7.0 Obstetric management

7.1 Antenatal management

7.1.1 Fetal ultrasound imaging should be performed as per national guidelines regardless of maternal HIV status. Grading: 1D

The National Screening Committee [11] and the NICE antenatal guidelines [12] recommend that ultrasound screening for fetal anomaly should be offered to all pregnant women between 18 + 0 and 20 + 6 weeks' gestation. There is no evidence to alter this for women infected with HIV.

In the past, because of a theoretical increased risk of anomaly due to first trimester ART exposure, more detailed ultrasound scanning (i.e. in a fetal medicine unit) has been considered. The evidence from prospective reports of first trimester ART exposure to the APR [7] does not support the need for increased surveillance with the most commonly prescribed therapies (listed in Appendix 4), although with newer medication the knowledge base is inevitably limited. APR reports on the frequency and nature of birth defects and ART are updated every 6 months (http://www.apregistry.com/).

7.1.2 The combined screening test for trisomy 21 is recommended as this has the best sensitivity and specificity and will minimize the number of women who may need invasive testing. Grading: 2C

Clinical Guidance 62 (CG62) [12] also recommends that all women should be offered screening for trisomy 21. The most effective screening is with the combined test at 11 + 0 to 13 + 6 weeks' gestation. This includes maternal age, nuchal translucency, βHCG and pregnancy-associated plasma protein A. In the general population this has a detection rate of 92.6% with a false positive rate of 5.2% [13].

For women who present too late for the combined test, the most clinically and cost-effective serum screening test (triple or quadruple test) should be offered between 15 + 0 and 20 + 0 weeks [12]. However, significantly increased levels of βHCG, α-fetoprotein and lower levels of UE3 (the elements of the ‘triple test’) have been observed in the HIV-positive population [[14][[15][#[16]]Ent]211] while a reduction in βHCG in patients treated with PI-based [17] or with NNRTI-based HAART has been reported. As Down's syndrome is associated with increased βHCG, theoretically, HIV infection per se may increase the false-positive rate in women and thus increase the number of invasive tests offered compared with the uninfected population. Pregnancy-associated plasma protein A and nuchal translucency are unaltered by HIV infection or ART [18] and are thus the preferred screening modality.

7.1.3 Invasive prenatal diagnostic testing should not be performed until after HIV status of the mother is known and should be ideally deferred until HIV VL has been adequately suppressed. Grading: 1C

  • Limited data suggest amniocentesis is safe in women on HAART. There are minimal data on other forms of prenatal invasive testing.
  • All clinicians performing a prenatal invasive test should know the woman's HIV status, and if necessary delay the invasive test until the HIV result is available.
  • Where possible, amniocentesis should be deferred until VL is <50 HIV RNA copies/mL.
  • The fetal medicine team should discuss management with an HIV physician if the woman is HIV positive and has a detectable VL.

7.1.4 If not on treatment and the invasive diagnostic test procedure cannot be delayed until viral suppression is complete, it is recommended that women should commence HAART to include raltegravir and be given a single dose of nevirapine 2–4 h before the procedure. Grading: 1D

The French Pediatric HIV Infection Study Group observed a relative risk of HIV transmission of 1.9 (95% CI 1.3–2.7; P = 0.003) with ‘antenatal procedures’ that included amniocentesis, cerclage, laser therapy and amnioscopy [19]. This study was conducted between 1985 and 1993 and, of the 1632 mother–infant pairs (overall transmission 19%), only 100 mothers had received zidovudine, mostly for advanced HIV infection.

There are few studies on the safety of invasive testing in the HAART era. A study of 9302 pregnancies in France in 2009 (of which 166 had an amniocentesis) showed that the risk of MTCT in the untreated rose from 16% to 25% in those who had an amniocentesis, in those on zidovudine alone the risk rose from 3.3% to 6.1% and in those on HAART there were no transmissions in 81 mothers who underwent amniocentesis [20]. VL data were not reported, but in other settings suppression of VL reduces transmission.

A further study of nine women in France on HAART in 2008 [21] and 17 women on HAART in Portugal (1996–2009) showed no transmissions, while transmission occurred in one of six women either not diagnosed with HIV prior to amniocentesis, or not treated before the procedure. There are no studies and few case reports in the HAART era reporting on chorionic villus sampling or cordocentesis [22]. For evidence relating to choice of ART to reduce transmission risk associated with amniocentesis, see Section 5.4 on late presentation.

7.1.5 ECV can be performed in women with HIV. Grading: 2D

  • ECV should be offered to women with a VL <50 copies/mL and breech presentation at >36 + 0 weeks in the absence of obstetric contraindications.

There is less obstetric risk to the baby and mother when the fetus is head-down at the time of birth. ECV is a procedure by which the fetus, which is lying bottom first, is manipulated through the mother's abdominal wall to the head-down position. If the fetus is not head down by about 36 weeks of pregnancy, ECV reduces the chance that the fetus will present as breech at the time of birth, and thus reduces the chance of CS. There is no published evidence that helps decision-making regarding ECV in the HIV-positive pregnant woman. For the general maternity population, ECV is recommended [12]. The question of whether ECV might increase the risk of MTCT of infections such as HIV is important and, in the absence of direct evidence, we have reviewed the relevant biological evidence and concluded that maternofetal transfusion, as a consequence of this procedure, is extremely rare, and unlikely to be precipitated by ECV [23]. It is also reassuring that in a randomized trial of fundal pressure to expel the baby during CS, no evidence of maternofetal transfusion was found [24].

7.2 Mode of delivery

7.2.1 Vaginal delivery is recommended for women on HAART with HIV VL <50 HIV RNA copies/mL plasma at gestational week 36. Grading: 1C

  • For women taking HAART, a decision regarding recommended mode of delivery should be made after review of plasma VL results at 36 weeks.
  • For women with a plasma VL <50 HIV RNA copies/mL at 36 weeks, and in the absence of obstetric contraindications, planned vaginal delivery is recommended.
  • For women with a plasma VL of 50–399 HIV RNA copies/mL at 36 weeks, PLCS should be considered, taking into account the actual VL, trajectory of the VL, length of time on treatment, adherence issues, obstetric factors and the woman's views.
  • Where the VL is ≥400 HIV RNA copies/mL at 36 weeks, PLCS is recommended.

Published cohort data from the UK and other European countries have shown MTCT rates of <0.5% in women with plasma VL <50 HIV RNA copies/mL taking HAART, irrespective of mode of delivery [[1],[4],[25],[26]]. These studies support the practice of recommending planned vaginal delivery for women on HAART with plasma VL <50 HIV RNA copies/mL.

Among HIV-positive women taking HAART in pregnancy and delivering between 2000 and 2006 in the UK and Ireland, there was no difference in MTCT rate whether they delivered by planned CS (0.7%; 17 of 2286) or planned vaginal delivery [0.7%; four of 559; adjusted odds ratio (AOR) 1.24; 95% CI 0.34–4.52]. Median VL on HAART was <50 HIV RNA copies/mL (IQR 50–184). MTCT was 0.1% (three transmissions) in 2117 women on HAART with a delivery VL <50 HIV RNA copies/mL. Two of the three infants were born by elective (pre-labour) CS (0.2%, two of 1135) and one by planned vaginal delivery (0.2%, one of 417); two of the three had evidence of in utero transmission (being HIV DNA PCR positive at birth). In this study there were no MTCT data for specific VL thresholds or strata >50 HIV RNA copies/mL plasma, but in the multivariate analysis, controlling for ART, mode of delivery, gestational age and sex, there was a 2.4-fold increased risk of transmission for every log10 increase in VL, with lack of ART and mode of delivery strongly associated with transmission [1].

Data from the ANRS French Perinatal cohort reported on 5271 women delivering between 1997 and 2004 of whom 48% were on HAART. In women on HAART with a delivery VL of <400 copies/mL there was no significant difference in MTCT rates according to mode of delivery, with three of 747 (0.4%) transmission in the ECS group compared with three of 574 (0.5%) transmissions in the vaginal delivery group (P = 0.35). The effect of mode of delivery was also analysed for women delivering with a VL >10 000 HIV RNA copies/mL and no significant protective effect of elective CS was seen (OR 1.46; 0.37–5.80). MTCT was low at 0.4% in women delivering with a VL <50 HIV RNA copies/mL but mode of delivery data for this subset were not provided [4].

In contrast, data from the ECS of 5238 women delivering between 1985 and December 2007 showed that in 960 women delivering with a VL <400 HIV RNA copies/mL, elective CS was associated with an 80% decreased risk of MTCT (AOR 0.2; 95% CI 0.05–0.65) adjusting for HAART and prematurity. There were only two transmissions among 599 women delivering with VLs <50 HIV RNA copies/mL (MTCT 0.4%) with one delivering vaginally at <34 weeks and one by ECS at 37 weeks, but further analysis was not possible [26].

A potential explanation for the differing conclusions of the effect of mode of delivery on MTCT in women with delivery plasma VLs <400 HIV RNA copies/mL in these two studies is that the true value of the plasma VL in studies that use assays with a lower limit of detection of 400 copies/mL, is not known. It is conceivable that there may exist a significant difference in the VL distribution <400 copies/mL between different cohorts, which could account for the contrasting findings. This highlights the fact that it is not possible to infer that MTCT rates from studies using a VL assay with cut-off <400 HIV RNA copies/mL can necessarily be applied to patients with plasma VLs of 50–399 HIV RNA copies/mL using current assays with lower limits of detection of 50 HIV RNA copies/mL or less.

There are no published data on the impact of mode of delivery on MTCT rates for women with plasma VLs between 50 and 399 HIV RNA copies/mL. Data from the NSHPC UK and Ireland cohort 2000–2011 (P Tookey and C French, unpublished data) and from the ECS 2000–2011 (C Thorne, unpublished data) have therefore been used to estimate the risk of MTCT and impact of mode of delivery for women on HAART with plasma VLs between 50 and 399 HIV RNA copies/mL. In the NSHPC, there were seven transmissions among 593 women with documented VL in this range: the transmission rate was 1% for those delivered by PLCS and 2.15% for those who delivered vaginally or by emergency Caesarean (P = 0.19). In the ECS cohort, of 405 women the transmission rates were 0.37% (95% CI 0.099–2.06) and 1.46% (95% CI 0.18–5.17), respectively. Although neither of these data sets show a significant difference in MTCT these findings suggest that for women with plasma VLs between 50 and 399 HIV RNA copies/mL, the risk of MTCT for women intending vaginal delivery is about 2%, and with PLCS it is 1% or less. We therefore recommend that PLCS should be considered in this group taking into account the actual VL, trajectory of the VL, length of time on treatment, adherence issues, obstetric factors and the woman's views.

Both sets of unpublished data again confirmed a lack of benefit for PLCS when the plasma VL is <50 HIV RNA copies/mL, MTCT being <0.5% irrespective of mode of delivery, supporting the recommendation of planned vaginal delivery for this group.

The UK, French and European cohorts described above all showed a protective effect of PLCS compared to vaginal delivery when applied to the entire cohort. The cohorts do not provide data to determine the viral threshold above which PLCS should definitely be recommended. However, given conflicting data regarding the effect of mode of delivery on MTCT in women with a VL <400 HIV RNA copies/mL, together with data from the UK study showing a 2.4-fold increased risk of transmission for every log10 increase in VL associated with mode of delivery, the Writing Group felt that until further data are available, PLCS should be recommended for all women with a VL >400 HIV RNA copies/mL.

7.2.2 In women for whom vaginal delivery has been recommended and labour has commenced, obstetric management should follow the same principles as for the uninfected population. Grading: 1C

Traditionally, amniotomy, fetal scalp electrodes and blood sampling, instrumental delivery and episiotomy have been avoided in HIV infection because of theoretical transmission risks. Data from the pre-HAART era have been reviewed. These show little or no risk for many of these procedures. Studies from the HAART era have not re-addressed these factors.

The French cohort (1985–1993) provides data on the risk of various obstetric factors in a predominantly untreated, non-breastfeeding population. Procedures, classified as amniocentesis, and other needling procedures, cerclage, laser therapy and amnioscopy were associated with an increased risk of transmission (RR 1.9; 95% CI 1.3–2.7). Fetal skin lesions (RR 1.2; 95% CI 0.7–1.8) and episiotomy tear (RR 1.0; 95% CI 0.7–1.3) were not associated with transmission [19]. In a retrospective study from Spain, in predominantly the pre-HAART era, HIV transmission occurred in 26.3% of infants exposed to fetal scalp monitoring (electrodes or pH sampling or both) compared with 13.6% who had neither (RR 1.94; 95% CI 1.12–3.37) [27]. However, prolonged ROMs was a significant contributor to the risk of transmission associated with this invasive monitoring. In the Swiss cohort neither fetal scalp electrodes (RR 2.0; 95% CI 0.58–6.91) nor pH blood sampling (RR 1.73; 95% CI 0.58–5.15) were confirmed as independent risk factors [28].

In the WITS cohort (1989–1994) artificial ROMs (RR 1.06; 95% CI 0.74–1.53) and exposure to blood during labour (RR 0.7; 95% CI 0.4–1.27) or delivery (RR 1.06; 95% CI 0.74–1.52) were not associated with transmission [5].

Induction has previously been avoided as there were concerns about the duration of ruptured membranes and risk of MTCT but recent evidence (see Section 7.3 Management of spontaneous rupture of membranes) would appear to be reassuring on this point.

Data from the predominantly untreated French cohort (1985–1993) showed no risk with instrumental vaginal delivery (RR 0.8; 95% CI 0.6–1.2) [19]. Data from the smaller Swiss cohort (n = 494, 1986–1996, transmission rate 16.2%) also failed to identify instrumental delivery as a risk factor (RR 1.82; 95% CI 0.81–4.08) despite <20% of the cohort taking any ART for prophylaxis [28].

In the absence of trial data for women with HIV infection who undertake a vaginal operative delivery, evidence to support a benefit of any type of operative vaginal delivery over CS for them or their infants is limited to expert judgement and extrapolation from other data sets and is subject to inherent biases. There are theoretical reasons why low cavity traction forceps may be preferred to a vacuum-assisted delivery (i.e. as it is generally accepted that they are associated with lower rates of fetal trauma than vacuum-assisted delivery).

In women with a VL <50 HIV RNA copies/mL it is unlikely that the type of instrument used will affect the MTCT and thus the one the operator feels is most appropriate should be used as in the non-HIV population (and following national guidance [29]).

The importance of the use of ART in the PMTCT of HIV is clear and undisputed. Good quality studies to determine the remaining contribution of obstetric events and interventions to MTCT in the setting of a fully suppressed HIV VL have not been performed and are unlikely to be performed in the near future. HIV DNA [30] and HIV RNA [2] in cervicovaginal lavage have been identified as independent transmission risk factors. Large cohort studies from the UK, Ireland and France have concluded there is no significant difference in MTCT in women with an undetectable VL when comparing those who have a planned vaginal delivery and those who have a PLCS. These studies provide some reassurance with regard to concerns raised about possible discordance between plasma and genital tract VL that have been reported in patients with an undetectable VL on HAART [[3],[31],[32]]. The clinical significance of this phenomenon is not clear and further research is warranted. Furthermore, there are reassuring results from the limited studies that have examined the effect on MTCT of amniocentesis and length of time of ROMs in women on HAART and in those with a VL <50 HIV RNA copies/mL. An association between MTCT and use of instrumental delivery, amniotomy and episiotomy is not supported by data from the pre-HAART era and there is a lack of data from the HAART era. Therefore, while acknowledging the potential for discordance between the plasma and genital tract VL, the Writing Group felt that there was no compelling evidence to support the continued avoidance of these procedures as well as induction of labour in women on HAART for whom a vaginal delivery had been recommended based on VL.

The data regarding fetal blood sampling and use of scalp electrodes also originate from the pre-HAART era and have yielded conflicting results. The Writing Group acknowledges a lack of data from the HAART era, but concluded that it is unlikely that use of fetal scalp electrodes or fetal blood sampling confers increased risk of transmission in a woman with an undetectable VL although this cannot be proven from the current evidence.

Electronic fetal monitoring should be performed according to national guidelines [29]. HIV infection per se is not an indication for continuous fetal monitoring, as there is no increased risk of intrapartum hypoxia or sepsis.

If the woman has no other risk factors, she can be managed by midwives either in a midwifery-led unit or at home. She will need to continue with her HAART through labour and adequate provision needs to be made for examination and testing of the newborn and dispensing of medication to the newborn in a timely fashion.

7.2.3 VBAC should be offered to women with a VL <50 HIV RNA copies/mL. Grading: 1D

In the absence of randomized trial data for women with HIV infection who undertake VBAC, evidence to support benefit of VBAC and vaginal birth over elective CS is limited to expert judgement that is subject to inherent biases.

The probability of a successful vaginal delivery remains dependent on current and past obstetric factors. In general, provided that the woman is being cared for in a consultant-led maternity unit and the labour properly monitored with rapid recourse to CS in the face of any difficulty, the outcome of trial of labour for mother and neonate is good, even if scar dehiscence occurs [33]. In the non-HIV population, 70% of VBACs manage a vaginal delivery with a uterine rupture rate of about 0.3%.

Therefore, where a vaginal birth has been recommended based on ART and VL, maternal management of the delivery, including a decision regarding VBAC, should be as for an uninfected woman.

7.2.4 Delivery by PLCS is recommended for women taking zidovudine monotherapy irrespective of plasma VL at the time of delivery (Grading: 1A) and for women with VL >400 HIV RNA copies/mL regardless of ART (see Recommendation 7.2.1) with the exception of elite controllers (see Section 5.5: Elite controllers). Grading: 1D

Zidovudine monotherapy with a planned pre-labour pre-ROMs CS is a proven option for women not requiring treatment for themselves, with a pretreatment VL <10 000 HIV RNA copies/mL plasma.

Observational studies conducted in the early 1990s, before the use of HAART, found a reduction in MTCT with PLCS. In 1999, a large international meta-analysis (n = 8533) [34] and an RCT of mode of delivery in Europe (n = 436) [9] both demonstrated a protective effect of PLCS, with reductions in MTCT of 50% and 70% respectively. In the latter study, the risk of transmission in women who were taking zidovudine monotherapy and who were delivered by PLCS was <1%. Cohort data from the UK and Ireland between 2000 and 2006 have shown that the MTCT rate in women on zidovudine monotherapy combined with PLCS was 0% (0 of 467 patients; 95% upper CI 0.8%) [1]. This was not significantly different from the 0.7% transmission rate with HAART plus PLCS (17 of 2337 patients; 95% CI 0.4–1.2%) or the 0.7% rate with HAART plus planned vaginal delivery (four of 565 patients; 95% CI 0.2–1.8%). These findings support the option of zidovudine monotherapy in women not requiring treatment for themselves with low VLs who either have an obstetric indication for, or are prepared to be delivered by, PLCS.

There is no evidence that women on HAART with a low VL have increased surgical morbidity compared with the HIV-negative population

A Cochrane review evaluating the risk of postpartum morbidity according to mode of delivery included five studies: the European randomized mode of delivery trial and five observational studies from North America and Europe [35]. This review found a higher incidence of minor postpartum morbidity, including fever and anaemia requiring transfusion, among HIV-positive women delivered by CS compared with those who delivered vaginally. Low CD4 cell count and co-morbidities such as diabetes were independent risk factors for postpartum morbidity. This review included women who were not on HAART.

More recent cohort data from Europe [[25],[36]] and from case-controlled studies in the USA [37] and UK [38] involving women on HAART with undetectable VLs have demonstrated very low rates of maternal morbidity, irrespective of mode of delivery.

7.2.5 Where the indication for PLCS is the prevention of MTCT, PLCS should be undertaken at between 38 and 39 weeks' gestation. Grading: 1C

Where PLCS is undertaken only for obstetric indications and plasma VL is <50 copies/mL, the usual obstetric considerations apply and timing will usually be at between 39 and 40 weeks.

The timing of PLCS is a balance between the risks of transient tachypnoea of the newborn (TTN) and the likelihood of labour supervening before the scheduled CS [39]. Where the indication for PLCS is PMTCT, the earlier timing reflects the importance of avoiding the onset of labour. In these cases, the risk of MTCT associated with labour and ROMs is considered to outweigh the risk of TTN. Where PLCS is undertaken only for obstetric indications, the optimal timing of PLCS is between 39 and 40 weeks [33]. The risk of TTN at this gestation is approximately 1 in 300 and this risk doubles for every week earlier that delivery occurs. The administration of steroids to the mother to reduce the risk of TTN should be considered for PLCS prior to 38 completed weeks.

7.3 Management of spontaneous rupture of membranes

7.3.1 In all cases of term pre-labour spontaneous ROM, delivery should be expedited. Grading: 1C

7.3.2 If maternal HIV VL is <50 HIV RNA copies/mL immediate induction of labour is recommended, with a low threshold for treatment of intrapartum pyrexia. Grading: 1C

7.3.3 For women with a last measured plasma VL 50–999 HIV RNA copies/mL, immediate CS should be considered, taking into account the actual VL, the trajectory of the VL, length of time on treatment, adherence issues, obstetric factors and the woman's views. Grading: 1C

7.3.4 If maternal HIV VL is ≥1000 RNA copies/mL plasma, immediate CS is recommended. Grading: 1C

In the pre-HAART era, several studies [[5],[6],[40]] suggested that prolonged duration of ruptured membranes, usually analysed as >4 h, in women who were either untreated or if treated were largely receiving zidovudine monotherapy, resulted in a significantly increased risk of MTCT. A widely quoted meta-analysis (not reporting VL data) subsequently showed a 2% increase in relative risk of transmission per hour of membrane rupture (AOR 1.02). Transmission increased from 12% with <1 h membrane rupture to 19% with >12 h of membrane rupture [41].

There are few published studies from the HAART era. A study from Spain of 500 HIV-positive women examined the effect of various obstetric risk factors on MTCT rates in women on no treatment, monotherapy or dual therapy, and finally in those on HAART. ROMs >6 h compared to <6 h was only significantly associated with MTCT in the group of women on no treatment (26.6% vs. 11.9%; P ≤ 0.01). Corresponding transmission rates for the mono–dual therapy group were 14.3% vs. 7.1% (P = NS) and in the women on HAART (0.8% vs. 0.0%; P = NS) [42].

The NSHPC study of HIV-positive women in the UK and Ireland reported on 1050 women where length of time of ROM was recorded from 2007. In 618 women delivering with a VL <50 HIV RNA copies/mL when comparing those with ROM ≤4 h to >4 h the MTCT rate was 0.3% (one of 326) and 0.0% (none of 292), respectively (P = 0.34). Restricting the analysis to the 386 women with a VL <50 copies/mL who delivered vaginally did not alter this conclusion [43]. Therefore, for women on HAART who rupture their membranes at term with a VL <50 HIV RNA copies/mL and who do not have an obstetric contraindication to vaginal delivery, a CS is not recommended.

As both acute and chronic chorioamnionitis have been associated with perinatal transmission [[6],[44][[45][#[46]]Ent]241], albeit from studies largely performed in the pre-HAART era, it is recommended that labour should be expedited for all women with ROM at term. Hence, women with ROM at term with a VL <50 HIV RNA copies/mL should have immediate induction with a low threshold for the treatment of intrapartum pyrexia. The NICE induction of labour guidelines [47] and NICE intrapartum guidelines [29] should be followed with regard to use of antibiotics and mode of induction.

NSHPC data for the effect of ROM greater or less than 4 h for women with a VL > 50 HIV RNA copies/mL are more difficult to interpret as the numbers are currently small. In women with VL 50–999 HIV RNA copies/mL there were two transmissions with ROM > 4 h (two of 51) and none in the women with ROM ≤ 4 h (none of 43). The two transmitters both had emergency CSs but the timing of this is not known. Although not statistically significant (P = 0.19), these limited unpublished data suggest a possible trend towards greater transmission risk with ROMs >4 h for those with VL ≥ 50 HIV RNA copies/mL, and until further data are available, it is the recommendation of the Writing Group that CS should be considered for women with a VL of 50–999 HIV RNA copies/mL at term. Again, if CS is not undertaken, delivery should be expedited, as above.

Data from the NSHPC for women with a VL > 1000 HIV RNA copies/mL are sparse at present, with one of 14 (7.1%) transmitting with ROM ≤ 4 h compared to three of 15 (20%) with ROM > 4 h. A single-centre study from Miami of 707 women on ART showed ROM > 4 h to be associated with an increased risk of MTCT if the VL was >1000 HIV RNA copies/mL. There was no association at <1000 HIV RNA copies/mL but it is not possible to determine the number of women with a VL > 50 and <1000 HIV RNA copies/mL in this group. Until further data are available, an urgent (category 2) CS is recommended where the VL is >1000 HIV RNA copies/mL regardless of treatment [48].

In women who have a detectable VL it may be possible to optimize their HAART regimen to reduce the risk of MTCT (See Recommendation 4.2.6).

7.3.5 The management of PPROMs at ≥34 weeks is the same as term ROM (see Section 7.3 Management of spontaneous rupture of membranes) except women who are 34–37 weeks' gestation will require group B streptococcus prophylaxis in line with national guidelines. Grading: 1C

7.3.6 When PPROM occurs at <34 weeks: Grading: 1C

  • Intramuscular steroids should be administered in accordance with national guidelines.
  • Virological control should be optimized.
  • There should be multidisciplinary discussion about the timing of delivery.

There are no data to inform the optimum management of preterm labour or early preterm pre-labour ROMs. Decisions regarding the optimum management of early preterm ROM require the assessment of a number of factors, including the exact gestation, facilities available, maternal VL and presence of other co-morbidities such as infection and pre-eclampsia. Corticosteroids to improve fetal lung maturation should be given as per the Royal College of Obstetricians and Gynaecologists guidelines [49] and (if delivery is to be delayed) oral erythromycin [50]. Decisions regarding timing of delivery should be made in consultation with the full MDT, including the neonatal unit.

There is no evidence that steroids for fetal lung maturation (with the associated 24-h delay in induction) are of overall benefit at 34–37 weeks' gestation in women with ROMs, thus delay for the optimization of fetal lung maturity is not recommended. For this reason, and to minimize the risk of developing chorioamnionitis, induction is recommended from 34 weeks' gestation in women with ROMs who are not in labour.

If the maternal VL is not fully suppressed, consideration should be given to the options available to optimize therapy. An additional concern is that the early preterm infant may be unable to tolerate oral therapy and therefore loading the infant through the transplacental route with maternal therapy is recommended (see Section 5: Use of antiretroviral therapy in pregnancy). There is most experience with maternal oral nevirapine 200 mg stat >2 h before delivery, but double-dose tenofovir and standard-dose raltegravir can also be considered.

7.4 Use of intrapartum intravenous infusion of zidovudine

7.4.1 Intrapartum intravenous zidovudine infusion is recommended in the following circumstances:

  • For women with a VL > 10 000 HIV RNA copies/mL plasma who present in labour, or with ROMs or who are admitted for planned CS. Grading: 1C
  • For untreated women presenting in labour or with ROMs in whom the current VL is not known. Grading: 1C
  • In women on zidovudine monotherapy undergoing a PLCS intravenous zidovudine can be considered. Continued oral dosing is a reasonable alternative. Grading: 1B

There are no data to support the use of intrapartum intravenous zidovudine infusion in women on HAART with a VL < 10 000 HIV RNA copies/mL plasma.

The use of intravenous zidovudine is suggested for women taking zidovudine monotherapy as per Recommendation 5.3.4. The use of intravenous zidovudine for women on HAART with a VL between 50 and 10 000 HIV RNA copies/mL can be considered regardless of mode of delivery. However, continued oral dosing of their current regimen is a reasonable alternative.

The effectiveness of zidovudine monotherapy in preventing MTCT was first demonstrated in the ACTG 076 RCT of non-breastfeeding women in which zidovudine was initiated orally before the third trimester, given intravenously during labour and delivery, and orally to the neonate for the first 6 weeks of life, reducing MTCT by 67% [8]. Intravenous zidovudine has therefore been included in the management of all women treated with zidovudine monotherapy. However, the data on the contribution of intravenous zidovudine are poor. In a prospective study of all women prescribed zidovudine monotherapy during pregnancy before the publication of the ACTG 076 findings (1988–1994) in which the 8.8% transmission rate among women with CD4 cell counts >200 cells/μL is similar to that of the zidovudine monotherapy arm of ACTG 076 (8.3%), intrapartum intravenous zidovudine was not associated with lower rates of transmission [51]. One rationale for intrapartum intravenous zidovudine in ACTG 076 was that labour would be associated with poor absorption of oral therapy. While not strictly comparable, the well-recognized rapid absorption of single-dose nevirapine during labour suggests that the impact of labour on absorption may be overestimated. Pharmacokinetic data from an RCT of oral zidovudine monotherapy vs. placebo indicate that adequate (therapeutic) zidovudine concentrations are achieved in cord blood with oral dosing. Although the concentrations are lower than have been reported with intravenous infusion, transmission was not associated with zidovudine cord blood concentration [52].

Intravenous zidovudine has historically been considered for women whose plasma VL has not been completely suppressed at the time of delivery. There is no evidence that the intravenous administration of zidovudine alters the rate of placental transfer but higher maternal plasma levels will be reflected in the cord blood concentrations.

Intravenous zidovudine (as part of an intervention package; see Section 5: Use of antiretroviral therapy in pregnancy) has also been recommended for women who present in labour, having not received ART. However, data from the New York State HIV diagnostic service (1995–1997) suggest that intrapartum intravenous zidovudine alone does not significantly reduce transmission (10%; 95% CI 3.3–21.8%), as, provided neonatal prophylaxis is commenced within 48 h of delivery (this being the only intervention accessed), the latter has similar efficacy (9.3%; 95% CI 4.1–17.5%) [10].

From the French data there is no evidence that intrapartum intravenous zidovudine further reduces the risk of MTCT in women on HAART unless maternal HIV VL is >10 000 copies/mL [4]. However, individual circumstances vary, and intravenous intrapartum zidovudine may be considered as one of a number of maternal intrapartum ART options for women with VLs > 50 HIV RNA copies/mL who present in labour, or with ROMs or who are admitted for planned CS provided this does not delay other interventions.

The evidence for the efficacy of intravenous zidovudine in the HAART era is generally poor. However, data from the French cohort support this practice for women on HAART with a VL >10 000 HIV RNA copies/mL. One could extrapolate that it may be of potential benefit in women presenting untreated in labour with an unknown current VL although this is not supported by the New York State data. Therefore in this setting, the Writing Group recommends the immediate administration of oral agents (see Section 5: Use of antiretroviral therapy in pregnancy) with intravenous zidovudine as an option.

In women on HAART with a VL between 50 and <10 000 HIV RNA copies/mL, intravenous zidovudine can be considered. Continued oral dosing is a reasonable alternative.

Intravenous zidovudine is not recommended for women taking HAART who have an undetectable VL at the time of labour or CS. Oral HAART should be taken at the normal dosing interval.