Using the polymerase chain reaction to genotype human papillomavirus DNAs in samples containing multiple HPVs may produce inaccurate results.

Compared with other laboratory techniques, the polymerase chain reaction (PCR) is a simple, rapid, sensitive method for detecting human papillomavirus (HPV) DNA in cervical samples. However, since many cervical samples contain multiple HPV types, we decided to investigate whether PCR results from such samples accurately reflected the relative amounts of each HPV type present. Theoretical calculations of product accumulation when multiple DNAs with different amplification efficiencies are present in the same sample were done. In addition a set of samples in which cloned HPV DNAs were mixed in varying proportions prior to PCR was tested. Finally, four clinical samples containing multiple HPV types by hybridization assays were subjected to PCR, using two different primer sets. Each of these lines of investigation showed that selective amplification of one HPV DNA over another can occur when mixed HPV types are present. This effect may lead to inaccurate information regarding both types and relative amounts of HPV DNAs in samples containing multiple HPV types. A protocol to avoid this problem is described.