Sub-Saharan Africa is the region most affected by the pediatric HIV pandemic. Approximately 2.3 million of the 2.5 million children living with HIV infection reside in this region. Global trends suggest access to services is increasing for the prevention of mother-to-child transmission, and that treatment for children with established infection is marginally improved. Between 2004 and 2009, annual numbers of new childhood HIV infections worldwide declined by 24% to 370 000, HIV-related child deaths decreased by 19% to 260 000, and, at the end of 2009, antiretroviral therapy (ART) coverage reached 354 000 children.[1]

Since 2003, the WHO has published global ART guidelines for children, based on a public health approach to scaling up care for HIV-infected children. This approach includes principles of simplicity, standardization, scientific rigor, and equity.[2] Unlike American and European treatment recommendations, WHO guidelines were primarily designed to provide technical guidance to pediatric treatment programmes in resource-limited settings (RLS), and have undoubtedly influenced treatment programmes throughout sub-Saharan Africa. In response to research developments, global treatment guidelines have become more complex. The most recent revision of the WHO guidelines, published in July 2010, grades the quality of evidence of each recommendation.[3]

The 2010 WHO guidelines recommend that all children <24 months of age with HIV infection be started on ART regardless of their clinical status or CD4 count results; although this approach is strongly recommended, the quality of evidence is moderate for those below 12 months of age and very low for those between 12 and 24 months of age.[3] To initiate treatment early, HIV-infected infants should be identified as soon as possible during the first few months of life. However, implementation of early infant diagnosis programmes has proved challenging. At present, most pediatric ART programmes in RLS still report an average age above 3 years at ART initiation. An analysis of seven large South African pediatric ART cohorts documented a median age of 3.6 years at the start of ART.[4] In another study conducted at 16 clinics in western, eastern, and southern Africa, the median age at ART initiation was 4.9 years.[5] The updated WHO recommendations should substantially improve access to early ART. However, attention should also be directed to optimizing treatment regimens with the goal of maintaining children and youth successfully on ART through to adulthood, while at the same time minimizing treatment side effects. This is particularly true of young HIV-infected children who will experience prolonged exposure to antiretroviral agents.

Many of the WHO recommendations regarding first- and second-line regimens are based on low or moderate quality evidence[3] New evidence has emerged, which may better guide the formulation of first- and second-line regimens. Future revision of global and national ART guidelines should consider these developments. In addition, newer agents and drug classes are increasingly being investigated. The role for these agents in pediatric treatment regimens should be comprehensively evaluated.

1. Aim of the Review and Literature Search

This review was based on an invited oral presentation delivered at the 26th International Pediatric Association Conference.[6] The aim of this review was to analyze the latest WHO first- and second-line regimen recommendations for HIV-infected children (table I), review the impact that recent developments may have on these recommendations, and provide selective commentary on the potential role of newer antiretroviral agents.

Table I
figure Tab1

Preferred antiretroviral therapy regimens recommended by the WHO (compiled from information from the WHO[3])

Full size image

Peer-reviewed research publications listed in the PubMed database, and abstracts of oral papers and posters presented at the International AIDS Society conferences, Conference on Retroviruses and Opportunistic Infections meetings, and International Workshop on HIV Pediatrics meetings were the primary information sources. The search was limited to all papers and abstracts published before 2 March 2011. Several combinations of search terms were employed, including ‘antiretroviral therapy’ or ‘regimen’ and ‘HIV infection’ and ‘children’; ‘tenofovir’ and ‘children’; ‘nevirapine’ and ‘children’; ‘nevirapine’ and ‘switching’ and ‘children’; ‘nevirapine’ and ‘exposure’ and ‘resistance’ and ‘children’; ‘antiretroviral therapy’ and ‘nevirapine exposure’ and ‘children’; ‘antiretroviral therapy’ and ‘tuberculosis’ and ‘children; ‘efavirenz’ and ‘rifampicin’ and ‘children’; ‘protease inhibitors’ and ‘rifampicin’ and ‘children’; ‘rifabutin’ and ‘antiretroviral therapy’; ‘second-line’ and ‘antiretroviral therapy’ and ‘children’; ‘darunavir’ and ‘children’; ‘raltegravir’ and ‘children’; ‘etravirine’ and ‘children’; ‘atazavavir’ and ‘children’; and ‘maraviroc’ and ‘children’.

2. First-Line Regimens

2.1 Preferred Nucleoside Reverse Transcriptase Inhibitor (NRTI) Backbone

Of the three WHO-recommended nucleoside reverse transcriptase inhibitor (NRTI) backbone options (table II), zidovudine plus lamivudine is preferred because this combination is widely available in RLS, and generally well tolerated in children. The stavudine-containing option is the least favored based on its safety profile; particular stavudine safety concerns include increased risk of fat redistribution and hyperlipidemia.[7] Relative inexpense makes the stavudine option very attractive to ART programmes in RLS.

Table II
figure Tab2

Nucleoside reverse transcriptase inhibitor (NRTI) backbone options recommended by the WHO (compiled from information from the WHO[3])

Full size image

A randomized controlled clinical trial comparing three different NRTI backbones in ART-naïve children demonstrated that lamivudine and abacavir was more effective than either zidovudine/lamivudine or zidovudine and abacavir in terms of virologic suppression rates, growth changes, and regimen switching rates. Superiority of lamivudine/abacavir persisted beyond 5 years of treatment.[8,9] Favorable pharmacokinetic data suggest that once-daily abacavir/lamivudine could be administered to children as young as 3 months of age.[10] Furthermore, an abacavir-containing backbone preserves the thymidine analogues for second-line regimens.

Tenofovir-containing regimens are not used in routine pediatric practice because of their potential adverse effects on bone mineralization. In this regard, research findings have yielded conflicting results, with some reports documenting reduced bone mineralization and others showing that tenofovir-containing regimens do not decrease bone mineralization.[11–14] However, tenofovir is an effective antiretroviral agent in children and adolescents.[15] In children aged 12 years and older with hepatitis B co-infection, WHO recommends that tenofovir is used in combination with either lamivudine or emtricitabine as the preferred backbone because of the additional antiviral activity against the hepatitis B virus (HBV). A randomized controlled trial in adults recently showed that tenofovir plus emtricitabine was superior to emtricitabine alone for treating hepatitis B co-infection. Although the study sample size was small, 90% of the patient group that received a fixed-dose combination of tenofovir/emtricitabine achieved a plasma HBV viral load determined by polymerase chain reaction below the detectable concentration (<170 copies/mL) after 48 weeks, compared with 33% of patients who received emtricitabine alone.[16] There are no similar randomized pediatric studies in the published literature.

In summary, WHO NRTI backbone recommendations are appropriate for pediatric practice in RLS, based on considerations of safety, efficacy, relative cost of NRTI backbone options, and accessibility. Abacavir-containing regimens should be more widely promoted because of the associated thymidine analogue-sparing benefit, improved toxicity profile, and lower risk of developing NRTI cross-resistance. The safety of tenofovir should be further explored.

2.2 Non-Nucleoside Reverse Transcriptase Inhibitor (NNRTI)- or Protease Inhibitor (PI)-Containing Regimens

The WHO guidelines recommend that the protease inhibitor (PI) fixed-drug combination lopinavir/ritonavir be used as the ‘third’ drug in initial therapy for children <24 months of age with prior exposure to nevirapine for the prevention of mother-to-child HIV transmission. Nevirapine is the recommended ‘third’ drug in the absence of prior nevirapine exposure.[3]

The International Maternal Pediatric Adolescent AIDS Clinical Trials group (IMPAACT) P1060 study, a phase II, randomized controlled trial, compared outcomes in young children treated with either nevirapine or lopinavir/ritonavir-containing regimens. In cohort 1, a lopinavir/ritonavir-containing regimen was shown to be more efficacious than nevirapine-containing ART in HIV-infected children who had previously been exposed to perinatal nevirapine. A total of 164 children were randomized; 152 aged 6–24 months and 12 aged 24–36 months were randomized to receive nevirapine, lamivudine, and zidovudine, or lopinavir/ritonavir, lamivudine, and zidovudine. The primary endpoint was virologic failure or discontinuation of treatment by week 24. More children in the nevirapine-treated group achieved this endpoint (39.6% vs 21.7%; p = 0.02), confirming the superiority of lopinavir/ritonavir-containing ART in this setting (table III).[17]

Table III
figure Tab3

Proportion of children reaching the primary endpointa in the two cohorts of the IMPAACT P1060 trial[17,18]

Full size image

Results from cohort II of the P1060 study indicated that lopinavir/ritonavir-containing ART was more efficacious than nevirapine-containing ART in HIV-infected children who had never previously been exposed to nevirapine. In this cohort, 287 children aged 2–36 months (77 were <12 months of age) were randomized to receive nevirapine, lamivudine, and zidovudine, or lopinavir/ritonavir, lamivudine, and zidovudine. An interim analysis showed that after 24 weeks, 40.1% of children receiving the nevirapine-containing regimen experienced inadequate virologic suppression or had discontinued therapy compared with 18.6% of the children receiving the lopinavir/ritonavir-containing regimen. The 24-week between-treatment differences in the primary endpoint were 22% (p = 0.03) in children <12 months of age and 21.3% (p < 0.001) in those ≥12 months of age, in favor of children receiving the lopinavir/ritonavir-containing regimen.[18]

Taken together, the results of both cohort studies provide unequivocal evidence of the superiority of lopinavir/ritonavir-containing regimens over nevirapine-containing regimens in young, HIV-infected children irrespective of whether or not they had prior nevirapine exposure (table III).

For children >24 months of age, WHO recommends a non-nucleoside reverse transcriptase inhibitor (NNRTI)-containing first-line regimen irrespective of whether or not there was prior NNRTI exposure. The Paediatric European Network Pediatric AIDS Clinical Trials-1 (PENPACT-1) study, a randomized controlled trial, compared the efficacy of NNRTI-containing (n = 132 patients) and PI-containing (n = 131) regimens in older children (median age of 6.5 years). Children who had received perinatal nevirapine were excluded from this trial. Long-term virologic, immunologic, and clinical outcomes were determined. After a 4-year follow-up period, the outcomes of the two regimen types were similar. In particular, the proportions of children with viral load <400 copies/mL and <50 copies/mL were similar in both treatment groups (p = 0.77 and p = 0.35, respectively), and the mean percent increase in CD4 count from baseline to 4 years was 13.7% in the PI-treated group and 15.2% in the NNRTI-treated group (p = 0.19).[19] While the results of this study clearly establish a role for NNRTI-containing first-line therapy in this patient group, there are fewer data evaluating this first-line option in HIV-infected children >24 month of age with prior NNRTI exposure. An observational study of 92 children (44 nevirapine exposed, median age 20.3 months; and 48 nevirapine unexposed, median age 94.3 months) showed that after 48 weeks of nevirapine-containing ART, 76% and 80% of the nevirapine exposed and unexposed children, respectively, achieved a viral load of <400 copies/mL.[20]

Thus, the WHO recommendation to treat children <24 months of age with an NNRTI-containing first-line regimen in the absence of prior nevirapine exposure is not based on strong evidence.

2.3 Nevirapine-Based Therapy

In December 2008, 78% of the world’s children taking ART were receiving a nevirapine-containing regimen.[21] Nevirapine is relatively inexpensive and palatable, and dosing has been established in young infants. It has been co-formulated in fixed-dose combinations (FDCs) that do not require cold chain management. Several generic pediatric FDCs have been developed that provide therapeutically adequate drug levels of nevirapine in children and are chewable or dispersible.[22,23]

Nevirapine is relatively safe for pediatric use.[24] In a controlled trial, children were randomized to start Triomune Baby or Junior tablets (Cipla; FDCs containing 6 and 12 mg of stavudine, respectively; 30 and 60 mg of lamivudine, respectively; and 50 and 100 mg of nevirapine, respectively) either at full dose from the start or at half the dose daily together with stavudine and lamivudine daily for 14 days prior to full-dose commencement. Adverse events were relatively few, even in those receiving the full dose from the start. More children in the full-dose group developed rash, which resolved in the majority of patients after dose reduction and subsequent re-escalation. There were no significant between-treatment differences in biochemical abnormalities, and these abnormalities were not considered clinically significant.[25] Adult reports suggest that long-term use of nevirapine is also safe.[26]

Previous exposure to nevirapine is associated with a high risk for viral resistance. The prevalence of nevirapine resistance in infants has been estimated to be 52.6% after single-dose nevirapine exposure, and 16.5% after the use of single-dose nevirapine in combination with other antiretroviral agents.[27] After extended postnatal nevirapine prophylaxis to reduce breastfeeding transmission (≥6 weeks), up to 92% of infants who were subsequently diagnosed with HIV infection were reported to have nevirapine resistance.[28,29] Extended postnatal nevirapine prophylaxis may cause delayed clearance of nevirapine-resistant mutants during the first year of life. A small study showed that more HIV-infected infants who received 6 weeks of nevirapine had nevirapine resistance at the age of 6 months than those who received single-dose nevirapine (58% vs 26%). Although this finding was not statistically significant, these results are concerning for clinical practice.[30] Because previous exposure to nevirapine clearly undermines the efficacy of subsequent NNRTI-based therapy, the role of NNRTIs in pediatric ART programmes should be reviewed in RLS.

2.4 Lopinavir/Ritonavir-Based Therapy

Despite the WHO recommendation to expand use of lopinavir/ritonavir in HIV-infected infants with prior nevirapine exposure,[3] implementation has been challenging. Currently there are no generic formulations or FDCs containing lopinavir/ritonavir, and lopinavir/ritonavir is more expensive than nevirapine. The ritonavir component gives the drug an unpleasant flavor, and liquid lopinavir/ritonavir requires a cold chain for storage. Although a pediatric solid formulation containing lopinavir 100 mg and ritonavir 25 mg has been licensed, it is only recommended in children weighing ≥10 kg.[3] These tablets are not breakable or crushable. Furthermore, these tablets are large and not easy for small children to swallow. A sprinkle formulation for younger children is in the concept phase but is unlikely to be available for several years. Long-term metabolic toxicities of PIs, including lipodystrophy, and lipid and glucose derangements have been described and are of particular concern for young children in whom exposure is likely to occur over many years.[31]

Nevertheless, lopinavir/ritonavir has a high genetic barrier to resistance requiring multiple resistance mutations in order to confer resistance to the agent.[32] Studies have demonstrated that in adults who experience virologic failure while taking boosted PIs, relatively few developed major PI resistance mutations (MRIMs).[33] Similarly, few MRIMs have been reported in children who had virologic failure while taking lopinavir/ritonavir.[34,35]

2.5 Nevirapine Switching Strategy

Several adult reports have described the consequences of switching from PI- to NNRTI-based regimens. These studies showed that switching was not associated with increased risk of virologic failure, cholesterol levels declined, adherence improved, and quality of life and body shape improved.[36–39] In 17 heavily pretreated children who were switched from a PI to efavirenz, 16 achieved viral suppression after a median follow-up period of 48 weeks, adherence improved, and triglyceride and cholesterol concentrations declined.[40]

The Nevirapine Resistance Study (NEVEREST), a randomized open-label clinical trial, addressed whether prior exposure to single-dose nevirapine would compromise the treatment outcomes of HIV-infected children who initially achieved virologic suppression on a lopinavir/ritonavir-containing first-line regimen and who were then switched to a nevirapine-containing regimen. Briefly, 195 HIV-infected children previously exposed to single-dose nevirapine and who achieved virologic suppression for at least 3 months while receiving lopinavir/ritonavir-containing ART were randomized to either continue the existing ART regimen (n = 99) or substitute lopinavir/ritonavir for nevirapine (n = 96). The median age at ART initiation was 11 months and at randomization was 20 months. Fifty-two weeks after randomization, more children in the control group (lopinavir/ritonavir-containing ART) experienced at least one viral load measurement >50 copies/mL (58% vs 44%; p = 0.02). However, fewer children in the nevirapine switch group maintained a viral load consistently below 1000 copies/mL. The probability of a confirmed viral load >1000 copies/mL was 2% in the control group versus 20% in the nevirapine switch group (p < 0.0001). The presence of NNRTI mutations detected by resistance testing in the viral population before ART was initiated was associated with confirmed viral load >1000 copies/mL in the nevirapine switch group.[41]

The results of this study suggest that provided an undetectable viral load or a viral genotype pattern indicating the absence of NNRTI resistance mutations is confirmed prior to the switch, and there is access to regular viral load monitoring, switching from a lopinavir/ritonavir- to a nevirapine-containing regimen can occur. However, viral load testing is rarely accessible and genotyping virtually unavailable in most RLS. Where viral load testing is available, this switch strategy may be considered, with pre- and regular post-switch viral load testing. This strategy may be attractive for countries considering lopinavir/ritonavir-containing first-line regimens after single-dose nevirapine exposure. The complexity of introducing a switch strategy in busy routine programmes, many of which are nurse driven, should not be underestimated. The effectiveness of this strategy after prolonged post-natal nevirapine exposure requires evaluation.

2.6 Summary

In summary, lopinavir/ritonavir-containing first-line regimens may be more advantageous in young children irrespective of whether or not they were previously exposed to NNRTIs. In older children with no prior exposure to nevirapine, NNRTI-containing and PI-containing regimens are equally efficacious. In HIV-infected children >24 months of age with prior NNRTI exposure, further research is needed to identify the preferred first-line regimens. Switching from lopinavir/ritonavir to NNRTI-based therapy in children with prior nevirapine exposure may be successful, provided careful monitoring with virologic testing is available.

3. Antiretroviral Therapy (ART) Regimens during Tuberculosis (TB) Co-Treatment

The estimated incidence of tuberculosis (TB) in HIV-infected infants is approximately 1600 per 100 000, >24-fold higher than the incidence rate among HIV-uninfected infants.[42] Co-treatment of HIV and TB among children is common. In South Africa, more than 25% of children starting ART are receiving concomitant anti-TB treatment.[43,44]

The WHO has published preferred ART regimens for children receiving rifampicin-containing anti-TB regimens (table I).[3] Rifampicin induces the metabolism of most NNRTIs and PIs by the cytochrome P450 (CYP) enzymatic system. In particular, the metabolism of nevirapine and lopinavir is markedly increased, and that of efavirenz moderately increased. By contrast, ritonavir is a potent inhibitor of the hepatic CYP3A4 and CYP3A5 enzyme isoforms, counteracting the effect of rifampicin on the CYP enzymatic system.[45] In clinical practice, several strategies have been proposed to overcome or circumvent the metabolic effects of rifampicin. The limited research that has evaluated these approaches in HIV-infected children is discussed in sections 3.1–3.4.

3.1 NNRTI-Containing ART

For HIV-infected children <3 years of age receiving rifampicin therapy, a nevirapine-containing regimen is one of the preferred options. In one study, adequate nevirapine exposure was achieved with a daily nevirapine dose range of 300–526 mg/m2 in children co-treated with rifampicin-containing therapy.[46] In another pharmacokinetic study, only three (43%) of seven children co-treated with rifampicin-containing therapy achieved an adequate nevirapine trough concentration of >3 mg/L (table IV).[47] However, the clinical, immunologic, and virologic outcomes of 26 children receiving nevirapine-containing ART plus concomitant rifampicin-containing anti-TB therapy were similar to those in children receiving the identical ART regimen without anti-TB therapy.[49] Finally, another pharmacokinetic study evaluated 21 children who had received at least 4 weeks of a nevirapine-containing FDC plus rifampicin-containing anti-TB therapy. The mean total nevirapine daily dose was 353 mg/m2. A suboptimal pre-dose concentration of <3 mg/L was documented in 57% of patients, and there was a 41% reduction in the area under the plasma concentration-time curve during anti-TB treatment (table IV).[48]

Table IV
figure Tab4

Summary of nevirapine pharmacokinetic studies in children co-treated with rifampicin-containing anti-TB therapy[46–48]

Full size image

Based on these results, WHO advises that where nevirapine-containing regimens are combined with rifampicin, a nevirapine dose of 200 mg/m2 twice daily should be used.[3] This recommendation has not been evaluated prospectively, and may be challenging to implement given that, in many RLS, nevirapine-containing FDCs are used. To ensure that the recommended nevirapine dose is administered during anti-TB treatment, additional nevirapine may be required, increasing the complexity of ART administration.

Rifampicin results in a moderate reduction in efavirenz plasma exposure of 22–26%.[50,51] A single study explored the effect of rifampicin on efavirenz pharmacokinetics in 15 HIV-infected children with a median age of 6.3 years at enrolment. Rifampicin-containing anti-TB therapy had no effect on the mean efavirenz concentration. However, 60% and 53% of the children had subtherapeutic efavirenz trough concentrations (<1 mg/L) during and after anti-TB therapy, respectively.[52] Further studies from South Africa and Burkina Faso confirmed that many children dosed according to recommended guidelines did not achieve adequate efavirenz plasma exposure.[53,54] Consequently, WHO has revised the weight-band dosing instructions for efavirenz in the latest global pediatric ART guidelines.[3] Additional efavirenz dose adjustments are not considered necessary for children receiving concomitant rifampicin.

3.2 Triple NRTI Regimens

For children who are receiving concomitant anti-TB therapy, irrespective of their age, an ART regimen consisting of three NRTIs is regarded as an acceptable alternative to an NNRTI-containing regimen (table I).[3] However, this recommendation is not supported by good quality evidence.

There are concerns regarding the potency of triple NRTI therapy. Only 8 of 15 (53.3%) ART-naïve children aged 2–12 years achieved virologic suppression after 24 weeks of treatment with a regimen comprising zidovudine, lamivudine, and abacavir.[55] Another study retrospectively reviewed the efficacy of triple NRTI regimens in 92 individuals, aged 5–23 years. Treatment with these regimens was associated with delayed virologic control and poor long-term control. Only 14% achieved >1 log10 reduction in viral load by 12 weeks and, after 48 weeks of therapy, 69% remained virologically unsuppressed. High baseline viral load, particularly >100 000 copies/mL, was associated with an increased odds of having an unsuppressed viral load at 48 weeks.[56] In another study of 20 children aged 2–18 years who initially achieved virologic suppression on a PI-containing regimen and were then switched to a triple NRTI regimen, 19 (95%) maintained a plasma viral load <50 copies/mL during a 96-week follow-up period, suggesting that virologically suppressed children are more likely to maintain virologic suppression when switched to an NRTI-only regimen.[57] Children in the studies mentioned in this paragraph did not receive anti-TB therapy.

There are no studies evaluating the performance of triple NRTI regimens in children co-treated with rifampicin-containing anti-TB therapy, and no data on the performance of these regimens in children <2 years of age when baseline viral loads are frequently markedly elevated.

3.3 Overcoming the Effect of Rifampicin on Lopinavir Metabolism

In children who develop TB while receiving a lopinavir/ritonavir-containing regimen, WHO advises that additional ritonavir be administered to achieve a 1 : 1 milligram ratio of lopinavir : ritonavir, during anti-TB therapy.[3] In adults co-treated with rifampicin and lopinavir/ritonavir-containing regimens, the metabolic impact of rifampicin may be overcome by augmenting the ART regimen with additional ritonavir or by doubling the dose of lopinavir/ritonavir.[58] These strategies have undergone limited evaluation in children.

A pharmacokinetic study in HIV-infected children with TB (aged 7 months to 3.9 years) showed that by providing additional ritonavir to children receiving lopinavir/ritonavir to achieve a lopinavir : ritonavir milligram ratio of 1 : 1, the effect of rifampicin on lopinavir metabolism may be attenuated. A minimum lopinavir concentration of >1 mg/mL was maintained in 13 of 15 children using this strategy, suggesting that ritonavir boosting is an important option for children receiving lopinavir/ritonavir and concomitant rifampicin.[59] This strategy did not completely compensate for the enhancement of lopinavir oral clearance.[60] Whether this approach is effective in infants <6 months of age is currently not known. Since TB may complicate the course of HIV-infection during early infancy, this is an important question to resolve.

Another pharmacokinetic study showed that double doses of lopinavir/ritonavir did not achieve adequate lopinavir exposure in young children, aged 0.98–1.93 years co-treated with rifampicin.[61] Retrospective assessment of children co-treated with rifampicin-containing anti-TB therapy and three different PI-modified strategies showed that those who received lopinavir/ritonavir boosted with additional ritonavir had better virologic outcomes than either children who received double-dose lopinavir/ritonavir-containing therapy or those on ritonavir-containing ART. Furthermore, the outcomes of children treated with lopinavir/ritonavir plus additional ritonavir were similar to those of a group of children taking lopinavir/ritonavir-containing ART but no anti-TB therapy.[62]

Ritonavir suspension required for boosting is not widely available in RLS and a heat-stable solid formulation suitable for pediatric practice does not currently exist. WHO has called for the development of heat-stable formulations suitable for boosting lopinavir/ritonavir in young children.[3]

3.4 Rifabutin-Containing Anti-TB Therapy

Rifabutin, an alternative to rifampicin, is a less potent inducer of the CYP system, which exerts minimal effects on antiretroviral metabolism. A recent Cochrane review concluded that there was insufficient evidence of its efficacy and safety in HIV-infected patients.[63] Research in children has been constrained by the absence of a suitable formulation. Consequently there are currently no dosing guidelines, or safety and efficacy data for rifabutin in HIV-infected children co-treated with ART.

3.5 Summary

Current global recommendations for treating children and adolescents with ART and rifampicin-containing anti-TB therapy are based on limited evidence. In particular, there is no prospective evidence confirming the efficacy of nevirapine at a dosage of 200 mg/m2 twice daily or triple NRTI combination therapy during rifampicin-containing anti-TB therapy. Current evidence supports the use of lopinavir/ritonavir-containing regimens boosted with additional ritonavir in children >6 months of age and efavirenz-containing regimens in children >3 years of age. Well designed research is urgently needed to optimize strategies for overcoming the effect of rifampicin on existing and newer antiretroviral agents in children of all ages.

4. Second-Line Recommendations

Although many ART programmes in RLS are in their infancy, increasing numbers of children in these programmes have already been initiated on second-line therapy. A cross-sectional survey of 17 sites in six Asian countries completed in April 2008 showed that 20% of 3606 children receiving ART had progressed beyond first-line therapy.[64] A more recent survey that analysed data from 10 Asian and 16 southern African sites showed that 10% of 1301 Asian children and 3.3% of 4561 southern African children had initiated second-line therapy.[65]

Switching to second-line therapy in RLS has been constrained by unclear criteria for diagnosing treatment failure, and the inaccuracy of clinical and immunological criteria for identifying children with treatment failure. In a provincial programme in South Africa that assessed the outcomes of 1741 children started on ART, the overall cumulative percentage of patients with an unsuppressed viral load at 18 months was 23.1%, yet only 2.73% had progressed to second-line therapy.[66] In the South African national ART programme, among 252 children with more than 1 year of follow-up after not responding to first-line therapy, only 38% had been switched to second-line therapy. In that study a PI-containing first-line regimen was negatively associated with switching to second-line therapy because many of these children had had prior NNRTI exposure and their attending medical practitioners were reluctant to switch them to an NNRTI-containing second-line regimen.[67,68] Further research on the role of monitoring CD4 counts and viral load, and genotyping to enable switching to second-line therapy is urgently required.

Preferred second-line regimens recommended by WHO, and listed in table I, are generally predictable and uncontroversial.[3] The major exception is the recommendation to administer a NNRTI-containing second-line regimen to HIV-infected children who were previously exposed to NNRTIs and who had treatment-failure with lopinavir/ritonavir-containing first-line therapy (see section 4.2).

4.1 Children Who Had Treatment Failure with NNRTI-Based Therapy

In general, for children with treatment failure with an NNRTI-based first-line regimen, switching to a boosted PI-based regimen is a good and potent option (table I). In the absence of virologic testing, treatment failure is usually diagnosed late, leading to accumulation of viral mutations, which may undermine second-line therapy.

The PENPACT-1 study compared ART outcomes in children who were treated with either PI- or NNRTI-containing regimens and had treatment failure. The effect of switching to second-line therapy at viral load concentrations of either >1000 copies/mL or >30 000 copies/mL was evaluated. PI resistance, at the time of diagnosing treatment failure, was rare, whereas NNRTI resistance was selected for early. There was no increase in NRTI resistance in the PI group who switched at a higher versus lower viral load threshold. By contrast, about 10% more children accumulated NRTI mutations in the NNRTI group who were switched at a viral load of >30 000 copies/mL than in those switched at a viral load of >1000 copies/mL.[19] These results suggest that regular virologic monitoring of children taking NNRTI-based regimens may ensure that early switching to second-line therapy occurs, lowering the risk of accumulating NRTI mutations.

4.2 Children with Treatment Failure with PI-Based Therapy

For children with treatment failure with a boosted PI regimen, the efficacy of the WHO second-line NNRTI plus two NRTIs option (see table I) remains unclear, especially in children who received extended postnatal nevirapine prophylaxis.[3] Boosted PIs such as lopinavir/ritonavir have a high genetic barrier to resistance, corroborated by results from the PENPACT-1 study.[19] Where virologic testing is routinely available, detection of viral escape in these children should trigger intensified adherence counseling. The recommendation to switch to an NNRTI-based regimen after PI-failure, if there was prior nevirapine exposure, is concerning since archived resistance mutations may affect outcomes and the first-generation NNRTIs have a low genetic barrier to developing resistance. If treatment failure to a lopinavir/ritonavir-containing regimen occurs, genotyping, if available, may be helpful to formulate second-line regimens, which could include newer antiretroviral agents.

Currently, genotyping and newer antiretroviral agents are not widely available in RLS. In these settings, in keeping with WHO recommendations, an NNRTI-containing second-line regimen may be employed. Because poor adherence is the more likely reason for apparent treatment failure in the presence of PI-based first-line therapy, maintaining the PI regimen while trying to improve adherence may be a safer initial step. However, research is urgently needed to define optimal and affordable second-line options for children with treatment failure with PI-based therapy after perinatal nevirapine exposure; current second-line ART recommendations are not supported by strong evidence.

5. Newer Antiretroviral Agents and Treatment Strategies

5.1 Newer Agents

Several newer antiretroviral agents have either entered clinical practice or are under investigation. Although there is a time lag between adult and pediatric trials, some new agents have been approved for pediatric use and several are currently undergoing dose-finding investigation in children.

5.1.1 Second-Generation NNRTIs

Etravirine has entered phase II trials in treatment-experienced children. It is active against HIV strains resistant to efavirenz, nevirapine, and delavirdine, having a higher genetic barrier to resistance. Etravirine plus an optimized background regimen (OBR) provided better viral suppression and improved immunologic response than the OBR alone.[69] Small percentages of children and adults in sub-Saharan Africa may harbor mutations to etravirine; these mutations accumulate during prolonged exposure to a failing regimen containing nevirapine or efavirenz.[70,71] Because etravirine has only thus far been studied with PIs, its performance in the absence of PI co-treatment remains to be determined.

Rilpivirine (TMC-278), with a similarly high genetic barrier to resistance, is also entering pediatric phase II trials.

5.1.2 PIs

Darunavir and tipranavir are two new PIs licensed for use in treatment-experienced children. Darunavir, a non-peptide PI, binds rapidly to protease at a unique site and disassociates slowly causing a 2-fold higher binding strength compared with other PIs. It is potent even against PI-resistant viral strains, and is recommended in salvage regimens, particularly in the presence of major resistance mutations.[72,73] Darunavir has demonstrated equivalence with lopinavir/ritonavir in drug naïve adults and has a favorable toxicity profile.[74] Results from 156 children who underwent resistance testing while receiving PI-containing ART showed a relatively low frequency of darunavir resistance mutations; 13% had one darunavir mutation, 3% had two mutations, and 2% had three mutations. Of 55 children who had taken lopinavir/ritonavir as the only PI, one had low-level darunavir resistance.[75] Although darunavir should be administered with low-dose ritonavir, co-formulated products do not exist. Darunavir is licensed for use from the age of 6 years.[76] Recent research has established dosing guidelines for children 3–6 years of age.[77] The bioavailability of darunavir following the administration of an oral solution to adult volunteers was shown to be comparable to that of the 300 mg tablet, and is under study in children.[78]

Tipranavir is licensed for use in children >2 years of age. Treatment failure with darunavir preserves the activity of tipranavir and vice versa, although darunavir has a lower pill burden and is associated with less serious adverse events.[69] Tipranavir-containing regimens were evaluated in treatment-experienced children aged 2–18 years. After 48 weeks, 39.7% administered low-dose tipranavir and 45.6% who received high-dose tipranavir had a viral load <400 copies/mL.[79] A liquid formulation has been produced for tipranavir.[80]

5.1.3 Integrase Inhibitors

Among this new antiretroviral class, raltegravir has been licensed in adults. Although not licensed for pediatric practice, studies in children are underway to optimize dosing, and evaluate the safety and efficacy of raltegravir in treatment-experienced children and adolescents.[81] Results from a phase I/II trial of 10 antiretroviral-experienced but integrase-näive children aged 6–11 years with a bodyweight ≥25 kg showed that raltegravir dosed at 400 mg twice daily in association with an OBR was well tolerated and that 56% of the children achieved virologic success after 24 weeks.[82]

A recent study showed that in 12 children aged 2–5 years a raltegravir chewable tablet administered at a dosage of 6 mg/kg twice daily achieved adequate drug exposure.[83] Co-administration of rifampicin in adults has been shown to cause a 40–60% decrease in raltegravir concentration, complicating concomitant treatment of TB.[84] This could be problematic should raltegravir become accessible to HIV-infected children in RLS, although this requires investigation.

A newer agent, elvitegravir, is also entering pediatric trials.

5.1.4 C-C Chemokine Receptor Type 5 Antagonists

This new antiretroviral class requires C-C chemokine receptor type 5 (CCR5) tropic HIV-1 virus to be active. Maraviroc has been licensed for use in adults. The role of CCR5 inhibitors is yet to be established in pediatric care, although maraviroc has shown promise when used in first-line regimens; CCR5-tropic HIV-1 strains are predominantly involved in transmission of the virus, and found in early infection.[85]

5.2 Potential Application of Newer Agents

Newer antiretroviral agents could fulfill several roles in future pediatric ART practice:

  • Optimal second-line regimens are required for nevirapine-exposed children who have treatment failure with lopinavir/ritonavir-containing first-line therapy. Regimens containing newer PIs and/or integrase inhibitors should be evaluated. This is a priority research area.

  • Integrase inhibitor-containing first-line regimens could be used in NNRTI-exposed children as a PI-sparing strategy during the initiation of ART. Alternatively, once virologic suppression with a PI-containing regimen is attained, an integrase inhibitor such as raltegravir could be substituted for the first-line PI. Raltegravir-containing first-line regimens are effective in ART-naïve adults.[86]

  • With the introduction of the integrase inhibitors it may be possible to sequence first- and second-line regimens with non-overlapping resistance patterns, thus reducing the need for viral load monitoring and viral resistance testing. This approach would be extremely attractive for RLS, with minimal access to these monitoring measures.

  • Optimal dosing of first-line regimens in early infancy, especially in premature infants, requires investigation. In a recent report, severe toxicity, including cardiac toxicity in a group of HIV-infected preterm neonates, was associated with the administration of lopinavir/ritonavir oral solution. Major constituents of lopinavir/ritonavir solution, namely lopinavir, ethanol, and propylene glycol have been implicated.[87] Alternative first-line regimens containing newer antiretroviral agents should be urgently evaluated for safety and efficacy in this patient group.

  • A challenge for RLS is the high frequency of TB co-infection. Optimal strategies for combining rifampicin with integrase inhibitors or newer generation PIs should be explored in co-infected children.

  • Although WHO guidelines focus on first- and second-line therapy, third-line regimens may be considered in older children and adolescents, depending on the availability of newer agents, including newer generation NNRTIs and PIs, CCR5 antagonists, and integrase inhibitors.[3]

6. Conclusions

This review has concentrated on first- and second-line regimens for children in RLS. As more light is shed on the management of children receiving ART, careful consideration should be given to the choices of treatment regimens that may improve the durability of first-line and subsequent therapy. Several aspects of pediatric ART regimens have been under-researched, including optimal strategies for administering concomitant ART and anti-TB medication. The arrival of the newer antiretroviral agents accompanied by another intense period of research should hopefully lead to further refinement of existing regimen options.