This is a good point and one that is applicable to all trials using placebos with double-blinding in which the active drug has distinctive side effects - there are cues that patients might be able to use to determine that they're on the active medication even with the blinding. Researchers are aware of it and it's hard to get around. Like a lot of these things, it's less likely as the explanation for the result of the trial if the difference between active drug and placebo groups is big. I think that in the case of this trial, the late-arriving improvement is also an argument against patients simply getting a bigger placebo/perceived improvement boost from Rituximab - I'd expect those effects to kick in pretty quickly. Personally, I've always been surprised that researchers are not required as standard in every trial to ask patients whether they think they have been administered the active drug or placebo. At least they'd then have a measure of patients' ability to tell which condition they're in, despite the attempts at blinding.