肺功能与强直性脊柱炎双向因果关系： 一项双样本孟德尔随机化研究

Bidirectional Causal Relationship between Pulmonary Function and Ankylosing Spondylitis: A Two-Sample Mendelian Randomization Study

摘要: 目的：观察性研究提示肺功能与强直性脊柱炎(AS)存在相关性，但因果方向尚不明确。本研究采用双向双样本孟德尔随机化(MR)方法，评估肺功能指标与AS之间的潜在因果关系。方法：从已发表GWAS中提取第1秒用力呼气容积(FEV1, n = 421,986)、用力肺活量(FVC, n = 422,876)和呼气峰流速(PEF, n = 421,986)的遗传工具变量，并与AS GWAS数据(n = 164,682)进行双向MR分析。工具变量按关联强度筛选并进行连锁不平衡剪枝。以逆方差加权法(IVW)为主要分析方法，并采用加权中位数法(WM)和MR-Egger回归进行敏感性分析；通过异质性检验、水平多效性检验、留一法及多变量MR分析评估结果稳健性。结果：正向MR显示，遗传预测的FVC降低与AS风险升高显著相关(IVW OR = 0.678, 95% CI: 0.518~0.889)。未发现FEV1或PEF与AS存在显著因果关联。反向MR显示，遗传预测的AS与FEV1轻度下降存在统计学关联；但该效应量极小(OR = 0.996)，其实际影响幅度有限，与FVC或PEF未见显著因果关联。结论：本研究提示，肺功能与强直性脊柱炎之间可能存在指标特异性的双向关联：遗传预测的FVC降低与AS风险增加相关，而AS与FEV1轻微下降之间存在统计学关联，但由于效应量极小，其临床意义仍需谨慎解释。上述发现仍需在实验研究和更大样本队列中进一步验证。

Abstract: Objective: Observational studies suggest an association between pulmonary function and ankylosing spondylitis (AS), but the causal direction remains unclear. We conducted a bidirectional two-sample Mendelian randomization (MR) study to assess potential causal relationships between pulmonary function indices and AS. Methods: Genetic instruments for forced expiratory volume in 1 second (FEV1, n = 421,986), forced vital capacity (FVC, n = 422,876), and peak expiratory flow (PEF, n = 421,986) were derived from published genome-wide association studies (GWAS) and applied to AS GWAS data (n = 164,682). Instruments were selected by association strength and linkage disequilibrium pruning. The inverse variance-weighted (IVW) method was the primary analysis, complemented by weighted median and MR-Egger regression. Heterogeneity and directional pleiotropy were assessed, and leave-one-out and multivariate MR analyses tested robustness. Results: Genetically predicted lower FVC was significantly associated with a higher risk of AS (IVW OR = 0.678, 95% CI = 0.518~0.889). No significant causal association was observed between FEV1 or PEF and AS. Reverse MR analyses indicated that genetically predicted AS was statistically associated with a slight reduction in FEV1 (IVW OR = 0.996, 95% CI = 0.993~0.998); however, the effect size was extremely small, suggesting a limited magnitude of impact. No significant causal association was observed between AS and FVC or PEF. Conclusion: This study suggests a potentially index-specific bidirectional relationship between pulmonary function and AS. Genetically predicted lower FVC was associated with an increased risk of AS, whereas genetically predicted AS showed a statistical association with a slight reduction in FEV1. Given the extremely small effect size, the clinical significance of this reverse association should be interpreted with caution. These findings require further validation in experimental studies and larger cohorts.