The impact of smoking on aneurysmal rupture in female patients: a nationwide retrospective cohort study
摘要
Intracranial aneurysms (IAs) are the most common cause of non-traumatic subarachnoid hemorrhage (SAH). Although multiple risk factors have been identified—including female sex and tobacco exposure—smoking remains a well-established and modifiable driver of aneurysmal rupture; however, contemporary, women-focused, nationally representative estimates of rupture risk and care patterns using balanced comparisons and neurosurgically relevant inpatient outcomes derived from large, adjusted cohorts using balanced comparisons remain limited. The Healthcare Cost and Utilization Project (HCUP) National Inpatient Sample (NIS) was queried to identify female patients diagnosed with IAs between January 1, 2003, and December 31, 2020. The primary focus was to assess association with ruptured presentation at admission. A 1:1 matching was used to compare outcomes between female smokers and nonsmokers. The K-nearest neighbor method was employed with the Mann–Kendall test to analyze the data trends. A Random Forest model was developed to classify ruptured presentation at admission, predict the rupture risk and to quantify feature contributions through an association plot and feature importance plot leveraging SHAP (Shapley Additive Explanations) values. All analyses were conducted using R and Python software. Using the HCUP National Inpatient Sample (NIS), a nationwide real-world inpatient database, we identified 32,530 hospitalized women with intracranial aneurysm diagnoses. Of the 32,530 patients identified, 64% were smokers (N = 20,753) and 36% were non-smokers (N = 11,777). Following KNN matching, each group had 11,777 patients with a mean age of 57 ± 11.4 years. Patients in the smokers’ group were most commonly White (N = 9,268; 78.7%), followed by Black (N = 1,390; 11.8%) and Hispanic (N = 648; 5.5%). A higher number of female smokers had a history of hypertension (61.2% vs. 49.7%, p < 0.001) and dyslipidemia (32.6% vs. 24.1%, p < 0.001). Female smokers were also more likely on long-term aspirin (20.8% vs. 14%, p < 0.001), NSAIDs (5.4% vs. 3.8%, p < 0.001), and steroids (1% vs. 0.7%, p = 0.005). More female smokers underwent surgical treatment (30.8% vs. 28.9%, p = 0.001), while more non-smokers received endovascular treatment (71.1% vs. 69.2%, p = 0.001). The incidence of aneurysmal rupture was significantly higher in smokers (27.3% vs. 11.8%, p < 0.001). Multivariable logistic regression showed higher odds of ruptured presentation in female smokers (OR = 1.1 [1.02–1.19], p = 0.012). Aging, dyslipidemia, diabetes, and long-term aspirin use were independently associated with rupture among smokers. Association analysis and SHAP-based feature importance analysis (AUC = 0.8) showed that smoking status was the most critical predictor of rupture. In a large-scale, women-only national cohort, documented tobacco exposure was associated with higher odds of aneurysmal rupture at presentation compared with non-smokers. This work provides contemporary, women-specific estimates that can inform risk communication and counseling and motivate future sex-informed risk stratification efforts, while reinforcing smoking cessation counseling as an important, clinically actionable target. Although the adjusted effect estimate is modest and should be interpreted cautiously within an administrative dataset—likely reflecting exposure misclassification (current vs former; no pack-years) and residual confounding inherent to claims-based analyses—the direction of effect is consistent with prior literature and supports smoking cessation counseling as an actionable risk-communication target for women with intracranial aneurysms.