Simulation of Pooled Nucleic Acid Testing to Identify Antiretroviral Treatment Failure During HIV Infection in Seoul, South Korea

摘要: Monitoring HIV RNA levels (viral loads) every 3–6 months is recommended in patients receiving antiretroviral therapy (ART) because replication the most important indicator of treatment response. In high-resource settings, regular viral load monitoring a standard care1 but not resource-constrained settings.2 Currently, number ART growing resource-limited regions and so are needs for failure development drug resistance. Because high cost testing, virological has been substituted by clinical immunologic those regions, effectiveness poor.2,3 There various efforts to identify less costly still accurate method response setting, pooling strategy proposed.4-7 In theory, pooled testing can decrease reducing assays needed screen population ART,4-7 similar nucleic acid (pooled NAT) used acute infection among individuals presenting or blood donation.8-10 The usefulness NAT detect could be affected several factors, including rate virologic failure, assay platform, level detection, inherent error type, laboratory space avoid contamination during processing, personnel availability expertise.5 Therefore, studies about use circumstances needed. Three approaches have evaluated previous studies, minipools, minipool + algorithm, matrix algorithm.4 Each these demonstrated reduction that need with minimal sensitivity is, “relative efficiency.”4,5 However, relative efficiency varied each approach, addition an algorithm determining deconvolution pools greatly enhanced efficiency. Specifically, approach showed lowest efficiency, 8 10 pool size was efficient. intermediated them, one 5 samples (5 algorithm) highest It also least likely technical errors.4 Based on potentially save more than platform technically demanding greater chance contamination.6 Moreover, requires before performed, which lead longer turnaround time feasible at our institution. we pursued evaluation approach. Based previously published data liberally assuming accuracy would 100%, calculated what demonstrate improvement over individually. point from “1 − ((1/N) NR)” R)” [R = individual samples, N per test (pool size)]. Accordingly, efficient when (R) lower (1/N2)). As example, useful <16%. We then performed simulations using collected >6 between January 2009 December 2010 urban Korean hospital. (Roche COBAS AmpliPrep/COBAS TaqMan) detection 20 copies/mL, 152,970 won (USD $136). Exceeding 200 copies/mL defined as according recent care guidelines.1 this demonstration evaluation, both actual simulated were assumed no measurement error; however, values based gamma distribution categories. During 24-month period, 1577 loads 351 who months. overall 9.7% cutoff value HIV-RNA while (copies/mL) 106(0.4%). data, selected following simulation. We arranged sample chronological order prescription. 730 tests simulation all representing 847 saved 0.54. Converting cost, total USD $115,192 saved. Since threshold defining study different study, study,4-7 investigated 1500 found reduced 6.2%, subsequent 0.62. In conclusion, seems very promising effectively monitor resources South Korea. conducted retrospectively simulation, analyze accuracy, impact approach; therefore, prospectively designed should conducted.