© World Health Organization
WHO Technical Report Series, No. 914, 2003

Antiretroviral medicines

Evidence provided to the WHO Expert Committee on the Selection and Use of Essential Medicines in support of the application to include 12 antiretroviral (ARV) medicines in the Model List of Essential Medicines was assembled from a series of literature searches, some of which were conducted by staff at the Cochrane Collaboration. For review purposes, the Expert Committee classified this evidence as follows:

• Level 1. Evidence from relevant high quality systematic reviews of unbiased randomized comparative clinical trials.

• Level 2. Evidence from at least one relevant unbiased randomized comparative clinical trial.

• Level 3. Evidence from relevant controlled observational studies.

Nucleoside reverse transcriptase inhibitors

The medicines proposed for listing were zidovudine (ZDV or AZT), lamivudine (3TC), stavudine (d4T), didanosine (ddI) and abacavir (ABC). Typically, two of the first four agents are used to assemble a dual nucleoside “core” to which a third agent, abacavir, a non-nucleoside reverse transcriptase inhibitor or a protease inhibitor, is added. The most commonly used dual nucleoside core is the combination, zidovudine + lamivudine (ZDV + 3TC).

Zidovudine + lamivudine

The literature search retrieved a large number of citations that described randomized controlled trials involving zidovudine + lamivudine, used as a dual nucleoside therapy or in combination with a range of other ARV agents. Of these, 3 meta-analyses and 11 randomized controlled trials provided comparative information on the performance of zidovudine + lamivudine relative to the alternative dual nucleoside regimens that are recommended in the WHO guidelines for ARVs (1); these studies were considered to constitute Level 1 and Level 2 evidence of efficacy.

The trials provided not only unambiguous evidence of the efficacy of zidovudine + lamivudine as a dual nucleoside core (which could be combined with a range of other ARVs), but also data which established the efficacy of the alternative regimens recommended in the WHO guidelines for ARVs (1). More specifically, several studies showed that the efficacy of zidovudine + lamivudine was similar to that of other commonly used dual nucleoside regimens. According to one large direct comparative study, the fixed-dose combination of zidovudine + lamivudine was as effective as the concomitant use of the individual medicines. However, the fixed-dose combination resulted in a higher level of adherence to the prescribed treatment regimen.

The trials also indicated that the combination was fairly well tolerated, although, qualitatively, the nature of the adverse reactions observed tended to vary between the different nucleoside combinations. In general, nausea, anaemia, neutropenia and thrombocytopenia were more common in patients receiving zidovudine-containing regimens, while neuropathic symptoms, and possibly lactic acidosis, were more likely to occur in patients taking other nucleoside combinations.

Stavudine

Nineteen randomized trials involving stavudine were retrieved by the literature search; in nine of these, stavudine was used as part of a three-medicine combination. Results of these trials were considered to constitute Level 2 evidence of efficacy.

Of those studies in which stavudine was combined with another nucleoside reverse transcriptase inhibitor in order to form a dual nucleoside core, five involved didanosine as the other medicine and in three, lamivudine was the second nucleoside reverse transcriptase inhibitor. In direct comparisons of dual nucleoside regimens that are generally considered to be effective, stavudine + didanosine was found to be either equivalent or superior to zidovudine + lamivudine in terms of efficacy; moreover, the dual combination was as well tolerated as zidovudine + lamivudine. Stavudine was also found to be effective when combined with lamivudine; when used in combinations comprising three or more medicines, the pairing stavudine + lamivudine demonstrated an efficacy similar to that of stavudine + didanosine. The combination, stavudine + zidovudine, was not especially effective in trials and thus stavudine was not recommended for use in combination with this particular ARV by the Committee. Overall, the combination stavudine + didanosine was reasonably well tolerated; although pancreatitis occurred more frequently in patients taking this combination compared with those on zidovudine + lamivudine, the latter was more likely to cause anaemia.

Didanosine

Sixteen randomized trials involving didanosine were reviewed, including five studies in which didanosine + stavudine was used as part of a three-medicine combination. These studies were considered to provide Level 2 evidence of efficacy.

In direct comparisons of dual nucleoside regimens that are considered effective, the combination didanosine + stavudine was found to be either as effective as or more effective than, and also as well tolerated as, the combination, zidovudine + lamivudine. When tested as part of a three-medicine combination, didanosine + stavudine in combination with indinavir was found to be as effective as, or more effective than, zidovudine + lamivudine + indinavir and as effective as stavudine + lamivudine + indinavir.

Generally, the combination, didanosine + stavudine + indinavir, was fairly well tolerated. Patients receiving the didanosine + stavudine combination were more likely to develop pancreatitis whereas anaemia was the more commonly observed complication associated with the use of zidovudine + lamivudine combinations.

Abacavir

Abacavir is usually used as a third medicine to complement dual nucleoside therapy and, as such, is used in an equivalent fashion to the non-nucleoside reverse transcriptase inhibitors and the protease inhibitors. Library searches conducted by the Cochrane Review Group for HIV/AIDS produced a total of seven randomized controlled trials and five uncontrolled single arm studies that collectively provided an insight into the efficacy and safety of abacavir for different indications. These studies were considered to constitute Level 2 and Level 3 evidence of efficacy.

In three of the controlled trials, abacavir was used as a component of an initial therapy regimen in ARV-naïve subjects. When used in this way, the efficacy of the three-medicine combination, abacavir + zidovudine + lamivudine, was reported to be superior to that of the two nucleoside reverse transcriptase inhibitors alone. Other studies have demonstrated that abacavir-containing regimens are equivalent to protease inhibitor-containing regimens in achieving viral suppression, with the possible exception of subjects with high baseline viral loads. Furthermore, there was some evidence to suggest that the use of combinations that include abacavir as opposed to protease inhibitors may lead to improved adherence to prescribed treatment regimens.

The effect of the addition of abacavir (relative to a placebo) was studied in patients with viral loads above 400 counts/ml who were receiving dual and triple-medicine regimens (i.e. intensification therapy). Overall, the trial showed that patients on abacavir were more likely to have a viral load of less than 400 counts/ml after 48 weeks of therapy than those taking the placebo. 13% of patients in the abacavir arm had viral loads that were less than 50 counts/ml; in comparison, none of the patients receiving the placebo had viral loads as low as 50 counts/ml. The remaining three trials examined the role of abacavir as an alternative to protease inhibitors (i.e. as replacement therapy) in patients receiving stable combined drug therapy. Results of all three revealed high rates of continued viral suppression in patients randomized to abacavir-based regimens.

The trials indicated that, overall, abacavir is tolerated reasonably well. Rates of adverse reactions were similar to, or less than, those observed in patients receiving protease inhibitor-containing regimens and treatment adherence rates were similar or higher. However, hypersensitivity reactions were associated with the use of abacavir in several trials, with reported rates varying between 2% and 10%. Such reactions were occasionally severe and several fatalities have been recorded. This finding highlighted the importance of proper training for health-care providers, in particular, with regard to the identification of early symptoms and signs of abacavir hypersensitivity reactions.

Abacavir is currently available as part of a fixed-dose combination product, comprising abacavir, lamivudine and zidovudine.

Non-nucleoside reverse transcriptase inhibitors

Two non-nucleoside reverse transcriptase inhibitors were proposed for listing: nevirapine (NVP) and efavirenz (EFV or EFZ). These medicines are usually added as a third agent to dual nucleoside core combinations.

Nevirapine

Library searches compiled in collaboration with the Cochrane Review Group for HIV/AIDS resulted in the retrieval of a substantial number of randomized controlled trials and five uncontrolled single arm studies involving nevirapine. These studies were classified as Level 2 and Level 3 evidence of efficacy.

Several of the randomized controlled trials were included in a published meta-analysis that documented a clear advantage of using nevirapine in combination with two nucleoside reverse transcriptase inhibitors over dual nucleoside therapy alone. The meta-analysis also demonstrated that combinations comprising nevirapine plus two nucleoside reverse transcriptase inhibitors were as effective, but possibly better tolerated, than highly active ARV therapies that included a protease inhibitor. Randomized trials not included in the meta-analysis provided further evidence in support of the benefits of nevirapine-containing regimens, which include once-daily dosing, an improved quality of life and a regression in lipid abnormalities.

As the combination, nevirapine + stavudine + lamivudine, has been the subject of only one randomized clinical trial, information from uncontrolled studies was used to evaluate this particular regimen. On the whole, such data supported the efficacy of this regimen. Evidence reviewed by the Committee for other ARV combinations, indicated that the stavudine + lamivudine nucleoside pair is as effective as the more widely used combination, zidovudine + lamivudine.

The principal adverse effect associated with the use of nevirapine was rash. In most trials, rash occurred in at least 20% of individuals; such rates are higher than those typically observed with the other ARVs. Nevertheless, withdrawal from therapy because of adverse drug reactions (ADRs) was not noticeably more prevalent among patients taking nevirapine than those on other ARVs. Whereas severe liver damage has been reported in some patients on other ARVs, it has been observed only rarely with nevirapine. Risk factors for the development of liver damage with ARVs include heavy alcohol use and co-infection with hepatitis C virus (HCV). Although the rare occurrence of severe liver toxicity is a legitimate concern when using nevirapine in post-exposure prophylaxis (a low-risk situation), the evidence reviewed on this occasion was considered insufficient to warrant discouraging its use in individuals who are infected with HIV.

Efavirenz

Library searches conducted by the Cochrane Review Group for HIV/ AIDS generated a total of 15 good quality randomized clinical trials involving efavirenz. These were considered to constitute Level 2 evidence of efficacy.

When used in initial therapy (six trials), the action of efavirenz in combination with two nucleoside reverse transcriptase inhibitors (zidovudine + lamivudine in five of the six trials) was superior to that of the two nucleoside reverse transcriptase inhibitors alone. Efavirenz-containing regimens were as good as, or better than, protease inhibitor-containing regimens in achieving long-term viral suppression, but efavirenz-containing regimens that lacked nucleo-side reverse transcriptase inhibitors were inferior to those that contained them.

The efficacy of efavirenz as a component of salvage therapy in failing regimens was examined in 5 of the 15 trials. These studies showed that efavirenz was an effective medicine for non-nucleoside reverse tran-scriptase inhibitor-naïve patients who were failing nucleoside reverse transcriptase inhibitor regimens. In such cases, regimens comprising efavirenz plus a protease inhibitor (either nelfinavir or indinavir) were superior to efavirenz given alone and also to a single protease inhibitor.

Efavirenz has also been tested (the remaining four trials) as an alternative to protease inhibitors in patients who are susceptible to the lipodystrophy syndrome that can be induced by these medicines (i.e. switch therapy). The three trials for which data were available showed that use of efavirenz in place of continued treatment with protease inhibitors resulted in comparable or superior viral suppression. These studies also demonstrated that efavirenz was as efficacious as nevirapine and abacavir in maintaining viral suppression but was more likely to be associated with hypercholesterolemia and hypertriglyceridemia.

Overall, efavirenz was tolerated quite well. Although certain side effects were common (e.g. rash, central nervous system effects including dizziness, impaired concentration and dreaming), rates of discontinuation because of toxicity were the same as, or lower than, those reported in patients using combination therapies that included pro-tease inhibitors.

Protease inhibitors

Protease inhibitors, like the non-nucleoside reverse transcriptase inhibitors, are usually added as a third agent to a dual nucleoside core combination. When used in such combinations, these medicines are amongst the most potent suppressors of HIV replication. For this reason, they are generally reserved for use later in the course of a patient’s treatment.

Five protease inhibitors were proposed for listing: nelfinavir (NFV), indinavir (IDV), lopinavir (LPV), ritonavir(r) and saquinavir (SQV). Ritonavir is often used at relatively low doses in combination with indinavir, lopinavir and saquinavir as a booster and not as a protease inhibitor in its own right. At low doses, ritonavir inhibits the metabolism of the companion protease inhibitor, greatly enhancing the latter’s bioavailability. When used in this manner, low-dose ritonavir allows the frequency of protease inhibitor dosing to be decreased and reduces the need for dietary restrictions that are otherwise associated with the use of protease inhibitors. For effective protease inhibition, much higher doses of ritonavir would be needed than those recommended in the present Model List.

Nelfinavir

Twelve randomized controlled clinical trials of nelfinavir were recovered by the literature search. Collectively, these trials evaluated a total of 21 treatment arms, 11 of which involved the use of nelfinavir as a part of triple therapy and 9 involved the use of nelfinavir in combination with two nucleoside reverse transcriptase inhibitors. These studies were considered to constitute Level 2 evidence of efficacy.

In these clinical trials, nelfinavir-containing combinations displayed similar efficacy to other protease inhibitor-containing medicine combinations, with some evidence of superior tolerability. The most common adverse effect was diarrhoea.

Indinavir + low-dose ritonavir

Library searches compiled by the Cochrane Review Group for HIV/ AIDS produced a number of studies documenting the nature of the pharmacokinetic interaction between indinavir and ritonavir. These studies confirmed that the interaction between indinavir and ritonavir results in higher minimum concentrations of indinavir which, in turn, allow the combination treatment to be given twice daily (as opposed to three times daily when indinavir is used alone) and without dietary restrictions.

At full treatment doses, indinavir is known to be an effective protease inhibitor and when used in combination with other medicines -usually two nucleoside reverse transcriptase inhibitors - to have a potent antiretroviral action. Studies on the effect of indinavir in combination with a dual nucleoside core were reviewed as part of the submissions for the nucleoside reverse transcriptase inhibitors (e.g. see Didanosine).

Studies involving the use of indinavir in combination with low doses of ritonavir are, however, limited in number. The literature search revealed only one randomized clinical trial (which compared the efficacy of indinavir + low-dose ritonavir (IDV/r) given twice daily with that of full-dose indinavir treatment three times daily), one non-randomized comparative study and seven uncontrolled studies, that is to say, predominantly Level 3 evidence of efficacy.

In all of these studies, the protease inhibitor combination was given together with other ARV agents, usually two nucleoside reverse tran-scriptase inhibitors. The results provided some evidence to suggest that, over the range of different doses of indinavir and ritonavir tested, the indinavir + low-dose ritonavir combination is as efficacious as indinavir at full dosages (1). As it was not clear whether indinavir + low-dose ritonavir combination therapy was associated with fewer adverse side effects than full-dose indinavir, it would be prudent to assume that, in this respect at least, the former combination offers no real advantage. The convenience of twice-daily therapy is, however, a clear advantage.

At present, indinavir + low-dose ritonavir is not available as a fixed-dose combination.

Lopinavir + low-dose ritonavir

Lopinavir + low-dose ritonavir (LPV/r) is an extremely potent pro-tease inhibitor. Library searches performed by the Cochrane Review Group for HIV/AIDS recovered details of five randomized clinical trials involving lopinavir + low-dose ritonavir which were considered to constitute Level 2 evidence of efficacy.

In three of the five trials, lopinavir + low-dose ritonavir in combination with two nucleoside reverse transcriptase inhibitors (typically stavudine + lamivudine) achieved high rates of viral suppression in ARV-naïve subjects. Lopinavir + low-dose ritonavir, in a single dose, was reported to be as effective as lopinavir given twice daily, and when administered in combination with stavudine + lamivudine, was superior to a corresponding nelfinavir-containing combination. The two remaining trials demonstrated the potency of lopinavir + low-dose ritonavir as a rescue therapy in subjects who had increased viral loads despite treatment with multiple ARVs.

Lopinavir + low-dose ritonavir is currently available as a fixed-dose combination from two manufacturers.

Saquinavir + low-dose ritonavir

The Cochrane Review Group for HIV/AIDS who conducted the library search for studies involving saquinavir + low-dose ritonavir (SQV/r), retrieved details of six randomized clinical trials which were considered to represent Level 2 evidence of efficacy.

Across the six trials, the most commonly used regimen was 400mg of each medicine, twice daily, although different doses of the two medicines were also tested. One trial documented an apparent therapeutic equivalence of a range of doses of the two medicines when combined with two nucleoside reverse transcriptase inhibitors. However, the 400mg + 400mg dose, being the lowest total dose of the medicines and also the dose with the best tolerance, was considered to be the most attractive option. In another large trial, the efficacy of saquinavir + low-dose ritonavir at a dose of 400mg + 400mg was compared with that of indinavir or ritonavir given in full doses; results indicated that the saquinavir + low-dose ritonavir combination was at least as effective overall and more effective in patients who had not been previously treated with ARVs than the single medicines. Although the remaining trials provided some additional insight into the potential of this particular combination, their usefulness for assessment purposes was impaired by a number of inherent study design shortcomings. For example, several trials involved suboptimal saquinavir + low-dose ritonavir combinations (e.g. saquinavir + low-dose ritonavir in combination with only one nucleoside reverse tran-scriptase inhibitor); another compared the saquinavir-containing regimen with one that is not recommended in the WHO guidelines (1). Overall, however, the evidence was considered to be of reasonable quality and probably better than that which supports the combined use of indinavir + low-dose ritonavir.

Adverse effects were broadly similar to those observed with other protease inhibitor combinations, but without the problems of neph-rolithiasis, an effect that is typically linked to the use of indinavir.

Artemether + lumefantrine (fixed-dose combination)

WHO recommends that when used for the treatment of acute uncomplicated multidrug-resistant falciparum malaria, artemisinin and its derivatives should be administered in combination with another effective blood schizontocide in order to reduce recrudescences and to slow the development of resistance (2). Such artemisinin-based combinations, which include artesunate + mefloquine and artemether + lumefantrine, have several distinct advantages over other anti-malarials in that: (i) they are capable of producing a rapid clinical and parasitological cure; (ii) as yet there is no documented parasite resistance to them; (iii) they reduce gametocyte carrier rates; and (iv) they are generally well tolerated. At present, only ad hoc combinations of artesunate + mefloquine are used operationally for the treatment of acute falciparum malaria in areas of multidrug resistance. In such settings, fixed-dose combinations would have a number of practical advantages over the ad hoc combinations, including ease of use and improved adherence to treatment regimens in the target populations.

At present, artemether + lumefantrine is the only artemisinin-based fixed-dose combination that has been both widely studied and registered for the treatment of acute multidrug-resistant falciparum malaria. It has been shown to be an efficacious and safe formulation when used for the treatment of acute uncomplicated falciparum malaria in Africa, China and Thailand. Studies in Thailand, for example, have demonstrated that artemether + lumefantrine provides similar cure rates to artesunate + mefloquine in areas prone to multidrug-resistant falciparum malaria and, as such, has the potential to serve as a suitable alternative to artesunate + mefloquine in these areas. At its meeting in 1999, the Committee rejected a proposal to add the fixed-dose combination, artemether + lumefantrine, to the Model List on a number of grounds, including lack of data on operational use and concerns about compliance and cost (see section 6.1.2). These issues have now been addressed as follows:

• Operational experience. Due to unacceptable levels of resistance to sulfadoxine + pyrimethamine in the region, KwaZulu Natal developed a malaria treatment policy based on the use of artemether + lumefantrine as its first-level medicine. The policy, the first of its kind to be adopted by a department of health in Africa, was implemented in February 2001 in conjunction with improved vector control measures. At the end of 2001, a dramatic 78% reduction in the total number of malaria cases (from 41786 cases in 2000 to 9443 cases in 2001) and an 87% decrease in the number of malaria deaths was recorded. This impressive improvement in the malaria situation was attributed to the combined effects of the increased use of residual household spraying and the replacement of an ineffective malaria treatment by the fixed-dose combination, artemether + lumefantrine.

• Compliance. In household surveys conducted following the introduction of the above policy in KwaZulu Natal, 95.4% of patients surveyed claimed that they had completed their course of treatment with artemether + lumefantrine and only 0.5% admitted that they had medicines remaining at home. The corresponding figures for Mpumalanga were 86.6% and 1.3%, respectively. New packaging designed to improve compliance in patients of low literary status has been developed and its effectiveness is currently being assessed.

WHO clinical guidelines on the use of artemether + lumefantrine, published in 2000, state that a 6-dose regimen should be adopted as the standard for all age groups and in all situations in order to avoid confusion and to ensure the highest efficacy and reliability possible with this combination (2). The manufacturer has agreed to these recommendations. As only the 4-dose regimen has been registered for use in selected malaria-endemic developing countries (mainly in Africa), further studies with the 6-dose regimen are required to support its registration in such countries.

• Prices. The manufacturer has agreed to provide the fixed-dose combination, artemether + lumefantrine, as Coartem® to WHO at a no-profit price. As of February 2002, the price to WHO of the range of treatment doses, based on a 6-dose regimen, are as follows:

- children 10-14kg = US$ 0.90;
- children 15-24kg = US$ 1.40;
- children 25-34kg = US$ 1.90;
- adult > 35kg = US$ 2.40.

These prices are approximately half those presently being charged for the medicine in KwaZulu Natal and equivalent to the lowest available price for the ad hoc combination of artesunate + mefloquine that is currently used in Cambodia. It is anticipated that further reductions in the price of artemether + lumefantrine may be possible in the future for the most vulnerable population groups.

References

1. Scaling up antiretroviral therapy in resource-limited settings: guidelines for a public health approach. Geneva, World Health Organization, 2002 (available from the Internet at http://www.who.int/hiv/topics/arv/ISBN 9241945674.pdf).

2. Guidelines for establishing DOTS-Plus pilot projects for the management of multidrug-resistant tuberculosis (MDR-TB). Geneva, World Health Organization, 2000 (document WHO/CDS/TB/2000.279).