Repeat antenatal HIV testing in the third trimester: a study of feasibility and maternal uptake rates
The study aimed to assess the feasibility and acceptability of third-trimester antenatal HIV testing within our service after two cases of HIV seroconversion in pregnancy were noted in 2008. North American Guidelines recommend universal third-trimester HIV testing in areas with an HIV prevalence of more than 1 per 1000. The HIV prevalence rate in our area is 3.01 per 1000.
Pregnant women prior to 28 weeks of gestation were recruited at booking between 1 September 2008 and 31 August 2009 and offered an additional third-trimester HIV test. Consent was obtained and testing was performed by hospital and community midwives. Information was entered into a modified existing electronic maternity database. A qualitative e-mail survey of midwives investigated barriers to participation in the study.
A total of 4134 women delivered; three (< 0.1%) declined first-trimester testing. Twenty-two women (0.5%) tested HIV positive, of whom six were newly diagnosed. Overall, 2934 of 4134 women (71%) were offered and accepted a third-trimester HIV test and had results available. Data were unavailable for 195 women (4.7%). A total of 663 of 4131 women (16%) were not offered a third-trimester test. Of 3273 women documented as having been offered a test, 3177 (97.1%) accepted. There were no positive third-trimester tests. Forty of 50 (80%) midwives surveyed responded with questionnaire feedback and cited lack of national policy and extra workload as barriers to performing third-trimester testing.
Third-trimester testing was feasible and consent rates were high in those offered repeat testing. Third-trimester testing has the potential to prevent paediatric HIV infection and universal testing should be considered in high-prevalence areas.
Full adherence to national prevention of mother-to-child transmission of HIV guidelines has reduced the transmission rate from around 25% without intervention to less than 1% annually . An audit exploring factors behind perinatal HIV transmission in 87 children between 2002 and 2005 concluded that around 70% were born to undiagnosed women and at least 20% of mothers had tested negative in pregnancy . Thus, identifying seroconversion in pregnancy could help reduce mother-to-child HIV transmission further.
US guidelines  recommend universal repeat third-trimester HIV screening for women receiving care in facilities where at least one HIV infection is diagnosed per 1000 pregnant women screened. In the last decade, approximately 1 in 250 women who booked with North West London Hospitals (NWLH) NHS Trust were HIV positive. Twenty-five HIV-positive women on average deliver per year in the Trust. Universal first-trimester opt-out HIV testing rates consistently exceed 99%.
The HIV prevalence of the adult population is 3.04 per 1000 averaged across the two boroughs of Brent and Harrow served by NWLH NHS Trust . Nationally, in 2010, about three-quarters of women diagnosed before delivery were already aware of their infection [5, 6].
Studies have noted that universal screening of pregnant women has been more successful than risk-based testing at diagnosing HIV infection and preventing perinatal transmission [7-9]. Current British HIV Association (BHIVA) guidelines 2012  support the position of the UK national screening committee  which decided against recommending universal repeat third-trimester HIV testing, instead recommending that ‘repeat tests could be offered to any woman who was thought to be at continuing risk of infection and to any woman who requested a second or subsequent test’. The Royal College of Obstetricians and Gynaecologists (RCOG) guidelines state that ‘if a woman tests HIV negative at booking but is judged by her clinician as being at continued high risk of acquiring HIV, offering a repeat HIV test should be considered’ .
In 2006, the perinatal HIV group at NWLH NHS trust observed two cases of HIV seroconversion in pregnancy. There is additionally one HIV-infected child in our service whose mother tested HIV negative in the first trimester of pregnancy and who reportedly mixed breast and bottle fed for 2 months after delivery.
Following discussion among the Trust's multidisciplinary perinatal HIV group (which includes paediatricians, pharmacists, midwives, specialist nurses, microbiologists and physicians in genitourinary medicine), it was decided to set up a pilot study to assess if third-trimester antenatal HIV testing was feasible in our population and to determine uptake rates. It was not anticipated that there would be a new HIV seroconversion and thus a new positive HIV test within the study period, and therefore cost−benefit analysis was not part of the study aims.
Third-trimester HIV screening was launched for women booking at NWLH NHS Trust from 1 September 2008 until 31 August 2009. Data were collected for the subsequent deliveries between 1 May 2009 and 31 April 2010. Third-trimester screening could be offered at any point up to delivery.
Women were recruited if their antenatal booking appointment was between 1 September 2008 and 31 August 2009 and if the gestation at booking was no more than 28 weeks. If a woman was more than 28 weeks of gestation at booking and/or had booked prior to 1 September she was not included in the study, and so was offered routine HIV screening along with her other booking bloods.
At the antenatal booking at 12 weeks of gestation, in line with local and national protocol , the booking midwife made a routine recommendation for a first-trimester HIV test. If this was accepted the midwife offered and, if accepted, obtained consent for a third-trimester HIV test also. Routine antenatal care continued either in the community or in the hospital as normal. After 28 weeks of gestation or when the pregnant woman attended for a routine third-trimester full blood count, a third-trimester HIV test was carried out.
The Central Maternity Information System (CMIS) is a computerized database that is used by NWLH NHS Trust to store all maternity clinical data. For the purposes of this study, third-trimester testing was added to the computer system with four options (accepted and done; declined; not offered; accepted and not done).
At the end of the study, the CMIS database was used to retrospectively collect figures on screening uptake.
The specialist midwives held two all-day training and education sessions for all midwives to attend and on-going ad hoc sessions for newcomers and doctors. The specialist midwives paid individual visits to all clinical areas and notified all general practitioners in referring areas of Brent and Harrow by letter. On-going reminders were issued in the form of discussion at clinical meetings, posters in the phlebotomy area and flyers.
Qualitative feedback on the feasibility and acceptability of the approach was obtained through anonymous individual questionnaires after the study was completed and group feedback during the study to the specialist midwives.
HIV testing was performed in the laboratory, using venous blood, on an Architect analyser (Abbott Diagnostics, Maidenhead, Berkshire, UK) with Abbott's fourth-generation HIV antigen/antibody combination assay.
This feasibility study is classified as health service evaluation research and as such ethical approval was not required.
Results for the deliveries during the data collection period from 1 May 2009 to 31 April 2010 are shown in Table 1.
Table 1. Results of third trimester screening 2008–2009
|Total number of deliveries||4134|
|Declined all HIV testing (n)||3|
|Offered and accepted third-trimester screening (result available) [n (%)]||2934 (70.9)|
|Offered and accepted third-trimester screening (result not available) [n (%)]||243 (5.9)|
|Offered and declined third-trimester screening [n (%)]||99 (2.4)|
|Not offered third-trimester screening [n (%)]||663 (16.0)|
|Data unavailable [n (%)]||195 (4.7)|
There were 22 HIV-positive women booked in that period (1 September 2008 to 31 August 2009), of whom six were newly diagnosed at the antenatal booking. There were no additional positive HIV tests in the third-trimester testing group.
All third-trimester screening was carried out by hospital or community midwives only.
There were up to 155 midwives involved in third-trimester testing. Fifty e-mail questionnaires requesting qualitative feedback on the third-trimester testing were sent out to the midwives involved, by one of the lead specialist midwives. Forty questionnaires were returned. Reported barriers to repeat testing included staff forgetting to perform a test as a consequence of the time elapsing between consent being obtained and eligibility being confirmed, perceived extra workload and the fact that the test was not a national recommendation. Factors associated with increased testing included practical triggers to remind staff to perform the test, information posters in the clinic area informing the mothers of the availability of the test, who then reminded their midwife to do it, the presence of an HIV testing champion and feedback on performance.
The cost of the HIV test was £8.28 and to achieve 100% testing of the 4316 deliveries at that time would have cost £35 736. Education sessions cost £935, extra administration cost £500 and clinical costs including counselling and auditing were £4491. Total costs for 1 year were therefore £41 662.
Third-trimester repeat HIV testing is feasible in our population of pregnant women and consent rates were high in those offered repeat testing. The Trust staff were broadly supportive of repeat testing. The main obstacles to repeat HIV testing in our study were mainly a consequence of staff forgetting and the logistics of documentation. These obstacles are easily surmountable with the use of testing champions, computer screen reminders and better documentation in the ‘hand held’ notes. Only 2.4% of women declined repeat HIV screening. Anonymized feedback sent to the specialist midwives stated that it was sometimes quicker to report ‘patient declined’ on the screen than to check whether the mother was still happy for repeat HIV testing to be performed, suggesting that the true percentage who would have declined is likely to have been even lower. Far more patients were not offered repeat testing by staff, suggesting that even higher uptake rates of repeat testing would be achievable if this were policy. The 243 (6%) of patients who accepted a test but for whom results were not available probably did not have the test performed by midwives.
Third-trimester testing did not affect the uptake of first-trimester screening. Only three women were untested in the first trimester as a consequence of declining all testing.
From the National Study of HIV in Pregnancy and Childhood (NSHPC) data  it may be inferred that four infected children are born per year nationally following maternal seroconversion in pregnancy or breast feeding. It is not known what proportion seroconvert during the later stages of pregnancy rather than during breast feeding and hence would have HIV infection detected by routine third-trimester screening. It is also not known how many children born to mothers who seroconverted in pregnancy have undiagnosed HIV infection in the population.
While routine third-trimester testing is not recommended nationally, both BHIVA pregnancy guidelines  and RCOG guidelines  advocate considering third-trimester testing for those perceived to be at increased risk. This strategy was initially adopted with first-trimester testing, before the introduction of universal testing, and is known to have had inconsistencies in interpretation of which groups or individuals are at higher risk. Unprotected sex is commonly practised by pregnant women, and where the partners' HIV status is not known, would constitute increased risk of transmission in pregnancy.
The costs of such a screening programme would, once established, be considerably reduced by eliminating the need for staff education and setting up an on-going auditing system (via the CMIS for those who use this system). A further significant cost reduction would be achieved through a cheaper HIV test.
In an 18-month period, out of seven infants who were diagnosed HIV-positive at a London teaching hospital, four (57%) were born to mothers who had unrecognized HIV seroconversion during pregnancy . Transmission of infection to such infants results in a lifetime of required care for HIV infection. There are currently no published data on lifetime cost of paediatric HIV infection in the UK, but North American data estimated $228 155 for 25 years of survival , more expensive than older Australian data . With advances in treatment, HIV-positive children are likely to live well beyond this age and the lifetime cost of their treatment will therefore be higher.
A study using mathematical modelling of mother-to-child transmission of HIV in South Africa  concluded that maternal seroconversion in late pregnancy and breast feeding was a significant contributor to the burden of childhood HIV infection, but this is an area of higher incidence than the UK. Enhanced perinatal surveillance in the USA showed that seroconversion during pregnancy increased between 2005 and 2010, and it was concluded that repeat third-trimester screening, which is already nationally recommended, should be enhanced . The US figure used as the threshold comes from a cost-effectiveness modelling paper which found that a second HIV test in the third trimester is as cost-effective as other common health interventions when HIV incidence among women of childbearing age (WCA) is > 17 HIV cases per 100 000 person-years .
The comparable figure for Brent and Harrow for 2008 based on 2001 census data was 22 cases in an estimated population of 60 000 WCA for Brent and Harrow, or 37/100 000 person-years 
Our study illustrates that the test is feasible and the uptake rate would suggest that it is also highly acceptable to women and does not result in reduced uptake of the standard first-trimester screening HIV test. The difficulties we identified with midwives offering and carrying out the test would be much less if the testing was part of routine care, where in our Trust 99% uptake of HIV screening is routine.
We understand that cost-effectiveness and projected low likelihood of detection are the main reasons for not currently offering routine repeat third-trimester HIV testing. However, we believe that assessing feasibility and uptake rates is an important initial step before planning larger studies which would also contribute economic data to a fuller discussion of the advantages and disadvantages of repeat testing in pregnancy in the UK, to add to data from other settings with varying HIV incidence and resources . Our study would be improved by inclusion of more detailed information such as the median (and range) week of gestation when the repeat test was undertaken to assess the available time to undertake interventions to reduce mother-to-child transmission.
Further studies across multiple sites with varying HIV prevalence should be performed to assess the cost-effectiveness of such additional testing in the third trimester, particularly in populations where universal first-trimester testing is consistently achieved. Our data suggest that third-trimester HIV testing in our population was feasible, was acceptable to staff and had high consent rates in pregnant women, and other centres in areas of similar HIV prevalence may wish to consider instituting it to help reduce transmission of infection to babies.
The study was funded by Brent PCT.