CPAP is the established treatment of moderate/severe OSA. The parameters capable to influence the level of effective CPAP in these patients are yet unknown. Previous predictive equations for CPAP (hereafter called eCPAP) included body mass index (BMI), neck circumference (NC) and baseline apnea-hypopnea index (AHI. None of these equations included arousal index (AI) or desaturation index (DI).
Fifty consecutive patients (20 females and 30 males) with moderate/severe OSA, who underwent a split-night or a CPAP titration polysomnographic study in our Sleep Center, were retrospectively analyzed. We used multiple regression analysis to compute a predictive equation of the effective level of CPAP and we analyzed the correlation between eCPAP and the prescribed CPAP (pCPAP).
The derived predictive equation for pCPAP was: pCPAP = 6.8 + 0.001AHI + 0.10BMI - 0.08NC + 0.06AI - 1.14DI (p<0.0001, R2=0.45, SD=2.0). The mean pCPAP was 10 (±2.6 SD); the mean eCPAP was 8 (±2.3 SD); mean difference was 2 cmH2O. The mean difference was slightly smaller in split-night vs. titration studies (2.0 vs. 2.6), in males vs. females (2.0 vs. 2.8), and in severe OSA vs. moderate OSA (2.0 vs. 3.0) cmH2O. The derived equation is distinct from previously published equations, which estimated eCPAP. There was a weak correlation between pCPAP and eCPAP [R2=0.48, p<0.0001]. This could be explained by the fact that our CPAP prescription protocol takes into account not only normalization of AHI, but also near-normalization of AI and DI.
We determined an equation for pCPAP based on several clinical parameters. The correlation between eCPAP and the pCPAP was weak, warranting new studies to explore the lack of correlation and the role of additional predictive parameters.
This is a pilot phase of a larger study for derivation and validation of the best predictive equation for effective CPAP in OSA. The predictive value of OSA severity parameters such as AHI, AI, DI, and body habitus characteristics such as NC and BMI will be evaluated extensively in the main study.
Octavian Ioachimescu, None.