A Clinical Prediction Model for Invasive Pulmonary Fungal Disease in Dialysis Patients: Development, Validation, and...

Area of Science:

• Infectious Diseases
• Nephrology
• Medical Informatics

Background:

• Patients on maintenance hemodialysis (MHD) are highly susceptible to invasive pulmonary fungal infections (IPFI) due to compromised immunity and frequent interventions.
• IPFI in MHD patients presents with non-specific symptoms and inconclusive microbiology, leading to diagnostic delays, misdiagnosis rates over 50%, and high mortality (up to 40%).
• Early and accurate identification of IPFI is crucial for timely antifungal therapy and improved patient outcomes in the MHD population.

Purpose of the Study:

• To develop and validate a clinical prediction model for the early identification of invasive pulmonary fungal infections (IPFI) in patients undergoing maintenance hemodialysis (MHD).
• To improve diagnostic accuracy and reduce delays in IPFI detection among MHD patients.
• To provide a tool for risk stratification to guide antifungal therapy and optimize drug use.

Main Methods:

• A retrospective cohort study utilizing two cohorts: a modeling cohort (2022-2023) and an external validation cohort (2024).
• Inclusion criteria: Maintenance hemodialysis patients admitted with pneumonia.
• Predictors were identified using least absolute shrinkage and selection operator (LASSO) regression from 22 clinical variables, followed by logistic regression for model construction. Standardization (z-score) was applied to predictors before LASSO regression.

Main Results:

• A four-variable prediction model was developed, incorporating highest body temperature, hypothyroidism, pancytopenia, and anti-neutrophil cytoplasmic antibody (ANCA)-associated disease.
• The model demonstrated strong internal validity (optimism-corrected AUC = 0.864) and robust external validation performance (AUC = 0.877).
• Pancytopenia and maximum body temperature were identified as the most significant predictors in the model.

Conclusions:

• The developed prediction model effectively identifies MHD patients with pneumonia at high risk for IPFI, showing strong and validated predictive performance.
• This tool can aid clinical pharmacists in early risk stratification, supporting ward decision-making and antifungal stewardship.
• Implementation of this model can optimize the prophylactic and targeted use of antifungal agents, promoting rational drug therapy in MHD patients.