Risk factor analysis and nomogram for predicting poor symptom control in smoking asthmatics

Abstract

Abstract Background Smoking induces and modifies the airway immune response, accelerating the decline of asthmatics’ lung function and severely affecting asthma symptoms’ control level. To assess the prognosis of asthmatics who smoke and to provide reasonable recommendations for treatment, we constructed a nomogram prediction model. Methods General and clinical data were collected from April to September 2021 from smoking asthmatics aged ≥14 years attending the People’s Hospital of Zhengzhou University. Patients were followed up regularly by telephone or outpatient visits, and their medication and follow-up visits were recorded during the 6-months follow-up visit, as well as their asthma control levels after 6 months (asthma control questionnaire-5, ACQ-5). The study employed R4.2.2 software to conduct univariate and multivariate logistic regression analyses to identify independent risk factors for ‘poorly controlled asthma’ (ACQ>0.75) as the outcome variable. Subsequently, a nomogram prediction model was constructed. Internal validation was used to test the reproducibility of the model. The model efficacy was evaluated using the consistency index (C-index), receiver operating characteristic (ROC) curve, calibration curve, and decision curve. Results Invitations were sent to 231 asthmatics who smoked. A total of 202 participants responded, resulting in a final total of 190 participants included in the model development. The nomogram established five independent risk factors (P<0.05): FEV1%pred, smoking index (100), comorbidities situations, medication regimen, and good or poor medication adherence. The area under curve (AUC) of the modeling set was 0.824(95%CI 0.765-0.884), suggesting that the nomogram has a high ability to distinguish poor asthma control in smoking asthmatics after 6 months. The calibration curve showed a C-index of 0.824 for the modeling set and a C-index of 0.792 for the self-validation set formed by 1000 bootstrap sampling, which means that the prediction probability of the model was consistent with reality. Decision curve analysis (DCA) of the nomogram revealed that the net benefit was higher when the risk threshold probability for poor asthma control was 4.5 − 93.9%. Conclusions FEV1%pred, smoking index (100), comorbidities situations, medication regimen, and medication adherence were identified as independent risk factors for poor asthma control after 6 months in smoking asthmatics. The nomogram established based on these findings can effectively predict relevant risk and provide clinicians with a reference to identify the poorly controlled population with smoking asthma as early as possible, and to select a better therapeutic regimen. Meanwhile, it can effectively improve the medication adherence and the degree of attention to complications in smoking asthma patients.