Shi Haolin, Ma Xiuhua
Beijing Friendship Hospital, Beijing Daxing District People's Hospital, Capital Medical University Daxing Teaching Hospital, Beijing, China.
BMC Pulm Med. 2025 May 3;25(1):213. doi: 10.1186/s12890-025-03684-z.
As a cardiovascular health (CVH) assessment tool, Life's Crucial 9 (LC9) is often associated with diverse chronic health indicators. However, no study has yet explored the association of LC9 with multifactorial components of lung health. Thus, this study aimed to investigate the correlation of LC9 with lung health.
This cross-sectional study used data from the National Health and Nutrition Examination Survey (NHANES), which covers individuals aged 40 years and older with complete LC9 and lung health data. Multiple regression was employed in linear relationships investigation, while Restricted Cubic Spline (RCS) was used to explore nonlinear relationships. Subgroup analyses and interaction tests demonstrated the stability of associations. Combining LC9 to build a Light Gradient Boosting Machine (LightGBM) machine learning (ML) model to predict lung health, Shapley Additive Explanations (SHAP) sorted the contribution of LC9 components to the model.
From a total of 10,461 study participants, 1725 had low CVH, 7476 had moderate CVH, and 1260 had high CVH. There was a strong positive correlation between LC9 score and lung health. This association remained consistent across subcomponent strata. RCS analysis revealed non-linear associations between LC9 and respiratory outcomes, including cough, asthma, and COPD. The LightGBM model incorporating LC9 demonstrated excellent predictive performance for lung health, with favorable metrics in Area Under the Curve (AUC), accuracy, and specificity. SHAP analysis identified depression, nicotine exposure, and BMI scores as the predominant contributors among LC9 components to the model's predictive capability.
Individuals with better CVH as assessed by LC9 tended to have better lung health. The combination of the LightGBM model could achieve better prediction results.
作为一种心血管健康（CVH）评估工具，生命关键9项（LC9）常与多种慢性健康指标相关联。然而，尚无研究探讨LC9与肺部健康多因素成分之间的关联。因此，本研究旨在调查LC9与肺部健康的相关性。
这项横断面研究使用了来自美国国家健康与营养检查调查（NHANES）的数据，该调查涵盖了40岁及以上且拥有完整LC9和肺部健康数据的个体。在研究线性关系时采用多元回归，而受限立方样条（RCS）用于探索非线性关系。亚组分析和交互检验证明了关联的稳定性。结合LC9构建轻量级梯度提升机（LightGBM）机器学习（ML）模型来预测肺部健康，SHAPley值法对LC9各成分对模型的贡献进行排序。
在总共10461名研究参与者中，1725人CVH较低，7476人CVH中等，1260人CVH较高。LC9评分与肺部健康之间存在强正相关。这种关联在各子成分层次中保持一致。RCS分析揭示了LC9与包括咳嗽、哮喘和慢性阻塞性肺疾病（COPD）在内的呼吸结局之间的非线性关联。纳入LC9的LightGBM模型对肺部健康具有出色的预测性能，在曲线下面积（AUC）、准确性和特异性方面具有良好指标。SHAP分析确定抑郁、尼古丁暴露和体重指数（BMI）评分是LC9各成分中对模型预测能力贡献最大的因素。
通过LC9评估的CVH较好的个体往往肺部健康状况较好。LightGBM模型的组合能够取得更好的预测结果。
