo Controlled Superiority Trials

Design Issues in ABS Trials

• Criteria for Study Endpoints

• Use of Surrogate Endpoints

• Non-Inferiority Design Issues

• Choice of the NI Margin• Bio-creep with Repeated NI Trials

• Active vs Placebo controlled trials

• Time to Event Analyses in ABS Trials

Time to Event Analyses in ABS Trials

• “In self-resolving diseases, it may be more appropriate to measure time to resolution or improvement of symptoms”

• To have 90% power to detect a reduction in time to resolution from 7 days to 5 days (with the standard 2.5% false positive error rate) one needs approximately 200 pts/arm.

(Stat Significance if one obtains about 1.3 days )

Conclusions

• Criteria for Study Endpoints

• A Correlate does not a Surrogate Make

• Non-Inferiority Design Issues

• Choice of the NI Margin• Bio-creep with Repeated NI Trials

• Placebo controlled trials

• Time to Event Analyses in ABS Trials