We have reported previously on the unpublished preliminary results of several prospective studies that evaluated the clinical utility of resistance testing, based on short-term virologic responses. In early 2002, the final results of four of these studies were published, bringing the number of published studies to six (see also the GART and VIRADAPT studies) [1-6]. The results of these four studies provide mixed conclusions regarding the virologic impact of resistance testing performed prior to sequent modification of antiretroviral therapy. The results of these four studies provide mixed conclusions regarding the virologic impact of resistance testing performed prior to sequent modification of antiretroviral therapy. All six prospective studies completed to date may be reviewed at a glance in Table 1.

The ARGENTA trial, which was conducted between April 1999 and February 2000, evaluated 174 subjects who had HIV-1 RNA levels > 2000 copies/mL or a reduction of < 1 log₁₀ copies/mL HIV-1 RNA after being treated for more than 2 months with their most recent ARV regimen [1]. Subjects had a median of 24 months of total ARV experience and had used a median of two HAART regimens, with 41% of subjects having received a nonnucleoside analog as part of their treatment. Subjects were randomized to receive a genotypic resistance test (TruGene, Visible Genetics) or no test (standard of care arm) prior to modification of their antiretroviral therapy. Both groups received expert guidance. The primary efficacy outcome was the percentage of subjects who achieved HIV-1 RNA levels below 500 copies/mL at 3 and 6 months.

The median baseline plasma HIV-1 RNA was 4.36 log₁₀ copies/mL in the genotype arm and 4.17 log₁₀ copies/mL in the standard of care (SOC) arm, with median baseline CD4+ T-cell counts of 264/uL and 266/uL, respectively.

In an intent-to-treat (missing equals failure) analysis, 27% of subjects in the genotype arm versus 12% in the SOC arm achieved a HIV-1 RNA level of < 500 copies/mL (P = 0.01) at 3 months. At 6 months, however, the difference was no longer statistically significant (21% vs 17%).

The Havana study was contemporaneous with the ARGENTA study (both were conducted between March 1999 and February 2000) and evaluated 326 subjects with HIV RNA levels > 1000 copies/mL [2]. Subjects had a median of 4.3 years of antiretroviral experience, with 22.6% having been exposed to a nonnucleoside analog. Subjects were randomized twice prior to modification of their current antiretroviral therapy: first, either to receive a genotypic resistance test (TruGene, Visible Genetics) or no test; second, either to receive expert guidance or no expert guidance. The primary efficacy outcome was the percentage of subjects with HIV-1 RNA levels below 400 copies/mL at 24 weeks.

The median baseline plasma HIV-1 RNA was 4.0 log₁₀ copies/mL in all four groups, and the median baseline CD4+ T-cell count was 387 cells/uL.

In an intent-to-treat analysis, 48.5% of genotyped patients versus 36.2% of SOC subjects achieved HIV-1 RNA level < 400 copies/mL (P < 0.05) at 24 weeks. There was no significant difference between the "no test" groups that did and did not receive expert advice in the intent-to-treat analysis. However, a multivariate analysis showed that subjects had a higher probability of achieving an HIV-1 RNA level < 400 copies at 24 weeks in the genotype alone arm (odds ratio of 1.92 [CI 1.16 - 3.17]) and in the genotype plus expert advice arm (odds ratio of 2.13 [1.3 - 3.53]). This analysis also revealed that subjects with a history of three or more prior antiretroviral regimens had an odds ratio of 0.3 (0.16 - 0.59) for achieving virologic suppression.

Conducted between January 1998 and September 1999, the VIRA 3001 study evaluated 272 subjects with HIV RNA levels > 2000 copies/mL and with previous exposure to two or more nucleoside analogs and at least one protease inhibitor [3]. Only 4% had been treated with a regimen that included a nonnucleoside analog. Subjects were randomized to receive a phenotypic resistance test (Antivirogram, Virco) or no test (SOC arm) prior to modification of their antiretroviral therapy. No expert advice was provided. The primary efficacy outcome was the percentage of subjects with HIV-1 RNA levels below 400 copies/mL at 16 weeks.

The median baseline plasma HIV-1 RNA was 4.18 log₁₀ copies/mL in the phenotype arm and 3.92 log₁₀ copies/mL in the SOC arm and the median baseline CD4+ T-cell count was 348 cells/uL versus 347 cells/uL.

In an intent-to-treat (observed) analysis, the results were statistically significantly different, with 59% of subjects in the phenotype arm and 43% in the SOC arm achieving a HIV-1 RNA level < 400 (P = 0.036). In another intent-to-treat (missing equals failure) analysis, this difference was no longer significant, with 46% and 34% in the phenotype and SOC arms achieving this HIV-1 RNA level. Secondary intent-to-treat/observed analyses showed that the median reduction in HIV-1 RNA level was higher in the phenotype arm and that subjects in the phenotype arm were more likely to receive two or more drugs in their subsequent antiretroviral regimen that would be considered phenotypically active. In the intent-to-treat (observed) analyses of subjects who did not have a nonnucleoside analog added to their subsequent antiretroviral regimen, 49% of subjects achieved plasma HIV-1 RNA level < 400 copies/mL compared with versus 23% who did not (P = 0.015); when a nonnucleoside analog was used in nonnucleoside analog-naive subjects, 75% of these subjects achieved plasma HIV-1 RNA level < 400 copies/mL compared with 60% who did not (P = NS).

Conducted between April 1999 and October 1999, the Narval study evaluated 541 subjects with HIV RNA levels > 1000 copies/mL and with previous exposure to one or more protease inhibitors [4]. Subjects had a median of 56 months of previous antiretroviral experience, with 26% having been treated with a regimen containing a nonnucleoside analog. Prior to modification of their antiretroviral therapy, subjects were randomized to receive either a genotypic resistance test (TruGene, Visible Genetics), a phenotypic resistance test (in-house assay), or no test. No expert advice was provided. The primary efficacy outcome was the percentage of subjects with HIV-1 RNA levels below 200 copies/mL at 12 weeks.

The median baseline plasma HIV-1 RNA was 4.3 log₁₀ copies/mL, and the median baseline CD4+ T-cell count was 280/cells uL.

There was no statistically significant difference in the proportion of patients with HIV-1 RNA < 200 copies at 12 weeks among the genotype arm (44%), the phenotype arm (35%) and the SOC arm (36%). Secondary analyses showed no difference among the study groups for the proportion of subjects achieving < 20 HIV-1 RNA copies/mL. There were no differences when analyses were performed with missing values considered as failures. In a secondary analysis of only those subjects with a first PI failure, response rates were higher across all groups and were statistically significantly different for the genotype group (65%) compared with the phenotype group (45%) or the SOC group (45%; genotyping versus standard of care, P = 0.022).

The recently published results of these four studies provide conflicting data about the clinical utility of resistance testing when modifying a virologically failing antiretroviral regimen. Using the primary endpoint in the NARVAL study of an HIV-1 RNA level below 200 copies/mL at 12 weeks, neither genotyping nor phenotyping offered a statistically significant advantage in the entire study group. Not surprisingly, the results were different in a secondary analysis of those subjects with less antiretroviral experience: in this population, genotyping provided an advantage. Yet access to a phenotypic test was still not associated with superior virologic response when compared to standard of care.

In contrast, in the VIRA 3001 study, the authors reported an advantage in those subjects who received a phenotype compared to those who received standard of care when an intent-to-treat (observed) analysis was performed. This difference did not hold up, however, in a more rigorous intent-to-treat (missing equals failure) analysis. The VIRA 3001 study population was also less experienced than the populations evaluated in the other three studies. Thus, the clinical utility of phenotyping is still not consistently demonstrated.

With regard to genotyping, the findings of the ARGENTA study were similar to the findings of the GART and VIRADAPT studies: benefit associated with genotyping was transient-seen at three months and gone at six. In some contrast, the Havana trial showed benefit in the genotyping group at 24 weeks. This trial also showed, in contrast to Narval, a significant benefit with resistance testing in the more antiretroviral-experienced subpopulation compared with that in subjects with less treatment experience. The Havana trial investigators took the question that the GART study posed regarding the influence of expert advice and showed that expert advice alone was not associated with increased benefit; however, in a multivariable analysis, both genotyping and advice were helpful.

Which study's results are to be believed? The truth probably lies somewhere in-between the concluding statements of the published reports of these four trials. The influence of cross-resistance and available therapeutic alternatives-especially the nonnucleoside analog class-in patients failing protease inhibitor-containing regimens cannot be underestimated. Neither can the expertise of the managing clinician in understanding these issues be ignored. As the Havana group concluded:

"the lack of significance of the interaction between genotype and expert advice in the multivariate analysis and the current limitations in the ARV drug arsenal to overcome resistance once it is acquired…suggests that the impact of expert committee recommendations on our study results could be primarily due to their overall experience in the management of HIV infection [2]."

And even if there is short-term virologic benefit associated with one or more resistance testing strategies, the long-term impact on CD4+ cell count and clinical outcome of changes driven by the results of resistance testing remains truly unexplored.

References

Cingolani A, Antinori A, Rizzo MG, et al. Usefulness of monitoring HIV drug resistance and adherence in individuals failing highly active antiretroviral therapy: a randomized study (ARGENTA). AIDS. 2002; 16:369-379.
Tural C, Ruiz L, Holtzer C, et al. Clinical utility of HIV-1 genotyping and expert advice: the Havana trial. AIDS. 2002;16:209-218.
Cohen CJ, Hunt S, Sension M, Farthing C, Conant M, Jacobson S, Nadler J, Verbiest W, Hertogs K, Ames M, Rinehart AR, Graham NM. A randomized trial assessing the impact of phenotypic resistance testing on antiretroviral therapy. AIDS. 2002;16:579-588.
Meynard JL, Vray M, Morand-Joubert L, Race E, Descamps D, Peytavin G, Matheron S, Lamotte C, Guiramand S, Costagliola D, Brun-Vezinet F, Clavel F, Girard PM. Phenotypic or genotypic resistance testing for choosing antiretroviral therapy after treatment failure: a randomized trial. AIDS. 2002;16:727-736.
Baxter JD, Mayers DL, Wentworth DN, et al. A randomized study of antiretroviral management based on plasma genotypic antiretroviral resistance testing in patients failing therapy. CPCRA 046 Study Team for the Terry Beirn Community Programs for Clinical Research on AIDS. AIDS. 2000;14:F83-F93.
Durant J, Clevenbergh P, Halfon P, et al. Drug-resistance genotyping in HIV-1 therapy: the VIRADAPT randomised controlled trial. Lancet. 1999;353:2195-2199.