Secondary Logo

Journal Logo

Epidemiology and Prevention

Outcomes of Antiretroviral Therapy in Children in Asia and Africa

A Comparative Analysis of the IeDEA Pediatric Multiregional Collaboration

Leroy, Valeriane MD, PhD*,†; Malateste, Karen MSc*,†; Rabie, Helena MD, PhD; Lumbiganon, Pagakrong MD, PhD§; Ayaya, Samuel MD, PhD; Dicko, Fatoumata MD; Davies, Mary-Ann MD#; Kariminia, Azar PhD**; Wools-Kaloustian, Kara MD, MS††; Aka, Edmond MD‡‡; Phiri, Samuel MD, PhD§§; Aurpibul, Linda MD‖‖; Yiannoutsos, Constantin PhD¶¶; Signaté-Sy, Haby MD, PhD##; Mofenson, Lynne MD*,**; Dabis, François MD, PhD*,† for the International IeDEA Pediatric Working Group1

Author Information
JAIDS Journal of Acquired Immune Deficiency Syndromes: February 1, 2013 - Volume 62 - Issue 2 - p 208-219
doi: 10.1097/QAI.0b013e31827b70bf



The HIV pediatric epidemic continues to expand,1 with an estimated 3.4 million children under 15 years of age living with HIV as of December 2010. However, although great strides have been made to expand access to care for adults, access and use of pediatric antiretroviral therapy (ART) have lagged behind.1 Providing ART to children poses unique challenges to HIV care programs. This may be due to weak linkage between prevention of mother-to-child transmission (PMTCT) and child health services, poor access to HIV diagnostic tests in children, and complexities of infant and child HIV care differing from those generally encountered among adults. Limited access to pediatric antiretroviral drugs and availability of drugs in appropriate formulations and the reliance on an individual other than the patients themselves to provide the medications are only a few of the challenges.2,3 Thus, overall, at the end of 2010, only 23% of the estimated 2,020,000 HIV-infected children <15 years of age eligible for ART were treated in low- and middle-income countries, with coverage rates varying from 2% in Sudan to 88% in Botswana.1,4

Characteristics at ART initiation and short-term outcomes (mortality, retention in program) are evolving in pediatric ART programs. Recent reports show decreased mortality in children after ART initiation,5–12 but children continue to initiate ART with advanced disease.13,14 In addition, programs that have rapidly scaled up ART have, as in adults, high loss to follow-up (LTFU) rates.6 In adults, high rates of LTFU correspond to underreported mortality15,16 suggesting that in children a significant proportion of child mortality may be unseen as well. Several site-specific and individual patient factors could explain the heterogeneity in effectiveness of HIV-care program outcomes.17 However, factors associated with mortality and cohort attrition in children have received little attention to date and need further investigation to evaluate the impact of ART scale-up on pediatric outcomes at the program level.17

In 2006, the US National Institutes of Health launched the International epidemiologic Databases to Evaluate AIDS (IeDEA) initiative to better describe trends in HIV epidemiology in the context of ART access by region of the world ( As part of the pediatric component of the IeDEA multiregional collaboration, we have combined data across 4 resource-constrained IeDEA regions to study 18-month mortality and loss to follow-up rates after ART initiation and explore their baseline individual and site-specific determinants.


Ethics Statement

Each participating pediatric HIV clinic formally agreed to be included in the IeDEA collaboration, with local Institutional Review Board and NIH approval.

Study Design

We conducted a multiregional cohort analysis within the IeDEA collaboration. All clinical centers from 4 (Asia Pacific, East Africa, Southern Africa, and West Africa) of the 6 IeDEA Collaboration regions collecting data on pediatric HIV care were eligible.

Study Population

All HIV-infected children (positive polymerase chain reaction <18 months or positive serology ≥18 months) aged <16 years at ART initiation, with documented gender, who were ART naive (except for exposure to perinatal PMTCT prophylaxis) and who initiated ≥3 antiretroviral drugs from November 21, 1995 (first ART initiation in Asia Pacific), to June 24, 2009, were included irrespective of the first-line ART regimen. The database was closed on December 20, 2009.

Each ART program within the consortium has its own protocol for clinical follow-up and laboratory testing. However, HIV-infected children were typically seen in clinics at least every 3 months, and their CD4 count was measured every 6 months to monitor their immunological response to ART. Routine viral load monitoring was not available at most sites.


The outcomes of interest were (1) mortality, defined as a death documented in the database; (2) loss to follow-up, defined as failure to return to the clinic for >6 months after the last clinic visit was reported in the database, the closing date of the database was >6 months since that visit, and the child was not known to be dead or have transferred to another treatment site.

Key Variables and Definitions

Baseline was defined as the date of ART initiation. Follow-up was censored at the earliest of the following: date of the last clinical contact up to 18 months on ART or date of transfer out or date of death. Explanatory baseline factors included the following: gender, age, CD4 count and percent, year of ART initiation, and first-line antiretroviral regimen [containing a nonnucleoside reverse transcriptase inhibitor (NNRTI) or a protease inhibitor (PI)]. The clinical stage of HIV disease was defined as either less advanced [CDC stage A/B or World Health Organization (WHO) stage I/II/III] or advanced (CDC stage C or WHO stage IV).

Age at ART initiation was categorized into 5 groups: <12 months, 1–2 years, 3–4, 5–9, and ≥10 years. The severity of immunodeficiency was defined according to the 2006 WHO criteria18: CD4% < 25% or CD4 count <1500 cells per cubic millimeter for children <12 months; CD4% < 20% or CD4 count < 750 cells per cubic millimeter for children between 12 and 35 months; CD4% < 15% or CD4 count < 350 cells per cubic millimeter for children between 36 and 59 months; CD4% < 10% or CD4 count <200 cells per cubic millimeter for children ≥5 years. According to the toxicity scales from the National Institute of Allergy and Infectious Diseases, severe anemia was defined as hemoglobin (Hb) <7 g/dL for children ≥57 days old. Baseline values for laboratory and clinical measurements were defined as the values closest to ART initiation that fell into a window of 90 days before ART initiation and 30 days after. Site-specific variables from a pediatric site survey conducted in 2009 were also analyzed19: IeDEA region, rural versus urban site, length of clinic experience with provision of pediatric ART, clinic size (total number of children included per site during the study period), family-centered care approach, use of generic drug combinations, free access to ARV drugs, free access to opportunistic infection (OI) prophylaxis, free access to laboratory tests, documentation of death events, and lost-to-follow-up tracking system.

Data Management

In collaboration with participating sites, each IeDEA regional data center is responsible for the development of data collection systems, the establishment of mechanisms for receiving and combining data from individual sites, verifying data quality, harmonizing lists and definitions of variables. In addition, each regional data center develops and implements methods for analyzing cohort data and performs training on data collection, processing, and cleaning. Four regional data sets were prepared and were transmitted to us by the regional data centers. These data were merged for the purpose of this analysis.

Statistical Methods

Baseline categorical data are presented as frequencies (percentage) and continuous data as medians and interquartile ranges (IQRs). Continuous variables were compared using the Kruskal–Wallis test and categorical variables using χ2 or Fisher exact tests. Heterogeneity of baseline differences between regions was tested. A competing risk model was used to analyze the independent risk of the 2 failures types: death and LTFU. LTFU is a competing cause of death, potentially increasing the risk of death because of ART interruption. Thus, assumptions about the independence of these 2 outcomes are not realistic. For this reason, we used a cumulative incidence function to estimate the cumulative probability of each outcome over 18 months of follow-up.20 Cumulative incidence functions were also estimated adjusted for baseline covariates.

To study correlates of the 2 outcomes, univariate analyses were first conducted. Second, we conducted multivariate competing risk analysis, using the fine and gray proportional subdistribution hazard regression model using the R statistical software version 2.11.1 (The R foundation for Statistical computing, Vienna, Austria) with the cmprsk package.21,22 Two multivariate models were created using this methodology. Variables were included in the full model if they were associated with each outcome in univariate competing risk analysis with P < 0.25.23 A final model was then created, by using a backward elimination procedure with a P value <0.05 considered statistically significant in the adjusted analysis. There was a substantial modification of the effect of the “search LTFU variable” in the full model. However, as subgroup analyses could not be run with too small sample sizes in some of the strata, we omitted the variable “search LTFU” from both mortality and LTFU models. The adjusted subdistribution hazard ratios (asHR) were reported with their 95% confidence intervals (95%CIs). Variables for which data were available on <70% of the patients were not included.


Description of Sites

From 2000 to 2009, the collaboration included 54 pediatric clinical centers from 4 IeDEA regions with 13,611 children overall: 1454 from 11 Asian sites, 3114 from 23 East-African sites, 6162 from 10 Southern-African sites, and 2881 from 10 West-African sites (Table 1).

Summary of Baseline Characteristics and Site Description by IeDEA Region (N = 13,611 Children in 54 Clinics)*

The sites were predominantly urban or semirural public sector clinics (Table 1). Significant between-region heterogeneity in site characteristics was observed. Asian and East-African cohort sites primarily had <250 children registered, whereas Southern and West-African cohort sites primarily had >500 active patients (P < 0.01). Across these 4 regions, 93.8% of sites had free access to all laboratory tests, and 93.3% had free access to first-line ART. Rates of access to free OI prophylaxis were lower overall (83.3%) with substantial regional heterogeneity: 77.2% in Asia, 100.0% in East Africa, 96.1% in Southern Africa, and only 41.0% in West Africa (P < 0.001). Free access to second-line therapy was available for 87.8% of sites overall but in significantly fewer sites in West Africa, 57.5% (P < 0.001). Most sites (65.1%) traced children who were lost to follow-up using home visit and telephone calls with, again, substantial regional heterogeneity observed: 81.4% of Asian, 91.5% of East-African, 64.4% of Southern-African, and only 29.9% of West-African sites (P < 0.001, Table 1).

Characteristics of Individual Patients and Antiretroviral Regimens at ART Initiation

At ART initiation, the median age was 5 years (IQR: 2–9); median CD4 percentage was 12% (IQR: 7–18) with 51.8% of children being immunodeficient, and 19.9% having clinical AIDS. NNRTI-based regimens were the most common first-line antiretroviral drug regimens (76.9%), followed by those with a PI (20.1%), with other combinations accounting for the remaining 3%. These baseline patient characteristics differed significantly between regions (Table 1). Median CD4 percentage and age were 7% and 7 years, respectively, in Asia; 12% and 6 years in East Africa; 14% and 4 years in Southern Africa; and 13% and 5 years in West Africa (P < 0.001). More than 90% of the children in Asia and East Africa started ART with an NNRTI and 2 NRTIs, compared with 66.4% of children in Southern Africa and 69.3% in West Africa. The proportions of children initiating ART before 2005 were 40.0% in Asia, 6.4% in East Africa, 21.6% in Southern Africa, and 31.6% in West Africa (P < 0.001).

Mortality and Loss to Follow-Up Rates Across Regions

Among 13,611 children followed up for a median length of 18 months (IQR: 7–32) and contributing 20,417 person-years of follow-up by 18 months after ART initiation, 5.7% of children had died, 12.3% were lost to follow-up, and 8.6% had been transferred out of the clinic (Table 2). Mortality varied from 4.3% in East Africa to 7.4% in West Africa, whereas LTFU varied even more widely from 4.1% in Asia to 21.8% in West Africa (P < 0.001). Transfer-out rates varied from 0.4% of children in East Africa to 16.4% in Southern Africa (Table 2).

TABLE 2-a:
Death, Transfer Out, and Loss to Follow-Up at 18 Months After ART Initiation by IeDEA Region and Cohort (N = 13, 611)
TABLE 2-b:
Death, Transfer Out, and Loss to Follow-Up at 18 Months After ART Initiation by IeDEA Region and Cohort (N = 13, 611)

Overall, using a competing risk model with death and LTFU as competing events, the cumulative incidence rates of death were 4.5% (95% CI: 4.2 to 4.8), 5.5% (95% CI: 5.1 to 5.9), and 6.3% (95% CI: 5.9 to 6.7) at 6, 12, and 18 months, respectively. The estimated crude cumulative incidence of death at 18 months was the highest in West Africa (7.9%, 95% CI: 6.9 to 8.9), followed by Southern Africa (6.2%, 95% CI: 5.6 to 6.8), Asia (5.7%, 95% CI: 4.7 to 7.0), and East Africa (4.9%, 95% CI: 4.2 to 5.8; P < 0.01). When adjusting for age, WHO clinical stage and CD4 at ART initiation, along with type of facility, and cohort size, the risk of death was 7.4% in West Africa, 5.7% in Southern Africa, 5.4% in Asia and 4.3% in East Africa (P = 0.02, Table 3).

Mortality and LTFU Adjusted Analysis Using a Competing Risk Model in 13,611 Children on ART*

The crude estimated 18-month LTFU rate was 23.5% (95% CI: 21.9 to 25.1) in West Africa, 16.4% (95% CI: 14.9 to 17.8) in East Africa, 10.8% (95% CI: 10.0 to 11.7) in Southern Africa, and 4.3% (95% CI: 3.3 to 5.5) in Asia (P < 0.01). When adjusting for age, WHO clinical stage and CD4 at ART initiation along with type of ART regimen, year of ART initiation, type of facility, free access to laboratory tests, and to ARV, nonurban site, and cohort size, the risk of loss to follow-up was 21.8% in West Africa, 14.0% in East Africa, 9.0% in Southern Africa, and 4.1% in Asia (P < 0.01, Table 3).

Factors Associated With the Risk of Dying

The crude estimated 18-month cumulative mortality was significantly higher in children <24 months old at baseline, compared with that in older children (P < 0.01, Fig. 1). It was also significantly higher (10.5%) in children with severe immunodeficiency (CD4 < 10%) at baseline, compared with that in the other groups: 4.7% in those with CD4 (10%–20%), 4.5% in those with CD4 ≥20%, and 5.4% in those with unknown CD4 percentage (P < 0.01, Fig. 2). We observed a significantly higher mortality in clinics with no LTFU search compared with those with any search in univariate analysis (see Table, Supplemental Digital Content 1,

Estimated cumulative incidence curves with death (left panel) and LTFU (right panel) as competing events in 13,611 children on ART, by age group at ART initiation in the IeDEA pediatric collaboration database.
Estimated cumulative incidence curves with death (left panel) and LTFU (right panel) as competing events in 13,611 children on ART, by CD4% at ART initiation in the IeDEA pediatric collaboration database.

The adjusted statistical analysis using a competing risk model (Table 3; see Table, Supplemental Digital Content 1, identified the following individual baseline correlates of higher 18-month risk of dying (asHRs and their 95% CIs) age <12 months (asHR 2.8; CI: 2.2 to 3.5; P < 0.01) and age 1–2 years (asHR 1.5; CI: 1.2 to 2.0, P < 0.01) compared with 10–15 years (the reference group); clinical AIDS or stage-4 disease versus earlier stages (asHR 2.1; CI:1.8 to 2.5, P < 0.01); CD4% < 10% versus ≥20% (asHR 3.0; CI: 2.3 to 3.9, P < 0.01); or missing CD4% (asHR 1.5; CI: 1.2 to 2.0, P < 0.01). Site-specific variables associated with higher mortality were the following: attending a private facility (asHR 1.5; CI: 1.1 to 2.0, P < 0.01) or an unknown type of facility (asHR 2.4; CI: 1.6 to 3.8, P < 0.01) compared with a public sector 1, being followed up in a large size cohort (500–800 children; asHR 1.8; CI: 1.4 to 2.3, P < 0.01) compared with <250 children; receiving care in West Africa, the site of highest mortality (asHR 1.3; CI: 1.0 to 1.7, P = 0.04) compared with East Africa, the region with the lowest mortality rates. Mortality in Southern Africa and Asia did not differ from that in East Africa. No other pairwise regional comparison was statistically significant when adjusting for all other patient-level and site-specific predictors.

Factors Associated With Loss to Follow-Up

The crude estimated 18-month cumulative incidence of LTFU was significantly higher in children <24 months at baseline compared with that in older children (P < 0.01, Fig. 1) and in children with CD4 percentage ≥20% or unknown percentage at baseline, compared with those with CD4 percentage <20% (P < 0.01, Fig. 2). In the adjusted statistical analysis, a number of patient-level and site-specific factors were identified as independent predictors of higher LTFU hazards adjusted for the competing risk of death (Table 3; see Table, Supplemental Digital Content 2, age <12 months versus 10–15 years (asHR 1.6; CI: 1.4 to 2.0, P < 0.01), having received a first-line PI-based ART (asHR 1.4; CI: 1.2 to 1.7, P < 0.01); or another regimen (asHR 1.7; CI: 1.3 to 2.1, P < 0.01) versus an NNRTI-based regimen; having clinical AIDS or stage-4 disease (asHR 1.4; CI: 1.2 to 1.6, P < 0.01), or unknown AIDS staging (asHR 1.7; CI: 1.5 to 2.0, P < 0.01) versus non-AIDS stage, having started ART in 2005–2007 (asHR 2.4; CI: 2.0 to 2.8, P < 0.01), or having started ART after 2007 (asHR 3.4; CI: 2.9 to 4.1, P < 0.01) versus having initiated therapy before 2005, receiving care in a nonurban clinic (asHR 1.7; CI: 1.4 to 1.9, P < 0.01) versus being treated at an urban clinic; having to pay for laboratory tests (asHR 2.6, 95% CI: 1.9 to 3.5, P < 0.01) versus receiving free access to laboratory tests and having to pay for first-line antiretrovirals (asHR 5.0, 95% CI: 3.4 to 7.3, P < 0.01) versus having access to free ART, being a member of a larger cohort (ie, 1 with 500–800 children; asHR 1.8, 95% CI: 1.5 to 2.3, P < 00.1) versus a smaller cohort (<250 children), and receiving care in East Africa (asHR 3.5, CI: 2.6 to 4.7, P < 0.01), or in West Africa (asHR 3.1, CI: 2.2 to 4.3, P < 0.01) versus Asia the region with the lowest rates of LTFU (see Table 5, Supplemental Digital Content 1,

Predictors of a less frequent LTFU were as follows: having CD4% (10%–20%, asHR 0.8; CI: 0.7 to 1.0, P = 0.03) versus ≥20%, receiving care in a private clinic (asHR 0.2 CI: 0.1 to 0.3, P < 0.01) rather than in a public clinic, and cohort size (250–500, asHR 0.8 CI: 0.6 to 0.9, P < 0.01).


As HIV treatment is rapidly scaled up toward universal access in the first decade of the 21st century, the IeDEA multiregional pediatric collaboration provides a unique opportunity to monitor, analyze, and compare children’s outcomes taking into account patient and program-level factors while accounting for regional heterogeneity in these factors, in large-scale care and treatment programs in low- and middle-income countries. This analysis included 13,611 children contributing with 20,417 child-years of follow-up in a large number of pediatric cohorts. The overall cumulative mortality was measured at 5.7%, whereas LTFU rate was 12.3% after 18 months on ART, with marked differences of these estimates between the 4 regions studied: Asia, East Africa, West Africa, and Southern Africa.

There are important observations gleaned from our study. First, the 18-month mortality in children on ART is much lower than the reported mortality estimates among untreated children before the ART era,24,25 arguing for a clear benefit of ART as a pediatric HIV intervention. Our mortality estimates are consistent with those of previous reports of studies conducted among HIV-infected children on ART in resource-constrained countries with reported mortality rates, ranging from 6.3% to 11.5%, usually obtained in small-size cohorts and in the early periods of the ART scale-up.5,7,9–14,26 Despite heterogeneity between regions, our findings consistently reflect the impact of overall delayed ART initiation in children when adjusting for individual, programmatic, and regional factors. Indeed, independent individual predictors of residual mortality on ART (advanced clinical stage of HIV disease, and severe immunosuppression) reflect late access to ART. In addition, most (71%) deaths in this data set occurred within the first 6 months after ART initiation, advocating for earlier and increased access to ART initiation in the pediatric population. These findings are consistent with those reported in previous studies5,7,9–14 and in individual IeDEA regions.8,26–28 A number of site-specific variables also explained part of the observed mortality: A greater cohort size increased mortality risk, possibly reflecting the impact of skilled health care staff shortages or work overload, receiving care in Western Africa independently increased the mortality by 40% compared with in East Africa. Although delayed ART initiation is common in all regions, there were significant differences in site-level factors such as having lower access to free HIV services in Western Africa compared with other regions (laboratory tests, OI prophylaxis, first-line, and second-line ART).

Second, the 18-month LTFU rates were unacceptably high and differed substantially across regions, with significant 3.5-fold and 3.1-fold increased hazards in East Africa and West Africa compared with the risk in Asia. Multiple sources of between-region heterogeneity could explain these higher LTFU rates including variations in (1) operational definitions of LTFU and patient tracing systems, (2) lack of access to free HIV services (laboratory tests, antiretroviral drugs, and OI prophylaxis) in West Africa compared with other regions, (3) failure to accurately capture patient transfers as documented transfer-out rates differed significantly between regions, as very few transfers (<1%) were recorded in East and West Africa compared with 16% in Southern Africa, (4) levels of decentralization of HIV pediatric care out of urban hospitals, and (5) reliability of drug supply chain management for non-NNRTI–based regimens, which are less affordable and accessible in East and West Africa compared with that in Southern Africa. Requiring the patients’ families to shoulder the fees for HIV care and clinical services was associated with higher LTFU rates as has previously been reported in adults.29

In addition to patient-level and site-specific factors, our analysis illustrated the substantial impact of the large-scale ART rollout in lower-income countries over the past decade. Initiation of ART after 2005 and site cohorts with (500–800) children were associated with increased hazard of LTFU, whereas cohorts with <500 children had lower LTFU rates. As antiretroviral treatment decreases mortality, we hypothesize that the number of HIV-infected children treated will continue to increase over time, particularly in regions where transfer-out rates are low, such as in East Africa and West Africa. This substantial increase in the number of HIV-infected children followed in individual health facilities and the associated increased workload may further negatively impact the standard of care. Finally, the political instability and violence crisis could have also increased the rates of LTFU in East and West Africa, as this was recently investigated in Kenya.30

We observed the seemingly contradictory finding that treatment at private clinics (where presumably care is not free) were associated with higher rates of death but lower rates of LTFU compared with being cared at public clinics (more frequently free of charge). Although not explicitly measurable with these data, we suspect that this dichotomy may reflect that record-keeping practices may be better in private settings with a significantly higher mortality rates documented and lower LTFU rates compared with public settings, with a substantial proportion of excess LTFU being unreported mortality.

Limitations must be acknowledged in our study. The mortality rate we observed at 18 months is likely to be underestimated because of the high rates of LTFU, with a proportion of those lost more likely to die soon after defaulting from the program. Indeed, mortality has been commonly underreported in African adult ART programs due to limited access to care and the frequent occurrence of death at home.16 The same finding was recently observed in children in East Africa where mortality and disclosure issues, including fear of family or community discrimination, were the most important reasons why children became lost to follow-up.31 Thus, it would be valuable to further ascertain outcomes at least in a sample of those children lost to follow-up in pediatric programs to correct the mortality rates as proposed in adults.32 Even though this information was available in a small number of cohorts in this study, this was not attempted here so our estimates are likely underestimating overall mortality rates, particularly in regions with high rates of LTFU. This means that pediatric outcomes may be much worse than reported, particularly in regions, such as East and West Africa, with both high mortality and LTFU rates.

When the cohort size is large, as in this case, statistical significance testing of baseline variables may produce highly significant statistical differences, even when they may have little clinical or programmatic implication. For example differences in gender between the regions was <5% (absolute) but was still highly statistically significant.

Data quality is another concern when analyzing such large collaborative data sets is. Lack of data completeness and substantial interstudy heterogeneity may reduce the accuracy of our results compared with studies coming from more homogeneous settings. Also, data on factors related to ART adherence or other social issues such as disclosure and orphan status, which were not generally available in our data, would help to better explain clinical outcomes in HIV-infected children. Smaller studies where these issues were documented have reported that disclosure is associated with ART adherence and needs to be monitored in programs.33 In addition, substantial rates of missing data on a number of target outcomes and their predictors and record-keeping practices have been varied over time and between regions. In the adjusted regression analyses, when missing variables were included as a separate category within variables, this may introduce bias in the effect estimates,34 and this was the reason why we did not included variables if missing data were >70%.

Finally, our study may not be representative of all children on ART in these regions as most of the data were gathered from urban sites, in which the standard of care may be higher than in rural areas. Thus, we hypothesize that the programmatic deficiencies identified here could be worse in less structured clinical settings or at lower levels of the health care system. Nevertheless, we are confident of our main findings, which raise significant operational concerns regarding the impact of both delayed ART start as measured by the high rates of advanced disease at the time of care initiation in our cohort, and the importance of long-term retention to care once enrolled into a program.

Despite the possibly superior level of care in our facilities, and concerns with missing data or quality of data collection, our mortality outcomes confirm that there is still a gap in achieving the standard of implementation of early ART in all HIV-infected children <2 years old as recommended in the 2010 WHO guidelines.35 Increasing early access to ART would decrease early on-ART mortality as demonstrated in the CHER trial.36 This is one of the primary operational challenges in pediatric care in these settings, pointing to the need for earlier identification of HIV-infected children, as soon as possible after birth, followed by expedited linkages from testing to ART initiation.

Retention in HIV care is one of the most important challenges faced by health care workers and HIV-implementing partners as the coverage of HIV care and ART have improved for children in low-income countries.37 In this study, high LTFU rates were associated with rapid scale-up, suggesting problems of adaptation of the health care organization and/or staff shortages. Evolving to smaller decentralized clinics rather than to expanding large single programs and task shifting could result in increased access to ART services and good program outcomes as recently reported in Malawi.38,39 Finally, LTFU was strongly associated with fees for ART services. This predictable finding, already reported in adults29 underscores the urgent need for universal free access to all ART services for children too.

Future research should assess whether health systems are meeting the challenges of providing care linkages between all points on the HIV care continuum from early infant diagnosis, to ART initiation and long-term retention of patients. It is also crucial to better document the causes of loss to program and to propose sustainable approaches to increase retention in HIV pediatric programs in lower-income countries.

With this aim, access to treatment through a family-centered approach should be considered. In this model, primary caregivers on treatment can be sources of continuity, knowledge, and strength for pediatric patients and other HIV-infected family members.40,41

Although data quality of such international data sets should be improved and documentation of clinical outcomes could have been more complete, this multiregional collaborative study offered a unique opportunity to improve our understanding of the rapidly expanding pediatric ART services taking into account both patient-level and site-specific factors. Such analyses should be repeated to explore secular trends in the second decade of wide ART use in these settings. Large-scale ART for children in resource-limited settings is feasible, but innovative and sustainable approaches are urgently required to improve early ART initiation, and retain children in ART programs in lower-income countries.


The authors acknowledge all the children and their families followed up in the participating pediatric centers. They also thank the staff from all participating pediatric centers. They warmly thank all the investigators and pediatric coordinators from the Pediatric IeDEA Regions contributing to the project: Asia (Annette Sohn), East Africa (Kara Wools-Kaloustian), Southern Africa (Mary-Ann Davies), Western Africa (Alain Azondekon), and the IeDEA Pediatric Working Group: Melanie Bacon, Rosemary McKaig, Robin Huebner, and Lori Schwarze. They also thank Denis Nash (Columbia University, Mailman School of Public Health) for helping with the regional site survey and Andrea Ciaranello for her helpful comments on earlier draft. Special thanks to Alioum Ahmadou (ISPED) for his statistical advice.

This study was led by the IeDEA West-Africa Group (principal investigators: François Dabis, Emmanuel Bissagnéné); Bordeaux staff: Eric Balestre, Didier K. Ekouévi, Charlotte Lewden, Valériane Leroy, Karen Malateste, Elodie Rabourdin, Rodolphe Thiebaut. Abidjan staff: Gérard Allou, Jean-Claude Azani, Patrick Coffie, Hughes Djétouan, and Bertin Kouadio.

The International Epidemiological Databases to Evaluate AIDS in West Africa (IeDEA West Africa) is supported by the National Cancer Institute, the Eunice Kennedy Shriver National Institute of Child Health and Human Development, and the National Institute of Allergy and Infectious Diseases as part of the International Epidemiologic Databases to Evaluate AIDS (IeDEA; grant no. 5U01AI069919-01 to 04). The content of this publication is solely the responsibility of the authors and does not necessarily represent the official views of any of the institutions mentioned above. No funding bodies had any role in study design, data collection and analysis, decision to publish, or preparation of the article.


1. World Health Organisation, UN AIDS, UNICEF. Global HIV/AIDS Response: Epidemic Update and Health Sector Progress Towards Universal Access, Progress Report. Geneva, Switzerland: WHO, UNAIDS, UNICEF; 2011.
2. Davies MA, Moultrie H, Eley B, et al.. Virologic failure and second-line antiretroviral therapy in children in South Africa—the IeDEA Southern Africa collaboration. J Acquir Immune Defic Syndr. 2011;56:270–278.
3. O’Brien DP, Sauvageot D, Zachariah R, et al.. In resource-limited settings good early outcomes can be achieved in children using adult fixed-dose combination antiretroviral therapy. AIDS. 2006;20:1955–1960.
4. UNAIDS. Global report. UNAIDS report on the global AIDS epidemic. Geneva, Switzerland: UNAIDS; 2010.
5. Anaky MF, Duvignac J, Wemin L, et al.. Scaling up antiretroviral therapy for HIV-infected children in Cote d'Ivoire: determinants of survival and loss to programme. Bull World Health Organ. 2010;88:490–491.
6. Arrive E, Kyabayinze DJ, Marquis B, et al.. Cohort profile: the paediatric antiretroviral treatment programmes in lower-income countries (KIDS-ART-LINC) collaboration. Int J Epidemiol. 2008;37:474–480.
7. Bolton-Moore C, Mubiana-Mbewe M, Cantrell RA, et al.. Clinical outcomes and CD4 cell response in children receiving antiretroviral therapy at primary health care facilities in Zambia. JAMA. 2007;298:1888–1889.
8. Davies MA, Keiser O, Technau K, et al.. Outcomes of the South African National Antiretroviral Treatment Programme for children: the IeDEA Southern Africa collaboration. S Afr Med J. 2009;99:730–737.
9. Fassinou P, Elenga N, Rouet F, et al.. Highly active antiretroviral therapies among HIV-1-infected children in Abidjan, Cote d’Ivoire. AIDS. 2004;18:1905–1913.
10. George E, Noel F, Bois G, et al.. Antiretroviral therapy for HIV-1-infected children in Haiti. J Infect Dis. 2007;195:1411–1418.
11. Puthanakit T, Oberdorfer A, Akarathum N, et al.. Efficacy of highly active antiretroviral therapy in HIV-infected children participating in Thailand’s National Access to Antiretroviral Program. Clin Infect Dis. 2005;41:100–107.
12. Reddi A, Leeper SC, Grobler AC, et al.. Preliminary outcomes of a paediatric highly active antiretroviral therapy cohort from KwaZulu–Natal, South Africa. BMC Pediatr. 2007;7:13.
13. Edmonds A, Yotebieng M, Lusiama J, et al.. The effect of highly active antiretroviral therapy on the survival of HIV-infected children in a resource-deprived setting: a cohort study. PLoS Med. 2011;8:e1001044.
14. Sauvageot D, Schaefer M, Olson D, et al.. Antiretroviral therapy outcomes in resource-limited settings for HIV-infected children <5 years of age. Pediatrics. 2010;125:e1039–e1047.
15. Anglaret X, Toure S, Gourvellec G, et al.. Impact of vital status investigation procedures on estimates of survival in cohorts of HIV-infected patients from Sub-Saharan Africa. J Acquir Immune Defic Syndr. 2004;35:320–323.
16. Brinkhof MW, Pujades-Rodriguez M, Egger M. Mortality of patients lost to follow-up in antiretroviral treatment programmes in resource-limited settings: systematic review and meta-analysis. PLoS ONE. 2009;4:e5790.
17. Williams PL, Van Dyke R, Eagle M, et al.. Association of site-specific and participant-specific factors with retention of children in a long-term pediatric HIV cohort study. Am J Epidemiol. 2008;167:1375–1386.
18. World Health Organisation. HIV prevention and treatment guidelines. Antiretroviral therapy of HIV infection in infants and children in resource-limited settings, towards universal access: recommendations for a public health approach (2006 version). Geneva, Switzerland: WHO; 2006.
19. Nash D, Farr A, McKaig R, et al.. Characteristics of HIV care and treatment programs in the International epidemiologic Databases to Evaluate AIDS (IeDEA) collaboration. TULBPE053 Sixth IAS Conference on HIV Pathogenesis, Treatment and Prevention, Rome, Italy, July 17–20, 2011.
20. Putter H, Fiocco M, Geskus RB. Tutorial in biostatistics: competing risks and multi-state models. Stat Med. 2007;26:2389–2430.
21. Scrucca L, Santucci A, Aversa F. Competing risk analysis using R: an easy guide for clinicians. Bone Marrow Transplant. 2007;40:381–387.
22. Scrucca L, Santucci A, Aversa F. Regression modeling of competing risk using R: an in depth guide for clinicians. Bone Marrow Transplant. 2010;45:1388–1395.
23. Hosmer D, Lemeshow S. Model-building strategies and methods for logistic regression. Applied Logistic Regression. 2nd ed. Wiley & sons; 2000: 92–99.
24. Newell ML, Coovadia H, Cortina Borja M, et al.. Mortality of infected and uninfected infants born to HIV-infected mothers in Africa: a pooled analysis. Lancet. 2004;364:1236–1243.
25. The Cross Continents Collaboration for Kids. Markers for predicting mortality in untreated HIV-infected children in resource-limited settings: a meta-analysis. AIDS. 2008;22:97–105.
26. Fenner L, Brinkhof MW, Keiser O, et al.. Early mortality and loss to follow-up in HIV-infected children starting antiretroviral therapy in Southern Africa. J Acquir Immune Defic Syndr. 2010;54:524–532.
27. Ekouevi DK, Azondekon A, Dicko F, et al.. 12-Month mortality and loss-to-program in antiretroviral-treated children: the IeDEA pediatric West African Database to evaluate AIDS (pWADA), 2000–2008. BMC Public Health. 2011;11:519.
28. Lumbiganon P, Kariminia A, Aurpibul L, et al.. Survival of HIV-infected children: a cohort study from the Asia-Pacific region. J Acquir Immune Defic Syndr. 2011;56:365–371.
29. Brinkhof MW, Dabis F, Myer L, et al.. Early loss of HIV-infected patients on potent antiretroviral therapy programmes in lower-income countries. Bull World Health Organ. 2008;86:559–567.
30. Yoder RB, Nyandiko WM, Vreeman RC, et al.. Long-term impact of the Kenya postelection crisis on clinic attendance and medication adherence for HIV-infected children in western Kenya. J Acquir Immune Defic Syndr. 2012;59:199–206.
31. Braitstein P, Songok J, Vreeman R, et al.. ‘Wamepotea’ (they have become lost): outcomes of HIV-positive and HIV-exposed children lost to follow-up from a large HIV treatment program in western Kenya. J Acquir Immune Defic Syndr. 2011;57:e40–e46.
32. Egger M, Spycher BD, Sidle J, et al.. Correcting mortality for loss to follow-up: a nomogram applied to antiretroviral treatment programmes in sub-Saharan Africa. PLoS Med. 2011;8:e1000390.
33. Vreeman RC, Nyandiko WM, Ayaya SO, et al.. The perceived impact of disclosure of pediatric HIV status on pediatric antiretroviral therapy adherence, child well-being, and social relationships in a resource-limited setting. AIDS Patient Care STDS. 2010;24:639–649.
34. Greenland S, Finkle WD. A critical look at methods for handling missing covariates in epidemiologic regression analyses. Am J Epidemiol. 1995;142:1255–1264.
35. World Health Organisation. Antiretroviral drugs for treating pregnant women and preventing HIV infection in infants. Recommendations for a public health approach. Geneva, Switzerland: WHO; 2010.
36. Violari A, Cotton MF, Gibb DM, et al.. Early antiretroviral therapy and mortality among HIV-infected infants. N Engl J Med. 2008;359:2233–2244.
37. Geng EH, Nash D, Kambugu A, et al.. Retention in care among HIV-infected patients in resource-limited settings: emerging insights and new directions. Curr HIV/AIDS Rep. 2010;7:234–244.
38. Bemelmans M, van den Akker T, Ford N, et al.. Providing universal access to antiretroviral therapy in Thyolo, Malawi through task shifting and decentralization of HIV/AIDS care. Trop Med Int Health. 2010;15:1413–1420.
39. McCollum ED, Preidis GA, Kabue MM, et al.. Task shifting routine inpatient pediatric HIV testing improves program outcomes in urban Malawi: a retrospective observational study. PLoS ONE. 2010;5:e9626.
40. Reddi A, Leeper SC. Antiretroviral therapy adherence in children: outcomes from Africa. AIDS. 2008;22:906–907.
41. Leeper SC, Montague BT, Friedman JF, et al.. Lessons learned from family-centred models of treatment for children living with HIV: current approaches and future directions.J Int AIDS Soc. 2010;13 (suppl2):S3.


The IeDEA West Africa Working Group is organized as follows: Primary Investigators: Pr François Dabis* (INSERM U897, ISPED, Bordeaux, France), Pr Emmanuel Bissagnene* (SMIT, CHU de Treichville, Abidjan, Côte d’Ivoire). Clinical Investigators by country and alphabetical order (*Member of the IeDEA West Africa Technical Committee): Jocelyn Akakpo, Alain Azondékon, Jules Bashi, Sagbo Gratien, Sikiratou Koumakpaï, Marcel D. Zannou* (Benin); Ye Diarra, Eric-Arnaud Diendere, Joseph Drabo,* Fla Koueta (Burkina Faso); Edmond Aka-Addi, Clarisse Amani-Bosse, Franck-Olivier Ba-Gomis, François Eboua-Tanoh, Serge-Paul Eholie,* Calixte Guehi, Kouakou Kouadio, Serge-Olivier Koulé, Eugène Messou, Albert Minga, Aristophane Tanon, Marguerite Timité-Konan, Pety Touré (Côte d’Ivoire); Kevin Peterson* (Gambia); Bamenla Goka, Lorna Renner* (Ghana); Hadizatou Coulibaly, Fatoumata Dicko, Moussa Maiga,* Daouda Minta, Mariam Sylla, Hamar Alassane Traoré; Man Charurat* (Man Charurat); Bernard Diop, Fatou Ly Ndiaye, Papa Salif Sow, Haby Signaté Sy,* Judicaël Tine (Senegal). Epidemiology and Statistical Unit (INSERM U897, ISPED, Université Victor Segalen, Bordeaux, France): Eric Balestre, Didier K. Ekouévi,* Antoine Jaquet,* Valériane Leroy,* Charlotte Lewden,* Karen Malateste, Annie Sasco, Rodolphe Thiebaut. Data Management Unit (PACCI, CHU Treichville, Abidjan, Côte d’Ivoire): Gérard Allou, Jean Claude Azani, Patrick Coffie. Pediatric clinical centers by city and country: Abidjan, Côte d’Ivoire: ACONDA-CEPREF, ACONDA-MTCT-Plus, CHU de Yopougon, Centre Intégré de Recherche Bioclinique d’Abidjan (CIRBA). Accra, Ghana: Korle Bu Teaching Hospital, Bamako, Mali: Hôpital Gabriel Touré. Cotonou, Benin: Centre National Hospitalo-Universitaire Hubert Maga, Hôpital d’Instruction des Armées. Dakar, Senegal: Hôpital d’Enfants Albert-Royer. Fajara, Gambia: Medical Research Council. Ouagadougou, Burkina-Faso: Centre Hospitalier Charles de Gaulle. Administration: Alexandra Doring and Elodie Rabourdin (ISPED), Hughes Djétouan, Bertin Kouadio, and Adrienne Kouakou (PACCI).

The IeDEA Southern Africa Steering Group is organized as follows: Member Sites (sites denoted with an asterisk contributed data to this analysis): *Cleophas Chimbetete, Newlands Clinic, Harare, Zimbabwe; Anna Coutsoudis, PMTCT Plus, Cato Manor, Durban; Diana Dickinson, Gaborone Independent Hospital, Gaborone, Botswana; *Brian Eley, Red Cross Children’s Hospital, Cape Town, South Africa; Lara Fairall, Free State Provincial ART Program, South Africa; Christiane Fritz, SolidarMed Zimbabwe, Zimbabwe;*Daniella Garone, Khayelitsha ART Program and Médecins Sans Frontières, Cape Town, South Africa; *Janet Giddy, McCord Hospital, Durban, South Africa; Christopher Hoffmann, Aurum Institute for Health Research, South Africa; Timothy Meade, CorpMed Clinic, Lusaka, Zambia; Patrick MacPhail, Themba Lethu Clinic, Helen Joseph Hospital, Johannesburg, South Africa; Lerato Mohapi, Perinatal HIV Research Unit, Johannesburg, South Africa; *Harry Moultrie, Wits Institute for Sexual Reproductive Health, HIV and Related Diseases, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, and Harriet Shezi Children’s Clinic, Chris Hani Baragwanath Hospital, Soweto, South Africa; James Ndirangu, Hlabisa HIV Treatment and Care Program, South Africa; Sabrina Pestilli, SolidarMed Mozambique, Mozambique; *Sam Phiri, Lighthouse Clinic, Lilongwe, Malawi; *Hans Prozesky, Tygerberg Academic Hospital, Stellenbosch, South Africa; Jeff Stringer, Center for Infectious Disease Research in Zambia, Zambia; *Karl Technau, Empilweni Service and Research Unit, Rahima Moosa Mother and Child Hospital, University of the Witwatersrand, Johannesburg, South Africa; *Paula Vaz, Paediatric Day Hospital, Maputo, Mozambique; *Robin Wood, Gugulethu and Masiphumelele ART Programs and Desmond Tutu HIV Centre, Cape Town, South Africa. Central Team: Matthias Egger, Claire Graber, Fritz Kaeser, Olivia Keiser, Institute of Social and Preventive Medicine, University of Bern, Bern, Switzerland; Andrew Boulle, Morna Cornell, Mary-Ann Davies, Nicola Maxwell, School of Public Health and Family Medicine, University of Cape Town, Cape Town, South Africa.

The TREAT Asia Pediatric Network of IeDEA Asia-Pacific is organized as follows: V. Saphonn,* S. Saramony, National Centre for HIV/AIDS Dermatology and STDs, Phnom Penh, Cambodia; U. Vibol,* S. Sophan, National Pediatric Hospital, Phnom Penh, Cambodia; J. Tucker, New Hope for Cambodian Children, Phnom Penh, Cambodia; F.J. Zhang, Beijing Ditan Hospital, Capital Medical University, Beijing, China; N. Kumarasamy,* S. Saghayam, Y.R. Gaitonde Centre for AIDS Research and Education, Chennai, India; N. Kurniati,* D. Muktiarti, Cipto Mangunkusumo General Hospital, Jakarta, Indonesia; SM Fong,* M. Thien, Hospital Likas, Kota Kinabalu, Malaysia; N.K. Nik Yusoff,* L.C. Hai, Hospital Raja Perempuan Zainab II, Kelantan, Malaysia; K.A. Razali,* J.M. Thahira, N.F. Abdul Rahman, Pediatric Institute, Hospital Kuala Lumpur, Kuala Lumpur, Malaysia; R. Nallusamy,* K.C. Chan, Penang Hospital, Penang, Malaysia; V. Sirisanthana,* L. Aurpibul, Chiang Mai University, Chiang Mai, Thailand; R. Hansudewechakul,* P. Taeprasert, Chiangrai Prachanukroh Hospital, Chiang Rai, Thailand; P. Lumbiganon,* P. Kosalaraksa, P. Tharnprisan, Khon Kaen University, Khon Kaen, Thailand; G. Jourdain, Program for HIV Prevention and Treatment, Chiang Mai, Thailand; J. Ananworanich,* C. Phasomsap, T. Suwanlerk, The HIV Netherlands, Australia, Thailand Research Collaboration (HIV-NAT), Bangkok, Thailand; K. Chokephaibulkit,* W. Phongsamart, O. Wittawatmongkol, Siriraj Hospital, Mahidol University, Bangkok, Thailand; H.K. Truong,* D.A.N. Mai, Children’s Hospital 1, Ho Chi Minh City, Vietnam; C.V. Do,* M.T. Ha, Children’s Hospital 2, Ho Chi Minh City, Vietnam; B.V. Huy,* V.L. Nguyen, National Hospital of Pediatrics, Hanoi, Vietnam; N.O. Le, Worldwide Orphans Foundation, Ho Chi Minh City, Vietnam; A.H. Sohn,* N. Durier, J. Pang, TREAT Asia, amfAR—The Foundation for AIDS Research, Bangkok, Thailand; D.A. Cooper, M.G. Law,* A. Kariminia, The Kirby Institute, University of New South Wales, Sydney, Australia (*Steering Committee members).

The IeDEA East Africa Steering Group is organized as follows: Dr. S. Ayaya: the United States Agency for International Development-Academic Model Providing Access To Health Care (USAID-AMPATH) Program; Dr. Lyamuya: Morogoro; Dr. Maruchu: Tumbi; Dr. Bukusi and Dr. Cohen: Family AIDS Care and Education Services; BS. Musick, C.T. Yiannoutsos, and E. Sang: East African Regional Data Center.


antiretroviral therapy; children; cohort studies; HIV infection; mortality; loss to follow-up; low-income countries; Asia; Africa

Supplemental Digital Content

© 2013 Lippincott Williams & Wilkins, Inc.