“…As crossover trials are subject to period intervention interactions, mor¬ tality and functional status effect mea¬ sures were also determined without in¬ cluding data from these studies. 30,35,37,40 Functional status was further examined without including data from the five tri¬ als29,31,33,39,40 that used symptom scales other than the New York Heart Associ¬ ation scale. In these analyses, effects again had the same direction, the same relative ranking, and the same statisti¬ cal significance as in the original analyses.…”