Risk factors and a prediction model of poor prognosis in patients with invasive aspergillosis in a general hospital

BACKGROUND: This study retrospectively analyzed the epidemiological trends of invasive aspergillosis (IA) at the Anhui Public Health Clinical Center. By evaluating the clinical risk factors associated with IA, an individualized nomogram prediction model was constructed. This study aims to improve early diagnosis and clinical management of severe invasive aspergillosis. METHODS: A retrospective analysis was conducted on the clinical data of 307 patients diagnosed with IA between January 2019 and March 2024. Patients were categorized into two groups based on clinical outcomes: the favorable outcome group (n = 268) and the unfavorable outcome group (n = 39). Least absolute shrinkage and selection operator (LASSO) regression and multivariable logistic regression analyses were employed to identify independent risk factors associated with poor prognosis. Subsequently, a nomogram was constructed to develop a risk prediction model for IA. RESULTS: Four variables associated with poor prognosis were identified via LASSO regression: ICU length of stay, elevated C-reactive protein (CRP), elevated blood glucose, and renal dysfunction (all P < 0.05). Multivariable logistic regression revealed that ICU length of stay (OR = 1.049, 95%CI: 1.003-1.096, P = 0.028), C-reactive protein (OR = 1.007, 95%CI: 1.003-1.012, P = 0.001), blood glucose (OR = 1.119, 95%CI: 1.041-1.206, P = 0.003), and renal failure (yes vs. no, OR = 3.187, 95%CI: 1.412-7.048, P = 0.004) were independent risk factors for poor prognosis. The area under the receiver operating characteristic curve (AUC) was 0.810, with a bootstrap-validated AUC of 0.7962, demonstrating favourable discrimination and stability. Calibration curves indicated favourable agreement between predicted and observed probabilities, and the Hosmer-Lemeshow test showed no significant deviation (P = 0.615), confirming favorable model fit. CONCLUSIONS: IA detection increased significantly over the past two years. We developed and validated a risk-factor-based nomogram that demonstrates good predictive accuracy and robustness in identifying poor prognosis among IA patients. This individualized tool, accessible via an interactive web-based application, provides clinicians with a good basis for early risk stratification and timely therapeutic intervention, ultimately aiming to improve clinical outcomes in high-risk populations.