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

Graeme Moyle, MD, for NATAP

As antiretroviral regimens do not eradicate HIV infection, the current therapy paradigm involves a commitment to indefinite therapy.

Long-term adherence to some regimens, particularly PI-based ones, is challenging due to requirements of high tablet volume, dietary restrictions and increased fluid intake. Improving the poor pharmacology of PIs by using booster co-therapies, such as low dose ritonavir improves administration characteristics but may also increase the number of important drug interactions. Additionally, ritonavir, and other protease inhibitors are known to cause elevation of blood lipids and may be less well tolerated in persons with hepatitis co-infections.

Short-term, effective therapy is associated with an increase in body weight and an improvement in nutritional status, although both plasma triglycerides, total and LDL cholesterol may elevate soon after therapy initiation. Longer-term (> 1 year) therapy has been associated with persistence of these lipid elevations, insulin resistance sometimes accompanied by new-onset diabetes mellitus, and abnormal body fat distribution. This may involve peripheral fat loss, localized or truncal fat gain or both.

It has not been clearly established if all these phenomena are inter-related although unifying theories have been proposed. Both PIs and nucleoside analogues (NAs) have been suggested to play a role in the pathogenesis of these changes and are hence both potential targets of switching strategies. Hypertriglyceridemia and reduced high density lipoprotein (HDL) cholesterol, but not diabetes mellitus or fat redistribution were recognized as complications of HIV-1 infection, and most typically wasting prior to the availability of PI.

A definition for the syndrome(s) to enable consistent case definition in studies does not currently exist. Reported clinical manifestations of the fat redistribution syndrome are heterogeneous and range from central or localised adiposity alone to peripheral fat wasting or combinations of both. Central adiposity, at least in males, appears mainly secondary to visceral fat accumulation and may be best assessed by CT scanning.

Localised problems include breast enlargement, or development of cervical fat pads (or ‘buffalo humps’). Peripheral fat wasting, generally presents with increased vein prominence as well as loss of facial fat pads such as the temporalis and naso-labial pads. The prevalence of these clinical manifestations is not known but appears to increase with time on antiretroviral therapy.

These changes have both significant psychological and social consequences for individuals with HIV infection. The conference included 2 studies evaluating the impact of fat redistribution on quality of life.

A German group evaluated 250 patients who had commence ART in 1996 (mean time since start of ART 36 months) using a standardized physician and patient questionnaire and visual analogue scales. Patients were mostly (80%) male with a mean age of 39 years. Physicians diagnosed lipodystrophy in 36% of the cohort. Patients with lipodystrophy all reported loss of quality of life in at least one of the measured domains: 63% for social contact, 68% on daily performance and sexuality and 83% on self esteem.

The authors also reported some specific physical changes were linked particular diminished ratings. For example, abdominal enlargement was significantly associated with diminished sexual and self esteem ratings, leg and arm changes with daily performance and social functioning [1].

In a survey of 74 attendees at a workshop on body shape 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 agent perceived as causative of the problem was changed. The authors, from Gay Men’s Health Crisis New York, concluded that patient perceptions around metabolic and body shape changes are driving treatment decisions and potentially therapy outcomes [2]. These needs are driving a host of studies evaluating the outcome of therapy switches.

Given the known, theoretical and observational associations of metabolic and clinical changes with PIs, most of the studies currently reporting data have evaluated changes in HIV disease markers, clinical and biochemical parameters following changes from a PI based to a PI sparing regimen.

The reasons for considering switching to a PI sparing regimen include to address metabolic and fat redistribution problems but also:

  • Improve adherence; reduce tablet load, remove food requirements, reduce dosing frequency
  • Reduce the risk of clinically important drug interactions and;
  • Managing an actual or possible PI toxicity including diarrhoea, recurrent renal calculi, etc.

In principle, two main outcomes should be looked at to evaluate the success or otherwise of switching. Firstly, that no harm is done and secondly, that benefit is gained. More specifically this can be measured as:

  • Maintaining of virological control (most reported studies have focused on patients with good virological control on PI-based therapy)
  • Maintaining of CD4 responses (and immune function) and;
  • Improving/resolving/preventing of toxicity – Improving adherence and quality of life.

Three drugs have been evaluated in switching studies. The non-nucleosides efavirenz (EFV) and nevirapine (NVP) and the potent nucleoside analogue abacavir (ABC). In general, the reports are not studies, per se, but observed clinical cohorts, thus the observations must be placed in the context of not knowing what would have had happened to a group of patients maintained on therapy.


Several large studies have looked at switching to nevirapine. In one Spanish study, 91 patients (80% male) were switched to nevirapine after at least 3 months undetectable (<50) on PI therapy. Sixty four percent of them were on their first ever regimen. 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/mm3 (range 80 to 1500). Ten patients (11 %) discontinued nevirapine: 8 due to intolerance and 2 because of sustained viral rebound at 24 and 36 weeks. Both responded to PI reintroduction. Additionally, at week 12, 7 patients showed minimal viral load rebounds (range 50 to 1700 copies/ml) returning to undetectable levels in 6 cases at 24 weeks. Analyses of risk factors for detectable VL were not reported. Mean CD4 rose at weeks 12, 24 and 36 by 1, 65 and 85 cells/mL, respectively. Whereas mean triglyceride at 12, 24 and 36 weeks fell by 125, 16 and 574 mg/dL, respectively [3].

A large 100 patients 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 [4].

A further cohort of patients (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 in Table 1.

After 36 weeks of switching therapy mean values of cholesterol, triglycerides, and glucose were reported to have returned to within normal ranges. Whilst there was 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 to nevirapine 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 week. With regard to tolerability, 6 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 [6].

Table 1.

Mean CD4 (cells/mm3) Viral load (% <80 copies/mL)
0 wks 12 wks 24 wks 36 wks 0 wks 12 wks 24 wks 36 wks
Arm A (28 patients) 393 425 446 410 100% 73% 73% 75%
Arm B (27 patients) 345 376 398 405 100% 75% 80% 77%


Similar data are available with EFV substituting for PIs. A German study evaluated this switch in 42 patients with median time on PI-regimen of 24 months (range: 6-41) (mean age: 43±9.4) with VL <50 cps/ml and CD4+ >200 cells/mm3 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 ZDV/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 66.9±17.1% at baseline to 60.5±18.1% at week 24. Non-fasting triglycerides showed a significant decrease from 322±212 mg/dl at baseline to 252±206 mg/dl at week 24, while values of HDL showed a significant improvement from 37±8 mg/dl to 43±9 mg/dl at week 24. Consistent with clinical trials data, 45.2% of patients reported nervous system symptoms during the first weeks of therapy with EFV [7].

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 in 12 cases. Most patients showed reduction in triglycerides after switch.

After 1 month, tryglycerides had fallen to 362 +/- 233 mg/dl (P< 0.004), at 4 moths 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 [8].

A retrospective analysis of 40 patients who had been switched for any reason from PI-containing antiretroviral treatment (ART) to a regimen of two NRTI plus EFV reported data suggesting lipid levels may not improve impressively after switch to efavirenz but than viral control is generally maintained after switching. All 11 patients in this cohort who had a viral load below 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) [9].

Efavirenz or nevirapine

In another prospective, multicentre, non randomised study of 103 Spanish patients switched 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 (efavirenz 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 [10].

Efavirenz and abacavir

A further approach is to use two of these agents, thus potentially also intensifying the regimen. The safety, efficacy and impact on metabolic parameters of abacavir (ABC) + efavirenz (EFV) when substituted for a PI in persons viral loads with <50 c/ml, was evaluated in an open label, single centre 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 wk 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 stabilisation and one worsening of signs. [11]


There appear to be considerable benefits to switching from a PI-based regimen to a PI-sparing regimen. However resolution of clinical and metabolic abnormalities, which may arise during prior therapy, are incomplete. Patients appear generally pleased with the improved administration characteristics of the new regimens and improvements in quality of life have been reported. Clinical and metabolic benefits include a reduction in insulin resistance and subsequently intra-abdominal fat, and, most prominently, with nevirapine and (previously) abacavir, improvements in total cholesterol and triglycerides.

Cholesterol and triglyceride abnormalities may not resolve after switching to efavirenz although HDL:LDL ratio may improve. Peripheral or subcutaneous fat mass improvements are not evident in all studies reported to date. Weight gain, probably in part due to removal of PI-related dietary restrictions, has been observed and may lead to improvements in appearance. Lack of peripheral fat gain may in part relate to a contribution from nucleoside analogues to peripheral fat wasting.

Maintenance of virological control varies between studies. This may be due to two factors: the relative potencies of the drugs used, and the extent of prior drug exposure (or drug resistance) in patients entering the studies. In switch studies, all agents appear effective at maintaining virological control in persons with no history of prior treatment failure.


Unless stated otherwise, references are to the Programme and Abstracts of the XIII International AIDS Conference, 9-14 July 2000, Durban, South Africa.

  1. Goetzenich A, Corzillius M, Mauss S, et al. The impact of lipodystrophy on quality of life. WePpB1381
  2. Kasper TB, Arboleda CH, Halpern M. The impact of patient perceptions of body shape changes and metabolic abnormalities on antiretroviral therapy. WePpB1380.
  3. Gorgolas M, Estrada V, Arranz A, et al. Efficacy of protease inhibitor switching to Nevirapine in patients on HAART with undetectable viral load. WeOrB485.
  4. Piroth L, Buisson M, Grappin M et al Simplified maintenance therapy with NNRTI (nevirapine) in patients with long term suppression of HIV-1 RNA: first results of a cohort study. WePeB4167.
  5. Ladisa N. Efficacy and tolerability of switching from a protease inhibitor (PI)-containing regimen to PI-sparing non-nucleoside reverse transcriptase inhibitor (NNRTI)-containing regimen in patients with undetectable viral load and metabolic disturbances. WePeB4157.
  6. Tebas P, Yarasheski K, Henry K , et al. Evolution of multiple metabolic parameters after the switch of protease inhibitors to nevirapine. ThPpB1485.
  7. Knechten H, Hšhn C, Sterner C, Braun P. Sustained viral suppression after substitution of Efavirenz (EFV) for protease inhibitor (PI) in a regimen with Lamivudine (3TC) and Stavudine (D4T) or Zidovudine (ZDV). WePeB4154.
  8. Maggiolo F, Migliorino M, Rizzi M, et al. PI-associated metabolic changes: switch therapy with efavirenz in virologically controlled HIV+ persons. ThPpB1484.
  9. C. Hoffmann, E. Jaegel-Guedes, E. Wolf et al. PI to Efavirenz (EFV) switch effect on lipids in HIV-positive patients. WePeB4185.
  10. Casado JL, Munoz V, Arrizabalaga J, et al Maintenance therapy with non-nucleoside reverse transcriptase inhibitors for HIV-infected patients with intolerance to a successful protease inhibitor-containing regimen (BEGIN study). WePeB4171.
  11. Bickel M, Rickerts V, Klauke, S et al The Protra Study: To evaluate the safety and efficacy of Abacavir (ABC) and Efavirenz (EFV) when substituted for a PI in HIV-1 infected subjects treated with 2 NTRIs and a PI with plasma HIV-RNA levels below the limit of assay quantification. TuPeB3210.

Links to other websites are current at date of posting but not maintained.