Don't jump the gun on this drug yet. More testing is needed to confirm efficacy:
Right. At this point, neither efficacy nor safety are known for hydroxychloroquine in the setting of Covid-19 patients.
Hydroxychloroquine was tested clinically in the past, probably for malaria patients, and maybe for prevention of malaria in healthy patients. So first, clinical testing has to establish what dose of hydroxychloroquine, alone and in combination with azithromycin, is safe and well tolerated in Covid-19 patients. This kind of safety study is called a Phase 1 Clinical Trial. It must determine safe doses for hydroxychloroquine alone and when combined with azithromycin (a commercially availble drug with well known dosing and safety profile). Usually these trials start testing at a low dose, and gradually increase the dose step wise, until dose-limiting-toxicities occur. The goal is to determine a maximum tolerated dose (MTD) and what types of toxicities occur.
After Phase 1 results, MTD and list of known toxicities, are known, a small Phase 2 trial (roughly 25 to 50 patients) can establish whether there is enough efficacy to make it worth while to run a much larger Phase 3 trial (roughly 250 to 500 patients) that actually can measure efficacy with reliable statistical significance. These Phase 3 trials take lots of time & money. This is the classic drug development strategy required by the FDA under normal circumstances.
With a true emergency, as we certainly do have, corners can be cut. The Phase 1 safety trial and Phase 2 efficacy trial can be combined into a Phase 1b/2 trial. Because hydroxychloroquine's safety is already known, test it in a small cohort (3-9 patients, 3 at each 3 dose levels) of Covid-19 patients to see if they require different doses than for malaria patients, and to see if toxicities are different in these patients. Again, this must be done ± azithromycin.
If results are similar to those found with malaria patients, move right along to a Phase 2 trial with three arms. It will test hydroxychloroquine alone, hydroxychloroquine plus azithromycin, and a suitable control arm, perhaps azithromycin alone. Patients with Covid-19 will be randomly selected to enter one of those 3 arms. Some one with better understanding of statistics than I have can predict how many patients will be needed in each arm to obtain statistically significant results. Just to spitball a number, lets pick 30 patients for each arm. Write in the usual early stopping rules in case of futility or unusually good efficacy. Also include a rule that allows patients to switch arms if one arm is much worse or much better than the others.
You have to also define a suitable medical endpoint to be able to judge success from failure. I know what that might be for cancer, but not for infectious diseases.
When such a randomized controlled multi-arm Phase 2 trial is done, the results can be presented to the FDA. The FDA can make the call to approve the drug as tested, or whether this trial should be continued with more patients in each arm. Again, the emergency we face can be reason to cut many corners, but not to eliminate clinical testing.
As you can see, that French paper that described preliminary (quick & dirty) results from 36 patients, 16 treated with no test drugs (control arm), vs. 20 with hydroxychloroquine, vs. 6 patients with hydroxychloroquine plus azithromycin merely suggests an approach for proper clinical development.
As a note, I worked for the last 20 years of my career as a scientist directly involved in a US Government agency that sponsored many hundreds of clinical trials of experimental cancer drugs. I know first hand what the FDA looks for in these trials, and I know how to design these trials so they provide clinically relevant answers. It can be a dreadfully slow, but essential, process. I'm very curious what the FDA can do to speed this up.
