HIV/AIDS continues to have a profound effect on the health of children worldwide. Despite advances in prevention of mother-to-child HIV transmission (MTCT), an estimated 330,000 children become infected through MTCT in resource-limited settings (RLSs) each year.1
In the United States and Europe, the effects of HIV on pediatric growth, morbidity, and mortality have been studied extensively among both HIV-infected and exposed uninfected children through prospective perinatal cohort studies. These studies have longitudinally tracked the growth and development, complications of HIV and treatment, hospitalizations, quality of life, and survival of children born to HIV-infected women,2–6 before and after the availability of potent combinations of pediatric antiretroviral treatment (ART).
However, in RLSs, with the largest pediatric HIV burden, there is a paucity of literature addressing the long-term growth and survival of infants born to HIV-infected women, including whether there are any late sequelae of exposure to perinatal antiretroviral (ARV) interventions. The limited numbers of published research studies have focused primarily on comparisons of infant morbidity and mortality in children aged younger than 36 months born to HIV-infected mothers.7–11
The HIVNET 012 clinical trial,9 which followed up HIV-exposed infants from birth to the age 18 months, and its companion rollover protocol, which followed up participant children from 24 months up to age 5 years, provided a unique opportunity to address longer-term growth, morbidity, and survival and to assess potential late sequelae from short peripartum ARV exposure. The overall aim of this analysis was to compare the long-term growth and survival among the HIV-infected and -uninfected children in the HIVNET 012 cohorts during a time period when ART was not widely available. In addition, we examined the most common causes of hospitalizations in HIV-uninfected and -infected infants. Finally, we monitored for any late sequelae over the first 5 years of life among children born to mothers in the short course zidovudine (ZDV) compared with the nevirapine (NVP) study arms of HIVNET 012.
HIVNET 012 was a phase IIB randomized trial conducted to evaluate the safety and efficacy of peripartum NVP or ZDV in HIV-infected Ugandan women and their infants for prevention of MTCT. The study design, methods, and outcomes were previously reported.9 Longitudinal data were collected prospectively on a cohort of mother–infant pairs enrolled in the primary HIVNET 012 study from pregnancy through the age 18 months. Additional data were collected prospectively from HIVNET 012 participants who consented and enrolled in a rollover extended follow-up observational study of children from the age 24 to 60 months. The Ugandan and Johns Hopkins institutional review boards approved both the primary and the extended follow-up protocols.
The extended follow-up study was conducted at the Makerere University–Johns Hopkins University Research Clinic in Kampala, Uganda from November 1999 to June 2004. This analysis includes all first-born HIVNET 012 infants followed up from birth through the age 18 months in the primary study and those subsequently enrolled and followed up in the extended follow-up study.
Children born to HIV-infected mothers in HIVNET 012 trial were prospectively followed up for 18 months to determine drug safety, HIV infection rate, and mortality.9 Children enrolled in the follow-up study were seen every 6 months from the age 24 months through 60 months. At each visit, a history was obtained including breastfeeding status, current and interim illnesses, interim medications, and serious adverse events (SAEs), and a routine physical examination was performed. The examination also included anthropometric measurements (weight, height, and head circumference) and neurologic assessment and neurodevelopmental screening, each done by trained staff using standardized procedures. Weight was measured using a balance beam scale, and height was measured using either a wooden stadiometer or height rod (0.1 cm scale). Laboratory evaluations were performed in HIV-infected infants only including complete blood count, CD4 cell count, and HIV-RNA polymerase chain reaction at each visit. The neurodevelopmental screening was done every 6 months at follow-up visits using the Denver Developmental Screening Test that evaluated 4 developmental domains: gross motor, fine motor, social contact, and language skills. The HIVNET 012 children’s performance was scored against the Denver Developmental Screening Test age and gender-normed age performance ranges as either pass or fail.
All study participants had access to free well and sick child care, growth monitoring, and diagnosis and treatment of illnesses at the research clinic. Cotrimoxazole prophylaxis was provided to all HIV-infected children and to 18.44% (92/499) of HIV-exposed infants during the first year of life. Combination pediatric ART was not available in Uganda during most of the period of follow-up. However, 3 children were started on ART in the final year of the 5-year follow-up. Study doctors collected all information during clinic visits using standardized medical forms. Children requiring hospitalization were referred to the Mulago National Referral Hospital. During hospitalizations, study staff abstracted pertinent clinical information, laboratory data, and discharge diagnoses from hospital records using a standardized data collection form. Attempts to determine cause of death through family interviews were done for all participants who died at home during the 60-month follow-up period.
Infant Growth Analysis
Weight, height, weight-for-height, head circumference, and body mass index were used to assess infant growth. Each outcome was standardized (Z score) for age and gender using World Health Organization (WHO) normal values.12
To estimate the effect of HIV-1 infection on infant growth through the age 5 years, for each standardized growth outcome, we included trend for infant age, HIV infection status (time dependent), and trend for time since HIV infection, allowing a random intercept for each infant. Similar analyses were conducted for comparison of randomization arms (NVP versus ZDV).
Kaplan–Meier methods were used to compute the 5-year survival rates of infected and uninfected infants. Because 79% of HIV infections occurred by the age 8 weeks and follow-up time was through 5 years, infants were classified as infected versus uninfected, regardless of the actual time of infection. In a second analysis, infected infants were divided into 3 groups—identified as infected at birth, infected after birth up to 8 weeks, and infected after the age of 8 weeks. Cox proportional hazards regression was used to estimate the hazard ratio for death for HIV-infected versus -uninfected infants. Infection status was defined as a time-dependent covariate, and adjusted for baseline maternal log10 viral load.
The frequency of infant causes of hospitalization and death were compared using χ2 tests. An infant could contribute to multiple categories, but each infant was only counted once within a category. Hematology results between ZDV and NVP arms were compared using t tests at each visit for infected infants.
Long-Term Safety Between 2 and 5 Years
The outcomes of hospitalizations, SAE, neurologic examinations and diagnoses, Denver developmental assessments, and mortality between the ZDV and NVP arms were compared using χ2 test and log rank tests.
All analyses were conducted using SAS version 9 (SAS Institute, Inc., Cary, NC).
There were 651 births in HIVNET 012, of which 627 were included in this analysis (Fig. 1). Four infants were stillborn, 13 were not first-born infants, and 7 were excluded because their HIV status was unknown. Of these 627, 128 became HIV infected before the age 18 months: 59 infected at birth, 42 after birth but before the age 8 weeks, and 27 after the age 8 weeks. At 18 months, 528 (84%) children were alive, and in follow-up, of which 491 (93%) were consented (426 uninfected and 65 infected) and enrolled in the 5-year follow-up protocol. Of these, 18 children died, 26 were lost to follow-up, and 447 completed follow-up to the age 5 years, giving 91% retention in the extended follow-up study. The median duration of breastfeeding was 9 months [95% confidence interval (CI): 8.8 to 10.3] for ZDV group and 8.8 months (95% CI: 7.9 to 9.7) for NVP group. There were 308 children randomized to the NVP arm, 302 to the ZDV arm, and 17 to the placebo. The mean birth weight of the uninfected children was 3.15 kg (SE, 0.019 kg) compared with 3.02 kg (SE, 0.64 kg) for the children infected at birth. There were no significant differences in infant gender, birth weight, and maternal age based on infant infection status.
Infant Growth Analysis
Mean growth in HIV-infected infants was significantly lower than uninfected infants for weight, height, and head circumference over the 5-year follow-up (Fig. 2). On average, an HIV-infected infant weighed 0.15 SDs less than an uninfected infant, in the first year after becoming infected (P < 0.0001). In the same period, they were 0.2 SD shorter (P < 0.0001). Our model predicts that at the age of 5 years, the weight of a study participant infant infected at birth would be 1.6 SDs less than the WHO normal, whereas an uninfected infant would be 0.9 SD less.
Table 1 shows the proportion of HIV-infected and -uninfected children with weight for age (WAZ) and height for age (HAZ) below −2 and −3 SDs at different time points. By the age of 12 months, 14% HIV-infected children had WAZ below −3 SD compared with 5.4% in the uninfected children (P = 0.0040). This difference peaked at 3 years with 14.3% HIV-infected children with WAZ below −3 SD compared with 1.1% uninfected children (P = 0.0072) and then steadily declined. The infected children were also significantly shorter than the uninfected children from 6 months through 5 years (6 months, P = 0.0020; 60 months, P = 0.0019). In regard to weight for height (WAH) Z scores, HIV-infected children had significantly lower WAH Z scores at 12 and 18 months: at 12 months, 7% HIV-infected children had evidence of wasting with WAH Z scores below −3 SD compared with 1.7% uninfected children (P = 0.0049). This difference peaked at 18 months with 7.3% HIV infected children with WAH Z scores below −3 SD compared with 0.5% uninfected children (P < 0.0001) and then steadily declined.
There was no difference between infants assigned to NVP or ZDV in any growth outcomes after adjusting for HIV-1 infection status (data not shown). Assessment of secondary microcephaly identified 9 infants with head circumference less than −3 SD, 7 infants randomized to NVP and 2 infants in the ZDV arm, hazard ratio = 3.2 (95% CI: 0.7 to 15.5), P = 0.14.
The 5-year survival in the HIV-exposed uninfected children was 93% (95% CI: 90% to 95%) compared with 43% (95% CI: 35% to 53%) in the HIV-infected infants (P < 0.0001) (Fig. 3A). Stratifying by time of infection, infants infected at birth and those after birth but before the age 8 weeks show a 5-year survival proportion of 39% (95% CI: 26% to 57%) and 39% (95% CI: 28% to 54%), respectively, whereas those infected after the age 8 weeks had a 61% (95% CI: 44% to 83%) survival (Fig. 3B). More than 50% deaths in the infected children occurred within the first 2 years of life. If an infected child survives to age 2 years, the chance of surviving to age 5 years were greater than 70%, irrespective of time of infection.
The leading cause of death during the 5-year period for the infected children was diarrhea and its complications (see Table S1a, Supplemental Digital Content, http://links.lww.com/QAI/A473), accounting for 42.9% deaths, followed by pneumonia (31.4%). In the uninfected children, malaria (22.9%) and diarrhea (22.9%) combined contributed to close to 50% deaths. There was no difference in the 3-year (99%) and 5-year (98%) survival in the uninfected children in ZDV and NVP arms (P = 0.90).
Cotrimoxazole Exposure and Infant Mortality
There was no difference in the survival between HIV-uninfected infants exposed to cotrimoxazole (93.5%) or not exposed to cotrimoxazole (94.8%) (P = 0.73).
There were significant differences in the proportion of infected and uninfected children hospitalized for each of the leading 5 causes of hospitalization (see Table S1b, Supplemental Digital Content, http://links.lww.com/QAI/A473). Malaria was the overall leading cause of hospital admission (21.9%). For the infected children, however, pneumonia was the leading cause of hospitalization (41.4%), followed by malaria (30.5%) and diarrhea (29.7%). Among the uninfected children, the leading causes of hospitalizations were malaria (19.6%), anemia (9.2%), and pneumonia (8.2%). There were no differences in the proportion of children ever hospitalized between the ZDV and the NVP arms (15.1% versus 12.7%; P = 0.46).
There were no significant differences in hemoglobin concentrations, platelet counts, absolute neutrophil counts, and CD4 cell counts in the HIV-infected infants in the ZDV and NVP arms throughout the 60-month period.
Developmental and Neurologic Evaluations
There were no differences between the proportion of children who passed the Denver Developmental Assessment evaluations in the 2 study arms, which were done at every 6-month visit, and based on a combined analyses of all serial neurodevelopmental assessments: gross motor assessment, 82.7% versus 84.8% for ZDV versus NVP (P = 0.55); fine motor assessment, 77.3% versus 80.1% for ZDV versus NVP (P = 0.50); social assessment, 86.3% versus 84.55% for ZDV versus NVP (P = 0.599); and language assessment, 7.94% versus 5.78% for ZDV versus NVP (P = 0.38). Similarly, there were no statistical differences between the study arms in the proportion of children who experienced a neurologic abnormality between 2 and 5 years (3.66% versus 3.96% for ZDV versus NVP; P = 0.87).
This study examined growth patterns, long-term safety of peripartum antiretroviral prophylaxis, morbidity, and 5-year survival of children enrolled in the HIVNET 012 trial. It is one of the few studies13 that has monitored the long-term growth and survival patterns of children born to HIV-infected mothers in RLSs in the era before pediatric ART was widely available. It is also unique in being the only study to date that has compared longitudinal growth patterns across treatment arms among children perinatally exposed to peripartum NVP and ZDV, the most commonly used ARVs for perinatal prevention in RLSs. Of importance, we found sustained lower weight and height compared with WHO growth standards in exposed but uninfected infants and in infected children. In contrast, normal brain growth based on head circumference measurements was preserved among the uninfected children over the 5-year follow-up. Not surprisingly, we found substantially lower weight, height, and head circumference among infected relative to uninfected children in this cohort.
The consistent poor growth through age 5 years in both HIV-infected and -uninfected children in the HIVNET 012 cohort likely reflects the high background rates of early childhood malnutrition, growth faltering, and stunting seen in RLSs such as Uganda.14 The pathogenesis of the poor growth is most likely multifactorial with widespread poverty, low calorie weaning foods, unhygienic food preparation, and feeding practices leading to diarrheal disease. Among the HIV-infected children, rapid HIV disease progression could also have contributed to the early severe malnutrition observed between the ages 12 and 24 months. The longitudinal growth patterns between the Ugandan-infected versus -uninfected children in the HIVNET 012 cohort showed similar patterns to cohorts of perinatally HIV-infected children in European and the US cohorts with infected children having lower Z scores for all parameters.15
Five-year survival among HIV-exposed uninfected children in the HIVNET 012 follow-up study cohort was significantly higher than that of the infected children (93% versus 43%) and slightly higher than overall Uganda national vital statistics for under 5-year survival (91% in 2011).16 The higher HIVNET 012 5-year survival of the uninfected children compared with these Uganda national data is encouraging and demonstrates that comprehensive medical care and close monitoring of children born to HIV-infected women in RLS can substantially reduce pediatric mortality. Our study participants had access to medications for acute medical conditions and cotrimoxazole prophylaxis for a proportion of exposed infants during the first year of life. This cheap and cost-effective intervention could also have contributed to the high survival rates in the uninfected children in HIVNET 012, potentially related to reductions in malaria and diarrheal diseases. A study done in Uganda found a protective role of cotrimoxazole in reducing malarial infections and improving survival in HIV-affected families.17
The poor survival of the HIV-infected children in the HIVNET 012 studies reflects an era when antiretroviral treatment (ART) was generally unavailable in Uganda. This low survival is similar to rates seen in other African studies of infected children in the pre-ART treatment era7–11 but contrasts with the 5-year survival among perinatally HIV-infected children in the United States and Europe, which was close to 70% before the availability of ART.18–20 Increased vulnerability of infected children to pneumonia, malaria, and diarrhea also contributed to the higher mortality seen among this group. The finding that children who became infected during breastfeeding compared with in utero or at birth had better survival is of interest. We hypothesize that this may be related to more mature immune response repertoire to the HIV among older infants and/or protective effects afforded by maternal humoral or other protective factors in breast milk.
The low mortality in uninfected children and the lack of differences in hospitalizations, SAE, neurologic examinations, and Denver Developmental assessment between the 2 drug arms that have different toxicity profiles are reassuring concerning the long-term safety of short-course peripartum ARV exposure. There were also no significant differences in the hematologic parameters in the HIV-infected children exposed to peripartum NVP and ZDV.
Important strengths of the HIVNET 012 follow-up study include its prospective nature, extremely low loss to follow-up over the 5-year period of the study; documentation of SAEs in a standardized manner, and the longitudinal physical growth measurements obtained by trained staff using calibrated measuring equipment.
There were certain limitations to the data available in the HIVNET 012 follow-up study. Although follow-up was generally high, 7% children from the original HIVNET 012 study were not available for the follow-up study. In addition, hospitalization and outpatient death diagnoses were based mainly on clinical diagnosis, as no autopsies were available to look more formally at specific causes of death. We also assumed that cotrimoxazole prescribed and dispensed to infants was ingested by the participants. Finally, complete blood count and CD4 cell counts were only available for the HIV-infected children after the age 18 months.
The growth, morbidity, and mortality findings in this study emphasize the heightened vulnerability of the HIV-uninfected infants and HIV-infected children born to HIV-infected women. The growth findings highlight the need for early nutritional monitoring, counseling, and support for HIV-exposed infants in RLS to reduce mortality and the risk of failure to thrive. Nutritional supplementation should be directed toward those with growth faltering or overt undernutrition. There is a critical need for early identification of the HIV-infected infants in the first several months of life, followed by rapid introduction of ARVs for infected infants to reduce the high mortality of the HIV-infected children.21
Given the availability of donor funding for both pediatric and adult ARV and increased availability of polymerase chain reaction diagnostics for early infant HIV diagnosis, there is cause for optimism that major improvements in under 5-year survival can be achieved in the next several years among the HIV-infected children in RLS such as Uganda. Early detection and aggressive management of pneumonia and diarrhea in HIV-infected children will also go a long way in reducing mortality in this population.
Careful attention must also be given to providing comprehensive long-term quality care to HIV-exposed uninfected children, who represent an extremely vulnerable population of children. Maternal child health clinics in high prevalence RLS should (1) emphasize early identification of HIV exposed but uninfected infants; (2) train clinic staff to routinely carry out low cost growth monitoring including the plotting and review of growth charts; (3) provide early nutritional interventions for uninfected infants and children who demonstrate a drop off in their growth parameters; (4) ensure high levels of immunization through regular attendance at well child clinics; (5) offer cotrimoxazole to all HIV-exposed infants in the first year of life; and (6) provide timely treatment for acute pediatric illnesses such as malaria, diarrhea, and pneumonia, which contribute significantly to morbidity and mortality in this group. To achieve this, international donor funding in concert within country resources will be required to strengthen the overall maternal child health infrastructure in most RLSs.
2. Moye J, Rich KC, Kalish LA, et al.. Natural history of somatic growth in infants born to women infected with human immunodeficiency virus. J Pediatr. 1996;128:58–69.
3. Bertolli J, Hsu HW, Sukalac T, et al.. Hospitalization trends among children
and youths with perinatal human immunodeficiency virus infection 1990–2002. Pediatr Infect Dis J. 2006;25:628–633.
4. Tovo PA, de Martino M, Gabiano C, et al.. Prognostic factors and survival in children
with perinatal HIV-1 infection. The Italian Register for HIV-1 Infections in Children
. Lancet. 1992;339:1249–1253.
5. European Collaborative Study. Fluctuations in symptoms in HIV infected children
: The first 10 years of life. Pediatrics. 2001;108:116–122.
6. Blanche S, Tardieu M, Duliege A, et al.. Longitudinal study of 94 symptomatic infants with perinatally acquired human immunodeficiency virus infection. Evidence for a bimodal expression of clinical and biological symptoms. Am J Dis Child. 1990;144:1210–1215.
7. Taha TE, Kumwenda NI, Broadhead RL, et al.. Mortality after the first year of life among human immunodeficiency virus type1-infected and uninfected children
. Pediatr Infect Dis J. 1999;18:689–694.
8. Taha TE, Graham SM, Kumwenda NI, et al.. Morbidity among human immunodeficiency virus-1-infected and -uninfected African children
. Pediatrics. 2000;106:.
9. Jackson JB, Musoke P, Fleming T, et al.. Intrapartum and single dose nevirapine compared with zidovudine for prevention of mother to child transmission of HIV- 1 in Kampala, Uganda: 18 months follow up of the HIVNET 012
randomised trial. Lancet. 2003;362:859–868.
10. Mbori-Ngacha D, Nduati R, John G, et al.. Morbidity and mortality in breast fed and formula fed infants of HIV-1 infected women. JAMA. 2001;286:2413–2420.
11. Newell ML, Coovadia H, Cortina-Borja M, et al.; Ghent International AIDS Society (IAS) Working Group on HIV Infection in Women and Children
. Mortality of infected and uninfected infants born to HIV infected mothers in Africa: a pooled analysis. Lancet. 2004;364:1236–1243.
12. WHO Multicentre Growth Reference Study Group. WHO Child Growth Standards: Length/Height-for-Age, Weight-for-Age, Weight-for-Length, Weight-for-Height and Body Mass Index-for-Age: Methods and Development. Geneva, Switzerland: World Health Organization; 2006. Available at: http://www.who.int/childgrowth/standards/technical_report/en/index.html
. Accessed June 4, 2013.
13. Spira R, Lepage P, Dabis F, et al.. Natural history of human immunodeficiency virus type 1 infection in children
: a five-year prospective study in Rwanda. Pediatrics. 1999;104:.
14. Bridge A, Kipp W, Konde-Lule J, et al.. Nutritional status of young children
in AIDS affected households and controls in Uganda. Am J Trop Med Hyg. 2006;74:926–931.
15. Newell ML, Borja MC, Peckman C, et al.; European Collaborative Study. Height weight and growth in children
born to mothers with HIV infection in Europe. Pediatrics. 2003;111:e52–e60.
17. Mermin J, Ekwaru JP, Liechty CA, et al.. Effect of cotrimoxazole prophylaxis, antiretroviral therapy and insecticide treated bed nets on the frequency of malaria in HIV-1 infected adults in Uganda; a prospective cohort study. Lancet. 2006;367:1256–1261.
18. Kline M, Paul ME, Bohannon B, et al.. Characteristics of children
surviving to 5 years of age or older with vertically acquired HIV infection. Pediatr AIDS HIV Infect. 1995;6:350–353.
19. Blanche S, Newell ML, Mayaux MJ, et al.. Morbidity and mortality in European children
vertically infected by HIV-1. The French Pediatric HIV Infection Study Group and European Collaborative Study. J Acquir Immune Defic Syndr Hum Retrovirol. 1997;14:442–450.
20. The European Collaborative Study. Natural history of vertically acquired human immunodeficiency virus-1 infection. Pediatrics. 1994;94:815–819.
21. Violari A, Cotton MF, Gib DM, et al.; CHER Study Team. Early antiretroviral therapy and mortality among HIV-infected infants. N Engl J Med. 2008; 359: 2233–2244.