Buspirone as a Potential Treatment for Recurrent Central Apneas

The hypothesis is that buspirone is a safe, effective drug to reduce the occurrence of recurrent central apnea and irregular breathing found in the setting of heart failure. A secondary hypothesis is that its effect will be similar to that or acetazolamide. Study Design: A one-dose double-blind crossover study of buspirone vs. placebo vs. acetazolamide will be performed to determine if active drug alters the number and/or severity of recurrent central apneas and hypopneas (AHI) in patients with heart failure. AHI is the primary outcome variable. In the initial phase of this study, we will recruit 18-20 patients to obtain ~15 complete studies, using the assumption of a ~20% drop-out, to reach a pre-set significance level of a 30% reduction in AHI in the drug groups with a power of 0.90 and a p=0.05 by post-hoc testing. Power estimates were calculated using the means and SDs derived from the population reported the study of acetazolamide by Javaheri et al (2006). A 30% reduction in AHI would be meaningful. A 15% dropout rate was present in the study by Javaheri et al (2006), but as our study is a three-way comparison, we chose a slightly higher rate. The reasons stated in these articles for a drop out included: viral illness, GI upset (on placebo or on theophyllin), tired of the sleep studies, and desire to terminate without cause. Statistical Analyses. Analysis of variance for repeated measures using Sidak's correction will be used to compare placebo, buspirone, and acetazolamide studies. For variables that are not normally distributed, Dunn's nonparametric test for multiple comparisons will be used. p > 0.05 will be considered significant. Mean values and SDs will be reported. This single dose, one night study is called Buspirone as a Potential Treatment for Recurrent Sleep Apnea I.

The randomization will be in a block design, and the analysis will take into account the blocked design. We will recruit 30 patients to obtain ~27 complete studies, using the assumption of a ~25% drop-out, to reach a pre-set significance level of a 50% reduction in AHI in the drug groups with a power of 0.90 and a p=0.05 by post-hoc testing (see Table C below). Power estimates were calculated using the means and SDs derived from the population reported the study of a one week trial of acetazolamide by Javaheri, values similar to those in the drug trial for theophyllin. Our reasoning is that a 50% reduction in AHI would be most meaningful. Our drop-out rate in the one-night study is estimated at ~25%.

Exclusion criteria of use of selective serotonin reuptake inhibitors (SSRIs) or antidepressants, while necessary because one of the drugs was buspirone, were too stringent for completion of this study in the VA setting. Of ~1000 patient charts screens, 8 were eventually entered into the trial, so that power criteria were not met. Records are being utilized to probe for hidden features in the PSG for use in future drug trials.