There may be long-term adverse health effects of in-utero antiretroviral therapy exposure. Data on children reported through national HIV surveillance were linked to routinely collected cancer and death data: a process known as ‘flagging’. Ninety-five per cent (2612) of reported children born in 2001–2004 in England or Wales who were uninfected or of indeterminate infection status were flagged. By the end of 2005, no cancers and 14 deaths (three uninfected and 11 indeterminate) had been notified.
aCentre for Paediatric Epidemiology and Biostatistics, Institute of Child Health, University College London, London, UK
bDepartment of Paediatrics, St Mary's Hospital, London, UK.
Received 25 August, 2006
Revised 29 November, 2006
Accepted 11 December, 2006
Most diagnosed HIV-infected pregnant women take antiretroviral therapy (ART) to reduce the risk of mother-to-child transmission of HIV, and some also take ART to delay HIV disease progression . HAART in pregnancy is associated with premature delivery , and there are concerns about other potential adverse outcomes and the long-term safety of in-utero ART exposure. It is reassuring that observational studies have provided no evidence of an increased risk of congenital abnormalities [3,4], nor of cancer in uninfected children in early childhood [5,6]. The carcinogenic potential of transplacental zidovudine exposure has, however, been demonstrated in animal studies , and the possibility of an increased risk of cancer later in life remains. Here we report on a unique protocol for the long-term monitoring of death and cancer in children born to HIV-infected women.
Since the 1980s, active surveillance of obstetric and paediatric HIV in the UK has been conducted through the National Study of HIV in Pregnancy and Childhood (NSHPC) . Pregnancies in HIV-infected women, children born to HIV-infected women and HIV-infected children are reported to the NSHPC through two parallel schemes. Information collected includes maternal demographic characteristics, ART use and perinatal details. Case identifiers include date of birth, partial postcode of residence and National Health Service (NHS) number (a unique identifier); names are not collected. Paediatric reports are followed up to establish infection status, which is usually determined in the first few months of life , and the progress of infected children is monitored long term.
Administered by the Office for National Statistics (ONS), the NHS Central Register is the central record of NHS patients registered in England and Wales. Summary data on births and deaths are stored on the Births/Deaths Registration Database (BDRD) at the General Register Office. Death and cancer registrations are recorded on the NHS Central Register. The ONS provides a medical research resource whereby study subjects are identified on the NHS Central Register (‘flagged’), so that details of deaths and cancers can be notified to specific studies . This system has been widely used, for example to follow up individuals exposed to treatments such as growth hormone  and gene therapy .
It is not possible to flag NSHPC children directly because the only identifying personal data on the NHS Central Register are NHS number, name, date of birth and sex. The NHS number is not always reported to the NSHPC, although its availability has increased in recent years from 42% for children born in 2001 to 73% for children born in 2004. Even if the NHS number is available, children cannot be flagged by NHS number alone because of the potential for transcription errors. Despite these constraints, a flagging mechanism has been established. Birth registration records of NSHPC children are identified on the BDRD using an algorithm consisting of variables recorded in both systems: child's date of birth, sex and NHS number and mother's date of birth and partial postcode of residence at delivery. A file listing is then returned to the NSHPC for review. Although most matches identified on the BDRD are unique, in a minority of cases two or more possible matches are identified for a child. A programme has been developed to confirm unique matches and to distinguish between multiple matches using two further variables: child's birth weight and mother's country of birth. Confirmed matches can then be flagged.
The majority of children reported to the NSHPC were born after 2000; 2266 uninfected children and 490 who were still of indeterminate infection status, born in 2001–2004 in England or Wales, were reported by the end of 2005 (Table 1). Using the matching algorithm on the BDRD, unique matches were obtained for 93% (2550/2756), multiple matches for 6% (176/2756) and no matches for 1% (30/2756). After the confirmation programme was run, 2612 confirmed matches (95% of 2756 children) were flagged on the NHS Central Register (Table 1).
Of the 2612 flagged children, 94% (2466) were exposed to ART in utero. Among exposed children, 79% (1949/2466) were exposed to HAART, including 20% (498/2466) exposed at conception or in the first trimester. By the end of 2005, the total length of reported follow-up time on the NHS Central Register was 7013 child-years (median 2.5 years, range 1.0–5.0), including 6593 child-years in 2466 exposed children.
By the end of 2005, no cases of cancer had been notified. Fourteen deaths were recorded, a mortality rate of 2.0 per 1000 child-years (95% confidence interval 1.1–3.3). Three deaths were in uninfected children: one unexposed child died at 9 months from a congenital malformation of the respiratory system; two children exposed to HAART died, one at one month from tuberculosis, the other at 8 months from congenital heart disease. The remaining 11 deaths were in children whose infection status had not yet been clarified, although in no case was clinical or laboratory evidence of HIV transmission reported. Three unexposed neonates died of complications associated with extreme prematurity. Three children exposed to zidovudine monotherapy died: one neonate from congenital heart disease, one at 3 months from a cerebral malformation, and one sudden unexpected death in infancy at 5 months. Among children exposed to HAART, two died of congenital heart disease, one of prematurity and sepsis, and one of group B streptococcal meningitis; these four were neonatal deaths. One HAART-exposed child died from an accidental injury.
Observing no cancers in this cohort of relatively young ART-exposed children is consistent with concurrent reported rates of cancer in 1–4 year olds in England (< 1/5000 children) . We cannot, however, exclude an elevated risk of malignancies at older ages. Procedures are now in place to enable the continued flagging of children reported to the NSHPC. This exercise also highlights the potential for using other sources of national routinely collected data to monitor long-term health outcomes in ART-exposed children, particularly with the proposed information technology developments in the NHS . Simple, non-invasive monitoring is particularly important because ART is increasingly used throughout pregnancy and in diverse drug combinations , and the number of exposed children continues to rise.
Acknowledgements and ethics approval
The authors would like to thank all those who report to the NSHPC, including paediatricians reporting through the Royal College of Paediatrics and Child Health, and obstetric respondents reporting through the Royal College of Obstetricians and Gynaecologists. Thanks also to colleagues at the ONS, and Marie-Louise Newell and Janet Masters at the Institute of Child Health. The NSHPC has approval from the London Multi-centre Research Ethics Committee. The flagging exercise has Patient Information Advisory Group support under Section 60 of the Health and Social Care Act 2001.
Sponsorship: This work was funded by the Medical Research Council and the Department of Health. Research at the Institute of Child Health and Great Ormond Street Hospital for Children NHS Trust benefits from R&D funding received from the Department of Health.
1. European Collaborative Study. Mother-to-child transmission of HIV infection in the era of highly active antiretroviral therapy. Clin Infect Dis
2. European Collaborative Study. Increased risk of adverse pregnancy outcomes in HAART-treated HIV infected women in Europe. AIDS
3. Townsend CL, Tookey PA, Cortina-Borja M, Peckham CS. Antiretroviral therapy and congenital abnormalities in infants born to HIV-1-infected women in the United Kingdom and Ireland, 1990 to 2003. J Acquir Immune Defic Syndr 2006; 42:91–94.
4. Antiretroviral Pregnancy Registry International Interim Report. Available at: http://www.apregistry.com
. Accessed: 26 July 2006.
5. Culnane M, Fowler MG, Lee SS, McSherry G, Brady M, O'Donnell K, et al
. Lack of long-term effects of in utero exposure to zidovudine among uninfected children born to HIV-infected women. JAMA 1999; 281:151–157.
6. Brogly S, Williams P, Seage GR III, Van Dyke R. In utero nucleoside reverse transcriptase inhibitor exposure and cancer in HIV-uninfected children: an update from the Pediatric AIDS Clinical Trials Group 219 and 219C cohorts. J Acquir Immune Defic Syndr 2006; 41:535–536.
7. Olivero OA, Anderson LM, Diwan BA, Haines DC, Harbaugh SW, Moskal TJ, et al
. Transplacental effects of 3′-azido-2′,3′-dideoxythymidine (AZT): tumorigenicity in mice and genotoxicity in mice and monkeys. J Natl Cancer Inst 1997; 89:1602–1608.
8. Hawkins D, Blott M, Clayden P, de Ruiter A, Foster G, Gilling-Smith C, et al
. Guidelines for the management of HIV infection in pregnant women and the prevention of mother-to-child transmission of HIV. HIV Med 2005; 6(Suppl. 2):107–148.
10. Swerdlow AJ, Higgins CD, Adlard P, Preece MA. Risk of cancer in patients treated with human pituitary growth hormone in the UK, 1959–85: a cohort study. Lancet 2002; 360:273–277.
11. Nevin NC, Spink J. Gene Therapy Advisory Committee long term monitoring of patients participating in gene therapy. Hum Gene Ther 2000; 11:1253–1255.