WHO recommends that ARV treatment programmes in resource-constrained settings choose one potent first-line ART regimen with which to start treatment in the majority of patients. Clinical trials of different triple-drug regimens have generally revealed comparable antiviral potencies ^{32, 36, 39, 42, 63-65}. The choice among these regimens therefore generally relies on other considerations, including side-effect profiles, potential drug interactions, comorbidities (e.g. tuberculosis, hepatitis), the maintenance of alternative options in the setting of treatment failure, and drug availability and cost.

Of the 16 approved ARV agents for the treatment of HIV-1 infection in the USA, six are NsRTIs, one is an NtRTI, three are NNRTIs and six are PIs. Thirteen of these agents have been incorporated into the present guidelines (Table 2).

Table 2. Approved antiretroviral agents included in WHO’s ARV guidelines^a

^aApproved and generally available in industrialized countries as of January 2002.
^bApproved for inclusion in WHO’s Essential Drug List as of April 2002.

The recommended regimens (Table 3) each contain a dual nucleoside component (backbone) to be combined with a PI, an NNRTI or the potent NsRTI, abacavir (ABC). In this context the following potential dual NsRTI components have to be considered: zidovudine(ZDV)/lamivudine(3TC), stavudine(d4T)/3TC, d4T/ didanosine (ddI), ZDV/ddI, ZDV/zalcitabine (ddC), and ddI/3TC. Of these, the first two have emerged as leading candidates for use in initial regimens because of their efficacy, toxicity profiles and years of clinical experience ^{30, 66, 67}. d4T/ddI remains an important dual nucleoside component of potent regimens but cautions have been raised about its potential to cause lactic acidosis, particularly in pregnant women, hepatotoxicity and neurotoxicity (both peripheral neuropathy and a condition resembling the Guillain-Barre syndrome)⁶⁸. Among the NsRTIs, d4T has been most strongly linked to the development of lipoatrophy in some studies ^{69 -72}. Substantial clinical endpoint data support the use of ZDV/ddI ^{73, 74}. ddI can be administered once daily, and the recently introduced enteric coated formulation of ddI substantially improves tolerance of this drug. ddI/3TC appears to have comparable antiviral potency to the other dual NsRTI combinations listed but the data sets supporting its use are more limited. ddC has not been recommended because of the requirement for dosing three times daily and the incidence of peripheral neuropathy. Although direct comparative data are limited, these five dual nucleoside components appear to possess comparable inherent antiviral activity in treatment-naïve persons ^{66, 67, 75, 76}. On the grounds of efficacy, toxicity, clinical experience and fixed-dose combination availability, ZDV/3TC is recommended as the initial dual NsRTI component of choice. However, country-specific decision-making is encouraged (Table 3).

It should be noted that ABC can also be used as part of a dual NsRTI backbone in initial regimens but for the purposes of these guidelines its recommended use is restricted to the cornerstone of triple NsRTI regimens. Certain dual NsRTI components should never be used because of antiviral antagonism (e.g. ZDV and d4T)⁷⁷.

Tenofovir disoproxil fumarate (TDF), the latest addition to the approved antiretroviral armamentarium, is an NtRTI, requiring only intracellular diphosphorylation in order to be active against HIV reverse transcriptase. Operationally, it can be viewed as expanding the NsRTI options. It is considered later in this text because the bulk of clinical trial data have been developed in treatment-experienced populations. Given its tolerance and once-daily dosing, it has substantial potential as a component of once-daily regimens in treatment-naïve persons as well.

Dual NsRTI therapy alone is not recommended as initial therapy because the regimen potencies are suboptimal and the emergence of drug resistance is predictable. However, it is recognized that many HIV-infected individuals in the developing world are being treated with dual NsRTI combinations because potent three-drug and four-drug combinations are not affordable. As dual NsRTI therapy is considered suboptimal in these and other published guidelines, persons currently doing well on dual NsRTI regimens should be considered for switching to one of the potent regimens outlined in this document. If this is not deemed feasible or advisable because of country-specific resource constraints or indiviudual physician/patient considerations, treatment should be continued for those persons who are deemed to be receiving continued benefit. This, however, is not an endorsement of the continued initial use of these regimens and every effort should be made to provide standard-of-care regimens for all patients in whom therapy is initiated.

In order to establish a potent ARV regimen a third drug must be added to the dual nucleoside backbone. The three combinations to consider involve the addition of one of the following: (1) an NNRTI; (2) abacavir; (3) a PI (+/- low-dose ritonavir for pharmacoenhancement). These combinations are PI-sparing, dual-class (i.e. PI- and NNRTI-sparing) and NNRTI-sparing respectively. This is important with respect to the maintenance of alternative treatment options following therapeutic failure. Although two other possible potent initial regimens exist, PI plus NNRTI plus NsRTI(s) or PI plus NNRTI, these are not considered here because they have been less well studied as initial regimens and are consequently not among the first-line options used for initial therapy in the developed world.

Suggestions of differences within and between these three basic regimens have begun to emerge: lopinavir/ritonavir (LPV/r) was associated with virological outcomes superior to those of nelfinavir (NFV) in one trial ⁷⁸. It should also be noted that there are only limited data on the use of ABC-containing triple NsRTI regimens in patients with advanced disease ³⁹. Although important, these clinical trial results are just one factor to be considered when deciding about which PI to employ or whether to use ABC in initial regimens. In general, therefore, the choice should rely on other considerations, including the side-effect profile, potential drug interactions, the maintenance of alternative options in the setting of treatment failure, and drug availability and cost.

Table 3. Recommended first-line antiretroviral regimens in adults and adolescents with documented HIV infection

^a ZDV/3TC is listed as initial recommendation for dual NsRTI component based on efficacy, toxicity, clinical experience and availability of fixed-dose formulation. Other dual NsRTI components can be substituted, including d4T/3TC, d4T/ ddI and ZDV/ddI, depending on country-specific preferences (see text). ZDV and d4T should never be used together because of proven antagonism. Fixed-dose formulations are preferred whenever possible as they promote enhanced drug adherence.

^b RTV-PI includes IDV/r, LPV/r or SQV/r.

When available, fixed-dose combinations are advantageous with respect to the simplification of regimens and consequent improved adherence. The major pharmaceutical manufacturers currently produce three fixed-dose combinations included in these guidelines: ZDV/3TC, ZDV/3TC/ABC and LPV/r. Fixed-dose formulations have also been produced by generic manufacturers (e.g. d4T/3TC/NVP and ZDV/3TC/NVP), which may facilitate simplified regimens, decrease cost and promote adherence if they can be legally used and their quality and bioequivalence have been demonstrated.

All of the initial regimens discussed are applicable for drug-susceptible HIV-1 infection with the main subtypes. For group O HIV-1 subtype or HIV-2 infections, only the triple NsRTI and PI-based regimens should be used because of the inherent resistance of these viruses to the NNRTI class of compounds.

Drug dosages for adults are listed in Annex 7. Relevant drug toxicities and major drug interactions for the agents recommended are listed in Table 10 and Annexes 8 and 11. (See also Chapter XV.)

A. NNRTI-based regimens

NNRTIs are very potent anti-HIV-1 agents in general but are inactive against the group O HIV-1 subtype and HIV-2. There are three approved NNRTIs, but because of the pill burden and the thrice-daily dosing associated with delavirdine (DLV), the two drugs recommended are efavirenz (EFZ) and nevirapine (NVP). Each should be administered in combination with two NsRTIs (Table 3, Annexes 3 and 6). Clinical trial and cohort data suggest that triple-drug combinations involving either drug are at least comparable to PI-based regimens. Although experts at WHO meetings considered that the data for EFZ were possibly the more convincing ^{36, 42, 62}, no definitive, comparative, randomized clinical trial data are yet available which allow differentiation between EFZ and NVP on the basis of potency. At present, therefore, the choice should be based on predicted tolerance and adherence, toxicity profiles, the presence of coexistent conditions such as pregnancy (or the potential to become pregnant), active tuberculosis or other significant coinfections, and drug availability. EFZ is now available as a single 600-mg tablet, and this further enhances the attractiveness of the drug. However, because in utero exposure to EFZ in primates has been shown to cause central nervous system and craniofacial abnormalities, the drug is contraindicated in women on ART who are pregnant or desire to become pregnant. For this reason, NVP should be considered the NNRTI of choice in women of childbearing age unless effective contraception is used. The potential side-effects of NVP, including rash and, particularly, hepatotoxicity, should be considered. EFZ may diminish the effectiveness of oral contraceptives. If this drug is used, therefore, alternative contraceptive methods must also be employed.

B. Triple NRTI-based regimens

For the purpose of these guidelines, triple NsRTI-based regimens are considered to be those containing ABC because of the potency of this agent (Table 3, Annexes 4 and 6)^{39, 79}. The bulk of the data for ABC-based triple NsRTI regimens have been developed in relation to the use of this agent combined with ZDV and 3TC but, given the comparable potency among the dual NsRTI components listed above, some flexibility in choice can be assumed. However, the availability of a fixed-dose combination containing ZDV, 3TC and ABC permits the delivery of a potent triple-drug combination with one pill administered twice daily. A low pill burden and a predicted high level of adherence are therefore major advantages of this approach. In addition, the lack of interaction with rifampicin and stability at room temperature are favourable characteristics. However, this regimen is of uncertain efficacy in patients with advanced disease, there is a risk of ABC hypersensitivity that may affect up to 5% of patients starting on the medication, and there are only limited data on the use of ABC in pregnancy. The risk of ABC hypersensitivity in regions with a high incidence of febrile illnesses such as malaria and tuberculosis could hinder accurate diagnosis of this potentially fatal side-effect.

C. Regimens based on protease inhibitors

Although there are six approved PIs, only four are recommended as first-line agents for reasons of tolerance, clinical trial experience and expert consensus on their applicability in resource-limited settings. These four agents are NFV, indinavir (IDV) combined with low-dose ritonavir (RTV), LPV/r, and saquinavir (SQV) combined with low-dose RTV (Table 3) ^{80, 81}. Amprenavir (APV) is not recommended for initial therapy because of the pill size and burden even when boosted with low-dose RTV; it is best used in the salvage setting pending the availability of the APV prodrug currently undergoing development. RTV in full dose as a single agent poses substantial tolerability problems that limit its utility; it is consequently best reserved for use in low doses as a pharmacoenhancer. A summary of the characteristics of the PI-based regimens is given in Annexes 5 and 6. Each of these PIs, in combination with two NsRTIs, offers sufficient potency to be recommended in this context, but each has drawbacks. NFV is widely prescribed, does not require refrigeration, may provide more alternative options if resistance is detected early and the virus mutates along the D30N pathway, and is safe in pregnancy ^{37, 82-84}. However, the pill burden is moderate, diarrhoea is a common side-effect and the drug cannot be combined with rifampicin. IDV offers the advantage of potency and a substantial data set of experience, including clinical endpoint trials, to support its use ^{29-32, 41, 82, 85}. When the drug is combined with low-dose RTV for pharmacoenhancement it no longer needs to be taken on an empty stomach and IDV/RTV can be given twice daily. The pill burden remains higher than optimally desired, however, and applicability in tropical climates is a concern because a high daily fluid intake is required in an attempt to avoid nephrolithiasis and because it is necessary to refrigerate RTV for longer-term stability. Moreover, both IDV and RTV are incompatible with rifampicin. LPV/r, the most recently approved PI, offers the advantages of potency, coformulation that improves its pharmacokinetics, and relatively good tolerance. On the negative side, experience with LPV/r in pregnancy is limited, and the drug is incompatible with rifampicin. SQV, in combination with low-dose RTV, is another PI option. This combination has been more recently studied and has the advantage of being well tolerated and administered as a twice-daily or possibly a once-daily regimen. Furthermore, it has been reported as compatibile with rifampicin (see Chapter XII). The use of RTV permits either the soft-gel or the hard-gel formulation of SQV to be used; it appears that the latter is associated with a reduced incidence of gastrointestinal side-effects when used in this combination. The hard-gel capsule of SQV should not be administered without RTV pharmacoenhancement because of its low bioavailability when administered alone. The disadvantage of SQV/RTV is the relatively limited data set that currently exists. The need for refrigeration of current RTV formulations in order to achieve stability beyond 30 days is a problem for severely resource-constrained areas. However, where refrigeration is not possible in an individual’s or family’s home, clinic dispensing facilities should be able to store and dispense the agent to their patients on a regular basis or should develop this capacity.

The metabolic toxicities associated with ART in general and PIs in particular are a major concern and the subject of intense investigation ^{86, 87}. Descriptions of these and other toxicities associated with PIs are detailed in Table 10 and Annex 11 (see Chapter XV).