Over a 4-year period, the Roche Amplicor kit was used in a United Kingdom reference laboratory for the detection or confirmation of human immunodeficiency virus (HIV) type 1 infection, particularly in infants born to HIV-infected mothers. Of 408 specimens from adults and older children tested, the 122 seronegative specimens were all Amplicor negative. Of the 286 seropositive specimens, 268 were Amplicor positive. On the basis of these results, the Amplicor assay has a specificity of 100% and a sensitivity of 93.7%. In addition, for 247 specimens from infants and young children, serological results may not have been diagnostic because of placental transfer of maternal antibodies. Forty-eight were Amplicor positive, and of the 199 Amplicor-negative specimens, 19 were assumed to be false negative on the basis of clinical data, serological markers (including p24 antigen), and/or results for previous or follow-up specimens. This represents a sensitivity of 75% for the Amplicor test for specimens from patients under 2 years of age. Of these 37 false-negative specimens plus 2 specimens from other laboratories, 31 could be characterized by amplifying extracted material from them by an in-house nested gag PCR spanning the Amplicor target region. The amplicons were sequenced and found to represent subtypes A (35.5%), B (22.6%), C (22.6%), D (16.1%), and G (3.2%). False-negative results by the Amplicor assay may be ascribed to low-target copy number, the physical behavior of one primer (SK462), and sequence variation in the target region of the other primer (SK431).