| Popis: |
BackgroundReflex Cryptococcal antigen (CrAg) testing in HIV-positive patients is done routinely at 47 laboratories in South Africa on samples with a confirmed CD4 count ObjectiveThis study aimed to verify the diagnostic performance of newer CrAg LFA assays against the predicate method.MethodsRemnant CD4 samples collected between February and June 2019, with confirmed predicate LFA CrAg results, were retested on settled plasma with the (i) IMMY CrAg semi-quantitative (SQ) LFA; (ii) Bio-Rad RDT CryptoPS SQ; and (iii) Dynamiker CrAg SQ assays, within 24 hours of predicate testing. Sensitivity/ specificity analyses were conducted comparing predicate versus the newer assays, with McNemar’s test’s p-values reported for comparative results (p values ResultsOf the 254 samples tested, 228 had comparative CrAg results across all assays. The predicate method reported 85 CrAg positive (37.2%), compared to between 35.08 and 37.28% for the Bio-Rad, IMMY SQ and Dynamiker assays. The IMMY CrAg SQ grading (+1 to +5) showed 67% of CrAg positive results had a grading ≥3, indicative of higher CrAg concentration (infection severity). False-negative results across all assays were 95% for all. False-positive results were highest for the Dynamiker LFA (14%) with a specificity of 77% (p=0.001). IMMY SQ and Bio-Rad assays specificities exceeded 90% (p=0.6 and 0.12). Internal quality control showed 100% accuracy for all assays.ConclusionPerformance verification of newer CrAg LFA assays under typical laboratory conditions varied, with best results by IMMY SQ and Bio-Rad. The high burden of HIV and cryptococcal disease in South Africa requires high specificity and - sensitivity (>90%) to prevent unnecessary treatment/hospitalization. The added value of positivity grading for patient management needs confirmation. |
