Highly implausible, IMO; hence, the results in one of the two ISIS-FXIRx trial arms were presumably a fluke.
The p value makes it clear that unless there were "protocol issues" the result isn't a fluke. It is the sine qua non purpose of a p value to provide a measure of how flukey it could be. (Note that unrepeatable "protocol issues" are inherently more likely in single site trials)
The concern IMO should be the text in the italics since it wasn't a large number of sites and, more importantly, almost all were in EE. I'd want to see the data from the two non-EE sites (in Canada).