In this prospective proof-of-concept clinical trial, combination therapy with ABC/EFV/ddI maintained modest suppression of HIV-1 RNA levels and increases in CD4+ cell counts through 48 weeks of therapy in a significant proportion of subjects who previously failed to respond to their initial NRTI/PI-containing antiretroviral regimens. Thus, this study showed that changing from a PI-based regimen to an NNRTI-based regimen for virologic failure is associated with a favorable outcome in some patients. While cross-study comparisons are limited by differences in subject populations, comparisons of the virologic findings of NZTA4008 with those from studies in subjects with similar baseline HIV disease characteristics and antiretroviral experience can suggest the relative therapeutic usefulness of ABC/EFV/ddI in this subject population. Thus, ABC/EFV/ddI resulted in more subjects achieving undetectable HIV-1 RNA levels and a more pronounced CD4+ cell count increase at 48 weeks than has been reported with d4T/ABC/EFV/ddI in antiretroviral-experienced subjects who had failed on PI-containing HAART [24
The addition of HU to the triple regimen did not enhance virologic suppression with ABC/EFV/ddI over the entire 48-week study period, although in the ITT: observed analysis (but not the ITT: M = F analysis), a tendency for greater suppression in the HU group was observed at 24 weeks (HIV-1 RNA <400 copies/mL: 89% vs 67% [non-HU]; <50 copies/mL: 74% vs 57%; P
>0.05). This trend reversed by week 48, with a slightly higher proportion of subjects in the non-HU arm achieving undetectable HIV-1 RNA at that time according to both assays. In contrast, Lafeuillade et al [24
] found that when HU 500 mg twice daily was added to an ABC/EFV/ddI/d4T regimen, the significantly enhanced virologic suppression over ABC/EFV/ddI/d4T alone observed at week 24 was maintained at week 48. As HU selectively depletes a greater number of purine, rather than pyrimidine, nucleotides in vitro
], theoretically greater inhibition of HIV-1 would be expected with purine analog reverse transcriptase inhibitors, such as ddI [25
]. Where the effect of HU on HAART in antiretroviral-experienced subjects has been compared to the effect in antiretroviral-naïve subjects, a greater HU-potentiating effect on virologic suppression was seen in the -experienced group [21
]. It is noteworthy that not all studies of antiretroviral-experienced subjects have shown even short-term enhancement of virologic suppression when HU is added to HAART regimens. Indeed, no change in virologic response was observed by Gonzalez et al [27
] in their case-control study in subjects receiving HU-containing HAART (HU dose: 500 mg twice daily) (n = 59) versus non-HU-containing HAART (n = 57) over a median of 18 weeks.
Subsequent to our study, Lori et al [28
] showed in RIGHT702 that HU, administered at the low dosage of 600 mg daily (lower than that in our study and most other clinical trials), had a better efficacy and safety profile than that seen with higher dosages. RIGHT702 was a randomized, controlled clinical trial in 115 HIV-infected patients comparing the efficacy and safety of HU at three different daily doses (600, 800–900, or 1200 mg/day) given as once-daily, twice-daily, or three-times-daily regimens with ddI and d4T. A pairwise comparison demonstrated a significantly greater proportion of patients on 600 mg daily than 800–900 mg daily attaining HIV-1 RNA <400 copies/mL at week 24 (primary endpoint) (P
= 0.027) and week 48 (P
= 0.03), and HIV-1 RNA <50 copies/mL at week 24 (P
= 0.013) and week 48 (P
= 0.028). HIV-1 RNA area under the plasma concentration-time curve (AUC) at week 24 (P
= 0.016) and week 48 (P
= 0.001) was also lower in the 600 mg daily groups. The twice-daily dosing interval groups were superior to the once-daily group for all virologic endpoints; however, for the CD4+ count there was a tendency favoring the once-daily dosing. The most efficacious combination of total daily dose and dosing interval for the primary endpoint was HU 300 mg twice daily (P
= 0.017). The total daily dose groups and the dosing interval groups were quite comparable with respect to adverse events. However, one case of lethal pancreatitis occurred in the HU 1,200 mg/day group.
Our study evaluated the effect on CD4+ count of delayed HU treatment (until 8 weeks post-baseline) compared to HU treatment initiated at the start of the study. HU had a cytopenic effect on the CD4+ cell count and blunted the CD4+ response to ABC/EFV/ddI. This effect was diminished if the addition of HU was delayed from baseline until week 8. Rutschmann et al [22
] previously compared the effect of immediate versus delayed (by 12 weeks) addition of HU 500 mg twice daily to a HAART regimen (ddI plus d4T), but they did not assess comparative CD4+ cell effects. The cytopenic effect of HU on CD4+ cell counts has been well documented in other studies of HU, especially those in which ddI was given concurrently [29
]. However, a few studies have reported little or no reduction in CD4+ cell counts, or even increases [30
]. The cytopenic effect appears to be a dose-related rather than a duration-related phenomenon [26
]. A strategy of delaying HU administration until several weeks after initiation of a new HAART regimen may have potential value in the treatment of HIV-infected subjects, especially those with low CD4+ cell counts pre-treatment.
ABC/EFV/ddI was generally well tolerated, with the primary adverse events being GI in nature. The addition of HU to this regimen did not affect the incidence of rashes, depressive disorders, diarrhea, cognitive function disorders, or possible ABC-related hypersensitivity reactions, although its addition did increase the incidence of nausea, nausea/vomiting, lack of appetite, headache, dizziness, neuropathy, and malaise/fatigue. Other studies have shown that HU at a dose of ≥500 mg twice daily produces adverse GI events that may be additive to those associated with other concurrently administered drugs [26
]. In combination with ddI, the incidence of peripheral neuropathy can also be expected to rise [33
]. The addition of HU 500 mg twice daily decreased the tolerability of the regimen, with a higher proportion of subjects discontinuing due to treatment-limiting toxicities. Other studies in which subjects received HU 1000 mg/day have similarly reported a high dropout rate due to adverse events [34
]. Thus, Biron et al [35
] found that approximately one-quarter of subjects receiving HU/ddI-containing antiretroviral therapy discontinued treatment within 12 months. However, unlike in our study, early withdrawal in these other studies was due primarily to hematologic toxicity (pancytopenia, neutropenia, anemia), elevations in amylase or liver function tests, or pancreatitis. In contrast to other long-term studies that evaluated HU in combination with ddI and d4T, we did not observe a greater incidence of hematologic toxicities in the HU arm compared with the non-HU arm. This may be due in part to the relative lack of myelosuppression with ABC and EFV [36
This study had several limitations in that it comprised a small sample size (especially at week 48: 10 and 11 in the HU and non-HU arms, respectively), included participants with relatively low baseline HIV-1 RNA values, involved differing numbers of subjects in the treatment arms, and had a high withdrawal rate. HU was randomized against a combination that was highly suppressive, and this could be viewed as a formidable situation in which to verify increased efficacy. As HU was added to the 3-drug combination regimen rather than substituted for one of the regimen components, the occurrence of additional adverse events in the HU arm compared to the non-HU arm is not surprising. The use of non-enteric-coated ddI in our study may have been responsible for greater safety concerns than would have been the case had an enteric-coated ddI formulation been given. Indeed, in an in vitro
study, Foli et al [38
] showed that HU increases mitochondrial toxicity when given with high doses of ddI. Once-daily non-enteric-coated formulation of ddI (which was used in the present study) results in higher maximal blood concentrations compared to the same doses of enteric-coated ddI, thus making mitochondrial toxicity more likely. Our study was also limited because it was not powered to show significant differences between subjects who received HU versus those who did not.