FIG. 29:  SERVE-HF randomized 1325 patients and reached its protocol-specified goal of 651 primary events,^[15] which represents a valuable contribution to the field as the largest study ever done.  The patients in SERVE-HF were sicker (30% with New York Heart Association (NYHA) class II and 69% with NYHA class III heart failure) than the patients in ADVENT-HF [see Figure 39]. 

ASV compliance was quite poor in SERVE-HF, with patients on treatment for a mean of only 3.7 hr/night, having been instructed to use the device for 5 hours each night.^[14] About 29% of patients in the ASV group discontinued use of the device, and 17% in the control group crossed over to PAP therapy, so there was a high rate of noncompliance.  In the intention-to-treat analysis, which included all patients, irrespective of whether they adhered to the protocol or not, there was no difference in the primary endpoint between the ASV treatment and control groups.^[15]  However, both all-cause and cardiovascular mortality in the ASV group were higher than in the control group, with hazard ratios (HRs) of 1.28 and 1.34, respectively, both highly significant differences.^[15]  So although there was no significant difference in the primary endpoint, for perhaps a more important endpoint, all-cause mortality and cardiovascular mortality, the results were worse in the ASV-treated group.