P399 Diagnostic value of Candida coloni zation index and serum Candida mannan antigen for candidemia in febrile episodes of pediatric lymphoreticular malignancies

Abstract Poster session 3, September 23, 2022, 12:30 PM - 1:30 PM Objective To evaluate the diagnostic performance of Candida colonization index and serum Candida mannan antigen predicting candidemia in febrile episodes of pediatric lymphoreticular malignancies Methods It was a prospective observational study done for 18 months, from November 2018 to April 2020 at the pediatric oncology unit of a multispecialty tertiary care center. Based on our patient load, duration of the proposed study, and available resources, a sample size of 49 (n = 49) was decided and 100 febrile episodes in children with lymphoreticular malignancy were studied. Children below 12 years, receiving chemotherapy for hematological malignancy having oral or axillary temperature ˃38.3°C for ˃1 h were included in this study. Children receiving the antifungal treatment in last 7 days were excluded from the study. Blood collected on day1 and day4 was cultured in BACTEC-9120. For colonization, swabs and samples were collected and cultured on SDA on day1, day4, and day8. All Candida isolates were subcultured on SDA and subjected to Gram's stain, germ tube test followed by Microscan identification. DNA sequencing followed by phylogenetic analysis was done for all the isolates of Candida recovered from blood. Antifungal susceptibility of yeast stains was done. Serum collected on day1 was used for C. mannan antigen detection using ELISA system. Results Prevalence of candidemia was 5%. Non-albicans Candida spp were isolated from blood cultures on day 4. Candida colonization decreased from day1 to day8. Colonization index (CI) day1 showed 80% sensitivity 98.9% specificity, and 98.9% negative predictive value. Significant colonization (CI ≥0.5) was seen in a larger proportion of cases that developed candidemia. There was a significant association of Candida colonization (CI ≥0.5) with occurrence of candidemia on day1 and day4. A total of 4 (80%) of candidemia episodes were positive for serum mannan antigen while 1 (20%) was negative. Mannan antigen was detected earlier with 80% sensitivity, 92.6% specificity, and 98.9% negative predictive value. All Candida isolates were sensitive to fluconazole, amphotericin-B, and caspofungin. Receiver operator characteristic curves for diagnostic performance of various parameters in predicting candidemia show the following trends: Best parameter in terms of AUROC is the CI (Day 1). Best parameters in terms of sensitivity are the CI (Day 1), CI (Day 4), and mannan antigen level. Best parameter in terms of specificity is the CI (Day 8). Best parameter in terms of positive predictive value is the CI (Day 1). Best parameters in terms of negative predictive value are the CI (Day 1), CI (Day 4), and mannan antigen level. Best parameters in terms of diagnostic accuracy are the CI (Day 1), CI (Day 8). Conclusion The CI can predict candidemia but the threshold value needs to be explored in pediatric patients with lymphoreticular malignancies. Mannan antigen detection gives early results with a high negative predictive value.