HIV viral load monitoring: The clinical impact of changing from RT-PCR to bDNA

Objective: To compare HIV-1 RNA levels obtained with reverse transcriptase polymerase chain reaction (RT-PCR) assay and branched DNA (bDNA) assay. Design: Retrospective chart review. Setting and participants: 194 HIV+ patients who had an HIV viral load measured by bDNA and by RT-PCR. Results: 127 (65.5%) had HIV RNA detectable by RT-PCR and bDNA. 53 (27.3%) had undetectable HIV RNA by both assays. 10 (5.1%) had an undetectable HIV RNA level by bDNA but a detectable level by RT-PCR, and 4 (2.1%) had an undetectable HIV RNA level by RT-PCR but a detectable level by bDNA. There was no statistically significant difference between the 2 assays in detecting viral copies (McNemar test for paired samples, 3.27; P = 0.0707). There was good correlation between the assays with a Pearson correlation of 0.97 (P<0.001) and a Spearman rank correlation of 0.96 (P<0.001). 2 of the 194 patients had additional testing done. Clinical decision making was not affected by the change from RT-PCR to bDNA assay. A Bland- Altman analysis of the 127 patients with detectable viral copies by both assays showed RT-PCR assay had a mean bias of 0.321 log ₁₀ viral copies/mL with limits of agreement of -0.452 to +1.098 log ₁₀ viral copies/mL. Conclusion: Switching from RT-PCR HIV viral load assay to bDNA assay did not affect clinical decision making and led to cost savings associated with HIV RNA testing. The 2 assays cannot be used interchangeably, as the RT-PCR gives higher viral load copies than the bDNA.