First-line antiretroviral therapy for HIV-infected children

Current WHO guidelines advise the use lopinavir/ritonavir as first-line treatment for HIV in children under 3 years of age but acknowledge that there are potential barriers to implementation, such as the need for a cold chain to point of dispensing, its unpalatability, cost and complexity of administration with tuberculosis treatment [1]. Ciaranello et al.[2] have undertaken a thorough modelling analysis comparing the cost-effectiveness of the ordering of antiretroviral therapy (ART) regimens for children less than 3 years of age using either lopinavir/ritonavir as first-line ART followed by nevirapine or vice versa as well as compared with no ART in South Africa and Cote d’Ivoire. Their analysis shows tremendous benefits of either ART order as compared with no ART. In South Africa, providing either ART regimen is shown to be cost-saving for the first 12 years, and highly cost-effective thereafter; although not cost-saving in Cote d’Ivoire, it is highly cost-effective. Extensive univariate and multivariate sensitivity analyses explore the impact of these potential implementation barriers. Firstly, they show that in most cases, lopinavir/ritonavir as first-line therapy is more cost-effective than nevirapine, albeit by relatively small margins. Secondly, even without viral load monitoring, it is still more cost-effective to provide a second-line therapy than only providing a first-line regimen, reiterating the importance of ensuring second-line therapy is available to children. Thirdly, in recognition that there are numerous delays in testing infants exposed to ART and initiating HIV-positive children on ART, the authors undertake an analysis of the cost-effectiveness of both regimens as initiated at 6, 12, 24 and 35 months and show that it becomes less expensive per life year saved when initiated at older ages. The challenge in the interpretation of these results is that lives lost by delayed initiation are not captured in their comparisons: life expectancy at older ages is longer because weaker children have died out prior to accessing ART. It is estimated about 20% of HIV-infected infants will die before the age of 3 months and over half before the age of 2 years without treatment [3–5]. Ciaranello's Figure 1 further illustrates this: delayed initiation from the age of 12 to 36 months would need to only initiate about 30% of the original cohort, as the other nearly 70% would not have survived the 2-year interval. Although their analysis does not set out to compare ART initiation at different ages, misinterpretation of these results risks providing a justification for low prioritization of strengthening systems for early infant diagnosis (EID) and early treatment initiation: the incremental cost-effectiveness of treatment is after all better when initiated later. In contrast (and consistent with their earlier work [6]), the important policy message from this analysis is how critical either of the treatment regimens is for saving lives and, in many cases, treatment saves money. The analysis by Ciaranello et al. [6] emphasizes that health systems must be strengthened to reduce the numerous barriers to EID and treatment. EID currently requires laboratory testing to confirm infection, and thus necessitates transportation of samples to central laboratories and return of results, both identified as high loss points in the EID testing cascade [7]. The coverage and practical implementation of EID therefore varies widely across resource-poor settings. For example, in rural Tanzania, EID coverage (health facilities providing EID services) in three regions ranged from 12 to 63% with average turn-around-times (TATs) from test to results being returned to the caregiver ranging from 19 to 49 days and does not have a systematic approach across regions for transporting samples and conveying results to health facilities [8]. Kenya has achieved a high coverage of 70% by transporting dried blot spot samples in weekly batches by courier to central laboratories for testing and SMS printers for remitting results in real time to rural and hard to reach clinics, with an average TAT of 25 days [9]. Returning results to clinics on time to feedback the results to families at routine immunization visits can mean the difference between HIV-infected infants initiating treatment or being lost to follow up. In recognition of these critical bottlenecks, there has been recent investment in new technologies by global funders such as PEPFAR and UNITAID. SMS printers reduce loss of paper-based test results and reduce TAT to health facilities, particularly in more remote locations [10,11]. This technology is rapidly being introduced and scaled up with pilots in at least 10 African countries (2010) [12]. Point of care (POC) diagnostics can provide decentralized EID with results in the same visit by most health workers and remove the need for transporting samples and results [7]. As yet, only one POC diagnostic has completed field testing in a low-resources setting, showing that EID can be feasible at the primary care level [13]. However, more diagnostic tools are in the pipeline for distribution in the near future [14]. To expedite introduction of such new POC diagnostics, UNITAID has been active in the development a toolkit to support policy development and adoption of these new technologies [15], and to increase access by leveraging price reductions for diagnostics [16]. In addition, they are supporting the development of new drug formulations that will be more palatable and not require a cold chain [17]. It would be a great contribution if the model by Ciaranello et al.[6] provides evidence on the most cost-effective uses of these new technologies. In sum, Ciaranello et al. [6] have shown how important it is to provide ART for infants infected with HIV. The lopinavir/ritonavir regimen provides benefits in terms of health outcomes as well as lower costs. However, LPV is not ideal. New formulations have the potential to increase access to treatment, improve retention in care and ultimately infant outcomes. However, ART formulation is not the full story. Although there remain numerous barriers to accessing diagnosis and treatment, such technologies may contribute to alleviating some of these barriers. It is recommended that in future such analyses do not ignore mortality attributable to delayed initiation, to ensure the full picture of relative cost-effectiveness and comparator is presented. I look forward to seeing the CEPAC model applied to better understand the cost-effectiveness of new POC technologies such as EID. Acknowledgements Conflicts of interest Fern Terris-Prestholt has received research funding from UNITAID.