| |
Early Virologic Nonresponse to Tenofovir, Abacavir, and Lamivudine in HIV-Infected Antiretroviral-Naive Subjects
|
| |
| |
The Journal of Infectious Diseases Oct 25, 2005;192:000 advance publication
..... The early failure of this regimen emphasizes the importance of treating patients on the basis of results of randomized, controlled trials rather than on assumptions about efficacy ....Pharmacokinetic interactions do not explain these results ...Inadequate potency of once-daily abacavir and lamivudine does not explain the poor response ...
.... The most likely explanation is the low genetic barrier to resistance produced by synergistic selection pressure from all 3 drugs for 2 point mutations, M184V and K65R ....
..... .... At 24 weeks after the switch, 101 (77%) of 131 subjects and 108 (82%) of 131 subjects achieved HIV-1 RNA levels of <50 and <400 copies/mL, respectively (ITT, M=F) .....
.... use of zidovudine in the second-line regimen was correlated with a greater decrease in HIV-1 RNA levels .....correlates of changes in HIV-1 RNA level were HIV-1 RNA level at the switch and the use of zidovudine in the second-line regimen .....The presence of the K65R mutation was not significantly correlated with decreases in HIV-1 RNA level ....HIV RNA levels were associated with response, either baseline or after switch ...
Joel E. Gallant,1 Allan E. Rodriguez,2 Winkler G. Weinberg,3 Benjamin Young,4 Daniel S. Berger,5 Michael L. Lim,6 Qiming Liao,6 Lisa Ross,6 Judy Johnson,6 and Mark S. Shaefer,6 for the ESS30009 Studya
1Division of Infectious Diseases, Johns Hopkins University School of Medicine, Baltimore, Maryland; 2University of Miami, Miami, Florida; 3Kaiser Permanente, Atlanta, Georgia; 4Rose Medical Center, Denver, Colorado; 5Northstar Medical Center, Chicago, Illinois; 6GlaxoSmithKline, Research Triangle Park, North Carolina
(See the editorial commentary by Kuritzkes, on pages XXXXX.)
ABSTRACT
Background. Antiretroviral combinations that reduce the number of pills and dosing frequency have the potential to simplify therapy. We compared 2 regimens dosed as 2 pills once daily.
Methods. This was a randomized, open-label, multicenter study of tenofovir disoproxil fumarate versus efavirenz, both administered once daily with the abacavir/lamivudine fixed-dose combination in treatment-naive human immunodeficiency virus type 1 (HIV-1)infected subjects. After reports of early nonresponse, an unplanned interim analysis was performed. Virologic nonresponse was defined as (1) a <2.0-log10 copies/mL decrease in HIV-1 RNA level by week 8, (2) an HIV-1 RNA rebound of >1.0 log10 copies/mL above the nadir, or (3) for subjects with 2 consecutive HIV-1 RNA measurements <50 copies/mL, a subsequent increase to >400 copies/mL on 2 consecutive occasions.
Results.
We randomized 340 subjects. Median baseline HIV-1 RNA level and CD4+ cell count were 4.7 log10 copies/mL and 251 cells/mm3, respectively; 194 subjects with HIV-1 RNA data from 8 weeks were included in the interim analysis.
Virologic nonresponse occurred in 50 (49%) of 102 subjects in the tenofovir disoproxil fumarate arm, compared with 5 (5%) of 92 of subjects in the efavirenz arm (P < .001).
Within 12 weeks, viral genotypes for nonresponders in the tenofovir disoproxil fumarate arm showed M184V or I/M/V mixtures in 40 (98%) of 41 subjects and K65R and M184V or mixtures in 22 (54%) of 41 subjects.
The protocol was immediately amended to modify the tenofovir disoproxil fumarate arm. The efavirenz arm continued unchanged; after 48 weeks, 120 (71%) of 169 subjects achieved HIV-1 RNA levels <50 copies/mL.
Conclusion. The tenofovir disoproxil fumarate/abacavir/lamivudine regimen resulted in an unexpected and unacceptably high rate of nonresponse and incidence of K65R and M184V/I. This 3-drug regimen should not be used.
Resistance. Baseline resistance was evaluated in 317 subjects. Genotypic mutations associated with NRTI, NNRTI, and primary PI resistance were detected in 3%, 6%, and 2% of subjects, respectively. Decreased phenotypic susceptibility to >1 drug within a class was noted in <1%, 18%, and 6% of subjects for NRTIs, NNRTIs, and PIs, respectively.
Genotypic and phenotypic resistance in the tenofovir DF arm was evaluated at baseline in 100 (98%) of 102 subjects included in the interim analysis and at week 12 in 41 (82%) of 50 subjects with nonresponse at the interim analysis. Week 12 samples from 9 subjects were not analyzed because of withdrawal before week 12, HIV-1 RNA levels that were too low for analysis, or failure to amplify virus.
Baseline genotypic resistance in the tenofovir DF arm was infrequent. At week 12, NRTI mutations were observed in all but 1 subject; 40 (98%) of 41 subjects had the M184V/I mutation, and 22 (54%) of 41 had both M184V/I and K65K/R. Baseline phenotypic resistance in the tenofovir DF arm was also uncommon; susceptibility to all NRTIs was observed in all 100 subjects. At week 12, phenotypic susceptibility to abacavir, lamivudine, and tenofovir was retained in 80%, 2%, and 95% of nonresponders, respectively.
Observational phase after switching.
After the interim analysis, 131 (77%) of 171 subjects in the tenofovir DF arm remained in follow-up and received a second-line regimen. These subjects were demographically similar to the overall population: 92% were male, the median age was 37.0 years (range, 17.069.0 years), 27% were black, 13% were Hispanic, and 60% were white. The median duration of tenofovir DF therapy before switching was 106.0 days (range, 29.0187.0 days). The median HIV-1 RNA level and CD4+ cell count at the time of switching were 3.26 log10 copies/mL (range, 1.696.28 log10 copies/mL) and 324 cells/mm3 (range, 191052 cells/mm3). At the time of switching, 24% and 41% were virologically suppressed to <50 and <400 HIV-1 RNA copies/mL, respectively. Of the 61 subjects with genotype data available at or before switching, 31 had a detectable K65R mutation.
Thirty unique regimens were used after the switch. Those used in at least 5% of subjects were efavirenz, abacavir, and lamivudine (n = 46); tenofovir DF, zidovudine, lamivudine, and abacavir (n = 18); efavirenz, zidovudine, and lamivudine (n = 13); efavirenz, zidovudine, lamivudine, and abacavir (n = 12); and efavirenz, zidovudine, lamivudine, and tenofovir DF (n = 7). The remaining regimens were classified into 4 exclusive categories: combinations that contained NNRTIs and NRTIs (n = 16); PIs and NRTIs (n = 12); NNRTIs, PIs, and NRTIs (n = 5); and NRTIs only (n = 2).
At 24 weeks after the switch, 101 (77%) of 131 subjects and 108 (82%) of 131 subjects achieved HIV-1 RNA levels of <50 and <400 copies/mL, respectively (ITT, M=F). Exploratory multivariate analyses examined independent predictors of response in the postswitch cohort and in a subset with genotype data. In the postswitch cohort with observed data (n = 113), the only statistically significant predictor of HIV-1 RNA levels <50 copies/mL at 24 weeks after the switch was baseline HIV-1 RNA level (odds ratio, 0.073 [95% CI, 0.0170.303]; P < .001). There were no significant predictors of HIV-1 RNA levels <50 copies/mL at 24 weeks after the switch in the genotyped subset (n = 53), and the K65R mutation was not associated with suppression to <50 copies/mL. In linear regression analyses that explored predictors of changes in HIV-1 RNA levels 24 weeks after switch, significant correlates were HIV-1 RNA level at the switch (parameter estimate [PE], -0.941; SE, 0.047; P < .001) and the use of zidovudine in the second-line regimen (PE, -0.192; SE, 0.094; P = .044), which was correlated with a greater decrease in HIV-1 RNA levels. In the genotyped subset, significant correlates of changes in HIV-1 RNA level were HIV-1 RNA level at the switch (PE, -0.910; SE, 0.093; P < .001) and the use of zidovudine in the second-line regimen (PE, -0.411; SE, 0.191; P = .036). The presence of the K65R mutation was not significantly correlated with decreases in HIV-1 RNA level.
DISCUSSION
In the present trial, treatment with once-daily tenofovir DF and abacavir/lamivudine FDC tablet was associated with an unprecedented rate of early virologic nonresponse accompanied by significant NRTI resistance, with detection of the K65R mutation in the majority of subjects and the M184V/I mutation in all but 1 subject. These findings are consistent with results of 2 small, uncontrolled trials that were concurrently studying the same regimen and reported early virologic failure in 33%58% of patients and similarly high rates of resistance [19, 20].
Several explanations for the poor response have been proposed, including clinical, pharmacokinetic, and virologic hypotheses. Although the dramatic findings of the present trial make it impossible to study this regimen further, some of these hypotheses can be refuted by existing data. Inadequate potency of once-daily abacavir and lamivudine does not explain the poor response this combination has demonstrated efficacy and durability when combined with efavirenz, both in the present study and in a large, randomized, double-blind trial [8]. Moreover, pharmacokinetic data support once-daily dosing of abacavir: the mean intracellular half-life of carbovir-triphosphate, abacavir's active metabolite, is >20 h [14].
Pharmacokinetic interactions do not explain these results, either. Tenofovir DF did not adversely affect plasma abacavir concentrations in a study of the single-dose regimen in healthy volunteers [21]. In a clinical trial of tenofovir DF, abacavir, and lamivudine, 32 (86%) of 37 subjects had adequate plasma trough concentrations of all 3 drugs, despite the poor response [20]. Interactions that affect intracellular concentrations of the active phosphorylated metabolites of these drugs also seem to be unlikely. In vitro studies in human peripheral blood mononuclear cells and T lukemic CEM lymphoblast cells found no changes in the phosphorylation of either drug, whether administered alone or in combination [22]. In a study of 15 HIV-infected subjects who received a regimen that included both abacavir and tenofovir DF, there were no changes in carbovir-triphosphate concentrations after the discontinuation of tenofovir DF and no changes in tenofovir-diphosphate concentrations after the discontinuation of abacavir [23].
The most likely explanation is the low genetic barrier to resistance produced by synergistic selection pressure from all 3 drugs for 2 point mutations, M184V and K65R. Both abacavir and tenofovir DF select for the K65R mutation, which reduces susceptibility to both drugs, as well as to lamivudine. M184V is selected for by lamivudine and abacavir, and it decreases susceptibility to both. Thus, the selection of 2 mutations, each of which may preexist as minority species, leads to virologic failure with this regimen. Although K65R was not universally detectable within 12 weeks in all subjects, clonal analysis of samples from selected subjects suggests the early presence of viruses containing mutations at either K65R or M184, followed by selection for viruses containing mutations at both sites [24]. If treatment with this regimen were continued, both mutations would presumably be detected. Further support for this hypothesis comes from a small, uncontrolled pilot trial of the combination of tenofovir DF, didanosine, and lamivudine in which 91% of subjects failed therapy [25]. Resistance issues were similar with this combination, given that didanosine also selects for K65R. Of those subjects for whom genotypic data were available, 100% had the M184V mutation and 50% had the K65R mutation.
While the present study was in progress, the interim results of AIDS Clinical Trial Group (ACTG) 5095 were reported, and they demonstrated that the triple-nucleoside regimen of zidovudine/lamivudine/abacavir was virologically inferior to a regimen that contained efavirenz and 2 or 3 nucleosides [10]. However, our finding of early nonresponse to tenofovir DF and abacavir/lamivudine appears to be a distinct phenomenon, considering the rapid development and high incidence of both virologic nonresponse and resistance. In ACTG 5095, 74% of subjects who received zidovudine/lamivudine/abacavir achieved suppression to <200 HIV RNA copies/mL at 48 weeks, whereas, in the present study, approximately one-half of subjects who received tenofovir and abacavir/lamivudine demonstrated nonresponse as early as 8 weeks into therapy. Futhermore, a combined analysis applied a modified nonresponse criteria to 6 trials that had zidovudine/lamivudine/abacavir arms and found nonresponse rates of 0%, 7%, 12%, 12%, 14%, and 24%, which were strikingly lower than that described with the tenofovir DF regimen studied in ESS30009 [26].
Despite the alarming results observed in the tenofovir DF arm, many subjects were able to achieve or maintain viral suppression in subsequent therapy. These observational data must be viewed with caution, because selection bias, the small sample size, and the short duration of follow-up limit their applicability. These subjects switched therapy early, and approximately one-fourth had HIV-1 RNA levels of <50 copies/mL at the time of switching. Conversely, several subjects who developed resistance, including the K65R mutation, after an early nonresponse did not remain in follow-up, and their outcome is unknown. Subjects in this cohort were also naive to both NNRTIs and PIs, so the availability of treatment options enhanced their likelihood of response. Treatment options for patients in whom the M184V and K65R mutations develop after the failure of regimens currently in use remains an important question that our study could not address.
It is clear from the present study that the 3-drug combination of abacavir, lamivudine, and tenofovir DF should not be used in clinical practice. The early failure of this regimen emphasizes the importance of treating patients on the basis of results of randomized, controlled trials rather than on assumptions about efficacy. This simple, convenient, and well-tolerated regimen was presumed to be effective on the basis of the potency of its component drugs and was already being used by some clinicians in clinical practice. Given the growing number of potent, well-studied combinations now available, there is no longer a rationale for the use of untested regimens outside of clinical trials in the treatment of therapy-naive HIV-infected patients.
RESULTS
Study subjects. A total of 460 subjects were screened, and 345 were randomized. The ITT population consisted of 340 randomized subjects who received >1 dose of study drug. The median baseline HIV-1 RNA level was 4.72 log10 copies/mL (range, 2.266.43 log10 copies/mL), and 30% had an HIV-1 RNA level >100,000 copies/mL. The median baseline CD4+ cell count was 251 cells/mm3 (range, 191171 cells/mm3). Baseline demographic characteristics were balanced between arms. The unplanned interim analysis was performed on a subset of 194 subjects with >8 weeks of virologic data available at the time of analysis.
Study subject disposition. After 48 weeks, 42 (25%) of 169 subjects in the efavirenz arm had prematurely withdrawn. Those in the tenofovir DF arm were allowed to continue in follow-up on a second-line regimen after the unplanned interim analysis. After 48 weeks, 58 (34%) of 171 subjects in this arm had prematurely withdrawn.
HIV-1 RNA level and CD4+ cell count. In the unplanned interim analysis of 194 subjects, HIV-1 RNA levels declined rapidly in the majority of subjects in the efavirenz arm, whereas many in the tenofovir DF arm had rebound or insufficient response within 812 weeks. Virologic nonresponse occurred in 49% of subjects in the tenofovir DF arm versus 5% in the efavirenz arm (P < .001). On the basis of these findings, investigators and subjects were immediately notified, and a protocol amendment was rapidly implemented. Subjects receiving tenofovir DF were allowed to switch to a second-line regimen at the discretion of the investigator.
Virologic nonresponse was further examined in post hoc subgroup and sensitivity analyses that used the unplanned analysis criteria. In subjects with baseline HIV-1 RNA levels <100,000 copies/mL, 35 (43.8%) of 80 subjects in the tenofovir DF arm met virologic nonresponse criteria, compared with 5 (6.8%) of 73 subjects in the efavirenz arm (P < .001). To explore the effect of the duration of follow-up, nonresponse criteria were applied to 125 subjects with at least 12 weeks of data. With this additional follow-up, nonresponse occurred in 30 (48%) of 63 subjects in the tenofovir DF arm versus 3 (5%) of 62 in the efavirenz arm (P < .001). Finally, in a sensitivity analysis that applied a less-stringent first nonresponse criterion of a 1.5-log10 copies/mL decrease in HIV-1 RNA levels by week 8, virologic nonresponse occurred in 41 (40%) of 102 subjects in the tenofovir DF arm versus 4 (4%) of 92 in the efavirenz arm (P < .001).
Longer term data were available for the efavirenz arm. In a planned ITT, M=F analysis, 120 (71%) of 169 subjects achieved HIV-1 RNA levels <50 copies/mL, and 127 (75%) of 169 achieved HIV-1 RNA levels <400 copies/mL after 48 weeks. The majority of subjects in the tenofovir DF arm switched regimens or withdrew before week 16, and data from after week 12 were censored. Subjects in the efavirenz arm had median increases in CD4+ cell counts of 111, 126, and 130 cells/mm3 after 8, 16, and 48 weeks, respectively, compared with increases of 62 and 101 cells/mm3 after 8 and 16 weeks in the tenofovir DF arm.
Adverse events and laboratory data. Grade 2 or higher adverse events were reported in 114 (67%) of 169 subjects in the efavirenz arm and in 102 (60%) of 171 subjects in the tenofovir DF arm. These are summarized, along with grade 3 or higher laboratory abnormalities.
|
|
| |
| |
|
|
|
