PI-dropping studies: still little impact on body fat changes

Switching from a protease inhibitor to an NNRTI or abacavir yields inconsistent results when the aim is the reduction of PI-related side effects such as lipid and body fat disorders, according to a host of studies presented yesterday at the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy in Toronto.

The switch studies were:

• A Spanish randomised study (1) of switching from PI to efavirenz or nevirapine-based therapy, compared with sticking to a PI

• An open label study of switching from PI to nevirapine-based therapy (2), with 18 months follow-up (Maintavir)

• A randomised study of switching from PI to efavirenz-based therapy, compared with sticking to a PI (DMP266-049) (3)

• A Swiss randomised study of switching from PI to abacavir-based therapy, compared to sticking with a PI, with 18 months follow-up (4)

• A Canadian randomised study of switching from PI to abacavir-based therapy (CNA30017)

The Spanish group (1) randomised 77 individuals with at least 9 months history of viral load continuously suppressed below 50 copies, and conducted DEXA scans at baseline to assess fat distribution. In this study there was no significant difference in the rate of viral rebound between the three arms after 9 months. Total cholesterol and LDL cholesterol levels fell significantly in the nevirapine arm, but not in the other arms. In a sub-study of 56 patients diagnosed with body fat changes at baseline (as defined by an objective scale developed from a cohort of Spanish patients with severe lipodystrophy), there was no significant improvement in body fat redistribution after 9 months when assessed by DEXA scan and anthropometric measures.

Francois Raffi reported on the Maintavir study (2), an open label French study which has now followed 73 patients for approximately 18 months after a switch from PI to an NNRTI-based regimen. All switchers had a history of viral load suppressed below 400 copies for at least one year prior to switching (69 of 73 were below 50 copies), and had been on PI treatment for an average of 22 months. Sixty-three switched to nevirapine and 10 to efavirenz, and in most cases did not make any changes to the NRTI backbone.

Ten out of 73 switchers experienced viral load rebound, which was significantly more likely in patients who had some experience of antiretroviral therapy, perhaps because NNRTI-based HAART was less able to control NRTI resistance, but since the study is uncontrolled we don’t know how they might have fared if they had stuck with PI-based HAART.

Seven of the 18 cases of lipodystrophy diagnosed at baseline were said to have improved during the study, but no baseline DEXA measurements were conducted that would allow an objective assessment. All the improvements were seen in the patients with fat accumulation. Three cases of lipoatrophy appeared after the switch, but no cases of fat accumulation developed.

There was no significant change in cholesterol levels in this study (although they were in the normal range to start with), but triglyceride levels did fall significantly by month 6 and returned to the normal range.

In DMP266-049 (3), 207 individuals with a history of more than 6 months viral load below 50 copies were randomised to switch to efavirenz or remain on a PI. This study reported 24 week data, and by a non-completer equals failure analysis, the efavirenz group had a significantly better viral load response (89% vs 81%). However, using an As Observed analysis, there was no difference in response. The data on lipid changes reported non-fasting measurements. Average baseline cholesterol measurements were within the normal range, and did not change during the follow-up period. HDL cholesterol levels increased significantly however; since the level of ‘good’ HDL cholesterol is often depressed in HIV infection, this is an encouraging result.

In the Swiss study, 163 individuals with a history of more than 6 months of continuous viral load suppression below 50 copies on PI-based HAART were randomised to switch to abacavir-based HAART, (and were able to take Trizivir when it became available), or to stay on PI-based HAART. After 68 weeks there was no difference in the rate of viral load rebound between the two groups (29% PI vs 25% ABC), but cholesterol levels were significantly more likely to have fallen within four weeks of randomisation in the abacavir group ( to the mid-average level), although HDL cholesterol levels did not fall.

A concern in this study was the greater risk of viral load rebound in abacavir recipients who had taken antiretroviral drugs prior to commencing PI-based HAART. In such cases AZT resistance mutations coupled with the M184V mutation appear combined to undermine the effectiveness of abacavir.

Julio Montaner of the University of British Columbia presented the CNA30017 study, which had a very similar design to the Swiss abacavir switch study. 211 patients were randomised to switch or stick to existing PI-based HAART. There was no difference in the rate of viral load rebound after 48 weeks, but PI recipients were significantly more likely to have stopped their PI therapy due to adverse events. Three out of four abacavir failures were due to the development of the M184V mutation, while the PI failures were associated with the development of PI resistance mutations.

There was a significantly greater improvement in non-fasting triglycerides in the abacavir group, although both groups had triglyceride levels within the normal range at baseline. Cholesterol levels did not improve significantly in either group over 48 weeks, but once again, fell within the normal range.

In all of the presented studies with the exception of CNA30017, there was no difference in the rate of adverse events between the arms, although Bonaventura Clotet reported that in the Spanish switch study, 5 cases of acute hepatitis occurred, and that liver enzyme elevations were strongly associated with hepatitis C co-infection.

At last week's Second International Workshop on Adverse Drug Reactions and Lipodystrophy in HIV Infection in Toronto, William Powderly noted, in a review of the lessons from switching studies, that studies to date had provided poor baseline data on body fat redistribution with which to evaluate claims that switching was associated with improvements.

Only five randomised studies had evaluated the impact of switching on lipodystrophy until the 40th ICAAC meeting, and none provided convincing evidence of improvement in body fat redistribution, he said.

He also noted that the design of studies had been inadequately thought out, with a failure to factor in the contribution of different nucleoside analogue combinations to fat changes and lipid elevations.

Key lessons

• Switching from a PI-based to NNRTI or abacavir-based therapy doesn’t pose a greater risk of viral load rebound compared to sticking with the supposedly more potent PI-based regimen.

• If the aim of the switch is the treatment of body fat changes, there is still no well controlled data to show that switching results in physical improvements. Well controlled means objective measurements at baseline, and a full description of any improvements. On the other hand, body fat redistribution doesn’t worsen once the switch occurs.

• If the aim of the switch is treatment of high cholesterol or triglyceride levels, it is tempting to assume that the switch will be helpful. However, an important caution is that all the studies reported yesterday started out with a population where the average lipid levels were within the normal range. An analysis of responses in those with abnormally high lipid levels is needed in order to show that very high lipid levels can be improved by switching to a particular drug.

• Most of the switch studies asked patients if they found it easier to adhere, or if the drugs were difficult to take, or if quality of life had improved. In each case they reported improved satisfaction with the new therapy, because it was easier to take.