Scholarlink

Background <p>The impact of ambient air pollution on perioperative outcomes in lung cancer surgery remains incompletely characterized. We synthesized available evidence while explicitly acknowledging severe structural limitations in the current literature, including extremely small evidence base and dominance by a single nationwide cohort.</p> Methods <p>We systematically searched PubMed, Scopus, Web of Science, and EMBASE through December 31, 2025, for observational studies examining associations between air pollution and post-surgical outcomes in lung cancer patients. Study quality was assessed using the Newcastle-Ottawa Scale (NOS) and ROBINS-I. Random-effects meta-analysis employed restricted maximum likelihood (REML) with Hartung-Knapp-Sidik-Jonkman (HKSJ) confidence interval adjustment. Because effect sizes from the same study share within-study correlation, we applied robust variance estimation (RVE) with the correlated hierarchical effects (CHE) model. Heterogeneity was explored through stratified analyses, meta-regression, and sensitivity assessments.</p> Results <p>Five studies contributing 24 effect sizes (<i>n</i> = 716,249) were included. 75% of effect sizes derived from a single nationwide South Korean cohort. The primary analysis showed no statistically significant overall association (hazard ratio [HR] = 1.039, 95% confidence interval [CI] = 0.995–1.084; <i>p</i> = 0.084), with extreme between-study heterogeneity (I² = 98.8%, τ² = 0.00835). The 95% prediction interval (0.757–1.424) spanned from substantially protective to substantially harmful, rendering the pooled estimate clinically uninformative. Excluding acute wildfire smoke exposure (Analysis B) yielded a marginally significant estimate (HR = 1.018, 95% CI = 1.001–1.036; <i>p</i> = 0.042; I² = 64.5%). Pollutant-specific analyses identified significant harmful associations for particulate matter ≤ 2.5&#xa0;μm (PM₂.₅) (HR = 1.031, 95% CI = 1.016–1.047), sulfur dioxide (SO₂) (HR = 1.076, 95% CI = 1.062–1.091), and wildfire smoke (HR = 1.430, 95% CI = 1.410–1.450). Conversely, ozone (O₃) exhibited a paradoxical protective association for overall mortality (HR = 0.924, 95% CI = 0.909–0.938). Short-term mortality (≤ 90 days) demonstrated a near-null effect (HR = 1.007, 95% CI = 0.971–1.044; <i>p</i> = 0.872; I² = 18.0%), representing the most reliable subgroup. Meta-regression identified geographic region as a significant moderator (North American versus Asian studies: β = 0.193, <i>p</i> = 0.002), though this comparison is limited by very few independent studies.</p> Conclusions <p>Aggregate evidence demonstrates no significant overall association, and the wide prediction interval precludes using the pooled point estimate for individual patient decision-making. However, pollutant-specific analyses identify PM₂.₅, SO₂, and wildfire smoke as potential perioperative risk factors. The paradoxical O₃ protective association likely reflects meteorological confounding rather than biological benefit. These findings are hypothesis-generating, not confirmatory, and are dominated by a single large cohort with limited geographic diversity. Standardized prospective investigations across diverse populations—including the Indian subcontinent and other high-pollution regions—are essential.</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

Ambient Air Pollution Exposure and Post-surgical Outcomes in Lung Cancer Patients: A Systematic Review and Meta-analysis

  • 期刊论文

  • 发表日期：

    2026年6月18日

查看全文
  • Seyed Masoud HaghighiKian,
  • Mehdi Khosravi-Mashizi,
  • Alireza Negahi,
  • Hossein Najdsepas,
  • Hossein Negahban,
  • Abolhasan Alijanpour,
  • Amirhossein Shahbazi,
  • Amirhosein Naseri,
  • Amirhossein Rahmani,
  • Kazem Aghili,
  • Ali Masoudi,
  • Hossein Neamatzadeh

摘要

Background

The impact of ambient air pollution on perioperative outcomes in lung cancer surgery remains incompletely characterized. We synthesized available evidence while explicitly acknowledging severe structural limitations in the current literature, including extremely small evidence base and dominance by a single nationwide cohort.

Methods

We systematically searched PubMed, Scopus, Web of Science, and EMBASE through December 31, 2025, for observational studies examining associations between air pollution and post-surgical outcomes in lung cancer patients. Study quality was assessed using the Newcastle-Ottawa Scale (NOS) and ROBINS-I. Random-effects meta-analysis employed restricted maximum likelihood (REML) with Hartung-Knapp-Sidik-Jonkman (HKSJ) confidence interval adjustment. Because effect sizes from the same study share within-study correlation, we applied robust variance estimation (RVE) with the correlated hierarchical effects (CHE) model. Heterogeneity was explored through stratified analyses, meta-regression, and sensitivity assessments.

Results

Five studies contributing 24 effect sizes (n = 716,249) were included. 75% of effect sizes derived from a single nationwide South Korean cohort. The primary analysis showed no statistically significant overall association (hazard ratio [HR] = 1.039, 95% confidence interval [CI] = 0.995–1.084; p = 0.084), with extreme between-study heterogeneity (I² = 98.8%, τ² = 0.00835). The 95% prediction interval (0.757–1.424) spanned from substantially protective to substantially harmful, rendering the pooled estimate clinically uninformative. Excluding acute wildfire smoke exposure (Analysis B) yielded a marginally significant estimate (HR = 1.018, 95% CI = 1.001–1.036; p = 0.042; I² = 64.5%). Pollutant-specific analyses identified significant harmful associations for particulate matter ≤ 2.5 μm (PM₂.₅) (HR = 1.031, 95% CI = 1.016–1.047), sulfur dioxide (SO₂) (HR = 1.076, 95% CI = 1.062–1.091), and wildfire smoke (HR = 1.430, 95% CI = 1.410–1.450). Conversely, ozone (O₃) exhibited a paradoxical protective association for overall mortality (HR = 0.924, 95% CI = 0.909–0.938). Short-term mortality (≤ 90 days) demonstrated a near-null effect (HR = 1.007, 95% CI = 0.971–1.044; p = 0.872; I² = 18.0%), representing the most reliable subgroup. Meta-regression identified geographic region as a significant moderator (North American versus Asian studies: β = 0.193, p = 0.002), though this comparison is limited by very few independent studies.

Conclusions

Aggregate evidence demonstrates no significant overall association, and the wide prediction interval precludes using the pooled point estimate for individual patient decision-making. However, pollutant-specific analyses identify PM₂.₅, SO₂, and wildfire smoke as potential perioperative risk factors. The paradoxical O₃ protective association likely reflects meteorological confounding rather than biological benefit. These findings are hypothesis-generating, not confirmatory, and are dominated by a single large cohort with limited geographic diversity. Standardized prospective investigations across diverse populations—including the Indian subcontinent and other high-pollution regions—are essential.