Fos-amprenavir (433908) shows signs of weakness in Kaletra comparison

Vertex Pharmaceuticals has released preliminary 48-week results from the CONTEXT study comparing its new formulation of amprenavir (433908) with lopinavir/ritonavir. The results show that when comparing the time-averaged viral load reduction from baseline over 48 weeks, 433908, or fos-amprenavir, boosted with ritonavir may be inferior to lopinavir/ritonavir (Kaletra). However, when comparing the proportions with viral load below 400 copies at week 48, no significant difference was seen between the twice daily fos-amprenavir/ritonavir and lopinavir/ritonavir regimens. Once daily fos-amprenavir appeared to be inferior by this measure too.

The study compared three regimens:

• Fos-amprenavir 1400mg once daily (two tablets) boosted with 200mg of ritonavir plus two nucleoside analogues selected by genotypic resistance test (n=105)
• Fos-amprenavir 700mg twice daily (one tablet) boosted with 100mg of ritonavir, plus two nucleoside analogues selected by genotypic resistance test (n=107)
• Lopinavir 400mg/ ritonavir 100mg (Kaletra) plus two nucleoside analogues selected by genotypic resistance test (n=103)

It recruited 320 protease inhibitor-experienced patients who had experienced failure of at least one PI-containing regimen. The primary endpoint of the study was time-averaged change in viral load minus baseline (AAUCMB), the endpoint recommended by the US Food and Drug Administration for evaluation of antiretroviral agents in treatment-experienced patients.

At 48 weeks, the mean AAUCMB (log₁₀ copies/mL) was -1.49 for patients on 908/r QD, -1.53 for those on fos-amprenavir/ritonavir BID and -1.76 for those taking LPV/r BID. Non-inferiority at 48 weeks using the primary AAUCMB endpoint could not be established because the upper limits of the 97.5% confidence intervals for the differences in mean AAUCMB between each of the fos-amprenavir/ritonavir treatment arms to LPV/r comparisons were above 0.5. In other words, there is a possibility that the difference between the lopinavir/ritonavir and fos-amprenavir/ritonavir groups could be larger that 0.5log₁₀ copies/ml.

However, in an analysis of the proportion of patients with viral load below 400 copies/mL, fos-amprenavir/ritonavir BID (58%) and LPV/r BID (61%) demonstrated comparable levels of antiviral activity. The proportion of patients who achieved vRNA below 50 copies/mL at 48 weeks was also similar in the fos-amprenavir/ritonavir BID (46%) and LPV/r BID (50%) arms. The proportions with viral load below 400 copies/ml in the fos-amprenavir/ritonavir QD arm was 50%.

Further data will be presented soon, probably at September’s Interscience Conference on Antimicrobial Agents and Chemotherapy in Chicago.

Vertex says that it is still confident that it can obtain marketing approval for fos-amprenavir in the United States before the end of 2003, and in Europe early in 2004. Vertex is co-developing fos-amprenavir with GlaxoSmithKline, which sponsored the Context study.

Vertex was not able to supply information on the following questions, citing the need the need to hold some information back for submission to a future scientific meeting:

• What were the actual confidence intervals for each comparison?
• What was the proportion below 50 copies/ml in the fos-amprenavir/ritonavir QD arm? At week 24 this was 40%
• Were the differences between the QD arm and the other arms statistically significant for the

Context in context

The results of the Context study need to be considered with respect to two issues.

Firstly, using time averaged viral load reduction minus baseline will underestimate the extent of viral load reduction because it can only detect the reduction that occurs from baseline to the lowest limit of detection used in the study. This may serve to mask differences in the intensity of viral load suppression below 50 copies, or to exaggerate the importance of relatively minor viral load variations in the band between 400 and 50 copies, depending on which you consider to be more important.

The fact that higher proportions of the fos-amprenavir treated patients had been exposed to at least two protease inhibitors may also be important. 49% and 38% of the fos-amprenavir/ritonavir QD and BID arms had been exposed to at least two protease inhibitors, compared with 32% of the lopinavir/ritonavir arm, according to data presented with the 24 week preliminary analysis of the Context study at the Tenth Conference on Retroviruses and Opportunistic Infections earlier this year in Boston. In this study variations in response due to sensitivity to the background nucleosides were likely to have been minimal, since everyone admitted to the study had shown sensitivity to two nucleoside analogues by phenotypic resistance testing. It should thus have been a clearcut test of response to the protease inhibitor component.

A poster presentation at the Twelfth International HIV Drug Resistance Workshop earlier this year showed that when analysed by response at week 8, patients with the a protease mutation at codon 82 had a better response if they received fos-amprenavir/ritonavir, whilst those with the I84V mutation had a better response if they received lopinavir (Yates). These preliminary findings suggest that it will be possible to individualise protease inhibitor treatment in PI-experienced patients to ensure that fos-amprenavir/ritonavir is used appropriately.

The results of the Context study will invite comparison with BMS 045, a randomized comparison of atazanavir/ritonavir and lopinavir/ritonavir. That study required patients to have experienced failure of at least one prior protease inhibitor-containing regimen,and to have fewer than four primary protease mutations. A more detailed analysis of the Context study will hopefully show whether its participants were more PI-experienced than those treated in the BMS 045 study. If they are, any marketing-driven comparisons between atazanavir/ritonavir and fos-amprenavir/ritonavir will be dubious.

The weaker performance of the once daily fos-amprenavir arm in this study is also confusing. Was this driven by greater PI experience in this group, or by an inherent weakness of once daily dosing of this drug? Ritonavir boosting of amprenavir delivers an extended half-life, but does not deliver increases in trough levels to the same extent as ritonavir boosting of atazanavir. A comparison of the inhibitory quotient (the ratio of trough levels to IC₉₀ (the concentration needed to inhibit 90% of virus replication)) for each regimen in this study would be instructive.