The Data Safety and Monitoring Board would intervene if the death rate is attributable to the drug, not placebo and yes they can unblind only to themselves. If they stopped the trial because of placebo death rates then why have a trial at all where death is a comparison measure.