The accuracy and difference of scoring rules and methods to score respiratory event-related leg movements in obstructive sleep apnea patients

Abstract

Objective: To compare manual scoring: 1) to the American Academy of Sleep Medicine (AASM) auto-scoring rules. 2) to the AASM and World Association of Sleep Medicine (WASM) rules, and evaluate the accuracy of the AASM and WASM for respiratory event-related limb movements (RRLM) in diagnostic and continuous positive airway pressure (CPAP) titration polysomnography (PSG).

Methods: We retrospectively, re-scored diagnostic and CPAP titration PSGs of 16 patients with obstructive sleep apnea (OSA), using manual re-scoring by the AASM (mAASM) and WASM (mWASM) criteria for RRLM, periodic limb movements during sleep (PLMS), and limb movements (LM), which were compared to auto-scoring by the AASM (aAASM).

Results: In diagnostic PSG, significant differences were found in LMs (p < 0.05), RRLM (p = 0.009) and the mean duration of PLMS sequences (p = 0.013). In CPAP titration PSG, there was a significant difference in RRLM (p = 0.008) and PLMS with arousal index (p = 0.036). aAASM underestimated LM and RRLM, especially in severe OSA. Changes in RRLM and PLMS with arousal index between diagnostic and titration PSG were significantly different between aAASM and mAASM, but there was no significant difference between scoring by mAASM and mWASM. The ratio of PLMS and RRLM changes between diagnostic and CPAP titration PSG was 0.257 in mAASM and 0.293 in mWASM.

Conclusions: In addition to the overestimation of RRLM by mAASM compared to aAASM, mAASM may also be more sensitive than aAASM in detecting RRLM changes in the titration PSG. Despite intuitive differences in the definition of RRLM between AASM and WASM rules, RRLM results between mAASM and mWASM were not significant and about 30% of RRLMs might be scored as PLMS by both scoring rules.