Dose reduction of antiretrovirals: a feasible and testable approach to expand HIV treatment in developing countries


Corresponding Author Sandro Vento, Department of Internal Medicine, School of Medicine, Faculty of Health Sciences, University of Botswana, Private Bag 00713, Gaborone, Botswana. Tel.: +267 3554190; Fax: +267 3105979; E-mail:


Il est très improbable que les systèmes de santé dans les pays en développement soient toujours en mesure d'inscrire de nouveaux patients et de maintenir les patients sous traitement antirétroviral dans le long terme. Cela à cause de la crise économique mondiale, les orientations des bailleurs de fonds en fonction des priorités pour la santé mondiale et les changements dans les nouvelles recommandations de l’OMS pour commencer les antirétroviraux à un seuil de 350 cellules CD4/mm3, qui influeront tous sur la disponibilité de l'aide pour la thérapie antirétrovirale. Nous proposons d'explorer de toute urgence la réduction de la dose des différents antirétroviraux comme une des options possibles pour aider et maintenir le traitement antirétrovirale dans les pays en développement. Des doses plus faibles de stavudine, éfavirenz, atazanavir, darunavir, lopinavir/ritonavir, ténofovir sont à l’étude ou sont à explorer et pourraient réduire les coûts et permettre un accès plus étendu à la thérapie anti-VIH pour les populations dans le besoin. La réduction de la dose pourrait également diminuer les effets indésirables, améliorer la tolérance et l'observance du traitement sans compromettre l'efficacité. Il est urgent de mettre en œuvre des essais de non-infériorité assez larges, bien suivis, comparant des doses plus faibles d'antirétroviraux à celles actuellement utilisées.


Es altamente improbable que los sistemas sanitarios de los países en vías de desarrollo puedan incluir nuevos pacientes y mantener a largo plazo a aquellos que ya están en tratamiento antirretroviral. Ello se debe a la crisis económica global, al cambio de énfasis en las prioridades en salud global de los donantes, y a los cambios en las nuevas recomendaciones de la OMS, de comenzar el tratamiento con antirretrovirales en un umbral de 350 células CD4/mm3, lo cual impactaría sobre el suministro de la ayuda para antirretrovirales. Proponemos que urgentemente se explore la reducción de la dosis de diferentes antirretrovirales, como una posible opción que ayude a mantener el despliegue del tratamiento antirretroviral en países en vías de desarrollo. Se están explorando, o valdría la pena explorar, unas dosis menores de estavudine, efavirenz, atazanavir, darunavir, lopinavir/ritonavir y tenofovir, que podrían reducir costes y permitir un mayor acceso a la terapia anti-VIH de poblaciones necesitadas. La reducción de las dosis también podría disminuir los eventos adversos, mejorar la tolerancia y favorecer la adherencia al tratamiento, sin comprometer la eficacia. Es urgente implementar ensayos clínicos de no inferioridad, razonablemente grandes y con un poder estadístico adecuado, en los que se comparen menores dosis de antirretrovirales versus aquellas actualmente utilizadas.

At the end of 2010, an estimated 34 million people were living with HIV worldwide. Antiretroviral therapy (ART) has been scaled up, and the most dramatic increases have occurred in sub-Saharan Africa, with a 20% rise between 2009 and 2010 alone (UNAIDS World AIDS Day Report 2011). It is estimated that at present 6.6 million people in low and middle-income countries are receiving HIV treatment, i.e. nearly half those eligible (UNAIDS World AIDS Day Report 2011), and between 7.1 million and 8.4 million people could receive ART by the end of 2012 (Renaud-Théry et al. 2011). Of these, 6.6% should be on second-line therapy (Renaud-Théry et al. 2011).

Can this increase be continued and expanded? The global economic crisis, the switch in emphasis of donors’ priorities to global health, and the changes in the new WHO recommendations to initiate ART at a threshold of 350 CD4 cells/mm3 will impact the supply of aid for ART. Thus, it is highly unlikely that the health systems in developing countries will be able to enrol new patients and to sustain patients on ART in the long term at the current cost of antiretrovirals. Within this complex scenario, the persistently high incidence of HIV (Prejean et al. 2011) will result not only in the need to cover the increasing necessity of ARTs for eligible patients, but also, despite the improvements in survival years, to ensure that infection rates remain constant. Drug costs indeed account for as much as 60% of antiretroviral treatment program costs in several countries (Hill et al. 2010), and the production costs of active product ingredient are the main determinant of generic antiretroviral drug prices (Hill et al. 2010).

We advocate for dose reduction of different antiretrovirals to be urgently explored as one of the possible options to help and sustain ART roll-out in developing countries. Antiretroviral drug dosing has not been unchangeable in the past, zidovudine (reduced from 1500 to 500 mg daily) (Volberding et al. 1990) being the clearest example. Indeed antiretroviral doses have often been initially chosen on a fairly arbitrary basis: for several antiretroviral drugs, phase 2 data had shown not only no difference in efficacy between different doses but also better tolerability of lower doses (Hicks et al. 1998; Murphy et al. 2001; Markowitz et al. 2007). Nonetheless, higher doses were chosen for further trials.

Stavudine is a particularly interesting example of dose reduction. A meta-analysis of dose-ranging studies showed that a 30 mg twice daily dose had the same efficacy of the 40 mg twice daily dose and a reduced risk of peripheral neuropathy, the drug’s main side effect (Hill et al. 2007) (Table 1). The dose of this nucleoside analogue was then reduced to 30 mg twice daily. Which antiretrovirals could be tested at lower doses?

Table 1. Studies of reduced doses of antiretrovirals
Antiretroviral agentReferenceMethodDoses studiedOutcomeConclusion
  1. NA, not applicable.

Zidovudine Volberding et al. 1990 Randomised, double-bind study1500 mg daily vs. 500 mg daily vs. placeboProgression to AIDS lower in the 500 mg and in the 1,500 mg than in the placebo groupLower dose showed equal efficacy and improved safety
Stavudine Hill et al. 2007 Meta-analysis30 mg twice daily vs. 40 mg twice dailyLower rates of peripheral neuropathy and lipoatrophy with lower doseLower dose showed equal efficacy and improved safety
McComsey et al. 2008 Randomised,open-label study20 mg twice daily vs. 40 mg twice daily, and 15 mg twice daily vs. 30 mg twice dailyImprovement in mitochondrial indices with lower dosesLower dose showed equal efficacy and improved safety
Efavirenz Hicks et al. 1998 Double-blind, placebo-controlled phase 2 clinical trial600 mg vs. 400 mg vs. 200 mg dailyNo difference between the proportion of patients with HIV-RNA <400 copies/ml at 24 weeks for all three dosesLower doses of efavirenz equally efficacious
ENCORE 1Double-blind, placebo-controlled clinical trial600 mg vs. 400 mg dailyOngoing 96 week study begun in September 2011NA
Lanzafame et al. 2012 Clinical cohort with pharmacokinetic analysis400 mg dailyHIV-RNA persistently <50 copies/ml with improved safetyViral efficacy with improved safety
Atazanavir Avihingsanon et al. 2009 Pharmacokinetic analysis of Thai patients300 mg plus 100 mg of ritonavir vs. 200 mg plus 100 mg of ritonavir daily Same efficacy and plasma atazanavir drug levels as seen in Caucasians on 300/100 mg daily200 mg plus 100 mg of ritonavir dosing sufficient in Thai patients
Giola et al. 2008 Pharmacokinetic analysis of Caucasian patients300 mg plus 100 mg of ritonavir vs. 200 mg plus 100 mg of ritonavir dailySame efficacy of standard dose of atazanavir (300/100 mg daily)Reduction of side-effect (hyperbilirubinemia) and persistence of viral control
Darunavir Lanzafame et al. 2011 Clinical case series with pharmacokinetic analysis600 mg of darunavir plus 100 mg of ritonavir dailyHIV-RNA persistently <50 copies/ml Viral efficacy of lower dose
Lopinavir Murphy et al. 2001 Prospective, randomised, double-blind trial400 mg plus 100 mg vs. 200 plus 100 mg of ritonavir100% of patients on lower dose had suppressed viral load (HIV-RNA <50 copies/ml) vs. 50% on higher doseBetter virological outcome probability related to greater tolerability of lower dose than standard dose
Hill et al. 2009 Pharmacokinetic meta-analysis200/50 mg twice daily vs. 200/150 mg twice daily vs. 400/100 mg twice daily200/150 mg twice daily dose of lopinavir/ritonavir showed similar lopinavir plasma levels to the standard doseHigher ritonavir dose can increase plasma concentration of lopinavir
Ramautarsing et al. 2012 Pharmacokinetic analysis of Thai patients200/50 mg twice dailyMost Thai patients had inadequate lopinavir plasma concentrations but undetectable HIV-RNA at week 12Reduced lopinavir and ritonavir doses do not allow adequate lopinavir plasma concentrations
Lattuada et al. 2011 Clinical case series with pharmacokinetic analysis200/50 mg twice dailyViral efficacy at 12 months even though not all patients had adequate lopinavir plasma concentrationsReduced lopinavir and ritonavir doses allow persistent control of viral replication

One small study showed that stavudine given at 20 mg twice daily can maintain viral suppression while reducing lipoatrophy and other mitochondrial adverse effects (McComsey et al. 2008), and this further reduced dose could be worthy of a trial (Innes et al. 2011).

Efavirenz, a common component of first-line treatment regimens the world over, is used at 600 mg once daily. However an initial phase 2 trial showed no difference in efficacy between 200, 400 or 600 mg once daily dose, combined with zidovudine/lamivudine (Hicks et al. 1998), and plasma drug levels can be much higher in patients with certain CYP2B6 allelic variants (present mostly in Africans) (Haas et al. 2004), or in individuals with low body weight. The ENCORE 1 trial comparing efavirenz (plus tenofovir and emtricitabine) at the standard 600 mg once daily dose vs. 400 mg once daily for 96 weeks in 630 naïve HIV-infected patients started in September 2011 and results will be available in 2014 (Encore1 Trial 2012). Meanwhile a small ongoing study in 33 HIV-infected patients treated with two nucleoside reverse transcriptase inhibitors plus efavirenz at reduced dose has shown no virological failure so far with a minimum follow-up of 15 months (Lanzafame et al. 2012).

Atazanavir is a protease inhibitor currently administered at the dose of either 300 mg with boosted ritonavir or, less frequently, 400 mg once daily. However a pilot study of atazanavir/ritonavir at a dose of 200/100 mg daily in Thai patients showed the same efficacy and plasma atazanavir drug levels as in Caucasians on 300/100 mg daily (Avihingsanon et al. 2009), and the same lower dose guided by therapeutic drug monitoring was effective in 31 Caucasians (Giola et al. 2008).

Darunavir was initially used at a 600/100 (ritonavir) mg twice daily dose in antiretroviral-experienced patients. Subsequently, the ARTEMIS (Boffito et al. 2008), MONET (Clumeck et al. 2011) and ODIN (Cahn et al. 2011) trials have all shown the efficacy of a darunavir/ritonavir 800/100 mg daily dose, which has also a more favourable safety profile, in both naïve patients and in patients with limited antiretroviral experience. We have recently successfully used in a small number of ART-highly experienced patients a further reduced dose of darunavir/ritonavir (600/100 mg once daily) (Lanzafame et al. 2011).

Lopinavir/ritonavir is commonly used at the dose of 400/100 mg twice daily but during its development lower doses have shown similar efficacy to the presently used one. In particular the Abbott 720 trial conducted in naïve patients evaluated three doses: 200/100 mg twice daily, 400/100 mg twice daily and 400/200 mg twice daily. The 200/100 dose showed a very good efficacy (Murphy et al. 2001) but the 400/100 mg was chosen for phase 3 development.

A meta-analysis estimating the effects of lopinavir dose vs. ritonavir dose on lopinavir pharmacokinetics in five pharmacokinetic trials of lopinavir/ritonavir indicated that a 200/150 mg twice daily dose of lopinavir/ritonavir could maintain very similar lopinavir plasma levels to the standard 400/100 mg dose (Hill et al. 2009). In a most recent small study in Thai HIV-infected patients, inadequate plasma concentrations were observed with lopinavir/ritonavir used at an even lower dose (200/50 mg twice daily) (Ramautarsing et al. 2012). However, at week 12, all patients had undetectable plasma HIV-1 RNA (Ramautarsing et al. 2012). We have also used 200/50 mg lopinavir/ritonavir twice daily in six HIV-infected patients and observed C through of lopinavir below the suggested minimum concentration of 1000 ng/ml; however again, all patients continue to have undetectable HIV-RNA levels after 12–20 months on this low dose (Lattuada et al. 2011).

Finally, as pharmacokinetic parameters of tenofovir are dose-proportional and reductions in plasma HIV-1 RNA are dose-related at doses of 75–300 mg daily, it could be worthwhile to test tenofovir at a reduced dose of 225 or 250 mg daily.

How cost effective would be a reduction in doses of antiretrovirals? We used the 2011 antiretroviral drug prices listed in the Medecins Sans Frontieres drugs report (Medecins Sans Frontieres 2011), and calculated a gamma distribution (Fay & Feuer 1997) in order to analyse, in a 1000 individuals sample simulation, the economic impact of lowering drug doses compared with standard therapies. As the cost of ART weights for around 70% of HIV-infected patients’ management costs (Rizzardini et al. 2012), the analysis was conducted considering only the drugs cost, as a significant proxy of resources absorption, and not considering hospitalisations and outpatient activities.

Lowering the dose of lopinavir/ritonavir led to a 100% probability to have a lower economic impact than standard therapies, and reduction of efavirenz doses led to a 99.5% probability to be less expensive than standard therapies. We then performed a budget impact analysis, considering the above two simplification strategies. In the absence of official, reliable data, we assumed that in developing countries 93% of patients are on first line treatment and of those 30% are on efavirenz (therefore overall 27.9% use efavirenz) and that 7% of treated patients are administered a protease inhibitor-based regimen (and of the latter 70%, i.e. 4.9% of all treated patients, receive lopinavir/ritonavir). Further considering that the total number of treated patients within developing countries is 6.6 million individuals (UNAIDS World AIDS Day Report 2011), in a conservative scenario (which does not take into consideration the potential reduction in toxicity of the reduced antiretroviral doses, and the related monetary savings), a therapy cost reduction of 38.56% for lopinavir/ritonavir and of 19.97% for efavirenz were observed.

Saved resources could be used to provide antiretrovirals to additional patients, increasing the number of individuals treated with lopinavir/ritonavir by 62.75% and the number of those treated with efavirenz by 24.96%. The total number of patients treated with either therapy would go from 2 164 800 to 2 827 293 with an increase of 30.6%.

We believe that these data indicate the feasibility of lowering the doses of a number of antiretrovirals without compromising efficacy of treatment. In addition to driving down costs and allowing wider access to anti-HIV therapy for population in needs, dose reduction would decrease adverse events, improve tolerability and favour adherence to treatment. The latter would likely compensate for the reduced forgiving of non-adherence to reduced dose regimens compared with the standard dose. Fewer adverse events would also result in costs savings and a better quality of life for patients.

Fewer treatments for adverse events would result in advantages both for national healthcare expenditure (reduction of the total patients’ management costs) and for pharmaceutical companies. Fewer outpatient consultations and hospitalisations (which comprise almost 30% of the total HIV positive patients’ costs; Rizzardini et al. 2012) due to HIV treatment-related adverse events and long-term toxicities could result in cost savings. Similarly, a decrease of switches due to drug-related adverse events and toxicities would lead to an advantage for pharmaceutical companies, as patients would maintain the same therapeutic regimen for longer periods, thus increasing the market share for providers of less toxic therapies. From the patients’ point of view, a reduction in terms of toxicities and adverse events would improve quality of life and reduce societal costs because patients would be better able to work.

It is urgent to implement reasonably large, well-powered non-inferiority trials comparing lower doses and the currently used ones, and we think that it would be in the best interest even of drug companies and regulatory agencies to propose and fund such trials, as it is ultimately more convenient to access a wider patient population. These trials should also consider economic data, in order to analyse real life-based models, and would also allow to reconsider currently established relationships between plasma levels and efficacy of antiretrovirals questioned by some studies (Langmann et al. 2002).

So far, ART related trials, have used the ability of drugs to maintain patients’ viral load below 50 copies/μl as a determinant variable for non-inferiority of a therapeutic regimen. However, from an economic point of view, a long term presence of viral load and the development of resistance mutations are more important. In a 48 or 96-week observational period, the fact that patients may present a blip with over 50 but fewer than 500 copies/μl, which is a low level, transient increase in viral load, should not define the failure of a treatment (Grennan et al. 2012). Indeed a blip would be clinically acceptable, not modifying the patients’ histories, and would not erode the economic advantage due to lower drug dosage and costs for a long period. In contrast, long term virological failure or a resistance mutation could lead to an unacceptable alternative both from a clinical and economic point of view.