Background:  CCR5 antagonists, a new class of antiretrovirals (ARV), require a tropism test to identify patients with only R5 HIV, who are most likely to clinically respond. The Trofile assay (Monogram Biosciences) was used and clinically validated in the maraviroc (MVC) development program. Recently, a new tropism assay (SensiTrop, Pathway Diagnostics) became commercially available in the USA. This study was undertaken to understand the performance characteristics of the SensiTrop vs Trofile assays.

Methods:  Stored screening samples (n =100) from HIV-infected, treatment-experienced patients in the MVC Expanded Access Program were retested for tropism using the SensiTrop assay and results were compared to results of the Trofile assay. To permit a greater sensitivity of the SensiTrop assay to detect D/M or X4 HIV, baseline samples were retested with Trofile if the screening sample tested R5 by the Trofile assay. Samples that tested R5 by Trofile at screening and baseline were considered R5; samples that had a D/M or X4 tropism result at screening were considered non-R5. Only samples with a confirmed R5 (screening and BL) or D/M or X4 (screening) result were used to determine the sensitivity/specificity and positive/negative predictive values to detect CXCR4-using virus for the SensiTrop assay compared to Trofile.

Results:  Samples sent for SensiTrop testing included 52 R5, 39 non-R5 (31 D/M; 8 X4), and 9 non-reportable (NR) samples as defined by Trofile. Of 49 screening samples, 5 (10.2%) with only R5 HIV-identified by Trofile tested D/M at baseline and 4 were NR (3 samples were unavailable for baseline testing). Hence, there were 40 samples with confirmed R5 results at both screening and baseline. SensiTrop testing of screening samples showed 69 R5, 20 non-R5, and 11 NR samples. Of the 39 samples assigned as D/M or X4 by Trofile at screening, SensiTrop identified 19 as R5, 14 as CXCR4 using (D/M or X4), and 6 as NR. Compared to the Trofile assay, sensitivity and specificity of the SensiTrop assay to detect CXCR4-using virus were 42.4% (95% confidence intervals: 25.6 to 59.3) and 92.5% (95% confidence intervals:  84.3 to 100.7), respectively. Assuming a 45% prevalence of non-R5 isolates among treatment-experienced patients, positive and negative predictive values to detect non-R5 HIV (as determined by Trofile) by SensiTrop were 82.2% and 66.3% for treatment-experienced patients.

Conclusions: Relative to the results obtained with Trofile, SensiTrop lacks adequate sensitivity, failing to detect D/M or X4 HIV in more than 50% of samples.