BHIVA 2003-4 audit on HIV and pregnancy

2 June 2004. Related: Conference reports, BHIVA 10th Cardiff 2004.

Polly Clayden, HIV i-Base

Data from the BHIVA pregnancy audit – a case note review of pregnancies among women with HIV in Britain between October 2002 and September 2003 – were presented at this meeting. This review was performed in order: “To enable BHIVA guidelines to be reviewed in the light of current practice and the most recent evidence.” It also gives an insight into the extent to which the current guidelines are being followed.

The audit evaluated questionnaires from 99 centres nationwide: 19 in London, 79 elsewhere and one unstated. Eighty centres submitted data for 504 pregnancies. Four pregnancies were excluded from the analysis due to: one maternal and foetal death at 24 weeks gestation caused by multi organ failure attributed to TB drugs and/or nevirapine; two pregnancies were terminated and one was not delivered during the audit period.

Management and communication

The majority of centres (87) managed pregnancy and delivery working with a multi-disciplinary team and 81 respondents were satisfied with communication among healthcare professionals. A large number (79) of respondents said that post-natal ward midwives would be informed of a woman’s HIV status, eight said problems had arisen through relevant staff not being told of a woman’s status and 11 through staff using this information inappropriately.

Diagnosis and patient demographics

The majority of centres (82) used an “opt out” system of antenatal HIV testing, 11 reported an “opt in” system and six did not know/answer. Of the group of women evaluated 208 (42%) knew their status before pregnancy; 249 (50%) were diagnosed in the first or second trimester; 233 (47%) by routine antenatal screening; 37 were diagnosed in the third trimester but >7 days pre-delivery; 3 (0.6%) within seven days pre-delivery and 2 (0.4%) post-delivery.

The majority of patients, 390 (78%) were black African; 60 (12%) were white; 21 (4%) were black Caribbean (20 patients were described as “other” and there were no data for nine).

Most of the women were either in the 20-30 year old age group – 239 (48%) or the 30-40, 215 (43%) age group. Only two (0.4%) were under 15, 23 (5%) were age 15-20 and 8 (2%) were over 40 years old. (There were data missing for 13 [3%] women).

Maternal health and use of antiretrovirals

Almost half (49%) of the women – both off and on treatment – had a CD4 count of >350 cells/mm³ near the beginning of their pregnancy, 2% had 0-50 cells/mm³; 18% 51-200 cells/mm³; 27% 2001-350 cells/mm³ and 4% were not known.

At the start of pregnancy, 104 (21%) women were already receiving antiretrovirals. Treatment varied widely and there were a few worrying reports: two women were receiving dual therapy (although had undetectable viral loads); 11 were on efavirenz (two with detectable viral loads) at the start of pregnancy and five at the end of pregnancy; six were receiving ddI and d4T together (two with detectable viral loads) and four were still using ddI/d4T at the end of their pregnancy.

At the end of their pregnancies, 484 women were using antiretroviral treatment or prophylaxis. The data were unclear for eight patients.

Half the women received ZDV/3TC/NVP; 14% ZDV monotherapy; 10% ZDV/3TC/NFV; 2% ABC/ZDV/3TC; 2% ZDV/3TC/LPVr. Two percent of women received no antiretrovirals, including one with HIV2 and four very late presenters. There was no analysis to evaluate how many women with higher CD4 counts who did not require treatment for their own HIV received triple therapy.

These findings are consistent with respondent’s replies when questioned about what they would do in various scenarios. The investigators noted that one respondent, when asked how they would treat a woman with a CD4 count of 562 cells/mm3 and a viral load of 5,700, proposed dual therapy of AZT and 3TC. (See Figure 1).

Respondents were also asked what they would do for a subsequent pregnancy for a woman who did not require treatment for her own health. Thirty-seven percent of respondents replied that they would base their strategy on a resistance test and/or adherence and viral load on previous ART and 20% that they would offer the standard therapy or the same as in the previous pregnancy.

Figure 1: Stated preference of ART by clinicians in three hypothetical scenarios

Click graphic to enlarge in new window.

Mode of delivery

In answer to a question about a hypothetical scenario involving a woman on HAART with stable, undetectable viral load, 55% of respondents reported that they would recommend an elective Caesarean section; only 9% and 7% would favour a planned vaginal delivery in women with previous uncomplicated births and first pregnancies respectively. Sixteen per cent were neutral and the remainder reported no policy or did not answer.

Of the reported pregnancies 422 (85%) were planned for Caesarean section and in 43 (9%) it was not considered to be indicated as 38 had a pre-delivery undetectable (<50 copies) viral load, three with viral load less than 1000 and two not known. The majority of women (35) in this group were on HAART, two were receiving ZDV monotherapy and one ZDV/3TC dual therapy. Nine women (2%) declined a Caesarean section, two had no plan and data were missing for 22.

Actual mode of delivery reports 335 (67%) by elective caesarean; 70 (14%) after onset of labour; 54 (11%) vaginally and 41 (8%) “Not known”. Fifty-eight (14%) of planned Caesarean sections resulted in Caesarean section after onset of labour and eight (2%) delivered vaginally.

Deliveries in women planned for elective Caesarean

The investigators added there is: “Possible need for guidance on when to perform planned CS – many women labour early, so best done at 38 weeks, not later.”

Pregnancy outcomes

The majority of deliveries (48%) occurred at 38/40 weeks, which reflects the use of elective Caesarean, 11% were at or before 36-40 weeks and 9% at 37-40 weeks. Initial results suggest 10 stillbirths: nine at or before 36-40 weeks and one at 37-40. The investigators are following this up.

Foetal and neonatal abnormalities

The investigators reported 15 abnormalities among foetuses and neonates, these were described as follows:

• 2 babies known to have HIV
• 2 (one a twin) died of neonatal TB
• 1 spina bifida, possible sacral myelomeningocele
• 1 trisomy 21 & AV canal defect
• 1 congenital jejunal atresia
• 1 cleft palate
• 1 diaphragmatic hernia
• 1 clicky hip, absent red reflex, later found normal
• 1 intra-uterine growth retardation
• 1 “small with infection”
• “flat” baby incubated in neonatal intensive care
• 2 unclear

Breast feeding

In answer to a question about support currently offered for infant feeding the investigators reported: “Centres varied greatly in the support offered for bottle-feeding.” And in answer to a hypothetical question about a mother’s reluctance to bottle feed: seven replied that this was patient choice; 14 respondents cited child protection; 21 referred to the use of antiretrovirals maintaining maternal viral load below detection at 50 copies/mL; four mentioned antiretroviral prophylaxis for the baby and three suggested pasteurising expressed milk.

Conclusions

The investigators described these findings as “broadly positive” but cited areas in which they felt stronger guidance to be needed. In particular: appropriate use and stopping of nevirapine; avoidance of ddI and d4T together; appropriate use of planned vaginal delivery; timing of elective caesarean section; management of women who wish to breast rather than formula feed and management of subsequent pregnancies.

Comment

Nevirapine has been widely prescribed in combination therapy during pregnancy in the UK and elsewhere. In the UK triple therapy has generally been prescribed only once the CD4 count has fallen below 350 and usually when closer to 200 cells/mm3. Since the recent caution from Boehringer this remains an option for pregnant women with CD4 counts below 250 cells/mm3 but the more difficult question is how best to manage those pregnant women with high CD4 counts for whom a short course of combination therapy is considered the best option.

The unquantified risk of fulminant hepatitis with nevirapine in pregnancy, as well as the difficulties of preserving future options when discontinuing combinations with widely differing half-lives – a troubling finding was that when asked about stopping nevirapine 15% of respondents said that they would stop all drugs together – means that other options, need to be found. While protease inhibitors appear to be the best alternative there is relative paucity of data on their pharmacokinetics during pregnancy, and the importance of transplacental transfer of antiretrovirals should also be considered especially if delivery is likely to occur before viral replication has been controlled.

Conversely these data probably highlight the relative safety of using nevirapine in pregnancy and it is important to note that this audit was performed prior to the updated safety warning.

Perhaps the most important initial finding of the audit is the observation of 10 stillbirths in 500, a rate almost 4 times the national rate of 5.6/1000 (http://www.statistics.gov.uk). Further analysis is urgently required to determine whether this is due to recognised social and medical factors, confusion over reporting and methodology (“miscarriages” were censored in this analysis) or whether current intervention strategies are in anyway implicated. The investigators are following this up.

Finally this audit also highlights the need for strong support packages for HIV positive mothers to encourage bottle feeding. This includes clear information about the risk of breastfeeding, counselling and practical interventions such as formula and sterilising equipment for those mothers with limited resources, until there are robust data to support the early studies suggesting that breastfeeding by a mother with an undetectable viral load using antiretrovirals carries zero risk of transmission.

Revision of the BHIVA Pregnancy Guidelines is currently underway.

Reference:

BHIVA clinical audit committee. Oral presentation, BHIVA Plenary Session 3, 10th BHIVA conference 2004.