Clinical trials in crisis Clinical Trial Commentary Dr Eric Topol Chairman and Professor, Department of Cardiology Director of the Joseph J Jacobs Center.

Clinical trials in crisis

Clinical Trial Commentary

Dr Eric TopolChairman and Professor, Department of CardiologyDirector of the Joseph J Jacobs Center for Thrombosis and Vascular Biology at the Cleveland Clinic

Dr Robert CaliffProfessor of Cardiology Associate Vice Chancellor for Clinical Research at Duke University

Rigging trials for marketingClinical trials in crisis

New York Times expose on underpowered trials made the argument that the pharmaceutical industry has grown sophisticated enough to deliberately engineer results of trials for greater marketing effect.

There may be some truth to those concerns. -Topol

Stolberg SG and Gerth J. Drug Makers Design Studies With Eye to Competitive Edge.

New York Times 2000 Dec 23:A1.

“The principle here that you’re alluding to is that finding a p-value of greater than .05 - or in other words, failing to find evidence of a difference - is not the same as finding evidence of no difference.”

Dr Robert CaliffProfessor of Cardiology Associate Vice Chancellor for Clinical Research at Duke University

Clinical trials in crisis

Evidence of difference

PowerClinical trials in crisis

It is not the number of patients you enroll in a trial that determines power, it is the number of events.

A trial of 1 million healthy college students for 5 years will still be underpowered for CV events.

-Califf

TARGET

4812 patients from 18 countries who were undergoing coronary stenting.

Primary endpoint: 30-day outcomes of death, MI, and urgent revascularization.

Tirofiban Abciximab OR p-value

7.55% 6.01% 1.26 0.037

OR = odds ratio

Clinical trials in crisis

PROVE-IT

PRavastatin Or atorVastatin Evaluation and Infection Therapy

Statins require a lag phase of 1 year to 18 months to show benefit against placebo.

PROVE-IT designed to last 2 years, with active control, and so wouldn’t be expected to show benefit.

-Topol

Clinical trials in crisis

Evidence-based marketing

Positioning a drug based on clinical trials is fine as long as the trials are well done.

Better to have marketing based on trials than on belief, testimony, or pathophysiological construct.

-Califf

There are dangers of marketing with pseudo-evidence.

-Topol

Clinical trials in crisis

Constructing an equivalence trial

For true equivalence you need hundreds of thousands of patients. The best you can do is declare “non-inferiority”.

-Topol

The best way to determine the boundary is to determine with experts and patients how much of a difference do you need in clinical outcomes to convince you to switch from one treatment to the other.

-Califf

Clinical trials in crisis

MIRACLPrimary endpoint: death, MI, and refractory recurrent ischemia requiring hospitalization.

3000 patients following an acute coronary event were randomized to either 80 mg atorvastatin or placebo for a period of 3 months.

Primary endpoint results: 14.8% with atorvastatin vs 17.4% on placebo (16% reduction, p=0.048).

Reduction was primarily due to effect of atorvastatin on recurrent symptomatic myocardial ischemia (26% reduction, p=0.02).

Clinical trials in crisis

PROVE-IT vs MIRACL

PROVE-IT active controlpravastatin vs atorvastatin

MIRACLplacebo controlatorvastatin vs placebo

Clinical trials in crisis

Sample sizeMethod A: $/cost per patient = # of patients enrolled

Then you figure out “how can we justify the sample size we can afford?”

Used more often than people like to admit.

Method B: “the right way”

Decide the minimal difference that would change clinical practice and size your sample to detect that difference in an event-driven clinical trial.

-Califf

Clinical trials in crisis

Deliberate underpowering?

The worst case scenario:

Deliberately designing a trial to have insufficient events in order to declare non-inferiority at the end of the trial.

-Topol

Clinical trials in crisis

GUSTO-III

Reteplase vs tPA

Had a limited number of patients in order to cut costs, which resulted in extreme difficulty in drawing any clinically useful conclusions.

Not deliberately done, but nonetheless caused problems with interpretation.

-Topol

Clinical trials in crisis

Some data better than none?

If you can generate some randomized data, but not enough to generate the definitive answer, is it better to do the non-definitive trial?

The danger is when a trial is non-definitive, but an effort is made to turn it into a definitive non-inferiority trial.

-Califf

Clinical trials in crisis

“I think that pharmaceutical companies have a lot of really bright folks working for them who have figured it out and can help to engineer the results of a trial … to be able to make the claim about non-inferiority, at least to the medical community. So we’ve gotta keep an eye on them for that.”

Dr Eric TopolChairman and Professor, Department of CardiologyDirector of the Joseph J Jacobs Center for Thrombosis and

Vascular Biology at the Cleveland Clinic

Clinical trials in crisis

Staying alert

ASSENT-3

ASsessment of the Safety and Efficacy of New Thrombolytic Regimens - 3

6,000 patients with AMI presenting within 6 hours of symptom onset randomized into 3 groups:

(Group A) – full dose tenecteplase with heparin sodium

(Group B) – full dose tenecteplase with enoxaparin sodium

(Group C) – half dose tenecteplase with abciximab and heparin sodium

combined clinical endpoints: death, AMI, refractory ischemia, intracranial hemorrhage and major bleeding

Clinical trials in crisis

Follow-up trials

There is the danger that the definitive follow-up trials will not be done despite promises to the contrary.

Will there be a follow-up to ASSENT-3?

Will there be a follow-up to ESPRIT?

-Topol

Clinical trials in crisis

Power of capitalism

Head-to-head trials will happen because the market demands it.

Market is set by what physicians demand in terms of this practice.

-Califf

Clinical trials in crisis

Clinical trials in crisis

Who will do active control trials?

These trials cost a great deal of money.

The trials are either not undertaken as they ought to be, or they are undertaken with insufficient amount of study.

-Topol

Clinical trials in crisis

PROVE-IT redux

There are lots of reasons to do a clinical trial comparing atorvastatin vs other statins:differences in effects on HDL

fibrinogininflammation

Comparing a potent LDL lowering statin vs one with other effects could be very important if it was done properly – with longer duration and greater numbers than PROVE-IT.

-Topol

“[Bristol-Myers Squibb] are a company that put everything on the line and did definitive mortality trials and the response of the clinical community has been to prescribe [pravastatin] at about one seventh or one eighth the rate as atorvastatin which has not had any definitive clinical outcome trial.”

Dr Robert CaliffProfessor of Cardiology Associate Vice Chancellor for Clinical Research at Duke University

Clinical trials in crisis

Doing the right thing

“The medical community has been seduced by the ease of lowering LDL. But it may not even have much to do with the LDL per se as to the clinical benefits. That’s why we need a definitive trial in this area and that’s why we don’t have an adequate governmental or honest broker support.”

Dr Eric TopolChairman and Professor, Department of CardiologyDirector of the Joseph J Jacobs Center for Thrombosis and

Vascular Biology at the Cleveland Clinic

Clinical trials in crisis

Honest brokers

Clinical trials in crisis

Critical trials

Certain trials are crucial but don’t get done:

Drugs with extensive track records put against newer drugs with lots of uncertainty.

Drugs that are commonly prescribed and worth vast amounts to the pharmaceutical industry, but there remain large gaps in our knowledge.

-Califf

Clinical trials in crisis

How do you solve the problem?

A government agency to force definitive trials?

Does the market work through physicians?

Or does the market control physicians?

Where these uncertainties exist, will any company give an objective answer?

Clinical trials in crisis

Comparing treatments

“[These trials] are so easy to do. […] We could get the answers to these questions in no time, don’t you agree?

-Topol

“It’s not rocket science. It’s things that doctors do every day and it is just randomly allocating whether you do A or B where you don’t know which is better.”

-Califf

Clinical trials in crisis

Consumer reports

An ideal world would have a system where patients and doctors would set the top 10 questions to be answered.

Need to reward the R&D dollars for the best quality, not the best marketing.

There’s a reason we have Consumer Reports for cars: you don’t just trust the

salesman.-Califf