EACS: Switching to atazanavir maintains viral suppression, even in patients with protease resistance

Two studies presented at the Eleventh European AIDS Conference (EACS) have found that switching from other protease inhibitor-based regimens to ones based on boosted atazanavir (ATV/r, Reyataz/Norvir) maintains viral suppression and results in some falls in lipids (blood fats).

A third study found roughly similar failure rates in patients switched from LPV/r regimens to either boosted or unboosted atazanavir, and also similar decreases in blood lipids, but found significantly higher CD4 increases in patients taking the unboosted drug.

The first study, the ATAZIP study, was a randomised one in which patients either switched from boosted lopinavir (LPV/r - Kaletra) to ATV/r or stayed on it.

The second, the SIMPATAZ study, was an open-label one in which patients in 27 Spanish HIV clinics switched from any current PI-containing regimen to an ATV/r-based one.

Main 48-week results from the ATAZIP study were presented at the International AIDS Society Conference in Sydney in July. In brief, 248 patients were randomised either to stay on LPV/r-based regimens or switch to ATV/r. After 48 weeks the proportions who had failed treatment for any reason and had to change or stop therapy were 17% on ATV/r and 20% on LPV/r, a non-significant difference. The proportion who failed virologically, meaning a viral rebound to over 200 copies, was 5% on ATV/r and 6% on LPV/r.

The poster presented at EACS was a sub-analysis of patients who had either failed on previous PI-containing regimens, had PI resistance mutations, or both. This group numbered 84 patients. Only three patients who underwent resistance testing had no PI mutations, though about a third of patients did not receive a resistance test at all, a surprising omission in a clinical trial. Nearly all of those tested had some PI resistance mutations and the majority of those tested (69% on ATV/r and 74% on LPV/r) had at least one primary PI resistance mutation. Only five patients discontinued the study, and only one for virological failure (in the LPV/r arm).

In this subset of experienced and/or resistant patients, 18% on ATV/r and 15% on LPV/r had failed treatment for any reason by 48 weeks, actually a slightly lower proportion than in the general study. But, not unexpectedly, the proportion who failed because of viral rebound was higher, with 11% of patients on ATV/r and 13% on LPV/r failing virologically. In terms of blood lipids, patients switching to ATV/r had, relative to the LPV/r patients, significant falls in triglycerides and total cholesterol, a non-significant fall in ‘bad’ LDL-cholesterol levels, and no change in ‘good’ HDL-cholesterol.

In the SIMPATAZ study, patients who had been on any PI-containing regimen for at least six months and with undetectable viral load for at least four months switched to ATV/r. This is a twelve-month study, but for the EACS conference, 8-month results (three out of four monitoring visits) were presented.

One hundred and eighty-one patients enrolled in the study and to month eight there were 17 discontinuations (9%). Six of these were patient choice, seven were lost to follow-up, and of the other four, two died (one lung cancer, one heart attack), one developed Hodgkin’s lymphoma and one patient failed virologically.

Eighty per cent of patients had had previous exposure to at least one PI, and 46% had failed at least one other ARV regimen, most of them (a third of the whole group) another PI-containing regimen. Previous PI regimens consisted mainly of either LPV/r-based of nelfinavir-based ones; 62% had switched from another boosted PI.

Of note, because this was a Spanish patient group, in 47% of patients the HIV risk factor was injecting drug use, 52% were co-infected with either hepatitis B or C and 17% were on methadone maintenance therapy. Liver enzymes did not increase during the study, except for the expected rise in unconjugated bilirubin, a well-known side effect of atazanavir.

Of note, NRTIs could be switched at the same time, and it was up to doctor and patient to choose the drugs; as a result the proportion taking AZT went down from 27% to 19% and the proportion taking tenofovir increased from 11% to 29%; in addition, 19% of patients started taking abacavir, none having done so before.

After eight months, the proportion with an ‘undetectable’ viral load fell from 100% to 91%. Undetectable meant below 50 in most cases but 15 patients (8%) were at clinics that still used assays with a 400 copy cut-off. When these were excluded, the proportion of the whole study group with viral loads under 50 copies/ml fell from 92% to 80%.

There were significant falls in triglycerides and total cholesterol and a slight, but significant fall in LDL-cholesterol; HDL-cholesterol stayed the same.

In the third study, Nicola Gianotti of the St Raffaele Institute in Milan switched 184 patients from LPV/r based regimens to either ATV/r (124 patients) or unboosted ATV (60). What to switch to was not randomised, and patients switching to the boosted drug were slightly more treatment experienced (an average of 18 months on ARVs versus 15.5 months) and had significantly higher CD4 counts (542 versus 473, p=0.04). The average length of time spent on ATV/r or ATV in the study was 18.5 months.

Roughly equivalent numbers of patients on ATV/r and ATV experienced a resurgence of virus; 9.8% rebounded on ATV/r and 10.3% on ATV, with the average time to virological rebound the same in both arms. Of note, only 8% on ATV rebounded if patients taking thymidine analogues (AZT or d4T) were excluded.

There were significant falls in triglycerides and total cholesterol, with the advantage to unboosted atazanavir, but only in the case of triglycerides was this advantage statistically significant.

There was a significantly greater CD4 increase in patients taking ATV compared with those on ATV/r, with increases of 14 cells (4%) on ATV/r and 62 (15%) on ATV. Gianotti was asked if this was simply due to the fact that patients on unboosted ATV were starting from a lower CD4 baseline, but she said other studies had shown similar differences between the boosted and unboosted drug.