To determine whether the status of human-papillomavirus (HPV) infection affects the clinical outcome of cervical carcinoma (CC), HPV genotype was prospectively determined in 94 consecutive CC cases subsequently followed for a median duration of 37.5 months. With a consensus PCR-RFLP method of HPV genotyping, 81 (86.2%) cancers were positive for HPV DNA. They were classified, according to the phylogenic similarities, into HPV-16-related (type 16, n ‫؍‬ 45; type 31, n ‫؍‬ 2), HPV-58-related (type 58, n ‫؍‬ 17; type 33, n ‫؍‬ 3; type 52, n ‫؍‬ 2) and HPV-18-related (type 18, n ‫؍‬ 8; type 68, n ‫؍‬ 1) groups, and analyzed in relation to clinical outcome. The following results were observed: (i) Type-58-related HPVs were more prevalent in the old age (older than the median age of 52) group than in the young age group (41% vs. 14.6%, p ‫؍‬ 0.045); (ii) 63% (5/8) of patients with advanced stages (III and IV) were HPV-negative, a figure much higher than that Cervical carcinoma (CC) is one of the most common malignancies both in incidence and in mortality in women worldwide. Epidemiologic and laboratory studies have revealed a multifactorial and multi-step process, with infection of oncogenic types of human papillomavirus (HPV) standing for the predominant role (zur Hausen, 1991; Muñoz et al., 1992). The integration of HPV DNA into host genome and over-expression of the E6 and E7 oncogenes appear to be critical steps in cancer development (Phelps et al., 1988). Consequently, HPV DNA is present in most high-grade squamous intra-epithelial lesions (HSIL) and invasive cancers of the uterine cervix (Muñoz et al., 1992).On the basis of differences in DNA sequences, more than 70 types of HPV have been identified, and about 20 of them are associated with CC. They were grouped, according to their prevalence in SIL and CC, into intermediate-risk types (31, 33, 35, 51, 52 and 58) and high-risk types (16, 18, 45 and 56) (Lorincz et al., 1992). But the biologic behavior and clinical relevance of this classification have not been confirmed. However, several reports have noted that HPV genotypes associated with lesions of similar pathologies are phylogenetically related (De Villiers, 1989;Chan et al., 1992;Bernard et al., 1994), and indicated amino-acid-sequencedependent oncogenic behavior of HPV. Accordingly, a phylogenetic study has related, by order of phylogenic branching, HPV types 58/33/67/52, types 16/31/35 and, in remote branches, types 56/51 and 18/68/45 (Bernard et al., 1994).A number of studies of the influence of HPV status on the clinical features of cervical cancers have produced conflicting results. HPV 16 was found to be more prevalent in welldifferentiated carcinomas (Riou et al., 1990), while HPV 18 was associated with adenocarcinoma (AC) and adenosquamous carcinoma (ASC) (Higgins et al., 1991;Chen et al., 1994) and younger age (Higgins et al., 1991;Lorincz et al., 1992;Nakagawa et al., 1996). In addition, HPV-18-positive CCs were found to have a poor prognosis (Rose et al., 1991;Nakagawa et al., 1996;Lombard et al...