Background

Since they were first published in 2002, WHO guidelines on ART have evolved as the body of evidence to support the earlier initiation of ART has progressively grown (1). The 2013 WHO consolidated ARV guidelines recommended initiating ART for all adults with HIV and a CD4 count at or below 500 cells/mm³, regardless of WHO clinical stage, giving priority to those with severe or advanced HIV disease (WHO clinical stages 3 or 4) or a CD4 cell count at or below 350 cells/mm³ (2). This strong recommendation was based on moderate-quality evidence from three randomized controlled trials (3–5) and 21 observational studies (6–27) showing that initiating ART at or below a CD4 threshold of 500 cells/mm³ compared with later initiation reduced the risk of progression to AIDS and/or death, TB and developing a non-AIDS-defining illness and increased the likelihood of immune recovery. In addition, high-quality evidence from one randomized controlled trial indicated that earlier ART can markedly reduce the risk of sexual transmission to HIV-negative sexual partners among heterosexual couples (4).

Mathematical models and ecological studies also suggested that initiating ART earlier could affect HIV incidence at the population level if there is high uptake of sustained testing, ART coverage and retention (28–32). For people with certain clinical conditions such as TB, hepatitis B (HBV) coinfection requiring HBV treatment, pregnancy, breastfeeding and HIV serodiscordant couples, the 2013 guidelines recommended initiating ART regardless of WHO clinical stage or at any CD4 cell count.

Global ART coverage for all individuals living with HIV had reached approximately 41% – or 15 million people – by March 2015 (33). According to the WHO Country Intelligence Database, by June 2014, more than half (60%) of the 58 WHO HIV focus countries had adopted the CD4 threshold of 500 cells/mm³ or less for initiating ART, and 7% (n=4) had already moved the CD4 threshold to above 500 cells/mm³ (Fig. 1) (34). Although the median CD4 count at the time of ART initiation is increasing, it remains significantly lower than 350 cells/mm³ in almost all settings, including high-income countries (35,36), and late presentation for treatment is associated with high early mortality rates, higher direct health-care costs and poor retention in care (37–39). Increasing knowledge of HIV status, strengthening links between testing and care, modifying health systems to manage patient volumes and ensuring optimal long-term retention and adherence remain significant challenges in many settings (40).

Fig. 1

Uptake of WHO policy on the threshold for initiating ART among adults and adolescents living with HIV in low- and middle-income countries, 2014.

Rationale and supporting evidence

Initiating ART among all adults living with HIV regardless of WHO clinical stage or at any CD4 cell count

Since 2013, evidence and programmatic experience have continued to favour earlier initiation of ART because of reduced mortality, morbidity and HIV transmission outcomes. Increasing evidence from systematic reviews and cohort analyses also indicates that untreated HIV infection may be associated with the development of several non-AIDS-defining conditions (including cardiovascular disease, kidney disease, liver disease, several types of cancer and neurocognitive disorders) (41–44) and that initiating ART earlier reduces such events and improves survival. Recent evidence from a large randomized clinical trial also shows that, as demonstrated for heterosexual serodiscordant couples, ART substantially reduces sexual transmission to HIV-negative sexual partners among homosexual couples (45).

The recommendation to initiate ART at any CD4 cell count was based on a systematic review with GRADE evidence profiles that assessed the quality and strength of the evidence from one randomized controlled trial (46) and 17 observational studies (11,12,15–19,23,27,30,47–52) reporting clinical, immunological and virological outcomes and HIV transmission.

In the analysis of data from the single randomized controlled trial (TEMPRANO), moderate-quality evidence (downgraded from high quality because of imprecision) showed that initiating ART at a CD4 count >500 cells/mm³, in the absence of other treatment indications, leads to less severe HIV morbidity (combined outcome of death, AIDS and severe non-AIDS diseases such as malignancies and bacterial diseases) compared with treatment initiation at a CD4 count at or below 500 cells/mm³ (hazard ratio = 0.56, 95% confidence interval (CI) 0.33–0.94)(46).

Limited data from another randomized clinical trial (INSIGHT-START study) unpublished at the time of the Clinical Guideline Development Group meeting were available but not incorporated into the systematic review or GRADE table (53). The Data and Safety Monitoring Board advised immediate dissemination of the findings from the START study because of predefined stopping rules. The START trial was not part of the systematic review because the comparison groups did not match the review PICO and was therefore not considered in relation to the quality of the evidence. Box 1 summarizes the key findings from the START study that were presented to the Clinical Guideline Development Group and are supportive of the new recommendation.

Box 1

The START study. The interim results of the Strategic Timing of Antiretroviral Treatment (START) trial were reviewed as a complementary assessment of the evidence. This study enrolled 4685 people at 215 sites in 35 countries. Twenty-seven percent of the (more...)

Analysis of the observational studies found a significantly lower risk of HIV disease progression (50), and the TEMPRANO randomized controlled trial data demonstrated from modelling the potential lower rates of HIV transmission to uninfected partners (46), but the evidence was rated as very low quality in both cases. However, interim data from the HPTN 052 clinical trial indicated that early ART is highly effective at prevention of sexual transmission of HIV (54). Similar to the START trial, relevant data from the HPTN 052 were unpublished at the time of the guidelines review meeting and were not incorporated into the systematic review due to a different comparator to that of the review. Box 2 summarizes these data.

Box 2

New data from HPTN 052. The HIV Prevention Trials Network (HPTN) 052 study showed that starting ART early reduced the overall risk of HIV sexual transmission to uninfected partners by 93% (54). In this Phase 3 randomized controlled trial, 1763 HIV serodiscordant (more...)

Moderate-quality evidence from the TEMPRANO trial showed that initiating ART at a CD4 count above 500 cells/mm³ was not associated with increased risk of grade 3 or 4 adverse events¹ (46). Low-quality evidence from observational data showed an increased risk of any severe laboratory adverse event and hepatic adverse events in individuals initiating ART at CD4 count above 500 cells/mm³, although this was not associated with treatment discontinuation (48).

Programmatic data from several countries that are offering earlier ART either to all people living with HIV or to specific populations have shown significant increases in ART uptake and linkage to care, reduction in the time between HIV diagnosis and ART initiation regardless of baseline CD4 cell count and an increase in the median CD4 value at ART initiation. Retention in care has not differed between individuals who start at CD4 counts above 500 cells/mm³ compared with those who started based on the standard of care (55–57).

Comparing benefits and harm

The benefits of earlier ART initiation include fewer events of severe HIV morbidity and disease progression, improved uptake and initial linkage to care, better immune recovery and decreased HIV transmission. However, not all observational studies have consistently demonstrated the beneficial effect of initiating ART at CD4 cell count at or above 500 cells/mm³ on mortality, the incidence of inflammation-related non-AIDS events and ongoing viral replication compared with initiation at CD4 at or below 500 cells/mm³. Follow-up will be needed to evaluate the potential harm and benefits of ART over a lifetime.

It is increasingly recognized that, in settings with a high burden of HIV and TB infections, increasing ART coverage is associated with decreasing TB case notifications, and this is likely to improve when ART is started earlier.

A modelling exercise based on national cohort data from four countries in sub-Saharan Africa concluded that programmatic gains and mortality reduction were accrued by eliminating the pre-ART period, suggesting that making treatment available to everyone will strengthen the continuum of care (62).

Cost and cost–effectiveness

The same modelling exercise indicates that expanding ART eligibility criteria to above 500 cells/mm³ or regardless of CD4 cell count and linking to HIV care could result in 6–14% fewer people dying from HIV-related causes during the next decade (62). In this exercise, the vast majority of the impact is caused by programmatic simplification leading to more people initiating ART in a timely manner and therefore avoiding adverse outcomes during the per-ART period rather any direct therapeutic benefits. The increased cost of earlier ART would be partly offset by subsequent reduced costs (such as decreased hospitalization and increased productivity) and preventing people from acquiring HIV infection. The modelling results suggest that such a change is likely to be cost-effective in many settings if people initiating ART adhere to treatment and retention in care is maintained. Costs will increase significantly but far less than if the additional outreach interventions required to maintain individuals in pre-ART care are also included.

According to UNAIDS estimates, expanding ART to all people living with HIV is projected to avert 21 million AIDS-related deaths and 28 million new infections by 2030 (63). However, these benefits require high testing uptake, high treatment coverage, sustained adherence and high rates of retention in care. The cost implications at the regional and country levels can also vary and should be further explored, since countries have different levels of current treatment coverage and local cost considerations depending on their context and resources.

Equity and acceptability

Concerns have been expressed that, given limited resources, very early treatment could result in some people who urgently need treatment being displaced by people for whom treatment would be beneficial but is less urgently needed. For this reason, initiating ART among adults with severe or advanced HIV disease or with a CD4 count at or below 350 cells/mm³ is recommended as a priority in this guideline. Additional concerns that mandatory or coercive approaches will be used among high-risk marginalized populations highlight the importance of adequate patient information, informed consent, appropriate health worker training and rights-based legal frameworks to facilitate access.

The community-led global consultation examining the acceptability of earlier initiation of ART at a higher CD4 count for people living with HIV, caregivers and service providers found that earlier initiation was considered acceptable. Participants in the consultation emphasized the need for collaborative decision-making with service providers to ensure that the ultimate decision to initiate ART rests with the person living with HIV. Motivation to start and adhere to treatment may be more difficult for people who feel well and have higher CD4 counts than for people who are or have been ill. Stigma and discrimination continue to act as barriers to treatment access and adherence. Critical factors in promoting ongoing engagement in care and adherence include ensuring access to a stable supply of free or affordable ARV drugs, facilities that are easily accessible and that ensure confidentiality, sympathetic providers and community adherence support.

A qualitative literature review showed that acceptability for earlier treatment is greater when people know that treatment reduces mortality risk. Service providers recognize the preventive benefits of earlier ART and the need for earlier ART initiation for asymptomatic people. Among people living with HIV, acceptance increases when they also have comorbidities or conditions associated with a higher risk of HIV transmission. Issues cited in the literature supported those identified in community consultations.

Feasibility

According to cohort and national programme data, the number of people needing treatment could increase by up to 35% if ART is initiated at any CD4 count rather than at or below 500 cells/mm³ (63). Modelling estimates predict that this increase would be lower, in the range of 7% to 21% over five years, because not all people living with HIV are diagnosed and present for care immediately. Country experience has also shown that moving to a higher CD4 threshold for ART initiation may not necessarily lead to a significant immediate increase in the numbers of people who actually access treatment in the absence of increased uptake of HIV testing, stronger linkage to care, adequate treatment monitoring and sustained adherence support. Late presentation for treatment is still quite common, with the median CD4 count at ART initiation below 350 cells/mm³ in the majority of settings, including high-income countries (59,60).

Implementation considerations

Regardless of the epidemic profile and disease burden, priority should be given to people with symptomatic HIV disease or with CD4 count at or below 350 cells/mm³ who are at high risk of mortality and most likely to benefit from ART in the short term.

Initiating ART at CD4 counts above 500 cells/mm³ may involve additional human, infrastructure and financial resources. Countries vary in health system capacity and are at different stages in ART coverage and quality of care. A phased approach to implementation may be needed, especially in settings with a high burden of HIV infection, lower ART coverage, less developed health systems, lower rates of testing, poor pre-ART care, weaker human resources, limited laboratory capacity, budget constraints and/or competing health priorities. In such settings, equity considerations and giving priority to those who most need treatment should guide implementation. The complete consolidated guidelines will provide further guidance in this area.

The increased need for ART associated with early initiation may place demands on the health system in some settings that could increase the risk of drug resistance, such as drug stock-outs, insufficient patient preparation and suboptimal adherence. To maximize the long-term effectiveness of first-line ART regimens and ensure that people are taking the most effective regimen, the scaling up of HIV treatment should be accompanied by measures to monitor and improve service quality at the site and programme levels.

In all settings, continued monitoring of the long-term safety profile of ARV drugs and the implications of earlier initiation for drug resistance, toxicity and adherence will be needed. It also remains essential to address structural and social barriers to accessing treatment faced by key populations, such as criminalization, discrimination and stigma (64–66).

Research gaps

Several ongoing implementation trials are evaluating the feasibility, acceptability, cost–effectiveness and impact of immediate treatment for all people living with HIV regardless of CD4 cell count at the population level (SEARCH and MaxART studies). Primary outcome results are not expected before 2017 or 2018 (67,68). Three large randomized trials are examining the population effect of early ART initiation on HIV incidence and mortality (Botswana Combination Prevention Trial and HPTN-071 (PopART) study and 12249 ANRS TasP trial), with results expected in 2018 or 2019 (69–71).

Other research priorities include assessing the incidence of short- and long-term severe adverse events as a result of increased exposure to ART, barriers to and enablers of adherence and long-term retention in care and the impact on the cascade of care and the magnitude of the prevention benefit of early initiation of ART, especially among key populations and adolescents.

Background

Programmes to prevent mother-to-child transmission were some of the earliest public health interventions that used ARV drugs to reduce the risk of HIV transmission. Initially, the regimen recommended by WHO was a single dose of nevirapine given to the mother during labour and to the infant in the first few days of life. With the launch of the “3 by 5” initiative and the rolling out of national HIV care and treatment programmes, the guidelines made an important shift, recommending that pregnant women should be tested for HIV and then assessed for treatment eligibility. Those considered eligible for treatment should be offered combination lifelong antiretroviral therapy for their own health, while those who were not eligible should receive short courses of antiretroviral prophylaxis for PMTCT. Although eligibility criteria have changed and the preferred regimens for ART and for PMTCT prophylaxis have evolved, this distinction between treatment and prophylaxis became a fixture of PMTCT programmes.

In 2013, the revised and consolidated ARV guidelines recommended that all pregnant and breastfeeding women be initiated on ART regardless of clinical eligibility. This recommendation was driven by practical experience from programmes showing that PMTCT prophylaxis (using different drugs at different times in the course of pregnancy, labour and delivery as well as long duration of infant prophylaxis while breastfeeding) was more challenging to implement in the field than giving ART to all pregnant women (especially if the ART regimen was a once-daily fixed-dose combination tablet). However, the guidelines still offered programmes the option to either continue ART lifelong in all women (option B+) or to stop ART after the period of mother-to-child transmission risk in women who did not otherwise meet eligibility criteria (option B). Option B+ was felt to be of the greatest benefit in settings with high HIV prevalence and high fertility in which initiating ART among all pregnant and breastfeeding women would reduce HIV incidence and prevent HIV transmission in both current and future pregnancies.

Following the release of the 2013 guidelines, many countries moved to adopt option B+ as the preferred approach for PMTCT programmes. The most recent Global AIDS Response Progress Reporting data show that the majority of countries, including almost all the 22 high-priority countries listed in the Global Plan towards the elimination of new HIV infections among children by 2015 and keeping their mothers alive (72), are now either piloting or implementing lifelong ART for all pregnant and breastfeeding women living with HIV at a national scale (Fig. 2).

Fig. 2

Adoption of policy on preventing the mother-to-child transmission of HIV in low- and middle-income countries, 2014.

Increasing evidence to support earlier ART initiation among all adults as described in the previous section, as well as widespread uptake of option B+ and emerging programme data on the success of option B+ in practice, all support a revised recommendation in 2015 that all pregnant and breastfeeding women living with HIV should initiate ART and remain on lifelong treatment regardless of clinical or CD4 stage of disease. As a result, option B is no longer relevant.

Providing ART to all pregnant and breastfeeding women living with HIV serves three synergistic purposes: (1) improving individual health outcomes, (2) preventing mother-to-child transmission of HIV and (3) preventing the horizontal transmission of HIV from the mother to an uninfected sexual partner.

Rationale and supporting evidence

The evidence on options B and B+ and the clinical and immunological impact of stopping ART among postpartum women were reviewed in the context of increasing data showing the benefit of ART at all stages of HIV disease and the strong, new recommendation to initiate ART among all adults with HIV at any CD4 count.

A systematic review was conducted to compare option B and option B+ in terms of maternal health outcomes. The review did not identify any randomized controlled trials or observational studies that directly compared options B and B+; however, 18 studies reported on option B outcomes – comprising four randomized controlled trials (73–76), three single-arm trials (77–79) and 11 cohort studies (80–90) – and 10 cohort studies reported on outcomes associated with option B+ (91–100). All the studies evaluating option B+ suggested that women experienced health benefits in terms of immunological and clinical parameters. None of the studies included in the review reported on how option B+ affected HIV transmission rates to partners, although this is an important likely benefit for women who remain on lifelong ART.

Since the key difference between option B and option B+ is not when ART is started but whether it is stopped, literature on the clinical and immunological impact of stopping ART among women during the postpartum period was also reviewed. Five cohort studies and one randomized controlled trial were identified that examined changes in clinical and immunological parameters following discontinuation of ARV drugs. Most of these studies used the historical threshold for ART initiation of CD4 count below 350 cells/mm³, and in several cases mothers were receiving ARV prophylaxis rather than ART, but all showed a gradual decline in immune function after ARV drugs were stopped. Using the time frame of six months after discontinuation, 6–20% of women with a baseline CD4 count below 500 cells/mm³ had become eligible, whereas only 1.5% of women became eligible if the baseline CD4 was above 500 cells/mm³ (101,102). A lower baseline CD4 count was also associated with a 2.5-fold higher risk of WHO stage 2 or 3 clinical events (103).

Apart from the impact on clinical and immunological outcome, there are also programmatic consequences of stopping ARV drugs among postpartum women. In one study from Malawi, loss to follow-up post-delivery was much higher in women with baseline CD4 above 350 cells/mm³ (who were not eligible for treatment) than those with CD4 below 350 cells/mm³ (who were eligible and started on ART) (104). The findings were similar in a South African cohort where the women who were considered ineligible for ART were twice as likely to be lost to follow-up at six months postpartum as the women who had started treatment (105).

The body of evidence demonstrates the advantages of lifelong ART for pregnant and breastfeeding women and adds to the compelling data from recent randomized clinical trials suggesting that all adults with HIV benefit from ART at any CD4 and regardless of their clinical stage of disease.

Comparing benefits and harm

Most countries are moving to adopt universal ART for all pregnant and breastfeeding women. The benefits include improved health outcomes, lower mortality and the potential for reduction of horizontal transmission of HIV. Women may be less likely to drop out of care after the end of the transmission risk period and may avoid some of the clinical and operational complexity of repeated cycles of stopping and starting ART in subsequent pregnancies.

The risks to offering lifelong ART to all pregnant and breastfeeding women living with HIV as opposed to ART only during the period of mother-to-child transmission risk include the potential for cumulative drug toxicity and the possibility of poor adherence with long-term use, potentially leading to the development of drug resistance. In general, these risks for pregnant and breastfeeding women are similar to those for non-pregnant adults. A further potential risk to the fetus may arise from exposure to ARV drugs, especially when given early in the gestational period. Early gestation exposure is more likely with the change in guidelines for adults, as a greater number of women of childbearing age will initiate ART across all CD4 counts. To date, no evidence suggests a significant increased risk of congenital anomalies associated with the currently recommended first-line ARV drug regimens (106). How the recommendations are implemented may also affect outcomes.

The costs associated with implementing option B+ will probably require increased resources, especially in the short term. Nevertheless, overall, the benefits of option B+ are considered to outweigh the potential harm.

Cost and cost–effectiveness

In a cost-modelling exercise, the total cost (including drugs, diagnostics and service delivery) of keeping a woman on option B+ was an estimated US$ 2069 over five years (107). However, since maintaining a woman off ART also incurs costs for monitoring and follow-up, the incremental cost of moving from option B to option B+ was relatively low and varied between US$ 92 and US$ 605 depending on baseline CD4 and breastfeeding status. Several model-based analyses have supported the cost–effectiveness of strategies for preventing the mother-to-child transmission of HIV, with many finding option B and B+ to be cost-saving or highly cost-effective compared with option A. When outcomes beyond the mother-to-child transmission of HIV are considered, such as maternal health, preventing the mother-to-child transmission of HIV in future pregnancies and preventing horizontal transmission, option B+ has been found to be highly cost-effective compared with option B (108,109).

Equity and acceptability

A qualitative literature review on the acceptability of option B+ indicated high acceptability of lifelong ART among pregnant and breastfeeding women as well as among service providers. Women have raised some concerns about fear of drug toxicity for themselves and their children, but women generally value the health benefits and the ability to protect their children and their partners from HIV (110,111). The review also highlighted some of the challenges of lifelong treatment, including disclosure to partners and employers, stigma, lack of support and costs and time off work associated with clinic visits and drug pick-ups.

Early programme experience from South Africa suggests that pregnant women find same-day initiation of ART (starting ART on the day of HIV diagnosis) acceptable, especially because in this setting, many women are already aware of their status, have high levels of treatment literacy and can access support services (112). By contrast, same-day ART initiation in Malawi has been associated with a high rate of early loss to follow-up, with many women failing to return for a second visit (113).

The recommendation to initiate ART for all adults with HIV at any CD4 count will likely add to the acceptability of ART for pregnant and breastfeeding women, since it normalizes what may previously have been considered stigmatizing. A simplified and harmonized approach to ART among pregnant and non-pregnant adults will also promote health equity.

Implementation considerations

Many of the concerns raised about the recommendation to give all pregnant and breastfeeding women lifelong ART relate to how this policy should be implemented. Programme managers must address the issues of ensuring the quality of HIV testing, determining when to initiate ART in a newly diagnosed woman and maintaining the continuity of ART services in the postpartum period when implementing this recommendation.

ART for pregnant and breastfeeding women should ideally be delivered within maternal, newborn and child health clinics by integrating HIV and antenatal care. Integrated services benefit mothers, and this is likely to be feasible in settings with a high burden of HIV infection. However, achieving integration will depend on the context and the resources available in terms of staff time and physical space. In one retrospective cohort study from Malawi, 45% of the women interviewed reported that, although they started ART in an antenatal care clinic, they were referred to separate ART services soon after delivery (114). There is no established model for when to transition mothers on ART out of maternal, newborn and child health services, but a recent report from South Africa highlighted the importance of this potential additional loss point in the PMTCT cascade. In a retrospective review of women referred to ART clinics in the postpartum period, up to 25% did not remain in care five months after referral (115).

Beyond issues related to clinical service delivery, a key consideration for national programmes is the need for strengthened data systems to track women on ART across multiple delivery sites, along with the need for targeted interventions to improve adherence and retention, such as community support and the use of peer counsellors.

Research gaps

Significant knowledge gaps remain about where and how best to implement option B+ to improve retention and follow-up of the mother and infant pair. For example, the integration of ARV drug delivery in antenatal care and maternal, neonatal and childcare services as opposed to referral to ART clinics requires further implementation and assessment.

Adolescents and young women living with HIV face unique challenges in preventing the transmission of HIV to their children and attending to their own health needs, including poor access to reproductive health services, poor uptake of testing and poor retention in care (116). Operational research is urgently needed to identify the drivers of poor outcomes among adolescents, to define how to provide adolescent-friendly maternal and newborn health services and to develop specific strategies to improve retention in care.

Although ART in pregnancy and during breastfeeding provides clear public health benefit in terms of maternal health and PMTCT, the potential long-term harm from fetal and infant exposure to maternal drugs is poorly understood. The risk of congenital birth defects is likely to be low for the currently recommended first-line ARV drugs, but little is known about newer drugs and the possible effects on growth, development and organ maturation resulting from exposure to ART absorbed across the placenta and through breast-milk.

Rationale and supporting evidence

The recommendation is also based on the strong operational and programmatic advantages of alignment with the criteria for initiating ART among adults and children and the clinical benefits demonstrated by evidence from adult studies (126).

A systematic review of the evidence did not identify any study investigating treatment initiation strategies specific to adolescents. This subgroup was also not captured by adult studies that assessed the clinical outcomes of immediate versus CD4-driven ART initiation (127,128).

Perinatally infected adolescents are more likely to experience chronic diseases and neurodevelopmental, growth and pubertal delays in comparison to their age-matched peers. Older adolescents who acquire HIV behaviourally do not present the same clinical features but face potentially greater challenges in dealing with stigma and lack of family and community support to access care.

There are some limitations in extrapolating from the evidence for adults and some uncertainty around the potential benefits that immediate initiation of ART may have on adolescent health outcomes given the unique challenges that may arise in achieving long-term adherence and retention among adolescents. For this reason, the quality of the evidence reviewed for adults was downgraded for indirectness, and the overall quality of the evidence for treating all adolescents living with HIV was rated as low.

Comparing benefits and harm

To assess the potential benefits of starting ART earlier, a causal modelling study of data from southern and western Africa and Europe was updated using prospective data. This examined 4553 ART-naive perinatally infected adolescents 10–15 years old (median age 12.4 years), of whom 14% presented with CD4 counts above the existing eligibility thresholds of 500 cells/mm³. In the analysis, median follow-up time was 656 days. Mortality appeared higher when ART was started very late. However, after four years of follow-up, the difference between immediate ART versus initiating ART at or below 500 cells/mm³ was insignificant. These differences were similar for the adolescents who presented with a CD4 count above 500 cells/mm³. Overall, the study did not show any clear survival or growth benefit from early treatment in this population (129).

Indirect evidence showed that perinatally infected adolescents for whom treatment initiation is delayed to 10 years of age are unlikely to normalize CD4 count (130) and, after onset of chronic lung disease, do not fully recover lung functioning (131), suggesting that adolescents who have survived through childhood untreated may have limited gains from initiating ART earlier compared with younger children.

The high risk of loss to follow-up in this age group, particularly among adolescents aged 15–19 years (132–136), is an important factor in assessing the trade-off between risks and benefits of earlier ART initiation. Adolescents are also known to be less adherent than adults and younger children (137). Two systematic reviews on adherence and viral suppression showed varying rates for adolescents (138,139), and treatment failure was observed among 10% of 1007 perinatally infected children in the COHERE cohort, with the risk being higher with a longer time on ART and when treatment was started in adolescence (140).

Despite the limited clinical evidence in support of earlier treatment initiation for both perinatally and behaviourally infected adolescents and potential concerns attached to the risk of drug resistance because of poor adherence, experience to date suggests that aligning with the initiation criteria for adults will contribute to simplifying programming and further expanding ART coverage (141) and present crucial opportunities to engage adolescents living with HIV in care.

Equity and acceptability

In community consultations, adolescents, service providers, parents and caregivers emphasized the importance of ensuring that priority be given to adolescents most in need of treatment as well as the challenge of adherence (142–145). The key challenges identified included forgetting to take medicines, having unstable lives not conducive to daily medication and relative lack of power in treatment decision-making. Effective interventions and services to support adherence among adolescents – including community interventions and the use of peer educators – are required.

Feasibility and resource use

Earlier initiation of ART among adolescents is likely to be feasible within existing health systems. Because of late diagnosis (146), many adolescents are already likely to be eligible based on 2013 initiation criteria, and the increase in the overall number of adolescents starting treatment would therefore be relatively small (141). Increased patient enrolments would nevertheless increase demands on supply chains and provider workload. The experience of some national programmes has shown that, although all adolescents 10–15 years old can be treated, challenges include ensuring that commodities are available, strengthening laboratory systems and conducting provider training (141).

Increased demand for commodities, human resources and infrastructure is expected to require additional funding. A costing analysis shows that ARV drugs are likely to be the most significant cost driver (147). Laboratory commodities are likely to be the second largest contributor to total cost, followed by human resources and co-trimoxazole.

Implementation considerations

Ensuring that adolescents are diagnosed and receive ART in a timely manner will require developing adolescent-friendly health services and providers and strongly emphasizing support for adherence and retention in care. The full update of the consolidated guidelines will contain further guidance on operational and service delivery approaches to support the implementation of adolescent-friendly health services.

Research gaps

How earlier ART affects retention, adherence and HIV drug resistance among adolescents with less advanced disease requires further investigation. Improved age disaggregation of existing cohort and surveillance data is needed to improve understanding of adolescent-specific issues and needs.

Background

A review of evidence, together with programmatic data and operational considerations, has led to revised recommendations in 2015 to initiate ART in all children with HIV, aligning the recommendation for children with the new recommendations for adults and adolescents.

Infants and young children living with HIV have an exceptionally high risk of poor outcomes, with up to 52% of children born living with HIV dying before the age of two years in the absence of any intervention (148). By five years of age, the risk of mortality and disease progression in the absence of treatment falls to rates similar to those of young adults (149,150). Improved access to early infant diagnosis has increased the identification of infants living with HIV, but rates of ART initiation among infants living with HIV – all of whom should initiate treatment – remain suboptimal. Overall, most children who are eligible for ART are still not being treated, and ART coverage among children lags significantly behind that among adults: 32% versus 40% globally in 2014 (151).

Diagnosing and retaining children exposed to and living with HIV in care present unique challenges because of their dependence on a caregiver. Loss to follow-up has been particularly high (152), with retention especially challenging for children who are in HIV care but not receiving ART (153).

The 2013 WHO ARV guidelines aligned clinical and immunological criteria for ART eligibility for children older than five years with those for adults: that is, treatment was recommended for WHO clinical stage 3 or 4 disease or CD4 counts at or below 500 cells/mm³. ART was also recommended for all children living with HIV younger than five years of age regardless of clinical or immunological status. For children between one and five years of age, it was recommended that those younger than two years of age with WHO stage 3 or 4 clinical disease or CD4 percentage below 25% or CD4 count below 750 cells/mm³ be given priority. For infants younger than one year of age, a strong recommendation to treat regardless of clinical and immunological conditions was maintained. The 2013 guidelines recognized the challenges of treating infants in their first two weeks of life because of lack of treatment options for which safe and effective dosing has been established and the lack of experience globally, further complicated by the high rates of prematurity and low birth weight in low- and middle-income countries.

Countries with a high burden of HIV among children have rapidly adopted the 2013 treatment criteria (154), and some countries have decided to expand ART to all children and adolescents younger than 15 years to simplify ART delivery (155).

Fig. 3

Adoption of ART initiation for infants and children in low- and middle-income countries from the Global AIDS Response Progress Reporting, 2014.

Rationale and supporting evidence

A systematic review (127) conducted in 2013 and updated in 2015 identified only one randomized clinical trial, PREDICT, that assessed the clinical benefit of early ART initiation among children (156). The trial enrolled 300 children (1–12 years old, median age 6.4 years) with CD4 percentage above 15% and without United States Centers for Disease Control and Prevention stage C disease, randomizing them to either early treatment or deferred treatment until the CD4 percentage fell below 15%. There was no difference in AIDS-free survival or neurodevelopmental outcomes between the two arms, but height gain was better among those initiating ART earlier (157).

A causal modelling study (158), also updated in 2015 using prospective data, assessed outcomes for 7358 ART-naive children 5–10 years old (median age 7.2 years), of whom 26% (1903) presented with CD4 counts exceeding the existing eligibility threshold of 500 cells/mm³. In this analysis, after five years of follow-up, early ART differed slightly but significantly in mortality from waiting for the CD4 count to fall below 500 cells/mm³. The causal modelling analysis also showed significantly better growth response among those starting ART immediately (159).

In addition, other evidence suggests that initiating ART earlier could mitigate the negative effects of HIV infection on growth and pubertal and nervous system development (160–166).

Earlier ART may also promote immune recovery. In a study of the long-term effects of ART on CD4 cell evolution in children receiving ART, children with a greater degree of immunosuppression at baseline did not recover to normal values (CD4% >25%) even after five years of ART, whereas the CD4 percentage among children starting ART at higher CD4 levels normalized within a year of receiving ART (167). As shown by the normalization of inflammatory markers, earlier ART initiation is also likely to reduce HIV-induced chronic immune activation, thus potentially limiting the onset of chronic lung disease and increased risk of cardiovascular disease for which clinical correlates are still missing among children (168).

The recommendation to start ART immediately is conditional for children living with HIV from 1 to less than 10 years old because of the paucity of evidence supporting ART initiation regardless of the clinical and immunological conditions in this population (160,161). However, this approach is expected to provide significant programmatic advantages, especially in settings with limited access to immunological testing, a high burden of HIV disease and low ART coverage among children.

Comparing benefits and harm

In addition to clinical considerations, earlier ART is likely to expand coverage in this age group. A rapid assessment conducted in May 2015 to assess the implementation of a policy to treat all children younger than 15 years in Uganda found a 74% increase in the number of children newly initiating ART and an increase in ART coverage among children from 22% to 32% between 2013 and 2014 (141). The proportion of children receiving ART at lower-level health facilities increased from 42% to 46%, suggesting that simplifying the criteria for initiating treatment could also be instrumental in effectively decentralizing ART services. In addition, the time from eligibility to ART initiation significantly decreased, suggesting that simpler initiation criteria allowed more rapid treatment initiation. Programmatic experience suggests that children receiving ART have better retention than those in care but not receiving ART (155). The retention rates in Uganda appeared to be comparable among children starting ART when eligible or with CD4 above 500 cells/mm³, but there was a reduction in retention at six months, highlighting the need to ensure that children and caregivers receive appropriate counselling and support to stay in care.

The potential harm of earlier ART initiation includes short-term side effects that may predispose children to suboptimal ART adherence and subsequently treatment failure (169,170), along with the emergence of drug resistance and the need for second- and third-line regimens, for which options suitable for children are still limited. Treatment failure was observed in 10% of cases in the COHERE cohort, with the risk being higher the longer children are on ART and the older they are when initiating ART (171).

Increasing demands on the health system, drug stock-outs and consequent treatment discontinuation may also contribute to treatment failure and HIV drug resistance. Long-term side effects and chronic disease may result in increased morbidity and affect quality of life in adulthood. On balance, the likely clinical and programmatic benefits of earlier ART are likely to outweigh these potential types of harm.

Equity and acceptability

Expanding ART to every child living with HIV is expected to increase equity and be well accepted. In community consultations, the acceptability of earlier treatment of children living with HIV from the perspectives of parents, caregivers and health-care providers was based on the perceived health benefits for the child. However, psychosocial support for parents and caregivers, especially for disclosure, was highlighted as critical to facilitating initiation and improving adherence (172).

Feasibility and resource use

Implementing this recommendation is likely to be feasible, since it represents a relatively small increased burden on current health systems (141). Late diagnosis is still common (173), and an estimated 80% or more of children identified as living with HIV would already be eligible for ART based on 2013 recommendations. However, increased numbers of children receiving ART may also lead to higher demand on supply chain systems and increased provider workload. Laboratory monitoring will need to be strengthened to monitor treatment efficacy and identify treatment failures among children. The experience of some national programmes has demonstrated that treating all children with HIV is feasible but also highlights the importance of secure commodity supplies, adequate health worker training and the need to ensure sustainability of resources (141).

Increased demand for HIV commodities, human resources and infrastructure may require increased funding. A costing analysis in Zambia has shown that ARV drugs are likely to be the most significant cost driver, accounting for 81% of total costs among children 0–14 years old. Laboratory commodities were the second largest contributor to total cost, followed by human resources and co-trimoxazole (141).

Implementation considerations

As ART is expanded to all children regardless of clinical and immune status, children younger than two years or children with WHO stage 3 or 4 disease or CD4 percentage below 25% or CD4 count at or below 750 cells/mm³ (if younger than five years) and CD4 counts at or below 350 cells/mm³ (if older than five years) should be given priority for treatment because of their higher risk of death and rapid disease progression.

Implementation approaches should include using opportunities to deliver ART for children in maternal, newborn and child health settings (175).

Expanding ART services for children will require strategies to improve retention in care and to support adherence. Careful clinical monitoring remains essential to assess the risk of treatment failure, but lack of laboratory monitoring should not be a barrier to initiating ART (176).

Research gaps

How earlier ART affects retention, adherence and potential HIV drug resistance among children with less advanced disease needs to be investigated further. Optimal service delivery models to ensure rapid identification and ART initiation among infants and children also need to be investigated. Strategies are needed to provide an integrated package of care to reduce overall child mortality.

HIV infection

HIV infection was measured in 11 randomized controlled trials comparing PrEP to placebo, three randomized controlled trials comparing PrEP to no PrEP (such as delayed PrEP or “no pill”) and three observational studies. Across data from 10 trials comparing PrEP with placebo, the results from a meta-analysis demonstrated a 51% reduction in risk of HIV infection for PrEP versus placebo (192).

Mode of acquisition

When studies were stratified by mode of acquisition (rectal, vaginal or penile exposure), PrEP showed similar effectiveness across groups. The relative risk of HIV infection for PrEP versus placebo for rectal exposure is 0.34 (95% CI: 0.15–0.80, P = 0.01). For penile or vaginal exposure, the relative risk of HIV infection for PrEP versus placebo is 0.54 (95% CI: 0.32–0.90, P = 0.02) (192). Parenteral exposure to HIV was not analysed separately because only one study explicitly included people who inject drugs, and their exposure to HIV arose from sexual practices and incomplete access to sterile injection equipment.

Sex and gender

Of the 10 randomized PrEP trials reporting HIV outcomes, women were included in six studies and men in seven studies. PrEP was effective for both men and women. The relative risk of HIV infection for PrEP versus placebo was 0.57 (95% CI 0.34–0.94; P = 0.03) among women and 0.38 (95% 0.20–0.60; P = 0.0001) among men. Two placebo-controlled trials that targeted women exclusively showed very low uptake of PrEP (less than one third) in the active arm and no effectiveness on an intent-to-treat basis (183,186). PrEP effectiveness among women in four trials that included both women and men was higher. For example, among women younger than 30 years in a trial that included both men and women, the effectiveness was 72% (95% CI: 29–92%, P = 0.01) for TDF and 77% (95% CI: 25–90%, P = 0.01) for FTC + TDF PrEP (180). The results from a recent study (HPTN 067) among young, predominately single South African women receiving open-label FTC + TDF as PrEP showed that young women can maintain adherence, with 80% having substantial concentrations of drug detected at week 4 and 65% at week 24 in the daily PrEP arm (193). More information about PrEP in transgender populations is needed.

Adherence

When all studies were analysed together, the results produced significant heterogeneity. The results from meta-regression conducted to evaluate whether certain variables moderated the effect of PrEP on reducing the risk of acquiring HIV infection demonstrated that adherence is a significant moderator.

When studies were stratified according to adherence levels (high, moderate and low), heterogeneity was greatly reduced within adherence subgroups, demonstrating that most heterogeneity between studies can be explained by differing adherence levels. Within adherence subgroups, PrEP is the most efficacious among the high-adherence group (defined as >70% drug detection, but all studies in this group had adherence at or above 80%) and significantly reduces the risk of acquiring HIV infection in studies with moderate levels of adherence (41–70% drug detection). Among studies with low adherence (40% or lower drug detection), PrEP shows no effect in reducing HIV infection (192).

Safety

Ten randomized controlled trials comparing PrEP with placebo presented data on any adverse event. The risk of experiencing at least one adverse event during follow-up was common in participants in all trials. Across studies, the rates of any adverse event did not differ for PrEP versus placebo. Similarly, there were no differences across subgroups, including mode of acquisition, adherence, sex, drug regimen, drug dosing or age.

Eleven randomized controlled trials comparing PrEP with placebo presented the results for any grade 3 or 4 adverse event. Across studies, there was no statistical difference in rates of any grade 3 or 4 adverse event for PrEP versus placebo and there were no statistical differences across subgroup analyses, including adherence, sex, drug regimen, drug dosing or age (192).

Several studies noted subclinical declines in renal functioning and bone mineral density among PrEP users (194–196). These subclinical changes did not result in clinical events and were not progressive over time.

Drug resistance

The risk of drug resistance to FTC was overall low (11 people with FTC- or TDF-resistant HIV infection among 9222 PrEP users, or 0.1%), and this occurred mainly among people who were acutely infected with HIV when initiating PrEP: 7 people with FTC- or TDF-resistant HIV infection among 9222 PrEP users. The proportion of people with drug-resistant HIV infection did not differ in the PrEP and placebo groups among everyone at risk, although the number of events was low (n = 6 people infected). Multiple HIV infections (8–50) were averted for every case of FTC resistance associated with starting PrEP in the presence of acute HIV infection (192). Modelling the HIV drug resistance resulting from ART is predicted to far exceed that resulting from PrEP (197). Although mathematical models inform the risk of resistance, their results rely on data from clinical trials and make assumptions about risk of drug resistance selection during PrEP. How implementing PrEP on a large scale affects resistance overall is unknown, and active surveillance during PrEP scale-up may therefore be warranted.

Sexual and reproductive health outcomes

No evidence indicated that PrEP led to risk compensation in sexual practices, such as decreased condom use or more sexual partners (198,199).

PrEP does not appear to affect the effectiveness of hormonal contraception, although two studies found trends towards higher rates of pregnancy among oral contraceptive users who also took PrEP. When multivariate analysis accounted for confounders, this relationship became nonsignificant. Oral PrEP was not associated with increased adverse pregnancy-related events among women taking PrEP during early pregnancy (180,186). More information is needed about interactions between PrEP and the hormone therapy used by transgender populations.

The review sought to evaluate the effectiveness of PrEP in preventing HIV infection in the context of access to a combination of standard approaches to HIV prevention (192). Across all trials, PrEP was provided in the context of a package of HIV prevention interventions, including regular HIV testing and counselling, provision of condoms, screening and treatment for sexually transmitted infections, adherence counselling and other options relevant to the study population, such as access to contraception for women and methadone maintenance for people who inject opioids.