Although children are doing well on HIV treatment, survival is affected by a number of factors, including delaying the start of treatment, long-term retention in care and preventing treatment failure so as not to compromise future treatment options, according to research presented on July 19 at the Eighteenth International AIDS Conference in Vienna.
A multiregional analysis in Africa and Asia over a period of 18 months (within the paediatric International Epidemiological Databases to Evaluate AIDS (IeDEA) collaboration) of over 13,000 HIV-infected children of 15 years of age and under looked at the probability of, and reasons for, death and programme loss after starting antiretroviral treatment according to the 2006 World Health Organization (WHO) criteria.
These regional cohorts are part of the larger IeDEA collaboration, which includes cohorts in the United States, Australia and Latin America. This global collaboration is intended to address broad questions about HIV that cannot be answered at the single cohort level.
Data of 13,611 children from 54 clinics in Asia, East Africa, Southern Africa and West Africa were analysed. The 18-month crude probability of death after the start of antiretroviral therapy differed slightly according to the region, Asia 5.8% (4.7 to 7.2), East Africa 5.5% (4.6 to 6.5), Southern Africa 6.4% (5.8 to 7.1) and West Africa 8.9% (7.8 to 10.1) (95% CI, p<0.0001).
Factors that put children at increased risk for poor survival included advanced illness, low weight for age, being under 12 months of age, being severely anaemic and having a CD4 percentage under 10 or missing at baseline.
Six percent of children died and over 12% were lost to follow-up, with considerable variations across regions and sites, Valeriane Leroy reported. Loss to follow-up was defined as not having been seen for more than six months since the last visit. The median age for starting treatment was five, with a median CD4 percentage of 12.
The 18-month crude probability of loss to follow-up varied significantly by region, from 4.5% (3.5 to 5.8) in Asia, to 17% (15.6 to 18.6) in East Africa, 11.5% (10.6 to 12.5) in Southern Africa and 24.7% (23 to 26.4) in West Africa (95% CI, p<0.0001).
Factors associated with a greater loss to follow-up included being under three years of age, advanced illness, a missing CD4 percentage at baseline, low levels of haemoglobin (being anaemic), and having started treatment after 2005.
Other factors associated with the increased probability of loss to follow-up included being in a public rural site, having cohorts of more than 250, and that patients being asked to pay for laboratory tests, as well as being on a protease-inhibitor-containing (PI) regimen. Over 75% were on a regimen of two nucleoside reverse transcriptase inhibitors (NRTIs) and one non-nucleoside reverse transcriptase inhibitor (NNRTI), with 20% on a regimen of two NRTIs and one PI. Treatment regimen varied significantly by region in Asia and East Africa: close to 100% were on the NNRTI-containing regimen; while in both Southern and West Africa a third were on a PI-containing regimen.
CD4 percentage greater than 20 at baseline improved outcomes, as did telephone calls to track the children. Dr. Leroy suggested that telephone calls were less stigmatising than home visits.
These trends over time, coupled with the size of the cohorts, Valeriane Leroy stressed, signify the measurable overload health facilities are experiencing in resource-limited settings with scale-up. The provision of free treatment and laboratory services, she noted, are critical incentives to assure that patients return to care.
Treatment programmes for children are also having to address the need for second-line treatment, which may be different in some settings due to the limited number of drugs available to treat children, and the higher costs of some products used in second-line treatment.
Determining when to switch treatment in children may have a critical impact on their future prospects of successful treatment.
Elizabeth Reddy and colleagues, in another presentation that looked at a cohort of HIV-infected children in Tanzania, showed that depending only on clinical and immunological guidelines to determine treatment failure resulted in long periods of time with undiagnosed treatment failure. This had a negative impact on the effectiveness of future treatment.
Viral load testing was done on a group of children in HIV care, who had been on non-nucleoside reverse transcriptase inhibitor-based (NNRTI) first-line treatment, for 180 days or more. First-line regimens included zidovudine/lamivudine (AZT/3TC) or stavudine/lamivudine (d4T/3TC) and nevirapine or efavirenz. The study was observational so the choice of regimen was the physician’s.
Genotypic resistance testing was done on those samples with viral loads equal to or greater than 1000 copies/ml. The children were then followed for 12 months. Clinical staging, antiretroviral regimens, viral load levels and CD4 counts were recorded. Clinical or immunological treatment failure was defined according to WHO 2006 guidelines. Virologic failure was defined as HIV viral load decrease of less than 1 log10 after 70 days, or 400 copies/ml or more after 180 days on antiretroviral treatment.
At baseline, out of a total of 206 children, 33.4% (69) had virologic failure, six (9%) of whom had clinical or immunological failure. Of the 45 with genotype results, over 90% had resistance mutations, and all had NNRTI mutations.
Of the 57 children who had been followed for six months or more, 82.4% (47) had started a second-line treatment regimen: abacavir/didanosine and lopinavir/ritonovir (ABC/ddI+LPV/rtv).
Viral load levels were subsequently available for 31 of the children, who had a mean of 170 (70 to 280) days on second-line treatment. Of these, eight had virological failure and two had clinical or immunological failure.
Most of the children who experienced treatment failure on first-line NNRTI-based treatment as part of a public health approach, in spite of numerous resistance mutations, did well on a protease-inhibitor based second-line treatment regimen. However, the risks of virological failure was five times greater in those children (8, or 32%) who had a viral load equal to or greater than 100,000 copies/ml before starting second-line treatment.
Dr. Reddy concluded that larger numbers and longer follow-up were needed to determine the cost:benefit ratio of viral load monitoring of paediatric antiretroviral treatment in resource-limited settings.
Effective means to identify treatment failure early, as well as creative and workable approaches to keep children in antiretroviral treatment programmes, are needed to improve survival of HIV-infected children.
Further information
Presentations and their related abstracts from Valeriane Leroy and Elizabeth Reddy are available on the official conference website.
Leroy V et al.18 month mortality and loss-to-program in ART-treated children in Asia and Africa: the International Epidemiologic Databases to Evaluate AIDS (IeDEA) paediatric multiregional collaboration. Eighteenth International Conference on AIDS, abstract MOAB0202, Vienna, July 2010.
Reddy E et al. Virologic outcomes on second line antiretroviral therapy (ART) for HIV-infected Tanzanian children with and without clinical or immunologic failure at ART switch. Eighteenth International Conference on AIDS, abstract MOAB023, Vienna, July 2010.