253. Evaluation of a Cryptococcal Antigen Lateral Flow Assay and the Burden of Cryptococcal Disease: A Cohort Study at Grady Memorial Hospital in Atlanta, Georgia

BACKGROUND: While Cryptococcus neoformans is a major cause of morbidity and mortality among HIV-infected persons worldwide, there is scarce recent data on disease prevalence in the United States, including in Southeastern states, where HIV rates are high. We sought to determine the prevalence of cryptococcal disease and compare the performance of a cryptococcal antigen (CrAg) lateral flow assay (LFA) vs. latex agglutination (LA) test. METHODS: All patients from Grady Memorial Hospital in Atlanta, Georgia who had a serum or cerebrospinal fluid (CSF) sample sent for CrAg LA testing as part of routine management from November 2017 to July 2018 were included. The LFA was performed on all samples by research staff; results were not available to clinicians. Rates of disease and agreement between the LA test and LFA were calculated. RESULTS: Among 467 patients, 570 LA tests were performed; 417 on serum and 153 on CSF (87 patients with multiple tests performed). Mean age was 44 years, and most were male (n = 322, 69%). Most patients had HIV (n = 371, 79%); median CD4 count was 73 cells/mm(3) and 77% were not receiving ART. Among HIV-infected individuals, testing was performed equally in the inpatient and outpatient setting. Cryptococcal testing was done in 53 persons without apparent risk factors. Thirty-three (7%) patients had a positive serum or CSF test. Five (1%) patients had both a positive serum and CSF LA test and LFA. While the overall agreement between the LA test and LFA was substantial to high for CSF (κ = 0.71) and serum (κ = 0.93), respectively, there were important discrepancies. Four patients with a negative serum LA test had a positive serum LFA. Five patients had false-positive CSF LA tests, determined by negative CSF LFA testing, India ink, and CSF and fungal cultures. All were treated with amphotericin and flucytosine with one patient experiencing a severe anaphylactic reaction to amphotericin. CONCLUSION: We found a moderately high rate of cryptococcal disease and important discrepancies between the LA test and LFA. The LFA appeared to be more sensitive for cryptococcemia and more specific for meningitis. Clinical implications of these findings include earlier detection and treatment of cryptococcemia, and averting unnecessary treatment of meningitis with costly medications associated with high rates of adverse events. [Image: see text] DISCLOSURES: All authors: No reported disclosures.