Systematic Review Exercise, Breathwork and Combined Exercise and Breathwork, Acute Interventions for Stress Reduction in Healthy Adults: A Systematic Review and Meta-Analysis

This review's primary analysis compared the relationship between exercise, breathwork and combined interventions in outcome measures that included any reference to acute stress measurements, as outlined in Table 1. Separate subgroup meta-analyses examined the relationship between intervention duration, categorised as short, medium and long and outcome measures. An additional subgroup meta-analysis was conducted to examine the relationship between low, moderate, vigorous and self-moderated exercise intervention intensity and outcome measures. All meta-analyses were conducted using random-effect models (Borenstein & Rothstein, 1999) on pooled studies to compare standardised mean differences between conditions pre-post exercise, breathwork, combined, and controls to compute an estimate Hedges' g ES, along with 95% confidence intervals (CI) for outcome measures. The Hedges' g threshold was set at 0.2 small, 0.5 medium and 0.8 large, along with 95% CI around the estimated ES (differences between intervention and control groups) (Cohen, 1992). An alpha level of P < 0.05 was used to determine significance. All statistics were analysed using Comprehensive Meta-Analysis (V3, Biostat, Englewood, USA) and presented as Forrest plots (where applicable). Heterogeneity between studies was measured using the I2 statistic (Higgins & Thompson, 2002), and the threshold for cut-off percentage points of 25% (low), 50% (moderate) and 75% (high) was used. The risk of publication bias was tested using Egger's regression significance test and funnel plot symmetry/asymmetry (Egger et al., 1997).

本综述的核心分析对比了运动、呼吸训练及联合干预方案与结局指标间的关联，结局指标涵盖所有涉及急性应激测量的内容，具体如表1所示。单独开展的亚组元分析（meta-analysis）分别针对短、中、长三类干预时长，考察其与结局指标的关联；另有一项亚组元分析，聚焦于低、中、高强度及自我调节型运动干预强度与结局指标间的关联。所有元分析均采用随机效应模型（Borenstein & Rothstein, 1999），对汇总后的研究数据进行分析，对比运动组、呼吸训练组、联合干预组与对照组在干预前后的标准化均差，以计算Hedges' g效应量（ES, effect size），并给出各结局指标的95%置信区间（CI）。Hedges' g的效应量阈值设定为：0.2为小效应，0.5为中等效应，0.8为大效应，同时计算干预组与对照组间效应量估计值的95%置信区间（Cohen, 1992）。本研究采用α=0.05作为显著性检验阈值。所有统计分析均通过Comprehensive Meta-Analysis（V3, Biostat, 美国恩格尔伍德市）完成，结果以森林图（适用于相关场景时）呈现。研究间的异质性采用I²统计量进行评估（Higgins & Thompson, 2002），并以25%（低异质性）、50%（中等异质性）、75%（高异质性）作为异质性划分阈值。采用Egger回归显著性检验与漏斗图对称性分析，检验本次研究的发表偏倚风险（Egger et al., 1997）。