Since the widespread implementation of routine antenatal HIV screening in the UK and Ireland from 2000 onwards, the proportion of HIV-infected pregnant women diagnosed before delivery has increased substantially, from less than one-third in the mid-1990s to over 90% since 2003 . Diagnosed women are offered interventions to reduce the risk of mother-to-child transmission (MTCT), including antiretroviral therapy and elective caesarean section delivery, and are advised not to breastfeed . As a result of improving detection rates and high uptake of interventions, transmission rates in the UK and elsewhere in Europe decreased from 15–25% in the early 1990s to less than 2% in recent years [3–5]. Highly active antiretroviral therapy (HAART) is routinely prescribed in pregnancy [2,6], and a substantial proportion of previously diagnosed women are on HAART at conception and throughout pregnancy . Although antiretroviral therapy provides clear benefits, these highly potent drugs could also have adverse effects on the pregnancy or developing fetus.
Prematurity is associated with increased morbidity and mortality , and has been identified as a risk factor for MTCT . A possible association between antiretroviral therapy and prematurity was initially identified in 1998 when premature delivery was reported in 10 of 30 women taking combination therapy in a Swiss study . Combined analysis of Swiss and European Collaborative Study (ECS) data confirmed that prematurity rates were higher among women on combination therapy compared with untreated women ; the ECS subsequently reported a two-fold increase in prematurity associated with HAART, compared with mono or dual therapy . A potential biological mechanism for this association has also been proposed . In contrast, reports from the United States (US) have been conflicting. Although several studies found no association between prematurity and combination therapy [13–15], Cotter et al reported a 1.8-fold increased risk associated with protease inhibitor (PI)-containing therapy . Differences in population characteristics, indication for treatment, data collection or analytical approaches could account for these discrepant findings [17,18].
The association between HAART and other pregnancy outcomes is also unclear; an increased risk of fetal death  and very low birthweight (< 1500 g)  has been reported in some studies, but not in others [15,16,20]. There is so far no evidence of an association between antiretroviral therapy (ART) and congenital abnormalities [21–23].
In the UK and Ireland, information on all pregnancies in diagnosed HIV-infected women is collected routinely through national surveillance: we explore the use of ART in pregnancy and its relationship with prematurity, birthweight, stillbirth and neonatal mortality in this unselected population.
The National Study of HIV in Pregnancy and Childhood
Surveillance of obstetric HIV infection in the UK and Ireland is carried out through the National Study of HIV in Pregnancy and Childhood (NSHPC). A confidential, voluntary, active reporting scheme was established in June 1989 under the auspices of the Royal College of Obstetricians and Gynaecologists; a designated respondent for each maternity unit (generally a midwife, obstetrician or genito-urinary physician) reports all pregnancies in HIV-infected women, regardless of outcome, and data are collected on a standard questionnaire. A parallel paediatric HIV reporting scheme has been in operation since 1986 : HIV-infected children and infants born to infected women are reported by paediatricians, mainly through the British Paediatric Surveillance Unit's active monthly reporting scheme.
This paper is based on pregnancies resulting in a singleton live birth or stillbirth delivered between 1990 and 2005, in women diagnosed with HIV before delivery and reported to the NSHPC by March 2006. Analyses were based on pregnancies rather than women, and some women were therefore included more than once.
Exposure to antiretroviral therapy in pregnancy was categorized as untreated, monotherapy, dual therapy or HAART (three or more antiretroviral drugs). As dual therapy regimens are uncommon, and of lower potency than HAART , they were grouped with monotherapy in these analyses. HAART was categorized according to whether PIs and/or non-nucleoside reverse transcriptase inhibitors (NNRTIs) were included. Timing of therapy was classified as ‘early’ if antiretroviral therapy was initiated before or during the first trimester of pregnancy (≤ 12 completed weeks gestation).
Year of delivery was grouped into 1990–1993 (pre-antiretroviral therapy era), 1994–1999 (early antiretroviral therapy era ) and 2000–2005 (HAART and routine antenatal screening era [6,27]). Mode of delivery was reported as vaginal, or elective or emergency caesarean section. Gestational age was in completed weeks and prematurity was defined as delivery at < 37 weeks gestation. Delivery by elective caesarean section should not impact on prematurity rates, as the British HIV Association (BHIVA) guidelines recommend that it take place at 38 weeks gestation . Stillbirth was defined as fetal death at ≥ 24 weeks gestation, and neonatal death as death in the first 28 days of life. Z-scores for birthweight standardized for gestational age were obtained using British standards .
Maternal clinical status was classified as symptomatic if AIDS or HIV-related symptoms were reported at any time in pregnancy. CD4 cell count (not routinely reported before 2000) was categorised as < 200, 200–499 and ≥ 500 cells/μl; latest antepartum CD4 cell count was used in the analyses. HIV RNA viral load (VL) (recorded from 1998 onward) was classified as < 1000 or ≥ 1000 copies/ml to allow for changes over time in assay detection limits; the closest VL to delivery was included in analyses, measured at least 1 week after antiretroviral therapy initiation and before 15 days postpartum.
Data were managed in a Microsoft Access 2002 database (Microsoft Corp., Redmond, Washington, USA) and analysed using Stata 9.0 (Stata Corp., College Station, Texas, USA) . Categorical variables were compared using χ2 tests or Fisher's exact test, and means using t-tests. Logistic regression models were fitted to obtain odds ratios (ORs) and 95% confidence intervals (CI). All adjusted ORs (AORs) were adjusted for the following potential confounders : injecting drug use (IDU) as the probable route of HIV infection, ethnic origin, maternal age at delivery and clinical status; interaction terms between these factors and antiretroviral therapy were also considered. Likelihood ratio tests (LRT) were used to compare nested logistic regression models. To allow for repeat pregnancies in the same woman, generalized linear mixed effects were used to fit logistic regression models accounting for random effects attributed to the mother . Only random effects on the intercept of the linear predictor were considered. All ORs and P-values were calculated using this class of models.
Over the study period, 5009 pregnancies were reported and the overall prematurity rate was 13.3% (667/5009). Seventy pregnancies (1.3%) had inadequate information on antiretroviral therapy; these did not differ from those for which treatment information was available in terms of prematurity, but a higher proportion occurred before 2000 (34 versus 17%; P < 0.001), and were in white women (27 versus 17%; P = 0.023) and women with IDU-acquired infection (12 versus 5%; P < 0.001).
There were clear baseline differences between antiretroviral therapy-treated and untreated women, and between women on mono/dual therapy and those on HAART, reflecting changes in the characteristics of HIV-infected women over time and trends in antiretroviral therapy use in pregnancy (Table 1).
Pregnancies in untreated women
There were 494 pregnancies in untreated women; the prematurity rate in this group was 11.0% (20/181) in 1990–1993, 16.7% (26/156) in 1994–1999 and 19.7% (31/157) in 2000–2005.
Since the introduction of antiretroviral therapy, women remaining untreated at delivery were likely to have delivered before antiretroviral therapy could be initiated, due to late diagnosis and/or premature delivery; for premature deliveries in 2000–2005, diagnosis was within 14 days of delivery in 29% (9/31) of untreated but only 2% (9/522) of treated women (P < 0.001). Due to biases introduced by these changes over time, it was not appropriate to use these pregnancies as a comparison group.
Pregnancies in antiretroviral therapy-treated women
Most of the 4445 pregnancies with antiretroviral therapy exposure were in black African women, and few were in women with IDU-acquired infection (Table 1). Median maternal age at delivery was 29.7 years [range, 14.8–47.4; interquartile range (IQR), 26.2–33.6]. Most deliveries were by elective caesarean section (Table 1), and median gestational age was 38 weeks (IQR, 38–39). Compared with HAART, mono/dual therapy was associated with being younger, white, acquiring HIV through IDU, VL ≥ 1000 copies/ml and CD4 cell count ≥ 500 cells/μl (Table 1).
Most women took HAART in pregnancy (Table 2), and HAART more frequently included an NNRTI than a PI. Over a quarter of women on HAART were on treatment early in pregnancy (Table 2), whereas most women taking mono/dual therapy started later (94.9%, 1007/1061). Actual date of initiation of antiretroviral therapy was available for 89.3% (3968/4445) of pregnancies: the proportion in which antiretroviral therapy was initiated late (within 2 weeks of delivery) did not differ by type of antiretroviral therapy, even among women who delivered prematurely (data not shown).
The overall prematurity rate in antiretroviral therapy-treated pregnancies was 13.1% (583/4445; 95% CI, 12.1–14.2); 51.8% (302/583) of premature deliveries were at < 35 weeks, including 23.3% (136/583) at < 32 weeks. Prematurity was not significantly associated with maternal age or ethnic origin (Table 3); significant risk factors were IDU-acquired infection, HIV-related symptoms in pregnancy, and CD4 cell count < 500 cells/μl (Table 3). Viral load was not associated with premature delivery in univariable analysis; 13.9% (53/382) of pregnancies in women with VL ≥ 1000 copies/ml delivered prematurely, compared with 11.4% (320/2797) with VL < 1000 copies/ml (P = 0.168).
The prematurity rate was 14.1% (476/3384) in pregnancies with HAART exposure and 10.1% (107/1061) with mono/dual therapy exposure (Table 3) (OR = 1.49; 95% CI, 1.18–1.89; P = 0.001). The association remained significant after adjusting for clinical status, IDU-acquired infection, ethnic origin and maternal age (AOR = 1.51; 95% CI, 1.19–1.93; P = 0.001). There were no significant interactions. Including CD4 cell count in the model reduced the AOR slightly, but the association remained significant (Table 3). The multivariable analyses were also repeated for 3179 pregnancies with recorded VL (excluding CD4 cell count to avoid collinearity and loss of power): both VL ≥ 1000 copies/ml (AOR = 1.45; 95% CI, 1.02–2.06; P = 0.039) and HAART (AOR = 1.47; 95% CI, 1.03–2.09; P = 0.032) were independently associated with prematurity. The prematurity rate in pregnancies with missing CD4 cell count was higher than in those with CD4 count (17 versus 12%, respectively, P < 0.001), and similarly for VL (16 versus 13%, P = 0.015), probably reflecting lack of opportunity for testing women in premature labour.
Information on other risk factors for prematurity, such as previous obstetric history and prior preterm delivery, was not available. However, parity was reported for 90.9% (4039/4445) of pregnancies: the magnitude of the association between HAART and prematurity was similar for parous and nulliparous women (AOR = 1.46; 95% CI, 1.08–1.98; P = 0.012, and AOR = 1.60; 95% CI, 1.06–2.43; P = 0.026, respectively).
The association between HAART and prematurity was consistent across time periods (1994–1999, AOR = 1.53; 95% CI, 0.85–2.75; P = 0.159; 2000–2005, AOR = 1.50; 95% CI, 1.13–2.00; P = 0.005). Excluding 157 pregnancies with dual therapy exposure did not substantially alter the findings (AOR = 1.39; 95% CI, 1.08–1.79; P = 0.010), nor did excluding 913 pregnancies with treatment (mostly HAART) in early pregnancy (AOR = 1.45; 95% CI, 1.13–1.87; P = 0.003). The association was also present in 1034 pregnancies with diagnosis before pregnancy, but no early antiretroviral therapy (AOR = 1.43; 95% CI, 0.93–2.21; P = 0.105). In HAART-exposed pregnancies, there was no association between timing of treatment initiation and prematurity: 16.4% (150/914) following early exposure, compared with 14.6% (188/1287) following initiation at 13–26 weeks gestation (AOR = 1.11; 95% CI, 0.85–1.44; P = 0.460).
The association between HAART and prematurity was more pronounced for deliveries at < 35 and < 32 weeks than at < 37 weeks gestation (Table 4). Delivery at < 35 weeks occurred in 7.8% (264/3384) of HAART-exposed pregnancies and 3.6% (38/1061) of mono/dual therapy-exposed (P < 0.001) pregnancies, and at < 32 weeks in 3.6% (121/3384) and 1.4% (15/1061) respectively (P = 0.001).
Among 3384 HAART-exposed pregnancies, premature delivery occurred in 8.7% (6/69) of those with exposure to NRTIs only, 14.3% (261/1831) of those with NNRTI exposure, 13.5% (169/1256) with PI exposure, and 17.5% (40/228) with both PI and NNRTI exposure (χ2 = 4.37, P = 0.224). The risk of premature delivery was not significantly different according to whether HAART included a PI or not (AOR = 0.96; 95% CI, 0.78–1.19; P = 0.738).
Mean birthweight was slightly but significantly lower for infants exposed to HAART in utero than for infants exposed to mono/dual therapy (2.98 versus 3.10 kg, respectively, P < 0.001). After standardizing for gestational age, HAART-exposed infants were significantly lighter than those exposed to mono/dual therapy (mean z-scores for birthweight standardized for gestational age: HAART, –0.06; mono/dual therapy, 0.06, P = 0.002). Results were similar for premature and non-premature infants, but not statistically significant for the 469 premature infants (mean z-scores: HAART, –0.08; mono/dual therapy, 0.08, P = 0.329).
The stillbirth rate was 12.7 per 1000 births (43/3384) following HAART exposure, compared with 5.7 per 1000 (6/1061) following mono/dual therapy exposure (OR = 2.26; 95% CI, 0.96–5.34; P = 0.062; AOR = 2.27; 95% CI, 0.96–5.41; P = 0.063). The neonatal mortality rate was 4.2 per 1000 live births (14/3341) following HAART, and 1.9 per 1000 (2/1055) following mono/dual therapy (OR = 2.20; 95% CI, 0.50–9.69; P = 0.298). Twelve of 21 neonates who died were born at < 32 weeks gestation, five with in utero exposure to HAART.
Through this national surveillance study, information is sought on all diagnosed HIV-infected pregnant women in the UK and Ireland; comparison with unlinked anonymous survey data suggests that over 90% of affected pregnancies were diagnosed and reported between 2003 and 2005 , and substantial case ascertainment bias is therefore unlikely. The demographic characteristics of HIV-infected women and antiretroviral therapy use in pregnancy changed over time. In particular, untreated women formed a heterogeneous group: almost 40% were reported before antiretroviral therapy was widely used and about 30% since 2000, when not receiving antiretroviral therapy was generally associated with late diagnosis, a more likely scenario where pregnancy is curtailed by premature delivery. Comparison of prematurity rates between treated and untreated women was therefore not appropriate. Only 70 pregnancies were excluded due to insufficient treatment information, some of which were undoubtedly in untreated women: their prematurity rate was similar to the overall rate, and excluding this group is unlikely to introduce significant bias.
Among pregnancies in treated women, HAART was associated with a 1.5-fold increased risk of premature delivery compared with mono/dual therapy, and the effect was stronger for delivery at < 35 or < 32 weeks gestation. The association was maintained after adjusting for IDU-acquired infection, ethnic origin, maternal age, and factors associated with maternal immunosuppression (HIV-related symptoms, CD4 cell count, VL). Prior preterm delivery has been postulated as a potential confounder in the association between antiretroviral therapy and prematurity [14,18]; although we had no information on prior preterm delivery, our findings were similar for parous and nulliparous women. Information on other potential risk factors for prematurity, such as socio-economic status, smoking, and use of alcohol or illicit drugs, is not collected through the surveillance system. However, despite substantial changes over time in the demographic characteristics of HIV-infected pregnant women, including risk factors for prematurity, the magnitude of the association with HAART was similar in 1994–1999 and 2000–2005.
Findings from other studies exploring the association between antiretroviral therapy and prematurity have been inconsistent, possibly due to differences in populations or study protocols. In contrast to cohort studies and clinical trials, which recruit selectively and require consent, these analyses are based on comprehensive population surveillance. Our results concur with other European studies showing an increased risk of prematurity associated with HAART [10,11]. A US study has also reported a similar association , but limited to PI-containing regimens. In our study, substantial numbers of women were on PI- and non-PI-containing regimens, and an increased risk of prematurity was observed in both groups.
An important source of bias in observational studies exploring antiretroviral therapy and premature delivery is indication for treatment . The BHIVA guidelines now suggest monotherapy or HAART for pregnant HIV-infected women who do not require treatment for their own health , but only HAART for women who do. Although information on indication for treatment was not available, when we adjusted for HIV-related symptoms and CD4 cell count as a proxy for maternal health, the association between HAART and prematurity remained. We also carried out a sub-analysis of pregnancies in women diagnosed before pregnancy but not already on treatment (and possibly less likely to require treatment for their own health than those on antiretroviral therapy at conception); the magnitude of the association (AOR = 1.43) was similar to that in the main model, although not statistically significant.
To explore the association between duration of treatment and prematurity, we compared pregnancies with early HAART initiation with those with initiation at 13–26 weeks gestation. We excluded pregnancies delivered after 26 weeks because of bias introduced by differences in opportunity to start treatment according to timing of delivery; our failure to detect a statistically significantly increased risk of prematurity with longer duration of treatment could have been due to the reduced sample size. Few studies have explored the effect of duration of HAART exposure [11,16], possibly because timing of initiation of treatment is not always available, and because analysis is complicated since premature delivery naturally shortens the duration of antenatal treatment.
The association between low CD4 cell count and prematurity has been reported elsewhere, as has the link with IDU , which we detected despite having information only on IDU-associated HIV acquisition, rather than IDU during pregnancy. Although evidence suggests that black African women in the UK generally have higher prematurity rates than white women [31,32], rates were similar in our study; however, white HIV-infected pregnant women are unlikely to be representative of the general white pregnant population with respect to current or past IDU, or other risk factors for premature delivery. The overall prematurity rate in this study (13.3%) was substantially higher than the 7–8% reported for the general population in the UK [32,33], probably due in part to differences in maternal characteristics. Although there are no national data on trends in gestational age, the rate of low birthweight (< 2500 g) in singleton infants in England and Wales remained relatively constant between 1983 (5.8%) and 2000 (6.1%) .
HAART-exposed infants were of lower birthweight (standardized for gestational age) than those exposed to mono/dual therapy, regardless of prematurity, although in clinical terms this difference was small. The association between HAART and birthweight standardized for gestational age has not been reported elsewhere, but birthweight < 1500 g was associated with combination therapy in one US study , and an increase in low birthweight over time has been reported in the ECS . Although stillbirth and neonatal mortality rates were higher in HAART-exposed infants, this difference was not statistically significant, possibly due to insufficient numbers.
Our findings, based on routine population surveillance in the UK and Ireland, support the premise that HAART in pregnancy is associated with an increased risk of premature delivery. The beneficial effects of prophylactic ART on MTCT rates are indisputable; observed transmission rates in this population have remained below 2% since 2000. Nevertheless, these findings raise important questions about the type of treatment to be recommended to pregnant women, particularly for those not needing HAART for their own health. The number of births to diagnosed HIV-infected women in the UK and Ireland continues to rise, and an increasing number of women are on HAART at conception. Monitoring adverse pregnancy and perinatal outcomes should remain a priority, and further research into the mechanisms leading to preterm labour in HIV-infected women is needed.
National surveillance of obstetric and paediatric HIV is undertaken through the National Study of HIV in Pregnancy and Childhood (NSHPC) in collaboration with the Health Protection Agency Centre for Infections, and Health Protection Scotland. We gratefully acknowledge the contribution of the midwives, obstetricians, genito-urinary physicians, paediatricians, clinical nurse specialists and other colleagues who report to the NSHPC through the British Paediatric Surveillance Unit of the Royal College of Paediatrics and Child Health, and the obstetric reporting scheme run under the auspices of the Royal College of Obstetricians and Gynaecologists. We thank Janet Masters who co-ordinates the study and manages the data. We also thank Icina Shakes from the NSHPC, and Marie-Louise Newell, Claire Thorne, Hermione Lyall and Annemiek de Ruiter for their helpful comments on drafts of this paper.
Ethics approval: Ethics approval for the NSHPC was renewed following review by the London Multi-Centre Research Ethics Committee in 2004 (ref. MREC/04/2/009).
Author contributions: C.L.T. and P.A.T. participated in the data collection and drafted the paper. C.L.T. carried out the statistical analyses with support from M.C.-B.. All authors participated in developing the concept of the paper and interpreting the results. All authors commented on all drafts of the paper and approved the final version. P.A.T. is responsible for the NSHPC, and is the guarantor.
Sponsorship: This work was undertaken at the UCL Institute of Child Health with funding from the Department of Health and the Health Protection Agency. Research at the Institute of Child Health benefits from Research and Development funding received from the Department of Health.
Conflict of interest statement: We declare that we have no conflicts of interest.
1. The UK Collaborative Group for HIV and STI Surveillance. A complex picture. HIV and other sexually transmitted infections in the United Kingdom: 2006. London: Health Protection Agency Centre for Infections. 2006. Available at: http://www.hpa.org.uk/publications/2006/hiv_sti_2006/
. Accessed: 13 December 2006.
2. British HIV Association. Guidelines for the management of HIV infection in pregnant women and the prevention of mother-to-child transmission of HIV. March 2005. Available at: www.bhiva.org
. Accessed: 28 July 2006.
3. Duong T, Ades AE, Gibb DM, Tookey PA, Masters J. Vertical transmission rates for HIV in the British Isles: estimates based on surveillance data. BMJ 1999; 319:1227–1229.
4. European Collaborative Study. HIV-infected pregnant women and vertical transmission in Europe since 1986
5. European Collaborative Study. Mother-to-child transmission of HIV infection in the era of highly active antiretroviral therapy
. Clin Infect Dis
6. Coll O, Fiore S, Floridia M, Giaquinto C, Grosch-Worner I, Guiliano M, et al
. Pregnancy and HIV infection: A European consensus on management. AIDS 2002; 16(suppl 2):S1–S18.
7. Slattery MM, Morrison JJ. Preterm delivery. Lancet 2002; 360:1489–1497.
8. European Collaborative Study. Vertical transmission of HIV-1: maternal immune status and obstetric factors
9. Lorenzi P, Spicher VM, Laubereau B, Hirschel B, Kind C, Rudin C, et al
. Antiretroviral therapies in pregnancy: maternal, fetal and neonatal effects. Swiss HIV Cohort Study, the Swiss Collaborative HIV and Pregnancy Study, and the Swiss Neonatal HIV Study. AIDS 1998; 12:F241–F247.
10. European Collaborative Study and the Swiss Mother + Child HIV Cohort Study. Combination antiretroviral therapy and duration of pregnancy.AIDS
11. European Collaborative Study. Increased risk of adverse pregnancy outcomes in HAART-treated HIV infected women in Europe
12. Fiore S, Newell ML, Trabattoni D, Thorne C, Gray L, Savasi V, et al
. Antiretroviral therapy-associated modulation of Th1 and Th2 immune responses in HIV-infected pregnant women. J Reprod Immunol 2006; 70:143–150.
13. Tuomala RE, Shapiro DE, Mofenson LM, Bryson Y, Culnane M, Hughes MD, et al
. Antiretroviral therapy during pregnancy and the risk of an adverse outcome. N Engl J Med 2002; 346:1863–1870.
14. Tuomala RE, Watts DH, Li D, Vajaranant M, Pitt J, Hammill H, et al
. Improved obstetric outcomes and few maternal toxicities are associated with antiretroviral therapy, including highly active antiretroviral therapy during pregnancy. J Acquir Immune Defic Syndr 2005; 38:449–473.
15. Watts DH, Balasubramanian R, Maupin RT Jr, Delke I, Dorenbaum A, Fiore S, et al
. Maternal toxicity and pregnancy complications in human immunodeficiency virus-infected women receiving antiretroviral therapy: PACTG 316. Am J Obstet Gynecol 2004; 190:506–516.
16. Cotter AM, Garcia AG, Duthely ML, Luke B, O'Sullivan MJ. Is antiretroviral therapy during pregnancy associated with an increased risk of preterm delivery, low birth weight, or stillbirth? J Infect Dis 2006; 193:1195–1201.
17. Thorne C, Fiore S, Rudin C. Antiretroviral therapy during pregnancy and the risk of an adverse outcome. N Engl J Med 2003; 348:471–472.
18. Tuomala RE, Yawetz S. Protease inhibitor use during pregnancy: is there an obstetrical risk? J Infect Dis 2006; 193:1191–1194.
19. Suy A, Martinez E, Coll O, Lonca M, Palacio M, de Lazzari E, et al
. Increased risk of pre-eclampsia and fetal death in HIV-infected pregnant women receiving highly active antiretroviral therapy. AIDS 2006; 20:59–66.
20. European Collaborative Study. Exposure to antiretroviral therapy in utero or early life: the health of uninfected children born to HIV-infected women
. J Acquir Immune Defic Syndr
21. Watts DH, Covington DL, Beckerman K, Garcia P, Scheuerle A, Dominguez K, et al
. Assessing the risk of birth defects associated with antiretroviral exposure during pregnancy. Am J Obstet Gynecol 2004; 191:985–992.
22. European Collaborative Study. Does highly active antiretroviral therapy increase the risk of congenital abnormalities in HIV-infected women?J Acquir Immune Defic Syndr
23. 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.
24. Nicoll A, Lynn R, Rahi J, Verity C, Haines L. Public health outputs from the British Paediatric Surveillance Unit and similar clinician-based systems. J R Soc Med 2000; 93:580–585.
25. British HIV Association. British HIV Association (BHIVA) guidelines for the treatment of HIV-infected adults with antiretroviral therapy. July 2003. Available at: www.bhiva.org
. Accessed: 28 July 2006.
26. Connor EM, Sperling RS, Gelber R, Kiselev P, Scott G, O'Sullivan MJ, et al
. Reduction of maternal-infant transmission of human immunodeficiency virus type 1 with zidovudine treatment. Pediatric AIDS Clinical Trials Group Protocol 076 Study Group. N Engl J Med 1994; 331:1173–1180.
28. Cole TJ, Freeman JV, Preece MA. British 1990 growth reference centiles for weight, height, body mass index and head circumference fitted by maximum penalized likelihood. Stat Med 1998; 17:407–429.
29. Stata Corporation. Stata 9.0 for Windows
. College Station, Texas: Stata Corporation; 2003.
30. Rabe-Hesketh S, Skrondal A, Pickles A. Reliable estimation of generalized linear mixed models using adaptive quadrature. The Stata Journal 2002; 2:1–21.
31. Patel RR, Steer P, Doyle P, Little MP, Elliott P. Does gestation vary by ethnic group? A London-based study of over 122,000 pregnancies with spontaneous onset of labour. Int J Epidemiol 2004; 33:107–113.
32. Aveyard P, Cheng KK, Manaseki S, Gardosi J. The risk of preterm delivery in women from different ethnic groups. BJOG 2002; 109:894–899.
33. Gardosi J, Francis A. Early pregnancy predictors of preterm birth: the role of a prolonged menstruation-conception interval. BJOG 2000; 107:228–237.