To date, few studies evaluated age-specific HPV prevalence patterns and determinants of HPV infection among women with normal cytology in countries with high incidence of cervical cancer. We report here the first investigation of this type done in Colombia. An overall HPV DNA prevalence of 14.9% was found, which is similar to that reported in other high-risk populations in Mexico (14.5%) (Lazcano et al, 2001
) and Costa Rica (16%) (Herrero et al, 2000
). In addition, it is similar to the HPV-DNA prevalence among control women (13%) in a concurrent case–control study in Cali, Colombia (Muñoz et al, 1996
Several studies have suggested that the prevalence of HPV infection decreases with age, HPV being uncommon in cytological normal women over age 35 years (Melkert et al, 1993
; Morrison, 1994
). However, a few recent large population-based surveys from Costa Rica and Mexico also presented some increase in HPV infection among women in peri- and post-menopausal ages. While in Costa Rica the peak was observed in women older than 55 years, with a predominance of LR HPV types, in Mexico the second peak started earlier, after 45 years of age, with predominance of HR HPV types and with an increase in detection of LR HPV types compared to younger ages.
In our study the HPV prevalence in the age group 45–54 years was low, but there was an increased prevalence of HPV infection in women aged 55 or more years, with a predominance of LR types and multiple HPV infections, but still 5.7% were HR HPV positive. If confirmed, the risk of HPV infection – most notably LR types – among post-menopausal women, it may be explained in different ways: (1) reactivation of latent HPV infections by decreased immune response. Some immunological studies show a decrease in circulating mature T cells in older people as a result of a decrease in CD8+ lymphocytes, a decline in the frequency of CD4+ T cells producing IL-2 and/or a decreased expression in IL-2 receptors (Lesourd and Meaume, 1994
; Gostout et al, 1998
; Ginaldi et al, 1999
). (2) Reactivation of latent HPV infections by hormonal changes related to the gradual decline of ovarian function around menopause (Lazcano et al, 2001
). (3) A cohort effect (Herrero et al, 2000
), where older women were exposed to the virus early in life and belonged to generations more heavily exposed to HPV infections.
An alternative explanation for this second peak lies in the fact that only women with normal cervical cytology were included in this study. Persistent infections with HR HPV types will lead to cervical intraepithelial lesions with subsequent treatment. It may result in an under-representation of HR HPV types and the proportion of them in the population will diminish at older age. In contrast, LR HPV infections that rarely give rise to cervical dysplasia will remain (Jacobs et al, 2000
We identified 32 different HPV types in our study population. This broad diversity of infections is consistent with previous studies showing a greater HPV heterogeneity in mild cervical dysplasias than in severe dysplasias (de Roda Husman et al, 1994
; Liaw et al, 1999
; Jacobs et al, 2000
; Lazcano et al, 2001
). It has been estimated that at least 50% of sexually active adults have ever had a genital HPV infection, most of these, however, being transient and resolving spontaneously (Koutsky, 1997
In our study HPV 16, 58, 56, 52, HPV 81 (CP8304), 51 and 18 were the most prevalent HPV types detected as single infections in all age groups. Although HPV16 is the predominant type detected in the majority of studies (Jacobs et al, 2000
; Lazcano et al, 2001
; Rolon et al, 2000
), remarkable differences were noted in our study population with a high prevalence of other HPV types such as HPV 56, 58 and the LR type HPV 81 (CP8304). This information should be taken into account when developing HPV vaccines tailored to this population.
In this study HPV 26, 53, 73, 34 and Iso 39 were analysed as types with oncogenic potential according their presence in patients with cervical cancer and the alignment analysis of the E6 gene (modified from Myers et al, 1996
). However the percentage of these HPVs is very low and for absolute HR or LR classification, transformation studies in combination with follow up epidemiological studies are necessary.
The presence of multiple infections (29.7%) was higher than previously observed among control subjects in the IARC studies done in Brazil (0%) (Eluf-Neto et al, 1994
), the Philippines (14.3%) (Ngelangel et al, 1998
) Thailand (9.8%) (Chichareon et al, 1998
), Morocco (5.3%) (Chaouki et al, 1998
) and Paraguay (16.7%) (Rolon et al, 2000
), lower than in a population-based study from Costa Rica (39%) (Herrero et al, 2000
) and similar to the results from a study done in The Netherlands (28%) (Jacobs et al, 2000
), where 3305 cytological normal cervical scrapes were analysed using the same laboratory technique as in the present study. These differences in prevalence of multiple infections could be due to differences in the technique used (sensitivity, specificity and types identified), or real differences in the prevalence of the HPV types searched for and identified in the populations studied. In addition, these populations differed in age-composition: mean age 32 years in this study, whereas in Thailand it was 49.7 years. Taking into account the age-dependent prevalence of multiple infections, this fact may also explain the discrepancy in results. In our study, 97% of the women with multiple infections presented a HR HPV type. The long-term follow up of these infections will help to clarify the role of HR multiple infections in the development of cervical lesions.
Besides age, number of regular sexual partners and OC use (specially in women below age 35 years) which were risk factors for HPV positivity, other reproductive and sexual behaviour factors considered by us were only weakly and inconsistently related to HPV infection. It might be explained by the relatively high frequency of HPV infections in the studied women and possibly the predominance of ‘male role’ in the transmission of HPV infections to women (particularly over age 35 years). Unfortunately, no information was available on the sexual behaviour of the partners in our study population. Our results suggest an age-dependent association between HPV infection and some risk factors. A possible explication of the highest prevalence of HPV infections among young, highly educated women and with more than one regular or casual sexual partner could reflect changes in lifestyle and sexual behaviour in younger generations (possibly reduced influence of religion and greater freedom).
We observed an increase of HPV infections and particularly of LR types in women older than 55 years and we tried to analyse if these women were different in terms of exposure to risk factors. Due to the low number of women in this group we could not find a clear pattern of association. We found that use of oral contraceptives may be a risk factor for both HR and LR HPV infections. Use of exogenous hormones – as contraceptives or hormone replacement – could influence HPV infections in different ways. Results from laboratory work have indicated that the HPV genome contains a hormone-recognition segment, which might show an interactive effect of oral contraceptives and HPV (Monsonego et al, 1991
). Additionally, steroid hormones do interfere with cellular gene function involved in cell cycle regulation and apoptosis, and they also might inhibit the immunologically mediated resolution of minor HPV-induced cervical lesions (Von Knebel Doeboritz et al, 1997
With respect to parity, in our study there was a trend of decreased risk of HR and multiple HPV infections with increase of number of births. Previous studies have presented conflicting results: an increase in risk with number of births (Hildesheim et al, 1993
) or decrease or no association with parity at all (Lazcano et al, 2001
). There is still insufficient data to give final conclusions about the effect of number of births on the risk of HPV infections.
In conclusion, our results showed a broad diversity of HPV infections in women with normal cytology, with HR types being at least three times more common than LR types. Whereas young women were particularly at risk of HR and multiple HPV infections, post-menopausal women showed an increased frequency of LR HPV types. In this population, high educational level and multiplicity of sexual partners seem to determine risk among young women. No other hypothesized risk factor was clearly associated with the risk of HPV infection.