|Home | About | Journals | Submit | Contact Us | Français|
Background. There are scarce data on the long-termsurvival of human immunodeficiency virus (HIV)—infected children receiving antiretroviral therapy (ART) in lower-middle income countries beyond 2 years of follow-up.
Methods. Previously untreated children who initiated ART on meeting immunological and/or clinical criteria were followed in a prospective cohort in Thailand. The probability of survival up to 5 years from initiation was estimated using Kaplan-Meier methods, and factors associated with mortality were assessed using Cox regression analyses.
Results. Five hundred seventy-eight children received ART; of these, 111 (19.2%) were followed since birth. At start of ART (baseline), the median age was 6.7 years, 128 children (22%) were aged <2 years, and the median CD4 cell percentage was 7%. Median duration of follow-up was 53 months; 42 children (7%) died, and 38 (7%) were lost to follow-up. Age <12 months, low CD4 cell percentage, and low weight-for-height z score at ART initiation were independently associated with mortality (P < .001). The probability of survival among infants aged <12 months at baseline was 84.3% at 1 year and 76.7% at 5 years of ART, compared with 95.7% and 94.8%, respectively, among children aged 1 year. Low CD4 cell percentage and wasting at baseline had a strong association with mortality among older children but weak or no association among infants.
Conclusions. Children who initiated ART as infants after meeting immunological and/or clinical criteria had a high risk of mortality which persisted beyond the first year of therapy. Among older children, those with severe wasting or low CD4 cell percentage at treatment initiation were at high risk of mortality during the first 6 months of therapy. These findings support the scale-up of early HIV diagnosis and immediate treatment in infants, before advanced disease progression in older children.
Access to antiretroviral therapy (ART) has scaled up rapidly in resource-limited settings, with the goal of universal coverage for all those in need of therapy. In 2008, >4 million human immunodeficiency virus (HIV)—infected persons received ART in lower-middle income countries, including 275,000 children, an estimated 38% coverage of children in need of therapy . Although knowledge is rapidly emerging on the long-term effectiveness of ART among adults [2, 3], much less is known about effectiveness among perinatally HIV-infected children .
Treatment programs in Africa and Asia, mostly composed of children with a median age of 5–7 years at ART initiation, have reported 12-month survival rates ranging from 84% to 97% [5, 6]. A multicountry African study reported 93% (95% confidence interval [CI], 92%–94%) survival at 2 years of ART . However, there are scarce data on survival beyond 2 years, particularly among children initiating therapy during infancy. Such data, along with causes of mortality, are essential to inform clinical practice and policy.
Thailand was one of the first lower-middle income countries to provide free access to ART, with pilot programs from 1999. By 2008, of the estimated 15,000 children living with HIV, >8700 received ART . We followed children receiving ART in a network of public hospitals throughout Thailand as part of the national program. The cohort includes a significant proportion of children followed since birth, many of whom initiated ART as infants. Here we report the survival at 5 years of ART and factors associated with mortality.
Study population. The Programs for HIV Prevention & Treatment (PHPT) cohort prospectively followed HIV-infected children receiving ART in 40 public hospitals in urban and periurban settings across Thailand. The cohort began in 1999 as a pilot program and expanded in 2002 as part of national scale-up efforts.
HIV-infected children born to mothers participating in clinical trials on prevention of mother-to-child transmission of HIV (PHPT-1 & PHPT-2) were enrolled from January 1999 [8, 9]. These children were followed from birth (birth cohort), received early HIV diagnosis by HIV type 1 DNA polymerase chain reaction, and received trimethoprim-sulfamethoxazole prophylaxis from 6 weeks of age. From 2002, any HIV-infected child presenting at the participating hospitals was enrolled (referred cohort). Antiretroviral therapy and monitoring were provided free of charge. Parents or guardians provided written informed consent at study entry, and from December 2006, assent was requested from children aged 8 years. The study was approved by the Thai Ministry of Public Health and local Ethics Committees.
Treatment criteria. Children initiated ART on meeting the following clinical or immunological criteria: documented HIV infection by antibody test or HIV type 1 DNA polymerase chain reaction if aged <18 months, United States Centers of Disease Control and Prevention HIV clinical disease stage B or C, or CD4 T cell percentage <20% if aged <2 years and CD4 T cell percentage <15% if aged 2 years [10, 11].
Follow-up. Children attended the clinic monthly and had a physical exam, drug refills, and adherence counselling conducted by a nurse. They saw a physician every month during the first 3 months of treatment and every 3 months thereafter; at other visits they were referred to the physician as needed. CD4 and virology testing was performed at start of ART and every 6 months thereafter. Loss to follow-up was defined as a missed scheduled visit and no contact for >6 months. Efforts were made to trace children who were lost to follow-up with telephone calls and home visits. Voluntary withdrawal was defined as notified exit from the study.
Antiretroviral treatment. Nucleoside reverse-transcriptase inhibitor (NRTI) dual therapy was provided from 1999, mostly comprising zidovudine and didanosine. Protease inhibitor—based highly active antiretroviral therapy (HAART) became available in 2002 and nonnucleoside reverse transcriptase inhibitor—based regimens in 2003. An adult generic fixed-dosed combination of stavudine, lamivudine, and nevirapine (GPO-vir) produced in Thailand was widely used, with tablets divided into halves or quarters according to the child's body weight [12, 13]. Children who initiated with dual NRTIs could switch to HAART as soon as it became available. Children who experienced adverse effects or treatment failure were switched to alternative regimens.
Inclusion criteria. Children (aged <18 years) who initiated ART from 1 January 1999 through 31 January 2009 were included in this analysis if they were previously antiretroviral naive (excluding prevention of mother-to-child transmission of HIV prophylaxis) and were followed up for >1 day.
Outcome and risk factors. The outcome of interest was all-cause mortality. For deaths which occurred in a hospital, the primary cause of death and contributing factors were reported by the physician. If a child died at home, the caregiver was interviewed by the physician about events leading up to death. Clinical records were reviewed by 2 independent physicians, and causes of death were classified based on CoDe coding (http://www.cphiv.dk/CoDe.aspx).
Potential risk factors were age, sex, CD4 cell percentage, viral load, Centers for Disease Control and Prevention disease stage, hemoglobin level, anthropometric measures, calendar year at start of ART (baseline), and initial regimen. Weight-for-age z score, height-for-age z score, and weight-for-height z score were calculated based on Thai reference curves [14, 15].
Statistical analysis. To estimate mortality rates, follow-up for each child was considered from start of ART until date of death or last visit. Kaplan-Meier probability of survival was estimated overall, by mode of entry, and by baseline characteristics; differences in survival curves were tested using logrank test. Analyses were based on intent-to-continue treatment, ignoring treatment changes, interruptions, or terminations.
Age was categorized a priori into 4 groups. Two of the age groups were <12 months and 12–23 months, to allow evaluation of risk among infants and young children. Children aged 2 years were divided by distribution. Because weight-for-age z score and weight-for-height z score were strongly correlated, only weight-for-height z score was included in analyses.
Cox proportional hazard models were used to estimate crude and adjusted mortality hazard ratios. Associated variables with P value .2 in univariate analysis were considered in multivariate analyses, where variables with P < .2 were included in the final model with use of forward selection because of the limited number of events. Nonlinearity of continuous variables were assessed using a cubic spline term .
The proportional hazard assumption was assessed using a test for interaction between variables and follow-up time, divided at before and after 6 months of ART. The effect of associated variables was examined for interaction with age at start of ART (<12 months vs 12 months of age) on the basis of natural history studies and the CHER trial, which showed poor prediction of mortality among HIV-infected infants [17, 18]. Where interaction was suggested (P < .1), adjusted hazard ratios were reported separately by age group.
To avoid loss of information and potential biased estimates because of missing data in Cox analyses, we imputed missing values with use of Multivariate Imputation by Chained Equations (MICE) based on 20 cycles [19, 20]. Data were analyzed using STATA, version 11 (StataCorp).
A total of 586 antiretroviral-naive children initiated ART; 10 children with 1 day of follow-up were excluded, leaving 578 for this analysis (Figure 1). Fifty-three percent were male; 111 children (19%) were enrolled in the birth cohort, and 467 (81%) were enrolled in the referred cohort, initiating ART at a median age of 13 months and 7.6 years, respectively (Table 1). Children in the birth cohort had higher baseline CD4 cell percentage, had a less severe disease stage, and were more likely to start ART with dual NRTIs before 2003, compared with the referred cohort. Overall, 59 children (10.2%) started ART during infancy.
The median duration of follow-up was 53 months (interquartile range, 34–72 months), and 243 children (42%) reached 5 years of follow-up after ART initiation. Forty-two children (7.3%) died, 38 (6.6%) were lost to follow-up, and 57 (9.9%) voluntarily withdrew from the study. The loss to follow-up rate was higher in the birth cohort than the referred cohort, whereas study withdrawal rates were similar. Most withdrawals (50 of 57 cases) involved self-referral to another treatment program or hospital.
The mortality rates during the first 6 months of receiving ART were similarly high in the 2 cohorts; the combined rate was 10.2 deaths per 100 person-years (95% CI, 7.1–14.8 deaths per 100 person-years). Thereafter, the rate decreased to 1.55 deaths per 100 person-years (95% CI, 0.78–3.11 deaths per 100 person-years) in the birth cohort and 0.35 deaths per 100 person-years (95% CI, 0.16–0.77 deaths per 100 person-years) in the referred cohort. The probability of survival at 5 years was 87.1% (95% CI, 78.7%–92.3%) in the birth cohort and 93.4% (95% CI, 90.5%–95.3%) in the referred cohort.
Causes of death. Twenty-one deaths (50%) occurred at the hospital, 19 (45%) at home, and 2 (5%) in other settings. Cause of death was recorded in 40 cases (95%). The following primary causes were mostly related to infection: pneumonia of any cause (n = 11), sepsis (n = 6), diarrhea (n = 4), tuberculosis (n = 3), wasting syndrome (n = 3), congestive heart failure (n = 3), respiratory failure (n = 2), meningitis (n = 2), and 1 each of cardiomyopathy, cerebral hemorrhage, cryptococcal meningitis, encephalopathy, sequelae from cerebral toxoplasmosis, and drowning. The median age at death was 6.8 years (range, 0.4–15.0 years). Six children died at an age <12 months; all deaths were attributed to infection.
Factors associated with mortality. In univariate analyses, all variables except for sex and initial regimen were associated with mortality (Table 2). In multivariate analyses, baseline age, CD4 cell percentage, and weight-for-height z score remained independently associated with mortality, whereas hemoglobin had a weak effect (P = .07). After adjusting for these factors, mode of entry to the study was no longer associated with mortality (adjusted hazard ratio [aHR] for birth vs referred cohort, 1.45; 95% CI, 0.4–4.6; P = .53).
Children starting ART during infancy (age <12 months) were at increased risk of mortality (aHR, 7.1; 95% CI, 2.7–18.6), compared with the reference group of children aged 8 years. There was no evidence that mortality differed between the 3 older age groups (P = .67). Figure 2 shows the probability of death by baseline age over follow-up time. Most deaths among children aged 1 year at baseline occurred within 6 months after ART initiation, whereas deaths among infants were observed throughout the follow-up time. The probability of survival among infants was 89.6% (95% CI, 78.4%–95.2%) at 6 months, 84.3% (95% CI, 71.9%–91.5%) at 1 year, and 76.7% (95% CI, 63.2%–85.8%) at 5 years of therapy. The combined survival estimates among older children were 95.7% (95% CI, 93.5%–97.1%) at 6 months and 1 year and 94.8% (95% CI, 91.7%–96.7%) at 5 years (Table 3).
CD4 cell percentage was independently associated with mortality, with a 67% risk increase per 5% decrease in baseline CD4 cell percentage (aHR, 1.67; 95% CI, 1.25–2.17). This association was more pronounced among children aged 12 months at baseline (aHR, 2.27; 95% CI, 1.30–4.0), compared with infants (aHR, 1.39; 95% CI, 1.02–1.89; P = .13, for interaction).
In addition, baseline weight-for-height z score was strongly associated with mortality among older children (aHR, 2.32; 95% CI, 1.64–3.33) but not among infants (aHR, 1.10; 95% CI, 0.78–1.56; P = .003, for interaction).
In analyses restricted to children whose initial ART regimen was HAART (n = 515), there were 36 deaths (7.0%). The mortality rate was 10.3 deaths per 100 person-years (95% CI, 6.9–15.2 deaths per 100 person-years) at 6 months and 0.58 deaths per 100 person-years (95% CI, 0.32–1.05 deaths per 100 person-years) thereafter. Factors associated with mortality were consistent with those reported in the overall cohort (data not shown).
In 2002, prior to the national scale-up of ART, mortality rates among infants and children aged <5 years in Thailand were 22 and 28 deaths per 1000 live births, respectively , and AIDS was the leading cause of child mortality, estimated to account for 14% of deaths . In our study, children receiving ART had a 95% probability of survival at 6 months of therapy, similar to reports from low- and middle-income countries [6, 7, 23, 24]. The mortality rates were 10.2 deaths per 100 person-years (95% CI, 7.1–14.8 deaths per 100 person-years) during the first 6 months, decreasing to 0.62 deaths per 100 person-years (95% CI, 0.37–1.05 deaths per 100 person-years) after 6 months of therapy. The latter rate is comparable to the 0.5–0.9 deaths per 100 person-years reported in pediatric HIV cohorts in developed countries [25, 26].
However, children initiating ART during infancy had significantly poorer survival, compared with older children, after adjusting for baseline CD4 cell percentage, wasting, and hemoglobin level, and the high mortality risk persisted even after the first year of treatment. In contrast, older children had high mortality during the first 6 months of ART and rapidly stabilized thereafter. The poorer survival among infants is consistent with reports where children initiated therapy after meeting immunological and/or clinical criteria [7, 24, 27, 28]. The higher survival among older children may be attributable to the selection of survivors; children starting ART at older ages had better HIV prognoses because they had survived infancy without receiving ART [29, 30]. For that reason, we will discuss the results among infants and older children separately.
The cumulative probability of death among children starting ART during infancy was 15% at 1 year, increasing to 23% at 5 years of therapy. There are few estimates of mortality among infants receiving ART in resource-limited settings to compare .
The CHER clinical trial in South Africa, which randomized HIV-infected infants to early treatment initiation at <3 months of age or deferred treatment until meeting clinical and/or immunological criteria, reported 4% mortality at 40 weeks after randomization in the early treatment arm, compared with 16% in the deferred treatment arm, a 75% reduction in mortality (P < .001) . The World Health Organization subsequently revised its guidelines for immediate initiation of ART in HIV-infected infants, irrespective of immunological or clinical status . Interestingly, mortality in the CHER deferred treatment arm was very close to that observed in our cohort at 1 year of ART. However, a recent study in Asia reported higher mortality among infants initiating ART after meeting immunological and/or clinical criteria of 27% (95% CI, 14%–49%) at 1 year of therapy, although the sample size was small (28 infants) .
Importantly, infants in our cohort faced an extended period of high mortality risk, well beyond the first year of therapy; the reasons for this are unclear, and additional studies are warranted. The extended CHER trial is currently assessing the longterm clinical benefits of early therapy in infants and may shed light on this. Recent analyses from the European observational cohorts reported the cumulative 5-year risk of AIDS or death among infants who initiated therapy at <3 months of age when asymptomatic was significantly lower at 4.6%, compared with 21.5% among children starting ART when symptomatic (P < .001) . Additional studies are needed, particularly in resource- limited settings implementing the new World Health Organization guidelines.
Furthermore, with the expansion of access to early HIV diagnosis [34–36], many countries have reported challenges in increasing uptake among HIV-exposed infants and retaining newly diagnosed infants [37, 38]. This highlights the need for strengthening counselling services for new mothers and referral linkages between prevention of mother-to-child transmission of HIV and treatment programs.
Children starting ART after 1 year of age had a probability of survival of 95.7% at 1 year, which stabilized around 94% from 2 to 5 years of therapy. This is higher than recent reports from Thailand where the 5-year survival rate was 88% (95% CI, 86%–90%), although that estimate included infant mortality, and the median duration of follow-up was only 1.7 years, with 10% of subjects lost to follow-up . A small cohort study of 76 children in Cote d'Ivoire also reported lower survival rates of 86% (95% CI, 77%–92%) at 4 years but again included infants . As we have shown, survival rates vary significantly with age at initiation of ART, independent of baseline characteristics. There is a need for long-term survival estimates stratified by baseline age from future cohorts to compare, because such estimates are critical for the evaluation of treatment programs and for modelling the impact at the population level .
Low CD4 cell percentage at ART initiation is commonly associated with mortality, and in resource-limited settings, low weight-to-height z score has also been reported to be an important risk factor [7, 23, 24]. In our cohort, low CD4 cell percentage and weight-for-height z score at baseline had strong associations with mortality among older children but weak or no association among infants. Similar results were observedwhen using CD4 cell count or weight-for-age z score (data not shown). These findings are consistent with reports of poor predictors of mortality among infants and further supports their need for immediate treatment when asymptomatic [17, 33].
Among older children, those with severe wasting (weightfor- height z score, <−2) at baseline had a high risk of mortality of 21.8% (95% CI, 10.9%–40.6%) at 6 months of ART, compared with 3.4% (95% CI, 1.9%–5.3%) among children with higher z scores. This emphasizes the need for improved access to HIV diagnosis and follow-up of all HIV-exposed children to ensure timely treatment initiation.
Surprisingly, there was no effect of initial ART regimen on mortality despite 11% of children initiating with dual NRTI regimens prior to availability of HAART. This may be partly attributable to the small sample size. Also, there were higher rates of loss to follow-up among children starting dual therapy, compared with HAART, which may include unreported deaths. On the other hand, these children switched to HAART as soon as it became available, after a median of 31 months (interquartile range, 19–54 months) of dual therapy, which may have reduced the long-term disadvantage in terms of survival that was observed in other studies .
There are certain study limitations to be considered. Subjects lost to follow-up may have been unreported deaths, resulting in an underestimation of mortality, as reported in adult studies [3, 43]. However, loss to follow-up in our cohort was low at 6.6% at 5 years of ART, and baseline factors associated with loss to follow-up were not consistent with those associated with mortality (data not shown).
One strength of this study was its implementation in a wide network of public hospitals in urban and periurban settings, with routine HIV care provided by health care professionals who implement the national treatment program. However, additional support was provided with regular virology tests and second-line treatments after 2002, which may not have been as widely or promptly available in routine-care settings at that time. Therefore, some caution is needed with regard to generalizability, although it provides an indicator of the long-term survival rates achievable in settings very similar to standard of care. Furthermore, we performed the analyses with use of multiple imputed data to avoid bias estimates attributable to missing data. Results were similar to those obtained using the original dataset (data not shown).
In conclusion, >94% survival at 5 years of therapy was achieved in children aged 1 year at ART initiation, narrowing the gap between resource-rich and resource-limited countries. However, children who initiated ART as infants after meeting clinical or immunological criteria had significantly higher risk of mortality that persisted throughout the follow-up time. Data collection among new birth cohorts is important to assess the impact of scale-up of early HIV diagnosis and treatment strategies on infant and child survival. Continued follow-up of this cohort as children enter adolescence and begin using secondand third-line regimens is also needed to inform future programs and policies.
PHPT (Thailand): site principal investigators (number of children enrolled in each hospital). Chiangrai Prachanukroh (141): R. Hansudewechakul, K. Preedisripipat, C. Chanta; Nakornping (101): S. Kanjanavanit; Prapokklao (48): C. Ngampiyaskul, N. Srisawasdi; Chonburi (42): S. Hongsiriwan; Bhumibol Adulyadej (32): P. Layangool, J. Mekmullica; Phayao Provincial (31): P. Techakunakorn, S. Sriminiphant ; Samutsakhon (22): P. Thanasiri, S. Krikaiornkitti; Kalasin (18): S. Srirojana; Lamphun (17): P. Wannarit, K. Pagdi, R. Kosonsasitorn, R. Somsamai; Mae Chan (15): S. Buranabanjasatean; Sanpatong (15): N. Akarathum; Somdej Prapinklao (12): N. Kamonpakorn, M. Nantarukchaikul; Phan (9): S. Jungpichanvanich; Phaholpolphayuhasena (9): P. Attavijtrakarn; Samutprakarn (9): A. Puangsombat, C. Sriwacharakarn; Rayong (7): W. Karnchanamayul; Buddhachinaraj (7): N. Lertpienthum, W. Ardong; Nakhonpathom (7): S. Bunjongpak; Health Promotion Region 6 Khon Kaen (6): S. Hanpinitsak, N. Pramukkul; Somdej Pranangchao Sirikit (5): T. Hinjiranandana; Mae Sai (4): S. Kunkongkapan; Chacheongsao (3): R. Kwanchaipanich; Chiang Kham (2): V. Wanchaithanawong, P. Jittamala; Pranangklao (2): P. Lucksanapisitkul; Health Promotion Region 10 Chiang Mai (1): W. Jitphiankha, K. Jittayanun; Hat Yai (1): B. Warachit, T. Borkird; Mahasarakam (1): S. Na-Rajsima, K. Kovitanggoon; Ratchaburi (1): C. Sutthipong, O. Bamroongshawkaseme.
PHPT clinical trial unit. Sites monitoring: P. Sukrakanchana, S. Chalermpantmetagul, C. Kanabkaew, R. Peongjakta, J. Chaiwan, Y. S. Thammajitsagul, R. Wongchai, N. Kruenual, N. Krapunpongsakul,W. Pongchaisit, T. Thimakam, R.Wongsrisai, J. Wallapachai, J. Thonglo, S. Jinasa, J Khanmali, P. Chart, J. Chalasin, B. Ratchanee, N. Thuenyeanyong, P. Krueduangkam, P. Thuraset, S. Thongsuwan, W. Khamjakkaew; Laboratory: P. Tungyai, J. Kamkorn, W. Pilonpongsathorn, P. Pongpunyayuen, P. Mongkolwat, L. Laomanit, N. Wangsaeng, S. Surajinda, W. Danpaiboon, Y. Taworn, D. Saeng-ai, A. Kaewbundit, A. Khanpanya, N. Boonpleum, P. Sothanapaisan, P. Punyathi, P. Khantarag, R. Dusadeepong, T. Donchai, U. Tungchittrapituk, W. Sripaoraya; PHPT data center: S. Tanasri, S. Chailert, R. Seubmongkolchai, A. Wongja, K. Yoddee, K. Chaokasem, P. Chailert, K. Suebmongkolchai, A. Seubmongkolchai, C. Chimplee, K. Saopang, P. Chusut, S. Suekrasae, T. Yaowarat, B. Thongpunchang, T. Chitkawin, A. Lueanyod, D. Jianphinitnan, J. Inkom, N. Naratee, N. Homkham, T. Thasit, W. Wongwai, W. Chanthaweethip, R. Suaysod, T. Vorapongpisan; Administrative support: N. Chaiboonruang, P. Pirom, T. Thaiyanant, T. Intaboonma; S. Jitharidkul, S. Jaisook, D. Punyatiam, L. Summanuch, N. Rawanchaikul, P. Palidta, S. Nupradit, T. Tankool, W. Champa; Tracking and supplies: K. Than-in-at, M. Inta, R. Wongsang; Drug distribution center: D. Chinwong, C. Sanjoom, P. Sawnchitta, P.Wimolwattanasarn, N. Mungkhala, N. Jaisieng.
We thank all the children and their families and all members of the hospital teams of the participating sites. We are also grateful for the advice and assistance from the Thai Ministry of Public Health Office of the Permanent Secretary, Department of Health, Department of Communicable Diseases Control, and Provincial Hospitals Division and, especially, M. Teeratantikanont, P. Amornwichet, V. Chokevivat, T. Siraprapasiri, S. Thanprasertsuk, N. Voramongkol, S. Pattarakulwanich, N. Aungkasuwapala, and A. Chitwarakorn. We thank W. Sirirungsi, P. Leechanachai, and A. Haesungcharern, from Chiang Mai University, and K. McIntosh from the Children's Hospital of Boston, Massachusetts. Special thanks to Suriyan Tanasri and Sanupong Chailert for the data preparation and management.
Financial support. The Global Fund to fight AIDS, Tuberculosis, and Malaria Thailand (Grant Round 1 sub recipient PR-A-N-008); Ministry of Public Health, Thailand; Oxfam Great Britain, Thailand; Institut de Recherche pour le Développement, France; Institut National d'Etudes Dé- mographiques, France; Department of Technical and Economic Cooperation, Thailand; International Maternal Pediatric Adolescents AIDS Clinical Trials Group (IMPAACT); the National Institutes of Health, United States; and United Kingdom Medical Research Council (Doctoral Training Account Studentship to I.J.C.).
Potential conflicts of interest. All authors: no conflicts.
Presented in part: First International Pediatric AIDS Conference, Cape Town, South Africa, July 2009 (abstract P_21).