In low- and middle-income countries, a growing number of perinatally infected children are reaching adolescence and adulthood because of the scale-up of antiretroviral therapy (ART) programs during the last decade. However, few pediatric cohorts have been sufficiently followed to investigate clinical outcomes during adolescence.1–3 Maintaining ART adherence throughout adolescence is challenging4,5 and poor outcomes during this vulnerable period have been reported.6–8
Using data from a cohort of HIV-infected children receiving ART over a 14-year period within a network of 39 public hospitals throughout Thailand, we estimated the incidence of mortality before (“early”) and after (“late”) 6 months of ART initiation and of a composite outcome defined as new/recurrent AIDS-defining events or death after 6 months of ART start (late AIDS/death). We also studied factors associated with the risk of late AIDS/death.
We included data on all perinatally HIV-infected children (≤18 years) who initiated ART between January 1, 1999 and June 30, 2013 and were followed until July 31, 2014, as part of the Program for HIV Prevention and Treatment prospective observational cohort (NCT00433030).9,10 Children were enrolled with parental/caregiver consent, and personal assent if age ≥8 years. ART and laboratory monitoring were provided free of charge.
The study protocol was approved by the ethics committees at the Thai Ministry of Public Health, local hospitals, and the Faculty of Associated Medical Sciences, Chiang Mai University, Thailand.
Clinical visits were scheduled at ART initiation, at 2 weeks, 1, 3, and 6 months, and then every 6 months thereafter. CD4 and virology testing was performed at start of ART and every 6 months thereafter. HIV genotyping was performed in case of virological failure using the Agence Nationale de Recherches sur le SIDA in-house technique (AC11-Resistance Study Group PCR and Sequencing Procedures).11 Lost to follow-up (LTFU) was defined as a missed scheduled visit and no contact for at least 9 months since last appointment. Attempt to contact LTFU children were made with telephone calls and home visits. Voluntary withdrawal was defined as notified exit from study, including notified transfer to adult care.
Nucleoside reverse transcriptase inhibitor (NRTI) dual therapy was provided from 1999, mostly zidovudine and didanosine. Protease inhibitor (PI) became available in 2002 and nonnucleoside reverse transcriptase inhibitors (NNRTIs) in 2003. Children initially on dual NRTIs could be switched to triple therapy when it became available. Children who experienced drug toxicity or treatment failure received alternative regimens as needed.
Outcomes and Analysis of the Risk Factors
Events occurring during the first 6 months after ART start were defined as “early,” and those occurring after 6 months were defined as “late.” The outcomes of interest were early deaths, late deaths, and a composite outcome defined as late new/recurrent AIDS-defining events or death (late AIDS/death). All events were documented by the responsible site physicians. Clinical records were reviewed by 2 independent physicians, and causes of death were classified according to the International Classification of Diseases 10.
Factors associated with mortality within this cohort after median follow-up of 4.5 years have been previously described.9 We assessed here factors associated with the risk of occurrence of late AIDS or death. Children were considered at risk from 6 months after ART start and their follow-up time was censored at the earliest date of first AIDS event, death, or at last visit.
Analysis of factors associated with late AIDS/death included baseline characteristics at ART start and time-updated variables. Baseline characteristics were based on the closest assessment available within 1 year before ART start or, if not available, within 15 days after. Weight-for-age and height-for-age z-scores were computed based on Thai references12 and BMI-for-age z-scores based on WHO curves.13,14 For time-updated variables, all measurements after 6 months of ART were used and dichotomized as follows: adolescent (current age ≥13 years, corresponding to the age at puberty in this population15), BMI-for-age z-score always <−2 SD, and ever severely anemic (hemoglobin <8 g/dL). HIV-RNA viral load was categorized in 3 categories: always suppressed (<400 copies/mL), occasional viremia (≥1 viral load measurement ≥400 copies/mL), and never suppressed (all viral load measurements ≥400 copies/mL).
Crude mortality rates are provided with their 95% confidence intervals (CIs) based on Poisson distribution. Incidences are reported in events per 1000 person-years of follow-up (PYFU).
Factors associated with the late composite outcome of AIDS or death used Fine-Gray competing risk regression models accounting for LTFU as a competing event,16 with a backward stepwise selection. In the multivariable analysis, only time-updated variables were considered if baseline and time-updated variables were collinear and both were associated with the outcome (P < 0.20) in the univariable analysis. Associations were reported using subdistribution hazard ratio.17
To avoid loss of information and biased estimates because of missing data, we imputed missing values with linear interpolation for time-updated variables.18 Sensitivity analyses on complete cases and with adjustment on age and calendar year at ART initiation, and using HIV-RNA load ≥1000 copies/mL threshold instead of 400 copies/mL, were also performed. All tests were 2-sided and P < 0.05 was considered statistically significant. Analyses were performed using STATA version 13.
Overall Incidence of Mortality
Over the 14-year study period, 619 children ≤18 years initiated ART within the cohort. The median [interquartile range (IQR)] follow-up duration was 7.7 years (3.5–9.8) corresponding to 4320 PYFU. Overall, 53 (9%) children died during follow-up, 144 (23%) were LTFU, and 152 (25%) voluntarily withdrew, most often because of notified self-referral to another hospital. The median follow-up duration among children who died was 0.4 years (0.1–2.2), 6.5 years (2.2–8.5) for those LTFU, and 4.9 years (1.5–8.7) for those who withdrew. The crude incidence (95% CI) of death was 12 (9 to 16) per 1000 PYFU (Table 1).
A total of 29 deaths occurred within the first 6 months after ART initiation corresponding to an incidence of early death of 99 (69–142) per 1000 PYFU (Table 1).
For the analysis of late events, 553 (89%) children were included, 66 children were excluded because of <6 months of follow-up after ART start (29 deaths, 15 LTFU, and 22 withdrawals). Characteristics at ART initiation are described in Table 2: 302 (55%) were female, median age 6.4 years (IQR, 2.2–9.6), and BMI-for-age z-score −0.8 (−1.9 to 0.1). Their median follow-up was 8.1 years (4.9–9.9). Half of the children were Centers for Disease Control and Prevention stage B or C, median CD4% was 8.2% (2.5–16.3), and HIV-RNA load 5.1 log10 copies/mL (4.7–5.6). The majority, 441 (80%) children initiated NNRTI-based ART, 56 (10%) PI, and 56 (10%) a dual NRTI regimen.
A total of 24 (4%) children died after a median time of 2.6 years (1.3–6.3). Crude incidence (95% CI) of late death was 6 (4 to 9) per 1000 PYFU (Table 1). Of these deaths, 14 (58%) occurred in hospitals, 8 (33%) at home, and 2 (8%) elsewhere. The median age at death was 11.9 years (IQR, 2.6–14.4). Causes of death were mostly HIV-related: wasting syndrome (5), cryptococcal meningitis (4), septic shock (3), respiratory failure (3), cerebral hemorrhage (2), bacterial pneumopathy (1), brain abscess (1), complete heart block (1), and congestive heart failure (1). Three external causes of death included drowning (2) and suicide (1).
Late New/Recurrent AIDS-Defining Event
Twenty-four patients experienced a total of 45 new/recurrent AIDS-defining events (Table 1). Median time to first new/recurrent AIDS-defining event was 5.1 years after ART initiation (1.7–7.2). Crude incidence (95% CI) of new/recurrent AIDS-defining events 6 (4 to 8) per 1000 PYFU. First AIDS-defining events were tuberculosis (6), pneumonia (6), penicilliosis (4), cryptococcal meningitis (2), septic shock (2), and one each of the following: brain abscess, meningitis, encephalopathy, and pneumocystosis.
Late Composite Outcome: New/Reccurent AIDS-Defining Event or Death
A total of 38 (7%) children met the late composite outcome of AIDS/death, the first event was AIDS in 24 children (10 went on to subsequently die) and 14 died without a preceding AIDS event. Median time to the first composite outcome was 3.3 years (1.3–6.3) after ART start, with a crude incidence 9 (7–12) per 1000 PYFU (Table 1).
Factors associated with the late AIDS/deaths are presented in Table 2. In multivariable analysis, after adjustment on time-updated variables, factors independently associated with the composite outcome were current age ≥13 years [adjusted subhazard ratio (aSHR), 4.9; 95% CI: 2.4 to 10.1], HIV-RNA load always ≥400 copies/mL (aSHR, 12.3; 95% CI: 4.0 to 37.6), BMI-z-score always below −2 SD (aSHR, 13.7; 95% CI: 3.4 to 55.7), and ever experiencing severe anemia (aSHR, 4.6; 95% CI: 2.0 to 10.5). There was no interaction between current age and HIV-RNA load. Results of sensitivity analyses were similar when using complete case analysis, or with adjustment on age and calendar year at ART initiation, or when using HIV-RNA load ≥1000 threshold instead of 400 copies/mL (data not shown).
HIV genotypic resistance testing was performed at least once in 30% (167/553) of children, 47% (18/38) of those who reached the composite outcome versus 29% (149/515) of those who did not (P = 0.026). Median time from ART initiation to first genotyping was 3.3 years (2.0–5.1) among those who reached the composite outcome versus 2.0 years (1.3–4.7) in the others (P = 0.200). At the time of first HIV genotyping, 80% of children had NRTI mutations, 77% NNRTI mutations, and 3% PI mutations with no difference whether they met the composite outcome or not (P = 0.758). However, there was a trend suggesting children with the composite outcome were more likely to carry resistance mutations to at least 2 ART classes (32%, 12/38) than the others (20%, 103/515) (P = 0.099).
Our study indicates that the incidence of mortality in HIV-infected children and adolescents dramatically decreased after the first 6 months of ART as it was 16 times lower than during the first 6 months. It emphasizes the importance of splitting early and late mortality analyses. After adjustment, being an adolescent, having persistent severely low BMI, never achieving viral suppression <400 copies/mL, and one or more episodes of severe anemia 6 months after ART start were independently associated with late new/recurrent AIDS-defining events or deaths.
The overall mortality incidence in our study was low at 12 per 1000 PYFU, within the range 6–15 per 1000 PYFU reported in other pediatric cohorts in developed countries2,19 but lower than the 17–144 per 1000 PYFU in developing countries.3,20–26 Also, the dramatic decrease in mortality after a few months of ART is consistent with previous studies.19–22 Yet, rates are difficult to compare because mortality depends on age at ART initiation, which varies across cohorts. Indeed, several studies indicate lower mortality in children initiating ART at older ages, a selected population of long-term survivors.3,19,21,25,26 In a previous assessment within the same cohort after 4.5-year follow-up, the late mortality incidence was similar than in this updated analysis, reflecting a stable mortality trend over time among those on long-term ART.9
Children meeting the composite outcome had their first HIV genotypic resistance test later than those who did not reach the outcome, which may reflect delayed detection of treatment failure leading to the accumulation of resistance mutations to ART classes.
As in any cohort, unreported AIDS-defining events or mortality among children LTFU may lead to an underestimation of events.27 Taking into account LTFU as a competing risk should limit this bias.
Since the number of deaths was relatively small and deaths were mainly related to AIDS-defining events, we used a combined outcome of AIDS-defining events or deaths to increase the statistical power in identifying most at-risk children.
In previous studies over shorter follow-up periods, baseline characteristics at ART initiation such as young age, advanced disease stage, poor immunological status, anemia, and poor nutritional status were associated with mortality.3,9,19–23,28 However, as follow-up lengthens, the significance of baseline characteristics fades while the relative contribution of time-updated variables increases. In our study, using time-updated variables, we found that children experiencing persistent poor nutritional status or severe anemia were at higher risk of adverse outcome. These factors are easily accessible to clinicians, even with limited laboratory facilities.
Our analysis clearly shows that adolescence was an independent predictor of late new/recurrent AIDS-defining event or death with a risk nearly 5 times higher than in younger children. This association remained even after adjustment on viral load, suggesting the potential role of unadjusted confounders related to poor adherence. In a cohort study in Uganda, over a median 28-month follow-up, the crude mortality rate was higher for adolescents than for younger children.6 Also, an increasing trend in mortality has been described among perinatally HIV-infected adolescents reaching adulthood in England.29 Maintaining ART adherence is challenging especially when children stop relying on their caregivers for their daily medication.4,30,31 In a context where third-line ART is not readily available, this may result in adverse outcomes, including death.32,33
Our results emphasize factors that should trigger closer follow-up and highlight the need for novel interventions to support the increasing group of adolescents at risk of poor clinical outcomes.
The authors thank all the children and families who participated in the Program for HIV Prevention and Treatment (PHPT) cohort study.
The authors thank the site principal investigators Lamphun: Pornpun Wannarit; Phayao Provincial Hospital: Pornchai Techakunakorn; Chiangrai Prachanukroh: Rawiwan Hansudewechakul; Chiang Kham: Vanichaya Wanchaitanawong; Phan: Sookchai Theansavettrakul; Mae Sai: Sirisak Nanta; Prapokklao: Chaiwat Ngampiyaskul; Banglamung: Siriluk Phanomcheong; Chonburi: Suchat Hongsiriwon; Rayong: Warit Karnchanamayul; Bhuddasothorn Chacheongsao: Ratchanee Kwanchaipanich; Nakornping: Suparat Kanjanavanit; Somdej Prapinklao: Nareerat Kamonpakorn, Maneeratn Nantarukchaikul; Bhumibol Adulyadej: Prapaisri Layangool, Jutarat Mekmullica; Pranangklao: Paiboon Lucksanapisitkul, Sudarat Watanayothin; Buddhachinaraj: Narong Lertpienthum; Hat Yai: Boonyarat Warachit; Regional Health Promotion Center 6, Khon Kaen: Sansanee Hanpinitsak; Nong Khai: Sathit Potchalongsin; Samutsakhon: Pimpraphai Thanasiri, Sawitree Krikajornkitti; Phaholpolphayuhasena: Pornsawan Attavinijtrakarn; Kalasin: Sakulrat Srirojana; Nakhonpathom: Suthunya Bunjongpak; Samutprakarn: Achara Puangsombat; Mahasarakam: Sathaporn Na-Rajsima; Roi-et: Pornchai Ananpatharachai; Sanpatong: Noppadon Akarathum; Vachira Phuket: Weerasak Lawtongkum; Chiangdao: Prapawan Kheunjan, Thitiporn Suriyaboon, Airada Saipanya.
The authors are also grateful to Kanchana Than-in-at, Nirattiya Jaisieng, Rapeepan Suaysod, Sanuphong Chailoet, Naritsara Naratee, and Suttipong Kawilapat who performed data management. The authors also thank all PHPT staffs who were involved in the PHPT cohort and this study.
1. French N, Mujugira A, Nakiyingi J, et al. Immunologic and clinical stages in HIV
-1-infected Ugandan adults are comparable and provide no evidence of rapid progression but poor survival with advanced disease. J Acquir Immune Defic Syndr. 1999;22:509–516.
2. Judd A, Doerholt K, Tookey PA, et al. Morbidity, mortality
, and response to treatment by children
in the United Kingdom and Ireland with perinatally acquired HIV
infection during 1996–2006: planning for teenage and adult care. Clin Infect Dis. 2007;45:918–924.
3. Kabue MM, Buck WC, Wanless SR, et al. Mortality
and clinical outcomes in HIV
on antiretroviral therapy in Malawi, Lesotho, and Swaziland. Pediatrics. 2012;130:e591–599.
4. Kahana SY, Rohan J, Allison S, et al. A meta-analysis of adherence to antiretroviral therapy and virologic responses in HIV
, and young adults. AIDS
5. Lowenthal ED, Bakeera-Kitaka S, Marukutira T, et al. Perinatally acquired HIV
infection in adolescents
from sub-Saharan Africa: a review of emerging challenges. Lancet Infect Dis. 2014;14:627–639.
6. Bakanda C, Birungi J, Mwesigwa R, et al. Survival of HIV
on antiretroviral therapy in Uganda: findings from a nationally representative cohort in Uganda. PLoS One. 2011;6:e19261.
7. Jobanputra K, Parker LA, Azih C, et al. Factors associated with virological failure and suppression after enhanced adherence counselling, in children
and adults on antiretroviral therapy for HIV
in Swaziland. PLoS One. 2015;10:e0116144.
8. Nachega JB, Hislop M, Nguyen H, et al. Antiretroviral therapy adherence, virologic and immunologic outcomes in adolescents
compared with adults in southern Africa. J Acquir Immune Defic Syndr. 2009;51:65–71.
9. Collins IJ, Jourdain G, Hansudewechakul R, et al. Long-term survival of HIV
receiving antiretroviral therapy in Thailand
: a 5-year observational cohort study. Clin Infect Dis. 2010;51:1449–1457.
10. Suaysod R, Ngo-Giang-Huong N, Salvadori N, et al. Treatment failure in HIV
on second-line protease inhibitor-based antiretroviral therapy. Clin Infect Dis. 2015;61:95–101.
11. The French ANRS (National Agency for AIDS
Research) AC11 Resistance group. HIV
-1 PCR and Sequencing Procedures. Available at: http://http://www.hivfrenchresistance.org
/ANRS-procedures.pdf. Accessed January 15, 2017.
12. World Health Organization. International Classification of Diseases (ICD). Available at: http://apps.who.int/classifications/icd10/browse/2016/en. Accessed July 8, 2016.
13. de Onis M, Onyango AW, Borghi E, et al. Development of a WHO growth reference for school-aged children
. Bull World Health Organ. 2007;85:660–667.
14. Working Group on Using Weight and Height References in Evaluating the Growth Status of Thai Children
. Manual on Using Weight and Height References in Evaluation in Growth Status of Thai Children
. Bangkok, Thailand
: Department of Health, Ministry of Public Health; 2000.
15. Rolland-Guillard L, de La Rochebrochard E, Sirirungsi W, et al. Reproductive health, social life and future plans of adolescents
born with HIV
: a case-control study in Thailand
. Presented at the 9th IAS Conference on HIV
Science; July 23–26, 2017; Paris, France.
16. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94:496–509.
17. Lau B, Cole SR, Gange SJ. Competing risk regression models for epidemiologic data. Am J Epidemiol. 2009;170:244–256.
18. Cokluk O, Kayri M. The effects of methods of imputation for missing values on the validity and reliability of scales. Educ Sci Theory Pract. 2011;11:303–309.
19. Brady MT, Oleske JM, Williams PL, et al. Declines in mortality
rates and changes in causes of death in HIV
during the HAART era. J Acquir Immune Defic Syndr. 2010;53:86–94.
20. Anaky MF, Duvignac J, Wemin L, et al. Scaling up antiretroviral therapy for HIV
in Cote d'Ivoire: determinants of survival and loss to programme. Bull World Health Organ. 2010;88:490–499.
21. Gebremedhin A, Gebremariam S, Haile F, et al. Predictors of mortality
on anti-retroviral therapy in Mekelle Hospital, Northern Ethiopia: a retrospective cohort study. BMC Public Health. 2013;13:1047.
22. Lumbiganon P, Kariminia A, Aurpibul L, et al. Survival of HIV
: a cohort study from the Asia-Pacific region. J Acquir Immune Defic Syndr. 2011;56:365–371.
23. Nugent J, Edmonds A, Lusiama J, et al. Predicting mortality
initiating highly active antiretroviral therapy in a resource-deprived setting. Pediatr Infect Dis J. 2014;33:1148–1155.
24. Puthanakit T, Aurpibul L, Oberdorfer P, et al. Hospitalization and mortality
after receiving highly active antiretroviral therapy. Clin Infect Dis. 2007;44:599–604.
25. Sanjeeva GN, Gujjal Chebbi P, Pavithra HB, et al. Predictors of mortality
rate in a cohort of children
living with HIV
from India. Indian J Pediatr. 2016;83:765–771.
26. Zanoni BC, Phungula T, Zanoni HM, et al. Risk factors associated with increased mortality
initiating antiretroviral therapy (ART) in South Africa. PLoS One. 2011;6:e22706.
27. Fenner L, Brinkhof MW, Keiser O, et al. Early mortality
and loss to follow-up in HIV
starting antiretroviral therapy in Southern Africa. J Acquir Immune Defic Syndr. 2010;54:524–532.
28. Ebissa G, Deyessa N, Biadgilign S. Predictors of early mortality
in a cohort of HIV
receiving high active antiretroviral treatment in public hospitals in Ethiopia. AIDS
29. Fish R, Judd A, Jungmann E, et al. Mortality
in perinatally HIV
-infected young people in England following transition to adult care: an HIV
Young Persons Network (HYPNet) audit. HIV
30. World Health Organization. Adherence to long-term therapies: evidence for action. Available at: http://http://www.who.int
/chp/knowledge/publications/adherence_full_report.pdf. Accessed July 8, 2016.
31. Xu L, Munir K, Kanabkaew C, et al. Factors influencing antiretroviral treatment suboptimal adherence among perinatally HIV
. PLoS One. 2017;12:e0172392.
32. Dow DE, Shayo AM, Cunningham CK, et al. Durability of antiretroviral therapy and predictors of virologic failure among perinatally HIV
in Tanzania: a four-year follow-up. BMC Infect Dis. 2014;14:567.
33. Nsheha AH, Dow DE, Kapanda GE, et al. Adherence to antiretroviral therapy among HIV
receiving care at Kilimanjaro Christian Medical Centre (KCMC), Northern Tanzania: a cross-sectional analytical study. Pan Afr Med J. 2014;17:238.