Int. J. Cancer: 103, (2003) 2002 Wiley-Liss, Inc.

Transcription

1 Int. J. Cancer: 103, (2003) 2002 Wiley-Liss, Inc. Publication of the International Union Against Cancer PREVALENCE OF SPECIFIC TYPES OF HUMAN PAPILLOMAVIRUS AND CERVICAL SQUAMOUS INTRAEPITHELIAL LESIONS IN CONSECUTIVE, PREVIOUSLY UNSCREENED, WEST-AFRICAN WOMEN OVER 35 YEARS OF AGE Long Fu XI 1, Papa TOURÉ 3, Cathy W. CRITCHLOW 2, Stephen E. HAWES 2, Birama DEMBELE 3, Papa Salif SOW 3 and Nancy B. KIVIAT 1 * 1 Department of Pathology, School of Medicine, University of Washington, Seattle, WA, USA 2 Department of Epidemiology, School of Public Health, University of Washington, Seattle, WA, USA 3 Department of Infectious Diseases, University of Dakar, Senegal Previous studies among women worldwide have demonstrated that infection with specific types of human papillomaviruses (HPV) is central to the pathogenesis of cervical neoplasia. There is little data, however, concerning the prevalence of specific HPV types and the association of each type with cervical neoplasia among women in sub-saharan Africa, who remain at very high risk of cervical cancer. We studied 2,065 consecutive patients aged 35 years or older, presenting to community health clinics in Dakar and Pikine, West Africa, who had not been screened previously for cytologic abnormalities or HPV. Cytologic diagnosis and HPV detection were accomplished using a ThinPrep Pap and a polymerase chain reaction-based reverse-line strip assay, respectively. Odds ratios (OR) and associated 95% confidence intervals (CI) were estimated using polynomial logistic regression. Cytologic abnormalities were found in 426 women (20%), including 254 (12%) with atypical squamous cells of undetermined significance, 86 (4%) with low-grade squamous intraepithelial lesions, 66 (3%) with high-grade squamous intraepithelial lesions (HSIL) and 20 (1%) with invasive cancer. HPV infection was detected in 18%. Among women with negative cytologic findings, the prevalence of high risk but not low risk HPV types increased with age. HPV16 (2.4%) and HPV58 (1.6%) were the most frequently detected HPV types in this population, as well as being the most strongly associated with risk of HSIL/cancer (HPV16: OR 88, 95% CI ; HPV58: OR 51, 95% CI ). These data suggest that in addition to HPV16, HPV58 should be considered in the strategic planning of vaccination against cervical cancer in this geographic region Wiley-Liss, Inc. Key words: human papillomavirus; squamous intraepithelial lesion; Africa Genital human papillomaviruses (HPV) are common sexually transmitted viruses and a subset, termed high risk types, are integral to the development of invasive cervical cancer 1,2 and its cytologic precursors, squamous intraepithelial lesions (SIL). 3 5 HPV16 is the most common HPV type found in the cervix and in cervical cancers, where it is detected in over 50% of cases. 6 Other HPV types detected commonly in cervical cancers include types 18, 45 and Development of appropriate HPV vaccines requires knowledge of the prevalence of various HPV types in neoplastic and nonneoplastic cervical lesions in populations at high risk of cervical cancer. Thus far, data concerning type-specific HPV prevalence has come primarily from studies among women in Europe and North and South America Although there appears to be little geographic variation in the prevalence of HPV16, this does not seem to be the case for many other HPV types. For example, HPV39 and HPV59 are confined largely to cervical cancers in Central and South America; HPV types 56 and 68 are relatively common elsewhere but not in Central and South America. 6 Further, HPV52 and 58 are rare in North America and Europe, but relatively prevalent in parts of East Asia Thus far, little data is available concerning the prevalence of specific HPV types among women with and without cervical lesions, in sub- Saharan Africa where Pap screening is not available routinely and the incidence of invasive cervical cancer remains high. 21 We conducted the current study among previously unscreened women in Senegal to determine the prevalence of specific HPV types in relation to cervical neoplasia. MATERIAL AND METHODS Subject eligibility and recruitment Consecutive women, aged 35 years or older, not pregnant currently and with an intact cervix, presenting to community health clinics in Dakar, Senegal or Pikine (a suburb of Dakar) were eligible to enroll. After obtaining informed consent, a face-to-face interview was conducted soliciting information on demographic characteristics and reproductive history and a general physical and detailed gynecological examination was carried out. Study procedures were approved by the institutional review boards of the University of Washington and the University of Dakar. Between January 1998 and August 2000, 2,420 women, representing approximately 75% of all eligible women, were approached, with all agreeing to participate. None had been screened previously for cervical cytologic abnormalities or HPV infection. On review, 53 women were found to be pregnant at the time of study entry and were excluded from further analysis. We additionally excluded 40 women for whom cytologic data were unavailable, 172 women whose cervical samples were unsatisfactory for cytologic evaluation, 49 women for whom HPV results were unavailable and 41 women whose cervical samples were insufficient for HPV testing, leaving 2,065 women (85.3% of the enrolled women) available for analysis. Specimen collection and cytologic diagnosis Each participant provided an exfoliated cervical cell sample that was obtained with a cervical brush and placed in PreservCyt Informed consent was obtained from all participants in accordance with guidelines of the US Department of Health and Human Services. This study was approved by the institutional review boards of the University of Washington, Seattle and the University of Dakar, Senegal. Grant sponsor: National Cancer Institute; Grant number: CA 75920, CA *Correspondence to: HPV Research Group, University of Washington, 1914 N 34th Street, Suite 300, Seattle, WA Fax: nbk@u.washington.edu Received 28 June 2002; Revised 17 October 2002; Accepted 18 October 2002 DOI /ijc.10876

2 804 XI ET AL. (Cytyc Corporation, Marlborough, MA). A sample for HPV testing was obtained with a Dacron swab, placed in Specimen Transport Medium (STM; Digene Corporation, Silver Spring, MD), stored at 20 C, frozen and shipped to Seattle. ThinPrep smears were prepared and reviewed by a cytotechnologist and a pathologist. Abnormalities were classified according to the Bethesda classification system 22 as within normal limits or reactive cellular changes (normal), atypical squamous cells of undetermined significance (ASCUS), low-grade SIL (LSIL), highgrade SIL (HSIL) or carcinoma. Detection and genotyping of HPV A 100- l aliquot of the cervical sample collected in STM was processed as described previously. 23 Briefly, the sample was precipitated with 400 l of absolute ethanol containing ammonium acetate, centrifuged and the pellet dried and re-suspended. HPV detection and typing analyses were carried out using a polymerase chain reaction (PCR)-based reverse-line strip test for HPV types 16, 18, 26, 31, 33, 35, 39, 42, 45, 51 59, 66, 68, 73, 82, 83 and 84, as well as HPV Types 6, 11, 40 and -globin as described previously. 24 In addition, a biotin-labeled generic probe was used to target all HPV DNA fragments generated using a pair of MY09- MY11 primers. 23 Samples hybridizing with the generic probe but not with any of the type-specific probes were considered to be positive for unclassified HPV types. HPV negative control samples (K562 human DNA cell line DNA) were interspersed between clinical specimens to evaluate potential inter-specimen contamination. Positive control samples served to monitor the overall amplification and hybridization process. Control specimens were processed with all clinical specimens and each individual hybridization reaction tray contained at least 1 positive and 1 negative control. HPV testing was carried out by personnel who had no knowledge of a subject s medical history or cytologic diagnosis. Statistical analyses The primary focus of the analysis was to determine the risk of cervical lesions associated with specific HPV types. HPV types previously shown to be associated with cervical cancer, including types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59 and 68, were categorized as high risk types. 6 Remaining types detected by the reverse-line strip test were categorized as low risk types. Each woman was classified as HPV positive or negative according to whether or not HPV DNA was detected in a cervical sample. Women in whom HPV DNA was detected were also characterized as having either a single HPV type or multiple HPV types detected. In analyses assessing cervical lesion risk associated with detection of multiple HPV types, women were classified hierarchically as infected with any high risk, as compared to only low risk HPV types. Odds ratios (OR) and 95% confidence intervals (CI) for associations between risk of cervical lesions (dependent variable) and HPV status (independent variable) were estimated using polynomial logistic regression 25 (Strata 5.0 for Windows, Strata Corporation, College Station, TX). Unless otherwise stated, OR were adjusted for age at study entry (35 39, 40 44, 45 49, 50 54, 55 years), marital status (monogamously or polygamously married, widowed, divorced, never married) and number of prior pregnancies (0 4, 5 8, 9). For analyses regarding risk of cervical lesions associated with number of HPV types detected, we excluded women who were positive for unclassified HPV types because the actual number of types present was unknown. In examining type-specific risk associations, we initially evaluated risk associated with a given HPV type after controlling for infection with HPV types other than the one of interest. We further evaluated this risk in analyses among women who were positive for that HPV type alone in comparison to women who were negative for all HPV types. A 2 test for trend was used to test whether an increasing proportion of HPV positivity was associated with increasing severity of cytologic diagnosis. All reported p- values are 2-sided. RESULTS Demographic and behavioral information as well as cervical samples adequate for cytologic diagnosis and HPV testing were available for 1,776 women from Pikine and 289 women from Dakar. The mean age of enrolled women was 42.7 (s.d. 6.4) years. Fifty-three percent of the subjects were in polygamous marriages and 10% were single, divorced or widowed. Eleven percent of women reported use of oral contraceptives, although the majority of subjects were not currently practicing contraception. Most women (98%) reported having been pregnant at least once, with 38% having 9 or more pregnancies. Forty-nine percent of gravid women had their first pregnancy before age 18 and 62% reported having had a miscarriage. Cytologic abnormalities were found in 426 (21%) of the 2,065 subjects, including 254 (12%) with ASCUS, 86 (4%) with LSIL, 66 (3%) with HSIL and 20 (1%) with cervical cancer. Women with HSIL or cancer (HSIL/cancer) as compared to those with normal cytologic findings were somewhat more likely to be 50 years of age or older, be in a polygamous marriage, have 9 or more pregnancies, a history of miscarriage and their first pregnancy before age 18 (Table I). Prevalence of HPV infection Overall, HPV DNA was detected in 366 (18%) of 2,065 subjects, including 13% of 1,639 women with normal cytology results, 21% of 254 women with ASCUS, 51% of 86 women with LSIL and 76% of 86 women with HSIL/cancer (p trend 0.001). The prevalence of infection with any HPV type was highest (23%) among women aged 55 years or older and lowest (14%) among those aged years. Among women with normal cytology findings (Fig. 1), HPV prevalence generally increased with age and was highest among women 55 years of age or older (20%, p 0.03). This age-specific prevalence pattern was also evident for high risk and unclassified HPV types. HPV DNA was somewhat more likely to be found in cervical swab samples from women in polygamous marriages and those with 5 or more pregnancies, but HPV DNA detection was not associated with oral contraceptive use, younger age at first pregnancy, or history of miscarriage. HPV prevalence and cervical cytology findings were similar between women from Pikine and women from Dakar (data not shown). We found that HPV16 and 58 were the most common types detected, both overall (2.4% and 1.6%, respectively) and among women with HSIL/cancer (23% and 13%, respectively; Table II). HPV18, 33, 52, 53 and 54 were detected in greater than 1% of women screened and among women with HSIL/cancer, in 4.7%, 8.1%, 8.1%, 2.3% and 1.2%, respectively. Although infrequently detected in the overall population, HPV31 and 51 were present in 5.8% and 3.5%, respectively, of women with HSIL/cancer. Types 11, 26, 42, 45 and 68 were extremely rare in this population, with each type found in only 1 or 2 samples. None of the samples examined was positive for HPV types 40, 55 or 57. Among the 246 women with typed HPV infections, 47 (19%) were positive for more than 1 HPV type, including 35 with 2 types, 11 with 3 types and 1 with 5 HPV types. Association of cervical lesions with detection of specific HPV types The presence of cervical abnormalities was highly associated with the detection of any HPV (ASCUS: OR 1.8, 95% CI ; LSIL: OR 7.8, 95% CI ; HSIL: OR 22, 95% CI 13 37; Table III). Most women with HPV detected were infected with high risk (with or without low risk) types that were detected in 181 (9%) of 2,065 women; low risk HPV types were detected in 65 (3%) women. Not surprisingly, HSIL/cancer was strongly associated with detection of high risk as compared to low risk HPV types, however, there was little difference in the risk

3 HPV IN RELATION TO RISK OF SIL 805 Characteristic TABLE I DEMOGRAPHIC AND REPRODUCTIVE CHARACTERISTICS OF STUDY PARTICIPANTS IN DAKAR, SENEGAL, , BY CERVICAL CYTOLOGY DIAGNOSIS 1 (n 1,639) ASCUS (n 254) Cervical cytology diagnosis LSIL HSIL/cancer Age (years) (35.1) 81 (31.9) 39 (45.3) 22 (25.6) (30.6) 92 (36.2) 24 (27.9) 29 (33.7) (21.7) 57 (22.4) 17 (19.8) 18 (20.9) (7.7) 11 (4.3) 2 (2.3) 8 (9.3) (4.9) 13 (5.1) 4 (4.6) 9 (10.5) Marital status Monogamous marriage 610 (37.3) 91 (35.8) 35 (40.7) 23 (26.7) Polygamous marriage 861 (52.6) 132 (52.0) 44 (51.2) 54 (62.8) Widowed, divorced, or single 166 (10.1) 31 (12.2) 7 (8.1) 9 (10.5) Current use of oral contraceptives No 1,465 (89.4) 220 (86.6) 78 (90.7) 79 (91.9) Yes 173 (10.6) 34 (13.4) 8 (9.3) 7 (8.1) Number of pregnancies (19.6) 60 (23.6) 25 (29.1) 8 (9.3) (42.8) 102 (40.2) 39 (45.3) 33 (38.4) (37.6) 92 (36.2) 22 (25.6) 45 (52.3) Age at first pregnancy 2 (years) (51.0) 137 (55.9) 46 (54.1) 31 (36.9) (49.0) 108 (44.1) 39 (45.9) 53 (63.1) Having a history of miscarriage 2 No 565 (37.8) 92 (39.8) 36 (43.4) 23 (28.7) Yes 930 (62.2) 139 (60.2) 47 (56.6) 57 (71.3) 1 Numbers may not add up to the total n due to missing values. ASCUS, atypical squamous cells of undetermined significance; LSIL, low-grade squamous intraepithelial lesion; HSIL/cancer, high-grade squamous intraepithelial lesion or cancer. Values are n (%). 2 Among women who had at least one pregnancy. FIGURE 1 Age-specific prevalence of human papillomavirus (HPV) infection among 1,639 Senegalese women with normal cervical cytology results. High risk types include HPV16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59 and 68. Low risk types include HPV6, 11, 26, 42, 53, 54, 66, 73, 82, 83 and 84. Co-infection with both high- and low-risk HPVs was counted in each of the risk categories. of ASCUS or LSIL associated with high risk, as compared to low risk HPV types. Of the 366 women in whom HPV was detected, 120 (33%) were type-unclassified. Infection with an unclassified HPV type was associated with a substantially increased risk of HSIL/cancer (OR 8.1, 95% CI ) and a modestly increased risk of LSIL (OR 3.1, 95% CI ) and ASCUS (OR 1.7, 95% CI ). Risk of HSIL/cancer tended to be higher among women in whom multiple HPV types as compared to only 1 HPV type were detected, although the difference was not statistically significant (p 0.23). Similar results were observed when the analysis was restricted to women with at least one high risk HPV type detected (data not shown). Last, we examined the risk of cervical neoplasia associated with each of the 7 most frequently detected HPV types (using women with normal cytology findings as the comparison group and adjusting for other HPV types; Table IV). HSIL/cancer was most highly associated with detection of HPV types 16 (OR 56, 95% CI ), 58 (OR 19.8, 95% CI 7 57) and 33 (OR 16, 95% CI 5 52). Risk of HSIL/cancer was also associated with infection with high risk HPV types 18 and 52, but not with detection of HPV53 or 54 (low risk types). LSIL was primarily associated with detection of HPV types 58 (OR 13.7, 95% CI ) and 16 (OR 9.4, 95% CI ), whereas ASCUS was not significantly associated with the presence of any of the 7 most common HPV types found in this population. We additionally compared women in whom only 1 HPV type was detected to women who were negative for all HPV types (data not shown). The risks for HSIL/cancer remained most strongly associated with infection with HPV16 (OR 88, 95% CI ) and HPV58 (OR 51.2, 95% CI ). LSIL was highly associated with detection of HPV58 (OR 30, 95% CI ) and HPV16 (OR 14.6, 95% CI ). Among women positive for a single low risk type (HPV53 or HPV54), none had a cytologic diagnosis consistent with HSIL/ cancer. DISCUSSION To our knowledge, our study with over 2,000 consecutive, previously unscreened Senegalese women aged 35 years or older, is the largest to date examining the prevalence of specific HPV types in relation to cervical neoplasia in Africa. Previous studies

4 806 XI ET AL. TABLE II PREVALENCE OF TYPE SPECIFIC HUMAN PAPILLOMAVIRUS (HPV) INFECTION BY CERVICAL CYTOLOGY DIAGNOSIS 1 HPV type (n 1,639) Cervical Cytology Diagnosis ASCUS (n 254) LSIL HSIL/cancer HPV6 4 (0.2) 0 (0.0) 2 (2.3) 2 (2.3) HPV11 0 (0.0) 0 (0.0) 1 (1.2) 0 (0.0) HPV16 17 (1.0) 6 (2.4) 7 (8.1) 20 (23.3) HPV18 14 (0.9) 4 (1.6) 4 (4.7) 4 (4.7) HPV26 0 (0.0) 0 (0.0) 1 (1.2) 0 (0.0) HPV31 6 (0.4) 2 (0.8) 1 (1.2) 5 (5.8) HPV33 12 (0.7) 4 (1.6) 2 (2.3) 7 (8.1) HPV35 0 (0.0) 1 (0.4) 3 (3.5) 1 (1.2) HPV39 2 (0.1) 3 (1.2) 0 (0.0) 1 (1.2) HPV42 1 (0.1) 0 (0.0) 1 (1.2) 0 (0.0) HPV45 4 (0.2) 1 (0.4) 0 (0.0) 0 (0.0) HPV51 5 (0.3) 0 (0.0) 1 (1.2) 3 (3.5) HPV52 9 (0.5) 1 (0.4) 3 (3.5) 7 (8.1) HPV53 12 (0.7) 3 (1.2) 3 (3.5) 2 (2.3) HPV54 16 (1.0) 6 (2.4) 1 (1.2) 1 (1.2) HPV56 5 (0.3) 1 (0.4) 1 (1.2) 1 (1.2) HPV58 12 (0.7) 2 (0.8) 8 (9.3) 11 (12.8) HPV59 6 (0.4) 0 (0.0) 2 (2.3) 0 (0.0) HPV66 3 (0.2) 1 (0.4) 1 (1.2) 2 (2.3) HPV68 1 (0.1) 0 (0.0) 0 (0.0) 1 (1.2) HPV73 5 (0.3) 2 (0.8) 1 (1.2) 0 (0.0) HPV82 1 (0.1) 0 (0.0) 0 (0.0) 2 (2.3) HPV83 13 (0.8) 2 (0.8) 0 (0.0) 2 (2.3) HPV84 0 (0.0) 3 (1.2) 1 (1.2) 1 (1.2) Single type detected 102 (6.2) 24 (9.4) 31 (36.0) 42 (48.8) Multiple types detected 20 (1.2) 8 (3.1) 6 (7.0) 13 (15.1) 2 genotypes 14 (0.9) 6 (2.4) 5 (5.8) 10 (11.6) 3 genotypes 6 (0.4) 2 (0.8) 1 (1.2) 2 (2.3) 5 genotypes 0 (0.0) 0 (0.0) 0 (0.0) 1 (1.2) 1 Infection with more than one HPV type is counted multiple times according to the number of types detected. ASCUS, atypical squamous cells of undetermined significance; LSIL, low-grade squamous intraepithelial lesion; HSIL/cancer, high-grade squamous intraepithelial lesion or cancer. Values are n (%). TABLE III ASSOCIATION OF CERVICAL NEOPLASIA WITH DETECTION OF HUMAN PAPILLOMAVIRUS (HPV) AND PRESENCE OF MULTIPLE HPV TYPES HPV status (n) Cervical Cytology Diagnosis ASCUS LSIL HSIL/cancer n OR (95% CI) 3 n OR (95% CI) 3 n OR (95% CI) 3 Any HPV type Negative 1, Positive ( ) ( ) ( ) Risk category 1 Low risk type ( ) ( ) ( ) High risk type ( ) ( ) ( ) Unclassified type ( ) ( ) ( ) Number of HPV types detected 2 Single type ( ) ( ) ( ) More than one type ( ) ( ) ( ) 1 High risk types include HPV16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59 and 68. Low risk types include HPV6, 11, 26, 42, 53, 54, 66, 73, 82, 83 and 84. For those infected with multiple types, HPV typing results are grouped hierarchically according to risk category, i.e., high risk types low risk types. 2 Excludes 120 women tested positive for unclassified HPV types. 3 Odds Ratios (OR) are estimated from polynomial logistic regression, adjusted for age, marital status, and the number of pregnancies, using the normal cytology category as a contrast for the outcome. ASCUS, atypical squamous cells of undetermined significance; LSIL, low-grade squamous intraepithelial lesion; HSIL/cancer, high-grade squamous intraepithelial lesion or cancer. among African women 17 20,26 28 have been limited by their relatively small sample size or use of clinic-based rather than general population samples. We found cervical HPV infection to be relatively common in this population, being detected in 18% of women (13% among women with normal cervical cytology and 76% among those with HSIL/cancer). Interestingly, HPV prevalence was greatest among women aged 55 years or older. Investigators conducting 2 recent population-based studies of women in Costa Rica 10 and Mexico 11 similarly reported an increased HPV prevalence among older women. As in our study, most women aged 55 years or older had not undergone previous routine pap screening. In studies among women receiving routine Pap screening, 7,8,29,30 cervical HPV prevalence is highest among younger women, then declines with increasing age, which most likely reflects the acquisition of HPV near the onset of sexual activity followed by the resolution of infection. The increased prevalence we observe among older women might reflect either higher rates of previous HPV exposure (cohort effect) or reactivation of latent HPV infection resulting from decreased immune surveillance associated with older age or HIV infection. Although we were unable to examine the role of HIV infection, there is no evidence to suggest that HIV infection is differentially associated with specific HPV types.

5 HPV IN RELATION TO RISK OF SIL 807 TABLE IV ASSOCIATION OF CERVICAL CYTOLOGY DIAGNOSIS WITH THE HUMAN PAPILLAMOVIRUS (HPV) TYPES MOST FREQUENTLY DETECTED IN THIS POPULATION (n) Cervical Cytology Diagnosis ASCUS LSIL HSIL/cancer n OR (95% CI) 1 n OR (95% CI) 1 n OR (95% CI) 1 High risk HPV types HPV16 1, ( ) ( ) ( ) HPV18 1, ( ) ( ) ( ) HPV33 1, ( ) ( ) ( ) HPV52 1, ( ) ( ) ( ) HPV58 1, ( ) ( ) ( ) Low risk HPV types HPV53 1, ( ) ( ) ( ) HPV54 1, ( ) ( ) ( ) 1 Odds ratios (OR) adjusted for infection with types other than the one examined are estimated from polynomial logistic regression using the normal cytology category as a contrast for the outcome. ASCUS, atypical squamous cells of undetermined significance; LSIL, low-grade squamous intraepithelial lesion; HSIL/cancer, high-grade squamous intraepithelial lesion or cancer. Moreover, the prevalence of HIV in this population is known to be 1% and as such, is unlikely to have had a significant effect on our estimate of HPV prevalence. The effect of changing hormone levels on HPV detection at the time of menopause is unclear. As was seen in 8 previous studies 7 11,31 33 examining prevalence of various HPV types among large numbers of women presenting for reasons other than cervical neoplasia or known gynecologic problems, we found HPV16 to be the most frequently detected HPV type, followed by HPV58. HPV58 is relatively common among Costa Rican, Japanese and Chinese women, 10,31,32 but uncommon in other female populations. 7,8 HPV58, which was originally isolated from a Japanese woman with invasive cervical cancer, 34 is located on the same branch of the HPV phylogenetic tree as HPV16, 35 and is also related closely to HPV types 31 and In studies of Dutch, Mexican and British women, 8,9,11 the second most frequently detected HPV type is HPV31, whereas among Brazilian women and women in the United States, 7,12 HPV53 is detected most frequently after HPV16. Although ASCUS, LSIL and HSIL were each associated with detection of any HPV (low risk, high risk and unclassified types), the strongest associations were observed between HSIL/cancer and detection of high risk HPV types, as has been seen in previous studies. 10,16,29,37 39 Given the limited number of HSILs/cancers included in previous studies, however, few investigators have been able to adjust for, or adequately assess the effect of, co-infection with other HPV types, or accurately assess the risk of cervical neoplasia associated with individual HPV types. 39 We found that HPV16 and HPV58 were the two HPV types most strongly associated with HSIL/cancer, both as a single infection or in the presence of other HPV types. Interestingly, although HPV16 and HPV58 were frequently detected in HSIL/cancer samples (22% and 13%, respectively), 60% of women with such pathology were positive for other HPV types. Our findings of a high risk of cervical abnormalities associated with HPV58 are consistent with those from a recent study in Mozambique 40 (that found that HPV58 was the second most common HPV type detected among HPV positive women with cervical lesions). Unlike HPV16, the prevalence of HPV58 and its association with HSIL/cancer appear to vary across different populations. It is possible that differing HPV types might predominate in a specific population simply as the result of the chance introduction of a specific virus into that population, or varying abilities of certain HPV types to sustain endemic infections in a particular geographic area. It is also possible that geographic-related sequence variations for a given HPV type might alter its neoplastic potential, resulting in varying risk associations across different populations. It is also possible that HLA polymorphisms present in various populations might affect the risk of HPV persistence and development of cervical neoplasia by altering the immune response to specific HPV-encoded epitopes. 41 Further, as has been reported in some, 31,42,43 but not all previous studies, 10,44 we found that infection with multiple HPV types (relative to a single HPV type) was associated with a somewhat increased risk of HSIL/cancer, although this difference was not significant statistically. Whether this trend reflects a synergistic effect between HPV types on neoplastic progression 43 or whether women with multiple HPV infections have deficient immune responses to HPV 40 is unclear. A recent study 45 found that the MY09 and MY11 primers that we used are less efficient in generating PCR products than are the newly developed PGMY09 and PGMY11 primers. For example, the detection rate of HPV types 26, 35, 42, 45, 52, 54, 55, 59, 66, 73 and 83 is approximately 25% greater when using PGMY09/11 as compared to MY09/11 primers. Thus, it is possible that the high prevalence of HPV58 found in this population is due to differences in primer-related type-specific amplification efficiencies. This is unlikely, however, because the same primers (MY09/11) that we used were also used in the majority of previous studies. Further, even if we assume that the true prevalence of HPV types other than HPV16 and 58 were 25% greater than that observed (to compensate for potential differences in type-specific amplification efficiencies between the two primer systems), HPV58 would remain the second most frequently detected HPV type in our population (data not shown). Our study is unique in that we examined HPV type-specific prevalence in regard to risk of cervical neoplasia in a population

6 808 XI ET AL. that had not been previously screened by cytology. In a screened population, HPV prevalence and the magnitude of associations with cervical neoplasia may be altered as a result of repeated screening and subsequent treatment of lesions. Estimates of HPV prevalence and risk associations in an unscreened population are important to vaccine development efforts in that a vaccine has the potential to prevent a greater number of cancers among women for whom cytology screening is unavailable. Another strength of our study is that HPV testing and cytologic examinations were centralized and carried out without knowledge of specific clinical information for each subject, which minimized the potential for ascertainment bias. A potential limitation is that our HPV prevalence estimates and risk associations could have been affected by the fact that approximately 25% of eligible women were not approached, although none of the invited subjects refused to participate in the study. But, eligible subjects were approached consecutively and we have no reason to think that eligible subjects who were approached significantly differed from those who were not approached. An additional potential limitation is that study participants were recruited from local community-based general medical clinics; although women attend these clinics for a variety of reasons, we do not know the degree to which we can generalize our findings to other populations. Because all subjects were recruited from a consecutive series of patients regardless of the reason for the clinic visit, it is unlikely that either study participation or reason for clinic visit was differentially related to the probability of HPV detection or presence of cervical lesions. In summary, among Senegalese women aged 35 years or older unscreened previously, we found that HPV58 was the second most common HPV type detected (after HPV16) with HPV16 and HPV58 most strongly associated with risk for HSIL/cancer. These data suggest that in addition to HPV16, HPV58 should be considered in the strategic planning of vaccination against cervical cancer in this geographic region. ACKNOWLEDGEMENTS We would like to thank C. Sarr, D. Rich, E. Reay-Ellers, M. Touré, F. Faye Diop and H. Agne for their clinical and field work on this project, the University of Dakar outpatient community health clinics for their support and all of the women who participated in this study. The authors recognize the valuable contributions by A. Starling for data management and J. Kuypers for the laboratory assays. REFERENCES 1. IARC. Monographs on the evaluation of the carcinogenic risks to humans. Human papillomaviruses. vol 64. Lyon: IARC, Walboomers JM, Jacobs MV, Manos MM, Bosch FX, Kummer JA, Shah KV, Snijders PJ, Peto J, Meijer CJ, Munoz N. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 1999;189: Schiffman MH, Bauer HM, Hoover RN, Glass AG, Cadell DM, Rush BB, Scott DR, Sherman ME, Kurman RJ, Wacholder S, Stanton CK, Manos MM. Epidemiologic evidence showing that human papillomavirus infection causes most cervical intraepithelial neoplasia. J Natl Cancer Inst 1993;85: Herrington CS. Human papillomaviruses and cervical neoplasia. I. Classification, virology, pathology and epidemiology. J Clin Pathol 1994;47: Cox JT. Epidemiology of cervical intraepithelial neoplasia: the role of human papillomavirus. Baillieres Clin Obstet Gynaecol 1995;9: Bosch FX, Manos MM, Munoz N, Sherman M, Jansen AM, Peto J, Schiffman MH, Moreno V, Kurman R, Shah KV. Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. International biological study on cervical cancer (IBSCC) Study Group. J Natl Cancer Inst 1995;87: Peyton CL, Gravitt PE, Hunt WC, Hundley RS, Zhao M, Apple RJ, Wheeler CM. Determinants of genital human papillomavirus detection in a US population. I Infect Dis 2001;183: Jacobs MV, Walboomers JM, Snijders PJ, Voorhorst FJ, Verheijen RH, Fransen-Daalmeijer N, Meijer CJ. Distribution of 37 mucosotropic HPV types in women with cytologically normal cervical smears: the age-related patterns for high-risk and low-risk types. Int J Cancer 2000;87: Cuzick J, Beverley E, Ho L, Terry G, Sapper H, Mielzynska I, Lorincz A, Chan WK, Krausz T, Soutter P. HPV testing in primary screening of older women. Br J Cancer 1999;81: Herrero R, Hildesheim A, Bratti C, Sherman ME, Hutchinson M, Morales J, Balmaceda I, Greenberg MD, Alfaro M, Burk RD, Wacholder S, Plummer M, et al. Population-based study of human papillomavirus infection and cervical neoplasia in rural Costa Rica. J Natl Cancer Inst 2000;92: Lazcano-Ponce E, Herrero R, Munoz N, Cruz A, Shah KV, Alonso P, Hernandez P, Salmeron J, Hernandez M. Epidemiology of HPV infection among Mexican women with normal cervical cytology. Int J Cancer 2001;91: Franco EL, Villa LL, Sobrinho JP, Prado JM, Rousseau MC, Desy M, Rohan TE. Epidemiology of acquisition and clearance of cervical human papillomavirus infection in women from a high-risk area for cervical cancer. J Infect Dis 1999;180: Huang S, Afonina I, Miller BA, Beckmann AM. Human papillomavirus types 52 and 58 are prevalent in cervical cancers from Chinese women. Int J Cancer 1997;70: Liaw KL, Hsing AW, Schiffman MH, You SL, Zhang T, Burk R, Chen CJ. Human papillomavirus types 52 and 58 are prevalent in cervical cancer from Chinese women. Int J Cancer 1997;73: Lin QQ, Yu SZ, Qu W, Cruz Y, Burk RD. Human papillomavirus types 52 and 58. Int J Cancer 1998;75: Chan PK, Li WH, Chan MY, Ma WL, Cheung JL, Cheng AF. High prevalence of human papillomavirus type 58 in Chinese women with cervical cancer and precancerous lesions. J Med Virol 1999;59: Chabaud M, Le Cann P, Mayelo V, Leboulleux D, Diallo AS, Enogat N, Afoutou JM, Anthonioz P, Coll-Seck AM, Coursaget P. Detection by PCR of human papillomavirus genotypes in cervical lesions of Senegalese women. J Med Virol 1996;49: Engels H, Nyongo A, Temmerman M, Quint WG, Van Marck E, Eylenbosch WJ. Cervical cancer screening and detection of genital HPV-infection and chlamydial infection by PCR in different groups of Kenyan women. Ann Soc Belg Med Trop 1992;72: ter Meulen J, Eberhardt HC, Luande J, Mgaya HN, Chang-Claude J, Mtiro H, Mhina M, Kashaija P, Ockert S, Yu X, Meinhardt G, Gissmann L, et al. Human papillomavirus (HPV) infection, HIV infection and cervical cancer in Tanzania, East Africa. Int J Cancer 1992;51: Castellsague X, Menendez C, Loscertales MP, Kornegay JR, dos Santos F, Gomez-Olive FX, Lloveras B, Abarca N, Vaz N, Barreto A, Bosch FX, Alonso P. Human papillomavirus genotypes in rural Mozambique. Lancet 2001;358: Parkin DM, Pisani P, Ferlay J. Global cancer statistics. CA Cancer J Clin 1999;49: The 1988 Bethesda system for reporting cervical/vaginal cytological diagnoses. National Cancer Institute Workshop. JAMA 1989;262: Kuypers JM, Critchlow CW, Gravitt PE, Vernon DA, Sayer JB, Manos MM, Kiviat NB. Comparison of dot filter hybridization, Southern transfer hybridization and polymerase chain reaction amplification for diagnosis of anal human papillomavirus infection. J Clin Microbiol 1993;31: Gravitt PE, Peyton CL, Apple RJ, Wheeler CM. Genotyping of 27 human papillomavirus types by using L1 consensus PCR products by a single-hybridization, reverse line blot detection method. J Clin Microbiol 1998;36: Ananth CV, Kleinbaum DG. Regression models for ordinal responses: a review of methods and applications. Int J Epidemiol. 1997;26: Womack SD, Chirenje ZM, Gaffikin L, Blumenthal PD, McGrath JA, Chipato T, Ngwalle S, Munjoma M, Shah KV. HPV-based cervical cancer screening in a population at high risk for HIV infection. Int J Cancer 2000;85: La Ruche G, You B, Mensah-Ado I, Bergeron C, Montcho C, Ramon R, Toure-Coulibaly K, Welffens-Ekra C, Dabis F, Orth G. Human papillomavirus and human immunodeficiency virus infections: relation with cervical dysplasia-neoplasia in African women. Int J Cancer 1998;18: Kreiss JK, Kiviat NB, Plummer FA, Roberts PL, Waiyaki P, Ngugi E, Holmes KK. Human immunodeficiency virus, human papillomavirus and cervical intraepithelial neoplasia in Nairobi prostitutes. Sex Transm Dis 1992;19: de Roda Husman AM, Walboomers JM, Hopman E, Bleker OP, Helmerhorst TM, Rozendaal L, Voorhorst FJ, Meijer CJ. HPV prevalence in cytomorphologically normal cervical scrapes of pregnant

7 HPV IN RELATION TO RISK OF SIL 809 women as determined by PCR: the age-related pattern. J Med Virol 1995;46: Melkert PW, Hopman E, van den Brule AJ, Risse EK, van Diest PJ, Bleker OP, Helmerhorst T, Schipper ME, Meijer CJ, Walboomers JM. Prevalence of HPV in cytomorphologically normal cervical smears, as determined by the polymerase chain reaction, is age dependent. Int J Cancer 1993;53: Sasagawa T, Basha W, Yamazaki H, Inoue M. High-risk and multiple human papillomavirus infections associated with cervical abnormalities in Japanese women. Cancer Epidemiol Biomarker Prev 2001;10: Chan PK, Chang AR, Cheung JL, Chan DP, Xu LY, Tang NL, Cheng AF. Determinants of cervical human papillomavirus infection: differences between high-and low-oncogenic risk types. J Infect Dis 2002; 185: Giuliano AR, Papenfuss M, Abrahamsen M, Denman C, de Zapien JG, Henze JL, Ortega L, Brown de Galaz EM, Stephan J, Feng J, Baldwin S, Garcia F, et al. Human papillomavirus infection at the United States-Mexico border: implications for cervical cancer prevention and control. Cancer Epidemiol Biomarkers Prev 2001;10: Matsukura T, Sugase M. Cloning of a novel human papillomavirus (type58) from an invasive cervical carcinoma. Virology 1990;177: Bernard HU, Chan SY, Manos MM, Ong CK, Villa LL, Delius H, Peyton CL, Bauer HM, Wheeler CM. Identification and assessment of known and novel human papillomaviruses by polymerase chain reaction amplification, restriction fragment length polymorphisms, nucleotide sequence and phylogenetic algorithms. J Infect Dis 1994;170: Kirii Y, Iwamoto S, Matsukura T. Human papillomavirus type 58 DNA sequence. Virology 1991;185: Moscicki AB, Shiboski S, Broering J, Powell K, Clayton L, Jay N, Darragh TM, Brescia R, Kanowitz S, Miller SB, Stone J, Hanson E, et al. The natural history of human papillomavirus infection as measured by repeated DNA testing in adolescent and young women. J Pediatr 1998;132: Liaw KL, Glass AG, Manos MM, Greer CE, Scott DR, Sherman M, Burk RD, Kurman RJ, Wacholder S, Rush BB, Cadell DM, Lawler P, et al. Detection of human papillomavirus DNA in cytologically normal women and subsequent cervical squamous intraepithelial lesions. J Natl Cancer Inst 1999;91: Schiff M, Becker TM, Masuk M, van Asselt-King L, Wheeler CM, Altobelli KK, North CQ, Nahmias AJ. Risk factors for cervical intraepithelial neoplasia in Southwestern American Indian women. Am J Epidemiol 2000;152: Castellsague X, Menendez C, Loscertales MP, Kornegay JR, dos Santos F, Gomez-Olive FX, Lloveras B, Abarca N, Vaz N, Barreto A, Bosch FX, Alonso P. Human papillomavirus genotypes in rural Mozambique. Lancet 2001;358: Apple RJ, Erlich HA, Klitz W, Manos MM, Becker TM, Wheeler CM. HLA DR-DQ associations with cervical carcinoma show papillomavirus-type specificity. Nat Genet 1994;6: Ho GY, Bierman R, Beardsley L, Chang CJ, Burk RD. Natural history of cervicovaginal papillomavirus infection in young women. N Engl J Med 1998;12: Liaw KL, Hsing AW, Chen CJ, Schiffman MH, Zhang TY, Hsieh CY, Greer CE, You SL, Huang TW, Wu TC, O Leary TJ, Seidman JD, et al. Human papillomavirus and cervical neoplasia: a case-control study in Taiwan. Int J Cancer 1995;4: Ho GY, Kadish AS, Burk RD, Basu J, Palan PR, Mikhail M, Romney SL. HPV 16 and cigarette smoking as risk factors for high-grade cervical intra-epithelial neoplasia. Int J Cancer 1998;29: Gravitt PE, Peyton CL, Alessi TQ, Wheeler CM, Coutlee F, Hildesheim A, Schiffman MH, Scott DR, Apple RJ. Improved amplification of genital human papillomaviruses. J Clin Microbiol 2000; 38: