Guidelines for the management of HIV infection in pregnant women and the prevention of mother-to-child transmission

British HIV Association


Dr E.G.H. Lyall, Department of Paediatrics, St Mary's Hospital, Imperial College, London, UK. E-mail:


Aims of the guidelines These guidelines, drawn up by a multidisciplinary group of clinicians and lay workers active in the management of pregnant women infected with HIV, aim to give up-to-date information on interventions to reduce the risk of mother to child transmission of the virus. The evidence on the use of interventions to prevent mother to child transmission of HIV has been graded according to the strength of the data as per the definitions of the US Agency for Health Care Policy and Research [1]. Weighted evidence on the use of combination antiretroviral therapy (ART) for the treatment of HIV infection per se is presented in the BHIVA guidelines for adults [2,3]. The highest level evidence (i.e. randomised controlled trials (RCTs) or large, well conducted meta-analyses) is only available for formula feeding, prelabour caesarean section and zidovudine monotherapy. The need to treat mothers for HIV infection has led to the widespread use of ART in pregnancy which in turn results in new questions such as how to deliver when the mother, on therapy, has no detectable plasma viraemia with the most sensitive assays. In addressing many common and/or difficult clinical scenarios in the absence of ‘best evidence’ the guidelines rely heavily on ‘expert opinion’.

Recommendations for management are given in the section on clinical scenarios, and summarized in Table 3. An expanded version of these guidelines with an appendix on safety and toxicity data is available on the BHIVA website The authors are available to discuss individual cases.

  • Sections:

  • 1 Summary
  • 2 Preconception and fertility management in women and men with HIV
  • 3 Sexual health of HIV infected pregnant women
  • 4 Psychosocial support for pregnant women with HIV
  • 5 Assessment of viral load in pregnancy
  • 6 Prescribing antiretroviral therapy in pregnancy
  • 7 Maternal drug toxicity in pregnancy
  • 8 Obstetric management of pregnancy and delivery
  • 9 Interventions to reduce mother to child transmission of HIV – clinical scenarios
  • 10 Management of infants born to HIV infected mothers
  • 11 Infant feeding
  • 12 Pregnancy in women with HIV-2
  • Tables

  • 1 Evidence of the efficacy of antiretroviral therapy to reduce the risk of mother-to-child transmission of HIV
  • 2 Evidence of the efficacy of prelabour caesarean section to reduce the risk of mother-to-child transmission of HIV
  • 3 Clinical scenarios
  • 4 Infant doses of antiretroviral therapy and PCP prophylaxis

1 Summary

The risk of transmission is related to maternal health, obstetric factors and infant prematurity. Overall there is a close linear correlation between maternal viral load and risk of transmission, but as yet the evidence for a threshold below which transmission never occurs is limited [4]. This may be due to discrepancies between plasma and genital tract viral loads. However, the major studies of viral load and transmission have used assays with a lower limit of detection of 500–1000 HIV RNA copies/mL, and a relative insensitivity for some non-B subtypes. CD4 lymphocyte counts and clinical disease stage have been shown in some cohorts to have an association with the risk of transmission even after controlling for viral load [5]. The only obstetric factors that consistently show an association with risk of transmission are mode of delivery and duration of membrane rupture but invasive procedures in labour are generally avoided as they pose a theoretical risk of iatrogenic transmission. Delivery before 34 weeks of gestation has been shown to be associated with an increased risk of transmission [6].

Formula feeding has been advocated for positive women since the association with breast feeding and increased transmission was noted in 1992 [7]. The protective role of caesarean section has been clarified with both a meta-analysis [8] and a RCT reported in 1999 [9].

The findings of the first RCT, published in 1994, showing that monotherapy with zidovudine (ZDV) could reduce transmission from 25% to 8% in a non-breast feeding population [10], have been supported by numerous observational studies confirming this reduction in clinical practice. Subsequent studies have shown equivalent benefit in mothers with more advanced disease and in those who are more heavily pretreated. As standard treatment for nonpregnant adults is now with at least three antiretrovirals more women are taking combination therapy in pregnancy [2]. There are still relatively few data, however, on the safety of antiretroviral therapy (ART) in pregnancy and the management of any HIV-positive pregnant woman requires a careful consideration of the balance between the mother's own health needs, the need to reduce vertical transmission and possible adverse effects of ART to the fetus.

2 Preconception and fertility management in men and women infected with HIV

There are three groups to consider: HIV-positive men with negative female partners, HIV-positive men with positive female partners and HIV infected couples. All three groups may have fertility problems but for the discordant couples there is also the risk of sexual transmission of HIV. The risk of transmission to the woman, of approximately 1 : 500 per sexual encounter [11,12] can be reduced, but not eliminated, by limiting exposure to the most fertile period [13]. ′Sperm washing′ is a process whereby spermatozoa are removed from the surrounding HIV infected seminal plasma by a sperm swim-up technique [14]. To date there have been no seroconversions in women inseminated with washed sperm [15]. HIV-positive men with low sperm counts may be offered intracytoplasmic sperm injection following the sperm washing process [16]. Sperm washing is available in London, at the Chelsea & Westminster Hospital and St Thomas's, and in Birmingham, but not currently on the NHS. To protect an uninfected male partner, couples are equipped with quills, syringes and Gallipots and instructed on how to perform artificial insemination by partner at the time of ovulation [17]. For positive couples, safer sex to reduce transmission of viral variants is usually recommended. Couples wishing to conceive are advised to limit unprotected intercourse to the fertile period, or to follow the advice for discordant couples. The latter being especially important when there is virological failure.

In the past, in vitro fertilization (IVF) has been much disputed for HIV infected couples with infertility. IVF is now ethically acceptable since the vertical transmission rate has fallen to < 1% along with an increased life expectancy for parents on highly active antiretroviral therapy [18].

3 Sexual health of HIV-positive pregnant women

There are few data regarding the prevalence of genital infections in HIV-positive women in the UK [19]. The majority of UK pregnant HIV infected women come from Sub-Saharan Africa where the prevalence of genital infections is high [20]. However, apart from the recommendation that all pregnant women should be screened for HIV, hepatitis B and syphilis, asymptomatic pregnant women are not routinely screened for genital infections.

Chorioamnionitis may lead to premature rupture of the membranes with the possibility of premature birth [21,22]. Chorioamnionitis, prolonged rupture of membranes and premature birth have all been associated with mother-to-child transmission (MCT) of HIV and may be interlinked [23–25]. Although both Chlamydia trachomatis and Neisseria gonnorrhoea have been associated with chorioamnionitis, the organisms usually implicated are those associated with bacterial vaginosis (BV) and Ureaplasma urealyticum[21,22]. A strong association between bacterial vaginosis and premature delivery has been reported [22,26]. Organisms associated with bacterial vaginosis have been shown to stimulate HIV expression in vitro[27,28]. There are preliminary data suggesting that BV may be associated with an increased risk of maternal HIV infection in pregnancy as well as premature delivery and HIV MCT [29]. Genital infections, particularly ulcerative diseases, are associated with sexual transmission of HIV [30,31]. This may be due to an increase in local HIV replication resulting in a higher viral load in genital secretions, secondary to the presence of specific organisms, and/or ulceration and inflammation [32,33]. A recent study from Kenya demonstrated a reduction in cervical mucosal shedding of HIV-1 RNA following treatment of both gonococcal and chlamydial cervicitis [34]. Viral load in cervico-vaginal specimens has been shown to be correlated with MCT of HIV-1 [35]. Usually, the genital tract viral load will mirror the plasma viral load [36], but discordance may occur with genital infection. Increased genital tract viral load secondary to infection could conceivably increase the risk of MCT.

In the absence of RCTs, but for the reasons outlined above, HIV-positive pregnant women should be screened for genital infections. This should be done as early as possible in pregnancy and repeated at around 28 weeks, any infection detected should be treated according to UK national guidelines [37].

4 Psycho-social support for pregnant women with HIV

To support women to take advantage of the interventions to reduce transmission of HIV to their infants, professionals should be aware of the psycho–social issues. Women who are pregnant are usually advised to avoid taking medication if possible, but women with HIV are encouraged to take therapy to help prevent MCT, and some women struggle with this conflict. Full discussions about this dilemma will increase the likelihood of adherence to ART. Advice to avoid breast-feeding and caesarean section is also contrary to the norm and questions from family and friends may be difficult to answer. Women may require help with the practicalities of formula feeding as well as financial support.

Living with an HIV diagnosis can lead to fear about a breach of confidentiality, especially if the woman has been diagnosed during the pregnancy or not yet told anyone of her status. There should be a clear local referral pathway for HIV-positive pregnant women. This should include: an HIV physician; midwife/obstetrician; paediatrician and may include a social worker, health advocate and voluntary groups. It is good practice for the woman to have met the paediatric team prior to delivery.

5 Assessment of HIV viral load in pregnancy

HIV RNA copy number varies depending on: the assay employed; biological variation of RNA; and specimen handling [38]. The contribution of these variables to HIV RNA concentrations appears to be of the order of 0.3–0.6 log10 copies/mL, although in some instances it may be as high as 1.0 log10. The Bayer bDNA assay (version 3.0; Bayer PLC, Newbury, UK) generally gives lower HIV RNA copy numbers than the Roche reverse transcriptase polymerase chain reaction (RT PCR) (version 1.5; Roche Products Ltd, Welwyn, UK) but the two assays have been shown to be highly correlated [39]. In order to ensure reliable and accurate quantification of HIV-1 RNA the same assay should be used to monitor viral load. In cases where there are discrepancies between viral load, CD4 cell number and clinical status it is advisable to re-test with another assay in which different nucleotide sequences are used to bind or amplify target RNA.

Accurate quantification of non-B subtypes of HIV-1 is an important requirement for monitoring pregnant women. Mismatches between the probes used in commercial assays and RNA target sequences may result in falsely low or undetectable viral loads among women infected with divergent subtypes. In the United Kingdom, 78% of infections among women attending antenatal clinics are non-B with 61% being subtype A and 29% subtype C [40]. Commercial assays were developed primarily using the B subtype of HIV-1 and their ability to quantify non B subtypes of the virus is variable [41–43].

Detection of antiretroviral drug resistance

In general, sequence based genotyping assays of resistance require approximately 1000 HIV RNA copies/mL and samples with low viral loads may not be sequenced successfully. Current assays may also fail to detect minority species. Drug resistant virus quickly reverts to wild-type in the absence of drug pressure, consequently, use of resistance testing to monitor treatment during pregnancy should ideally be conducted while the woman is still on therapy or within a few weeks of stopping [44]. Sexual transmission of drug resistant HIV is now well documented with prevalence rates, following recent infection, of approximately 10–20% in Europe and North America [45–47]. A recent UK study demonstrated a prevalence of 5% following primary HIV infection [48]. As with viral load assays, commercial resistance assays have been developed using the B subtype of HIV. Non-B subtypes may therefore be amplified and sequenced less efficiently than subtype B. However, non-B subtypes generally have a drug resistance profile similar to subtype B [49].

Any women presenting in pregnancy on a nonsuppressive ART regimen should have resistance testing undertaken [50]. Testing should also be considered for ART naive women, especially if there is an epidemiological risk of primary resistance (e.g. an ART exposed partner).

6 Prescribing antiretroviral therapy in pregnancy

In the UK 15 compounds are currently available for treatment of HIV-1 infection, either licensed or through named patient access. Only ZDV is specifically indicated for use in pregnancy (excluding the first trimester) to prevent HIV MCT. For most ART, prescription in pregnancy is cautioned.

Recommending antenatal HIV testing has significantly increased the number of HIV infected women identified in pregnancy; in addition an increasing number of women already known to have HIV are conceiving on ART. At a preconception consultation or some weeks into the first trimester of pregnancy, they will wish to know whether they should interrupt, continue or change therapy. Few studies have addressed current practice. In this section we will summarize the efficacy data from observational and controlled studies (Table 1a and Table 1b) and make weighted recommendations on the use of ART in pregnancy which balance the needs of the mother and infant with the limitations of the available data.

Table 1.  Evidence of the efficacy of antiretroviral therapy to reduce the risk of mother-to-child transmission of HIV infection. 1a Studies of antiretroviral therapy to prevent mother-to-child transmission in nonbreast feeding populations
Study nameCountriesStudy
Treatment componentsAge HIV
rates (%)
% Reduction
(initial gestation week)
  1. *Discontinued early because of higher rate of transmission (Short/Short arm).

ACTG 076/
ANRS 024 [10]
402ZDV 100 mg
5 × daily
18 months
22.6 placebo
7.6 ZDV
Bangkok Trial [52]Thailand392ZDV 300 mg bd
Nil6 months
18.9 placebo
9.4 ZDV
PHPT [51]
 Long/Long arm
Thailand1437ZDV 300 mg bd
180 days

57.6% cf short/
short interim (0.004)
 Short/Long arm  35Oral6 5.7 
 Long/Short arm  28Oral3 days 8.4 
 Short/Short arm  35Oral3 days 10.6* 
AI455-094 [66]Soweto197ZDV 300 mg bd
d4T 40 mg bd
ddI 200 mg bd
d4T & ddI
ZDV oral
d4T & ddI
d4T & ddI
6 weeks

6 weeks
On-going study
Table 1b .  Studies of antiretroviral therapy to prevent mother-to-child transmission in breast feeding populations
Study nameCountriesStudy
Treatment componentsAge
HIV assessed
rates (%)
% Reduction (P)
(initial gestation week)
DNA PCRPlacebo vs. ZDV
  1. IV = intravenous; ZDV = zidovudine; NVP = nevirapine; 3TC = lamivudine.

RetroCI [57]Côte d'Ivoire230ZDV 300 mg bd
OralNil3 months
6 months
12 months
18 months
24 months
26.1 vs. 16.5
26.1 vs. 16.9
28.5 vs. 18.5
30.1 vs. 21.6
30.1 vs. 22.1
37% (0.07)
27% (< 0.5)
DITRAME [58]Burkino Faso,
Côte d'Ivoire
421ZDV 300 mg bd
Oral1 week
6 months
15 months
27.5 vs. 18
30.6 vs. 21.5
PETRA [74]Republic of
South Africa,
1042ZDV 300 mg bd +
3TC 150 mg bd





6 weeks
18 months
6 weeks
18 months
6 weeks
18 months
19.2 vs. 9.2
26.6 vs. 20.7
19.2 vs. 12.6
26.6 vs. 24.4
19.2 vs. 18.4
26.6 vs. 25.7
50% (0.001)
22% (0.07)
37% (0.02)
8% (0.5)
−3% (0.8)
3% (0.8)
012 [68]
300 mg stat
+ 3 hly vs.
200 mg stat
1 week
stat 48 < 72 h
6–8 weeks

12 months
20.0 vs. 11.8
24.1 vs. 15.7
41% (0.006)

SAINT [72]Republic of
1306NilOral ZDV
+ 3TC vs. NVP
1 week8 weeksZDV + 3TC
vs. NVP
9.5 vs. 12.7
(> 0.10)

Monotherapy for reduction of mother-to-child transmission of HIV

The only published studies are for ZDV, nevirapine (NVP) and ritonavir (RTV).

Nucleoside analogue reverse transcription inhibitors (NRTIs) – monotherapy

The efficacy of ZDV to reduce MCT of HIV-1 has been demonstrated in several large randomized controlled studies [10,51,52] and supported by epidemiological surveys [53–56]. The efficacy of ZDV ranges from: 67%, when started before the third trimester and given to the mother by IV infusion during labour, and to the neonate; to 50% with shorter courses (started at week 36) without a neonatal component, in non-breast-fed babies; to 30% with a similar regimen in breast-fed babies [57,58]. In a non-breast feeding population the transmission rate with ZDV has been reduced to 6–8% [10,54]. When ZDV and pre-labour caesarean section (PLCS) were combined in a cohort of women irrespective of viral load, transmission was reduced to < 2% [9]. Zidovudine will transiently decrease the plasma viral load. In ACTG 076 when ZDV was commenced between weeks 14 and 28 of gestation, therapy was associated with a 0.24 log10 reduction in plasma viraemia at delivery [59]. In the Bangkok study ZDV was commenced at week 36 resulting in a 0.57 log10 reduction in plasma viraemia at delivery [52]. This was considered to account for 80% of the efficacy of ZDV to reduce transmission.

Viral load is an important predictor of transmission and ZDV reduces transmission at all levels of maternal viraemia. However, in mothers with very high viral load (> 100 000 RNA copies/mL) the transmission rate may be > 60% and even with a 2/3 reduction in transmission the risk to the infant would still be around 20%. Additional measures are therefore required for these babies and probably for any mother with a viral load > 10–20 000 copies/mL.

Resistance to ZDV in pregnancy

Sequence changes in the HIV-1 reverse transcriptase associated with decreased viral sensitivity to ZDV have been found, at the time of delivery, in only small numbers of women [60,61]. Higher plasma viral load and longer duration of therapy have been associated with the presence of resistance mutations at delivery [62,63]. In comparison with some other antiretroviral compounds ZDV related mutations develop slowly; therefore shorter courses and restricting the use of ZDV monotherapy to mothers with low viral load and high CD4 counts may limit the emergence of resistant strains. With increasing use of antiretroviral therapy primary (at the time of infection) or secondary (following therapy) acquisition of viral strains with reduced sensitivity to ZDV may become increasingly important [64,65]. Whether resistant mutants are more or less transmissible remains controversial.

The safety and efficacy of didanosine (ddI) and stavudine (d4T) separately and combined is part of an on-going investigation [66].

Protease inhibitors (PIs) – monotherapy

Protease inhibitors are highly protein bound and the data indicate that placental transfer in humans is limited. In a study of 86 pregnant women RTV monotherapy was initiated at gestation week 36 for a mean of 20 days. Median viral load reduction was 2.8 log10 and the transmission rate was 9.5%, but 12 women discontinued treatment, 10 because of elevated liver enzymes [67].

Nonnucleoside reverse transcriptase inhibitors (NNRTIs) – monotherapy

The rapid placental transfer and long half life of NVP have led to studies of the efficacy of NVP to reduce the risk of HIV MCT. In HIVNET 012 two doses of NVP, the first given to the mother in labour and the second to the neonate age 48–72 h, were compared with ZDV initiated in labour and prescribed to the neonate for one week. Transmission was reduced by 50% with NVP [68]. As with short-course ZDV the transmission rates at one year are less than expected (15.7% with NVP and 24.1% with ZDV) and the increased protection of NVP persisted even though the infants were breast-fed [69].

Mutations associated with decreased susceptibility to NVP occur rapidly and frequently on monotherapy [70]. In HIVNET 012, 19% (21/111) of mothers and 11/24 infected infants had genotypic evidence of NVP resistant virus [71].

In the SAINT study transmission rates at eight weeks were not significantly different from the HIVNET 012 study regimen (14%) with ZDV 300 mg plus lamivudine (3TC) 150 mg in labour and twice daily to mother and infant for one week postpartum (10.8%) [72]

Combination therapy for reduction of mother to child transmission

There are no large RCTs of combination therapy use in pregnancy, most studies reported are observational and some with very small numbers.

Dual nucleoside analogue therapy

In a French prospective nonrandomized study of 440 women treated with ZDV plus 3TC from week 32, maternal plasma HIV viraemia was reduced by 0.95 log10 and transmission rate was 2.6% (compared historically with 6.5% on ZDV monotherapy). Treatment was well tolerated by mothers and infants but at 6 weeks postpartum the M184V 3TC mutation was detected in 52/132 women but not in any women treated for less than four weeks [73]. Similar findings were demonstrated in a smaller cohort [61]. In an international RCT in breast-feeding women there was a 22% reduction in transmission at 18 months' follow up compared with placebo in children perinatally exposed to ZDV plus 3TC from 36 weeks gestation to 1 week postpartum, although this did not quite reach statistical significance [74].

Triple therapy with a PI

The most recent analysis of the US WITS cohort (Women and Infants Transmission Study) demonstrated a reduction in transmission with PIs, from 7.8% in mother-infant pairs receiving ZDV monotherapy to 1.1% with ART including a PI [75]. In a study of 76 women taking a PI as part of combination therapy during pregnancy there were 15 preterm deliveries (PTD) (< 37 weeks) but 60% of the mother had identifiable risk factors for PTD, such as a history of PTD, smoking and substance abuse [76]. A possible association between PI use in pregnancy and PTD has also been suggested by other studies [77,78], but was not shown in 462 women participating in ACTG studies during 1998–9 [79].

Triple therapy with an NNRTI

In the ACTG 316 study NVP was added at labour to maternal and neonatal therapy whether it be none, mono, dual or triple. The 1.5% transmission rate among 1174 mother-infant pairs, was less than anticipated at study design (5%), and confirms the potency of current management strategies. Forty-nine percent of mothers had no detectable plasma viraemia at delivery. The study was closed when it became clear that it was not powered to demonstrate any benefit from NVP used in this way [80]. Genotypic studies demonstrated resistance mutations to NVP in 11% (5/46) of mothers with detectable (> 400 copies/mL) plasma viraemia at delivery [81].

7 Maternal drug toxicities in pregnancy

Pre-clinical and clinical safety data has been reviewed elsewhere [82] and can be found on the website. Physiological changes in pregnancy may alter the kinetics of drug absorption, distribution, metabolism and elimination, thereby affecting the drug dosing and side-effects.

Nucleoside reverse transcriptase inhibitors (NRTIs)

Nucleoside analogue drugs are generally well tolerated in pregnancy; reported incidences of adverse effects are similar to those reported in nonpregnant HIV-infected individuals [77]. Nucleoside reverse transcriptase inhibitors can cause mitochondrial dysfunction as they have varying affinity for mitochondrial DNA γ polymerase, which may, result in mitochondrial DNA depletion [83]. Toxicity related to mitochondrial dysfunction has been reported in patients receiving long-term treatment with NRTIs and although this generally resolves with discontinuation of the drug or drugs, fatalities have been reported.

Early in 2001 the US Food and Drugs Administration [84] and the European Agency for the Evaluation of Medicinal Products (EMEA/CPMP/228/01) and Bristol Myers Squibb advised that three pregnant women had died of lactic acidosis following treatment with d4T and ddI (as part of triple therapy) and that a further four cases of lactic acidosis in pregnancy had been reported with this combination. It is not clear whether the frequency of this complication is higher in pregnant than with nonpregnant women. The use of ddI/d4T in pregnancy should for the time being be restricted to woman with resistance to or intolerance of other NRTIs. Monitoring liver function and blood lactate in pregnant women on this combination is recommended.

Protease inhibitors

Hyperglycaemia, new onset diabetes, exacerbation of existing diabetes mellitis and diabetic ketoacidosis have been reported with administration of protease inhibitors [85–87]. Women taking ART which includes a PI have been reported to have a higher risk of developing diabetes mellitus during pregnancy (3.5%) than HIV-negative women or HIV-positive women taking either NRTIs or on no therapy (1.35%) [88].

In a study of 86 HIV-positive, treatment naive women, side-effects of RTV monotherapy commenced in the 36th week of pregnancy were prominent [67]. Other observation studies have demonstrated similar side-effects to those seen in the nonpregnant population [76,89–91].


The use of NVP as part of combination antiretroviral therapy was retrospectively reviewed in a cohort of 46 HIV-infected pregnant women. Although, initiated in 30 during pregnancy, NVP was well tolerated and the only adverse effects probably related to NVP were rash (2) and biochemical hepatitis (2) [92].

Other drug treatments

Women on antiretroviral therapy are commonly on other therapies. In a multicentre retrospective study of 148 infants exposed to ART in utero the risk of congenital malformation was significantly raised in those exposed in the first trimester to folate antagonists used for Pneumocystis carinii pneumonia prophylaxis combined with ART [93]. The need for prophylactic therapies in women of child-bearing potential should be regularly reviewed as immune function recovers on ART.

8 Obstetric management of pregnancy and delivery

Since the majority of vertical transmission occurs during the intrapartum period, it was proposed that prelabour caesarean section with avoidance of labour and the birth canal might reduce the risk of transmission [94–96]. Definitive support for the protective effect of PLCS came from a trans-Atlantic meta-analysis of 15 prospective cohort studies [8] and a randomised controlled study of mode of delivery in Europe [9](Table 2).

Table 2.  Evidence of the efficacy of prelabour caesarean section to reduce the risk of mother-to-child transmission of HIV
Transmission rate (%)OR
(95% CI)
Transmission rate (%)OR
(95% CI)
Transmission rate (%)OR
(95% CI)
& other
SVD, other
  • PLCS = pre-labour caesarian section; SVD = spontaneous vaginal delivery; MOD = mode of delivery; ZDV = zidovudine; NS = not significant by Fisher's exact test; EMCS = emergency caesarean section.

  • a

    Full 076 protocol with antepartum, intrapartum, and postpartum ZDV.

  • b

    Advanced maternal disease – an AIDS diagnosis and/or CD4 < 200/ < 14%.

15 cohorts
Not given
pre 1997
US & Europe
n = 8533 [96]

Mode of
Delivery study
38 weeks [52]
n = 436



Actual CS
to SVD


Actual SVD
(95% CI)


(95% CI)
+ ZDVb

Actual CS
to SVD

Actual SVD
(95% CI)


(95% CI)
to PLCS +
CD4 < 200

Actual CS
+ CD4
< 200
to SVD +
CD4 < 200

Actual SVD
+ CD4
< 200


The meta-analysis of 8533 mother-child pairs confirmed that the vertical transmission rate was 50% lower in women who underwent caesarean section before the onset of labour or rupture of the membranes. This protective effect persisted when ART was used. In the RCT, there was an over all reduction in transmission of 70%, in a cohort of women with all levels of CD4 and disease status. The maternal complication rate of PLCS was very low.

In a further meta-analysis of the same cohorts, the risk of transmission increased approximately 2% for every hour of rupture of membranes (ROM) up to 24 h [97]. In women with an AIDS diagnosis the risk of transmission increased from 8% after 2 h of ROM to 31% after 24 h of ROM.

Maternal viral load data were not available for the meta-analyses or the mode of delivery RCT. It has subsequently been suggested that mothers with very low/undetectable viral load might consider vaginal delivery. In a recent meta-analysis of seven prospective studies from the US and Europe there were 44 transmissions in 1020 deliveries where maternal plasma viral load was < 1000 HIV RNA copies/mL at or around delivery [4]. The transmission rate for mothers on ART was 1% compared to 9.8% for no treatment. In multivariate analysis transmission was lower: with ART; caesarean section; greater birth weight; and higher CD4 count. These data were collected at a time when HIV-RNA PCR assays were less sensitive than currently, and suggest a protective effect of both ART and caesarean section even at very low viral loads. However, whether caesarean section has an additive effect at a delivery viral load of < 50 HIV RNA copies/mL is not known. Cost effectiveness analysis suggests that the vertical transmission rate would have to be less than 0.5% before caesarean section would be no longer cost effective [98].

The benefit of PLCS in reducing the risk of vertical transmission is now clear, especially for women with detectable viral load. However, possible surgical complications must be included in any discussion regarding mode of delivery. Some studies have suggested that postoperative complications are increased in HIV infected women compared to the background population [99–101] others [9] found no difference. Reported complication rates are related to the level of maternal immunocompromise.

Management of delivery

A plan for delivery management should be made in good time. Delivery by PLCS should be undertaken at 38 weeks [102], but there is some evidence that women infected with HIV do labour early and if there are concerns that delivery may occur sooner, the date for the caesarean section could be brought forward [24,103]. If labour starts prior to the planned delivery date, intravenous ZDV should be commenced if this is part of the ART regimen, but the caesarean section should not be delayed to complete the induction course of ZDV. If there is premature rupture of membrane, with or without labour, the risk of HIV transmission must be balanced with the risk of premature delivery. There is no known contra-indication to the use of short-term steroids to promote fetal lung maturity in women with HIV.

Although a PLCS is the recommended mode of delivery for women with HIV, some women may still wish for a vaginal delivery. This may be an important consideration for women who are planning to return to a country where subsequent caesarean section deliveries may not be possible or safe. If a vaginal birth is planned the membranes should be left intact until delivery is imminent. Fetal electrodes and blood sampling must be avoided. If there are signs of fetal distress consideration should be given to performing an emergency caesarean section as the risk of vertical transmission of HIV is increased after emergency vaginal obstetric intervention.

Other pregnancy issues

Prenatal diagnosis

HIV infected women contemplating invasive prenatal diagnosis should be counselled in a specialist fetal medicine unit and the best noninvasive screening tests available should be employed in the first instance. Data on the risks of transmission with chorionic villis sampling or second trimester amniocentesis are sparse. Administration of ART to cover the procedure should be considered.

Women who become unwell in the third trimester of pregnancy

Presentation in the third trimester with signs and symptoms of pre-eclampsia, cholestasis or other liver dysfunction may be due to these complications of pregnancy, but may also be adverse effects of ART. Any woman presenting with vomiting, malaise or oedema should be investigated for acidosis, hepatitis, pancreatitis, and disseminated intravascular coagulation whether or not she has hypertension or proteinuria. If there is lactate acidaemia/acidosis, discontinuation of ART, even at this crucial time, must be seriously considered.

9 Interventions to reduce mother to child transmission of HIV

Table 3 details nine clinical scenarios where a different approach to therapy in pregnancy may need to be considered and the issues relating to each scenario are discussed in this section. Pre-labour caesarean section at 38 weeks is recommended as the mode of delivery in all scenarios. Consideration for vaginal delivery may be given to women on stable therapy with undetectable viral load, but, there are, as yet, insufficient data for this to be formally recommended.

Table 3.  Clinical scenarios
Clinical scenariosRecommendations
ScenarioCD4 countViral loadTreatment
Gestation at
Antiretroviral therapyMode
  • Resistance testing should be carried out in women failing therapy. Also consider resistance testing in ART naive mothers.

  • Closely monitored / actively managed vaginal delivery may be considered in women with VL < 50.

  • START = short term combination antiretroviral therapy; HAART = highly active antiretroviral therapy; ZDV = zidovudine; MOD = mode of delivery; ART = antiretroviral therapy; IV = intra venous; 3TC = lamivudine; PI = protease inhibitor; NNRTI = non-nucleoside transcriptase inhibitor; d4T = stavudine; VL = viral load.

  • a

    An alternative to regimen (1) would be to use regimen (2) START for women with ‘good CD4, low VL, ART naive.

  • b When choosing the PI/NNRTI, consider the short-term toxicities of nevirapine (rash, hepatitis), and short term/long term toxicities of e.g. nelfinavir (diarrhoea, lipid derangement's etc).

  • c If ZDV unexposed, consider changing one NRTI to ZDV after 1st trimester. Substitute efavirenz with another drug, if presents in 1st trimester.

  • d

    If not on ZDV, not necessary to give IV ZDV infusion in labour, as long as all regular doses given at correct times. It may be contra-indicated to give ZDV if d4T is part of the regimen.

  • e

    When changing treatment consider the following: VL trend; absolute VL; potential toxicities; available drug choices; At stable low viraemia e.g. < 5000 it may be possible to continue current treatment regimen.

  • f

    Use NVP in preference because of rapid anti-viral effect and high transplacental concentrations.

  • g

    See table 4 for appropriate infant doses of ZDV, 3TC, NVP for 4-6 weeks of treatment.

  • h Premature or sick infants may not tolerate oral therapy, only available IV preparation is ZDV, see table 4 for appropriate infant doses.

  • i

    If there is extreme prematurity with rupture of membranes, a period of conservative management with steroids, antibiotics and START, may be more appropriate for the infant, seek expert advice.

(1) Woman does> 200 /< 10 000Naive< 32 weeksZDV monoZDV monoZDV monoStop therapyPLCS atMOD – A
not need> 21%   076 regimen076 regimen076 regimenafter delivery38 weeks 
treatment for    (BD dosing)IV at delivery(BD dosing)  ART – A for
own health    commence late for 4-6  prevention
(asymptomatic    2nd / early 3rd weeks  of
good CD4)    trimester, but    transmission
Low VL    before 32
(2) Woman does> 200 /> 10 000Naive< 32 weeksSTARTZDVZDV monoIf stoppingPLCS atMOD – A
not need> 21%   suggest: inc076 regimen076 regimentherapy after38 weeks 
treatment for    ZDV by 076IV at delivery(BD dosing)delivery, give ART – C for
own health    regimen ++for 4-6consideration prevention
(asymptomatic    3TC +Oral doses ofweeksto timing of of
good CD4)    PI / NNRTIbother ART as stopping NVP transmission
High VL    Commence late
2nd trimester
 in relation to
NRTIs, e.g.
3–5 days prior
(3) Woman< 200 /AnyNaive< 32 weeksHAARTZDVZDV monoContinuePLCS atMOD – A
needs< 21%   suggest: inc076 regimen076 regimenmaternal38 weeks 
treatment for    ZDV by 076IV at delivery(BD dosing)regimen after ART – A for
own health(200–300Any  regimen ++for 4–6delivery. maternal
but steep   3TC +Oral doses ofweeksMake sure no health. C for
slope of   PI/NNRTIbother ART as doses missed prevention
decline)   Defer until
after 1st
usual pre-
delivery time
(200–300)High  trimester, if
(4) WomanAny< 50OnAny timeContinue:cIf on ZDVMono-ContinuePLCS atMOD – A
presents  effective  076 regimentherapymaternal38 weeks 
pregnant on  ART  IV at deliverydcomponentregimen after ART – A for
effective ART     +
Oral doses of
other ART as
usual pre-
of the
for 4–6
Make sure no
doses misssed
delivery time
health. C for
(5A) On non-AnyVL > 1000May be< 32 weeksWith resistanceIf on ZDV,Mono-therapyContinuePLCS at 
suppressive resistance testmultiple data.076 regimencomponent ofmaternal38 weeks 
ART  drug ConsiderIV at deliverydthe mother'sregimen  
VL = 50–1000classes change to best+regimen toafter MOD – A
monitorexposed optioneOral doses ofwhich nodelivery.  
closely for   other drugs inresistanceMake sure ART – B for
trend, if ↑  Seek expertART as usualfor 4–6 weeksno doses material
resistance test  advicepre-delivery missed
delivery time
 health. C for
of transmission
(5B) Failing
therapy in late
Any> 1000 > 37 weeksSelect best combination from therapy history
Seek expert advice
Seek expert
38 weeks
ART – B for
health. C for
of transmission
(6) LateCD4 < 200 + any / no VLNaive>32 weeks,STARTZDVIf > 4 wksContinuePLCS atMOD – A
presentation,CD4 > 200 + VL > 10 000(usually)but beforesuggest: inc076 regimenmaternal Rx:maternal38 weeks 
but before   delivery.ZDV by 076IV at deliveryZDV mono.regimen ART – C for
deliveryNB take blood for base line Time to getregimen ++If < 4 wksafter prevention
 CD4, VL and resistance prior CD4, but3TC +Oral doses ofmaternal Rxdelivery of
 to any ART maybe not VLNVPf
other ART as
usual pre-
especially if
BL VL high:
Triple ARTg
until get
CD4 + VL.
Ideally, do
not stop
ART until
VL < 50
(7) Presents inUnknownUnknownNaiveIn labourNoneZDVTripleIf labour32 wks – term IfMOD – D
labour –  (usually)Any gestation 076 regimenARTg,hstopslabour progresses 
membranesNB take blood for base line intact IV at delivery continueconsider CSART - C for
intactCD4, VL and resistance prior membranes + Oral doses START. Prevention
 to any ART   of NVP + 3TC If delivered
- continue
ART until
get CD4 +
VL. Ideally,
do not stop
ART until
VL < 50
< 32 wks
IV antibiotics,
tocolysis, steroids.
If labour
consider CS
(8) PresentsUnknownUnknownNaiveIn labourNoneIV ZDVTripleContinueIf labourMOD – D
with rupture  (usually)Any gestation 076 regimen +ARTg,hmaternalprogresses 
of membranes,NB take blood for base line   Oral doses of STARTconsider SVDART – C for
+/- labourCD4, VL and resistance prior to   NVP + 3TC until get+ antibioticsprevention
 any ART No labour
Any gestation
   CD4 + VL.
Ideally, do
not stop
ART until
VL < 50

No labour
+ antibiotics
(9) PresentsUnknownUnknownNaiveAfter deliveryNoneNoneTripleAssess ART – C
after delivery (usually)    ARTg,hmaternal  
NB take blood for base line    Start ASAP,need for  
CD4, VL and resistance prior to    less likely toART  
 any ART    be effective
if > 48 hrs

Scenario 1: Women who do not yet require treatment for their HIV disease

Asymptomatic women with low plasma viral loads (< 10–20 000 HIV RNA copies/mL) and good CD4 counts (> 200–350 cells/µL) do not require ART for their own health [3], ZDV monotherapy with PLCS can be advocated for them to reduce vertical transmission as well as ART exposure in pregnancy. An alternative regimen for those wishing to deliver with no detectable viraemia would be short-term triple antiretroviral therapy (START; see below).

Scenario 2, 3, 4: Women who require treatment for their HIV disease

US [104] and UK [3] guidelines currently recommend that women with advanced HIV, who would normally be treated with ART, should be managed as if they were not pregnant. For these women, for asymptomatic women with high viral load and for women with prior ZDV exposure and ZDV resistant virus combination therapies are recommended. Until recently it was thought that, with the exception of the third group, ZDV should always be included as it was the only compound shown to have reduced transmission. However, there are now data for both RTV and NVP monotherapies which suggest equivalence with ZDV (see above). Since triple combinations reduce plasma viraemia to less than 50 copies/mL, vertical transmission, unless occurring prior to the initiation of therapy or in mothers in whom the viral load is underestimated, is likely to be a rare event. Although biologically plausible this assumption still needs to be confirmed epidemiologically.

For treatment naive mothers requiring combination therapy (scenarios 2 & 3), consideration should be given to safety and efficacy data, tolerability, and whether treatment is likely to be continued after delivery. The most extensive safety and prevention of transmission data are for ZDV and 3TC; however, the rapid development of 3TC resistance precludes dual therapy and optimally suppressive therapy should be recommended. Limited experience with triple therapies including a PI or NVP have been reported. If treatment discontinuation is planned, e.g. for a mother who has elected to take START in pregnancy (scenario 2), a protease inhibitor such as nelfinavir (NFV) might be preferred because the long half-life of NVP might result in the emergence of resistance. Alternatively the nucleoside backbone of the regimen may be continued for a few days after stopping NVP to avoid inadvertent ‘NVP monotherapy’. However there are no trial or observational data that address this issue.

Scenario 4,5: Women who conceive on therapy

Women who conceive whilst on antiretroviral therapy may wish to discontinue therapy during the first trimester. There are no data to support or refute this. Consideration should be given to maternal health and immune status at the time of initiating therapy as well as at the current time – viral rebound will occur within two to three weeks and ‘strategic treatment interruptions’ have been associated with significant CD4 lymphocyte decline. This may not only jeopardize maternal health but in theory may result in reactivation of infections associated with congenital abnormalities, e.g. cytomegalovirus. Many women will not realize or report their pregnant state until well into the period of organogenesis. If the mother's treatment is failing then this should be changed in time to ensure the lowest possible viraemia at the time of delivery. Resistance testing can help to identify the best options. Only exceptionally should antiretroviral therapy be initiated in or changed during the first trimester. Reasonable exceptions include serious illness for which antiretrovirals are the only recognized therapy.

Scenario 6,7,8: Women who present late in pregnancy

For women who present very late in gestation or in labour, for whom no transmission risk assessment has been possible it seems sensible to include compounds that rapidly cross the placenta and have reliable pharmacokinetics in the neonate. In this situation the most effective antiretroviral therapy is NVP. Given the high risk of resistance developing in the mother with even a single dose of NVP monotherapy [71,81], two other compounds, usually nucleoside analogues such as ZDV and 3TC, should be started. During labour ZDV should preferably be infused IV and all treatment should be continued after delivery until the mother's clinical, immunological and virological status have been determined. Consideration should be given to continuing triple therapy until maternal plasma viraemia has become undetectable which would theoretically reduce the risk of development of resistance to current therapy (see above for details on stopping NVP in relation to other drugs).

Scenario 9: Presentation of the infant after delivery

Where it is only ascertained after delivery that an infant has been born to an HIV infected mother, postexposure prophylaxis should be offered as soon as possible. There is observational data that ZDV can reduce transmission in this situation if given within 48 h of delivery [56]. Although there are no data, it would seem logical and consistent with other current PEP recommendations for high risk exposure to offer triple combination therapy for four weeks [105] (see below for ART choices).

10 Management of infants born to HIV infected mothers

Exposure to ART

Most neonates born in the UK to mothers known to have HIV will be exposed to ART in utero, during delivery and after birth for the first 4–6 weeks of life. Neonatal drug metabolism is generally slower than that of older infants or children, and premature neonates have even less efficient metabolism. Neonatal dosing regimens have been developed for most of the nucleoside analogues and for NVP (Table 4). Adequate neonatal blood levels are difficult to achieve with NFV and there is little experience of other PIs [106]. In Europe, the only ART available for intravenous (IV) use in sick and/or premature neonates, unable to take oral medication, is ZDV. Reduced oral and IV dosing schedules for premature infants have only been developed for ZDV. Neonatal metabolism of NVP is induced where there is antenatal exposure, so if this drug is subsequently given to the neonate a different regimen is required [107,108].

Table 4.  Infant doses of antiretroviral therapy and PCP prophylaxis
Zidovudine (ZDV)Term infant: oral dosing
2 mg / kg every 6 hours or,
4 mg / kg every 12 hours
Premature infant: oral dosing
1.5 mg / kg every 12 hours for first 2
weeks, then 2 mg / kg every 8 hours to
Sick infant: unable to oral feed
Term infant IV dose, 1.5 mg / kg every 6
Premature infant IV dose 1.5 mg / kg every
8 hours.
ACTG 076 study [10]
(*see below dosing
with 3TC)
Under study in
PACTG protocol
Anaemia and neutropaenia, more common with
combination therapy in mother and infant. May
need to stop therapy early, or substitute another
Should not be administered with d4T, potential
Negative competitive effect.
Lamivudine (3TC)2 mg/kg every 12 hoursMoodley et al., JID
1998: 178, 1327-33
Didanosine (ddI)50 mg/m2every 12 hoursACTG 249
Livingston et al.,
5th CROI 1998 #226
Difficult to separate dosing from feeding in
very young infants. May cause GI symptoms.
Stavudine (d4T)1 mg/kg every 12 hoursUnder study in
PACTG protocol
Should not be administered with ZDV,
potential negative competitive effect.
Abacavir (ABC)2 mg/kg every 12 hoursJohnson et al.
7th CROI #720
Dosing still under
Zalcitabine (DDC)No known dose  
Nevirapine (NVP)Stat dosing regime: single dose to mother
in labour and to infant at 48-72 hours of
age 2 mg/kg
Continuous dosing regime:
5 mg/kg once daily for 2 weeks, then 120
mg/m2 every 12 hours for two weeks, then
200 mg/m2 every 12 hours to completion
Musoke et al., AIDS
1999, 13: 479–86.
Under study in
PACTG protocol
Consider whether mother has received NVP
before delivery. This regime may under dose
infants with exposure to NVP in utero and
hepatic enzyme induction. It may be necessary
to start with the higher dose, as if from 2 weeks
of age. (It may be helpful to monitor levels)
Efavirenz (EFV)No known doses  
Delavirdine (DLV)No known doses  
Ritonavir (RTV)Still under study, no dose available yetPACTG 354 
Amprenavir (APV),
Indinavir (IDV),
Saquinavir (SQV),
Nelifinavir (NFV),
No known doses
No known doses
No known doses
No known doses
No known doses
Combination therapy
3TC 2 mg/kg every 12 hours,
ZDV 4 mg/kg every 12 hours
*Mudd et al. 1999 2nd
Conf Global
Strategies #360
3TC +
NFV 10 mg/kg every 8 hours,
3TC 2 mg/kg every 12 hours,
ZDV 2.6 mg/kg every 8 hours
Bryson et al. 2000
7th CROI #715
Dosing of NFV was inadequate, now
trialing dose of 40 mg / kg 12 hourly.
Co-trimoxazoleCo-trimoxazole 900 mg/m2,
< 0.25 mg m2– 120 mg
0.25–0.39 m2– 240 mg
once daily Monday, Wednesday, Friday
Simonds et al.
NEJM 1995; 332:
May cause rash, or bone marrow suppression.
Most important to give if no interventions to
reduce transmission have been undertaken.

When to consider combination ART in neonates

There have been very few studies of combination therapy in neonates [90,106,109]. Whether combination ART to the neonate has any additional benefit over single drug treatment is not known. Where a mother on combination therapy delivers with no detectable viraemia, our current practice is to use monotherapy for the neonate, as this is easier for the family and may reduce the incidence of adverse events in the neonate. The drug chosen from the maternal combination is usually the NRTI with the best-known infant pharmacokinetics (e.g. ZDV, 3TC, etc.). With infant feeding patterns, it is difficult to separate drug dosing from feeds, so drugs without food restrictions are preferred. Zidovudine should not be given to an infant born to a mother who is receiving d4T because of their theoretical competitive interaction.

There are two situations where combination treatment for neonates should be considered:

  • 1Where an unplanned delivery occurs, either prematurely prior to starting ART, or after a late presentation when details of maternal HIV parameters may not be available (scenarios 6,7,8).
  • 2Where the mother is only found to be HIV infected after delivery (scenario 9).

There have been no trials of combination ART for infants in these situations, but as in other postexposure prophylaxis cases, it should also be considered for neonates. We have used ZDV, 3TC and NVP for infants born to ART naive women, but for non-naive mothers other combinations might be required if there is a possibility of resistance. Resistance testing should be carried out in the mother in such a situation and as early as possible in the infant as well, if infected.

Duration of treatment for neonates

In PACTG 076, ZDV was administered for 6 weeks after birth and this subsequently became standard of care [10]. However, in a recent Thai study, a three day course of neonatal ZDV was equivalent to one months' therapy provided the mother had received ZDV from 28 weeks' gestation [51]. In UK centres, neonates are currently treated for 4–6 weeks.

Side-effects of treatment


Long-term side-effects of perinatal exposure to ART can be considered in four main categories: teratogenic; carcinogenic; developmental; and mitochondrial. Teratogenicity is most likely to be a problem with first trimester exposure to ART ± other drugs. To date, no increase in total number nor any specific fetal abnormality have been identified, but the voluntary reporting rate is disappointingly low [110]. Detailed fetal anomaly scanning at 18–20 weeks is advised after first trimester exposure. NRTI exposure could theoretically lead to a long-term risk of carcinogenicity, but no increased rate has been identified [111]. In the UK, the register of cancer and deaths is linked anonymously to the register of infants born to mothers with HIV to highlight any relationship. So far, no adverse developmental effects of ART exposure have been demonstrated in children [112]. Mitochondrial toxicity, including fatal encephalopathy, after perinatal ART exposure was first reported from the prospectively followed French cohort [113], and in an updated analysis toxicity was suspected in 18/2547 (0.7%) of exposed uninfected children [114]. Deaths related to mitochondrial toxicity have not been identified in other large cohorts [115]. Long-term follow up of infants from PACTG 076 found two ZDV exposed children with unexplained retinopathy and cardiomyopathy, both potential mitochondrial toxicities [112]. Another study did not demonstrate any evidence of cardiomyopathy [116].


Short-term, acute mitochondrial toxicity may also be a problem in the new-born, exacerbating the metabolic stress of delivery. A small number of sick infants have been reported with severe lactic acidosis, multisystem failure and anaemia, not attributable to any other cause, all have recovered with supportive care [117,118]. Elevated lactic acid levels have also been found in asymptomatic ART exposed infants [119].

Symptomatic neonatal anaemia is increasingly reported in infants exposed to ART, and may be worse with combination ART [120]. Transfusion is rarely required and most children respond to discontinuation of marrow suppressive therapy. Abnormal liver function has been reported in infants exposed to ZDV with 3TC [120]. In a small study, infants exposed to PIs in utero had significantly higher γGT levels than therapy naive or ZDV monotherapy exposed infants [121]. In a study of the safety and tolerance of RTV in combination with 3TC and ZDV given to mothers, 3/6 infants were born prematurely, two with severe hypoglycaemia, whilst the third infant, delivered severely preterm, died. One infant had grade 3/4 hyperbilirubinaemia, one had neutropenia and two were significantly anaemia [90]. High triglyceride levels were documented transiently in 2/6 infants treated with NFV (combined with d4T plus ddI), despite the fact that inadequate plasma levels of NFV were achieved [109].

In view of the metabolic abnormalities increasingly reported with ART, exposed neonates should have base-line investigations including: full blood count; pH; lactate; glucose; urea and elecrolytes; liver function tests; trigylcerides and amylase; as well as diagnostic HIV PCR tests. These tests may be repeated with each set of HIV diagnostic samples.

Laboratory diagnosis of HIV infection in non-breast fed infants

The gold standard test for HIV infection in infancy is HIV DNA PCR on peripheral blood lymphocytes [122]. As most infants are infected intrapartum and blood levels may still be very low, HIV DNA is not amplified from all infected infants at birth. Indeed a positive HIV PCR result within 72 h of birth has been taken as evidence of intrauterine transmission [96]. However, by 3 months of age > 95% of non-breastfed HIV infected infants will be PCR positive. In view of the genomic diversity of HIV a maternal sample should be amplified with the infant sample to confirm that the primers used can detect the maternal virus. If maternal virus cannot be detected by HIV DNA PCR, then different primer sets, or a different test, e.g. HIV culture/HIV-RNA PCR should be used [123]. Our current practice is to test infants at one day, one month, and three months of age. If all these tests are negative and the baby is not being breast fed, the parents are informed that the child is not infected. Loss of maternal antibodies is subsequently confirmed at 18 months of age. It is not necessary to carry out other surrogate or less sensitive tests (e.g. CD4, immunoglobulins, p24 antigen) unless there is a concern about the sensitivity of the HIV DNA PCR. Use of HIV-RNA PCR (viral load tests) for infant diagnosis has been increasingly reported, but these tests give false positive results of low copy number, causing unnecessary worry to families [124,125].

There is no current evidence that perinatal ART exposure delays the detection of HIV infection in infants. If an infant is found to be HIV infected despite perinatal ART exposure, then urgent HIV resistance testing is required to delineate the reasons for treatment failure and guide further treatment.

Prophylaxis, immunizations and clinical monitoring

Primary Pneumocystis carinii pneumonia (PCP) in infants with HIV remains a disease with a high mortality and morbidity [126]. However, as the risk of neonatal HIV infection has fallen to < 1% with antenatal interventions, the necessity for PCP prophylaxis has declined and in most European countries is no longer prescribed. PCP prophylaxis should be prescribed to infants born to mothers who received no interventions.

Infants born to HIV infected mothers should follow the routine immunization schedule, except that, BCG should not be given until the infant is confirmed un-infected. The risk of live oral polio vaccination (OPV) to HIV infected infants and their carers has not proved significant, so it is unnecessary to substitute killed injected polio vaccine. The hepatitis status of the HIV infected mother should be ascertained, so that Hepatitis B vaccination can be carried out if necessary. PCR amplification is required for early diagnosis of HCV infection in infants born to HCV infected mothers.

Reporting and long-term follow up

Clinicians caring for women with HIV and their children have a responsibility to report them to the UK national study of HIV in pregnancy*, via the Royal College of Gynaecologists antenatally and the BPSU after birth. All women who receive ART in pregnancy should be registered prospectively with the International Drug Registry. BPSU follow up of HIV exposed uninfected infants will now extend beyond 18 months to at least 5 years of life to assess for longer term affects of ART exposure (Pat Tookey, personal communication).

11 Infant feeding

Breastfeeding is an important route of transmission. In the UK, where safe infant feeding alternatives are available, HIV-infected women are advised to refrain from breast feeding. Where a mother acquires primary HIV infection after delivery, whilst breast feeding, the risk of transmission through breastfeeding is about 30% [7]. In established infection, the additional risk of transmission through breast feeding, over and above the intrauterine and intrapartum contribution, is estimated to be between 7 and 22% [7]. These estimates agree with those from prospective studies, and from an RTC [127].

If a woman chooses to breast feed despite the evidence, then she should be advised to breastfeed exclusively as this may reduce the risk of transmission [128], and for a shorter rather than longer period. If she is taking antiretroviral medication it should be explained that there is no evidence that this will protect the infant. Although ART is likely to reduce free virus, its effect on cell associated virus in the milk is not known.

12 Pregnancy in women with HIV-2 infection

HIV-2 is endemic in West Africa and other areas of high prevalence include parts of India and Portugal. Thirty-two cases of HIV-2 infection have been reported in the UK and 11 of these infections were in women [129]. HIV-2 appears to be less pathogenic than HIV-1 with prolonged periods of asymptomatic infection and slower rates of disease progression reflecting a lower rate of viral replication [130,131]. Vertical transmission rates of HIV-2 are also low, 0–4% in breast fed infants, in the absence of any interventions [132–134]. Interventions to reduce transmission of HIV-2 in pregnant women have not been clearly defined.

Treatment is indicated in pregnancy if the woman is symptomatic and CD4 cell numbers are < 300 cells/µL as this is usually associated with a detectable viraemia [135]. NNRTIs have little inhibitory activity against HIV-2 and are therefore not recommended, but the virus is susceptible to NRTIs and some PIs. Although currently there is no evidence to support interventions such as caesarean section or ART in women with HIV-2, they should probably be managed in a similar way to HIV-1 infected women. The risk from breast milk is probably lower than for HIV-1 but it may be advisable to avoid this method of feeding. Although quantification of HIV-2 RNA is the preferred method for monitoring disease and responses to treatment no commercial assays are currently available. There is one laboratory in the UK which can provide this service (contact Dr Judy Brewer, Royal London Hospital and St Bartholomew's). Infants born to infected women can be monitored for HIV-2 DNA PCR and loss of HIV-2 antibodies by 12–18 months of age.


  • *

    National Survey of HIV in Pregnancy – Royal College of Obstetricians and Gynaecologists. British Paediatric Surveillance of HIV in Children.

  • Antiretroviral Pregnancy Registry (in Europe managed by GlaxoSmith Kline) GlaxoSmithKline Ltd, Greenford Road, Greenford, UB6 0HE Tel.: no: 020 89664500; Fax: 020 89662338. Co-ordinator for both Ms Pat Tookey 020 78298686, E-mail: