I forgot to mention that, for our case, the 19/23 would apply (our PE is 28 days survival and there where 2 patients still hospitalized, the rest recovered or death).

So the 17.4% death rate would apply meaning a p-value of app. 0.0033 (2-tailed).

Of course, all of the above extrapolating the paper's numbers to our trial.

Just saying the numbers look good ...