Antiretroviral Therapy Highlights from the 7th Conference on Retroviruses and Opportunistic Infections by Keith Henry, MD with Editing by Jules Levin

The NATAP web site has about 50 reports from the Retrovirus Conference (

Dr Henry's commentary of the Conference reflects his unique perspective to antiretroviral therapy for HIV. He questions the orthodoxy on when to begin treatment and with what, and when to intervene with salvage and with what. He has been raising these questions for about two years. In the February 15 2000 Annals of Internal Medicine, Dr. Henry and Oren Cohen, MD, offer their opinions with supportive arguments on when to begin therapy and with what.

"The Case for More Cautious, Patient-Focused Antiretroviral Therapy", by Keith Henry, MD--

An Editorial article written by Oren Cohen, MD, NIH, containing commentary on Henry's article is also in the same issue of Annals of Internal Medicine--

"Antiretroviral Therapy: Time To Think Strategically"

This summary represents interpretation of data presented at the 7th CROIC from my perspective as a clinical researcher and clinician caring for over 430 mostly challenging inner city HIV+ persons. The topics I will touch on include: studies in treatment naÔve and experienced patients, salvage therapy, strategy studies, metabolic issues, and the new drug pipeline.

Studies in treatment naÔve patients- it is my view that new drug combinations used in this population should have one year success rates (defined as % < detection level with an intent-to-treat design) of at least 80-90% to possibly represent a true advance in the field. A meta-analysis of 22 studies (#519) found @ 48 weeks overall 55% < 400 copies/ml and 46% < 50 copies. The success rates for combinations using PIs, NNRTIs, or triple RTIs was similar. The COMBINE study (#516)presented 24 wk data comparing ZDV/3TC + nevirapine or nelfinavir (CD4 average= 355; RNA= 4.8 log, n=142 patients).

By intent to treat analysis 50% in the NVP arm < 20 copies/ml versus 33% NFV arm while the on treatment analysis found 80% NVP and 45% NFV arm < 20 copies. These early results suggest that the NNRTI arm can perform as well or better than the PI arm though the overall results are generally still sub-optimal. More impressive results were achieved in the ABT-378/r study (#515) wherein @ 72 weeks by intent-to-treat analysis 82% < 400 and 80% < 50 copies RNA/ml. The on treatment analysis found 98% < 400 and 96% < 50 copies/ml. That level of success is more of what we should expect from treatment though it is disconcerting that14% had quite elevated lipid levels.

Studies in NRTI experienced patients- the three year follow-up data for a pivotal ritonavir/saquinavir trial (# 533) was presented. At three years, by intent-to-treat analysis 55% < 200 copies RNA/ml while by the on treatment analysis the results were a robust 88% < 200 copies/nl. Lipid and fat distribution abnormalities were common in this group of patients with the risk going up with increased age or use of nucleoside reverse transcriptase inhibitors. Results from ACTG 370 (525) involving patients with prior use of D4T, ddI or double NRTIs adding indinavir +/- delavirdine (DLV) found @ 48 wks that the DLV arm performed better (by intent-to-treat analysis 70% < 50 copies RNA/ml versus 40%). However, since the average CD4 count = 500 the question would be was there any need to use up the PI and NNRTI class when the risk for AIDS related events was so low. ACTG 364 (#531) also looked at adding a PI (nelfinavir=NFV) or an NNRTI (efavirenz=EFV) or both to a group with detectable virus on NRTI therapy (average RNA= 7776 and average CD4 count= 350). At 48 weeks, 67% of the EFV + NFV versus 44% of the EFV versus 22% of the NFV group had < 50 copies RNA/ml. Again, since that group had a very low overall short-term risk for AIDS related events the question is whether utilizing all of the remaining classes of drugs is best from the long-term clinical perspective. Preliminary data suggests that their may have been differences in baseline genotypic NRTI resistance between the three treatment groups. This may have affected the virologic response rates between the three arms. So, until we have more conclusive resistance analysis we may want to withhold drawing conclusions from this study.

Studies in protease inhibitor experienced patients- in some limited data presented (ACTG 373 # 526) suggested that patients experiencing virologic rebounds while on amprenavir could by ITT @ 48 weeks 59% < 500 copies/ml utilizing indinavir + nevirapine + 3TC/AZT. Using an as-treated analysis 78% had <500 at week 48. Patients were either taking amprenavir monotherapy (n=36; 64%) or amprenavir+AZT/3TC (n=18; 32%). Again, since the average CD4 count= 346 and RNA= 4.19 log, the question is whether use of all available classes at that stage is in the best long-term interest of the patient. A key factor in the successful suppression of viral load after the treatment switch may be that the treatment change was made quickly--HIV-RNA was 15,488 copies/ml, 5 patients had HIV-RNA <500. This was one of the first studies suggesting that response to a second PI regimen could be successful if the switch in treatment was made soon after failure. An additional factor in the success of the patients after switching could be that amprenavir's resistance profile appears to be unique. As a firstline therapy, it appears to lend itself well to being able to suppress viral load back down to undetectable after virologic failure. As well, its unique resistance profile appears to lend itself well in a second-line rescue regimen. However, at a dose of 1200 mg twice daily the pill burden is difficult (8 pills per dose). As well, the GI-related side effects of amprenavir can be difficult. In ACTG 373, 8 (14%) dis-continued the study prematurely, and 11 (20%) experienced severe or greater intensity ( grade 30 signs/symptoms.

Use of ABT-378r + a NNRTI in first PI failures (# 532) found @ 48 weeks 70% < 400 copies and 60%<50 copies/ml. Although these results are among the best salvage results seen to date (25% had significant increases in lipid levels) the question again is what is the best stage to intervene for the best long-term clinical outcome considering the failure of this regimen would likely leave very few options open for subsequent regimens. Results of two other salvage trials (ACTG 359 # 235 and ACTG 398 # LB7) generally found high levels of toxicity and generally disappointing virologic results again highlighting the difficulty to achieve high level suppression with a reasonably well tolerated regimen after patients have virologically failed a potent PI such as indinavir.

The state of salvage therapy year 2000- It is still not clear to me how well resistance tests will provide long-term clinical benefit in this setting. Results of the VIRA 3001 study (#237) found @ 16 weeks 38% of patients who had therapy guided by phenotyping < 400 copies versus 23% without. Results using MEGA-HAART (using up to 8 drugs # 536) finding @ 48 weeks 40% < 400 copies RNA/ml provide some hope that some activity can be squeezed out of aggressive use of the available drugs but the ability of most patients to tolerate such regimens is questionable. The modest short-term virologic benefits seen with resistance testing need to be scrutinized as to whether there indeed is long-term clinical benefit. Patrick Yeni (#S33) estimated that there may be generally a 70% success rate for re-suppression after the first viral rebound and only a 30% success rate after the 2nd rebound. He reasonably stated that the goal early in treatment is high level suppression (thus failure = return of even low level of virus) but that later failures may need to rely more on the CD4 count or clinical parameters. In other words, in a person with multiple treatment failures, Yeni suggests the maintenance of a relatively high CD4 may be a reason for continuing a person on a regimen despite detectable HIV-RNA. In general, short-term virologic results are worse if the CD4 count is lower, the RNA level higher, or if drug levels are lower arguing for earlier intervention after virologic failure. However, if improved drugs and more compact regimens are likely to be available for a given patient situation within a clinically relevant time-frame, then trying to avoid resistance to yet unused drugs or classes of drugs may be a better approach than using the best possible regimen immediately.

Treatment strategies- intensification of a failing regimen either by improving pharmacokinetics by adding ritonavir to indinavir (#534) or adding the nucleotide RTI tenofovir (#740) seemed to be beneficial enough to warrant more efforts in how to selectively intensify a regimen without a wholesale overhaul of the drugs. As well, previous to this Retrovirus conference studies have suggested abacavir is a good candidate for intensification of a first-line regimen. Switching a successful regimen in order to minimize metabolic complications or to preserve the use of a PI for later was the subject of numerous studies. The best of the switch studies (#457) compared a switch to abacavir versus continuing the PI. There were improvements in the lipids in the abacavir group and comparable virologic success rates. Patients in the abacavir group who did experience virologic rebound successfully re-suppressed on a PI based regimen. NATAP's detailed review of 8 switching studies reported at Retrovirus (4 EFV, 2 abacavir, and 2 NVP) is available in the Retrovirus Summary Report on our web site. The particular article can be located at the following address-

Most of the switch studies were based in Europe or Australia which suggests to me that investigators there have been more quick to recognize the importance of the metabolic concerns and long-term strategy approaches . The story from various treatment interruption studies remains intriguing but unclear. There seems little doubt that viral rebound is seen in a fairly uniform manner once therapy is stopped. What is not clear is whether, after several cycles of start and stop, there is any return of more effective anti-HIV immunity. The data is too limited at this time to comment further. The most interesting observations came from Steve Deeks (#256) who looked at a group of 18 patients (RNA= 4.6 log and CD4=256) with long-standing PI failure. After stopping therapy @ 12 weeks there was an average loss of 94 in the CD4 count and an increase of 0.82 log RNA level. After about 8 weeks there was a shift in the viral fitness profile from an impaired virus (possibly related to drug resistance mutations) to a more fit "wild-type" virus and possibly an acceleration in the deterioration of the surrogate markers. Those observations make it difficult to know what the role of treatment interruptions will be in the setting of salvage.

Metabolic issues- there was more of an emphasis on the metabolic issues at this meeting than the actual complications of AIDS which is a sign of where therapy was been and where it is going. There is increasing data that the problem is multi-factorial. The data presented regarding risk for atherosclerosis suggest that the concern is real but not so overwhelming that any major change in approach to treatment is necessary for now. The pharmacokinetic interaction study of some PIs with the statins (#LB6) highlights the potential problems with a poly-pharmacologic approach to managing drug related metabolic problems. The NRTI class is being linked to lactic acidosis possibly through mitochondrial toxicity (#55,56,57). My view is that persons at higher risk for lactic acidosis (women, overweight, pre-existing liver disease, long-term NRTI use, ? D4T) need to be asked about symptomatolgy and screened with a properly collected lactic acid level and liver enzymes on a periodic basis (i.e. every 6-12 months) until more is known about this problem. David Cooper (#S21) believes that the NRTI class is associated more with lipo-atrophy and the lactic acid syndrome while the PI class is associated more with fat accumulation and lipid elevations. Data was presented suggesting that rates of lipo-atrophy might be lower in African-Americans and in men. Other factors such as increased age, time on therapy, and magnitude of viral suppression appear to play a role. One interesting question is whether the linkage with magnitude of viral suppression may also relate to more drug exposure with attendant more toxicity and effectiveness at the same time. The new kid on the block for concerns is osteoporosis (# 45 and 46). More data is needed at this point before any conclusions can be reached about causality.

New drugs- Joe Eron did a good job summarizing the new drug pipeline (#S34). This meeting highlighted the scientific foundation for new drug classes that will impact HIV integrase, chemokine receptors, and fusion inhibitors. However, only the latter class will likely be helpful to patients over the next several years so I wouldnít count on any white knight drugs to fix major problems. Whether the fusion inhibitors T-20 or T-1249 can move from proof of concept drugs to providing clear clinical benefit will have to await the results of studies over the next several years. Two drugs from old classes that caught my attention were DAPD (#668) and BMS-232632 (#504). DAPD is a NRTI with some activity against NRTI resistant virus and BMS-232632 is a convenient 2-3 pill once per day protease inhibitor that looks reasonably active in the initial trials. Although we need new drugs and there are many drugs in the pipeline, most represent drugs from the same classes as we now have available. Patients and clinicians need to optimally use the available drugs without counting on a new class or super drug to save the day (though that would be nice).

How to optimally use the available drugs- the debate about when to start therapy seems to be heating up again. Two presentations (#522 and 521) supported the early use

HAART. Another study (# 523 by Tebas et al including me) looked at a mathematical model over a 10 year period utilizing generally optimistic results from clinical trials. Even using data from # 521 and 522 we would conclude that patients would do better over the long-term if HAART were initiated in a more conservative manner ( we know there are significant failure rates to initial therapy and that salvage therapy is sub- optimal). A very telling study was presented by Margaret Fischl who looked at success rates with HAART in the setting of prisons with directly observed therapy (DOT) versus the clinical trial setting (#71). At 48 weeks the % suppressed < 400 copies was 100% in the DOT group versus 68% in the clinic setting with fewer adverse events in the DOT group. Those results were seen even though the DOT group had higher RNA levels and lower CD4 counts than the clinic group. That study clearly shows that our current drugs can be remarkably more effective if used optimally.

My overall impressions of the 7th CROIs- this conference continued the tradition of quality presentations and good science. The fields of vaccine development and viral surface receptors seem poised to provide clinically meaningful data in the near future.The conference was well organized with plenty of opportunity to view the posters and interact with colleagues. The future of audio-visual presentations was evident with 4 or 5 different laptop computers with either Mac or IBM platforms being utilized in one session (not without glitches but impressive nonetheless). The new drug pipeline seems relatively modest so outcomes with the currently available drugs will remain the lynch pin of the approach to using antiretroviral therapy. The attention given to side effects of long term therapy and the generally disappointing results of salvage studies underscores the need to learn how to best employ the current drugs over a period of time to be measured in decades. Bill Powderlyís summation of the metabolic symposium (#S24) seemed to hint at a possible paradigm shift towards a more judicious and strategic approach to therapy. It seemed to me that European and Australian investigators have more readily grasped the need for more creativity in approaching long-term therapy since studies based on those continents were prominent at this meeting. The interest in immune-based therapies also seems to be increasing. Although the problems facing the field are many, the opportunity to help protect patients against AIDS related complications over a time period measured in decades seems to be greater than ever. A disheartening undertone to the meeting was the number of HIV specialists whose practice is threatened by low reimbursement. Shifting resources from over utilization of drugs and laboratory tests to better support for HIV specialists and the team helping them (i.e to maximize adherence in the patients at greatest risk for AIDS complications) would seem to be a high priority in order to cost-effectively optimize patient outcomes.

Keith Henry, MD
HIV Program, Regions Hospital and the University of Minnesota AIDS Clinical Trials Unit
St. Paul, Minnesota
February 13, 2000

When to begin therapy; excerpts and comments from articles written by Keith Henry, MD, and Oren Cohen, MD, in recent medical journals