High incidence of cervical human papillomavirus infection in women during their first sexual relationship


*Mr S. Collins, Centre for Cancer Epidemiology, The University of Manchester, Kinnaird Road, Withington, Manchester M20 4QL, UK


The prevalence of cervical human papillomavirus increases with increasing numbers of sexual partners, leaving the impression that this infection is acquired only as a result of high risk sexual behaviour. Using longitudinal data from 242 women who had only had one sexual partner, we found that the risk of acquiring cervical human papillomavirus infection was 46% (95% CI 28–64) at three years after first intercourse and that the median time from first intercourse to first detection of human papillomavirus was only three months.


Case–control and cross sectional studies have reported that the prevalence of cervical human papillomavirus infection, which has been strongly and consistently associated with cervical neoplasia, increases with increasing numbers of sexual partners1. Popular reporting of this association has left the impression that cervical human papillomavirus infection is acquired as a result of high risk sexual behaviour and the adverse psychosocial consequences of a diagnosis of human papillomavirus infection are well-documented2. We report for the first time, using longitudinal data, the risk of cervical human papillomavirus infection associated with a woman's first sexual relationship.


Between 1988 and 1992, two thousand and eleven women aged 15 to 19 were recruited from a family planning clinic for a longitudinal study of the natural history of early cervical neoplasia. At recruitment a sexual history was taken and cervical samples taken and stored for future virological examination. Women were asked to reattend at intervals of six months, when their sexual history was updated and further samples taken. Follow up ended on 31 August 1997. The study protocol was approved by the appropriate ethical committee and informed oral consent was obtained from all women.

The study population for this analysis comprises a subset of 242 women who were recruited within six months of first having sexual intercourse and who had had only one sexual partner. Their baseline characteristics are given in Table 1. As described elsewhere3, after all clinical follow up had ended, cervical samples were tested for human papillomavirus DNA using general primer (GP5+/GP6+) mediated polymerase chain reaction; positive samples were further tested using type-specific primers to detect human papillomavirus types considered low risk (6/11), or high risk (16, 18, 31, 33, 52, 58) because of their association with cervical neoplasia. Women were assumed to be human papillomavirus negative on the date of first intercourse. Subsequent observations on human papillomavirus status were interval censored: the time of acquisition of human papillomavirus infection was known to lie in the six-month interval between the date of the first test which yielded a human papillomavirus positive result and the date of the immediately preceding test which yielded a negative result, but the exact time was unknown. Turnbull's modification of the product-limit estimator for interval-censored data was used to estimate the cumulative risk of human papillomavirus infection from first intercourse and to construct confidence intervals4. Follow up ended on the earliest of: date of acquisition of a second sexual partner or date of last visit.

Table 1.  Distribution of demographic and behavioural characteristics at study entry.
Mean age (SD)17.04(1.13)
Ethnic origin
South Asian4(1.7)
Socio-economic class according to father's occupation
III, non-manual19(7.9)
III, manual98(40.5)
Inadequately described14(5.8)
Smoking status
Alcohol consumption (units per week)
1 to 14202(83.5)
Age at first intercourse
Attended sexually transmitted disease clinic2(0.8)


Seventy-eight women tested positive for human papillomavirus; 26 for one or more high risk types, five for only low risk types, and 47 for other unidentified types. The cumulative risk at three years of cervical human papillomavirus infection was 46% (95% confidence interval 28% to 64%) (Fig. 1), and the median time from first intercourse to first detection of human papillomavirus was 2.6 months (range 0.3 to 59.0).

Figure 1.

The cumulative risk of cervical HPV infection during a woman's first sexual relationship.


There is compelling evidence to suggest that a cervical human papillomavirus infection, as distinct from human papillomavirus infections at more superficial sites which may follow vertical or horizontal transmission, is only acquired as a result of penetrative vaginal intercourse5. Our findings suggest that for many women this occurs shortly after beginning their first sexual relationship. Some women may have already had and cleared a cervical infection before recruitment, but because all women were recruited within six months of first intercourse it is unlikely that incidence has been substantially underestimated. Underreporting of partners is possible, but interviews were undertaken in an environment where confidentiality and anonymity have always been a priority; only two clinical research fellows and two research nurses conducted interviews during the entire study period and women appeared to show little reticence in discussing their sexual history. Women were also seen every six months, minimising reliance on recall. However, we cannot exclude the possibility that a woman remained monogamous within a relationship which was not itself monogamous, and may have become infected as a result of her partner acquiring the infection from an intercurrent relationship. Given that 180 of these women (74%) reported using barrier contraception, albeit intermittently, during this, their first relationship, it is difficult to see what further advice could be given to reduce the incidence of human papillomavirus infection. Perhaps cervical human papillomavirus infection should now be considered an inevitable consequence of sexual activity. Certainly, no stigma should be attached to its acquisition.


This study was supported by the Cancer Research Campaign. The authors would like to thank the following: the staff at the Birmingham Brook Advisory Centre; C. Meanwell for his contribution to study design; R. Lancashire for programming support; M. Whitlock for clerical assistance; J. Selby and P. Tilstons for technical assistance; M. Whitlock for clerical support; A. Sylvester and J. Cousins, the nurses who worked on the study, and the many others who facilitated the study and most importantly the women who participated in it.