Department of Otorhinolaryngology, Head and Neck Surgery, Inselspital, University Hospital and University of Bern, Freiburgstrasse 18, 3010, Bern, Switzerland.
Department of Otorhinolaryngology, Head and Neck Surgery, Kantonsspital Baselland, Liestal, Switzerland.
Sleep Breath. 2023 Mar;27(1):229-237. doi: 10.1007/s11325-021-02531-9. Epub 2022 Apr 2.
The objective was to analyze the clinical implications of manual scoring of sleep studies using peripheral arterial tonometry (PAT) and to compare the manual and automated scoring algorithms.
Patients with suspected sleep-disordered breathing underwent sleep studies using PAT. The recordings were analyzed using a validated automated computer-based scoring and a novel manual scoring algorithm. The two methods were compared regarding sleep stages and respiratory events.
Recordings of 130 patients were compared. The sleep stages and time were not significantly different between the scoring methods. PAT-derived apnea-hypopnea index (pAHI) was on average 8.4 events/h lower in the manually scored data (27.5±17.4/h vs.19.1±15.2/h, p<0.001). The OSA severity classification decreased in 66 (51%) of 130 recordings. A similar effect was found for the PAT-derived respiratory disturbance index with a reduction from 31.2±16.5/h to 21.7±14.4/h (p<0.001), for automated and manual scoring, respectively. A lower pAHI for manual scoring was found in all body positions and sleep stages and was independent of gender and body mass index. The absolute difference of pAHI increased with sleep apnea severity, while the relative difference decreased. Pearson's correlation coefficient between pAHI and oxygen desaturation index (ODI) significantly improved from 0.89 to 0.94 with manual scoring (p<0.001).
Manual scoring results in a lower pAHI while improving the correlation to ODI. With manual scoring, the OSA category decreases in a clinically relevant proportion of patients. Sleep stages and time do not change significantly with manual scoring. In the authors' opinion, manual oversight is recommended if clinical decisions are likely to change.
分析使用外周动脉张力测定（PAT）进行睡眠研究的手动评分的临床意义，并比较手动和自动评分算法。
怀疑患有睡眠呼吸障碍的患者接受了 PAT 睡眠研究。使用经过验证的基于计算机的自动评分和新的手动评分算法对记录进行分析。比较两种方法的睡眠阶段和呼吸事件。
比较了 130 例患者的记录。两种评分方法的睡眠阶段和时间无显著差异。手动评分数据中的 PAT 衍生的呼吸暂停低通气指数（pAHI）平均低 8.4 次/小时（27.5±17.4/h 与 19.1±15.2/h，p<0.001）。130 个记录中有 66 个（51%）的 OSA 严重程度分类降低。对于 PAT 衍生的呼吸干扰指数，也发现了类似的效果，从 31.2±16.5/h 降至 21.7±14.4/h（p<0.001），分别用于自动和手动评分。手动评分的 pAHI 在所有体位和睡眠阶段均较低，且与性别和体重指数无关。手动评分的 pAHI 绝对差值随着睡眠呼吸暂停严重程度的增加而增加，而相对差值则减小。手动评分时，pAHI 与氧减饱和度指数（ODI）的 Pearson 相关系数从 0.89 显著提高到 0.94（p<0.001）。
手动评分导致 pAHI 降低，同时改善了与 ODI 的相关性。使用手动评分时，OSA 类别会降低，而这在临床上是一个相关的比例。手动评分不会导致睡眠阶段和时间发生显著变化。作者认为，如果临床决策可能发生变化，则建议进行手动审查。
