translation agency

Gay Men's Health Crisis
The Great Salvage Therapy Drug Juggle
Jill Cadman
April 1, 1998
GMHC Treatment Issues 1998 Apr 1; 12(4): 1

Estimates of the number of people on protease inhibitors who fail to achieve stable viral loads below quantifiable limits (less than 400 copies/ml by PCR) range as high as 70% in pretreated populations. Those who do not reach this plateau face the daunting prospect of either continuing on a regimen that allows viral breakthrough or putting together some backup or "salvage" therapy. The chances of successfully switching therapies are dimmed by the realities of HIV's capability of developing simultaneous cross-resistance to groups of drugs as well as each individual drug's side effects, which make them intolerable for many people.

When to Switch Deciding when to switch therapy is a largely unsettled issue. The Public Health Service Guidelines for the Use of Antiretroviral Agents devotes several sections to this topic. The PHS document recommends switching if less than a 0.5 to 0.75 log (68% to 82%) reduction in viral load occurs four weeks after initiating therapy - or less than a 1 log (90%) reduction by eight weeks. The eventual goal is to achieve levels below the limit of quantification (BLQ) of the standard PCR assay (400 copies/ml) within four to six months. If viral loads do not go that low or eventually rebound above that level, the Guidelines again advise altering therapy. (The Guidelines recognize that people with very high initial viral loads may never go below 400 copies/ml. Also, some people who rebound may stabilize, at least for a time, at modest viral loads of a few thousand. For such individuals, continued monitoring may be acceptable until their HIV rebounds to a level greater than three times the lowest stable viral load measurement.) The Guidelines do not endorse the clinical use of the unapproved ultrasensitive PCR viral load assay, with its lower limit of detection of 20 to 40 copies/ml. Yet some studies (e.g., the INCAS trial of AZT/ddI/nevirapine and Merck 035, which tested AZT/3TC/indinavir) report that a viral load below 40 copies is strongly associated with long-term antiviral effect.

Driving viral loads below 400 copies/ml is frequently problematic. Many people whose viral loads have been reduced to below 400 copies really have no way to modify or intensify their antiviral regimes to restrict HIV still further because of the absence of effective treatment options.

At the International Workshop on Salvage Therapy for HIV Infection held in New Orleans on April 17-18, many of the researchers and physicians felt that switching early from a partially suppressive regimen (one that allows a low but detectable viral load) to a maximally suppressive regimen is likely to provide a better long-term outcome. The longer an individual stays on a treatment that allows some viral replication, the more likely that additional resistance mutations will accumulate. The correlation between baseline mutations and treatment response is not yet fully understood, but it does appear that the more resistance mutations present, the more likely subsequent drugs will fail.

Some workshop participants felt that an early switch strategy did not always pay off in their experience. Speaking after the workshop, Judith Falloon, M.D., of the National Institute of Allergy and Infectious Diseases (NIAID), stated that about a year ago she was very aggressive in changing treatments immediately but results were mixed. She said, "I switched a lot of people for a relatively limited benefit set." Although a viral load of 1,000 copies/ml used to be enough to trigger a change in regimen, she now waits to see if there is a definite upward trend in viral load or downward trend in CD4 count. A stable viral load in the 1,000 to 5,000 range may not warrant a change, especially if the patient does not have three new drugs available.

In addition, there have been cases of continued benefit even after viral breakthrough. At the 5th Conference on Retroviruses and Opportunistic Infections last winter, Steven Deeks, M.D., reported on 143 San Francisco General Hospital patients with CD4 count rises of 50 to 100 cells and viral loads below baseline a median of 18 months after initiation of therapy and a median of 12 months after clear evidence of virologic failure (see abstract 419 and Treatment Issues, March 1998, page 5).

Since this improvement probably will prove transitory in the end, Dr. Deeks stated that continuing on therapy after viral breakthrough should really only apply to people who do not have the option of switching to a maximally suppressive regimen. "If there is a salvage regimen that might achieve durable suppression, then it makes sense to switch early. If in all likelihood, you're only going to get transient viral suppression, then it may not be a good idea to switch because that would limit the ability to get durable suppression down the line, by burning up drugs now. This argument, in my mind, pertains mainly to the use of NNRTIs [non-nucleoside reverse transcriptase inhibitors, such as nevirapine] in salvage regimens. If someone has not been on an NNRTI, then they still have a chance of getting a very potent response to drugs in that class. However, this response will only be durable if the NNRTI is used in combination with other drugs that are effective. Often those drugs are not yet available." A similar cross-resistance issue may apply to protease inhibitors.

Data regarding the efficacy of salvage protocols are beginning to emerge from small clinical trials. Several such studies presented at recent conferences reported mixed results. Most salvage regimens seem to provide initial high-level viral suppression, but the antiviral effect does not last long in a large percentage of cases. Some studies have attempted to determine predictors of response, so that strategies with the best chance of success can be developed.

Backup for Nelfinavir Failure Finding a backup therapy for nelfinavir failure has been touted by nelfinavir's manufacturer (Agouron Pharmaceuticals) as comparatively easy. The key HIV mutation conferring nelfinavir resistance (at codon 30 on the protease gene) does not confer cross-resistance to other protease inhibitors. But the appearance of codon 30 is often accompanied or shortly followed by secondary mutations that do reduce response to protease inhibitors as a group.

At the Retrovirus Conference, Pablo Tebas, M.D., from Washington University in St. Louis, Missouri, gave an update on a small cohort of volunteers who had failed nelfinavir treatment in three phase II clinical trials (abstract 510 and Treatment Issues, Oct. 1997, pages 1-2). The participants were switched to a combination of d4T/3TC/ritonavir/saquinavir after two consecutive viral load tests above 5,000 copies/ml by Chiron's bDNA assay. All participants had taken nelfinavir for a median of 55 weeks.

Twenty-four volunteers in the substudy had little or no antiretroviral experience prior to enrolling. Dr. Tebas had 24-week data on 19 of these participants. All 19 achieved viral loads BLQ (less than 500 copies/ml by bDNA) after switching to the salvage protocol. This decrease was sustained at week 24 in 13 of the 19 (68%), five of whom were BLQ by the ultrasensitive assay (40 copies/ml by PCR). At the Salvage Therapy Workshop, Keith Henry, M.D., of Regions Hospital in St. Paul, recounted that the remaining five participants had also reached 24 weeks. According to the most recent viral load data, 17 out of 24 participants (71%) were still BLQ of the standard assay.

An additional seven participants in the crossover substudy had extensive prior antiretroviral therapy. They also had higher mean viral loads and lower mean CD4 counts than the less experienced group. Only three of the seven (43%) in this more advanced group achieved viral loads BLQ.

The crossover substudy design could not employ an early switch strategy because viral load results of the participants in the original studies were blinded (unknown to the investigators and participants). After unblinding the study results, it was determined that the median time on nelfinavir with detectable viral load was 48 weeks. An earlier change in therapies might have prevented the gradual accumulation of multiple resistance mutations.

The most frequent baseline mutations in the HIV protease gene prior to switching to ritonavir/saquinavir were at codon 30, found in 17 of 25 participants. Five of 25 participants harbored HIV with a mutation at codon 90 that sometimes appears after exposure to nelfinavir and is commonly associated with resistance to saquinavir. However, in this small cohort, the presence or absence of these mutations was not predictive of a short-term virologic response. The most significant predictor of failure was a higher baseline viral load.

The Mysteries of Saquinavir Failure Second-line therapy in the case of saquinavir failure has always been controversial because the key codon 90 resistance mutation does not by itself cause indinavir to fail. In concert with other mutations, it does contribute to indinavir resistance, though. It also may affect nelfinavir (see above). In addition, most salvage therapy studies have been conducted using the old formulation of saquinavir (Invirase). It is not possible to extrapolate these results to the new formulation of saquinavir, Fortovase, which achieves much higher blood levels of saquinavir. The low saquinavir levels resulting from Invirase may have been insufficient to cause saquinavir-resistant HIV to develop in many patients (although small undetectable subpopulations of resistant virus may have arisen). Last year, ACTG trial 333 found that switching to indinavir or Fortovase after lengthy Invirase use are not feasible salvage therapies (see Treatment Issues, Apr. 1997, page 6).

At the recent Retrovirus Conference, Jody Lawrence, M.D., of Stanford University, and colleagues described the sequential use of nelfinavir and indinavir/nevirapine in 16 participants who had received at least six months of treatment with a saquinavir-containing regimen and had viral loads above 5,000 copies/ml (abstract 422). Median baseline CD4 count was 156 cells and viral load was 16,716 copies/ml. Because all participants had received extensive antiretroviral therapy (an average of four prior agents), only 6 of 16 (38%) received one or two new reverse transcriptase inhibitors when switching to the new protease inhibitor-containing regimens.

Eleven participants experienced only a transient benefit after switching to nelfinavir. At two weeks they had a median viral load drop of 0.59 log (74%), with most participants' viral loads returning to baseline by week 12. The group then switched to indinavir plus nevirapine in combination with two reverse transcriptase inhibitors. This change decreased viral load by a median of 1.58 log (97%) after four weeks. Six of the 11 (55%) taking this third sequential protease inhibitor-containing regimen achieved a viral load below 400 copies/ml. However, only 3 (27%) maintained maximum viral suppression beyond week 20.

Dr. Lawrence stated that the total number of mutations at baseline was correlated with decreased durability of treatment response. Specifically, the AZT-associated resistance mutations at codons 67, 70 and 219 along with the saquinavir-associated codon mutation 90 were quite highly associated with rapid failure.

Indinavir Failure Rule: Switch Early Finding a backup therapy for those experiencing viral load rebound on indinavir is the most difficult problem of all. Indinavir regimens that fail to completely suppress HIV trigger the emergence of a series of mutations conferring broad cross-resistance to other protease inhibitors. Still, there is some hope.

Joel Gallant, M.D., and colleagues from Johns Hopkins in Baltimore, conducted a retrospective analysis of patients taking ritonavir/ saquinavir (plus new reverse transcriptase inhibitors in most cases) soon after reemergence of detectable HIV on an initially successful regimen (Retrovirus Conference abstract 427). Seventeen study subjects had failed indinavir-containing regimens. Twelve of the 17 (70.5%) achieved viral loads below the limit of quantification (400 copies/ml) at a median of 33 weeks (with some participants out to 50 weeks) after switching to the ritonavir/ saquinavir combination.

Four participants in this study had failed nelfinavir-containing combinations. Three of the four still had viral loads below 400 copies/ml at a median of 33 weeks after starting the salvage regimen.

Dr. Gallant feels that these participants experienced a better antiviral response than those in other studies did because they switched fairly early with a relatively low median viral load of 13,507 copies/ml. They thus avoided the accumulation of resistance mutations in their HIV and the resulting cross-resistance to new protease inhibitors. He stated, "I think that unless you make an early switch, you may be guaranteeing that patients will never get another protease inhibitor that works. In my experience, this is the only way that you maybe get a second round of protease inhibitors." Although the availability of new reverse transcriptase inhibitors for most trial participants may have helped too, Dr. Gallant feels that this was not enough to account for the durable response.

Saquinavir/Nelfinavir Gets Mixed Reviews Nelfinavir inhibits the liver's breakdown of saquinavir in a manner similar to ritonavir. Administering nelfinavir at the standard dose along with reduced doses of saquinavir yields saquinavir levels equivalent to those achieved with a standard Fortovase regimen. Nelfinavir/saquinavir may be a better salvage therapy for indinavir failure than most current possibilities because of reduced cross-resistance issues.

Stephen Kravcik, M.D., of Ottawa General Hospital in Canada, presented results of a 14-person pilot study at the Salvage Therapy Workshop. Participants had previous protease inhibitor treatment with indinavir, ritonavir and/or saquinavir. Median viral load was 84,500 prior to switching to nelfinavir/saquinavir. None of the participants were able to add two new reverse transcriptase inhibitors. This regimen did not work well as salvage therapy: there were only transient improvements in viral load and CD4 count. The three participants who experienced a good response (two had viral loads below 500 copies/ml and one had a sustained 1 log [90%] decrease in viral load) were all saquinavir-na�ve. Dr. Kravcik believes this combination only has a limited role as salvage therapy.

At the 6th European Conference on Clinical Aspects and Treatment of HIV Infection held in Hamburg last fall, a German group presented a poster on nelfinavir/saquinavir salvage therapy with better results (late breaker 925). J�rgen Lohmeyer and colleagues designed a salvage protocol using 1,000 mg of saquinavir (Invirase) and 1,250 mg of nelfinavir twice a day in conjunction with two reverse transcriptase inhibitors. The 25 participants had at baseline a mean CD4 cell count of 118 and a median viral load of 20,000. All had been previously treated with protease inhibitors. Seventeen participants completed eight weeks of treatment and experienced a mean viral load decrease of 1.8 logs (98%) and a mean increase of CD4 cells of 113. Viral load decreased to below 500 copies/ml in 10 of the 17.

Combining Three Hot Drugs is not so Hot One possible way to improve salvage therapy's chances is to administer potent new drugs that have not yet been approved and will therefore be completely new to patients regardless of treatment history. Dr. Falloon presented data at the Salvage Therapy Workshop on two salvage therapy protocols with one such combination using the three experimental drugs: efavirenz (DuPont Merck's new NNRTI), amprenavir (Glaxo Wellcome's new protease inhibitor) and abacavir (Glaxo Wellcome's new nucleoside analog). (For more information on these drugs, see Treatment Issues, Winter 1997/1998, pages 18-25.) The first trial was a small NIH pilot study of 11 participants who had failed on indinavir-containing regimens. In the initial part of the study, abacavir (300 mg) and amprenavir (1200 mg) were given twice a day. Efavirenz was added later (600 mg once a day). While using the abacavir/amprenavir combination, three of the 11 participants experienced a 1.5 to 2 log (97% to 99%) drop in viral load from baseline. The response lasted for only six and eight weeks in two of the three but went on for more than 28 weeks in the third. After efavirenz was added, there was a better response, with five of nine participants achieving at least a 1 log (90%) drop from baseline (two of the five had a short duration of response). Efavirenz also activated liver metabolism, decreasing the peak levels of amprenavir by 46% and the trough level by 59% in eight participants. When efavirenz and amprenavir were dosed together, the efavirenz levels were not significantly different than expected.

Despite this effect, the amprenavir dose was not increased for the larger salvage study, known as protocol 2007, because of the already high amprenavir pill burden (eight large pills twice a day). The open-label 2007 trial enrolled 101 participants who had detectable viral load after at least 20 weeks of protease inhibitor treatment. At baseline, median viral load was 122,851 copies/ml and median CD4 count was 160. Participants were categorized as NNRTI-na�ve or -experienced. Highly preliminary eight-week data are available for some of the participants. In the NNRTI-na�ve group, a 0.5 log (68%) viral load decrease was achieved in 81% at week two and 58% at week eight. In the experienced group, a 0.5 log decrease occurred in 53% at week two and 30% at week eight. Using a more stringent measure of benefit, a 1.5 log (96.8%) decrease or a decrease in viral load to less than 400 copies/ml occurred in 45% of the NNRTI- na�ve group at week two and in 42% at week eight. In the NNRTI-experienced group, fewer participants achieved this decrease, 18% at week two and 17% at week eight.

Only 5% of NNRTI-experienced participants versus 24% of NNRTI-na�ve participants went below 400 copies/ml at any time in the study's first eight weeks. Many of the NNRTI-experienced volunteers probably had developed NNRTI cross-resistance that affected efavirenz, which appeared to be central to the triple combination's effect. For all the participants, previously developed cross-resistance apparently presented a major challenge to both the abacavir and amprenavir. And whether efavirenz was effective or not in a patient, it lowered amprenavir levels, making this protease inhibitor still more problematic.

This regimen was also complicated by the three drugs' overlapping side effects: 47% of participants suffered skin rash or irritation, 59% had digestive complaints and central nervous system side effects were evident in 67%. Such side effects could cover up a relatively rare systemic syndrome that leads to a potentially life-threatening hypersensitivity to abacavir. However, Dr. Falloon stressed that this syndrome has only occurred upon discontinuing and then reintroducing the drug (see Treatment Issues, Winter 1997/98, page 22).

Mega-HAART There have been several accounts of salvage therapy using six or seven agents. Cassy Workman, M.D., of Australia, presented a report on 12 heavily pretreated study subjects at the Retrovirus Conference (abstract 426). Participants had failed ritonavir, saquinavir (Invirase) and indinavir as well as all the nucleoside analogs. Mean viral load at baseline was 170,065 copies/ml. Salvage therapy consisted of six drugs: d4T (40 mg twice a day), ddI (400 mg once a day), 3TC (150 mg twice a day), nevirapine (200 mg twice a day), nelfinavir (at an elevated 1,000 mg - 750 mg is standard - thrice daily) and saquinavir (Invirase, at the standard 600 mg three times a day). Nine of the 12 were able to stay on this onerous combination and achieved viral loads BLQ (400 copies/ml) at week 12. CD4 counts increased between 30 and 370 cells.

It is not too surprising that this combination should be effective, at least initially. The nevirapine and the extra-strength nelfinavir were both new, and the nelfinavir boosted the saquinavir (taken as Invirase) blood levels enough so that it was now an effective drug. Some of Dr. Workman's patients also may have still been sensitive to d4T and ddI since HIV resistance to these two is comparatively infrequent.

At the Salvage Therapy Workshop, Schlomo Staszewski, M.D., of Goethe University in Frankfurt, proposed a similar strategy that he termed "mega-HAART." He has combined six to eight drugs in 37 failing patients to strategically increase antiviral activity and plasma drug levels. Dr. Staszewski used at least three nucleoside analogs, one or two NNRTIs and two to three protease inhibitors. Thirty-two patients experienced a viral load reduction of over 2 logs (90%), 26 of 37 achieved viral loads BLQ on the standard assay at least once and 11 of 37 went BLQ on the ultrasensitive assay at least once. Dr. Staszewski is now testing maintenance therapies with a reduced number of drugs in patients whose response goes below quantification on the ultrasensitive assay.

"Henry's Best Shot at It" Meanwhile, in Minnesota, Dr. Henry is conducting an exploratory salvage strategy study in his clinic that combines information from resistance assays, viral load tests, treatment history and analysis of the protease inhibitor levels in the blood. Dr. Henry uses a computer program to devise a flow chart for each individual. He determines the major mutations and the categories of drugs that are most likely to be ineffective. He recycles d4T and ddI, adds hydroxyurea to boost the ddI and prescribes the best-looking protease inhibitor at the best possible dose. If the results are promising, Dr. Henry supplements the combination early with efavirenz or another NNRTI. He has used this strategy in about ten patients for only a short period to date, and so far, responses have been good.

He cautions that, "short-term virologic responses in the overall scheme of things is not something to brag about. If patients are suppressed and then rebound, I'm not sure what you've accomplished exactly other than burning more bridges." Dr. Henry stressed that an individual approach taking such factors as clinical status and adherence into consideration is extremely important. He added, "I'm tolerating people with less than total suppression because I think that in a year or two I will be able to offer more potent, sound regimens that will work if I don't burn through everything now. I think I have time to wait in some types of patients and don't have time to wait in others." Ongoing and Developing Trials Obviously, there is a great need for larger clinical studies and more hard data. Several such studies are either ongoing or in development. The most advanced is ACTG 359, which is a 24-week safety and efficacy study (with a 24-week extension) in 300 indinavir failures. Salvage regimens for this protocol include Fortovase plus ritonavir or nelfinavir in combination with delavirdine and/or adefovir. A unique feature of this study is that it involves no FDA-approved nucleoside analogs. An interim analysis is underway on preliminary data.

Two other large ACTG salvage protocols are expected to open for enrollment by July of this year. ACTG 398 is a phase II trial of amprenavir with or without another protease inhibitor in combination with abacavir, efavirenz and adefovir in 460 volunteers with protease inhibitor failure. Subjects will be allowed to have prior exposure to up to three protease inhibitors and one-third will be allowed prior use of nevirapine or delavirdine.

ACTG 400 is a phase II open-label salvage therapy trial for 300 nelfinavir failures. Participants must be na�ve to at least two nucleoside analogs and all of the NNRTIs. Treatment regimens will include combinations of the other protease inhibitors (including amprenavir) with efavirenz and two new nucleoside analogs. Anyone interested in further information on these trials can call 800/TRIALS-A.

Two small pilot studies are in the works through the Intercompany Collaboration for AIDS Drug Development (ICC) looking at salvage regimens with double NNRTI combinations, including Triangle's new compound MKC422 and delavirdine or nevirapine, plus ddI/d4T with or without hydroxyurea.