icon_folder.gif   Conference Reports for NATAP  
  3rd International Workshop on Adverse Drug Reactions and Lipodystrophy in HIV
Athens, Greece - October 2001
Back grey_arrow_rt.gif
Abacavir: Fat Redistribution, and Changes in Cholesterol & Triglycerides
  Two studies were reported comparing combivir/abacavir or Trizivir to PI regimens for differences in effects on lipids, glucose, body changes, and viral suppression. Trizivir is one pill twice per day and contains 3 drugs-AZT, 3TC, and abacavir. These 2 studies have limitations described in the next paragraph. But, as previous studies have shown, abacavir triple-nuke regimen show better effects on certain lipid measures than PI regimens, and there is a suggestion or trend towards less effect on causing body changes in patients who are treatment-naïve. In the second study viral suppression was maintained after switching from HAART to Trizivir.
The second study is open-label and randomized. In patients who already have body changes the second study below suggests that there may be some measure of improvement in a percentage of the patients in the study. But this study and previous studies show switching therapy from a PI to a PI sparing regimen is limited in reversing body changes, and in many cases does not reverse body changes. From the tools currently available to us, it appears the best approach in trying to avoid body changes is in the choice of the first treatment regimen. PI sparing regimens appear to be less likely to cause body changes. Body changes do occur with NNRTI+ 2 nuke regimens but not as quickly and perhaps not as severely. You can say the same for abacavir/combivir or Trizivir. Results from preliminary studies suggest rosiglitazone, an anti-diabetic drug, may improve or prevent body changes but further studies are ongoing and we must wait for the results.
The first study compared combivir/abacavir to nelfinavir/combivir and 3TC/d4T/nelfinavir (NFV) in treatment-naïve patients. The second study randomized patients who had undetectable viral load on a PI regimen (50% indinavir, 50% nelfinavir) to either remain on the PI regimen or switch to the Trizivir regimen. The results from these studies show generally better cholesterol and triglycerides responses to the abacavir regimens. And the results suggest that patients taking abacavir regimens can experience body changes but perhaps with less incidence and severity than when compared to taking a PI regimen. The results also show switching from a PI regimen to an abacavir regimen improved cholesterol and triglycerides while maintaining viral suppression. Previous studies have also shown similar findings but these studies appear to be more intensive looks at metabolics. A limitation of these reports is that we do not have available objective measures of fat redistribution. The first study is using DEXA to evaluate fat redistribution, but since the report is preliminary this data is not yet available. It appears that evaluation of body changes in the second study is by observation and not by objection measures such a DEXA. Another limitation of these studies is that they did not report dropouts due to viral failure. The second study reported dropouts for adverse events but not for viral failure. Patients in the second study had been on their PI regimen for >6 months but the average duration of PI therapy was not reported. The ability to reverse body changes may in part be dependent on how long a patient has had the body changes. Also, it is very important to remember that if a patient has prior treatment experiences and drug resistance, switching from a PI or NNRTI regimen to an abacavir triple nuke regimen entails the risk of losing viral suppression due to the underlying nuke resistance. Abacavir may not be as effective in viral suppression if there is nuke resistance present from prior treatment.
The primary endpoint in the first study is a change from baseline in fasting LDL (bad cholesterol). Key secondary endpoints the study looked at were: viral load and CD4 change, fasting total cholesterol, HDL (good cholesterol), triglycerides, development of genotype and phenotype resistance in viral failures, and fat redistribution by DEXA. Princy Kumar reported the preliminary 24 week data for a planned 96-week study. 50% of the study patients are women and there are 34 study sites in the US, Panama, Dominican Republic, Guatemala and Puerto Rico. The study is looking at gender differences but did not present any data in the preliminary 24-week look. The results look at patients with below 100,000 copies/ml and with patients 100,000 to 200,000 copies/ml. Patients were randomized equally to one of the 3 study treatment arms. The study is powered for statistical comparisons of the primary endpoint at 96-weeks.; no statistical comparisons were performed on this 24-week data. Among the total 258 patients in this study 20% were Caucasian, 37% Hispanic, and 40% African-American. Potential racial differences in lipids, glucose, and body changes between races were discussed at this meeting. There are few studies on this subject and the potential differences are not well understood, but Hispanics have a greater tendency for diabetes. Several researchers suggested here that African-Americans may not get lipoatrophy nearly as much as Caucasians. One researcher from a hospital in Harlem reported anecdotal experiences that African-Americans tend to put on weight in both the periphery and the belly. While we know the lipoatrophy is a common experience among Caucasians. It is suggested that the differences may be genetic but we don't know.
In the first study, the baseline characteristics appear to be comparable between the 3 treatment arms. LDL remained about the same at week 24 compared to the baseline value of 99.3. But LDL increased 15-20% in the NFV arms. In this preliminary 24-week data, the abacavir arm suppressed viral load <50 copies/ml (ITT m=f) as well as the 2 NFV arms. In the group with baseline viral load of 1,000-100,000, 58% (41/71) had <50 copies/ml compared to 51% (36/70) in the NFV/combivir arm and 51% (35/69) in the NFV/3TC/d4T arm. In the group with baseline viral load of 100,000 to 200,000 the number of evaluable patients was small-67% (10/15) in the abacavir/combivir arm had <50 copies/ml compared to 58% (11/19) in the NFV/combivir arm, and 6/14 (43%) in the 3TC/d4T/NFV arm. Baseline total cholesterol was 160 in the abacavir arm, 169 in the NFV/combivir arm, and 166 in the 3TC/d4T/NFV arm. HDL (good cholesterol) increased 3.6 in the abacavir arm, 2.6 in the NFV/CBV arm, and 3.8 in the d4t/3TC/NFV arm. Baseline triglycerides were 121 in the abacavir arm, 141 in the NFV/CBV arm, and 137 in the d4t/3TC/NFV arm. Fasting triglycerides increased 24 in the d4T/3TC/NFV arm compared to about 7.6 in the other 2 arms. Insulin: 2.6, 1.8, 0.5 uIU/ml, respectively in the abacavir, NFV/CBV, and d4T/3TC/ NFV groups. CD4 increased 90 in the abacavir arm and about 135 in the 2 NFV arms. Diarrhea was reported as 46-51% in the 2 NFV arms compared to 3% in the abacavir arm, but the authors did not say how they defined diarrhea. Nausea was 29% in the abacavir arm vs 31% in the NFV.CBV arm, and 17% in the d4T/3TC/NFv arm. There was more vomiting in the ABC/CBV arm, and more fatigue, anemia, and reduced white cell counts in the AZT arms. There were about the same number of total withdrawals in each arm. But they did not report withdrawals due to viral failure. Kumar reported that gender-related differences in the development of elevated lipids continue to be studied.
The second study (TRIZAL) was reported by Alan Lafeuillade and is a comparison of "metabolic disorders and clinical lipodystrophy 48 weeks after switching from HAART to Trizivir vs cintinued HAART". Patients had <400 copies/ml and were on their current therapy for 6 months or more. They had <50 copies/ml at screening for the study. They were randomized for 48 weeks to continue HAART or switch to Trizivir. The HAART regimen was their firstline therapy. At week 48, 75% in the Trizivir arm vs 69% in the HAART arm had <50 copies/ml (ITT). Adherence and the number of viral failures for each arm were not reported. Previous studies comparing indinavir to abacavir have shown that the rates of viral suppression were about the same, and it was mostly due to better adherence and tolerability. There were 106 patients in the Trizivir arm and 103 in the HAART arm. Both arms had about 25 months on prior therapy. And about 60% were on a PI regimen.
At baseline 40% in the Trizivir arm reported body changes and 50% in the continuing HAART arm:
Fat Accumualtion   6%   7%
Fat Loss   11%   16%
Fat Accumualtion + Fat Loss   8%   16%
Lipodystrophy possible related symptoms   14%   12%
Average Number of Symptoms Per Patient   2   3
  You will note that the HAART arm had 50% more cases of fat loss.
The number of subjects with at least 1 lipodystrophy symptom decreased 12% from baseline to week 48 in the Trizivir arm (from 40% [42/108] to 28% [27/97], and 8% in the HAART arm (from 50% [51/103] to 42% [42/98]. I don't think the authors reported if the difference between 12% and 8% was statistically significant.
As mentioned above it appears that lipodystrophy symptoms were reported by patients and/or care providers, objective measures such as DEXA were not used. And it appears that by switching to Trizivir body changes tend to resolve more often or not emerge as often compared to continuing on the HAART regimen. The authors reported that the number of lipodystrophy symptoms per subject was significantly lower in the Trizivir group (n=2) compared to the continued HAART group (n=4) (p<0.002).
  • The TZV group (n=97) had 4% new cases of fat accumulation vs 9% in the HAART group

  • The TZV group had 9% new fat loss cases vs 18% in the HAART group

  • Regarding symptom resolution, in the TZV arm (n=24, patients with symptoms at baseline) 38% experienced resolution of fat accumulation vs 26% in the continuing HAART arm. 42% in the TZV arm resolved fat loss vs 32% in the HAART arm. The differences in improvement were not reported to be statistically significant

  • The authors reported 12% in the TZV group (n=97) had emergence of fat accumulation and fat loss without resolution vs 20% (n=99) in the HAART arm, again this was not statistically significant. 63% in the TZV group (n=24) had resolution of fat loss and fat accumulation compared to 34% in the HAART group (n=38), but this was statistically significant (p=0.029)

Peripheral fat wasting remains the most frequent clinical manifestation at week 48 in both groups.
Cholesterol and Triglycerides Improve At Week 48
Statistically significant improvements occurred in cholesterol and triglycerides for the patients switching to Trizivir compared to those remaining on HAART. Median triglycerides was 1.6 mmol/l in both arms at baseline and decreased by about 10% (-0.17) in the Trizivir arm vs a gain of 0.01 (p=0.006) in the HAART arm. Cholesterol was on average 5.9 mmol/l in the Trizivir arm and decreased by about 17% (-0.80). Cholesterol was 5.6 mmol/l in the HAART group at baseline and decreased about 8.5% (-0.44) (p=0.001). For the patients who experienced increases in toxicity grades in cholesterol it appeared to occurr more often in the HAART group (29% vs 9%), p<0.001. Decreases in toxicity grade appeared to occur more often in the TZV group (40% vs 10%), p=0.001. Similar changes occurred regarding triglycerides: increases in toxicity grade 25% TZV vs 44% HAART (p<0.01); decreases of toxicity grade 22% TZV vs 4% HAART (p=0.05). The authors did not mention whether or not lipids were fasting and there was no report on glucose evaluation.
The authors concluded by saying further studies are needed to confirm the demonstrated benefits of switching to Trizivir.