Durban World AIDS Conference
July 9-14, Durban, South Africa

REPORT 34

"Viral Load "Blips": Reported by Jules Levin & Mike Norton

The natural history and clinical significance of intermittent virological "blips" in patients who attain an initially undetectable viral load (VL) on HAART

The authors of this study found CD4 increases were not as good in people with viral load blips as compared to those remaining undetectable. Easterbrook, Ives & Gazzrad from London, UK reported on this study of the prognostic significance of intermittent  virological breakthrough or "blips" in patients with an otherwise  undetectable VL on HAART remains unclear. We compared the outcome of patients whose VL remained below detectable limits (BLD) after  initiation of HAART compared to those who experienced intermittent  VL blips.

Eligible HIV + patients were those who had received an initial PI or NNRTI containing regimen between 01/01/96-31/12/98,  and had a viral load below the level of detection within 6 months of starting HAART. Patients were categorized into those whose VL remained BLD (£ 400 copies/ml)  and those who had at least 1 VL blip (>400) during follow-up.  Outcomes were % with VL BLD, median CD4 cell count rise at 12 and 18 months following HAART initiation and change or discontinuation  of the initial HAART regimen.

Of 860 patients identified from 3 London HIV centres  Easterbrook performed a preliminary analysis on 137 patients with „ 12 months follow-up after their initial VL BLD. The VL remained BLD in 77 (56%) and 60 (44%) had at least 1 VL blip during follow-up, and the level ranged from 401-1000 copies/ml in 30 (50%); 1001-5000  in 14 (23%); to >5000 in 16 (27%). The median time from undetectability to the first blip was 6 months.

At 18 months from HAART initiation, 64% of those with VL blips had a VL BLD. The median CD4 cell count rise at 12 and 18 months was significantly lower in those with VL blips than those whose VL remained BLD (133 cells vs 182,  p = 0.03 and 155 cells vs 269, p = 0.001). A similar % of patients  in the 2 groups (63%) changed or discontinued their initial HAART  regimen over a median follow-up of 14 months.

The authors concluded that this lends support to the adoption of a more pro-active  approach to treatment intensification in these patients.

In a follow-up Easterbrook forwarded data to me.

The cumulative percent with sustained VL>400 copies/ml was:

The median CD4 cell count change was:

Here is a report on viral load blips from the NATAP Summaries of the Resistance Workshop posted to the NATAP web site.

Transient Relapses or Blips in Viral Load in Patients on HAART

At the Workshop AMJ Wensing (University Medical Centre, Utrecht, The Netherlands) reported on a study regarding the association between viral load "blips" above detection and resistance and viral load rebound. In patients who achieve plasma HIV RNA <50 copies/ml during HAART, transient relapses ("blips") of the plasma viral load to levels above 50 copies/ml are regularly observed in some individuals. The goal of this study was to determine the mechanisms underlying these transient relapses of the plasma (blood) viral load during HAART.

Fifteen patients with a transient viral load relapse during HAART (3 or 4 drugs) were selected. The regimens of all patients included 3TC. All patients achieved viral load <50 copies/ml prior to relapse. Using an ultra-sensitive sequencing approach, the presence of resistance mutations was determined at the moment of relapse in both the protease and RT genes. The PCR product was sequenced according to the Big-dye-terminator protocol (Perkin-Elmer). Sequence analysis was performed using the ABI automated sequencer.

The median plasma viral load at the time of relapse was 72 copies/ml (range 50-1253). The genotype of the HIV plasma virus could be determined in 11 of the 15 patients. Primary mutations NRTI, NNRTI or protease inhibitors were seen conferring resistance to 1 or more of the administered drugs were observed in 8 of 11 patients, all of whom had the M184 3TC mutation in thr RT gene. Some of the 8 patients were treatment naÔve and some were experienced prior to the regimen they were taking at the time of this study. The median duration of follow-up after the relapse was 12 months. Failure of HAART defined as plasma levels above 50 copies/ml ocurred in only one patient after the transient relapse. Some patients were treatment naive and some were not.

The authors concluded that 2 mechanisms account for the blips during HAART. In approximately half of the cases resistant viruses are selected, indicating that viral replication occurred due to incomplete suppression, probably caused by a temporary decrease in the active drug concentrations. In the remaining cases, blips were due to the production of wild-type viruses most likely caused by activation of pre-therapy infected memory cells. In both cases a transient relapse of the plasma viral load did not preclude successful inhibition of viral replication <50 copies/ml by HAART therapy for at least 1 year after the relapse. As I previously mentioned, I spoke with one of the authors of this study and additional researchers. Their instinct was that viral load might rebound given enough time. I think the question is not answered until individuals are followed for a much longer period of time.

Here is the second study on this topic reported at the Workshop and previously reported in NATAP reports on this meeting.

Intermittent Viral Load Blips and Viral Load Rebound

Diane Havlir delivered an oral presentation  on intermittent viral load blips (HIV RNA 50-200) and if they are predictive of virologic rebound (>200 copies/ml) in patients receiving initial combination therapy.

She defined intermittent viremia, for this study, as HIV RNA >50 copies/ml with a subsequent measure <50 copies/ml after confirming virologic suppression (<50 copies/ml) with 24 weeks of therapy with indinavir, AZT and 3TC. Virologic rebound is 2 consecutive HIV RNA >200 copies/ml. (ACTG 343 maintenance failures were excluded). The median duration of therapy of the patients in the study they used for this analysis (ACTG 343) was 84 weeks. In ACTG 343 they found:

• intermittent blips (>50 c/ml) in 96/241 (40%)

• intermittent viremia (>200 copies/ml) in 47/241 (20%)

• greater than 1 episode (>50 c/ml) in 24/241 (10%)

• 2 consecutive HIV RNA (>50 c/ml) in 32/241 (13%)

Predictors Of Intermittent Viremia (>50 c/ml)

Baseline HIV RNA and maintenace therapy were predictors of blips. Baseline resistance at 215 position and time to <200 c/ml HIV RNA were borderline significant in predicting blips.

They used a modified Roche PCR viral load test with a low detection limit of 2.5 copies/ml. They found that median viral load was higher in subjects with intermittent viremia (23 copies/ml) compared to those without intermittent viremia (P<0.001). Only 8% of RNA measures in patients with intermittent viremia were <2.5 copies/ml compared to 52% of patients without intermittent viremia (P=0.013).

Havlir concluded that in this study population in 343 intermittent blips did not lead to viral rebound within the time frame they looked at (84 weeks): 9/96 (9.3%) of patients with intermittent viremia had viral rebound; 20/145 (13.8%) with viral suppression had viral rebound. Baseline HIV RNA was the only predictor of virologic failure in the model they used.

They also looked patients in the Merck 035 study in which patients received indinavir+AZT/3TC. Again they found those with intermittent blips had higher median RNA (7.3 copies/ml) than those with suppressed viremia (2.5 copies/ml). But individuals with suppressed viremia also had blips, its just that they were blips were <50 copies/ml. In this patient group with a median duration of 4.5 years of observation 0/6 patients with intermittent viremia had viral rebound, and 0/7 patients with viral suppression had viral rebound.

Intermittent viremia (>2.5 copies/ml) was present in all patients treated for as long as 5 years. One resistant researcher I spoke with at this meeting found resistant virus when he observed blips. This resistant virus was not present before therapy. He and I think it's intuitive that given enough time viral rebound will occur if there are blips. Havlir's study extends to 5 years in small numbers of individuals.  

Why do patients have persistent intermittent viremia? Havlir said it could be ongoing infection, latently infected cells, and sanctuaries.

Here is Durban report on blips from earlier contribution by Mike Norton for NATAP Summaries, also posted to web site at NATAP Durban Summaries.

INTERMITTENT VIRAL LOAD BLIPS BETWEEN 50 to 400 COPIES/ML

Doug Ward and colleagues had a poster titled ìThe significance of low-level Viremia in patients with previously undetectable HIV-1 RNA levels.î  They presented 32 patients who had had 2 previous viral loads of <50 copies/ml, who then presented with a viral load between 50 and 400 copies/ml.   Upon discovery of these viral load blips, patients were asked to repeat the viral load on follow-up visit.  On subsequent viral loads, without change of therapy, 75% returned to <50 copies/ml, all but 2 patients of those on the very next viral load.  25% of the 32 patients who blipped did not return to undetectable or <50 copies/ml.  The authors concluded that adherence can play an important role in why patients blip.   That if a patient were on his/her first regimen, they are more likely to return to <50 copies/ml with simply adherence advice and the continuance of the same antiretrovirals.  This is important given the limited options available once a patient switches regimens.  On the other hand, those that were already on their 2^nd or 3^rd regimen and blipped, were most likely not to return to <50 copies/ml.  Once again showing that that initial regimen is the regimen most likely to deliver durability of viral suppression.

DOES VIRAL LOAD HAVE TO BE BELOW 50 COPIES/ML?

Jeff Greene and colleagues in New York reached once again into their large observational database to share an interesting item that runs contrary to prevailing wisdom.  They presented 95 patients who were divided into 2 groups starting in 1998 when the more sensitive HIV RNA assays (ultrasensitive assay <50 copies/ml) were implemented in their clinic.  In group I, 54 patients were below the level of quantification for by both the 1^st generation and 2^nd generation assay, meaning their viral loads were <50 copies/ml.  In group II, 41 patients were detectable by the 2^nd generation assay but not by the 1^st generation, meaning their viral loads were between 50 and 400 copies/ml.  The groups were similar when looking at age, race, risk factor, time from HIV diagnosis, sex, time on HAART, and the total # of HAART regimens they had been on.  The results show that there was no difference among the two groups over the examined period of approximately 2 years when comparing mean change in cd4ís, incidence of OIís, and death rates (there was no HIV related deaths in either arm).  Most strikingly, there was no difference in virologic failure, defined as 2 consecutive viral loads > 400 copies/ml, between the two groups.  The authors concluded that optimal clinical response may not require viral suppression to < 50 copies/ml. 

Commentary: These observations, while real, are in conflict with the body of work that exists showing that < 50 copies/ml lends to a more durable virologic response when compared to those who nadir is between 50 and 400 copies/ml.  Perhaps with more time those not < 50 copies/ml will virologically fail at higher rate, but it is equally conceivable that if sustainable between 50 and 400 copies/ml of HIV RNA will translate into equally efficacious clinical outcomes.