Skip directly to content

CPCRA v. Syntex

Effect of ddI on Ganciclovir May Explain Disparity Between Two Prophylaxis Studies

And CMV PCR Makes its Debut

November 1996

Readers of TAGline are asked to harken back to the duelling ganciclovir studies unveiled this past winter (see the March TAGline, vol. 3, issue 2). Where the Syntex study #1654 (now Syntex/Roche) declared a huge benefit to the its drug in preventing the onset of CMV retinitis, the community-based CPCRA 023 study found no benefit at all. These conflicting results have remained a favorite subject of controversy at the years large AIDS meetings, especially in light of the drug's exorbitant price ($25,000-$30,000 annually). Now that the final analyses of these two studies have recently been completed, clues to the puzzling discordance have begun to emerge.

Antagonism Between Oral Ganciclovir and ddI Suspected to Diminish Anti-CMV Effect

Although one study has reported ganciclovir to decrease ddI's antiretroviral effect, ganciclovir is generally known to increase blood levels of ddI when the two drugs are taken together. The literature on ddI's effect on the pharmacokinetics of ganciclovir, however, is quite limited. This is about to change. For when the CPCRA 023 team combed through their data in their oral ganciclovir CMV prophylaxis study, they uncovered a curious (and antagonistic) relationship between concomitant use of oral ganciclovir and ddI. In their study, participants who received placebo ganciclovir and were also taking ddI experienced a very low rate of CMV disease. (That in itself seems odd.) By contrast, study participants who were actually receiving ganciclovir and were also taking ddI experienced a very high rate (7.5-fold increased risk, p= 0.02) of CMV disease. (Odder still.) In fact, of the 63 individuals in the placebo arm who were also taking ddI, only one developed CMV disease (that comes out to 1.9 cases of CMV per 100 person years). But for the 269 individuals in the placebo arm who were not taking ddI, 34 developed CMV (18.9 per 100 person years). So the CPCRA 023 results need to be looked at with this unexpected drug-drug interaction in mind; when ddI-taking individuals are removed from the analysis, prophylactic use of oral ganciclovir reduced the risk of developing CMV disease by 38%, compared to placebo (p= 0.04). (The Syntex study #1654, you may recall, reported that oral ganciclovir reduced the risk of CMV by 50%.)

The interaction between oral ganciclovir and ddI has baffled many. "Does ddI have some as yet undiscerned prophylactic effect against CMV?" became the obvious question. The curious prophylactic results seen with ddI had not been noted before. In fact, when the Syntex #1654 study team performed a post hoc analysis of all the #1654 patients who were taking ddI, they did not find the ddI effect that the CPCRA 023 team had reported. For their part, the 023 team stands by their analysis, but study chair Carol Brosgart cautions that the results should be interpreted carefully and that the negative ddI and oral ganciclovir interaction should be investigated further. Some people are now joking about getting Bristol (manufacturer of ddI) to design a CMV prophylaxis study with ddI alone or maybe in combination with its recently rediscovered anti-infective lobucovir.

Using CMV DNA PCR to Individualize Prophylactic Strategies

An interesting feature of the Syntex #1654 study was the monitoring of CMV-DNA by new PCR techniques. Virology specimens were available for 619 of the study's 725 participants, and the results are fascinating. For all study participants who were CMV PCR positive at study entry (approximate 55% positivity rate in both arms), CMV disease developed in 43% of the placebo recipients and 26% of the oral ganciclovir recipients (p=0.017). Among the CMV PCR negative patients at study entry, CMV disease developed in 1% of the oral ganciclovir recipients and 14% of the placebo recipients (p= 0.00001).

Quantitatively, CMV-DNA PCR positive persons were stratified into three groups: fewer than 2,500 copies; 2,500-50,000 copies; and 50,000-150,000 copies. Study participants with fewer than 50,000 copies of virus benefitted from oral ganciclovir: 20% of the oral ganciclovir patients developed CMV disease vs. 40% of those on placebo, p= 0.004). For those with greater than 50,000 copies of CMV virus, there was no protective effect of oral ganciclovir compared to placebo. Of note, there were 10 patients with over 150,000 copies (all randomized to the oral ganciclovir arm) and they all went onto develop CMV disease.

This viral load sub-study tells us three things: 1) CMV-DNA PCR negative patients can benefit from oral ganciclovir; 2) If a person is CMV-DNA PCR positive and has a low viral load (less than 50,000 copies) s/he will still get some benefit with using oral ganciclovir; and 3) If a person has high CMV-DNA viral load (over 50,000 copies), oral ganciclovir won't help. Curiously, the only other CMV prophylaxis study to use PCR techniques to assess qualitatively the CMV-DNA status of participants was the NIAID's valacyclovir study (ACTG 204). In that study, exactly the opposite predictive parameters were reported; that is, CMV PCR-positive persons responded very well to prophylactic therapy with valacyclovir, while those PCR-negative showed little benefit to anti-CMV prophylaxis.

Still, the #1654 study results point to some obvious questions: For someone who is CMV-DNA PCR positive with a CMV-DNA copy number over 50,000, is it worth (both in terms of medical economics and quality of life) prophylaxing with oral ganciclovir? Similarly, in a CMV-DNA negative person, what is the value of prophylaxis with oral ganciclovir (at $30,000 a year) when the risk of developing CMV disease is only 14%?

In many ways, this viral load sub-study of #1654 acts as a natural history study which tells us a great deal about who will might develop CMV. For a long time we knew that a CD4 cell cut-off of 50 cells was not sufficiently sensitive. If we look at all patients at baseline (those with under 50 CD4+ cells), approximately 50% will be CMV PCR positive and 50% will be CMV PCR negative. Approximately 15% of all the negatives and 45% of all positives will develop CMV end-organ disease. If a patient has high CMV viral load, his or her risk of developing CMV end-organ disease is ostensibly greater. These data have also given rise to the prospect of treating a CMV PCR positive person with a short course of intravenous ganciclovir (to get him or her CMV PCR negative) followed by oral ganciclovir, or nothing at all.