Rational use of antiretroviral therapy in low-income and middle-income countries: optimizing regimen sequencing and switching

Introduction During the 4 years to the end of 2007, the number of people in low-income and middle-income countries (LMICs) receiving antiretroviral therapy (ART) increased from 400 000 to 3 million [1,2]. Although early mortality [3] and retention in care [4] remain significant challenges, the majority of reports from LMICs have shown encouraging immunological, virological and survival outcomes [5–12]. Reported rates of switching to second-line ART regimens have been lower than expected [13–15], in part due to actual rates of treatment success, but mainly because of limited access to both virological monitoring [16] and second-line drugs [14]. Clinicians have also been reluctant to switch therapy [15] due to regimen cost, complexity, inconvenience and lack of subsequent treatment options. As cohorts mature and expand and access to virological monitoring and second-line regimens increase, however, rates of diagnosed treatment failure and switch to second-line regimens will increase [17]. As the cost of second-line regimens are currently three to 20 times more expensive than that of first-line regimens [18], these increases will challenge the cost-effectiveness [19,20] and sustainability [21] of HIV-treatment programmes. An effective response to the challenges of HIV treatment failure in LMICs must include reductions in the cost of second-line agents [22], but changes to commercial regulations, particularly in India, suggest the scale of price reductions seen with first-line agents are unlikely to occur with second-line agents. Strategies to maximize the effectiveness of first-line and second-line regimens and optimize the timing of regimen switching are required to fully utilize the survival benefit of available treatment options, maintain programme cost-effectiveness and enable achievement of universal access to HIV treatment. A comprehensive strategy must be evidence based and focused on the rational long-term use of ART at a population level. The objective of this review is to support the development of these strategies by providing an overview of available evidence with an emphasis on regimen sequencing and switching. How can we prolong effective use of first-line antiretroviral therapy? Drug resistance prior to initiation of combination antiretroviral therapy Primary (transmitted) resistance to nonnucleoside reverse transcriptase inhibitors (NNRTIs) increases the risk of NNRTI-based regimen failure [23], whereas the equivalent relationship remains to be defined for ritonavir-boosted protease inhibitor-based regimens. Although the use of mono-therapy and bi-therapy has been uncommon in LMICs, previous experience of sub-optimal ART is a significant contributor to virological failure in some settings [24,25]. The use of nevirapine monotherapy for the prevention of perinatal transmission is more common and is associated with development of resistance to NNRTIs [26] and reduced virological response to subsequent nevirapine-based therapy [27]. Virological responses can be improved by delaying therapy after delivery [28]. Alternatively, resistance can be limited with the use of alternative or additional [29] antiretroviral agents. Recent analyses suggest primary NNRTI resistance is unlikely to have a significant population-level impact on the effective use of first-line NNRTI regimens in most LMICs in the near future [30,31]. Nevertheless, adequately resourced HIV drug resistance surveillance programmes [32] have the potential to inform treatment strategies and determine the utility of drug resistance testing for treatment of naive patients. Timing of treatment initiation The timing of initiation of ART has important implications for optimizing clinical benefit of first-line regimens. Low CD4 cell count at the time of initiation of ART is common in LMICs [3,5–11] due to constraints on HIV testing and treatment and limited access to CD4 cell counting to guide treatment initiation. This results in increased mortality [33,34], virological failure [35] and drug resistance [36], whereas use of ART is associated with reduced mortality even at high CD4 cell counts [37]. These relationships favour initiation of therapy when the CD4 cell count falls below 350 cells/μl or earlier, but this strategy depends on early diagnosis and referral and the ability of health systems to manage larger cohorts. Given treatment interruption is not currently recommended [37–40], this strategy would also lead to longer duration of drug exposure, with attendant costs and risks of virological failure, drug resistance and subsequent clinical progression in settings with limited treatment options [41]. The long-term risks and benefits of alternative initiation strategies in LMICs remain unknown, but data from treatment interruption studies suggest earlier treatment initiation may be associated with similar or greater reduction in clinical events compared with the use of this strategy in high-income countries [38,42]. Selection of first-line regimen First-line antiretroviral regimens are highly effective in suppressing viral replication. Clinical studies of two nucleoside reverse transcriptase inhibitors (NRTIs) combined with either a NNRTI or a boosted protease inhibitor published since 2003 demonstrate a probability of virological success (HIV RNA <50 copies/ml at 48 weeks) between 59 and 83% by intent-to-treat analysis [43–53]. Efavirenz and nevirapine are both potent agents from the NNRTI class, but in a randomized comparison, equivalence could not be demonstrated [43], and some observational data suggest superior virological outcomes with efavirenz [54–57]. Moreover, mutations associated with reduced activity of etravirine, such as Y181C, are more commonly selected in patients failing nevirapine than in those failing efavirenz [57]. Randomized studies of efavirenz versus a boosted protease inhibitor, lopinavir/ritonavir, have shown better virological suppression in the efavirenz arm, but greater CD4 cell count increase in the lopinavir/ritonavir arm [59]. Ritonavir-boosted protease inhibitors have a high genetic barrier to resistance and are associated with fewer resistance mutations at the time of virological failure compared with NNRTI-based regimens [52,59]. Boosted protease inhibitors are an option when NNRTIs are contra-indicated: infection with HIV-2, intolerance of both nevirapine and efavirenz and in pregnant women requiring treatment for tuberculosis during first trimester. These advantages are, however, counterbalanced by their toxicity, drug interactions, greater potential for sub-optimal adherence [60], storage requirements of some agents and, despite recent price reductions, continuing high cost. A combination of two NRTIs and a NNRTI therefore continues to be recommended by WHO as standard first-line treatment [61]. These assessments may need to be revisited if further price reductions or formulation changes occur. NRTIs are used in combinations due to increased potency and genetic barrier to resistance. Randomized comparative studies have shown significantly greater virological response with tenofovir + emtricitabine as compared to zidovudine + lamivudine [49] and equivalence of virological response when abacavir + lamivudine was compared to zidovudine + lamivudine [50] and stavudine + lamivudine was compared to tenofovir + lamivudine [46]. In all these studies, the NRTI backbones were combined with efavirenz. Comparative data on NRTI combinations administered with nevirapine or boosted protease inhibitor are limited. The selection of an NRTI agent for use in combination with lamivudine/emtricitabine and a NNRTI is driven primarily by cost and toxicity in most LMICs. Despite its significant long-term toxicity, stavudine remains more widely used than zidovudine [14] owing to its lower cost and limited short-term toxicity [62]. Some studies suggest that the risk of zidovudine-associated anaemia can be reduced by starting with a stavudine-containing regimen and routinely substituting stavudine with zidovudine after 6 months of therapy [63,64]. Tenofovir is associated with less toxicity in randomized trials than are thymidine analogues [46,49] and emerging data from populations in LMICs with lower body weight are reassuring [65]. Although there are concerns regarding the prevalence of renal disease in Africa [66] and the feasibility of renal monitoring in many LMICs, the main barrier to more widespread use of tenofovir in first-line regimens is cost: currently a median three-fold increase in total regimen cost in LMICs compared to current first-line regimens [62]. An alternative to two class regimens is a nucleoside-only regimen. Randomized studies in high-income countries have shown that triple NRTI regimens lacking a thymidine analogue fail rapidly [67,68] and zidovudine + abacavir + lamivudine is inferior to standard regimens. Combining tenofovir and zidovudine has potential benefit due to antagonistic resistance pathways. A pilot randomized study comparing zidovudine + lamivudine + efavirenz and abacavir + lamivudine + zidovudine + tenofovir showed equivalent virological suppression [69] and in a large randomized trial ongoing in Uganda and Zimbabwe, the use of zidovudine + lamivudine + tenofovir was associated with HIV RNA more than 1000 copies/ml in 24% at 48 weeks [70], which is comparable to outcomes seen with the use of dual class initial regimens. Within the latter trial, a randomized comparison has shown superior virological suppression, but borderline increase in new WHO stage 3/4 events or death when zidovudine + lamivudine was combined with nevirapine compared to combination with abacavir [71]. The mechanism for this finding is not clear and further comparative studies are warranted. Agents from two new classes have been trialled in treatment-naive subjects. When compared with efavirenz in subjects with pretreatment CD4 cell count above 100 cells/μl, the integrase inhibitor raltegravir demonstrated comparable activity and less toxicity in combination with tenofovir and lamivudine [72]. The potential role of this drug class in LMICs within second-line or even first-line therapy is significant and will be shaped by cost and accrual of additional efficacy, toxicity and drug interaction data. Maraviroc, a CCR-5 antagonist, showed similar, but slightly inferior efficacy when compared with efavirenz [73]. The restriction of efficacy of this agent to patients with R5-tropic only virus and the need for pretreatment tropism assays suggests the role of this class will be limited in most LMICs. Early reports suggested subtype C viruses rarely utilize CXCR-4, but more recent reports from Zimbabwe [74] and South Africa [75] have shown substantial use of this co-receptor. From the NNRTI class, rilpivirine is an investigational agent with potent in-vitro antiretroviral activity and high genetic barrier to resistance [76]. Its low dosage, lack of potential interactions and the possibility of daily dosing suggest this agent may play an important role in LMICs. Adherence to antiretroviral therapy Adherence to ART is a major determinant of the longevity of first-line regimens [77]. A recent meta-analysis of adherence to ART in Africa and North America found a pooled estimate of adherence of 77% [95% confidence interval (CI) 68–85] for African studies compared to 55% (95% CI 49–62) for North American studies [78]. The African data were largely drawn from small cohorts (median n = 100) and additional data from large treatment programmes are needed. Understanding the determinants of adherence behaviour across populations and over time remains critical to the success of first-line data are currently limited The main to ART adherence in LMICs to are or to or of and lack of of treatment Despite the widespread of adherence support programmes in LMICs, a recent meta-analysis of randomized trials of adherence not studies in these settings and are also to be important of treatment these relationships remain may include the of drug and the of clinical of ART treatment and support and and in settings and in How can we first-line failure and optimize switch to second-line therapy? and switch strategies The rational use of ART in LMICs is on the of treatment failure and optimizing the timing of switch to alternative regimens. and switch strategies to the risks of HIV drug resistance and efficacy of second-line and clinical progression and early such as the monitoring that are the risks and benefits of alternative regimens and current and future drug can be monitoring available data regarding the of current WHO of treatment failure for of virological failure Although these data are suggest these a of patients with virological failure and that the majority of patients these are not failing Some studies suggest a CD4 cell count increase of less than cells/μl by 6 of therapy may more rapidly with potential virological failure than current significant remains due to and virological responses to initial ART study suggested use of CD4 cell count of more than between months and was associated with an improved of but similar and studies have some in but have below of WHO for the diagnosis of virological HIV viral testing HIV viral testing can be to patients with an increased risk of virological failure a CD4 cell count or increase of less than or cells/μl over months is for virological failure Alternatively, HIV viral testing can be to patients current early switching and testing but with low for virological studies of testing strategies have been HIV viral testing from high-income countries suggest that more HIV RNA viral monitoring virological outcomes and universal testing months has been the standard of care in high-income countries for many years largely from African however, have shown evidence that access to HIV viral mortality during early ART [3] and a published suggested survival benefit from of CD4 or viral monitoring to clinical monitoring A randomized trial of monitoring strategies in Uganda found in an analysis over a median 3 years of that CD4 cell counting to clinical monitoring was associated with a reduced risk of new or but or mortality benefit was seen with the further of HIV viral monitoring randomized trials of treatment monitoring are ongoing in and Uganda and Zimbabwe The advantages of universal testing are high and for treatment failure and of data. Although feasibility is currently limited in many settings this is to as or assays of viral suppression are an need The cost-effectiveness of viral testing will be on survival and resistance reduction of costs associated with increased use of second-line regimens and costs of viral testing study suggested universal HIV viral testing was in South Africa whereas suggested cost-effectiveness data are to support analyses in a of of universal and testing strategies have been and their cost benefits are HIV viral testing has also been used as an of ART adherence this with virological failure an adherence and regimen if viral of to has been but the of virological suppression and of of NRTI resistance is not studies with of viral suppression and drug resistance are needed. HIV viral and switch for HIV viral testing strategies are the of virological monitoring required to and virological progression and of drug resistance mutations and the viral for regimen switching. A number of studies of patients receiving protease inhibitor-based therapy in high-income countries have shown that progression is unlikely if the HIV viral is below 000 copies/ml or copies/ml or more below Moreover, the risk of disease progression or death has been shown to be lower if the HIV viral is below 000 copies/ml the only randomized trial to this which randomized patients to versus not fully and was not to demonstrate In is clear that resistance even at low of HIV RNA and studies of patients failing protease inhibitor-based regimens have largely shown that the of of drug resistance mutations is of HIV viral study found a of resistance for patients with HIV or HIV viral the rates of drug resistance was of patients over a median months with an of new mutations of and over 6 months and mutations These data may however, be to patients NNRTI-based regimens in LMICs, despite the use of similar NRTI backbones to the cohorts In an analysis of patients with triple class failure and viral published by the demonstrated a CD4 cell count with use of NNRTI-based therapy cells/μl CI to compared with a with protease inhibitor-based therapy cells/μl CI to NNRTI resistance may be more associated with increased mortality and may have less impact on viral when compared with protease inhibitor resistance. the that patients protease regimens NNRTI were less to experience a response in the of HIV during early ART studies from LMICs suggest of drug resistance mutations virological failure of NNRTI-based therapy and an between resistance mutations and subsequent failure HIV viral is associated with of resistance mutations in these settings remains In the of more the of viral testing be Although this may be on such as time on duration of prior viral suppression and is that in the between 3 and in the of more current WHO have suggested a HIV RNA of 000 copies/ml as a for first-line regimen switching in LMICs. Drug resistance testing at time of treatment failure results in improved short-term virological response in patients and is in high-income countries when based on HIV viral increases its benefit systems may the need for for many patients is to be in LMICs and to the monitoring and switch strategies A substantial of patients with virological failure of first-line therapy lack resistance mutations that resistance testing may costs by patients to adherence support than to second-line Although resistance testing is currently expensive this cost could be by use of commonly selected mutations in to demonstrate that failure is by resistance and not only by and switch strategies are by the agents in first-line and second-line regimens. with current treatment however, of data the development of rational can be that mutations associated with resistance to NNRTIs and lamivudine/emtricitabine will be rapidly selected in patients failing The which monitoring and switch strategies most early and long-term and clinical progression and the development of reverse transcriptase by mutations such as thymidine analogue mutations and and mutations associated with efficacy of The of LMICs a are and on data from second-line regimen be The efficacy of second-line regimens will be a of the first-line regimen the duration of viral in the of these drugs and the of drugs available to a second-line regimen. at resistance may the selection of this has not been found to be a major in treatment outcomes to is currently limited evidence to guide the selection of second-line regimens in LMICs after failure of first-line NNRTI-based therapy Clinical data suggest benefit from NRTI despite the of resistance due to reduced of virus NRTIs therefore be as a of second-line therapy even in the of resistance if potent combination regimens are not Ritonavir-boosted protease inhibitors have shown results when used in patients in high-income countries and a genetic barrier to resistance to the need for support from fully drugs in the regimen. on these and the costs and of WHO for a health currently a boosted protease inhibitor with two of nucleoside reverse transcriptase inhibitors The of NRTIs used within a second-line regimen depends on those used in the first-line regimen of or stavudine-containing first-line regimens for commonly of two and or and combined with drugs with a low genetic barrier has been associated with a probability of of two or more mutations has been most with NRTI resistance and to a potent second-line NRTI with lamivudine or emtricitabine that for the or high genetic barrier triple NRTI combinations such as zidovudine + lamivudine/emtricitabine + tenofovir may or selection Nevertheless, and NRTI resistance will with all thymidine analogue regimens if viral continues of first-line regimen and strategies on activity of nucleoside reverse transcriptase inhibitor agents in second-line antiretroviral mutations have a greater impact on NRTI activity than mutations increased risk of with subtype C data on activity in of data on early failure with this reverse transcriptase is a resistance associated most commonly with but also with abacavir and rarely stavudine The prevalence of this viruses from subtype receiving virological monitoring has below in clinical trials with tenofovir [46]. and the occur rarely in the of and in recent study of patients with over additional resistance or data suggest that the combination of the and as seen at time of failure of tenofovir + is associated with reduced viral clinical of this are lacking to zidovudine is increased in the of and selection of resistance will zidovudine mutations if these mutations are This the use of zidovudine failure of tenofovir + Although data on stavudine are less some clinical data suggest has similar This support for the use of a thymidine analogue tenofovir is that in-vitro data suggest may be selected more commonly in subtype C than and a of has been in subtype C patients failing or on the above first-line regimens zidovudine and lamivudine/emtricitabine may therefore be associated with greater efficacy of NRTIs in second-line regimens than stavudine-containing regimens and tenofovir is to be an effective second-line agent in this on the prevalence of NRTI resistance and Alternatively, there are some resistance data to support a sequencing strategy that tenofovir by a thymidine a option for a second-line as some of efficacy can be of the of resistance testing and cost and toxicity are of both thymidine analogue and tenofovir may increase the probability of regimen efficacy in the of resistance testing and may the of additional and abacavir are to but the toxicity of the and cost of the latter not In the increases tenofovir but increases resistance to abacavir and, to a resistance to of boosted protease inhibitor Ritonavir-boosted protease inhibitors have shown utility in patients and use in early therapy from high-income countries efficacy for many of these agents Comparative data are by the use of study populations and of boosted protease inhibitors for second-line treatment in LMICs are not Early data from cohorts universal access to HIV viral testing have shown short-term that the boosted protease inhibitor is providing the majority of treatment efficacy The of protease inhibitor is mainly by cost and requirements and to lopinavir/ritonavir is the only boosted protease inhibitor that not in available in combinations will be in LMICs. has been shown to be to lopinavir/ritonavir in this has a more and is daily has shown improved efficacy and similar in patients when compared with protease or lopinavir/ritonavir WHO has lopinavir/ritonavir and as the main protease inhibitors to be used in the of a second-line regimen in LMICs with a boosted protease inhibitor is an alternative to treatment of patients with failure of initial NNRTI-based This is for LMICs, as efficacy is of to first-line NRTI toxicity is and drug costs are available to however, suggest reduced efficacy of this strategy compared with standard protease therapy and subtype virus has been associated with lopinavir/ritonavir monotherapy failure the use of dual boosted protease inhibitors is a strategy that would be of concerns regarding to first-line agents. A number of small studies have this with encouraging and early clinical but significant toxicity data support the use of a protease inhibitor only strategy for second-line regimens at this reduction is strategy that potential for cost in LMICs. studies and clinical studies have shown virological responses for lower than those currently for lopinavir/ritonavir and Although this strategy can toxicity as as cost, greater of adherence may be required and the risks of treatment failure and resistance need to be in adequately and efficacy of treatment at or has been in a number of studies Although some studies in which therapy was at high CD4 cell counts suggested a role for treatment interruption the and studies showed increased in patients treatment and the study demonstrated that treatment interruption was associated with an increase in disease and major renal or disease this be recommended at this time and additional clinical studies are Recent studies have shown the of integrase inhibitors and and new nucleoside and nonnucleoside reverse transcriptase inhibitors to the of patients in high-income Although there is unlikely to be a need for access to treatment regimens in LMICs for many years these data suggest the possibility of effective second-line therapy from two classes not used in first-line the impact of diagnosis of first-line failure and NRTI and NNRTI efficacy trials of this strategy in patients in LMICs are a The of ART by the by has to a defined by the of universal access treatment and longer have to an increased on programme which is by the of significant increases in drug as larger of patients switch to second-line agents. Although reduction in the price of protease inhibitors in low-income and middle-income countries is an programme these price reductions are to be A comprehensive to the long-term use of ART is to maximize effectiveness and costs over the The of a comprehensive strategy be and short-term treatment and transmission of drug long-term and mortality and programme is but many important evidence and As new drugs available and monitoring and drug costs to strategies for the use of ART in LMICs will to The challenge is but are the benefits of in rational and comprehensive for rational use of antiretroviral therapy in low-income and middle-income for rational use of antiretroviral therapy in low-income and middle-income for with and of the of and the of the in HIV drug resistance and use of ART in low-income and middle-income countries and to the of the and in HIV drug resistance and use of ART in low-income and middle-income countries and to the of the have of has been as of or from and in the has support from and and been for for and has been for in or from and has or as a or for and has and and support from and has as a or for and and support from and