Initially used in salvage regimens, dual protease inhibitor therapy is now increasingly used earlier in treatment. (For discussions of dual protease inhibitors in salvage therapy, see Treatment Issues, "Salvage Therapy: Still More Intuition Than Data" January 1999 and "The Great Salvage Therapy Drug Juggle" April 1998.) The primary reasoning behind this strategy is that combining protease inhibitors can help overcome some of their shortcomings, such as poor biavailability and short half-lives. This is possible because they all are metabolized by the cytochrome P450 3A4 isoenzyme system. Ritonavir, in particular, inhibits this system, which means that administering ritonavir with other protease inhibitors can increase their blood levels. This allows lower or less frequent dosing and can remove food restrictions, resulting in fewer toxicities and improved adherence. Some combinations, such as ritonavir/saquinavir, and possibly ritonavir/indinavir, also appear to be very potent and long lasting. Nevertheless, federal guidelines do not recommend dual protease inhibitor therapy as first-line treatment, presumably due to concerns about side effects and drug-drug interactions, particularly since ritonavir interacts with a wide variety of drugs. There are also concerns that the use of two or more protease inhibitors could increase the risk of lipodystrophy and other metabolic toxicities. At this point, there are very little data on the long-term consequences of using dual protease inhibitor therapy.

Recent conferences have included several studies of dual protease inhibitor therapy in both treatment-naive and treatment-experienced populations, which are discussed below.


The combination of ritonavir and saquinavir has been in use for several years now, most commonly at a reduced dose of 400 mg each, twice a day. Last winter, Cameron presented three year follow-up data on the combination in 144 patients (Retrovirus Conference, January 30 to February 2, 2000, abstract 533). Patients started on various doses of the two protease inhibitors, but all eventually went to 400/400 mg. Intensification with nucleoside analogs after week twelve was allowed for those with a viral load over 200 copies, and at week 48 for anyone who wished to do so. Through week 144, 120 patients (85%) had a suppression of viral load to below 200 copies (thirteen results were obtained after treatment intensification). However, 41 patients (34%) in this group had a viral rebound. Fifteen of the 41 then intensified treatment with nucleoside analogs, twelve of whom subsequently stayed below 200 copies through 144 weeks. That is, 91 (63%) of the 144 eventually maintained a viral load below 200 copies, and 53 (58%) were only taking ritonavir and saquinavir. While 50 (35%) of the original 144 patients discontinued treatment, 20% because of adverse reactions, no new safety issues were uncovered during the study.

In another study involving intensification, 208 protease inhibitor- and d4T-naive patients were randomly assigned to receive ritonavir and saquinavir (400/400 mg) with or without d4T, twice daily (Gisolf, Retrovirus Conference, January 30 to February 2, 2000, abstract 527). If patients had a viral load over 400 copies after twelve weeks, nucleoside analog intensification was allowed. At 48 weeks on an intent-to-treat analysis, 63% of the ritonavir/saquinavir group had an undetectable viral load (limit not stated), compared to 69% of those who also received d4T (p=0.38). Twenty-eight patients who initially received only the protease inhibitors intensified therapy, as did three in the d4T group. Of those who intensified treatment, 97% had an undetectable viral load at their last visit. There was a ten percent dropout rate due to adverse events. The authors concluded that starting with ritonavir/saquinavir and intensifying the treatment if necessary was comparable to starting with three drugs, but longer follow-up is needed to determine this strategy's long-term efficacy.

These reports follow a study presented last October at the Lisbon Conference (Lundgren, ECCAT, October 23 to 27, 1999, abstract 487). In that trial, 318 protease inhibitor-naive patients took two nucleoside analogs plus either ritonavir, indinavir, or ritonavir/saquinavir. At 48 weeks, 42% of those taking indinavir, 40% of those taking ritonavir, and 58% of those taking ritonavir/saquinavir had viral loads below 20 copies. Researchers noted that ritonavir as the only protease inhibitor was not well-tolerated, with twice as many patients dropping out of that arm as the other two (about 66% compared to 33%).


Combining ritonavir with indinavir significantly increases the blood levels of indinavir and allows indinavir to be dosed twice daily without food restrictions. A recent study also concluded that low doses of ritonavir can triple the amount of indinavir in cerebrospinal fluid and increase semen indinavir levels seven-fold, which may affect HIV in these sanctuaries (van Praag, Retrovirus Conference, January 30 to February 2, 2000, abstract 312). While Merck favors combining 800 mg of indinavir with 200 mg of ritonavir, Abbott argues that a 400/400 combination is most effective. Currently enrolling, ACTG A5055 will compare the 400/400 and the 800/200 doses in patients who are failing amprenavir, nelfinavir, saquinavir, or nelfinavir/saquinavir combination therapy.

Two studies at the Lisbon Conference examined the 400/400 combination. The first included 92 treatment-naive patients who took ritonavir, indinavir, and two nucleoside analogs (Rockstroh, ECCAT, October 23 to 27, 1999, abstract 452). At 48 weeks, data were available for 88 patients, 24 of whom dropped out of the study. Ninety percent of the on-study population (70% including dropouts) had viral loads below 80 copies. These results, however, are less impressive than a ritonavir/saquinavir combination at 48 weeks. In the second study, there were 48 week data for 14 of 22 patients taking d4T, 3TC, ritonavir, and indinavir (Workman, ECCAT, October 23 to 27, 1999, abstract 620). Twelve of the 14 had viral loads below 50 copies. In both studies, patients had high blood lipid levels -- including seven serious cases in the second study -- and difficulties tolerating ritonavir, which, at that time, was only available in its notoriously horrible tasting liquid formulation.

In an intensification study, Shulman presented data on adding ritonavir to patients with a detectable viral load on indinavir and two nucleoside analogs (Retrovirus Conference, January 30 to February 2, 2000, abstract 534). Indinavir Cmin increased 272% and Cmax decreased 51% while AUC was maintained. Of 17 patients who had reached 16 weeks, the median viral load dropped to below 100 copies.

ABT-378/Ritonavir (Lopinavir)

At the Retrovirus conference, Gulick had results from an ABT-378/ritonavir trial involving 100 treatment-naive patients (abstract 515). For the first 48 weeks, patients took d4T, 3TC and ABT-378/ritonavir in a dose of either 200/100 mg or 400/100 mg (Group I) or 400/100 or 400/200 mg (Group II); after 48 weeks, all patients switched to the 400/100 mg dose of ABT-378/ritonavir. Median baseline viral loads were 100,000 and 80,000 copies and median CD4 cell counts were 421 and 301 cells for Groups I and II, respectively. At week 72, 82% had a viral load below 400 copies (intent-to-treat analysis) and 80% were below 50 copies. Median CD4 increase was 275 cells. While only one percent of patients discontinued the trial due to adverse events (12% discontinued for other reasons), grade two to four events (moderate to severe side effects) included diarrhea (21%), nausea (15%), abnormal stools (8%), weakness (7%), headache (7%), and vomiting (5%). Grade three or four laboratory abnormalities were increased cholesterol (14%), increased triglycerides (12%), and increased liver enzymes (8%, half of whom had chronic hepatitis B or C).


The SPICE study randomized 157 protease inhibitor-naive patients to receive saquinavir/two nucleosides, nelfinavir/two nucleosides, saquinavir/nelfinavir/two nucleosides, or saquinavir/nelfinavir (Johnson, 6th Retrovirus Conference, 1999, abstract 389). Nelfinavir was dosed at 750 mg twice daily, and saquinavir was at 1200 mg three times daily in all but the dual protease inhibitor arm (800 mg, twice daily). At 48 weeks, the proportions of patients with a viral load below 50 copies were 42%, 42%, 51%, and 35% in the four arms, respectively. Of 22 patients with virologic failure who switched to quadruple therapy, 24% achieved a viral load below 50 copies.

In the TIDBID study, three saquinavir-containing regimens were compared: saquinavir 1200 mg three times daily and two nucleoside analogs, saquinavir 1600 mg twice daily and two nucleoside analogs, or saquinavir 1200 mg twice daily, nelfinavir 1250 mg twice daily and one nucleoside analog (Cohen, 39th ICAAC, September 26 to 29, 1999, abstract 508). Of the 838 patients enrolled, 75% were treatment-naive (treatment-experienced patients received at least one new nucleoside analog). Data at 24 weeks showed that all three groups had similar viral suppression, with 40% and 60% below 50 copies on an intent-to-treat and as-treated analysis, respectively. CD4 cell changes were also similar in all groups, with about a 150 cell increase. Gastrointestinal side effects were most common, and diarrhea was more frequent in the dual protease arm (13% to 15% in the saquinavir arms and 23% in the saquinavir/nelfinavir arm).

Dual, or Triple, PIs in Salvage Therapy

Forty-eight week data from a salvage therapy trial involving ABT-378/ritonavir were presented at the last Retrovirus conference (Deeks, abstract 532). On an intent-to-treat analysis, 70% of patients had a viral load below 400 copies and 60% were below 50 copies. However, as Treatment Issues has noted previously (see September/October 1999), patients also took efavirenz (in addition to two nucleoside analogs) and were NNRTI naive, so it is impossible to tell how much ABT-378/ritonavir contributed to the results.

Results from ACTG 398 -- a study to determine whether a second protease inhibitor, combined with amprenavir, abacavir, efavirenz, and adefovir, would improve virologic response in protease inhibitor-experienced patients with a detectable viral load -- were presented as a Retrovirus late-breaker (Hammer, LB 7). Based on prior protease inhibitor experience, 481 patients were randomized to also receive saquinavir (1600 mg, twice daily), indinavir (1200 mg, twice daily), nelfinavir (1250 mg, twice daily), or a placebo. Median baseline viral load was 51,600 copies and CD4 count was 202 cells. At week 24, 35% in the saquinavir arm, 38% in the indinavir arm, and 39% in the nelfinavir arm had viral loads below 200 copies, compared to 28% in the placebo arm (p=0.02). Prior NNRTI use was significantly associated with virologic failure, but previous protease inhibitor use (one vs. two or more) was not. Also at week 24, seven percent of the patients had stopped treatment, 33% due to toxicity and 19% due to virologic failure. Grade three to four adverse events were similar in all four arms, with gastrointestinal symptoms, hypertriglyceridemia, and hypophosphatemia being the most common.

Two retrospective studies at the Third Salvage Therapy Workshop looked at indinavir (800 mg)/ritonavir (200 mg) in a salvage therapy setting. Grossman reported on 41 heavily pretreated patients (73% with NNRTI experience, 95% with prior indinavir or ritonavir experience) with a median viral load of 30,000 copies and CD4 count of 258 cells (abstract 27). At three, six, and nine months, 51%, 56%, and 62% of the patients, respectively, had viral loads below 400 copies. Median CD4 increase was about 75 cells at six months. Although there were no reports of kidney stones, nineteen of the 41 (46%) had adverse events, including nausea and vomiting (17% of the 41 patients), rash and dry skin (12%), diarrhea (5%), and paresthesia (3%).

In the second study, Campo provided results for 27 heavily pretreated patients who took at least one nucleoside analog and/or NNRTI in addition to the protease inhibitors (abstract 7). Fifteen patients (56%) attained viral suppression (<400 copies). In addition, four of four with baseline phenotypic resistance and ten of thirteen with baseline genotypic resistance to indinavir and ritonavir attained a viral load below 400 copies. Responders had higher initial CD4 counts (283 vs. 150) and lower viral loads (156,500 vs. 228,000). Of interest, however, baseline resistance to indinavir and ritonavir was associated with a better response to therapy. To explain this counter-intuitive result, Campo suggested that resistance could have be an indication of adherence. That is, those who were adherent on failing regimens developed resistance, and they were more likely to be adherent on the salvage regimen. Given the small numbers and retrospective nature of the study, however, it is difficult to draw any strong conclusions about the relationship between baseline resistance and adherence.

Also at the Salvage Therapy Workshop, researchers from Glaxo Wellcome proposed a salvage therapy regimen of amprenavir (600 mg), saquinavir (800), and ritonavir (100) twice daily, with or without other classes of antiretrovirals (Furfine, abstract 16). It is crucial to note that this proposal is currently theoretical: the pharmacokinetic interaction studies have not been carried out in humans, and what happens in the body can show great variance from what happens in a test tube. Nevertheless, there are several (again, theoretical) reasons to think the combination might be effective. First, laboratory tests have shown that amprenavir and saquinavir are synergistic. Second, there is little cross-resistance between amprenavir and saquinavir; in fact, in vitro studies show that the amprenavir I50V mutation is more susceptible to saquinavir. Moreover, when HIV with amprenavir-resistant mutations 46I/47V/50V is exposed to saquinavir, it develops the 84L mutation, but this quadruple mutation leads to greater susceptibility to amprenavir. Third, people who have failed nelfinavir or indinavir are often still sensitive to amprenavir and saquinavir. Finally, ritonavir enhances the blood levels of amprenavir and saquinavir.

Just as the benefits of this triple protease inhibitor combination are still unproven, the related adverse events are unknown. With some cooperation among the drug manufacturers, future drug interaction trials could shed some light on the feasibility of using this regimen in salvage therapy.

Back to the GMHC Treatment Issues Spring 2000 contents page.