Human papillomavirus DNA in invasive cervical carcinomas and its association with patient survival: a nested case-control study.

This study sought to examine the association between the presence of human papillomavirus (HPV) DNA in invasive cervical cancer and prognosis. A case-control study was undertaken nested in a cohort of 208 patients with invasive cervical carcinoma in Montreal. All 40 deceased patients formed the case groups. A control group of equal size was selected by matching to each case (1:1) a patient of the same age and year of admission who had survived her disease. HPV DNA was detected by the use of a PCR protocol. The odds ratio (OR) for cervical cancer death was computed by logistic regression. Detection of HPV, particularly of types 16 and 18, was negatively correlated with disease stage and histological grade. The OR for death was 0.27 [95% confidence interval (CI), 0.1-0.8] for those whose tumors were positive for HPV DNA versus those in whom HPV DNA was not detected. After adjusting for the confounding effects of stage and grade, the prognostic effect was somewhat reduced, with an OR of 0.34 (CI, 0.1-1.1), which was no longer significant. The magnitude of the HPV-survival association was not altered when the analyses were restricted to carcinomas of stages I and II. Regardless of the mechanism for the prognostic role, detection of HPV DNA in primary tumors may play an important adjunct role in assessing prognosis of patients with invasive cervical cancer.