The prevalence of human papillomavirus type 58 in chinese patients with cervical carcinoma and its influence on survival

To assess the prevalence of human papillomavirus (HPV) type-specific infections and its influence on prognosis and survival rate of cervical cancer patients treated with radiotherapy alone or in combination with chemotherapy or hyperthermotherapy or both. During 1993 and 1994, 120 patients with cervical cancer were consecutively assigned into four therapeutic modalities. One hundred and eight formalin-fixed and paraffin-embedded tissue samples available at the time of this study were subjected to DNA extraction. The DNA was analysed for the presence of HPV and HPV type-specific infections with polymerase chain reaction. The experimental data were co-analysed with the follow-up data to evaluate prognosis and survival in relation to the HPV infection and treatment. All samples were found to be positive for HPV; 32.40% (95% confidence interval 23.72-42.09) of samples were HPV 16 single positive, 16.67% (95% confidence interval 10.19-25.06) were HPV 58 single positive, 34.26% (95% confidence interval 25.4-44.01) were HPV 16/58 double positive and 16.67% (95% confidence interval 10.19-25.06) were positive for other HPV types. The 5-year survival rates were 88.89, 80.0, 67.57 and 61.11% in patients with HPV 58 single positive, HPV 16 single positive, HPV 16/58 double positive, and other HPV types, respectively (P = 0.18), irrespective of the treatment modalities. Univariate Cox regression analysis results showed that compared with the HPV 16 single positive group, the relative risk of death was high in the HPV 16/58 double positive group (hazard ratio 1.81, 95% confidence interval 0.71-4.60) and the other HPV types group (hazard ratio 2.09, 95% confidence interval 0.73-5.96), whereas it was lower in the HPV 58 single positive group (hazard ratio 0.53, 95% confidence interval 0.11-2.56), but this was not statistically significant (P = 0.43). All results remained statistically non-significant after adjustment for age and tumour stage with multivariate Cox regression. No apparent differences were observed in the survival rates among the four groups with regard to the treatment protocols. No effect was seen of either treatment protocols or HPV type on survival, except for a slightly high survival in HPV 58 single-infected patients, which was statistically non-significant.