Human Papillomavirus Genotypes and the Cumulative 2-Year Risk of Cervical Precancer

Transcription

1 MAJOR ARTICLE Human Papillomavirus Genotypes and the Cumulative 2-Year Risk of Cervical Precancer Cosette M. Wheeler, 1,2 William C. Hunt, 1 Mark Schiffman, 3 and Philip E. Castle, 3 for the Atypical Squamous Cells of Undetermined Significance/Low-Grade Squamous Intraepithelial Lesions Triage Study (ALTS) Group a 1 Department of Molecular Genetics and Microbiology and 2 Department of Obstetrics and Gynecology, University of New Mexico Health Sciences Center, School of Medicine, Albuquerque; 3 Division of Cancer Epidemiology and Genetics and National Cancer Institute, National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland Background. Prospective data on the risks of cervical precancer associated with specifi human papillomavirus (HPV) genotypes are limited. Methods. In 5060 women participating in the Atypical Squamous Cells of Undetermined Significance/Low Grade Squamous Intraepithelial Lesions Triage Study (ALTS), we determined the cumulative 2-year risks of cervical intraepithelial neoplasia (CIN) grade 2 or more severe ( CIN2) and of grade 3 or more severe ( CIN3) for 38 individual HPV genotypes, as detected by polymerase chain reaction. Results. The most common HPV genotypes detected at baseline, in descending order of prevalence, were 16, 52, 51, 31, 18, 53, 39, 56, 62, 59, and 58. When detected as a single-type HPV infection, HPV-16 had a 2-year cumulative risk of 50.6% (95% confidenc interval [CI], 44.1% 57.2%) for CIN2 and 39.1% (95% CI, 32.9% 45.7%) for CIN3. For other singly detected carcinogenic HPV types, the risk of CIN2 ranged from 4.7% (for HPV-59) to 29.5% (for HPV-31), and the risk of CIN3 ranged from 0.0% (for HPV-59) to 14.8% (for HPV- 31). Multiple infections with HPV genotypes of different risk classes resulted in a risk that was similar to, and not significantl different from, the risk observed for the HPV genotype of the highest risk class. Conclusions. Genotype-specifi HPV testing may be useful for identifying women with atypical squamous cells of undetermined significanc and low-grade squamous intraepithelial lesions who are at higher and lower risk of prevalent and incipient cervical precancer. Cervical cancer and precancer are caused by persistent infection with human papillomavirus (HPV) [1 4]. More than 40 genotypes (hereafter referred to as types ) of HPV can infect the cervix. Of these, 15 types belonging to 4 related species in the genus a-papillomavirus [5] can cause cancer, but the strength of carcinogenicity of individual types varies greatly [6 9]. For example, HPV-16 (of species a-9) is clearly a Received 29 March 2006; accepted 11 May 2006; electronically published 27 September Potential conflicts of interest: C.M.W., W.C.H., M.S., and P.E.C. report no conflicts of interest. Financial support: National Cancer Institute, National Institutes of Health, Department of Health and Human Services (contracts CN-55153, CN-55154, CN , CN-55156, CN-55157, CN-55158, CN-55159, and CN-55105). a ALTS group members are listed after the text, along with information on potential conflicts of interest for these members. Reprints or correspondence: Dr. Cosette M. Wheeler, Dept. of Molecular Genetics and Microbiology, UNM HOPE Clinic, 1816 Sigma Chi Rd. NE, Albuquerque, NM (cwheeler@salud.unm.edu). The Journal of Infectious Diseases 2006; 194: by the Infectious Diseases Society of America. All rights reserved /2006/ $15.00 much stronger viral carcinogen than any other type [10 12]. HPV-16 persists longer than most other HPV types [4] and, therefore, has higher prevalence. When it persists for 11 or 2 years, HPV-16 is more likely to cause cervical precancer and cancer than the other potentially carcinogenic types [4]. Globally, HPV-16 causes approximately one-half of cervical cancer cases [6 9]. It has been difficul to assess the carcinogenicity of HPV types other than HPV-16. The other types are less common among the general population and among women with cervical cancer or precancer outcomes, thus limiting the precision of risk estimates. Determining risk estimates for precancer and cancer when multiple HPV infections are detected is particularly difficul when HPV-16 is one of the coinfecting types and, therefore, dominates the risk. Addressing these concerns requires a very large study population of infected women with complete typing for the full range of HPV types and a large number of rigorously define disease outcomes. No currently established study population is big enough to evaluate HPV Genotypes and Cervical Precancer JID 2006:194 (1 November) 1291

2 precisely the risk posed by the least common and weakest of the carcinogenic HPV types. However, in one of the biggest efforts to date, we present data here from the Atypical Squamous Cells of Undetermined Significanc (ASCUS)/Low-Grade Squamous Intraepithelial Lesions (LSIL) Triage Study (ALTS) [13], in which we have characterized the 2-year cumulative risk of cervical intraepithelial neoplasia (CIN) grade 2 or more severe ( CIN2) and of grade 3 or more severe ( CIN3) for 14 individual carcinogenic HPV types and 24 noncarcinogenic or uncharacterized HPV types. SUBJECTS, MATERIALS, AND METHODS Study design and population. ALTS was a randomized trial comparing the following 3 management strategies for 5060 women with ASCUS ( n p 3488) or LSIL ( n p 1572) [13]: (1) immediate colposcopy (referral to colposcopy regardless of enrollment test results); (2) HPV triage (referral to colposcopy if the enrollment HPV testing result was either positive by Hybrid Capture 2 [Digene Corporation, Gaithersburg, MD] or missing or if the enrollment cytological diagnosis was high-grade squamous intraepithelial lesion [HSIL]); or (3) conservative management (referral to colposcopy at enrollment if the cytological diagnosis was HSIL). At enrollment, all women received a pelvic examination with collection of 2 cervical specimens; the firs specimen was collected in PreservCyt for ThinPrep cytological analysis (Cytyc Corporation, Boxborough, MA), and the second specimen was collected in specimen transport medium (STM; Digene). Women in all 3 arms of the study were reevaluated by cytological analysis every 6 months for 2 years of follow-up and were sent for a colposcopy if the cytological diagnosis was HSIL. An exit examination with colposcopy was scheduled for all women, regardless of study arm or prior procedures, at the completion of the follow-up. We refer readers elsewhere for details on randomization, examination procedures, patient management, and laboratory and pathological methods [13 17]. The National Cancer Institute and local institutional review boards approved the study, and all participants provided written, informed consent. HPV DNA testing. HPV typing was performed using an L1-based polymerase chain reaction (PCR) assay that employs a primer set designated PGMY09/11 and was performed on the STM specimen [18]. Amplimers were subjected to reverse-line blot hybridization for detection of 27 individual HPV types (6, 11, 16, 18, 26, 31, 33, 35, 39, 40, 42, 45, 51 59, 66, 68, 73 [PAP238a], 82 [W13b], 83 [PAP291], and 84 [PAP155]) [19]. We also tested for an additional 11 HPV types (61, 62, 64, 67, 69 72, 81, 82v [IS39], and 89 [CP6108]) in approximately onehalf of the specimens (58%) at enrollment and in all specimens collected at the follow-up visits [20]. Pathological analysis and treatment. Clinical management was based on clinical center pathologists cytologic and histologic diagnoses. In addition, all referral smears, ThinPreps, and histology slides were sent to the Pathology Quality Control Group (PQCG), based at the Johns Hopkins Hospital, for review and secondary diagnoses. The outcomes of interest for these analyses were CIN2 (which was based on the clinical center diagnosis, because it triggered treatment by loop electrical excision procedure) and CIN3 (which was based on the PQCG diagnosis) as the preferred scientifi surrogate for cancer risk. Statistical methods. Of the 5060 women enrolled in ALTS, 4915 (97.1%) had successful HPV testing of 27 HPV types at entry into the study. Ten women received 2 HPV PCR tests at study entry, and the results from both tests were combined for these analyses, such that types detected by either assay were included. A subset of 2833 women had successful testing for an additional 11 HPV types. For purposes of these analyses, we considered HPV types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, and 68 as primary carcinogenic types and HPV types 6, 11, 26, 40, 42, 53, 54, 55, 57, 61, 62, 64, 67, 69 73, 81, 82, 82v, 83, 84, and 89 (CP6108) as noncarcinogenic types. Women were assigned to an HPV risk group according to a priori established cervical cancer risk: positive for HPV-16, else positive for any carcinogenic HPV type and negative for HPV-16 (hereafter, carcinogenic types without HPV-16 ), else positive for any noncarcinogenic HPV type and negative for all carcinogenic types, else PCR negative. Cumulative 2-year risks for CIN2 and CIN3 were computed by considering any diagnosis made at entry, during the 2 years of follow-up, or at exit from the study. Of the 4915 women, only the 3944 (80.2%) who completed the 24-month follow-up visit were included in these analyses. We computed HPV type specifi risks of CIN2 and CIN3 for women who tested positive at entry for the specifi type only. We also computed risks for classes of HPV on the basis of their carcinogenic potential as described above. In further analyses, we divided the study period into 2 intervals: (1) enrollment visit up to the 12-month visit (hereafter, year 1 period ) and (2) 12-month visit through the 24-month visit or exit from the study (hereafter, year 2 period ). The HPV risk status for each interval was based on the presence of the HPV type in the highest HPV risk group during that interval. We then computed the risks of CIN2 and CIN3 during the year 2 period among women who had not received a diagnosis of or treatment for CIN2 during the year 1 period. We restricted these analyses to the women in the immediate colposcopy and HPV triage study arms, because of the insensitive diagnosis of CIN2 during the year 1 period in the conservative management arm [21]. Women without HPV testing for both time periods and without an exit colposcopy were excluded. The total number of women included in these analyses was Confidenc intervals (CIs) were computed using exact pro JID 2006:194 (1 November) Wheeler et al.

3 cedures, and comparisons of estimates of risk were tested for significanc using the Pearson x 2 statistic or the Cochran-Armitage test of trend, with an exact P value. SAS (version 9.1; SAS Institute) was used for all analyses. RESULTS The type-specifi HPV prevalence at study entry for this population of women with ASCUS and LSIL is shown in table 1. Overall, 68.4% of the women ( n p 3362) tested positive for at least 1 type of HPV. Of those women who were HPV positive, 43.7% were positive for only a single type, 28.0% were positive for 2 types, 15.4% were positive for 3 types, and 12.9% were positive for 4 12 types. For any given type, 20% occurred as a single-type HPV infection. The most common HPV types were, in descending order of prevalence, 16 (16.8%), 52 (9.4%), 51 (8.1%), 31 (7.1%), and 18 (6.6%). Table 1 also presents the cumulative 2-year risk for a clinical center diagnosis of CIN2 and a PQCG diagnosis of CIN3 by type of HPV detected at study entry for women with singletype HPV infections. The risk of CIN2 or CIN3 associated with HPV-16 was greater than that observed for any other HPV type. Among women who were infected only with HPV-16, the cumulative 2-year risk of CIN2 was 50.6% (95% CI, 44.1% 57.2%), and the risk of CIN3 was 39.1% (95% CI, 32.9% 45.7%). Other known carcinogenic types had risks of CIN2 and CIN3 that were lower than those for HPV-16 but were generally higher than those for noncarcinogenic or uncharacterized types. Among women with single-type infections of a carcinogenic type other than 16 (18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, and 68), the 2-year cumulative risk of CIN2 ranged from 4.7% (for HPV-59) to 29.5% (for HPV-31). Overall, the risk of CIN2 among women with a single carcinogenic type other than 16 was 18.8%. Women who were initially negative for HPV by PCR had a risk of 2.4% for CIN3. Of these 38 PCR-negative women with CIN3, 19 tested positive by use of a second cervical specimen and HPV test for carcinogenic HPV (Hybrid Capture 2; Digene Corporation). The risk of CIN3 for the remaining 19 PCR-negative, Hybrid Capture 2 negative women was 1.4%. The 2-year cumulative risk of CIN2 and CIN3 was greater for the HPV risk groups more strongly associated with cancer (table 2). For example, the 2-year risk of CIN2 for single-type HPV infections was 8.2% for noncarcinogenic types, 18.8% for carcinogenic types without HPV-16, and (as shown also in table 1) 50.6% for HPV-16. The 2-year risk of CIN3 for single-type HPV infections was 3.9% for noncarcinogenic types, 7.9% for carcinogenic types without HPV-16, and 39.1% for HPV-16. All differences in risk were statistically significan ( P!.01). Women whose infections were categorized as multiple carcinogenic types without HPV-16 had a nonsignificantl greater risk of CIN3 than did women whose infections were categorized as single carcinogenic types without HPV-16 (10.9% vs. 7.9%). However, the converse was true for infections with noncarcinogenic HPV types (1.4% vs. 3.9%). Except for HPV-16, infections with multiple HPV types of different risk groups resulted in risks of CIN2 and CIN3 that were close to the risk observed for the higher of the risk types. For example, women infected with a single noncarcinogenic type and a single carcinogenic type without HPV-16 had a 2-year cumulative risk of CIN2 of 17.3% and a 2-year cumulative risk of CIN3 of 8.3%, which were similar to the risks of CIN2 (18.8%) and CIN3 (7.9%) for women infected with a single carcinogenic type without HPV-16. By contrast, there was an apparent trend toward lower risk in women infected with HPV-16 in combination with single or multiple noncarcinogenic types. The 2-year risk of CIN2 decreased from 50.6% to 40.0% to 36.1% among women infected with HPV-16 only, with 1 additional noncarcinogenic type, or 11 additional noncarcinogenic type, respectively. A test of trend yielded P p.04. A similar pattern of decreasing risk (from 39.1% to 28.9% to 27.8%) was also observed using an end point of CIN3, with a test of trend yielding P p.07. We observed that as the overall risk of CIN2 decreased, the ratio of CIN2 to CIN3 increased. HPV-16 infections conferred the greatest risk of CIN2 and the lowest ratio of CIN2 to CIN3 (i.e., most HPV-16 positive women with CIN2 had CIN3). For single-type HPV-16 infections, the ratio was By contrast, women infected with noncarcinogenic HPV types had ratios of 1 or greater. Women infected with multiple noncarcinogenic types had the highest ratio (6.0). In logistic regression modeling (data not shown), age (!30 or 30 years) and referral cytological diagnosis (LSIL or ASCUS) were not confounding variables of the risks of either CIN2 or CIN3 associated with HPV. Having LSIL (vs. ASCUS) was significantl associated with having CIN2 or CIN3 (odds ratios [ORs], 1.38 and 1.40, respectively). Women 30 years of age had a significantl lower risk of CIN2 but not of CIN3 (ORs, 0.74 and 0.81, respectively), compared with women!30 years of age. In table 3, we present the risk of CIN2 diagnosed during the last 12 months of the study associated with the HPV risk status at 2 time intervals: 0 to!12 months (the year 1 period) and 12 to 24 months (the year 2 period). Women who were reclassifie in the year 2 period into a risk group that was lower or higher than the year 1 period had a decreased or an increased risk of CIN2, respectively, compared with those who were not reclassified For example, women who were HPV negative during both time periods had a risk of 0.6% (95% CI, 0.1% 1.9%), whereas women who were negative during the year 1 period but HPV-16 positive during the year 2 period had a risk of 14.3% (95% CI, 3.0% 36.3%). Among women who were HPV Genotypes and Cervical Precancer JID 2006:194 (1 November) 1293

4 Table 1. Prevalence of human papillomavirus (HPV) at study entry and 2-year risk of cervical intraepithelial neoplasia (CIN) grade 2 or more severe ( CIN2; clinical center diagnosis) and of grade 3 or more severe ( CIN3; Pathology Quality Control Group diagnosis), among polymerase chain reaction (PCR) negative women and women with single-type HPV infection. HPV genotype detected Tested, no. Single or multiple HPV infection a No. (%) Risk % (95% CI) Single HPV infection only b CIN2 CIN3 PCR negative (31.6) 4.3 ( ) 2.4 ( ) (3.2) 31 (19.6) 9.7 ( ) 3.2 ( ) (0.9) 3 (7.1) 0.0 ( ) 0.0 ( ) (16.8) 235 (28.4) 50.6 ( ) 39.1 ( ) (6.6) 82 (25.2) 18.3 ( ) 6.1 ( ) (1.0) 5 (10.4) 40.0 ( ) 20.0 ( ) (7.1) 88 (25.4) 29.5 ( ) 14.8 ( ) (3.3) 43 (26.9) 23.3 ( ) 14.0 ( ) (4.8) 64 (26.9) 23.4 ( ) 7.8 ( ) (5.9) 47 (16.2) 17.0 ( ) 8.5 ( ) (1.9) 13 (14.1) 0.0 ( ) 0.0 ( ) (3.6) 37 (20.7) 13.5 ( ) 8.1 ( ) (4.5) 41 (18.6) 24.4 ( ) 9.8 ( ) (8.1) 89 (22.3) 16.9 ( ) 5.6 ( ) (9.4) 124 (26.8) 18.5 ( ) 7.3 ( ) (6.1) 59 (19.8) 8.5 ( ) 3.4 ( ) (4.3) 30 (14.2) 3.3 ( ) 3.3 ( ) (2.4) 17 (14.2) 0.0 ( ) 0.0 ( ) (5.9) 54 (18.7) 5.6 ( ) 1.9 ( ) (0.0) 2 (100.0) 0.0 ( ) 0.0 ( ) (5.5) 67 (24.8) 26.9 ( ) 13.4 ( ) (5.6) 43 (15.6) 4.7 ( ) 0.0 ( ) (4.9) 28 (20.0) 10.7 ( ) 3.6 ( ) (5.7) 23 (14.2) 0.0 ( ) 0.0 ( ) (0.3) 0 (0) (4.9) 52 (21.4) 13.5 ( ) 3.8 ( ) (2.2) 15 (23.8) 20.0 ( ) 13.3 ( ) (3.3) 26 (16.1) 7.7 ( ) 7.7 ( ) (0.4) 1 (9.1) 0.0 ( ) 0.0 ( ) (3.6) 25 (24.3) 4.0 ( ) 0.0 ( ) (0.5) 3 (23.1) 0.0 ( ) 0.0 ( ) (1.1) 4 (12.9) 0.0 ( ) 0.0 ( ) (2.5) 23 (18.7) 0.0 ( ) 0.0 ( ) (2.3) 5 (7.7) 20.0 ( ) 40.0 ( ) (2.3) 18 (16.1) 27.8 ( ) 11.1 ( ) (3.6) 24 (13.6) 0.0 ( ) 0.0 ( ) (3.4) 27 (16.3) 7.4 ( ) 3.7 ( ) 82v (0.4) 2 (18.2) 50.0 ( ) 0.0 ( ) (5.3) 20 (13.3) 10.0 ( ) 0.0 ( ) NOTE. CI, confidence interval. a Percentages are those of women infected with the individual genotype shown. b Percentages are those of women infected only with the individual genotype among all women infected with that type. 1294

5 Table 2. Two-year risk of cervical intraepithelial neoplasia (CIN) grade 2 or more severe ( CIN2; clinical center diagnosis) and of grade 3 or more severe ( CIN3; Pathology Quality Control Group diagnosis), by human papillomavirus (HPV) infection category at study entry. CIN2 CIN3 HPV at study entry No. Risk % (95% CI) P Risk % (95% CI) P Pearson x 2 test reference group CIN2/ CIN3 a PCR negative ( ) 2.4 ( ) 0.63 Any noncarcinogenic ( )! ( ).3 PCR negative 1.5 Single noncarcinogenic ( ) ( ).1 PCR negative 0.94 Multiple noncarcinogenic ( ) ( ).2 Single noncarcinogenic 6.0 Any carcinogenic without HPV ( )! ( )!.0001 Any noncarcinogenic 1.3 Single carcinogenic without HPV ( )! ( ).007 Single noncarcinogenic 1.5 Multiple carcinogenic without HPV ( ) ( ).1 Single carcinogenic without HPV Single noncarcinogenic + single carcinogenic without HPV ( ) ( ).9 Single carcinogenic without HPV Multiple noncarcinogenic and/or carcinogenic without HPV ( ) ( ).2 Single carcinogenic without HPV Any HPV ( )! ( )!.0001 Any carcinogenic without HPV HPV-16 only ( )! ( )!.0001 Single carcinogenic without HPV HPV-16 + single noncarcinogenic ( ) ( ).09 HPV-16 only 0.46 HPV-16 + multiple noncarcinogenic ( ) ( ).2 HPV-16 only 0.10 HPV-16 + single carcinogenic without HPV ( ) ( ).9 HPV-16 only 0.31 HPV-16 + multiple carcinogenic without HPV ( ) ( ).06 HPV-16 only 0.61 HPV-16 + single noncarcinogenic + single carcinogenic without HPV ( ) ( ).3 HPV-16 only 0.52 HPV-16 + multiple noncarcinogenic and/or carcinogenic without HPV ( ) ( ).7 HPV-16 only 0.42 NOTE. The noncarcinogenic category includes all noncarcinogenic HPV genotypes (6, 11, 26, 40, 42, 53, 54, 55, 57, 61, 62, 64, 67, 69 73, 81, 82, 82v, 83, 84, and 89 [CP6108]), and the carcinogenic without HPV-16 category includes all carcinogenic HPV genotypes excluding HPV-16 (18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, and 68). CI, confidence interval; PCR, polymerase chain reaction. a Ratio of CIN2 cases to CIN3 cases, both as diagnosed by the Pathology Quality Control Group. 1295

6 Table 3. Risk of cervical intraepithelial neoplasia (CIN) grade 2 or more severe ( CIN2; clinical center diagnosis) and of grade 3 or more severe ( CIN3; Pathology Quality Control Group diagnosis) in women with!cin2 during the year 1 period, by human papillomavirus (HPV) risk status during the year 1 period and HPV risk status during the year 2 period. CIN2 CIN3 HPV status, year 1 period HPV status, year 2 period Women, no. Received diagnosis, no. Risk % (95% CI) Received diagnosis, no. Risk % (95% CI) PCR negative PCR negative ( ) ( ) Noncarcinogenic ( ) ( ) Carcinogenic without HPV ( ) ( ) HPV ( ) ( ) Noncarcinogenic PCR negative ( ) ( ) Noncarcinogenic different type(s) ( ) ( ) Noncarcinogenic same type(s) ( ) ( ) Carcinogenic without HPV ( ) ( ) HPV ( ) ( ) Carcinogenic without HPV-16 PCR negative ( ) ( ) Noncarcinogenic ( ) ( ) Carcinogenic without HPV ( ) ( ) different type(s) Carcinogenic without HPV ( ) ( ) same type(s) HPV ( ) ( ) HPV-16 PCR negative ( ) ( ) Noncarcinogenic ( ) ( ) Carcinogenic without HPV ( ) ( ) HPV ( ) ( ) NOTE. Women were assigned an HPV risk status according to the highest-risk HPV genotype detected during each time period. The noncarcinogenic category includes all noncarcinogenic HPV genotypes (6, 11, 26, 40, 42, 53, 54, 55, 57, 61, 62, 64, 67, 69 73, 81, 82, 82v, 83, 84, and 89 [CP6108]), and the carcinogenic without HPV-16 category includes all carcinogenic HPV genotypes excluding HPV-16 (18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, and 68). CI, confidence interval; PCR, polymerase chain reaction. HPV-16 positive during the year 1 period and remained HPV- 16 positive during the year 2 period, the risk of CIN2 was 27.4% (95% CI, 19.5% 36.6%), whereas those who were HPV negative during the year 2 period had a risk of only 4.7% (95% CI, 0.6% 15.8%). Table 3 also shows the corresponding risk estimates for CIN3. Although the numbers were small for many categories and the resulting risk estimates are imprecise, the pattern of risk was similar to that observed for CIN2. The average duration between the diagnosis of CIN and the last PCR HPV test was 92 days for CIN2 and 73 days for CIN3. DISCUSSION We have previously reported that, in this population of mostly young women with either equivocal or mild cervical cytological abnormalities, detection of HPV-16 at entry into the study was associated with a very high risk of CIN3 over a 2-year period [11]. Here, we report the 2-year cumulative risk of CIN2 and CIN3 for individual HPV types and various groupings of HPV types. We found that the recognized carcinogenic HPV types other than 16 (18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, and 68) had a collective risk of CIN3 that was approximately one-fift of that for HPV-16 (7.9% vs. 39.1%). The remaining HPV types had a collective risk of CIN3 that was approximately one-tenth that of HPV-16 (3.9%). Although the risks for the individual HPV types varied greatly, the limited precision of the estimates did not allow us to further distinguish the HPV types within these broad groupings. Women in ALTS were referred for equivocal or mild cytologic abnormalities. Of note, the risk of subsequent CIN3 would be lower in the general population, particularly among those with normal cytological f ndings. The low but nonzero subsequent risk of CIN3 among those who tested negative for HPV at enrollment shows that HPV testing is not 100% sensitive. Nonetheless, recent screening guidelines have included repeat cytological and HPV testing every 3 years. It is expected that multiple negative HPV tests would defin an extremely low-risk group. A recent report found a 10-year elevated risk of cervical precancer and cancer for HPV-18 that was similar in magnitude 1296 JID 2006:194 (1 November) Wheeler et al.

7 to that observed for HPV-16 [12]. The lower risk among HPV- 18 positive women in the present study might be explained by the short duration of follow-up. Among the non HPV-16 carcinogenic types, we observed the greatest risk of CIN3 (14.8%) for HPV-31. In this study, perhaps by chance, relatively low risks of CIN3 for HPV-56 and -59 were observed, with risks similar to those observed for the noncarcinogenic types. Noncarcinogenic types were associated with increased risk of CIN2, compared with the risk in women who were PCR negative. However, the risk of CIN3 for noncarcinogenic types did not differ significantl from the risk of CIN3 in women who were PCR negative. Thus, noncarcinogenic HPV types seem to cause CIN2 but not CIN3, which is consistent with the rare detection of noncarcinogenic HPV types in cohort studies of longer duration and in studies of invasive cervical cancer [4, 6 8, 10]. The occasional findin of CIN3 in women who were HPV negative might be due to false-negative results for carcinogenic HPV types or to histologic overcall. We also observed relatively high risks for HPV-26, -82, -67, and -42; these risks were of comparable magnitude to those observed for the carcinogenic types other than HPV-16. HPV- 26 and -82 have been suggested as potential carcinogenic types [7], and, thus, our finding of relatively high risks for these types might have been expected. HPV-67, which has not been identifie in other studies as carcinogenic, resides, interestingly, in the same phylogenetic species (a-9) as the carcinogenic HPV types 16, 31, 33, 35, 52, and 58. It is unclear whether HPV-67 is truly carcinogenic and was missed by the large case series [7] or whether it causes CIN3 but not cancer. The findin of an association between the risk of cervical precancer and HPV- 42 is surprising but warrants caution, given the wide CIs for these estimated risks and possible false-negative results for coinfection with carcinogenic types. HPV-42 was not found in cancers in large surveys [7]. In the present investigation, we also attempted to estimate the risks of combinations of HPV types. However, the low frequency of multiple infections of specifi HPV types made this impossible for all but a few combinations. Rather than look at combinations of specifi HPV types, we chose to look at combinations of types within broad risk classes. A few patterns emerged from these analyses. First, the risk classes had distinct and significantl different risk levels. Among women with a single infection, those infected with HPV-16 had the highest risk of CIN3 (39.1%), followed by those infected with a carcinogenic type without HPV-16 (7.9%) and by those infected with a noncarcinogenic HPV type (3.9%). Previous reports found that multiple infections with non HPV-16 carcinogenic types significantl increased the risk of CIN3, compared with infection with single-type non HPV- 16 carcinogenic infections [22, 23]. In the present study, we observed a similar yet nonsignifican increase in risk, compared with the risk observed for single infections of that risk class. For example, women infected with a single carcinogenic type without HPV-16 had a risk of 7.9% for CIN3, compared with a risk of 10.9% for women infected with 2 or more carcinogenic types without HPV-16. The likely explanation for this increase in risk is that the non HPV-16 carcinogenic types are not homogenous with respect to risk. If the risk of CIN3 in women with multiple infections is set by the HPV type with the greatest risk, then women with multiple infections will, on average, have a greater risk of CIN3 than will women with single infections. A third pattern observed was that multiple infections with HPV types of different risk classes resulted in a risk that was similar to, and not significantl different from, the risk observed for the highest risk class. Women infected with a single noncarcinogenic type and a single carcinogenic type without HPV- 16 had an 8.3% risk for CIN3, which was essentially the same risk observed in women infected with a single carcinogenic type without HPV-16. Similarly, women infected with HPV-16 and either a single noncarcinogenic type or an additional single carcinogenic type had risks that were close to that observed for HPV-16 alone (29%, 39%, and 39%, respectively). There was one interesting pattern that deviated from the above generalizations. A decreasing trend in risk was observed in the risk estimates among women infected with HPV-16 and either no, one, or multiple noncarcinogenic types. This may result from consistent antagonism between types. Previous work has suggested that antagonism between the noncarcinogenic HPV types 6 or 11 and HPV-16 results in a reduced risk for both CIN and invasive cervical cancer [24, 25]. We are currently exploring type-type interactions by use of Markov chain modeling. Another possibility is that this trend may reflec residual age confounding, given that younger HPV-16 positive women were significantl more likely to have additional carcinogenic HPV types detected than were older women ( P p.0003), who on average would be expected to have had their infections longer and, therefore, to be more likely to have developed precancer or cancer. Another objective of these analyses was to investigate how a change in HPV risk status over the course of follow-up could change the risk of CIN2 and CIN3. By dividing the study period into 2 intervals (0 to!12 months [the year 1 period] and 12 to 24 months [the year 2 period]) and characterizing a woman s HPV risk status for each interval, we were able to show that a change in HPV risk status between the 2 relatively short periods was associated with a concordant change in the risk of CIN2 and CIN3. For women who had a change in HPV risk level, the results of these analyses show that the HPV risk status during the second half of the study was much more closely related to the diagnosis of CIN2 than was the HPV risk status 1 year earlier. Persistence of HPV infection did, however, have an effect on HPV Genotypes and Cervical Precancer JID 2006:194 (1 November) 1297

8 risk, as the highest risk (27.4%) was observed for women who were HPV-16 positive for both years of the study. A lower, but still increased, risk was observed when a non HPV-16 carcinogenic type persisted. Finally, it should be stressed that the ALTS population consisted of women who were referred to 1 of 4 clinical centers in the United States with a cytologic diagnosis of ASCUS or LSIL from a community laboratory. The type-specifi prevalence of HPV and the incidence of CIN2 and CIN3 that we observed during the 2-year follow-up may not be generalizable to other populations with different characteristics. If the present results are confi med, we will need to determine the clinical uses of HPV type specifi testing. Given the limitations in sensitivity of colposcopy [21], in some circumstances, persistent infection with a carcinogenic HPV type might serve as an adjunct to colposcopy and defin treatment in the absence of obvious cervical precancer. Such a change would require the development of reliable, quality-controlled HPV typing kits for general use and careful avoidance of overuse. ALTS GROUP MEMBERS National Cancer Institute, Bethesda, MD. D. Solomon, project officer M. Schiffman, co project officer S. Wacholder, statistician; P. Castle. University of Alabama at Birmingham. E. E. Partridge, principal investigator; L. Kilgore, co principal investigator; S. Hester, study manager. University of Oklahoma, Oklahoma City. J. L. Walker, principal investigator; G. A. Johnson, co principal investigator; A. Yadack, study manager. Magee Womens Hospital of the University of Pittsburgh Medical Center Health System, Pittsburgh, PA. R. S. Guido, principal investigator; K. McIntyre-Seltman, co principal investigator; R. P. Edwards, investigator; J. Gruss, study manager. University of Washington, Seattle. N. B. Kiviat, co principal investigator; L. Koutsky, co principal investigator; C. Mao, investigator. Colposcopy Quality Control Group. D. Ferris, principal investigator (Medical College of Georgia, Augusta, GA); J. T. Cox, coinvestigator (University of California at Santa Barbara, Santa Barbara); L. Burke, coinvestigator (Beth Israel Deaconess Medical Center Hospital, Boston, MA). HPV Quality Control Group. C. M. Wheeler, principal investigator (University of New Mexico Health Sciences Center, Albuquerque); C. Peyton-Goodall, laboratory manager (University of New Mexico Health Sciences Center, Albuquerque); M. M. Manos, coinvestigator (Kaiser Permanente, Oakland, CA). Pathology Quality Control Group. R. J. Kurman, principal investigator (Johns Hopkins Hospital, Baltimore, MD); D. L. Rosenthal, coinvestigator (Johns Hopkins Hospital, Baltimore, MD); M. E. Sherman, coinvestigator (National Cancer Institute, Rockville, MD); M. H. Stoler, coinvestigator (University of Virginia Health Science Center, Charlottesville). Westat, Coordinating Unit, Rockville, MD. J. Rosenthal, project director; M. Dunn, data management team leader; J. Quarantillo, senior systems analyst; D. Robinson, clinical center coordinator. Quality of Life Group. Diane Harper (Dartmouth Medical School, Hanover, NH); A. T. Lorincz, senior scientifi office (Digene Corporation, Gaithersburg, MD); B. Kramer, senior programmer/analyst (Information Management Services, Silver Spring, MD). Potential conflict of interest relating to these ALTS Group members are as follows: A. T. Lorincz is the senior vice president of research and development and chief scientifi office of Digene Corporation, the maker of the Hybrid Capture 2 test, and owns stock in the company; J. T. Cox is on the speakers bureau and is a consultant for Digene Corporation; M. E. Sherman has previously received research funding from Digene Corporation; and M. H. Stoler has consulted for Digene Corporation, Roche Molecular Systems, and Cytyc Corporation. Acknowledgments We thank Digene Corporation, Cytyc Corporation, National Testing Laboratories, Denvu, TriPath Imaging, and Roche Molecular Systems, for donating or providing at reduced cost some of the equipment and supplies used in this study. References 1. Hildesheim A, Schiffman M, Gravitt P, et al. Persistence of type-specifi human papillomavirus infection among cytologically normal women. J Infect Dis 1994; 169: Ho GYF, Bierman R, Beardsley L, Chang CJ, Burk RD. Natural history of cervicovaginal papillomavirus infection in young women. N Engl J Med 1998; 338: Schlect NF, Kulaga S, Robitaille J, et al. Persistent human papillomavirus infection as a predictor of cervical intraepithelial neoplasia. JAMA 2001; 286: Schiffman M, Herrero R, Desalle R, et al. The carcinogenicity of human papillomavirus types reflect viral evolution. Virology 2005; 337: de Villiers EM, Fauquet C, Broker TR, Bernard HU, zur Hausen H. Classificatio of papillomaviruses. Virology 2004; 324: Bosch FX, Manos MM, Munoz N, Sherman M, Jansen AM, Peto J. 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: Munoz N, Bosch FX, de Sanjose S, et al. Epidemiologic classificatio of human papillomavirus types associated with cervical cancer. N Engl J Med 2003; 348: Clifford GM, Smith JS, Plummer M, Munoz N, Franceschi S. Human papillomavirus types in invasive cervical cancer worldwide: a metaanalysis. Br J Cancer 2003; 88: IARC monographs on the evaluation of carcinogenic risks to humans. Vol. 90. Human papillomaviruses. Lyon, France: International Agency for Research on Cancer, Khan MJ, Castle PE, Lorincz AT, et al. The elevated 10-year risk of cervical precancer and cancer in women with human papillomavirus 1298 JID 2006:194 (1 November) Wheeler et al.

9 (HPV) type 16 or 18 and the possible utility of type-specifi HPV testing in clinical practice. J Natl Cancer Inst 2005; 97: Castle PE, Solomon D, Schiffman M, Wheeler CM. Human papillomavirus type 16 infections and 2-year absolute risk of cervical precancer in women with equivocal or mild cytologic abnormalities. J Natl Cancer Inst 2005; 97: Peto J, Gilham C, Deacon J, et al. Cervical HPV infection and neoplasia in a large population-based prospective study: the Manchester cohort. Br J Cancer 2004; 91: Schiffman M, Adrianza ME. ASCUS-LSIL Triage Study: design, methods and characteristics of trial participants. Acta Cytol 2000; 44: The Atypical Squamous Cells of Undetermined Significance/Low Grade Squamous Intraepithelial Lesions Triage Study (ALTS) Group. Human papillomavirus testing for triage of women with cytologic evidence of low-grade squamous intraepithelial lesions: baseline data from a randomized trial. J Natl Cancer Inst 2000; 92: Solomon D, Schiffman M, Tarone R. Comparison of three management strategies for patients with atypical squamous cells of undetermined significance baseline results from a randomized trial. J Natl Cancer Inst 2001; 93: The Atypical Squamous Cells of Undetermined Significance/Low Grade Squamous Intraepithelial Lesions Triage Study (ALTS) Group. Results of a randomized trial on the management of cytology interpretations of atypical squamous cells of undetermined significance Am J Obstet Gynecol 2003; 188: The Atypical Squamous Cells of Undetermined Significance/Low Grade Squamous Intraepithelial Lesions Triage Study (ALTS) Group. A randomized trial on the management of low-grade squamous intraepithelial lesion cytology interpretations. Am J Obstet Gynecol 2003; 188: Gravitt PE, Peyton CL, Alessi TQ, et al. Improved amplificatio of genital human papillomaviruses. J Clin Microbiol 2000; 38: 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: Peyton CL, Gravitt PE, Hunt WC, et al. Determinants of genital human papillomavirus detection in a US population. J Infect Dis 2001; 183: Guido R, Schiffman M, Solomon D, Burke L. Postcolposcopy management strategies for women referred with low-grade squamous intraepithelial lesions or human papillomavirus DNA-positive atypical squamous cells of undetermined significance a 2-year prospective study. Am J Obstet Gynecol 2003; 188: Herrero R, Castle PE, Schiffman M, et al. Epidemiologic profil of type-specifi human papillomavirus infection and cervical neoplasia in Guanacaste, Costa Rica. J Infect Dis 2005; 191: Trottier H, Mahmud S, Costa MC, et al. Human papillomavirus infections with multiple types and risk of cervical neoplasia. Cancer Epidemiol Biomarkers Prev 2006; 15: Luostarinen T, af Geijersstam V, Björge T, et al. No excess risk of cervical carcinoma among women seropositive for both HPV 16 and HPV 6/ 11. Int J Cancer 1999; 80: Silins I, Zhaohui W, A vall-lundqvist E, et al. Serological evidence for protection by human papillomavirus (HPV) type 6 infection against HPV type 16 cervical carcinogenesis. J Gen Virol 1999; 80: HPV Genotypes and Cervical Precancer JID 2006:194 (1 November) 1299