Switching Studies

The studies discussed below look at switching from a PI regimen to a NNRTI or an abacavir triple nuke regimen. One key point when considering a switch is that you want to maintain viral suppression. If a person has undetectable viral load, you want to make sure that the regimen you switch to is capable of maintaining viral suppression. NRTI experience prior to the person's initial PI regimen may affect this. In two studies discussed below, NRTI experience prior to the person's PI regimen increased the risk for viral rebound for some individuals. In particular, in an abacavir study described below persons with AZT experience prior to starting a PI HAART regimen, had a higher failure rate than persons who did not have prior AZT experience. Otherwise, there were few surprises reported in these studies. The results reported here are similar to switch study results previously reported.

Efavirenz shows a trend toward better likelihood for maintaining viral suppression. When considering a switch to a NNRTI, you should consider the potential side effects and toxicities as well as the long-term durability of viral suppression. Accompanying efavirenz are CNS side effects which in general tend to disappear within a few weeks after starting the drug, but in a small number of cases the side effects can remain at a low level. In which case, a person can switch to another drug if they wish. If a person has hepatitis C, they may not want to take nevirapine because of what appears to be a higher rate of hepatoxicity than from abacavir or efavirenz. Total cholesterol & triglycerides more consistently improve after switch to nevirapine or abacavir. HDL (good cholesterol) improves after switch to efavirenz without change in total cholesterol. Efavirenz is approved to be taken once daily. Nevirapine is approved in the USA only to be taken twice daily but is used sometimes once daily.

Although the studies reported at ICAAC and discussed below did not address glucose and insulin, William Powderly from Washington University in St Louis discussed this in a general talk he delivered at ICAAC. He said that there is a considerable amount of consistency from studies showing improvements in measures evaluating glucose and insulin after switching from a PI to efavirenz, nevirapine or abacavir regimens. Triglycerides improve after switch to NVP in most studies. Several studies in EFV show mixed results regarding improved triglycerides--improved in some but not in others. In general, even when lipids improve they do not return to normal. Triglycerides appear to improve after switch to abacavir. Although usually cholesterol improves after switch to NVP, one or two studies did not show improvement. These types of improvements address the concerns that the risk for premature heart disease may increase for HIV-infected individuals experiencing elevations in cholesterol, triglycerides, and glucose and/or insulin resistance.

It is very difficult to assess body change improvements, because often the studies did not properly characterize baseline patient situations or properly evaluate patients throughout the study. However, when objective measures of evaluation have been used so far none of these studies have shown improvements in body changes. Improvements may take longer to emerge and the longest switch studies are only out to about 1 year. Several studies have reported subjective improvements reported by patients or doctors. Sometimes you hear of anecdotal reporting of improvements for some individuals after switching to PI sparing regimen.

From PI to Efavirenz

Study DMP 266-049 is a randomized, open-label comparison of switching from a PI regimen to an efavirenz regimen or remaining on the PI regimen. The objectives of this study are to see if suppression of viral load (<50 copies/ml) can be maintained after the switch and to evaluate tolerability and adherence. Patients were required to be <50 copies/ml at screening. Patients remained on their PI arm an overlap period of 7 days after starting efavirenz. Patients were randomized 2:1--226 to the efavirenz (EFV) arm and 120 to remaining on the PI.  

About 90% in both arms were men. About 15-20% were Black and 8% Hispanic. Mean baseline CD4 was about 570 in each arm. The mean duration of previous PI treatment was 20 months in the EFV arm and 22 months in the PI arm. Most study participants were taking either indinavir (44-46%) or nelfinavir  (26-30%). 12% were taking a double PI arm in the EFV arm: 8 patients were taking RTV/SQV (4%); 6 patients (3%) were taking NFV/SQV; 11 (5%) patients were taking NFV/IDV. In the PI arm 7% of patients were taking a double-PI regimen: 3% NFV/IDV; 1% RTV/SQV; 3% NFV/SQV. A little over 50% were taking AZT/3TC and about 30% were taking d4T/3TC; about 5% were taking d4T/ddI.

Through 24 weeks, 9 (4%) and 15 (12.5%) patients were randomized but not dosed in the EFV and PI arms, respectively. Discontinuing for any reason were 6.9% (15) in the EFV arm and 12.4% (13) in the PI arm. Discontinuing for adverse events were 2.8% (6) in the EFV arm and 2.0% (2) in the PI arm. Losing virologic control (>50 copies/ml on 2 occasions) were 3% (6 patients) in the EFV arm and 10.2% (9 patients) in the PI arm (=0.011). Using an ITT non-completer equals failure analysis which excludes patients randomized but not dosed, 89% (187/210) in the EFV arm and 81% (80/99) maintained viral suppression <50 copies/ml at week 24 (p£0.05). Using an observed data analysis, the difference between the 2 arms was not statistically significant: 95% (189/199) in the EFV arm and 92% in the PI arm (83/90) maintained viral suppression <50 copies/ml at week 24. There was no difference between the 2 arms in CD4 change: mean CD4 in EFV arm increased from 578 to 611 and in PI arm 566 to 621. Lipids were tested non-fasting and there were no changes in cholesterol levels in either the EFV or PI arm. HDL levels (good cholesterol) increased significantly in the EFV arm but not in the PI arm. And LDL (bad cholesterol) decreased non-significantly in the EFV and in the PI arm. Triglycerides rose non-significantly in both groups but more in the PI arm. There were no significant changes in mean liver enzymes (AST & ALT) but there was a significant increase in GGT in the EFV arm, which presenter said is consistent with the known association of EFV with this marker of hepatic enzyme induction. The reported number of patients with adverse events „Grade 2 reported in „2% of patients: 30% (65) in the EFV arm and 17% (18) in the PI arm. As expected the side EFV side effects reported were: abnormal dreaming 7%, rash 5%, dizziness 4%, insomnia 3%, concentration impaired 3%, hypertriglyceridemia 3% (versus 6% in the PI arm, neuropathy 2%, diarrhea <1%.

At Durban, a similar but smaller study with similar findings was presented (DMP-027). 165 patients who were well suppressed virologically were randomized to remaining on a PI regimen or switching to EFV. The NRTIs were maintained. Patients were on their PI regimen for about 2 years and had 3 consecutive monthly <50 copies/ml viral load measures. Using an ITT analysis at week 24, 97% (67/69) in the EFV group and 83% (54/65) of the individuals remaining on the PI regimen had <50 copies/ml (p=0.0076). Total cholesterol did not improve over the course of 24 weeks in either the EFV or PI arms. But HDL cholesterol significantly improved in the EFV arm but not in the PI arm.

From PI to Abacavir: with & without prior NRTI experience

Below are 2 abacavir studies comparing either a switch to an abacavir regimen or patients remaining on their PI regimen. Both studies are randomized and open-label. In the first study patients were permitted to enroll with prior NRTI experience, but in the second study patients were treatment naÔve.

Miles Opravill from the University Hospital in Zurich Switzerland reported the final results of this randomized open label study (SMT Study- Simplified Maintenance Therapy) which compares continuing on a current PI regimen or switching to an abacavir based regimen. There is a minimum of 48 weeks follow-up for each patient. His data shows overall treatment failure to be similar between the two arms, but virologic rebound occurred more often in the abacavir arm. The similar overall treatment failure rate appears likely due to improved adherence and tolerability/adverse events in the abacavir arm compared to the PI arm. Prior AZT use before starting HAART increased  virologic failure rate about 4-fold in both arms: 2.4% to 10.5% in PI arm, and 7.5% to 29% in the abacavir arm. Adherence <95% was twice as good in the abacavir arm compared to the PI arm by both patient and nurse assessment.

Patients were on a PI containing HAART regimen, and could have had prior NRTI experience (46% had prior mono/dual NRTI experience). They had undetectable HIV viral load for >6 months. At screening they were negative for the 215 AZT resistance mutation (PBMC DNA), and had <50 copies/ml. Patients were randomized to continue on their PI regimen or switch to abacavir+Combivir and in March 2000 were able to switch to Trizavir, which is one pill containing all 3 drugs--abacavir, AZT, and 3TC.

Virologic failure was defined as HIV viral load >400 copies/ml on 2 consecutive visits (usually 4 weeks apart)--ITT analysis done (switch included). Treatment failure was defined as a change/stop of randomized treatment or virologic failure, and ITT analysis performed in which switch=failure. Additional measures and substudies: adherence (self and study nurse reported); blood lipids; genotypic resistance; CD8 activation markers (CD38 and HLA-DR; HIV-RNA and DNA in lymphoid tissue.

BASELINE

There were 79 and 84 patients randomized to continuing PI and simplified abacavir regimen, respectively. About 80% of study participants were male. CD4s were 512. The HIV-RNA was below the level of quantification for about 15 months in both groups. Opravill did not specify if this meant they were <50 copies/ml for 15 months. Earlier in his talk he said viral load was undetectable for >6 months and <50 at screening. The abstract in the program said patients had plasma HIV-RNA <50 copies/ml for „6 months, implying patients had viral load <50 copies/ml for 15 months. The mean duration of prior therapy was 28.6 and 23.3 months in the comtinuation & simplified arms, respectively. Patients had been on the PI regimen for a mean 21.6 and 19.8 months in the continuation & simplified arms, respectively. 51% in the continuation arm & 42% in the simplified arm had prior mono/dual NRTI experience. Prior AZT use was 46% in the continuation arm & 37% in the simplified arm. Prior 3TC use was 18% and 15% in continuation & simplification arms, respectively. And all these patients had prior AZT therapy.

RESULTS

The median duration of follow-up is 68 weeks in both arms. There were 2 patients in the PI continuation arm with virologic failure, who did not change therapy. And there were 13 virologic failures in the abacavir arm who had not changed therapy (ITT, switch=failure). There were 25% treatment failures in abacavir arm and 29% in the PI arm, about the same.

The number of virologic rebounds above 50 copies/ml was not reported, only >400 copies/ml on 2 consecutive visits. More patients in the PI arm than the abacavir arm (21 vs 8) prematurely discontinued treatment due to changing one or more drug for adverse events, discontinued or lost to follow-up. But more patients had virologic failure on their originally randomized treatment on the abacavir arm than the PI arm (13 vs 2). Overall, there were 6% (5) virologic failures in the PI continuation arm vs 15% (13) in the abacavir arm because this included patients who switched to a different regimen (switch included).      

*Abacaivir hypersensitivity reaction.

Opravill reported time to probability of treatment failure (ITT; non-completers=failure) was no different between the two arms (p=0.73) over 2.5 years. The Kaplan-Meier Curve (ITT) Time to Virologic Failure showed a trend to favoring less likelihood of failure in the PI continuation arm (p=0.061). These data reflect the percentages noted above suggesting that patients tend to tolerate abacavir better (less changing therapy due to AE or discontinuation/lost to follow-up), but there were more virologic rebounds in the abacavir arm. And when you consider both AEs/tolerability and virologic failure the two arms perform equally. When considering which drug to switch to, you should look at the efavirenz and nevirapine data from various studies. Consider and compare virologic failure rates and discontinuations due to side effects, toxicities, tolerability and adherence. And consider your clinical experience with tolerability for all these drugs, and the patient's individual situation. Of course, we are always warned not to compare between studies but when there are not direct head-to-head comparisons, we are left to such comparisons. The ACTG is planning a comparing of a switch from a PI to nevirapine, efavirenz or abacavir, but I don't think there will be a PI continuation arm. It is difficult to randomize individuals to continuing PT or switching because patients tend to either want to remain on their PI regimen or switch to a different regimen.

PRIOR AZT THERAPY

12 of the 13 failures either had prior AZT experience often accompanied by 2 or more AZT/3TC mutations, or had no reported prior AZT use but had AZT mutation(s). Two individuals had 17 weeks prior AZT experience, 1 individual had 27 weeks prior experience, but the others had 38, 43, 44, 75, 100+, and 100+ weeks experience.  The AZT experience of the 13 patients with virologic failure in the simplified arm (abacavir), 9 had prior AZT experience (17-105+ weeks). When looking at PBMC DNA at baseline 5/9 with prior AZT experience, no AZT or 3TC mutations were seen.  In the 4 individuals with AZT use prior to HAART in whom mutations were seen, 3/4 had 2 or more mutations (41/184, 67/70, 67/215/219, 184). As you can see 1 patient had a 215 mutation (person had extensive prior AZT experience, 3+ years). Two individuals for whom no prior AZT use was noted had AZT and/or 3TC mutations. Some of the 13 failures had Prior to the study switch to abacavir they had the following AZT & 3TC mutations in PBMC DNA- 41, 67, 70, 184, 215, 219; the 215 mutation was present in only one person at baseline prior to switch. 4 out of 13 with virologic failure had no prior AZT experience and Opravill reported 2/4 had good adherence. Most of the patients who had reportedly good adherence who failed (n=8) had prior AZT experience (6/8) (sometimes with 3TC experience). And individuals who failed virologically within 8 weeks after switch all had prior AZT experience but 3 of the 8 had no AZT mutations in PBMC DNA at baseline. After failure in this study, individuals almost invariably had the 215 AZT mutation. Since the study required for entry that no 215 be found, AZT resistance had evolved. The patients had developed additional AZT & abacavir mutations to the ones they had at baseline in PBMC DNA. One patient who failed and was reported to have good adherence and had reportedly no prior AZT experience had at failure only the 184 mutation.

Pre-Treatment With Prior AZT mono/dual Therapy Pre-HAART Predicts Virologic Failure

Viral rebound was in part explainable by prior AZT/3TC use, illustrating that simplification from a PI to any NNRTI or an abacavir regimen in a person with prior NRTI experience and/or resistance must be approached with caution. In such experienced individuals you risk losing viral control and developing resistance to NNRTI or abacavir. In this study, patients with prior AZT mono/dual therapy had a higher failure rate in the abacavir simplification arm than patients with no prior AZT experience (29% vs 7.5%).                               

Failure Rate

Exposure to AZT before HAART was started was a significant predictor of failure. Among all patients the prior AZT use increased the odds ratio for rebound by 4.13 (CI 1.28-15.3; p=0.010). Opravill reported that the odds ratio for viral rebound was increased by 5.01 (CI 1.21-24; p=0.013) in the simplified arm and 5.00 in the PI arm (CI 0.46-252; p=0.18 NS).

Tonsillar biopsies were taken from 23 patients (9 PI arm, 14 abacavir arm) from patients with <50 copies/ml between 24 & 48 weeks into study.  Opravill reported that after one year of study HIV RNA and HIV DNA were the same in lymph tissue in tonsil for both the PI and abacavir arms. Median RND & DNA were not statistically different and were approximately (as best I could see from slide) below 10 copies/ml.

Changes in CD4 and CD8 lymphocytes were about the same in both groups. CD38+ and HLA-DR activation markers on CD8 lymphocytes activation markers were looked at in a subset of patients, and did not differ between the groups. These markers are used as a surrogate marker for persisting viral activity.

Non-fasting cholesterol decreased significantly in the simplification arm (-1.2 mmol/L at week 48). On graph it looked like cholesterol went down from 225 mg/dL at baseline to about 185 at week 96. Change from baseline value between arms was statistically significant (p<0.001 week 4-48; p<0.05 up to week 90). It appears as if improvement occurred 4 weeks after the switch, and remained so after two years of study follow-up. Cholesterol in PI arm was about 225 mg/dL at baseline and was about the same at weeks 48 and 96. There was no significant change in HDL fraction (good cholesterol).

There was a trend towards a decrease in non-fasting triglycerides, but significance was not at all time points. At week 48 triglycerides were decreased from baseline in abacavir arm by 1.0 mmol/L, or as it appeared by visual observation of slide to be from about 250 mg/dL at baseline to 175 at week 48 and 150 and week 96. In the PI arm triglycerides appeared on slide to be about 225 at week 48 and 190 at week 96.

Adherence was good overall (mean >97%) in both arms based on patient and study nurse assessment. The number of follow-up visits with adherence <95% for any study drug during previous 4-8 weeks was-

Julio Montaner from British Columbia in Vancouver reported on study CNA30017, which compares rates of treatment failure at 48 weeks in patients who continued on PI therapy or were switched to an abacavir regimen. This is a randomized open-label study, but is different than the Swiss SMT Study described above because patients in CNA30017 did not have prior NRTI experience. In this study patients either continued on their 2 NRTIs plus PI regimen (n=106) or continued on their 2 NRTIs and switched the PI to abacavir (n=105). In the previous SMT Swiss Study virologic failure was measured only by two consecutive viral loads >400 copies/ml but in this study failure was measured by both >400 and >50 copies/ml. This study also looked at adherence and treatment "satisfaction". Patients had to have a minimum of 6 months on their current PI regimen and have <50 copies/ml at study screening.

At baseline, the two arms had comparable characteristics: median age (38), about 80% men, median CD4s 500, median non-fasting triglycerides about 1.80 mmol/L, and cholesterol 5.2 mmol/L. Median time taking antiretroviral therapy was 20 months for both arms. About 10 patients in each arm escaped screening and had prior mono/dual NRTI experience.

TREATMENT FAILURE (ITT exposed)

Treatment failure was defined as two consecutive viral loads >400 copies/ml, or premature discontinuation of study treatment for any reason. There were 4 virologic failures in the abacavir arm and 2 in the continued PI arm. Using the <50 copies/ml test, 11 in each arm had viral rebound. This may be due to patients not having NRTI experience prior to the switch. Drug discontinuations due to AE were 5 in the abacavir arm and 15 in the PI arm. Drug discontinuation due to "other": 1 in abacavir and 3 in PI arms, respectively. Three study discontinuations in abacavir arm and 5 in PI arm. And 1 lost to follow-up in PI arm. Overall, 13 in the abacavir arm and 26 in the PI arm discontinued including viral rebound to >400 copies/ml (ITT exposed). The difference in time to treatment failure between the two arms was statistically significant (ITT exposed) at week 48 favoring the abacavir arm (13 vs 26 failures, p<0.028).

VIROLOGIC REBOUND (>400)

Interestingly, in the 4 abacavir patients with viral rebound 3/4 did not have the 3TC 184 mutation at baseline (HIV DNA) but had the 184 mutation after viral rebound (HIV RNA). One patient had prior AZT experienced but escaped screening and had AZT mutations at baseline in HIV DNA (41, 215). After 4 weeks in study viral load was 475 and mutations detected were 41, 215 and 184. The study regimen this person was receiving was d4T/ddI+abacavir. Patient 2 had 12,800 copies/ml HIV-RNA after rebound. At baseline no mutations were detected but after rebound only the 184 mutation was seen. Patient 3 had viral rebound at week 18, no mutations were detected at baseline, and only the 184 was seen after rebound. In patient 4, viral rebound was reported at week 36 with 43,000 copies/ml. No mutations were detected at baseline nor after rebound was detected. No change in therapy was made (study regimen was d4T/ddI/ABC) and subsequently patient achieved <50 copies/ml viral load. This and resistance profile suggest patient may not have been adherent.

In the 2 PI virologic failures (>400 copies/ml), patient 1 had reported rebound at week 8 with a viral load of 4500 copies/ml. There were 4 PI related mutations at baseline in DNA (35, 63, 71, 93), none of which are primary PI mutations and the PI resistance profile was the same after rebound (RNA). This patient was taking Combivir+Nelfinavir. No change to therapy was made and subsequent viral load measure was <50 copies/ml. Patient 2 had 3 PI mutations in DNA at baseline (10, 35, 36) and after rebound to 30,000 copies/ml mutations at 35 and 36 were seen. Study regimen for this patient was d4T/ddI/IDV and switched to Combivir+ABC. Subsequent viral load was <50 copies/ml. Further follow-up is required.

ADVERSE EVENTS

In the abacavir arm, 44% (n=48) reported any drug-related AE, and 33% (n=35) did so in the PI arm. Reporting serious adverse events were 10% (n=11) and 12% (n=13) in the ABC & PI arms, respectively. Suspected hypersensitivity was reported for 2 patients (2%) in the ABC arm and none in the PI arm.

CD4s increased 26 in the PI arm and 13 in the Abacavir arm at week 48.

Patients were asked to respond to: none of the drugs in the randomized treatment regimen were difficult to take. Montaner said this doesn't represent adherence so much as patient satisfaction. At baseline about 50% in both arms agreed that none of the drugs in their current regimen were difficult to take. After week 48 92% in the abacavir arm said drugs were easy to take versus 69% in the PI arm.

At week 48 (ITT exposed), median cholesterol decreased from baseline -0.51 mmol/L in the abacavir arm versus -0.11 mmol/L in the PI arm (p<0.001). Improvement occurred quickly within a few weeks. Montaner did not specify if the cholesterol changes were non-sating. But the non-fasting triglyceride median changes from baseline were small but statistically significant (-0.14 mmol/L in ABC arm versus +0.04 mmol/L in the PI arm at week 48).

LIPODYSTROPHY

At the Lipodystrophy Workshop in Toronto just prior to ICAAC, W Rozenbaum reported 48 week results from the French Lipodystrophy Sub study of CNA30017. At week 48 median cholesterol and triglycerides showed improvements in cholesterol & triglycerides while patients remaining on PI did not show improvements. But these improvements were not statistically significant. He reported that 1/2 of patients who switched to abacavir normalized their triglycerides. He reported improvements in insulin and glucose parameters compared to persons staying on PI regimen: insulin tolerance test, glucose tolerance test (OGTT). But, median change in fasting glycemia between baseline and week 48 was +0.16 mmol/L in the abacavir group vs +0.06 mmol/L in the PI group (not statisticaly significant). Rozenbaum concluded that there was a tendency towards a stabilization of insulin metabolism in the abacavir group. There was a suggestion from the data that some body changes may improve after a switch to abacavir from PI, but again as Powderly said the data is very unclear.

From PI to Efavirenz or Nevirapine: prior NRTI experience matters

An important point was made in the Maintavir Study presented by  F Raffi from University Hospital in Nantes, France. This was a prospective observational study at 3 sites for NNRTI naÔve individuals who were switched to either nevirapine or efavirenz. The same NRTIs were continued. The key point here is that some patients had prior NRTI experience. You will see that individuals with prior NRTI experience did not maintain viral load suppression as well as persons without NRTI experience. This study enrolled patients on a PI regimen for at least 1 year and with viral load <400 copies/ml for >6 months on a PI regimen. Virologic failure was defined as 2 consecutive viral loads >200 copies/ml. 73 patients switched therapy and 23 had AIDS. The reasons for switching were: to simply regimen (n=39), lipodystrophy (n=18), GI intolerance-diarrhea (n=17), adherence problems (n=4), ALT elevation (n=1), renal colic (n=4). 31 patients were treatment naÔve and 42 had prior treatment. At the time of PI initiation, median CD4 was 200 and viral load was 4.7 log (50,000 copies/ml). At the time of switch, median CD4 was 470, median time on PI regimen was 22 months (4-37 months range) and median duration <400 copies/ml was 19 months. 71/73 continued on the same NRTIs. 10 switched to EFV and 63 switched to NVP: 42 taking NVP once daily (400 mg once daily after 200 mg once daily for 2 weeks induction) and 21 twice daily. At the time of switch, 69/73 had <50 copies/ml, but 4 had viral load between 50-150.  

Median follow-up is at 82 weeks (26-112 range), and 51/73 are at 18 months or greater.

There were 5/73 discontinuations. 86.3% (63/73) had undetectable viral load: 53 had <80 copies/ml and 10 <200 copies/ml. Median viral load at rebound was 915 copies/ml (range 525-2500). In 5/10, intercurrent infection occurred a few weeks before failure. All the 10 failures were switched back to the PI regimen and within 2-6 weeks had undetectable viral load. Raffi reported that there was a trend towards a higher viral rebound rate for individuals with NRTI experience prior to initiating their PI HAART therapy compared to persons without any prior NRTI experience (19.2% vs 6.5%). He is examining these patients blood samples to see if they had pre-existing resistance. There were little or no changes in body fat distribution. Raffi reported 7/18 had subjective improvements in lipodystrophy and 3/55 patients developed lipodystrophy after the switch to nthe NNRTI. He reported no significant change in cholesterol but a significant reduction in triglycerides.

From PI to Nevirapine

The SENC Study (Spanish Efavirenz versus Nevirapine Comparison Trial) is a small randomized, open-label controlled trial in treatment naÔve. This study was conducted at one site: the Hospital Carlos III, an HIV center in Madrid. Patients with CD4s >100 and viral load 500-100,000 copies/ml were eligible for study. Patients were randomized to ddI 400 mg once daily + d4T 40 mg twice daily and either nevirapine 400 mg once daily or efavirenz 600 mg once daily. So, every drug was once daily except d4T. A short course of steroids was given with nevirapine for the purpose of preventing rash but recent study results have suggested this does not help in preventing rash. HIV-RNA was measured using bDNA and lipids & LFTs were followed. Primary endpoints of study were % of patients <50 copies/ml and % of patients with drug-related toxicities leading to stopping NNRTIs.

VIROLOGIC RESPONSE

ITT analysis is non-completer=failure and the OT (on treatment) analysis included patients only on original medications. The presenter said patients with viral load >50,000 copies/ml at baseline had similar responses but she did not show the data. This study drew sharp criticism because the presenter concluded that the study showed that EFV & NVP were equivalent. The criticism was that a much larger study is needed to show equivalence than to show a difference between regimens. The smaller a study is the less likely it is to show a difference between 2 arms. As well, the baseline viral load was low, maybe too low to detect differences between the 2 arms.

Subjects with HIV-RNA <50 copies/ml :

The increase in CD4s were equivalent as evaluated at week 12 (+140) in 6 evaluable patients in each arm. Presenter reported that there were 3 virologic failures (>50 copies/ml) in the NVP arm at maximum follow-up (12, 9, and 6 months, respectively). Viral loads were 470, 2400 and 287 copies/ml respectively. She said compliance was poor in 1/3. One of the 3 never reached undetectable.

DRUG RELATED TOXICITIES

None of the differences in side effects were statistically significant probably due to the small size of study, except with regards to CNS symptoms- 0 in NVP arm, 46% (n=11) in EFV arm (2 discontinuations). liver enzyme elevations- 60% NVP arm (15/25), 55% (12/22) in the EFV arm. Peripheral neuropathy- 16% (n=4) in NVP arm, 8% (n=2) EFV arm. Elevated cholesterol- 52% (n=13) NVP, 54% (n=13) EFV. Triglycerides- 28% (n=7) NVP, 37% (n=9) EFV. GI symptoms- 8% (2) NVP, 12% (3) EFV.

Abnormal LFTs at baseline in NVP arm were 8/15 and 3/12 in the EFV arm. Most LFT elevations were grade 1-2. Grade 3-4 2/15 NVP, 2/12 EFV. Presenter reported that positive serology for HCV was the only significant risk factor for NNRTI hepatotoxicity. Having HCV at baseline increased likelihood of having elevated LFTs 4.6 times. None of the patients discontinued due to LFT abnormality. There was spontaneous resolution in 2/13 and 6/9 in the NVP & EFV arms, respectively. There was no change in 5/13 in the NVP arm and 1/9 in the EFV arm. At this point the presenter said 60% in NVP arm (15/25) had direct drug induced hepatotoxicity and 45% (10/22) in the EFV arm.

The percent of patients with elevated cholesterol increased from 18% at baseline to 33% (n=39). The percent with elevated triglycerides was 21% at baseline and increased to 55% (n=38), respectively. There were no differences between treatment groups. Only the difference in cholesterol was significant (p=0.001).

A second small Spanish study was presented by B Clotet, also conducted at a single study site in Barcelona. But this small study compared switching from a PI regimen to either a NVP (n=26) or EFV (n=25) regimen, to remaining on the PI regimen (n=25). And patients had prior treatment experience. The main study goal was to assess viral and immune response, and secondary objectives were to examine metabolic abnormalities, body fat changes, adherence, quality of life and adverse events. Photographs were taken at baseline & 6 months. In this randomized, open-label study, follow-up was 9 months. Nucleosides were not changed. DEXA scan was done at baseline & 6 months. In this small study, viral suppression was maintained in all 3 arms. Total cholesterol and triglycerides improved in NVP arms but not in other arms. There was no improvement in body changes by DEXA scans & anthropometric measures. There were 5 cases of acute hepatitis (1 discontinuation) reported in NVP arm out of 26 patients, and none in the other arms. As a result, Clotet recommended monthly liver function monitoring of patients who begin NVP is required, mainly in HCV coinfected patients. There were 9 patients reporting CNS symptoms in the EFV arm out of 25 patients (3 discontnued). Quality of life was reported improved in the 2 NNRTI arms but not in the PI arm.

BASELINE CHARACTERISTICS

Patients were on PI triple regimen for at least 12 months, and <80 copies/ml for about 30 months equally in all 3 arms. Prior experience with antiretroviral therapy was 64-67 months in 3 arms. Prior time on HAART was 23 months in all 3 arms.  CD4s were 641 in NVP, 595 in EFV and 660 in PI arms, respectively. CD8s were 1361 in NVP, 1112 in EFV , and 1207 in PI arm. Patients had to be NNRTI naÔve and not received lipid lowering agents during the 6 months prior to study. At baseline, mean age was 37-40 years, 6-8 females were in each arm (about 1/3), and 9-12 IVDUs were in each arm.

TREATMENT FAILURE

Treatment failure was defined as 2 consecutive measures >80 copies/ml. Viral rebound reported was 1 (4%) in the NVP arm, 2 (8%) in the EFV arm, and 2 (8%) in the PI arm. Discontinuation due to adverse events was 2/26 (8%) in NVP arm, 3/25 (12%) in the EFV arm, and 2/25 (8%) in the PI arm. Lost to follow-up: 1 in NVP, 0 EFV, 1 PI. The total failure rate for any reason was 4 (15%) in NVP, 5 (20%) in EFV, and 5 (19%) in PI arm. There were no significant differences between groups.

ADVERSE EVENTS

Rash: 2 in NVP arm (1 discontinued); 1 in EFV arm, and 0 in PI arm.

Acute Hepatitis: 5 in NVP arm (1 discontinued), 5-fold above baseline values; 0 in EFV arm, and 0 in PI arm. These findings were different than in the SENC study where they saw no difference between EFV & NVP arms regarding LFT elevations.

CNS Symptoms: 0 in NVP arm; 9 in EFV arm (3 discontinued), 0 in PI arm.

Diarrhea (>3x/day): 0 in both EFV & NVP arms; 3 in PI arm had nelfinavir associated diarrhea (1 discontinued).

Nephrolithiasis: 1 in PI arm (IDV related) and none in other arms.

CD4s increased significantly in all 3 arms but there was no significant difference between the arms. The same occurred with CD8s--there was significant decline within each arm but no significant difference between the arms. GGT increased 3 fold in the NVP arm from baseline to 9 months (from 44 to 116), and 2-fold in the EFV arm (from 41-72); change in PI arm was from 57 to 72. The GGT increases were statistically significant in NVP & EFV arms (p<0.02). GGT elevation is thought to be related to induction of hepatic enzymes. GPT/ALT increased in NVP arm from 49 at baseline to 55 at 9 months and was statistically significant (p<0.028). GPT/ALT increased I EFV arm from 41 at baseline to 51 at 9 months but was not statistically significant. Increase in GPT/ALT in PI arm also was not significant but was 39 to 61. Clotet commented that hepatic abnormalities were seen in HCV coinfected patients.

LIPIDS

Total cholesterol significantly decreased in NVP arm, but not in the other 2 arms. The improvement in total cholesterol was due to a significant decrease in LDL cholesterol (bad cholesterol) while HDL cholesterol remained the same. Clotet did not report actual changes in HDL cholesterol. In other studies it's usually reported that HDL cholesterol improves in EFV, although total cholesterol may not change. Triglycerides significantly improved in the NVP arm, but not in the EFV & PI arms.

BODY CHANGES

Classified as having lipodystrophy at baseline: 20 in NVP arm, 19 in EFV arm and 19 in PI arm. About 75% across all arms. Receiving DEXA scans were 20 in NVP, 17 in EFV and 19 in PI arms. Clotet reported that no patient suffering lipodystrophy at baseline experienced any significant improvement in fat redistribution in any of the 3 arms. DEXA scans were performed at baseline & 6 months. DEXA and anthropometric measures confirmed the lack of improvement. Actual data was not shown.

Patients receiving EFV 7 NVP reported statistically significant improvement in quality of life throughout the study, while patients on PI did not report any change in quality of life. Reports of adherence were high and about equal in al 3 arms.