Genotypic resistance testing in individuals with an 'undetectable' viral load may be more reliable than currently thought, according to a report published in the September issue of the Journal of Acquired Immune Deficiency Syndromes. They argue that resistance testing in those with very low detectable HIV RNA levels (e.g. between four and 75 copies/ml) may be useful to inform treatment decisions in individuals who are experiencing viral load 'blips' or an unexpectedly slow response to a change in their antiretroviral regimen.
Current treatment guidelines suggest that standard genotyping resistance tests require a viral load of at least 500-1000 copies/ml to reliably provide a result.
However, a recent report from London's Chelsea & Westminster Hospital suggested that genotypic resistance testing in individuals with viral loads above 50 copies/ml was around 70% successful.
Investigators from Stanford University now argue that genotypic resistance testing is not only successful at very low viral loads, but also may have clinical utility.
During the past eight years, they have been performing genotypic resistance testing on plasma samples with HIV RNA levels that were below the level of quantification (i.e. 'undetectable') using the standard measurements of either Roche's Amplicor (50 copies/ml) or Bayer's Versant (75 copies/ml) HIV RNA assays.
In a brief report, they describe their experiences with patients enrolled in the Kaiser-Permanente Medical Care Programme in Northern California between July 1998 and August 2005.
During this time 122 samples from 116 individuals that had 'undetectable' viral loads (i.e. below 75 copies/ml) using the Versant assay were tested for genotypic resistance. This represented 3.4% of the total number of samples tested for genotypic resistance during this period (n=3631).
To exclude the possibility that the undetectable Versant measurements were wrong, 20 of the 122 samples were retested using the more sensitive Amplicor assay. Of these 20 samples, five had HIV RNA levels below 3-4 copies/ml, eight had HIV RNA levels below the standard limit of quantification (i.e. less than 50 copies/ml), and six had HIV RNA levels between 50 and 120 copies/ml.
They then phylogeneticallly analysed current and previous genotypic results from the 50 patients who had previously undertaken genotypic resistance testing. They did this in order to exclude the possibility that the current samples were contaminated or had been mixed up. Their phylogenetic analysis found that former and current samples from 49 of these 50 patients appeared to be genetically related, and they excluded the one patient whose samples did not appear to be related from further analyses.
Of these 49 remaining patients, a treatment history that included previous viral load levels were available for 45.
Doctors had ordered drug resistance testing for 20 patients after their viral loads had rebounded to detectable levels and then returned to 'undetectable' levels (also know as viral load 'blips'). Drug resistance testing was ordered for seven patients whose viral loads had declined slower than expected after an antiretroviral regimen change. A further 18 patients were suspected of having poor adherence prior to their 'undetectable' viral load results.
In all of these cases doctors wanted to find out whether resistance had accumulated despite low-level viraemia.
The investigators then describe the clinical outcome of these 45 patients after having been informed by the resistance tests that had been done when their viral loads were 'undetectable'.
For 23 of the 45 patients, their subsequent viral load levels remained below 75 copies/ml without a change in antiretroviral therapy.
A further five patients changed their antiretroviral therapy after receiving the genotype results and their viral load subsequently remained below 75 copies/ml.
For thirteen patients who continued on the same regimen, viral load levels subsequently rebounded to detectable levels.
A further two patients discontinued therapy, and viral load levels were not available for the remaining two patients.
The investigators then inputted the resistance test results obtained when the patients had an 'undetectable' viral load into the Stanford HIVdb genotypic resistance interpretation system, which is able to predict sensitivity to specific antiretroviral drugs.
Among the 28 patients who had a sustained viral load after being informed by the results of resistance testing done when they had an 'undetectable' viral load, twelve had majority virus populations with intermediate or high-level resistance to one (n=6), two (n=4), or three or more drugs (n=2).
However, just five of these 28 patients had changed therapy after being informed by the results.
The investigators write that “it is not unusual to sequence virus from plasma samples reported to have virus levels below the level of quantification and demonstrates the validity of these sequence results in nearly all patients (49/50) in whom a previous genotype was available for comparison.”
They argue that the patient population that are most likely to benefit from this are those experiencing viral load 'blips' and “patients with a slow virologic response to a preceding change in antiretroviral therapy”, they write. “In each case, the clinician's rationale for ordering a genotypic resistance test was to determine whether the virologic rebound or slow virologic response was caused by drug resistance that would require a change in therapy.”
They add that “[a]lthough this study does not prove the clinical usefulness of genotypic resistance testing on samples with low-level viraemia, our data in combination with other reports showing a high sensitivity of sequencing protease and [reverse transcriptase] in plasma samples with less than 1000 copies/ml suggest that the role of resistance testing in patients with virus that is detectable but less than 1000 copies/ml should be reassessed.”
Mitsuya Y et al. HIV-1 drug resistance genotype results in patients with plasma samples with HIV-1 RNA levels less than 75 copies/mL. JAIDS 43(1): 56-59, 2006.