 |
|
|
| |
Cutoffs Suggested for Predicting Efavirenz Failure With Low-Level K103N
|
| |
| |
Mark Mascolini
XVIII International Drug Resistance Workshop, June 9-13, 2009, Fort Myers, Florida
Studies of virologic outcome in people who begin a nonnucleoside with low-frequency K103N mutations yielded conflicting results [1-3]. A new analysis of low-level K103N in people starting efavirenz in a trial comparing tenofovir/emtricitabine (TDF/FTC) with zidovudine/lamivudine (AZT/3TC) suggested that more than 2000 copies/mL of K103N before efavirenz--or K103N making up more than 2% of the viral population--independently boosts the odds of virologic failure [4].
Evguenia Svarovskaia and Gilead Sciences colleagues analyzed the impact of K103N on virologic response in 509 people enrolled in GS-01-934. The original primary efficacy analysis excluded 22 people (4%) in whom standard genotyping detected a nonnucleoside mutation before treatment began. Of the remaining 487 study participants, 48 (10%) with resistance data available endured virologic failure through week 144, including 19 taking TDF/FTC and 29 taking AZT/3TC.
The Gilead team reanalyzed these 487 pretreatment viral samples with an allele-specific PCR assay that detects K103N representing as little as 0.5% of a person's viral population. Almost all of those samples--476 of 485 (98%)--yielded allele-specific PCR results. Sixteen of those 476 (3%) had detectable K103N missed by standard sequencing. Thus the percentage of all study participants with a nonnucleoside mutation before treatment rose to 7.5% (38 of 509 samples).
Among the 16 people with low-level K103N, efavirenz failed in 6 (37.5%), including 5 taking AZT/3TC and 1 taking TDF/FTC. Pretreatment K103N correlated strongly with virologic failure in the 5 people taking AZT/3TC with efavirenz (P = 0.005).
In these 16 people with low-frequency K103N, the percentage of K103N detected by the ultrasensitive assay and K103N copy count were higher in the 6 with virologic failure (0.8% to 15%, 1254 to 16,071 copies/mL) than in the 10 virologic successes (0.6% to 3.2%, 51 to 5535 copies/mL). Five of 6 people with more than 2000 copies/mL of K103N had a virologic failure, compared with 1 of 10 people without virologic failure, a significant difference (P = 0.008).
Efavirenz plus TDF/FTC failed in 1 person with a K103N load under the 2000 mark (1254 copies/mL), but this person began treatment with a CD4 count of only 20.
Multivariate statistical analysis that considered predictors as categorical variables determined that a K103N copy number at or above 2000 copies/mL hoisted the risk of virologic failure nearly 50 times (odds ratio [OR] 47.4, 95% confidence interval [CI] 5.2 to 429.2, P = 0.0006). In the same analysis, randomization to TDF/FTV versus AZT/3TC, pretreatment viral load above 100,000 copies, and pretreatment CD4 count above 199 did not independently predict failure. In a similar analysis, K103N making up more than 2% of a person's viral population upped the failure risk 25 times (OR 25.5, 95% CI 4.6 to 142.1, P = 0.0002).
Svarovskaia and coworkers suggested their findings point to a pretreatment K103N threshold that predicts virologic failure with efavirenz. However, that conclusion rests on a sample of only 16 people. The investigators point out that only 6 of 476 study participants with allele-specific PCR results (1.3%) had a pretreatment K103N population above 2000 copies/mL.
References
1. Metzner KJ, Walter H, Rauch P, et al. The prevalence of drug-resistant virus as a minority quasispecies before initiating art is not associated with therapy failure in persons initiating therapy with Truvada plus PI/r or NNRTI. 15th Conference on Retroviruses and Opportunistic Infections. February 3-6, 2008. Boston. Abstract 879.
2. Garcia-Diaz A, Johnson JA, Fox ZV, et al. Low-frequency mutations strengthen the impact of transmitted drug resistance on virological responses to first-line efavirenz or nevirapine based antiretroviral therapy. 7th European HIV Drug Resistance Workshop. March 25-27, 2009. Stockholm. Abstract 113.
3. Braun P, Ehret R, Wiesmann F, Metzner KJ, Knechten H. Prevalence of the K103N mutation at low frequencies in antiretroviral treatment-naive patients in Germany 2008. 7th European HIV Drug Resistance Workshop. March 25-27, 2009. Stockholm. Abstract 71.
4. Goodman DD, Margo NA, McColl DJ, et al. Pre-existing low-levels of the K103N HIV-1 RT mutation above a threshold is associated with virological failure in treatment-naive patients undergoing EFV-containing antiretroviral treatment. XVIII International Drug Resistance Workshop. June 9-13, 2009. Fort Myers, Florida. Abstract 41.
|
| |
|
|
|
|
|
