The French study? Aside from the fact that it wasn't a randomized study and 6 patients dropped out from the treatment group (1 died, 3 went to ICU)?

30 patients would be statistically significant. Problem is that they were non randomized, open-label, and they seemingly discarded some of the bad results. They also measured viral load, not cure, and the measurement was not done correctly in all cases.

It's basically a "there could potentially be a causal link here; needs way more research" type of study.