 |
|
|
| |
Firstline Nukes Evaluated: What Makes a First Regimen Work? Answers From 79 Trials and Cohorts
|
| |
| |
15th Conference on Retroviruses and Opportunistic Infections
February 3-6, 2008
Boston
Mark Mascolini
Systematic review of nearly 80 trials and prospective cohort studies yielded more than one surprise about the ingredients of first-line antiretroviral success [1]. The plain-Jane nucleoside ddI, for example, emerged as a good-looking partner for 3TC or FTC. And white race or a higher CD4 count trimmed the chance of reaching an undetectable viral load. At the same time, this analysis by Andrew Carr and Janaki Amin (St. Vincent's Hospital and the University of New South Wales, Sydney) nosed out some potential shortcomings in trial conduct and antiretroviral guideline writing.
Carr and Amin scrutinized all randomized trials and prospective cohort studies lasting at least 24 weeks and presented since January 1996. Included studies had to compare two or more regimens in adults with intention-to-treat analyses. Exclusion criteria included dual-nucleoside regimens, nonrecommended triple-nucleoside regimens, four-drug regimens, alternating or monotherapy, or regimens for primary infection. Of 119 reviewed studies, Carr included 79 (mostly randomized trials) involving 143 treatment groups and enrolling 23,067 previously untreated people.
Most study participants came from the Americas (56.3%), followed by Europe or Australia (25.8%), Africa (8.3%), and Asia (4.2%). About three quarters of enrollees were men, 62.4% were white, and 29.5% were black. The most common transmission modes were sex between men (44.5%) and sex between men and women (41.5%). Starting CD4 counts across all studies averaged 266 cells and viral load 4.9 log (about 80,000 copies).
More than 90% of the studies lasted over 12 months, and 90% were randomized. AZT/3TC was the most frequent nucleoside duo (35.6%), followed by 3TC/abacavir (15.9%), d4T/3TC (15.7%), ddI/d4T (11.1%), and tenofovir/emtricitabine (TDF/FTC) (6.7%). The most common third drug was a nonnucleoside (43.5%), followed by an unboosted PI (24.0%), a boosted PI (21.5%), and a third nucleoside (7.5%).
After an average 14.3 months of therapy in the analyzed studies, 59.1% of enrollees had an undetectable viral load (under 200 copies or lower) by intent-to-treat analysis. Multivariate analysis picked out several factors that correlated with an undetectable load:
· Compared with longer studies, a duration 6 to 12 months made an undetectable load almost 16 times more likely (estimate 15.64, 95% confidence interval [CI] 7.22 to 24.07, P = 0.0004).
· Compared with no food restrictions, dosing after fasting made an undetectable load 4 times less likely (estimate -4.23, 95% CI -10.74 to 2.28).
· Dosing after fasting with some drugs and after food with others made undetectability 14 times less likely (estimate -14.08, 95% CI -21.47 to -6.58,P = 0.0011).
· Compared with a reference combination of 3TC/TDF, ddI/FTC made undetectability 13 times more likely (estimate 12.95, 95% CI -0.66 to 26.56), while ddI/3TC made success 6 times more likely (estimate 6.03, 95% CI -6.27 to 20.33) and TDF/FTC made success 4 times more likely (estimate 4.12, 95% CI -3.77 to 12.01).
· Nucleoside duos that made an undetectable load less likely than 3TC/TDF were d4T/3TC (estimate -6.68, 95% CI -13.39 to 0.59), 3TC/abacavir (estimate -6.40, 95% CI -13.25 to 0.44), AZT/3TC (estimate -6.40, 95% CI -13.39 to 0.59), and ddI/d4T (estimate -2.60, 95% CI -11.69 to 6.50).
· Compared with a nonnucleoside regimen, a boosted PI more than doubled the chance of undetectability (estimate 2.31, 95% CI -4.32 to 8.94), while a nonnucleoside plus a PI, an unboosted PI, or a third nucleoside made success less likely.
· Compared with other races or ethnicities, white race made success less likely (estimate -0.13, 95% CI -0.21 to -0.05, P = 0.002).
· A higher starting CD4 count made success less likely (estimate -0.03, 95% CI -0.06 to -0.01, P = 0.014).
Compared with 3TC/TDF, nucleosides that lower the risk of side-effect dropouts were ddI/3TC (estimate -4.95, 95% CI -16.54 to 6.64) and TDF/FTC (estimate -1.15, 95% CI -12.24 to 9.94), whereas ddI/d4T, AZT/3TC, d4T/3TC, and abacavir/3TC (in that order) made side effect dropouts more likely. Older age boosted the chance of dropping out because of toxicity almost 25% (estimate 1.24, 95% CI 0.69 to 1.79, P < 0.0001). Side effect results were less robust than virologic results because fewer than half of the studies analyzed presented relevant data.
Carr and Amin noted several other limitations of the data they analyzed--low enrollment of women (under 30%), prior injecting drug users (10%), and people from low-income countries (12%). Few studies reported hepatitis status, adherence, or nonopportunistic infections, and no studies used HLA-B*5701 screening to predict abacavir hypersensitivity reaction.
The identified factors explained 83% of the variability in antiretroviral efficacy and 76% of the variability in tolerability.
The findings indicate that most antiretroviral studies are too short, Carr and Amin proposed, and include too few data on adverse events. The finding that ddI plus 3TC or FTC is more potent and better tolerated than more widely recommended nucleoside backbones "suggests that guideline panel recommendations might be improved by systematic review of all available data."
Although "largely ignored," Carr and Amin added, ddI is an "effective, cheap, first-line ART option, with immediate relevance to resource-limited settings."
Reference
1. Carr A, Amin J. Reasons for treatment success with initial ART: an analysis of 22,635 participants in 64 randomized, controlled trials and 14 prospective cohorts. 15th Conference on Retroviruses and Opportunistic Infections. February 3-6, 2008. Boston. Abstract 782.
|
| |
|
|
|
|
|
