Treatment in children and adolescents

Optimising nevirapine dosing in breastfed HIV-exposed infants

With South Africa’s current PPTCT guidelines, breastfeeding infants have a much better chance of staying HIV-negative too. Giving nevirapine to infants for the duration of breastfeeding has proven efficacy, as demonstrated by the Breastfeeding, Antiretrovirals and Nutrition (BAN) study, which reported a transmission rate of only 1.8%, and is both WHO and South African policy. However, the nevirapine dosing recommendations in “the WHO and South African guidelines might not be appropriate for preterm and low birthweight infants,” said Dr Renée De Wall of the University of Cape Town, who presented data from a pharmacokinetic study of nevirapine in these small infants.¹

Currently, WHO guidelines recommend nevirapine dosage should increase according to age and weight — based on data from PPTCT efficacy studies, which were mostly conducted in full-term infants who weighed 3 kg, on average. For low birthweight neonates, WHO recommended doses of 2 mg/kg plus therapeutic drug monitoring (TDM) of nevirapine plasma trough concentrations.

“Those guidelines raised several concerns with clinicians as to their implementation in preterm and low birth-weight infants because of the limited safety, efficacy and pharmacokinetic data,” said Dr De Wall. One issue is that prematurity is associated with immature drug metabolism which might increase toxicity from some drugs. Furthermore, nevirapine undergoes auto-induction of its metabolism, which is why in adults, the dose is increased after two weeks. However, WHO guidelines do not recommend an increase in the dose after two weeks, which might cause sub-therapeutic concentrations. Meanwhile, South Africa’s guidelines, which recommend starting with 10 mg, might be associated with toxicity in preterm and low birthweight babies.

Western Cape nevirapine dosing recommendations

Alternative recommendations accounting for low birth weight and 14 day dose adjustments were developed by a group of paediatricians and pharmacokineticists and made policy in the Western Cape Province in February this year, though some facilities began using these doses earlier.

Therapeutic drug level monitoring (TDM) is also recommended on days 7, 14, 28 and 42, with a target nevirapine concentration in the range of 0.1–10 mg/L.

But clearly TDM is not something that can be done everywhere. So Dr De Wall and colleagues analysed samples that have been submitted in the Western Cape to see whether nevirapine trough concentrations were in the target range for preterm and LBW infants dosed according to Western Cape guidelines, and to see whether there was an association between nevirapine concentration and age post-delivery or weight. Unfortunately, most of the samples couldn’t be used in the analysis, mostly because of poor documentation of weight, dose, and timing of the  blood draw.

However, 65 samples from 56 infants could be included in the study from infants with a median gestation of 32 weeks at birth (range 28-36) and with weights mostly in the 1000-1799 g (though 18% weighed less than 1000 grams).

“We found no sub-therapeutic nevirapine concentrations, and no toxicity was reported by requesting clinicians,” said Dr De Wall. In addition, they found that nevirapine trough concentrations were significantly lower after 14 days of age, so the recommended dosage adjustment is apppropriate. 

“Current Western Cape guidelines for dosing result in adequate nevirapine concentrations, but the WHO guidelines might result in sub-therapeutic concentrations,” said Dr De Wall. However, that doesn’t necessarily mean that the doses recommended by WHO are less effective, but this study does suggest that further studies may be needed to be certain that nevirapine 2mg/kg is adequate in preterm and low weight infants.

A dramatic increase in the initiation of ART in children with HIV in South Africa

“From 2005-2010, the South African Government, with assistance from PEPFAR, implemented a substantial scale-up of ART for children,” said Elysia Larson of the US Centers for Disease Control (CDC) who is also a fellow with the Association of Schools of Public Health.² In fact, this is something of an understatement — since 2005 the number of children under 15 newly initiated on ART each year has increased 16-fold from 1,851 in 2005 to over 31,000 in 2010 — and this seems to be making a dent in South Africa’s high under-5 mortality rates.

But there is still room for improvement.

“The HIV/AIDS epidemic has been hard on children in South Africa,” said Larson. Between the mid-1990s and the mid-2000s, under-five mortality in South Africa had increased dramatically (almost doubling) — making a mockery of Millenium Development Goal 4, reducing child mortality by two-thirds between 1995 and 2015. “According to the South Africa Every Death Counts Writing Group, most of those deaths are caused by conditions that are either preventable or treatable, at least 35% are attributable to AIDS,” said Larson. “About 3% of children aged 0-18 years old are HIV-infected, and just over one-third of those children are under the age of five.”

The number would actually be much higher, were it not that most children with HIV who go untreated die quickly — 67.3% within the first five years according to UNAIDS estimates. A rather depressing poster presentation at the conference backed these estimates up with local data on the sad outcomes of many infants who became infected over the course of the Vertical Transmission (VT) Study, conducted from 2001 to 2004.³ This was before ART was made widely available to HIV-positive children in South Africa — even though treatments were available elsewhere, and despite the pleas of the late Nkosi Johnson to his government to provide life-saving treatment at the opening plenary of the 2000 World AIDS Conference in Durban.

In the VT cohort, at the end of two years, only 102 out of 259 of the HIV-infected infants were still alive and on study at the end of two years, though 12 had ‘moved’ and a handful were lost-to-follow-up (and one can probably assume dead). There were 126 confirmed deaths overall, with a median age of 6.1 months at death. The authors noted a difference in the risk of mortality between children who were infected perinatally (only 28.9% of whom were still in follow-up at the end of the study — 76 confirmed dead out of 128 infants — compared to those who were infected postnatally (59.5% who still remained in study at the end of two years). However, the postnally-infected children appeared to have a higher risk of serious morbidity, which the authors said underscored the need to maintain close follow-up of HIV-exposed who are uninfected earlier in life.

Since that time, putting HIV-infected infants on ART has been shown to dramatically reduce mortality — by at least 75% if started in early infancy. Those findings have given a major impetus to scaling up early infant diagnosis at the first immunisation visit — and making effective linkages into care and treatment services for those children testing positive.

South Africa’s Department of Health eventually produced ART guidelines for children in 2004 and the PEPFAR collaboration began about the same time. At the start of 2005, there were 7,000 children on ART in the country but Stats SA estimated that about 183,000 children would be in need of ART by 2010. In the National Strategic Plan 2007-2011, the goal was to try to put over 150,000 on treatment by March  2011.

PEPFAR’s support for the paediatric ART programme has grown substantially, accounting for 20% of children receiving ART in 2005 and over 75% of those receiving treatment by September 2010. This amounts to quite a large number of children — beginning with about 2,412 children in September 2005 to 79,416 children in September 2010, a 32-fold increase.

“Our most recent data from January-March of 2011 shows that there are 89,421 children under fifteen currently on treatment in PEPFAR-assisted facilities,” said Larson. But this figure is still only half of those estimated to need treatment.

PEPFAR’s support is channelled through non-governmental organisations providing infrastructure, equipment, staff, mentorship and training primarily for the public sector programme — and does not usually involve buying the medications.

PEPFAR impact projections

One thing PEPFAR is known for, however, is an emphasis on data collection, monitoring and evaluation — and Larson reported on a retrospective analysis of routinely collected PEPFAR monitoring data from October 2005 through September 2010.

The analysis was restricted to outcomes in the children under five starting ART , and looked at trends in number of children receiving services, and the proportion of children remaining on treatment, to estimate the number of years of life gained by treatment for children under five.  It included 26,207 children under five on PEPFAR-assisted ART programmes at the end of the analysis period.

For the analysis, 10% were removed from the total number as they could be expected to either stop treatment, transfer, or be lost to follow-up. The 10% figure was based on an analysis of periodic cross-sectional data of children in the programme from April 2008 to September 2009 which found that 90% of children remained alive and on treatment after 1.5 years. That analysis had also found 0-4 year olds were 1.4 times as likely to transfer out of the programme or die as 5-14 year olds (95% CI: 1.3-1.5) and males were 1.3 times as likely to stop treatment as females (95% CI: 1.2-1.5).

“Then, using UNAIDS estimates for mortality of children under five who are receiving treatment versus those not on treatment, we calculated the number of years these children were expected to live on ART versus the number that would be expected to live if they were not on ART,” said Larson. Again, UNAIDS estimates are that 67.3% of children would die without ART, while they project that on ART approximately 15% of HIV-infected children will die the first year, followed by about 7% each subsequent year.

Of the 23,582 children under five expected to be retained in care, about 15,875 children would have died without ART compared to 9,877 deaths on ART.

“That means that providing therapy to these children will avert about 6000 deaths in the first five years on treatment,” she said. The analysis also found that after thirty years, the 26,207 children on ART will have lived a total of 255,123 years compared to 136,920 years if they were not given ART — a gain of 118,203 years of life. “Increasing access to ART for children under five has almost doubled the number of years these children will live,” said Larson. “And as more children are provided with treatment, more under-five deaths will be averted.”

This decrease in mortality has already begun to affect under-five mortality in the country, with a levelling off in the number of deaths after 2006, and what appears to be a decrease from 2007 to 2008, according to data from Statistics South Africa.

“This leads us to believe that the paediatric ART scale-up has been a tremendous public health accomplishment, however there is still a great unmet need for treatment,” she said, adding that she hoped that nurse-initiated and managed ART (NIMART) should help increase the number of children starting on treatment.

Remarkably consistent findings at Kheth’Impilo-supported sites

These nurses must be trained and mentored to fast-track HIV-infected infants onto ART, according to another poster presentation from the team at Kheth’Impilo, which reported findings remarkably consistent with Larson’s study.⁴ In addition, “closer integration of maternal, child health and HIV/ART clinics with efficient systems to perform early infant PCR testing and rapid retrieval of results are critical,” the poster’s authors concluded.

The Kheth’Impilo poster described a study looking at temporal trends in baseline characteristics and programmatic outcomes in children starting ART at thirty sites (18 primary care, twelve hospitals) that the NGO supports in four provinces (Western Cape, Eastern Cape, Mpumalanga, and KZN). As already noted KI provides infrastructure support, capacity development, clinical mentoring, community-based adherence support and electronic data collection systems at these sites.

The study included all children under 16 years of age enrolled on ART between January 2004 to September 2009 followed until March 2010, or until Kheth’Impilo finished supporting the site. This amounted to 3007 children. Consistent with the PEPFAR study, the number of children enrolling monthly increased from 1.9 per month when the study started to 106 in 2009. Amongst the 146 children lost to follow-up, 68 (46.5%) had a valid civil identification number, of whom 23 (33.8%) were registered as deceased in the national registry by November 2010. The corrected mortality was 6.1% (CI: 5.10-7.3%) after 24 months on ART and retention in care was 89.6% (CI: 81-91%) with no differences between annual enrolment cohorts — which suggests that despite the rapidly increased numbers going onto treatment each year, South African paediatric ART programmes are quite effective. After four years of ART, 84.1% (CI: 80.9%—86.7%) of children remained in care.

As Ms Larson noted in her talk, the proportion of children continuing on treatment is high (89.7% after 1.5 years in the PEPFAR study) compared to adults and UNAIDS estimates.

“Nevertheless, because the primary reasons for attrition were loss-to-follow-up and transferred care, it suggests a patient tracking system, such as a national system utilising a unique patient identifier, could greatly facilitate the care of patients and the monitoring of patient outcomes, though it must protect human rights and patient confidentiality,” she said.

It may be more important to implement such a system in order to make certain that young children are appropriately screened for HIV at both their immunisation visits and at 18 months of age (at which point, ELISA tests can be used since weaned children should no longer express their mother’s antibodies). Such a system could also be used to make certain that the mother gets her child’s test results and that effective referrals are made to the ART clinic and to community-based support for the mother and her child.

Can these responses be sustained?

If there is a dark cloud on the horizon, it is that these children will require effective antiretroviral therapy for life — and careful monitoring and support to make sure that it remains effective.

While the prospect of sustaining lifetime antiretroviral treatment is daunting enough in adults who may have access to only first-line and second-line treatment, the prospect is even more challenging in children, who may experience failure of first-line treatment due to very high viral load, adherence problems or sub-optimal drug concentrations.

For young children, the options for second-line treatment are limited due to the lack of suitable formulations for children, and third-line treatment is even more difficult – and expensive – to piece together. Drug resistance that develops after the failure of first-line treatment critically restricts later options and further complicates the assembly of an effective second or third-line combination..

So issues of adherence, drug resistance, virological monitoring and drug access are particularly acute in this group of patients.

But monitoring for treatment failure, whether in adults or children in resource limited settings, is difficult because “current immunological definitions of immunological failure are poor at detecting viral rebound,” said Dr Mary-Ann Davies, who presented data on the use of different immunological criteria to monitor children receiving ART in the South African International Epidemiologic Databases to Evaluate AIDS (IeDEA) paediatric sites.⁵

The study included 2513 children below the age of sixteen, whose viral load had been completely suppressed during the first year on treatment.

These children were then monitored with CD4 cell count and viral load measurements every six months to determine sensitivity, specificity, positive and negative predictive values (PPV) and (NPV) of WHO and US definitions of immunological failure for identifying virological rebound in a child on ART for at least 18 months.

After following the children out until 3.5 years of ART, the WHO definition was found to have a very low sensitivity although PPV was better (42%); the sensitivity was higher for the US definition (though still low), and the PPV was very low (20%).

To make these definitions work better, Dr Davies said that one could get another CD4 count to confirm failure, which would increase PPV, but decrease sensitivity. Furthermore, she noted that this simply isn’t done in practice in South Africa when viral load is available. So the researchers also looked to see whether the sensitivity and positive predictive value could be improved by combining the most sensitive immunological criteria with a single HIV-RNA measurement to define failure (a targeted viral load strategy). This approach increased the PPV for both the WHO and US criteria considerably, but the diagnostic accuracy was still sub-optimal.

Methods for determining immunological failure

Essentially, immunological failure and virological failure seem even less likely to correspond in children than adults. It is important to point out however, as Dr Davies noted, that her study was merely an algorithm applied retrospectively to the data set, which had no clinical data on the participants. So it wasn’t possible to exclude the possibility that low CD4 cell counts may have been due to a clinical cause such as tuberculosis. Nevertheless, the researchers concluded that better access is needed to viral load monitoring in order to accurately diagnose treatment failure in children — otherwise, virological failure may simply go unnoticed.

Unfortunately, this would present some of the same challenges as it does in adults, where studies have shown that routine viral loads do not appear to improve patient management and are not cost-effective — at least not during the first few years when treatment works well in most people.

In the Kheth’Impilo study, the proportion of children achieving virological suppression after six months was fairly high, at 83% (95% CI: 80-85.1) and 79.7% (95% CI 76.6-82.6%) after twelve months on ART. The study did note however, that children aged over two years when they went onto treatment had significantly better virological suppression 81.8% (95% CI: 80.2-83.3%) compared to 72% (95% CI 67.4-77.7%, p<0.0005) in children aged under two years of age at any time-point on treatment. However, other cohorts in similar settings, perhaps where children and their caregivers receive less support, have described higher rates of virological failure. For instance, one recent report described 38% of children on ART for more than a year experienced virological failure, and the authors’ remarked that the “correlation between virologic failure and immunologic decline was nearly absent.”⁶

Treatment failure could reflect the inadequacy of antiretroviral formulations for infants and small children — and/or the difficulty young mothers may have making sure that their infants get the doses they need (see more on adherence below). But it also means that a significant proportion of the children will need to switch to second-line therapy at a fairly young age. Over time the number of children failing first and then second-line treatment can only be expected to increase.

But the available treatment options for switching are even more limited in children than adults in most resource-limited settings. Consequently, there are questions about how best to manage treatment failure. One approach could be to keep them on virologically failing failing treatment regimens as long as their CD4 cell count and/or clinical response is good. The danger, of course, is that over time on a failing regimen, they will develop drug-resistant virus that will be less likely to respond to subsequent ART regimens as they become available. Children who became infected in utero or despite PPTCT may already be at a disadvantage due to exposure to antiretrovirals that could have led to resistance.

Drug resistance increasing in children

Indeed, drug resistance is emerging in children on ART in South Africa.

One study described a large number of resistance mutations found in thirteen children at Kalafong Hospital in Pretoria, who were started on a protease inhibitor-based regimen when they were younger than 36 months of age (mean age 22.4 months), some of whom had already been switched to a second line NNRTI regimen.⁷ The mean duration of treatment at testing was 45.9 months. “All these children were deemed to be at high risk as they suffered co-morbid conditions and had documented poor adherence,” according to the poster’s authors. Three of the children had WHO Stage 3 disease, and ten had WHO stage 4 disease; while 62% of their caregivers reported poor adherence, and 54% had a history of missed clinic appointments. Only one had ever achieved an undetectable viral load.

Resistance testing found that all of these children had virus resistant to nucleoside analogues (the 3TC-related M184V mutation being the most common), five had major protease inhibitor resistance mutations, and nine had resistance to mutations to non-nucleoside analogue reverse transcriptase inhibitors (NNRTI, i.e. efavirenz or nevirapine) — only one of these had been exposed to nevirapine as PPTCT.

A second poster presentation found drug resistance was more common in paediatric than adult patients experiencing virological failure on their second line protease inhibitor-based regimen at Tshwane Academic Hospital (Pretoria).⁸ All patients had been started on d4T/3TC with either nevirapine or efavirenz, and they were all switched to either AZT/ddI/ or a variation (AZT/abacavir) with ritonavir-boosted lopinavir (Kaletra / Aluvia).

About half of the failing adult patients showed no evidence of resistance at all, suggesting that they simply were not taking their treatment, but the virus from 91% of the children (30 out of 33) of the paediatric patients failing treatment had resistance mutations. In these children, virus in seven out of 33 (21%) had major PI mutations, five had more than three PI mutations, and 27.3% still showed evidence of the K103 mutation conferring resistance to NNRTIs (note, the mutation would probably have been detected in more children if they were still taking nevirapine or efavirenz). Finally, 75.8% of the children had virus that was resistant to 3TC, and 18.2 % had multiple thymidine analogue mutations (TAMs) that significantly decrease the chances of responding to subsequent nucleoside analogues.

Notably, some of these children had only moderate levels of viral load which — in the presence of less than universal resistance to the PIs — suggested that they may have been taking their treatment, but that adherence was suboptimal.

Nevertheless, the authors concluded that children exhibited increased levels of resistance to all three drug classes, perhaps because of higher levels of viral replication that typically occur in children or perhaps because some of them had previous exposure to unboosted PI therapy (ritonavir monotherapy).

Dr Theresa Rossouw of the University of Pretoria, who was the lead author of the latter poster, also gave an oral presentation describing findings from the Southern African Treatment and Resistance Network (SATuRN), which is monitoring emerging patterns of drug resistance in the region.⁹ She described higher rates of resistance among 49 children in SATuRN (though it is not clear to what extent these patients overlap) with PI resistance being documented in almost one third of the patients and three or more PI resistance mutations in 20.4%. NNRTI Resistance was seen in 21/49 (42.86%). Resistance to the nucleoside analogues was detected in 41/49 (83.67%), with eleven out of 49 (22.45%) with TAMS, and eight out of 49 (16.28%) with three or more TAMs.  She noted several other South African cohorts have also documented high levels of resistance among children failing treatment.^{10, 11, 12, 13}

Resistance, and limited treatment options will make it more difficult for these children to live with HIV into adolescence, when other challenges to treatment success seem to emerge.

High rates of virological failure in adolescents living with HIV compared to young adults

There are significantly poorer rates of virological suppression and higher rates of virological failure in adolescents compared to young adults, according to an observational cohort study in Gugulethu reported at the South African AIDS Conference, even though rates of mortality and loss-to-follow-up were similar in the two populations. “Further investigation is needed to determine what factors are associated with virological failure in adolescents,” according to Dr. Mweete Nglazi of the Desmond Tutu HIV Foundation, though the underlying cause seems to be poorer adherence.

“Adolescence is a unique bio-psychosocial and behavioural stage of life, with characteristics — that may be exacerbated by HIV infection — that often result in poor adherence to medical programmes and chronic medication,” said Dr. Nglazi.¹⁴   While studies in other settings have reported poor adherence and outcomes in adolescents living with HIV, she noted that there had been few studies in the region and far fewer that directly compared adolescents with adults.

So she and her colleagues looked at data from an ART cohort database derived from patient records to compare adherence, virological responses, loss to follow-up and mortality among adolescents and young adults receiving ART at the Hannan Crusaid clinic in Gugulethu, a Cape Town township. The study included 69 adolescents who started ART between nine and 19 years of age, and 818 young adults who started when they were between 20-28 years old. All of the patients received ART from between September 2002 to June 2009, with the follow-up data censored as of June 2010.

At baseline, a significantly lower percentage of the adolescents were female (66.2% vs. 86.6%) which would be expected since there would be an almost an equal number of boys among the adolescents infected as infants — while in later life sexual transmission puts more young women at risk. Median CD4 cell counts (133 and 116 respectively) and viral load (~4.8 log/mL) were similar for both groups. However, the adolescents were significantly more likely to be taking a regimen containing efavirenz (81.5 vs 49.2%) than nevirapine. Data on virological responses on the first year of treatment were available for a slightly smaller treatment-naïve cohort (46 adolescents and 716 adults) as some clients had transferred in from other facilities

Mortality and loss to follow-up were actually more common in the young adults than the adolescents, though these differences were not statistically significant.

Notably, a poster presentation reported similar findings of generally lower rates of mortality and loss to follow-up among adolescents (10-19 years old) than young adults (19.1-24 year old) or adults (24.1-28 years old) starting ART in a decentralised HIV/TB programme in rural Zimbabwe.¹⁵ This cohort included 1855 patients, 748 of whom were adolescents. Of the adolescents, 52.3% were female so this group was even more likely to have been infected as infants). The adolescent were also slightly less likely to have WHO stage 3 or 4 disease at baseline (62.2%, vs 71.4% and 78.3% in adolescents, young adults and adults respectively). Participants were initiated on ART between 2005 and 2010 (but mostly between the years of 2007 and 2009).

The young adults were the most likely to be lost to follow-up in this cohort, probably, the authors surmised, because they were more likely to move to seek out economic opportunities. Meanwhile adolescents were most likely to be retained in care. More importantly, the mortality rate was 2.6 (95% CI 1.9-3.6), 5.2 (95% CI 3.6-7.6 and 6.5 (95% CI 5.3-81) per 100 person years in adolescents, young adults and adults respectively. However, the study did not include data on virological responses.

Virological data were available for the Gugulethu adolescents, and frankly the results were somewhat puzzling. Initially, virological suppression was similar in both populations, with 97.3% of the adolescents and 89.6% of the young adults achieving viral loads below 400 copies/ml at week 16. However, something happened between week 16 and week 32, at which point only 37.5% of the adolescents had undetectable viral loads versus 75.1% of the young adults, which was a highly significant difference (p<0.001). By week 48 only 27.3% of the adolescents versus 63.1% were virally suppressed (p<0.001).

Mortality and loss to follow-up by age range in Gugulethu

The likelihood that some of the adolescents had been exposed to nevirapine for PPTCT is extremely low, given that they were all over the age of nine and PPTCT wasn’t available in South Africa when they were born. Likewise, it is unlikely that there would be a profound difference between regimens containing efavirenz versus nevirapine, unless side-effects were an issue.

The most likely factor would seem to be poorer adherence. Ironically, the Zimbabwean study concluded that its model of decentralised care had successfully managed the adolescents with less frequent clinic visits and adherence interventions than the norm in other programmes. In this model, ART was initiated by a mobile team, with intensive follow-up provided by nurses for three months, then three-montly clinic visits with a clinician and counsellor out to 18 months, after which time the patient would be provided with a three-month supply of medicine and a clinic visit would only be triggered by signs of clinical or immunological failure and poor adherence. But that study didn’t monitor viral load responses — and it may be a mistake to assume that better survival rates among this cohort, who have already survived living with HIV for most if not all of their lives, indicates a better treatment response.

Or possibly, there are different barriers to adherence in periurban Gugulethu than rural Zimbabwe. Perhaps the family structures supporting adolescents with HIV are different in these settings. However, it is notable that in another presentation from Zimbabwe, that looked at ways to better monitor adherence on ART in slightly younger children (median around 8 years of age), “the children reported lower adherence levels than their caregivers,” according to Dr Linnetie Mugore of Rhodes University.¹⁶

Regardless, almost all of these studies stressed that specific adherence strategies targeting children and adolescents need to be developed. The authors of the Kheth’Impilo study wrote that “scale-up of community adherence support for young children initiating ART is… imperative.” Indeed, caregivers need specialised training to dose young children appropriately. 

And there may be a gap between childhood and young adulthood where children are too old to be given doses, and yet, not old enough to take on the responsibility of taking chronic medications for the rest of their lives. Many adolescents may rebel against the notion, while for others, there are issues related to peer pressure and self worth.

The social pressures in the lives of children living with HIV were examined in more depth at the following International Conference on HIV and Social Science and Humanities that took place after the AIDS conference in Durban and will be discussed in a future HATIP.