Scholarlink

Background <p>Snoring is a common marker of upper-airway obstruction. However, how the temporal load of snoring relates to OSA severity from non-OSA to severe disease across the full AHI spectrum remains incompletely understood.</p> Methods <p>We conducted a retrospective, single-center study of 1,423 adults who underwent overnight portable sleep monitoring (PSM; Type III HSAT). Temporal snoring features included snore frequency (events/night), mean snore duration (s), total snoring time (min/night), and the percentage of estimated sleep time spent snoring (%). The primary outcome was the device-derived apnea–hypopnea index (AHI) as a continuous measure of OSA severity. Associations were evaluated using multivariable linear regression, with non-linearity explored using smooth curve fitting (generalized additive models) and segmented (threshold) regression.</p> Results <p>In fully adjusted models, higher snore event count and greater cumulative snoring burden (total snoring time and percentage of TST spent snoring) were associated with higher AHI. Mean snore duration showed an overall inverse association in linear models, while non-linear modeling indicated feature-specific patterns. Smooth curve fitting and segmented regression identified inflection points at ~ 15&#xa0;s (mean duration), ~ 75.6&#xa0;min/night (total snoring time), and ~ 24.3% and ~ 74.5% (snoring percentage). Subgroup analyses showed no interaction by sex; BMI modified the mean snore duration–AHI association, while associations for other snoring metrics were broadly consistent across BMI strata.</p> Conclusions <p>Temporal snoring metrics derived from PSM were associated with device-derived AHI and exhibited feature-specific non-linear patterns with interpretable inflection points. These candidate thresholds are hypothesis-generating and are not intended as diagnostic cut-points; prospective PSG-based replication and external validation are required before clinical application.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Temporal snoring metrics and obstructive sleep apnea severity: nonlinear associations and data-driven inflection points in a clinical cohort

  • 期刊论文

  • 发表日期：

    2026年3月12日

查看全文
  • Zhenxing Yu,
  • Fen Wang,
  • Zhipeng Zhang,
  • Jing Zha,
  • Cunlin Gong,
  • Liying Deng,
  • Jianglong Tu,
  • Yongmin Ding

摘要

Background

Snoring is a common marker of upper-airway obstruction. However, how the temporal load of snoring relates to OSA severity from non-OSA to severe disease across the full AHI spectrum remains incompletely understood.

Methods

We conducted a retrospective, single-center study of 1,423 adults who underwent overnight portable sleep monitoring (PSM; Type III HSAT). Temporal snoring features included snore frequency (events/night), mean snore duration (s), total snoring time (min/night), and the percentage of estimated sleep time spent snoring (%). The primary outcome was the device-derived apnea–hypopnea index (AHI) as a continuous measure of OSA severity. Associations were evaluated using multivariable linear regression, with non-linearity explored using smooth curve fitting (generalized additive models) and segmented (threshold) regression.

Results

In fully adjusted models, higher snore event count and greater cumulative snoring burden (total snoring time and percentage of TST spent snoring) were associated with higher AHI. Mean snore duration showed an overall inverse association in linear models, while non-linear modeling indicated feature-specific patterns. Smooth curve fitting and segmented regression identified inflection points at ~ 15 s (mean duration), ~ 75.6 min/night (total snoring time), and ~ 24.3% and ~ 74.5% (snoring percentage). Subgroup analyses showed no interaction by sex; BMI modified the mean snore duration–AHI association, while associations for other snoring metrics were broadly consistent across BMI strata.

Conclusions

Temporal snoring metrics derived from PSM were associated with device-derived AHI and exhibited feature-specific non-linear patterns with interpretable inflection points. These candidate thresholds are hypothesis-generating and are not intended as diagnostic cut-points; prospective PSG-based replication and external validation are required before clinical application.