Risk factors of COVID-19 associated pulmonary aspergillosis in a high endemic setting and development of a bedside clinical risk prediction score

ABSTRACT

Background. India has a high burden of invasive fungal infections. Although invasive aspergillosis was also reported during the COVID-19 pandemic, the real world data on the risk factors and outcome of CAPA are limited. Our aim is to determine risk factors and clinical outcomes of CAPA and develop a prediction model for patient stratification. Methods. A retrospective, case-control study was conducted at a 1300-bedded South Indian tertiary care academic centre from June 1st, 2020 to May 31st, 2021. CAPA cases were defined by 2020 ECMM/ISHAM consensus criteria as possible, probable, and proven infection. Age and admission period matched control group with COVID-19 but without aspergillosis was selected in a 11 ratio. A risk scoring stratification for CAPA was developed based on the significant CAPA risk factors using logistic regression model. Figure 1 Receiving operating characteristic curve of CAPA incidence score for predicting CAPA in the study cohort Results. 95 CAPA cases, of which 75(79%) were probable and 20(21%) possible, were diagnosed during the study period. 84(88.4%) patients had moderate to severe COVID-19, and 75(78.9%) were treated with steroids, most commonly dexamethasone (Table 1). The time from COVID-19 diagnosis to CAPA was 13 days (IQR 12). 40(42.1%) of patients were on mechanical ventilation at CAPA diagnosis. Outcome measures (MV, NIV and hospital/ICU stay) were significantly higher in CAPA patients compared to controls(Table 2). Neutropenia, use of steroids, broad spectrum antibiotic use, fluconazole prophylaxis and absence of co-infecting pathogen were found to be significant factors associated with CAPA(p< 0.05) (Table 3). An optimal risk score of >=10.00 predicted CAPA with a sensitivity of 84.2% and a specificity of 59% with an area under the curve of 0.77 (PPV=67.23%, NPV=78.87%)(Fig 1). 75(78.9%) patients had positive serum aspergillus galactomannan with an average value of 1.89+/-1.65. 28-day (41.1% vs 33.7%, p=0.13) and 6-week all cause-mortality (48.4% vs 37.9%, p=0.07) were higher, but not statistically significant, for CAPA. Significant univariate variables were included in the multivariate logistic regression model and predicted probabilities based on the beta co-efficients of the significant variables in the model were transformed to generate CAPA incidence score for patients. Conclusion. Risk factors of CAPA in Indian were similar to those reported previously in other countries. CAPA can be seen in severe COVID-19 patients who are not mechanically ventilated. A CAPA risk scoring system, that needs external validation, is a simple and feasible tool that could be useful in stratification of patients suspected of CAPA.