Comparison of hypopnea definitions in lean patients with known obstructive sleep apnea hypopnea syndrome (OSAHS)

In the interest of improving inter-rater reliability and standardization between sleep laboratories, hypopnea definitions were recently changed to place less emphasis on arousal scoring and more emphasis on oxygen desaturations. We sought to determine whether these changes would affect detection and treatment of OSAHS in lean patients-a group known to desaturate less-than-obese patients.

Thirty-five lean subjects (15 male, 20 women, five post-menopausal) diagnosed OSAHS and a documented benefit from treatment had diagnostic polysomnograms (PSG) originally scored using the American Academy of Sleep Medicine (AASM) rule from 1999 (referred to as "Rule C"). These patients had appropriate clinical care based on those results. PSG records were then re-scored in a randomized and blinded fashion utilizing hypopnea Rule A and B of the 2007 AASM guidelines.

Baseline mean (SD) apnea hypopnea indices (AHI) for rules A, B, and C were 6.4 (3.1), 20.6 (8.2), and 26.9 (7.3), respectively (p < 0.0001). Mean (SD) BMI was 24.4 (1.0). By design, all subjects were treatment responders. Eighty-six percent with CPAP, 83% with oral appliance, and 100% with surgical intervention reported resolution of their initial daytime or sleep complaint. Post-treatment AHIs for rules A, B, and C were 0.8 (0.9), 1.8 (1.2) and 2.3 (1.6; p < 0.001). In all three scoring conditions, the AHI was reduced significantly with treatment (p < 0.001). A repeated measures ANOVA of the difference between scoring methods indicated statistically significant differences between all three strategies at both pre- and post-treatment (p < 0.001). Sleepiness on the Epworth sleepiness scale decreased from a mean of 10.9 (2.3) to 5.7 (1.3) with treatment (p < 0.001). This change in subjective rating of sleepiness was more strongly correlated with rules B and C (r = 0.6) and more modestly correlated with Rule A scoring (r = 0.4).

Response to treatment was more tightly correlated with arousal based scoring rules B and C in this group of lean subjects. The1999 hypopnea rule was used at baseline to detect this cohort of patients with OSAHS that ultimately benefitted from treatment. Rule B detected OSAHS and correlated well with response to treatment, but many more were categorized as mild (5 < AHI < 15) at baseline. Since 40% of the subjects had an AHI less than 5 with Rule A, lack of sensitivity should be considered before applying Rule A to the scoring of sleep studies in lean patients.