Clinical Trials - how often do they allow crossovers?

[davebliz]

Clinical Trial Crossovers. Do you know if in blind trials, if the people getting the new drug get better results how often do they then allow the people that were getting the placebo to start getting the new drug? Thanks.

[Boywonder56]

Im in the titan trial...for erleada....they unblinded the tria a year early because of good results..i found out i had been on the erleada not placebo...am now in open label ...thankful as the 16,400.oo per month would be prohibitive...you can google all this under titan trial..bw

[Tall_Allen]

Some trials include cross-overs, some don't. They set it up that way in advance if they do. That's different from ending the blinding early (which some trials allow for - if they do, they set a recruitment/ follow-up target for that) and allowing everyone to have the study drug.

[Concerned-wife in reply to Tall_Allen]

Have you a rebuttal to Vinayak K. Prasad’s critique of the control arm in the Vision trial?

[Tall_Allen in reply to Concerned-wife]

I don't know who he is. What is his critique?

[Concerned-wife in reply to Tall_Allen]

He has concerns about the control group. m.youtube.com/watch?v=pAsPz...(Not working as expected?)

[Tall_Allen in reply to Concerned-wife]

I'm sorry- I don't get my info from youtube videos and can't take the time to watch it. If you want my comments, can you please summarize his points.

[Concerned-wife in reply to Tall_Allen]

Sorry. I thought you might have known him since he is an oncologist and prof at UC SF. I believe his main point was the trial had an extremely poor control arm because ,although the authors say they allowed Standard of Care, they didn’t allow the control group , for example, to receive Cabazitaxel. They thus had one of the largest drop out rates from the control group of cancer trials. He doesn’t doubt that Lutetium will have an impact. Just that the weakness of the control group was a huge weakness in the study.

[Tall_Allen in reply to Concerned-wife]

I think he is mistaken.

All patients in the VISION trial had to have had one and up to two regimens of chemo just to qualify. Patients who had received only one taxane were ineligible if they were deemed at baseline to be candidates for receiving a second taxane. While they were in the trial neither group (treatment or control) was allowed to have chemo. "These constraints were used because of a lack of safety data on combining the investigational drug with these agents... Patients whose condition was deemed to be appropriate for additional chemotherapy could discontinue the trial treatment and receive chemotherapy at the discretion of their physician. "

nejm.org/doi/full/10.1056/n...

So it was anticipated from the start that patients would end their involvement in order to receive chemo, Xofigo, or other disallowed therapies. 38% had received cabazitaxel before entering the trial (treatment or control). An additional 19% in the control group and 13% in the treatment group received post-protocol cabazitaxel.

Initially, there were a lot of drop-outs from the control group because they were disappointed that they were not randomized to treatment:

"After the trial started (May 29, 2018), a high incidence of withdrawal from the trial was noted in the control group at certain sites and was attributed principally to patient disappointment (see the Supplementary Methods section). After discussion with regulatory authorities, we implemented enhanced trial-site education measures on March 5, 2019 to reduce the incidence of withdrawal. The high incidence of withdrawal could have affected the interpretability of radiographic end points. Therefore, the primary analysis of imaging-based progression-free survival and the analyses of key secondary end points were amended to include only the patients who had undergone randomization on or after March 5, 2019. "

So they increased the sample size and only analyzed the results among patients in the study (treatment and control) after the new measures were in place.

[MateoBeach]

There is something that seems very cruel in many of the trials for advanced PC where median overall survival (OS which is the ultimate test) is the primary outcome. That means that half of all those in the placebo must die. And then start the clock to see how much longer to reach half of all those treated lived. Or else a significant statistic of survival “not reached”. They need to show this, added months of OS, in order to win FDA approval.Also why they often choose to test in only the most advanced, such as mCRPC after prior treatments. Then they won’t have to wait too long for half to die. Perhaps it is the only way to be perfectly certain. Yet it makes me sad.

The drugs enzalutamide, apalutamide, and darolutamide are mostly so similar in action. Why did they need separate placebo controlled trials for OS, rather than just comparing them to each other in a randomized comparison trial????

Crossover standards need to be specifically defined and included in trials. Not ethical otherwise IMO.

[Tall_Allen in reply to MateoBeach]

They only have to wait for half of the control group to die, not the treatment group. "Not reached" means the median or the 95% confidence interval was not reached in the time that the median control group mortality was reached.

They test the most advanced cases first because (1) they need it the most (2) have the least to lose if it is unsafe/ineffective (3) it takes the least time to prove, making the drug available off label.

While you think those drugs are similar (all that they have in common are that they are antiandrogens), there are several antiandrogens that have failed in Phase 3 trials and even more that never made it to Phase 3.

But the FDA is open to proof of effectiveness by metastasis-free survival (MFS). All of the drugs approved for non-metastatic CRPC were approved on that basis. It is usually a good surrogate endpoint for OS, and the median occurs sooner.