Drug Resistance after Failure of Initial Antiretroviral Therapy in Resource-Limited Countries

The report in this issue of Clinical Infectious Diseases by Sungkanuparph et al. [1] from Thailand provides important confirmation of a long-standing concern regarding failure of initial antiretroviral therapy in resource-limited countries. The combination of stavudine (or sometimes zidovudine), lamivudine, and nevirapine is a widely available, inexpensive, generic regimen that is typically coformulated and widely used for first-line treatment in most developing countries that provide such therapy. At one time, this was also an acceptable regimen in the developed world. However, stavudine is no longer recommended by treatment guidelines from either the Department of Health and Human Services or the International AIDS Society-USA, and nevirapine is categorized as an alternative agent in both sets of guidelines. These changes have come about primarily for reasons of toxicity: stavudine is associated with peripheral neuropathy and mitochondrial toxicity, including lipoatrophy and potentially fatal lactic acidosis/hepatic steatosis. Nevirapine use can cause severe skin reactions and/or hepatic toxicity, including potentially fatal hepatic necrosis, especially in women with CD4+ cell counts >250 cells/mm3. Despite these toxicities, this regimen continues to be widely used because of its convenience, its low cost, and its inclusion in guidelines from the World Health Organization and from many individual countries. The study by Sungkanuparph et al. [1] addresses another important drawback of this regimen: the concern about resistance and its effect on the availability of second-line therapy. This concern is more relevant for the developing world than for resource-rich countries where there are more options for subsequent therapy and where frequent viral load monitoring can identify virologic failure before extensive resistance has developed. In this study, 98 patients who experienced virologic failure while receiving stavudine, lamivudine, and nevirapine were identified. Virological failure was defined as confirmed viral rebound to an HIV RNA load >1000 copies/mL after virologic suppression. Patients had been receiving antiretroviral therapy for a median duration of 20 months and had a median viral load of 4.1 log10 copies/mL at the time that virologic failure was determined to have happened. Genotypic analysis found high levels of resistance to lamivudine (89% of patients) and nonnucleoside reverse-transcriptase inhibitors (NNRTIs; 92% of patients), which is not surprising, given the low genetic barrier to resistance for these agents. Of greater concern, however, was the degree of resistance demonstrated to nucleoside analogue reverse-transcriptase inhibitors (NRTIs) other than lamivudine. Thirty-seven percent of patients had ⩾1 thymidine analogue mutation (TAM)—mutations that are selected by stavudine and zidovudine and that cause cross-resistance to all other NRTIs. Thirteen percent of patients had ⩾3 TAMs, and 8% had the Q151M mutation, which causes extensive NRTI cross-resistance. NRTI mutations were more common in patients with HIV RNA loads >4 log10 copies/mL than among those with HIV RNA loads ⩽4 log10 copies/mL. The authors estimated that 48% of patients would have limited options for second-line therapy using the drugs available in Thailand, and this number increased to 63% among patients with HIV RNA loads >4 log10 copies/mL. Regardless of where they are treated, patients who experience treatment failure with an NNRTI-containing initial regimen typically have NNRTI resistance and require the use of a protease inhibitor (PI) in their next regimen. Lamivudine resistance is also extremely common with early treatment failure. However, in the developed world, there has been a movement away from the use of thymidine analogues (stavudine and zidovudine) in favor of tenofovir and abacavir, which are safer and which do not select for TAMs. When thymidine analogs are used, NRTI cross-resistance can be prevented by identifying virological failure and modifying therapy early, before TAMs have accumulated. Thus, it is easy to design fully suppressive second-line regimens after treatment failure. However, tenofovir and abacavir either are not available or are prohibitively expensive in most resource-limited countries, and viral load testing is either performed infrequently or not at all. In the Thai study [1], viral load testing was performed every 3–6 months, but this is by no means the standard elsewhere. Thus, by the time that treatment failure has been identified, often on the basis of a decrease in the CD4+ cell count or clinical progression, virological replication has been ongoing for a long period, and the extent of TAM-mediated NRTI cross-resistance may be much greater than it was in this study. As a result, there is great concern that a large proportion of patients in the developing world whose first-line regimens fail will have no good options for second-line therapy. A partial solution to this problem would be to increase the number of drugs available for second-line therapy. The authors point out that, if tenofovir and abacavir were available, the proportion of patients with limited second-line options would decrease dramatically. This may apply to the patients in this study, who underwent viral load testing on a regular basis, but in places where viral load testing is prohibitively expensive or unavailable, more extensive NRTI cross-resistance may also result in resistance to those 2 agents, leaving only the PIs as fully active drugs. Another option is to learn more about the ability of monotherapy with ritonavir-boosted PIs to fully suppress viral replication in PI-naive patients. Data from a number of small observational studies and clinical trials suggest that monotherapy with lopinavir-ritonavir (and possibly with other boosted PIs) may be suppressive in a large proportion of patients who receive it [2,3–4]. However, these data are preliminary, and there has been some evidence of higher rates of virological failure and PI resistance with this strategy. Moreover, boosted PIs remain expensive, and with the exception of the lopinavir-ritonavir tablet formulation, they require refrigeration. Sungkanuparph et al. [1] emphasize the importance of preventing resistance by identifying virological failure early. Frequent viral load testing would go a long way toward preserving future treatment options, especially if therapy could be modified before the emergence of TAMs. In the developed world, viral load testing became available at about the same time as HAART, but that has not been the case in resource-limited settings, where initial therapy is inexpensive, but the cost of viral load testing remains high. Instead, therapy is monitored using infrequently measured CD4+ cell counts or even clinical indicators, which are crude and insensitive measures of virological response. Efforts are underway to develop less expensive methods for monitoring viral load—an urgent need, if we are to be able to offer effective second-line therapy in the developing world. Another option is to reconsider our choice of initial therapy for resource-limited settings. No one argues that the regimens in use today are optimal: their flaws with respect to both toxicity and resistance are now apparent. However, the preferred drugs, tenofovir and abacavir, are either too expensive or not available in most of the developing world. Until that changes, the only other option is didanosine, which is available generically and is often used in second-line therapy. Didanosine is not without its drawbacks: it can cause pancreatitis, and like stavudine, can cause peripheral neuropathy and mitochondrial toxicity. It must also be given on an empty stomach, whereas the other agents used for first-line therapy can be given without food restrictions. However, didanosine has a distinct advantage over the thymidine analogues in that it does not select for TAMs. Unlike thymidine analogues, for which resistance is cumulative with ongoing viral replication, didanosine is more like tenofovir and abacavir, with selection of the K65R or L74V mutations. Although these mutations can cause some loss of susceptibility to abacavir (K65R and L74V) and tenofovir (K65R), they increase susceptibility to zidovudine, leaving at least 1 fully active NRTI for second-line therapy. The authors describe the current sequencing approach in Thailand in which patients who experience treatment failure with stavudine, lamivudine, and nevirapine might switch to a PI with zidovudine and didanosine. They correctly point out that this would be an effective regimen for patients who had developed resistance to only lamivudine and NNRTI. However, the combination of zidovudine and didanosine is rarely used in the developed world, in part because it does not include lamivudine, which is a component of virtually all zidovudine-containing regimens for resistance reasons. In addition to being a potent antiretroviral agent in its own right, resistance to lamivudine increases zidovudine susceptibility and helps to slow or prevent the emergence of TAMs and multiple-nucleoside mutations, which were observed with the zidovudine-didanosine combination in the pre-HAART era. Moreover, patients who develop TAMs while taking a stavudine-containing combination regimen would be expected to have as much resistance to zidovudine as they did to stavudine, and susceptibility to didanosine might also be compromised. In contrast, failure of an initial regimen containing didanosine plus lamivudine would typically result in M184V plus either K65R or L74V, each of which would enhance zidovudine activity in the subsequent combination. Assuming that the second-line regimen could not contain more than 3 drugs for cost reasons, lamivudine might be continued in order to maintain the beneficial effects of M184V. Thus, while a second-line regimen containing zidovudine, lamivudine, and a PI might only be considered a 2-drug combination because of lamivudine resistance, that represents a substantial improvement over the status quo, in which PIs are often being used without adequate NRTI back-up because of extensive NRTI resistance. Access to antiretroviral therapy in the developing world has had a dramatic impact on HIV-associated morbidity and mortality. However, if we hope to offer more than just a single drug regimen to HIV-infected patients, we must recognize the limitations of the drugs being used and the resistance problems that can result from suboptimal laboratory monitoring. As we work to increase access to antiretroviral therapy to the millions who need it, we must also strive to expand the number of low-cost treatment options, develop technologies for affordable viral load monitoring, and study alternative strategies for initial therapy that may preserve future options. Potential conflicts of interest.J.E.G. has been a consultant for Boehringer/Ingelheim, Bristol-Myers Squibb, Gilead Sciences, GlaxoSmithKline, Merck, Panacos, Pfizer, Vertex, Tibotec, and Mongram Biosciences and has received honoraria from Abbott Laboratories, Gilead Sciences, Tibotec, and Mongram Biosciences.