Prevalence of HIV antibody and pregnancy in Tayside, 1984-9: background to screening.

Objective: To determine age specific prevalence of HIV antibody, incidence of pregnancy, and likelihood of detection and correct assignment to risk category by antenatal screening of women known to be positive for HIV antibody, from 1984 to 1989.
Design: Retrospective analysis of reproductive history and risk behaviour of women positive for HIV antibody and prediction of detection by screening on the basis of blood group samples, Guthrie tests, and rubella tests.
Setting: City of Dundee, where the prevalence of HIV is high, since the appearance of HIV in 1984, predominantly among heterosexual intravenous drug users.
Patients: All (61) women known to be positive for HIV antibody who had had clinically indicated tests, for whom case notes were available for 60.
Main Outcome Measures: Risk assessment according to case notes and reported to the laboratory, incidence of infection, geographical location, age, date of positive test result, and reproductive history.
Results: With 61 infected women the overall minimum prevalence among women within the city of Dundee was 0.67/1000 and 2.9/1000 among women in their third decade. Of the 60 women whose reproductive history was available, 35 had 57 pregnancies, 36 of which occurred after seroconversion was known to have taken place, representing 8.7% of the total number of affected pregnancies reported for the United Kingdom. If antenatal screening for HIV had been performed between 1984 and 1989 it could not have detected positivity for HIV antibody in 25 (42%) women who had no pregnancies during this time. Among the remaining 35 women, screening samples taken for blood grouping could have identified a maximum of 34 (57%), samples taken to check rubella susceptibility a maximum of 22 (37%), and blood spots on Guthrie cards a maximum of 19 (32%). Retesting would have occurred in 14 women 33 times with samples taken for blood grouping, but three and four women would have been tested twice using samples taken for rubella testing and Guthrie cards respectively. Anonymous screening would have been unable to determine risk category as a history of intravenous drug use was known in 47 (79%) women before testing but this was increased by a further 5 (8%) who admitted to it after the test result was known.
Conclusion: Interpreting the results of antenatal screening programmes will be complex and will underestimate overall prevalence of HIV antibody among women; this will be exaggerated by strategies based on anonymous testing with Guthrie cards or on samples taken for rubella testing, which do not include women who have had an earlier loss of pregnancy. Only open testing with consent will permit satisfactory attribution to