Ritonavir (Abbott Protease Inhibitor) Proves Survival Benefit in Late-Stage AIDS

The most important single presentation at the Retroviruses conference was the report from Abbott Laboratories that its protease inhibitor, ritonavir, reduced the risk of death by 43% in a clinical trial with 1090 volunteers. This is the first time that a protease inhibitor has been proven to extend life. Abbott ran this study remarkably quickly; the volunteers were recruited between April and July 1995, and yet the researchers were able to report survival results in late January.

All the volunteers had a CD4 (T-helper) count of 100 or less; the median CD4 count was about 30, and about a quarter of the volunteers had counts under 10. All had used approved antiretrovirals (mostly AZT and d4T) for more than nine months. These volunteers were randomly assigned to receive either ritonavir (600 mg twice daily) or placebo for one year. But after four months, anyone who developed an AIDS-defining event was allowed to end the placebo phase of the study and use the drug.

In addition to the study drug, the volunteers were permitted but not required to continue using up to two other anti-HIV drugs; if they chose to do so, they needed to be on these drugs for at least six weeks before entering the study. 3TC was not permitted, since it was not officially approved when this trial started, although it was widely available in an expanded-access program. Also, volunteers were not allowed to use any other protease inhibitor.

The double-blind portion of this one-year study was stopped early because it reached a predetermined stopping number of 191 clinical outcomes. The 43% reduction in risk of death was found after the volunteers had been in the study for a median of 6.1 months. (The improvement was even more impressive after the first month of the study, with a reduction in the risk of either an AIDS-defining event or death of 58%. The risk reduction was somewhat less for CMV retinitis than for other AIDS-defining conditions.) Both the six-month survival result, and the one-month progression-or-survival result, were statistically significant. On the other hand, 17% of the volunteers on ritonavir had to discontinue treatment due to adverse events, as compared to a 6% discontinuation rate on the placebo.

Viral load, CD4, and CD8 measures were studied in a subset of the volunteers in this trial, and were reported in a separate paper, reviewed below.

Ritonavir is the only protease inhibitor so far to have proven survival benefit. It is widely suspected that other protease inhibitors would have comparable results if they were tested the same way. But in view of Abbott's result, as well as new information about viral load, there would be ethical problems in running the same placebo trial with other drugs.

Comment

In all past trials it has been very difficult to prove a survival benefit from a drug. (The earliest controlled study of AZT did show a survival benefit of AZT vs. placebo, but that result is not entirely credible because of problems with the trial.) How could Abbott prove survival benefit today in such a short time?

There are basically two reasons. First, the new protease inhibitors, especially in combination with other antiretrovirals, are much more effective against HIV than previous treatments were, making it easier to show a difference between treated and untreated patients.

Also, Abbott overcame the widespread industry prejudice against allowing volunteers with very low CD4 counts in trials. Many researchers had come to believe that these patients were too advanced to benefit from antiviral drugs; companies excluded them so that their drugs would look good. But the belief that these people could not benefit came from experience with much less effective drugs. In addition, for several years it has been known that when persons with AIDS are receiving excellent medical care, almost all the deaths occur in those with a CD4 count below 50. Abbott realized that the very people who were being excluded by other companies were the ones it needed to prove that its drug could prolong life.

Recruiting patients is often the critical weakness of clinical trials. The ritonavir trial recruited quickly because persons with a low CD4 count had no other trials available, because Abbott opened 67 trial sites in the U.S., Europe, and Australia, and because persons in the trial could continue the anti-HIV medications they were already using. This trial has been well accepted by patients and activists.

Ritonavir does have a major disadvantage, however. It interacts with many other drugs by preventing them from being metabolized by the body, causing the normal dose of some drugs to become a dangerous overdose. Some medications can be used with appropriate dose adjustments; others cannot be combined with ritonavir at all.

The Merck protease inhibitor indinavir (Crixivan(R)) has much less of a drug-interaction problem, and probably at least as much ability to suppress HIV as ritonavir does. But it does not have data proving that it prolongs survival.

Where do we go from here in proving clinical benefit? Must every new protease inhibitor prove that it can extend life? We think not, for several reasons:

* Any proof of survival benefit requires that people die in the trial. These trials become increasingly problematic ethically, as evidence accumulates that viral load changes predict clinical response to therapy.

* It is clearly unworkable to prove survival benefit EARLY in HIV infection, because the trial would have to run for many years before deaths occurred. It would be nearly impossible to run such a trial, and unconscionable to deny early treatment until its results were in. We will have to find other ways of investigating the effects of early treatment, based on better understanding of the disease process.

* All indications to date are that the key contribution of ritonavir was in lowering viral load. If the viral load could be lowered as much by another protease inhibitor, by other drugs, or even by herbal treatment, without introducing major toxicities, all indications are that the survival benefit would be about equal.

* Viral load trials can be run much more rapidly than clinical-endpoint trials, with far less resources, allowing screening of many drugs and combinations at all stages of HIV disease, and without ethical problems as no one has to die or suffer significant harm.

We believe that clinical-endpoint trials should be used very selectively, and no longer be required by the FDA for every new drug. Resources should be shifted to rapid trials which look at viral load, CD4 counts, and other measures of HIV disease status, to determine how well treatments are working -- trials which include long-term followup. Resources now spent on confirming final certainty in regulatory decisions should be focused instead on answering physicians' practical questions on how best to use the drugs.

What about the risk of long-term drug toxicity? This is always a danger; but is the best way to look for it a randomized trial designed to determine whether or not the drug works against the disease? If the benefit and harm of the drug are nearly balanced, then a clinical-endpoint trial might be necessary to tell which was greater. But today there are many protease inhibitors and combinations with major impact on HIV; a drug which caused equivalent damage in side effects would be abandoned long before such a trial could be completed.

This issue remains controversial, however, because in theory at least, clinical benefit trials still offer the most certain proof that a drug does more good than harm. Many people are philosophically attached to this promise of certainty.

[Note: The ritonavir survival results were presented in the "late breaker" session of the conference, abstract #LB6a. The late breaker session is for results which are too recent to have been submitted by the regular deadline, but important enough to be accepted anyway.]

Other Ritonavir Results

The following additional results on ritonavir were presented at the Retroviruses conference:

* Ritonavir plus AZT plus ddC produced very good viral load suppression. A study in France treated 29 patients with advanced, untreated HIV infection with this combination; 21 of them were able to tolerate the treatment for at least six months. Median viral load decreased to less than 1% of its starting value by month two, but was slightly over 1% at month six. Median CD4 count went from 156 at baseline to 303 at month six. Twelve of the 21 patients had the number of infectious cells reduced by over 99.99%. [Abstract #285]

* A substudy of the ritonavir survival trial reported viral load, CD4, CD8 changes in the first 159 patients enrolled who started with a viral load over 15,000 copies (which included most patients, since the typical viral load was about 200,000 copies). They were studied for the first four months of the trial (that is, before the point at which some patients were allowed to leave the placebo-controlled treatment and begin open-label ritonavir). The greatest difference in viral load between the ritonavir and placebo groups was 1.3 logs (20 fold), measured at two weeks; at the end of the four-month substudy, the viral load of the treatment group was about 0.6 logs (4 fold) lower than that of the placebo group, which had changed little during the trial, as expected. The greatest average difference in CD4 count was 45, measured at 16 weeks. The greatest difference in CD8 count was 363, measured at eight weeks (over an average baseline value of about 500). All these differences were
highly statistically significan
t. [abstract #LB6b, with additional information from the oral presentation.]

* An Australian paper looked at immunological changes in 21 patients treated with ritonavir, using two-color flow cytometry, and found improved immunologic function, as well as increased cell number. [abstract #232]

* A study in animals found that when ritonavir was combined with other protease inhibitors, the blood level of the other drug was greatly increased. This effect could vary greatly; the increase in blood level of saquinavir (the Roche protease inhibitor) was 29,000 percent (290 times), while the increase in indinavir (the Merck protease inhibitor) blood level was 800%. [abstract #143]

* A resistance study compared viral load and particular mutations of HIV in seven patients treated with a low dose of ritonavir without other drugs. After two weeks, average viral load had decreased by 1.62 logs (42 fold); but later the viral load increased, and it returned to baseline values in six of the seven patients. The rise in viral load was strongly associated with mutations, especially at positions 82 and 63 of the protease gene. The mutant virus appears to have been present before the patients ever saw the drug, but at very low levels, less than one one hundredth of one percent of the total virus. [abstract #201]

* A retrospective study compared the CD8 count changes in eight previous trials, of nucleoside analog reverse transcriptase inhibitors (such as AZT, or 3TC), non-nucleoside analog reverse transcriptase inhibitors (such as nevirapine), and the protease inhibitor ritonavir. Patients using ritonavir had large CD8 increases, up to 892 in higher-dose groups -- compared to small increases, or even decreases, with the other therapies. [abstract #451]