Herpes simplex virus type 2: A possible interaction with human papillomavirus types 16/18 in the development of invasive cervical cancer

Abstract

A case-control study of 766 histologically confirmed incident cases of invasive cervical cancer and 1,532 hospital and community controls was conducted in Latin America to evaluate the etiologic role of herpes simplex virus type 2 (HSV-2) and to examine whether HSV-2 interacts with other risk factors. In addition to a personal interview, all subjects were asked to donate blood samples and cervical swabs for assessment of exposure to HSV-2 and human papillomaviruses (HPVs) respectively. Ninety-eight percent of cases and 91% of controls agreed to the interview and blood collection. Women testing positive for HSV-2 antibodies were found to have a 60% increased risk of cervical cancer compared with seronegative women (95% CI = 1.3, 1.9). Control for education, sexual and reproductive behavior, prior Pap-smear screening, smoking, oral contraceptive use, HPV-6/11 DNA, or HPV-16/18 DNA detection did not materially affect this estimate. No effect modification of HSV-2 by age, HPV-6/11 DNA, pregnancies, oral contraceptive use or cigarette smoking was observed. However, a significant interaction was detected between HSV-2 and HPV-16/18. Compared with women testing negative to both virus types, those positive for HSV-2 alone had a RR of 1.2 (95% CI = 0.9, 1.6), those positive for HPV-16/18 DNA alone had a RR of 4.3 (95% CI = 3.0, 6.0), and those positive for both viruses had a RR of 8.8 (95% CI = 5.9, 13.0). These findings corroborate recent laboratory evidence of a possible biological interaction between HSV-2 and HPV-16/18 in the development of cervical cancer. Further confirmatory studies are needed, given concerns with potential misclassification of exposure by the laboratory assays utilized.

Likelihood Based Confidence Intervals for the Difference of Proportions in Two Doubly Sampled Data with a Common False-Positive Error Rate

“…However, if we note that the false-positive error rate is a characteristic of a fallible device, not a characteristic of population, there are cases in which it may be logical to assume ϕ 1 = ϕ 2 . For instance, if the same fallible device is applied to obtain Y 1 and Y 2 as the case-control study of Hildesheim et al (1991), then the data sets probably have a common error rate and the joint likelihood of λ and…”

“…However, when the cost of precise classification is expensive, a researcher often uses an inexpensive but fallible classifier with a supplementary inerrant classifier. For instance, the case-control study of Hildesheim et al (1991) aimed to examine that invasive cervical cancer can affect exposure to Herpes Simplex Virus(HSV). For the study, western blot procedure, which was known to be relatively inaccurate in detecting the infection of HSV, was applied to about two thousand women in case and control groups.…”

Viruses, chemicals and co-carcinogenesis

“…Others have postulated a role for HSV-2 infection and cervical cancer, while acknowledging the primary role for HPV (Hildesheim et al, 1991;Daling et al, 1996). For example, Hildesheim et al studied women with invasive cervical cancer in Latin America and compared viral and behavioral characteristics with controls.…”

“…For example, Hildesheim et al studied women with invasive cervical cancer in Latin America and compared viral and behavioral characteristics with controls. Compared to women negative for both HPV 16/18 and HSV-2, those positive for HSV-2 alone had a relative risk of 1.2, those positive for HPV16/18 DNA alone had a relative risk of 4.3 (95% CI ¼ 3.0, 6.0), and those positive for both HSV-2 and HPV16/18 had a relative risk of 8.8 (95% CI ¼ 5.9, 13.0), suggesting a possible biological interaction (Hildesheim et al, 1991). Furthermore, HSV-2 was found to be persistent in some cervical cancer tumors (Frenkel et al, 1972).…”

Bayesian adjustment for exposure misclassification in case–control studies

“…Table 1 described the data structure. While each cell a ij in the validation data is fully specified (i = 0, 1,j = 1, 2, 3, 4), only margins a 05 , a 06 , a 15 , a 16 in the main data are recorded.…”