Polysomnography scoring for sleep apnea. Use of a sampling method

Area of Science:

• Sleep Medicine
• Respiratory Physiology
• Diagnostic Accuracy

Background:

• Polysomnography (PSG) is essential for diagnosing sleep apnea severity through quantitative analysis of respiratory and sleep parameters.
• Manual scoring of PSG data is time-intensive, necessitating research into efficient yet accurate assessment methods.

Purpose of the Study:

• To evaluate the impact of reduced polysomnogram scoring percentages on the accuracy of sleep apnea severity estimation.
• To determine the minimum scoring percentage required to reliably assess key sleep apnea metrics.

Main Methods:

• Twenty adult polysomnograms with a wide range of apnea-hypopnea (AH) indexes were scored at varying percentages: 100%, 50%, 25%, 16.7%, 12.5%, and 10%.
• Scoring methods were compared against full (100%) scoring.
• Studies were categorized as mild, moderate, or severe based on established criteria for AH index, AH time, mean AH duration, mean and lowest oxygen saturation, and longest AH event.

Main Results:

• High accuracy in categorizing sleep apnea severity was maintained with 50% and 25% scoring for the AH index.
• 50%, 25%, and 16.7% scoring achieved 100% accuracy for AH time.
• Accuracy for mean AH duration, oxygen saturation, and longest AH events remained high with 50% and 25% scoring, with some reduction at lower percentages.

Conclusions:

• Significant reductions in polysomnogram scoring (down to 25% or 50%) can accurately estimate sleep apnea severity.
• This finding suggests potential for more efficient and cost-effective sleep apnea diagnosis without compromising clinical accuracy.
• Further validation may support the adoption of reduced scoring protocols in clinical practice.