The trial was powered to detect a 43% reduction in mortality,...

The trial was powered to detect a 43% reduction in mortality, picking such a large reduction means that a smaller sample size was needed to prove the effect. a smaller effect would mean (possibly many) more patients. Therefore more time, costs etc.

I can’t say exactly the process that was followed but the first couple of futility analyses wouldn’t have been able to say for sure that the study was not going to reach an outcome with such a small proportion of the target sample size completed.

This trial might end up showing a reduction in mortality which, due to the sample size, does not reach statistical significance. Unfortunately the flip side is that it such a difference may be due to chance alone for the same reason. If this happens then Novartis might decide to do a larger trial.

Obviously if there’s no difference at all then Novartis will almost certainly try and get out having handed over the bare minimum amount of cash. Who knows how much that will be.