Blood collection on filter paper: a practical approach to sample collection for studies of perinatal HIV transmission

Abstract

The use of dried blood spots lends itself to widespread application in large field studies, especially in remote areas. We present experience gained during a perinatal HIV transmission study in southern Africa in which dried blood spot samples were used for polymerase chain reaction (PCR) tests. In this study, 15,810 filter paper cards with dried blood spots were collected. Infants were seen at age 6 and 12 weeks, and PCR was routinely done in duplicate on each sample. Of 186 negative controls (infants born to HIV-negative women), two (1.1%) had a single strongly reactive PCR result; the repeated duplicates were both negative. In contrast, all 24 known positive samples were strongly positive in both tests. Results were available from 1,976 duplicate tests on 1,235 infants born to HIV-infected women. Based on the PCR result on a later sample, the positive predictive value was 97.6% if both replicates were strongly positive (absorbance: 0.8 OD450 U), 100% when one of the replicates was strongly positive, and 27% when one or both replicates were weakly positive (but none strongly positive). When both replicates were negative, the negative predictive value was > or = 96.2%. Thus, when a single HIV PCR test has a strongly positive result, the infant is very likely to be infected. A positive PCR result after age 1 month was 98.9% accurate in predicting antibody positivity after 15 months. Suggestions for sample collection, storage, and PCR testing are provided.

PIP: 15,810 filter paper cards with dried blood spots were collected during a perinatal HIV transmission study in southern Africa to be subjected to polymerase chain reaction (PCR) testing. Infants were seen at ages 6 and 12 weeks, with PCR routinely done in duplicate on each sample. Of 186 infants born to HIV-negative mothers, two had a single strongly reactive PCR result, while the repeated duplicates were both negative. All 24 known positive samples were strongly positive in both tests. Results were available from 1976 duplicate tests on 1235 infants born to HIV-infected women. Based upon the PCR result of a later sample, the positive predictive value was 97.6% if both replicates were strongly positive, 100% when one of the replicates was strongly positive, and 27% when one or both replicates were weakly positive. When both replicates were negative, the negative predictive value was greater than or equal to 96.2%. Therefore when a single HIV PCR test has a strongly positive result, the infant is very likely infected. A positive PCR result after age 1 month was 98.9% accurate in predicting antibody positivity after 15 months.