An improved algorithm for determining HIV type 1 subtypes in a primary laboratory in Uganda.

A pilot study was undertaken with the objective of developing a simple, economical, and efficient algorithm through which to subtype HIV-1 in a large epidemiological cohort study in Uganda. A peptide enzyme immunoassay (PEIA) employing both V3 and gp41 regions and a heteroduplex mobility assay (HMA) were evaluated in comparison with DNA sequencing. Of 146 samples selected, 115 (79%) were successfully sequenced. Taking sequence data as the "gold standard," other assays were compared with these data. The HMA correctly identified 95 (83%) of the samples, and only 1 sample was wrongly identified. The V3 PEIA alone and in combination with gp41 peptides correctly identified 76 and 78% of the samples, respectively; however, the number of wrongly identified samples was four times less with the combination compared with V3 peptides alone (4 versus 16%). The sensitivity, specificity, and positive and negative predictive values for serotype A and D samples were greater for the combination than V3 peptides alone. We have described a new algorithm to segregate subtypes A and D. This algorithm uses the two peptide assays followed by HMA and then DNA sequencing for untypable samples, giving an accuracy of 95% at a cost of 37 and 21% for consumables compared with subtyping all the samples by HMA or DNA sequencing, respectively. This proposed approach is suitable for epidemiological studies in Uganda and other regions with a predominance of A and D subtypes.