Letter: Project Inform nelfinavir potency concerns article

15 June 2000. Related: Correspondence.

To the Editor:

I am responding to the above article on nelfinavir in the Bulletin volume 1 number 2 May 2000, which suggests there is a potency problem with nelfinavir.

In your comment you state that the current federal US guidelines lists nelfinavir-based combinations as preferred treatment, despite the findings of the three studies that you discuss. One probable reason for this is that the US guidelines will have taken into account all the currently available clinical data on nelfinavir, not just selective data that at first appearance might suggest less efficacy and durability of nelfinavir versus other PIs or NNRTI combinations.

Firstly the studies you refer to need to be interpreted carefully, and considered in the context of other recent studies and the overall database on nelfinavir. Thus besides these three reports that appear to cast doubt on nelfinavir, there have been other recent publications that should be taken into account which do not suggest nelfinavir is less potent or durable than other ARTs; in fact some of them have the opposite conclusion.

Thus Easterbrook et al (Lisbon 99 abs 237) in an analysis of the Chelsea and Westminster database showed that nelfinavir ‘achieved the highest rate of undetectability and lowest rate of discontinuation’ compared to other PIs and nevirapine. Ribeiro et al (also Lisbon abst 595) showed nelfinavir to ‘promote the most marked and sustained virologic response’ in a retrospective analysis. Also, Montaner’s analysis of Avanti 2 / 3 and Incas (ICAAC 99 abs 1992) suggested that patients on nelfinavir or indinavir combinations had better durability of response compared to patients on nevirapine.

A recent full report in AIDS of a randomised prospective comparison of nelfinavir versus indinavir (Roca et al. AIDS 2000,14:157-161) showed similar efficacy but better adherence for a regimen of nelfinavir compared to indinavir. So recent data published within the same time frame, as the studies you quote do not raise concerns about nelfinavir’s potency.

Moreover, on this general point of efficacy the pivotal study of BID nelfinavir (542 study) shows as good a result as for most other triple therapy treatments. This was updated at Lisbon last year (abs 205) and showed on an ITT analysis 55% of patients below 50 copies at 72 weeks.

It is of interest also that at the retrovirus conference this year Bartlett et al (CROI abs 520) presented a meta-analysis of 19 treatment regimens from major trials for initial treatment, including nelfinavir, which showed consistency in efficacy across drug classes and certainly no suggestion of a problem with nelfinavir.

Turning now to the three studies you quote which show nelfinavir in a poor light. Firstly I am not aware of Eurosida data showing nelfinavir worse than other PIs. You don’t reference the study so I am not sure which publication you are quoting. However there is a recent Eurosida publication (Paredes et al. Arch Intern Med 2000:160; 1123-32) which shows no difference between nelfinavir and a control group. Moreover Eurosida data presented by Mocroft et al (CROI 2000 abs 539) on response to second line treatment showed no difference in the relative hazard of success comparing PIs individually or versus double PI.

On the second study quoted you tended to write off the explanation given by Agouron for the poor results of ACTG 364, but this was a trial that recruited heavily pre-treated patients who had rolled-over from more than one other ACTG study. If, therefore, there is an imbalance in the groups in terms of nucleoside resistance (as suggested by Agouron) then this could certainly explain why nelfinavir does less well than efavirenz especially when the response to nelfinavir is so much worse than has been seen before in previous trials. There was certainly an imbalance in the resistance profile in those 146 patients who had baseline VL >2000 copies/mL who were genotyped in the trial as seen in the table below which I have composed from data from the poster. See table 1.

Table 1.

It is pertinent that the authors of the poster suggest that all patients should be genotyped at baseline before deciding on the treatment regimen.

Finally I come to COMBINE. This is a randomised study in which the primary outcome measure was defined as the proportion of patients with HIV-1 RNA below 200 copies /mL at 48 weeks. The results presented at CROI 2000 are an interim 24 weeks’ analysis so the primary endpoint has not been achieved.

Of interest the results at 12 weeks published in the abstract of the meeting showed no difference between the groups in terms of virological or immunological response even in patients with baseline viral loads >100,000 copies/mL. The results at 24 weeks show a difference but it is not clear whether all patients have reached 24 weeks.

One other possible explanation for the difference in outcome between the two treatment arms is an imbalance in baseline characteristics. There were significantly more males in the nevirapine arm than in the nelfinavir arm (81 vs 67% p=0.029). There are more heterosexual patients in the nelfinavir arm and half the number of ‘homosexuals’ (18.6 versus 37%).

The mean CD4 count is slightly lower in the nelfinavir arm at baseline, and the mean VL is higher by 46430 copies. The median VL was higher by 6108 copies and there were more patients with VLs >100,000 (37 vs 29%).

Such cumulative differences (although not all individually significant) may be important especially at interim analyses. As far as I can see they were not taken into account in the analysis. Again the results for nelfinavir are out of line with what has been seen before in pivotal studies, which also may be due to the interim nature of the analysis.

In conclusion I hope that my comments alleviate the ‘concerns’ about nelfinavir and put these studies into perspective. An overview of all studies on nelfinavir would suggest no need for concern. Studies published concurrently with the three you quote show positive results (some suggesting superiority of nelfinavir). The published results of Eurosida do not show a problem with nelfinavir either as first or second line therapy. ACTG 364 may, by chance, have randomised more patients resistant to nucleosides into the nelfinavir arm and Combine has not reached its primary end-point.

Yours Sincerely

John Drake

Associate Medical Director, Roche Products Ltd.

Editors note:

We also received a letter from Jens D. Lundgren, M.D., D.M.Sc., Principal investigator for the EuroSIDA study, disagreeing with the interpretation of the EuroSIDA data in the Project Inform article. His concerns mirror those of the letter above. Prof. Lungren states that ‘the EuroSIDA study does not conduct randomised trials, and thus will never be able to conclusively show differences in efficacy of various regimens (but merely associations). ‘The EuroSIDA group has recently published in Arch Intern Med (Paredes et al, Arch Intern Med 160: 1132 – 32) that in people naive to PI/NNRTI (but not necessarily antiretroviral treatment naive) no difference between nelfinavir and ritonavir (the comparison group) was observed.’ ‘In a formal analysis presented at the Retrovirus conference in San Francisco, and contrary to ACTG 364, we did not find nelfinavir as part of a second-line regimen to perform poorer (either for virological success or subsequent failure) compared with other PI regimens in patients failing virologically a first line PI therapy.’

Clearly, head-to-head, prospective randomised trials are the only way to conclusively determine if differences between various combination regimens do exist. These will not, however, be performed by the pharmaceutical companies with their tendency to perform studies of incestuous combinations of their own products for regulatory approval purposes. National agencies should prioritise trials to address these concerns.

Roche Products Ltd. are the marketing licence holders for nelfinavir in Europe.