Cargando…
HIV false positive screening serology due to sample contamination reduced by a dedicated sample and platform in a high prevalence environment
Automated testing of HIV serology on clinical chemistry analysers has become common. High sample throughput, high HIV prevalence and instrument design could all contribute to sample cross-contamination by microscopic droplet carry-over from seropositive samples to seronegative samples resulting in f...
Autores principales: | , , , , , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
Public Library of Science
2021
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7799780/ https://www.ncbi.nlm.nih.gov/pubmed/33428663 http://dx.doi.org/10.1371/journal.pone.0245189 |
Sumario: | Automated testing of HIV serology on clinical chemistry analysers has become common. High sample throughput, high HIV prevalence and instrument design could all contribute to sample cross-contamination by microscopic droplet carry-over from seropositive samples to seronegative samples resulting in false positive low-reactive results. Following installation of an automated shared platform at our public health laboratory, we noted an increase in low reactive and false positive results. Subsequently, we investigated HIV serology screening test results for a period of 21 months. Of 485 initially low positive or equivocal samples 411 (85%) tested negative when retested using an independently collected sample. As creatinine is commonly requested with HIV screening, we used it as a proxy for concomitant clinical chemistry testing, indicating that a sample had likely been tested on a shared high-throughput instrument. The contamination risk was stratified between samples passing the clinical chemistry module first versus samples bypassing it. The odds ratio for a false positive HIV serology result was 4.1 (95% CI: 1.69–9.97) when creatinine level was determined first, versus not, on the same sample, suggesting contamination on the chemistry analyser. We subsequently issued a notice to obtain dedicated samples for HIV serology and added a suffix to the specimen identifier which restricted testing to a dedicated instrument. Low positive and false positive rates were determined before and after these interventions. Based on measured rates in low positive samples we estimate that before the intervention, of 44 117 HIV screening serology samples, 753 (1.71%) were false positive, declining to 48 of 7 072 samples (0.68%) post-intervention (p<0.01). Our findings showed that automated high throughput shared diagnostic platforms are at risk of generating false-positive HIV test results, due to sample contamination and that measures are required to address this. Restricting HIV serology samples to a dedicated platform resolved this problem. |
---|