Merck Protease Inhibitor: More News from Retroviruses Conference

Our last article on the 3rd Conference on Retroviruses and Opportunistic Infections (January 28 to February 1, Washington D.C.) focused on results of the Abbott protease inhibitor, ritonavir (Norvir(TM)), especially the proof that the drug substantially reduced the death rate in late-stage AIDS. The other highest-profile result from the same meeting concerned the Merck protease inhibitor, indinavir (Crixivan(R)), and its effect on viral load, especially in combination therapy.

The indinavir results must be viewed in context of their limitations, however. The two combination trials which attracted much attention at the Retroviruses conference have a total of 78 volunteers each; one is a six-month trial, the other will last one year but has not been completed yet. Also, there are no results today on clinical outcome, since trials to get that proof are ongoing, and no data are yet available.

The following outlines the major indinavir reports at the Retroviruses conference (and also includes information from elsewhere):

* The most impressive result was from the ongoing trial of the triple combination of indinavir plus AZT plus 3TC [late breaker abstract #LB7]. All the volunteers included in this study were AZT experienced, with at least six months prior treatment with that drug. All had T-helper counts of 50-400, and a viral load of at least 20,000 copies per ml. of blood serum, when they entered the study.

This study mainly looked at the percentage of the volunteers whose viral load dropped to an undetectable level (below 500 copies per ml., in the test which was used). Emphasizing this outcome -- instead of the more usual number of logs (factors of 10) that the viral load dropped -- is justifiable, because there is a problem with reporting the average log drop: since the lower limit of detection was 500 copies, it would be mathematically impossible for anyone who started with under 50,000 copies to record as much as a two-log drop, meaning that this study could not show a drop bigger than about two logs, even if it occurred. (Merck also reported the median viral load drop, despite this problem, which tends to make the drug look less active than it really is.)

At 16 weeks, 24 of 26 volunteers taking indinavir plus AZT plus 3TC had undetectable viral load -- compared to 13 of 26 on indinavir alone, and none of 26 on AZT plus 3TC. (Merck gave us this 16-week data after the Retroviruses conference; at that meeting, Merck presented 12-week results, which were
comparable.)

Only a few of the volunteers have yet reached the 24-week point, since the trial is still ongoing. But the percentages at that time do suggest that the results may be sustained; at 24 weeks, six of seven volunteers in the triple combination arm, four of nine using indinavir alone, and none of eight in the AZT plus 3TC treatment arm, had undetectable viral loads.

Median CD4 increases for the triple combination were 79 at 12 weeks and 146 at 24 weeks; the latter figure is less reliable because it is based on a small number of patients.

* Another study tested a similar combination, only indinavir plus AZT plus ddI was used instead of indinavir plus AZT plus 3TC [Abstract #200]. This study also differed in that only antiretroviral-naive volunteers were included. CD4 counts allowed were anything under 500, and as in the study above, viral load had to be greater than 20,000 copies. (The actual median CD4 count at entry to this study was 150 cells per ml, and the actual median viral load was about 5 logs, or 100,000 copies.)

At 20 weeks, 59% of those on the triple combination had an undetectable viral load.

The maximum median decline in viral load with the triple combination was 3.1 logs. (This may seem impossible if the starting median was 5 logs and no value under 2.3 logs (200 copies) could be recorded, but computations with medians can give unexpected results. Those with over 200,000 copies at the start of the trial could record drops of over three logs.)

The median CD4 increase for the triple combination therapy was 90 at 24 weeks.

* Early studies of indinavir alone, in doses that are now known to have been too low, found that the resistance developed fairly rapidly; the virus usually returned to about its starting value within 24 weeks. Persons who developed resistance to indinavir were also resistant to other protease inhibitors. This problem of resistance and cross resistance is the reason for the advice to start protease inhibitors with an adequate dose, and in combination with other anti-HIV drugs; the hope is to reduce HIV replication enough to greatly delay the development of resistant mutants. This is also the reason patients are advised to take these drugs on schedule and not interrupt treatment. (The resistance and cross resistance problems appear to be less with saquinavir, the Hoffmann-La Roche protease inhibitor, although resistance to that drug does develop.) This information was well known before the Retroviruses conference.

Indinavir might continue to have some benefit even after resistance develops. In the early low-dose studies, while the virus had come back by 24 weeks, CD4 count increases of a median of 80 to 100 were maintained for at least 52 weeks. It is possible that the resistant viruses are defective in some way and may cause less damage than the patients' original viruses, but this is not known for sure.

At the conference, results were presented from mathematical modeling of the changes in CD4 count before and after patients received indinavir [abstract #148]. This research suggested that "CD4 return is determined in large part by starting CD4 count."

* Side effects of indinavir include kidney stones or associated symptoms in about 2% to 3% of cases, and elevation of bilirubin. Usually patients are able to continue using the drug even if these effects occur. It is important to drink enough water when using indinavir, to minimize the risk of kidney stones.