Daclatasvir had been tested in many more patients and in different combinations including nuc before. Safety issue hadn't come up. In addition, BMY singled out INX189 trials rather than daclatasvir trials to stop, thus at minimum they didn't think daclatasvir likely to be responsible.
In addition to what jq1234 pointed out (on the more extensive clinical data available for daclatasvir), there may also have been signs for the very same tox observed in the phase II trial, in the preclinical data of the nuke but not the NS5A. That would strongly suggest which drug is the issue.