PIP: This paper addresses HIV detection and its variation using degenerate polymerase chain reaction (PCR) and sequence analysis among African children. A total of 6 genomes of HIV-1 from AIDS patients, representing geographically distinct regions of Zambia and the major circulating genotypes (A, B, D, and O) were examined. Sequence multiple alignment was used to determine matches of HIV-1 and its variation, but none was suitable; hence, alignments were re-examined to design new primers. Standard PCR conditions were used with a modified cycling protocol. The new primers were tested on blood DNA from 53 HIV-negative, 60 HIV-positive febrile infants, and 9 HIV-positive children with Kaposi's sarcoma (KS). HIV status was determined in an enzyme-linked immunosorbent assay. 6 of the HIV-negative infants, 43 of those who were seropositive, and all children with KS were positive for HIV-1 proviral DNA. One PCR-positive, HIV-seronegative infant and the 9 KS samples were examined further using automated DNA sequencing and showed no evidence for contamination. Multiple alignment and phylogeny analyses using the Clustal program showed most similarity (94%) to the published adult Zambian strains. Variations in PCR detection rate in the HIV-positive infants (72%) and the KS children (100%) were brought about by confounding factors such as maternal HIV-1 infection during pregnancy without seroconversion, impaired B-cell function, and diversity of HIV-1 genotype circulating in Zambia.