Study of nucleoside-sparing regimens stopped early due to high failure rates; NRTIs likely to remain therapeutic backbone for the time being

A French study designed originally as a two-year experiment to see if giving patients antiretroviral regimens that excluded the nucleoside reverse transcriptase inhibitors (NRTIs) would result in lower rates of lipoatrophy (fat loss) has been stopped after a year because of higher failure rates in the NRTI- sparing arm.

The French Hippocampe (‘Seahorse’) study took 117 drug-naive patients and gave half of them regimens containing no NRTIs, a quarter standard NRTI plus non-nucleoside (NNRTI) regimens and the other quarter standard NRTI plus protease inhibitor (PI) regimens.

The NNRTI and PI drugs were restricted to efavirenz (Sustiva) or nevirapine (Viramune) plus boosted lopinavir or indinavir (Crixivan). All NRTIs were allowed in patients that took them except for d4T and ddC, because of their greater mitochondrial toxicity and association with lipoatrophy. About two-thirds of patients took efavirenz and a third nevirapine (63% vs. 37%); about four-fifths of patients on PIs took boosted lopinavir and a fifth boosted indinavir (78.5% vs 21.5%).

By the 48^th week of the study, when it was stopped by its safety monitoring board, 78.9% of patients in the NRTI-containing arms had viral loads under 50 copies/ml compared with only 61.7% in the NRTI-sparing arm, on an intent-to-treat analysis meaning that any change or discontinuation of the study regime was judged a failure. The NRTI-sparing regimens had been less potent throughout the study, with the proportion with their viral load under 50 copies/ml after 12 weeks, for instance, being 57.9% in the NRTI-containing groups but only 36.7% in the NRTI-sparing one.

The percentage of viral loads under 400 copies/ml was not significantly different between the NRTI-containing and NRTI-sparing arms (83% versus 75%).

Fourteen patients in the NRTI-sparing arm either discontinued the study or had their regimen modified compared with three in the NRTI-containing arms.

On an on-treatment analysis, where only patients remaining on the given study regimen were counted, there was still a significant difference in viral failure rates up to week 36 of the study but this had become non-significant by week 48 because so many had dropped out of the nucleoside-sparing group.

An audience member commented that the results of the NRTI-sparing regimens were surprisingly bad, given that trials using boosted lopinavir as monotherapy had shown better results.

Possible explanations include:

• A high proportion of patients had low CD4 cell counts at baseline and study co-ordinator Dr Christine Katlama commented that “you need a very potent therapy for CD4 counts under 200 to start with.”
• Adherence rates were different in the two arms, though not to a degree that reached statistical significance. Seventy-seven per cent of patients taking NRTIs had high (>95%) adherence compared with 72% not taking them.
• The ethnic mix of the study groups varied. There were 49% black Africans in the NRTI-sparing group compared with 37% in the NRTI-taking group.
• Resistance tests were not taken at baseline so more patients in the NRTI-sparing group might have had pre-existing resistance; however the number failing with drug resistance was actually surprisingly low, with eight people showing drug resistance, four to non-nucleosides and only one to PIs.

Whether or not these are confounding variables, these results back up those from another study whose results were announced earlier this year (Fischl) where 74% of patients taking efavirenz plus two NRTIs achieved a viral load under 50 versus only 66% of those taking efavirenz plus boosted lopinavir.

The poorer performance of NRTI-sparing regimens means that nucleosides are likely to remain an essential part of highly active antiretroviral treatment (HAART) regimens for the time being.