Schlomo Staszewski delivered a talk at the Glaxo Wellcome sponsored symposium on Monday evening following the first day of the conference. In his talk on --What is Treatment Failure?-- he related a rationale for the primary endpoint of HIV therapy being below 50 copies/ml: maximal suppression of viral replication is most likely to delay the emergence of drug resistant virus. He proposed-- This is based on studies of protease inhibitors and NNRTIs. Is this applicable to all therapies? The following data he presented suggests a hypothesis that abacavir (ABC) may be different. Possibly, for some individuals taking a triple nucleoside ABC-based regimen if their viral load nadir (lowest point reachyed) on therapy is reduced to low levels but not below 50 copies/ml, they may be ble to maintain viral load for a prolonged period of time at low detectable levels below 5000 copies/ml. And clinical HIV progression can be held in check if viral load remains 5000 copies/ml.

Staszewski used data from CNA3005, which compared IDV + Combivir to ABC+Combivir, to make a case that failure of ABC regimen may lead to viral load remaining between 50-400 for extended period of time. However, other doctors have reported some individuals on a PI regimen have similar experience where their viral load rebounds above 400 but remains at a relatively low level of detection which could be several thousand.

At week 48 in CNA3005, the proportion of patients below 50 copies/ml was comparable in both arms (IDV & ABC)—about 45% (ITT) with baseline viral load 10,000 to 100,000 c/ml had below 50 copies/ml. In sub-group of patients with baseline viral load above 100,000 c/ml at week 48 there was a significant difference between the two treatment arms—46% in IDV and 31% in ABC arms had below 50 c/ml.

In an analysis comparing viral load rebound rates did not detect differences between the two treatment arms. The difference in maximal suppression (below 50 c/ml) between the IDV and ABC arms did not translate into higher rebound rates. The rebound rate to above 400, 1000, and 5000 was comparable in two treatment arms.

In the IDV arm 10% of persons reaching nadir viral load below 50 c/ml rebounded above 5000 c/ml at 48 weeks, while at week 48 90% with nadir above 400 c/ml rebounded virologically above 5000 c/ml. By week 8 70% with nadir above 400 had rebounded to over 5000. At week 48 of study 75% of those with nadir 50-400 had rebounded to above 5000.

Compared to 75% rebound rate in IDV arm (50-400 nadir), 30% with VL nadir 50-400 rebounded by 48 weeks. Staszewski suggested the absence of significant rebound despite measurable viral replication may be specific for this regimen. A question I have—what about the other 45% (75% minus 30%)? Since they rebounded above 5000 it is difficult to appraise in advance who will rebound and who may not.

In a previously presented meta-analysis of 1488 individuals receiving NRTIs in 3TC studies those with below individuals with disease progression accumulated mainly in individuals with high viral load-low CD4 defined by above 5000 c/ml and below 200 CD4s. During observation period of 1 year "no relevant" disease progression occurred up to 5000 c/ml. Staszewski said this data allows hypothesis that VL levels up to 5000 c/ml may be tolerated without significant disease progression for a period of time.

Staszewski said we need studies of ABC regimen to see if it’s possible to delay or possibly prevent significant viral evolution and the emergence of resistance despite low level viral replication. He said clinical observations suggest this may be the case. I think, if this hypothesis has merit an ABC regimen with d4T/ddI may be more effective since d4T and ddI resistance may not evolve as easily as for AZT and 3TC. He said, I

In the future different regimens may be associated with different definitions of treatment failure. Staszewski went on to show data that there were few virologic failures in CNA3005. Most "failures" were due to discontinuations due to adverse events and other reasons besides virologic failure. This suggests to me at first blush that more tolerable regimens will promote individuals ability to adhere.

Staszwski went on to show more data suggesting treatment "failure" is more likely due to tolerability or adherence than virologic failure. Treatment strategies should consider this element, he said.

He showed data from the "recently presented" I.C.O.N.A Italian cohort study of 805 treatment naïve individuals who received HAART. During the median follow-up was 41 weeks (1-89), 35% of patients discontinued therapy. The main reason for discontinuation was toxicity: 21% toxicity; 4.6% virologic failure; 6.3% non-adherence; 2.4% other; 63% continued on therapy.

The final data he showed was a comparison between ICONA, CNA3005 and DMP-006 (EFV+AZT/3TC).