Resistance, Fitness, Adherence, and Potency: Title and subTitle BreakMapping the Paths to Virologic Failure

A near-uniformly fatal clinical syndrome, acquired immunodeficiency syndrome (AIDS), has been transformed during the past 5 years into a treatable infectious disease. The availability of potent antiretroviral agents coincided with the ability to measure levels of circulating virus in vivo. When used in tandem, an understanding of human immunodeficiency virus (HIV) replication dynamics in vivo was made possible, forming the scientific basis for the use of combination antiretroviral therapy.¹ However, the treatment of HIV infection remains far from perfect, and new issues arise with regularity. Critical to achieving optimal therapeutic outcomes is an understanding of treatment failure. Early clinical trials of protease inhibitor monotherapy suggested that the pathway to treatment failure was exclusively via drug resistance.^{2 - 3} Viral rebound was thought to reflect failure of all components of a regimen. Furthermore, it was assumed that the absence of resistance-conferring genotypic changes reflected patient nonadherence.

In this issue of THE JOURNAL, articles by Descamps and colleagues⁴ and Havlir and colleagues⁵ question these assumptions in the context of 2 large clinical trials, Trilège⁶ and AIDS Clinical Trials Group 343.⁷ The inferior outcomes observed in patients randomly assigned to receive less intensive maintenance therapy have been recently published.^{6 - 7} In the articles in this issue, the authors seek to understand the findings.

In the Trilège Study, a 3-month induction phase with zidovudine, lamivudine, and indinavir was followed by randomization to either zidovudine and lamivudine, zidovudine and indinavir, or continued triple-drug therapy if the level of HIV RNA in plasma was less than 500 copies/mL. The primary end point was virologic failure, defined by 2 consecutive plasma measurements above 500 copies/mL on 2 consecutive visits, 6 weeks apart. Fifty-eight (20.8%) of the 279 randomly assigned patients met this end point, 29 receiving zidovudine and lamivudine, 21 receiving zidovudine and indinavir, and 8 receiving triple therapy. Fifty-eight study patients with durable virologic suppression were carefully selected by investigators as case-controls. The results of genotypic studies revealed the presence of the lamivudine resistance-conferring M184V substitution in reverse transcriptase in nearly all patients treated with lamivudine. However, primary-resistance mutations associated with reduced susceptibility to indinavir did not emerge during combination therapy with zidovudine and indinavir or triple therapy. Similarly, zidovudine-associated resistance-conferring mutations were rare and when present were confined to changes at codons 41 and 70 of reverse transcriptase.

Adherence as measured by pill counts revealed a statistically significant difference in median adherence rates between cases and controls for patients prescribed either zidovudine or indinavir during maintenance therapy. Furthermore, patients randomly assigned to receive zidovudine with indinavir only demonstrated statistically significant differences in adherence rates compared with controls. Plasma indinavir levels were found to be lower than expected in 2 groups, those failing triple therapy and those failing zidovudine and indinavir maintenance in association with a greater loss in antiviral efficacy. Indinavir levels tended to be in the expected range in those patients in the zidovudine and indinavir group in whom virologic failure was associated with a modest loss of antiviral activity. Of note, plasma indinavir levels were clearly higher in controls compared with cases in both the triple therapy and zidovudine and indinavir groups.

In the AIDS Clinical Trials Group 343 study, after a 6-month induction with the same triple combination regimen as used in Trilege, patients with plasma HIV RNA levels below 200 copies/mL were randomly assigned to receive zidovudine and lamivudine, indinavir monotherapy, or continued triple therapy. Patients were followed up monthly and the study end point, virologic failure, was defined as a subsequent plasma HIV RNA level of 200 copies/mL or greater. Plasma indinavir levels and resistance testing by both genotypic assay and a novel recombinant phenotypic assay were performed retrospectively in 9 of 23 patients in the indinavir monotherapy group who reached the study end point, as well as in 17 of 75 patients who experienced virologic failure during the induction phase, and 10 controls with sustained suppression throughout the course of study. In those failing indinavir monotherapy, plasma HIV RNA levels of 10³ to 10⁵ copies/mL were found at the time of viral rebound. In all 9 patients, resistance testing showed no reduced susceptibility to indinavir or resistance-conferring genotypic changes. Among patients receiving triple therapy, the M184V codon substitution in reverse transcriptase was observed in 14 of 17 patients. In 1 patient, a primary-resistance mutation in the HIV protease (M41L) was associated with rebound. Otherwise, resistance testing using both assays was consistent with retained indinavir susceptibility.

Plasma indinavir levels were available for 2 patients who received indinavir monotherapy, and 7 who received triple therapy during both suppression and virologic rebound whereas the remainder had drug levels available only during the period of suppression. No differences in weighted mean indinavir concentrations were observed among the 3 groups. However, the proportion of patients with at least 1 extremely low indinavir level was significantly higher in the group failing triple therapy.

These studies indicate that virologic failure is indeed multifactorial and not solely the result of multidrug resistance. Undoubtedly, adherence to a treatment regimen is essential. The time and degree of failure observed in the Trilege study were associated with the degree of adherence. Less toxic, simpler, and more patient-friendly regimens are urgently needed, but as these studies point out, not at the expense of the regimen potency.

Reductions in the potency of antiretroviral regimens during the maintenance phase allowed for the higher incidence of virologic rebound. In patients receiving zidovudine and indinavir and indinavir monotherapy, rebound occurred in the absence of readily demonstrable virus-mediated resistance to the antiviral agents being used. One explanation is that the rebounding virus population is a mixture of indinavir-susceptible and indinavir-resistant quasi species and the more fit population, wild-type, predominates. However, an additional consideration is drug potency.

Bonhoeffer and colleagues⁸ and Perno and colleagues⁹ have suggested that drugs may exhibit differential efficacy in different cellular populations. Reduction of potency of 1 of the maintenance regimens may have allowed ongoing wild-type virus replication in populations of infected cells in which indinavir, or perhaps indinavir and zidovudine, do not exert a strong selective pressure for the emergence of resistant virus. Multiple investigators have reported ongoing viral replication during therapy without demonstrable resistance.^{10 - 12} Potential mechanisms of cellular resistance recently have been identified and include interactions between inducible cellular gene products such as p-glycoprotein (MDR-1)¹³ and multiple drug-resistance proteins¹⁴ with substrates known to include protease inhibitors and nucleoside reverse transcriptase inhibitors, respectively.

Taken together, these studies demonstrate that virologic failure is complex and not exclusively mediated by viral resistance. Furthermore, these studies point out the relevance of resistance testing in the setting of virologic rebound. Nonadherence is clearly a critical factor but cannot be assumed to be the origin of treatment failure in the presence of rebound with wild-type virus. Understanding issues surrounding drug potency and cellular resistance seem critical at this juncture. Perhaps with better understanding of these issues, the elusive universal response to HIV therapy may be achieved.