Detection and typing of human papillomaviruses combining different methods: polymerase chain reaction, restriction fragment length polymorphism, line probe assay and sequencing.

The identification of the etiological factor of many cervical precancerous lesions and cervical cancer, the human papillomavirus (HPV) is widely used. In this study, we evaluated the consensus and type-specific (TS) polymerase chain reaction (PCR), restriction fragment length polymorphism (RFLP), line probe assay (LiPA, Innogenetics) and sequencing to determine the HPV types in cervical specimens. Out of 690 High-grade Squamous Intraepithelial Lesion (HSIL) samples, 86.7% were HPV positive and 13.3% HPV negative by consensus primers (MY09/MY11, L1C1/L1C2-1/L1C2-2 and/or GP5/6) directed PCR. Out of 598 HPV positive samples, 85.3% were typed by TS-PCR being HPV 6/11, 16, 18, 31 and/or 33, while 14.7% remained untyped. The most prevalent HPV type in the study group was HPV 16, identified in 35.5% cases, while HPV 31 was the second most frequent HPV type with a prevalence of 10.5%. They were followed by HPV types 6/11, 33 and 18 with a prevalence of 7.4%, 6.2% and 4.9%, respectively. Multiple HPV infections with two or more HPV types (6/11, 16, 18, 31 and/or 33) were found in 9.4% cases. A subset of 88 samples was further typed by RFLP and LiPA to determine the rare HPV types in HSIL samples. The most frequent low abundant HPV types in single infections in decreasing order were HPV 53, 58, 66, 56 and 52, while HPV 51 was the most frequent low abundant HPV type found in multiple HPV infections. Multiple HPV infections were mostly found by LiPA in 27.3% cases versus 14.8% cases found by RFLP. The perfect agreement between RFLP and LiPA assay pair was observed only for HPV types 16, 18, 34 and 59 (kappa value of 1). For other HPV types, the inter-assay agreement ranged from very good to no agreement indicating that neither assay is perfect. Sequencing was performed for 33 samples in cases where both RFLP and LiPA were inconclusive. Sequencing was shown to be a very good method in case of single HPV infection but not applicable in case of multiple HPV infections. Both RFLP and LiPA are good assays for epidemiological studies, although RFLP being cumbersome and time-consuming is less applicable than LiPA. Careful consideration has to be made before the implementation of either HPV typing methods in clinical laboratories.