Switching to Protease Inhibitor Sparing Studies

Can viral load be maintained and metabolic or fat redistribution be significantly improved?

Two studies at Durban studies reported an evaluation of the impact of fat redistribution on quality of life using a standardized physician and patient questionnaire and visual scales. In the German study (n=250) patients were mostly (80%) male with a mean age of 39 years. Physicians diagnosed lipodystrophy in 36% of the study group. Patients with lipodystrophy all reported loss of quality of life in at least one of the measured parameters: 63% for social contact, 68% on daily performance and sexuality and 83% on self-esteem. Abdominal enlargement was significantly associated with diminished sexual and self-esteem ratings, leg and arm changes were associated with daily performance and social functioning. In a survey of 74 attendees at a workshop at GMHC in NYC on body changes where 78% of respondents reported body shape or metabolic changes thought related to ART, 30% of these individuals had changed therapy and 7% interrupted ART specifically due to these problems. In general (83% of cases), amongst those changing therapy, only the drug perceived as causing the problem was changed.

In general, the studies switching to PI-sparing are not comparative. It's difficult to compare to another arm, because patients either want to switch or they don't, therefore randomization is difficult. The ACTG is planning a study comparing a switch to either an abacavir, EFV, or NVP regimen. In general, switches to NVP or EFV arms do not appear to improve body changes so far in follow-up of ongoing studies which extends to about 1 year. A Spanish NVP switch study (Lydia Ruiz reported at Lisbon Conference in Nov. 1999, details on NATAP web site) suggested improvements in lipoatrophy at 1 year. It's possible that longer follow-up is required to see improvements. Preliminary data on a switch to an abacavir regimen at the February Retrovirus Conference suggested improvements in body changes, and follow-up data is expected to be presented at the end of September at ICAAC. One study below switches from PI to EFV+abacavir regimen and patient reports indicate improved body changes. Switching to NVP or abacavir does seem to generally improve lipid abnormalities but not complete resolution, while improvements to lipids when switching to EFV is inconsistent (2 studies below reported reduction in triglycerides) and in general does not appear to improve lipids, although HDL cholesterol appears to improve. The ACTG is planning an interesting study combining Metformin and Rosiglytazone, two anti-diabetic drugs. Metformin may decrease fat accumulation, but it also may decrease fat in the periphery. Rosiglytazone may increase periphery fat. So, it's hoped the two together may improve body changes.


In a Spanish study, 110 patients (80% male) were switched to nevirapine after at least 3 months undetectable (<50) on PI therapy. 72% were naive to antiretrovirals prior to starting the PI regimen they were receiving at the beginning of the study. 62% had viral load of 100,000 copies/mL or more when their PI-containing regimen was started. The reasons for changing PI included treatment simplification (41 patients), renal complications (24), lipodystrophy (20), hypertriglyceridemia (3), and gastro-intestinal disturbances (3). The mean CD4 count at change was 601 cells (range 80 to 1500). After 1 year, 21% discontinued nevirapine. Three were lost to follow-up, 3 were receiving methadone and elected to discontinue NVP, 2 had hepatitis, and 3 had an increase in hepatic enzymes. 7 patients had a severe rash and 5 had viral rebound, of which 4 were naive to therapy prior to their PI-containing regimen and 2 had HIV RNA higher than 100,000 copies/mL at that time. Mean CD4 rose at weeks 12, 24 and 36 by 1, 65 and 85 cells/_L, respectively. Whereas mean triglyceride at 12, 24 and 36 weeks fell by 125, 16 and 574 mg/dL, respectively.

A large 100 patient French cohort also evaluated switching to NVP in PI treated patients with plasma VL <20 copies/ml for >1 year. Data at month 3 on 72 patients indicated a significant rise in CD4 percentage (28.4% to 30.6%) and a rise in CD4 cells of 39/ mm3. Seven patients (10%) had detectable viral load levels at month 3; available resistance data suggesting resistance only to NNRTI was present. There were no significant changes in cholesterol or triglycerides. Twenty-one patients reported side effects following switch with 5 patients discontinuing nevirapine.

Another cohort of patients who mostly switched to NVP yielded somewhat disappointing results. In this single centre Italian study, a total of 28 NNRTI-naive, PI-treated patients with a VL< 80 copies/ml for mean 21 months (range 7-33), and with metabolic toxicities were switched to NVP (n= 24) or EFV (n=4). These patients (Arm A) were compared to a cohort who continued the PI-containing regimens (Arm B). The groups were well matched for age, sex, HIV risk factors, CD4 count when initiating PI therapy and switching therapy, and for the duration with an undetectable viral load. Results after 36 weeks are reported below: (See Table 14)  

After 36 weeks of switching therapy mean values of cholesterol, triglycerides, and glucose were reported to have returned to within normal ranges. While there were no statistical difference between the percent of patients with undetectable viral load in the two treatment arms, it is concerning that as many as 25% of patients in this cohort lost virological control, the majority in the first 12 weeks after switch (5).

Concern that switching therapy may be more risky in persons with prior therapy experience led one group to investigate this approach only in persons on their first regimen. In this study, 40 patients on PI regimens with undetectable VL for at least 6 months, switched the PI for NVP 200-mg bid. Nucleosides were maintained. 90% of the patients were men, with a median CD4 of 511 cells/mm3. The median follow up was for 30 weeks. Six patients (15%) developed severe rash and were switched to EFV. One patient discontinued due to hepatitis. Only 1 patient (2.5%) lost viral control. Metabolic benefits were observed with 24 week results showing triglycerides had declined 44% (p > 0.05), HDL-cholesterol increased 26% (p > 0.001), and glucagon decreased 23% (p = 0.01). LDL cholesterol, glucose, insulin and proinsulin did not alter significantly. Hip and spine bone mineral density and truncal adiposity did not change.


A German study evaluated 42 patients with median time on PI-regimen of 24 months (range: 6-41) (mean age: 439.4) with VL <50 copies/ml and CD4+ >200 cells at baseline. Baseline therapy switch included one PI in 23 patients (IDV: n = 13; NFV: n = 9; RTV: n = 1), and two PIs in 19 patients (RTV/IDV, n = 15; RTV/SQV, n = 1, NFV/SQV, n = 3). NRTIs were D4T/3TC(n = 29), or AZT/3TC (n = 13). VL was undetectable for a median time of 12.5 months before substitution. Therapy was changed due to adverse events to PI (n = 11), lipodystrophy (n = 20) and/or wish to simplify regimen (n = 15). VL of all patients (39 with available data, 2 lost to follow up) has remained <50 copies/ml and a significant reduction of CD8+/CD38+ cells from 66.917.1% to 60.518.1% at week 24. Non-fasting triglycerides showed a significant decrease from 322212 mg/dl at baseline to 252206 mg/dl at week 24, while values of HDL showed a significant improvement from 378 mg/dl to 439 mg/dl at week 24. Consistent with clinical trial data, 45.2% of patients reported nervous system symptoms during the first weeks of therapy with EFV.

An Italian study specifically investigated the effect on switching in persons with elevated triglycerides and viral load <500 copies/ml from PI to efavirenz. Patients continued with the same NRTIs. Thirty-one patients receiving either IDV (14), RTV (4), SQV (2), NFV (3) SQV+RTV (7) or SQV+NFV (1) for 19.8 +/- 8.2 months were enrolled. Before PI treatment, their mean plasma triglyceride was 219 +/- 132 mg/dl. At switch this value was 718 +/- 453 mg/dl and was associated with lipodystrophy in12 cases. Most patients showed reduction in triglycerides after switch. After 1 month, tryglicerides had fallen to 362 +/- 233 mg/dl (P< 0.004), at 4 months 325 +/- 157 (P<0.008) and at month 8 359 +/- 149 mg/dl (P> 0.03). Total cholesterol blood levels were unchanged. Mean HIV-RNA dropped from 90 copies/ml to 49.3 copies/ml after one month. At 4 months all patients were <50 copies/ml and at 8 months all but one patient was <50 copies/ml. CD4 T-cell levels increased from 525 cells/mm3 to 625 cells/mm3. Treatment with efavirenz was generally well tolerated; only one patient interrupted therapy because of dizziness.

A retrospective analysis of 40 patients who had been switched for any reason from PI-containing antiretrovital treatment (ART) to a regimen of two NRTI plus EFV reported data suggesting lipid levels may not improve after switch to efavirenz but that viral control is generally maintained after switching. All 11 patients in this cohort who had a viral load <50 copies/ml prior to switch maintained control over an observation period of 6 months. Median non-fasting cholesterol rose from 201 mg/dL (range: 71-426) at baseline level to 218mg/dl (range of change 49 to +156)(p = 0.01); and triglycerides from 221 mg/dL (range: 66-1957) to 238mg/dl (range of change -433 to+853)(p = ns).

Substitution of Efavirenz for PI

C. Katlama reported on this French study of 165 randomized to continue their PI regimen or switch to EFV (600 mg once daily). These were individuals who were likely to have been on their first-line regimen upon entering study, although this was not mentioned in the poster. They were on their PI regimen for about 2 years, were NNRTI nave, and had 3 consecutive <50 copies/ml viral load measures before switching. So this group was well suppressed virologically and was likely adherent.

The NRTIs were maintained. This was a multi-center open-label study whose objective was to see the effect of switching on the durability of viral suppression in individuals who had viral suppression <50 copies/ml on a PI regimen. The mean duration of PI treatment was 23.4 (7.8) months in the EFV arm and 24.1 (7.0) months in the PI arm. There was a third arm in the study consisting of a substitution regimen of EFV + PI(s) which was discontinued because of poor acceptance by patients and investigators. Virologic failure was defined as confirmed on two consecutive assays plasma viral load levels ≥50 copies/ml.

At baseline mean CD4s were about 500 in both arms (range 113-1319). Over 90% of participants were white. Over 80% were men. The PI(s) individuals were on in the EFV arm (n=69): IDV 64%, NFV 26%, RTV 4%, RTV/IDV 2%, and RTV/SQV 2%. In the PI arm (n=65): IDV 60%, NFV 26%, RTV 12%, RTV/SQV (2%), and RTV/IDV (0).

At week 24 in the ITT-analysis, 67/69 (97.1%) (n=69) in the efavirenz group had <50 copies/ml; 54/65 (83.1%) in the PI(s) arm had <50 copies/ml (statistically significant difference between groups, p=0.0076). Using the Observed Data analysis, 67/68 (98.5%) in the EFV group had <50 copies/ml; in the PI(s) arm, 54/58 (93.1%) had <50 copies/ml (no statistical significance, p>0.05). CD4 counts were maintained in both groups --495 in EFV arm at week 24 and 554 in PI arm at week 24.

Regular cholesterol did not improve after switching to EFV. The investigators compared cholesterol changes from baseline between the 2 groups. In the EFV arm, mean baseline cholesterol was 228 mg/dL (standard deviation 43), 244 (SD 41) at week 12, and 239 ( SD 45) at week 24. In the PI arm, mean baseline cholesterol was 210 (SD 45), 222 (SD 56) at week 12, and 220 (SD 51) at week 24. Increases in mean HDL concentration were significantly greater at week 24 in the EFV arm: in the EFV arm HDL (good cholesterol) was 45.7 mg/dL (SD 15.1) at baseline, 49.7 (SD 15.5) at week 12, and 49.0 (SD 13.8) at week 24; in the PI arm, HDL was 42.1 at baseline (SD 14), 43.2 (SD 15.6) at week 12, and 42.1 (SD 12.5) at week 24.

CNS related adverse experiences reported in the EFV arm (there was statistical significance between treatment groups for these adverse experiences): dizziness (29%); insomnia (18.8%), 7.7% reported in PI arm; dreaming abnormally (23.2%); vertigo (20.3%); depression (10.1%), 3.1% in PI arm and the difference in depression was not statistically significant; concentration impairment (10.1%).

Efavirenz or nevirapine

In a prospective, multicentre, non-randomized study of 103 Spanish patients with a viral load <200 copies/ml (86% <50) (mean 12 months, range 1-34) on a PI regimen (mean 18 months range 4-35) switched to either efavirenz or nevirapine. Reasons for NNRTI choice were not stated. Twenty three percent of patients were on their first regimen. After 24 weeks, 93% of 33 efavirenz patients and 87% of 70 nevirapine patients remained <50 copies/ml in the as-treated analysis, and 78% and 75% in the intent-to-treat, respectively. Thirteen percent of patients withdrew due to adverse events. Median CD4 count increased from 502 to 591 cells/mm3. Patients with a longer previous time on NRTI therapy (42 vs, 33 months), and a shorter time with undetectable HIV load (8 vs 12 months), had a higher rate of viral failure (efavienz and nevirapine groups analysed together). Adverse events secondary to NNRTI were observed in 27 patients (31%), mainly hepatitis (9%) or rash (6%). Triglycerides and cholesterol levels significantly improved after switching to the NNRTI. For triglycerides from 321 to 246mg/dl (p = 0.01) and for cholesterol from 239 to 219mg/dl, p = 0.0001). Clinical manifestations of lipoatrophy were not noted to improve.

Efavirenz and abacavir

Use of two of these agents intensifies the regimen. The safety, efficacy and impact on metabolic parameters of Abacavir (ABC) + Efavirenz (EFV) when substituted for a PI in persons with viral loads <50 c/ml was evaluated in an open label, single center study. To establish tolerability, ABC was added at baseline, then at week 6, EFV replaced the PI. Twenty-six patients have been followed-up for a mean 24 weeks. Four discontinued at week 2 (3 for ABC-hypersensitivity, 1 for virological failure due to non-compliance). No grade 3 or 4 laboratory toxicities were reported. After stopping PI, all patients have remained <50 with stable CD4 counts. Mean fasting-triglycerides, HDL and LDL cholesterol from the first 16 patients who reached week 24, improved significantly compared to baseline. Two of 4 diabetic patients saw resolution of diabetes. Five of 8 patients with lipodystrophy self reported improvements, 2 stabilization, and one worsening of signs.