Effect of HIV Infection on Atypical Squamous Cells of Undetermined Significance

Transcription

1 MAJOR ARTICLE HIV/AIDS Effect of HIV Infection on Atypical Squamous Cells of Undetermined Significance Ann Duerr, 1 Pangaja Paramsothy, 2 Denise J. Jamieson, 3 Charles M. Heilig, 3 Robert S. Klein, 4 Susan Cu-Uvin, 5 Paula Schuman, 6 and Jean R. Anderson, 7 for the HIV Epidemiology Research Study 1 Fred Hutchinson Cancer Research Center, Seattle, Washington; 2 Contraceptive Research and Development, Arlington, Virginia; 3 Centers for Disease Control and Prevention, Atlanta, Georgia; 4 Montefiore Medical Center and Albert Einstein College of Medicine, Bronx, New York, 5 Brown Medical School, Providence, Rhode Island, 6 Wayne State University School of Medicine, Detroit, Michigan, and 7 Johns Hopkins University School of Medicine, Baltimore, Maryland Background. Detection of atypical squamous cells of undetermined significance (ASCUS) is a cervical cytologic finding that is suggestive but not definitive of squamous intraepithelial lesions (SILs). Methods. We examined the risk, characteristics, and progression of ASCUS in with and without human immunodeficiency virus (HIV) infection. Cervical Papanicolou (Pap) test and colposcopy data were obtained at the first 10 semiannual visits for the HIV Epidemiology Research study of 774 and 480 demographically similar, HIV-uninfected in the United States. Multiple logistic regression models and Cox proportional hazards models were utilized. Results. ASCUS was more common among (odds ratio [OR], 1.6 [95% confidence interval {CI}, ] to 2.6 [95% CI, ]) after adjustment for human papillomavirus (HPV) infection and other risk factors (e.g., race, condyloma, and prior Pap test result). Among with normal Pap test results at enrollment, the cumulative incidence of ASCUS was 78% among and 38% among HIVuninfected. and HIV-uninfected with ASCUS did not differ by prevalence of indices of inflammation (inflammation on Pap test and leukocytes on cervical gram stain). HPV infection, including high risk types, was more common among with ASCUS. Among with ASCUS, 60% of and 25% of HIV-uninfected developed SILs ( P!.01 ). Compared with with higher CD4 + lymphocyte counts, with CD4 + lymphocyte counts!200 cells/ml were more likely to present subsequently with a SIL detected by Pap test (OR, 1.7; 95% CI, ). Conclusions. Higher risk of SIL following the appearance of ASCUS among, especially with low CD4 + lymphocyte counts, supports the need for follow up with colposcopy and histologic examination, as indicated, to allow early detection and treatment of SIL. Both the etiology and outcome of atypical squamous cells of undetermined significance (ASCUS) detected by cervical cytologic testing are uncertain, as the name implies. Although it is recommended that laboratories avoid this diagnosis when a more precise diagnosis can be given [1], the inflammatory and predysplastic changes that lead to a diagnosis of ASCUS are typically reported in 5% 10% of Papanicolaou (Pap) tests [2 4]. For some, ASCUS may be an indication of coexisting dysplasia. Moreover, a diagnosis of ASCUS Received 12 August 2005; accepted 22 November 2005; electronically published 10 February Reprints or correspondence: Dr. Ann Duerr, 1100 Fairview Ave. N. LE-500, P.O. Box 19024, Seattle, Washington (aduerr@fhcrc.org). Clinical Infectious Diseases 2006; 42: by the Infectious Diseases Society of America. All rights reserved /2006/ $15.00 may be the result of early cytologic findings in eventually diagnosed with severe dysplasia, with more than one-third of cases of biopsy-proven cervical intraepithelial neoplasia being associated at the preceding Pap test with ASCUS, rather than with squamous intraepithelial lesions (SILs) [3]. Current management guidelines [6] propose follow-up of ASCUS using any of the following methods: repeat cytological testing, colposcopy, or DNA testing for high-risk human papillomavirus (HPV) types. Referral for colposcopy is recommended for all with ASCUS. These guidelines recommend referral for colposcopy for all with ASCUS when high-grade SIL cannot be ruled out. ASCUS remains an incompletely described entity among HIV-uninfected, even more so among. In contrast, the increased risk of HIV/AIDS CID 2006:42 (15 March) 855

2 SIL among has been well documented, and several groups [7, 8] have reported the increased risk of progression of SIL in this population. Fewer studies have reported on ASCUS [5]. Because there are few data on the comparative prevalence of ASCUS among and HIVuninfected and on the risk of cervical intraepithelial neoplasia among with ASCUS, it is difficult to know whether current management algorithms for ASCUS are optimal for. Using data from a large prospective study, we examined AS- CUS in and HIV-uninfected with similar risk profiles. We looked specifically at whether the risk or characteristics of ASCUS differ between the 2 groups and whether HIV infection alters the risk of subsequent progression to SIL among with ASCUS. METHODS The HIV Epidemiology Research Study enrolled 1310 (871 and 439 at-risk, HIV-uninfected ) from 4 US cities (Bronx, NY; Baltimore, MD; Providence, RI; and Detroit, MI) between April 1993 and January 1995; details of the design and enrollment procedures have been reported [9]. Institutional review boards approved the study protocol at each site and at the Centers for Disease Control and Prevention; all participants gave written, informed consent. Each site contributed approximately one-fourth of the participants. HIVinfected without AIDS-defining conditions and HIVuninfected were selected in such a way that about one-half reported risk related to injection drug use, and the remainder had increased risk due to sexual behavior but not injection drug use. Pap tests were performed with a cytobrush and a wooden spatula on all with a cervix. All Pap test results were evaluated using predefined criteria [10; an expansion of the 1988 Bethesda scoring system for cervical cytology] by a single commercial cytologic laboratory (Kyto Diagnostics, New York) [11, 12]. At 1 study site (Baltimore), cervical colposcopy was performed routinely at each study visit; at the other 3 sites, were referred for colposcopy if they had 1 Pap test result showing SIL or 2 consecutive Pap test results showing ASCUS. Cervical and/or vaginal secretion specimens and exfoliated cells were collected by cervicovaginal lavage and tested for HPV using PCR [9 11]. The gynecologic exam also ascertained the presence of infectious and inflammatory conditions, discharge, visible ulcers, and condylomata. Gram stains of cervical secretion specimens were scored for the presence of polymorphonuclear leukocytes, monocytes, and bacteria. Gram stains of vaginal secretion specimens were scored for bacterial vaginosis and for the presence of yeast, sperm, RBCs, polymorphonuclear leukocytes, and monocytes. Trichomoniasis was initially defined using wet mount (at visits 1 3); culture in modified Diamond s medium was introduced for all at visit 4. Women with symptomatic vaginal or cervical infections were either treated or referred for treatment. HIV infection status was ascertained through ELISA and Western blotting, with assessment of immune system compromise using CD4 + cell counts, as described elsewhere [13]. HIV load was determined via a third-generation, branched DNA signal amplification assay (Chiron) with a lower quantification limit of 50 copies/ml. Antiretroviral use was ascertained by self-reporting; treatment regimens were classified as HAART, other antiretroviral therapy, or no antiretroviral therapy. Regimens not listed in published guidelines [14] were classified as HAART if they were believed to have a high likelihood of effective and sustained viral suppression. Statistical methods. Pearson s x 2 test or Fisher s exact tests were used to compare proportions, and Student s t test was used to compare means between groups. The Wilcoxon ranksum test was used for variables that were not normally distributed. The cumulative incidence of ASCUS was estimated over the course of this study using the product-limit survival estimator. Logistic models provided estimates of overall prevalence by means of methods that compensated for within-subject correlation across multiple observations. These prevalence estimates averaged the characteristics of participants over the course of the study. Multiple logistic regression models that corrected for repeated measures were used to identify correlates of abnormal cervical cytologic findings. Initial models examined covariates one at a time and adjusted for repeated measures and 2 design variables: study site and risk category (sexual behavior or injection drug use). Factors suggesting at least a moderate ( P!.2) association with ASCUS were included in the subsequent models. Interaction between HIV and other covariates was assessed and included if P!.2. Inferences were based on SEs estimated by robust methods using an m-dependent working correlation structure. Cox proportional hazard models were used to calculate hazard ratios for the risk of ASCUS in with normal Pap test results at baseline, as well as the risk of SIL in with Pap test results showing ASCUS at baseline. Factors suggesting at least a moderate ( P!.2) association with the outcome were included in the subsequent models. The multivariate model also included the design variables study site and risk category. Interaction between HIV and other covariates was assessed and included if P!.2. All calculations were performed using SAS software, version 8.2 (SAS Institute). 856 CID 2006:42 (15 March) HIV/AIDS

3 Table 1. Inflammation and HPV infection among with atypical squamous cells of undetermined significance (ASCUS). Among No. (%) of visits Among HIV-uninfected Inflammation detected by Pap test 175 (15) 57 (20).28 Inflammation with cellular change detected by Pap test 49 (4.2) 15 (15).23 Mononuclear cell count b.53!1 642 (82) 152 (78) (10) 24 (12) 3 64 (8) 18 (9) PMN count b.93!1 247 (32) 59 (30) (23) 42 (22) (22) 39 (20) (24) 54 (28) HPV infection status determined by hybrid capture c!.01 Infected 522 (58) 59 (24) Uninfected 376 (42) 185 (76) HPV risk type determined by PCR!.01 High 260 (23) 39 (14) Intermediate 267 (24) 34 (12) Low 188 (17) 19 (7) Untypeable 196 (17) 44 (16) Not available 217 (19) 146 (52) Histologic finding d.33 Normal biopsy and normal colposcopy results 138 (81) 37 (80) CIN I 31 (18) 2 (4) CIN not otherwise specified 2 (1) 7 (15) NOTE. HPV, human papillomavirus; Pap, Papanicolou; PMN, polymorphonuclear leukocytes. a Multiple logistic regression models were used to adjust for repeated measures. b Determined by cervical Gram staining, which was restricted to visits 1 5 for with ( n p 782) and without ( n p 194) HIV infection. Data are no. of cells per 5 oil immersion fields. c Restricted to 898 visits among and 244 visits among HIV-uninfected. Samples that tested positive for any HPV type by hybrid capture were considered to demonstrate infection; samples that tested negative or indeterminate by hybrid capture or negative by PCR were considered to demonstrate noninfection. d Restricted to from 1 study site where colposcopy was routinely performed at each visit, regardless of Pap test result. P a RESULTS We considered cervical Pap test data from the first 10 semiannual study visits (over a period of up to 4.7 years) and included results from 8185 Pap tests. Data were available for 795 and 404 HIV-uninfected (N p 1199) with intact cervices at baseline. A median follow-up time of 4.4 years did not differ by HIV serostatus. At enrollment, were more likely to be African American and were less likely to have finished high school. For the 6 months prior to enrollment, were more likely to have reported always using a condom, less likely to have reported using crack, less likely to have reported having 11 male sexual partner than HIV-uninfected. HIVinfected and HIV-uninfected did not differ by the following factors at enrollment: risk cohort (drug use or sexual behavior), age, income, number of live births, smoking (currently), exchange of sex for drugs or money, injection drug use, and number of female sex partners (in the previous 6 months). Characteristics of ASCUS. In general, markers of cervical inflammation observed on Pap tests or on cervical specimens collected concurrently were of similar frequency among HIVinfected and HIV-uninfected with ASCUS (table 1). The presence of vaginal infections that might affect Pap test HIV/AIDS CID 2006:42 (15 March) 857

4 Table 2. Factors associated with atypical squamous cells of undetermined significance (ASCUS) versus normal Papanicolaou (Pap) test result. Adjusted OR (95% CI) a All (N p 1199) (n p 795) HIV status and cervical ectopy, cervical ectopy 2.6 ( ) 1.7 ( ), no ectopy 1.6 ( ) 1.0 b HIV-uninfected, cervical ectopy 0.9 ( ) HIV-uninfected, no ectopy 1.0 b CD4 + cell count, cells/ml! ( ) ( ) b HIV load, copies/ml 150, ( ) 10,001 50, ( ) , ( ) ( ) Undetectable 1.0 b Current HAART use c HAART 1.4 ( ) Other antiretroviral use 1.2 ( ) No antiretroviral use 1.0 b HPV risk type determined by PCR High 5.1 ( ) 5.4 ( ) Intermediate 4.0 ( ) 4.0 ( ) Low 2.4 ( ) 2.6 ( ) Untypeable 2.1 ( ) 1.9 ( ) No types detected 1.0 b 1.0 b Trichomoniasis present 0.8 ( ) 0.7 ( ) Cervical, vulvar, or vaginal condyloma 1.3 ( ) 1.3 ( ) Race and/or ethnicity African American 0.7 ( ) 0.7 ( ) Hispanic 0.8 ( ) 0.7 ( ) Other 0.7 ( ) 0.2 ( ) White 1.0 b 1.0 b No. of male sex partners within the previous 6 months ( ) 0.9 ( ) ( ) 0.8 ( ) b 1.0 b Pap test result at last visit SIL 12.6 ( ) 11.4 ( ) ASCUS 6.7 ( ) 5.8 ( ) Normal 1.0 b 1.0 b NOTE. Multiple logistic regression models that corrected for repeated measures were used to calculate ORs. HPV, human papillomavirus; SIL, squamous intraepithelial lesion. a ORs are adjusted for all other risk factors in the table, as well as for study site, risk cohort, and repeated measures. b Referent. c As defined by the Department of Health and Human Services guidelines.

5 Figure 1. Data shown for with normal Papanicolaou (Pap) test results at enrollment. The cumulative incidence of atypical squamous cells of undetermined significance (ASCUS) was 78% among and 38% among HIV-uninfected ( P!.01). Kaplan-Meier and product limit estimator method were utilized. results also did not differ markedly according to status of HIV infection. with ASCUS were more likely to be colonized with Candida species than were HIV-uninfected (36.7% of visits vs. 27.5% of visits; P!.01), but they did not differ in the presence of the following 2 vaginal infections: bacterial vaginosis (42.1% vs. 40.4%; P p.64) and trichomoniasis (15.1 vs. 13.2%; P p.43) (data not shown). HPV infection, including infection caused by high-risk virus types, was more common among with ASCUS than it was among HIV-uninfected with ASCUS (P!.01). Despite this, routine colposcopy performed for all HIVinfected and HIV-uninfected at 1 site did not reveal a higher prevalence of aceto-white, punctation, mosaicism, or atypical vessels among with ASCUS (data not shown). with ASCUS did not have a higher prevalence of cervical intraepithelial neoplasia (table 1). Prevalence of and risk factors for ASCUS. Although the characteristics of ASCUS appeared not to differ according to status of HIV infection, it was more common in both at the baseline visit (adjusted prevalence ratio, 1.6; 95% CI, ) and during the follow-up period. For all, the multivariate model presented (table 2) contains all variables found in initial analyses to be significantly associated with the risk of ASCUS ( P!.2). Factors examined in initial analyses and not found to be significantly associated with ASCUS included age, education, number of pregnancies or live births, recent sex with a woman (during the previous 6 months), condom use (during the previous 6 months), ever having exchanged sex for drugs or money, injection drug use, douching, vaginal colonization with Candida species, and bacterial vaginosis. In the final adjusted model for all, ASCUS was associated with HIV infection; cervical ectopy present during physical exam; cervical, vaginal, and/or vulvar condyloma present during physical exam; HPV infection; race and/ or ethnicity; and Pap test result from the previous visit. The association between ASCUS and HIV infection was modified by the presence of cervical ectopy. In the multivariate model of (table 2, right column), the level of ASCUS was significantly elevated in with CD4 + cell counts!200 cells/ml (adjusted OR, 1.5; 95% CI, ). Risk of developing ASCUS. Compared with HIV-uninfected, were more likely to present with ASCUS after receiving a normal Pap test result (figure 1). Among the 426 and 275 HIV-uninfected with normal Pap test results at enrollment, the cumulative incidence of ASCUS was 78% among the and 38% among the HIV-uninfected. In a proportional hazards model ( n p 701), the adjusted relative hazard for AS- CUS associated with HIV infection was 1.8 (95% CI, ) (table 3). HPV risk type at baseline was also associated with increased risk of progression to ASCUS. The hazard ratio of high-risk and intermediate-risk type was 3.2 (95% CI, ). For low-risk and untypable HPV, the hazard ratio was 1.6 (95% CI, ). Women who were culture positive for Candida species at baseline were also at risk. There was no statistical evidence of interaction between HIV infection and any other risk factor in the model. Among, those with CD4 + cell counts!200 cells/ml had a hazard ratio of 1.9 (95% CI, ). If ASCUS was defined as findings of ASCUS on 2 consecutive Pap tests, the cumulative incidence was reduced (27% and 7% in and HIV-uninfected Table 3. Risk of atypical cells of undetermined significance for with normal Papanicolaou test results at baseline. Adjusted hazard ratio (95% CI) All (n p 701) (n p 426) HIV infected 1.8 ( ) CD4 + cell count, cells/ml! ( ) ( ) a HPV risk type High or intermediate 3.2 ( ) 3.1 ( ) Low or untypeable 1.6 ( ) 1.8 ( ) No type detected 1.0 a 1.0 a Culture positive for Candida species 1.3 ( ) b 1.4 ( ) NOTE. Cox proportional hazard models were used to calculate hazard ratios. Models were also adjusted for the study-design variables site and risk cohort. HPV, human papillomavirus. a Referent. b P p.029. HIV/AIDS CID 2006:42 (15 March) 859

6 Figure 2. Data shown for with atypical squamous cells of undetermined significance (ASCUS) detected by Papanicolaou (Pap) test at the enrollment visit. The cumulative incidence of squamous intraepithelial lesion (SIL) was 60% among and 25% among HIV-uninfected ( P!.01). Kaplan-Meier and product limit estimator method were utilized., respectively), but the associations seen did not change (data not shown). Progression of ASCUS. Among with ASCUS at the enrollment visit (148 of whom were HIV infected and 48 of whom were HIV uninfected; figure 2), the cumulative incidence of SIL, according to Pap test results, was 60% among HIVinfected and 25% among the HIV-uninfected ( P!.01). Among with ASCUS at baseline who developed SIL during the follow-up period, the median time to SIL was 12 months among and 18 months among HIV-uninfected. Among with ASCUS at baseline, were at 2.4 times the risk of subsequently presenting with SIL, compared with HIV-uninfected. Women with high- or intermediate-risk types of HPV were significantly more likely to later present with SIL than were with no HPV present (table 4). DISCUSSION This study found that, particularly those with low CD4 + cell counts, were more likely to develop ASCUS, according to Pap test results. with ASCUS were more likely to present subsequently with cytological evidence of SIL, particularly with low CD4 + cell counts. We found that, among, increased risk of ASCUS demonstrated by Pap test and increased risk of SIL were associated with infection with high- or intermediate-risk types of HPV. A longitudinal analysis from the Women s Interagency HIV Study [15] found that, among with normal Pap test results at baseline, a greater proportion of (73.0%) than of HIV-uninfected (42.3%) had at least 1 abnormal Pap test result during up to 5 years of follow up, and 30% of all had ASCUS as a Pap test finding during this time period. In contrast to our findings, the reported prevalence of ASCUS in the Women s Interagency HIV Study did not vary significantly by status of HIV infection. By contrast, a greater proportion of in the study had low-grade SIL, compared with HIV-uninfected. ASCUS, the most common abnormal finding by Pap test, is not reliably reproducible among cytopathologists. ASCUS has been used in the Bethesda cytological classification system to indicate abnormal squamous cells that do not meet (quantitatively or qualitatively) criteria for a diagnosis of SIL [16]. Although the majority of Pap tests demonstrating ASCUS probably reflect a benign, reactive process, they may also reflect either underlying histologic abnormalities, or they may herald the future development of a histologic abnormality. With respect to management, the current guidelines state that the following 3 algorithms are considered to be acceptable for with ASCUS: repeat cytological examination, HPV DNA testing, and colposcopy [6]. By contrast, for with ASCUS for whom high-grade SIL cannot be ruled out, referral for colposcopy is the only recommended option. For, referral for colposcopy is recommended, regardless of immune status, despite limited evidence that addresses the appropriateness of this recommendation. Few studies have compared the incidence of ASCUS and the risk of SIL after an ASCUS Pap test among. The strengths of the present study include the large number of who were observed; the fact that the cohorts of HIV- Table 4. Risk of squamous intraepithelial lesion for with atypical cells of undetermined significance (ASCUS) detected by Papanicolaou test at baseline. Adjusted hazard ratio (95% CI) All (n p 196) (n p 148) HIV infected 2.4 ( ) CD4 + cell count, cells/ml! ( ) ( ) a HPV risk type High or intermediate 2.4 ( ) 2.4 ( ) Low or untypeable 1.6 ( ) 1.6 ( ) No type detected 1.0 a 1.0 a NOTE. Cox proportional hazard models were utilized to calculate hazard ratios. Models were also adjusted for the study-design variables site and risk cohort. HPV, human papillomavirus. a Referent. 860 CID 2006:42 (15 March) HIV/AIDS

7 infected and HIV-uninfected resembled each other with respect to HIV risk, and demographic characteristics; the long follow-up period with frequent collection of data regarding sexual behavior, and gynecological outcomes, and other medical history. This study s limitations include the fact that data were not available on the subclassification of ASCUS as ASC-US (atypical squamous cells undetermined significance) and ASC- H (atypical squamous cells cannot exclude high-risk squamous intraepithelial lesion [HS12]), in accordance with the most recent classification system, mainly because data collection was completed by Another limitation is the inclusion of colposcopic and histologic data for only a subset of study participants. Our finding that with ASCUS are more likely than seronegative to progress to SIL provides empirical support for the current recommendation that undergo colposcopic evaluation, with histologic examination as indicated, after the first finding of ASCUS. HIV EPIDEMIOLOGY RESEARCH STUDY GROUP Robert S. Klein, Ellie Schoenbaum, Julia Arnsten, Robert D. Burk, Chee Jen Chang, Penelope Demas, and Andrea Howard (Montefiore Medical Center and the Albert Einstein College of Medicine); Paula Schuman and Jack Sobel (Wayne State University School of Medicine); Anne Rompalo, David Vlahov, and David Celentano (Johns Hopkins University School of Medicine); Charles Carpenter and Kenneth Mayer (Brown University School of Medicine); Ann Duerr, Lytt I. Gardner, Charles M. Heilig, Scott Holmberg, Denise Jamieson, Jan Moore, Ruby Phelps, Dawn K. Smith, and Dora Warren (Centers for Disease Control and Prevention); and Katherine Davenny (National Institute of Drug Abuse). Acknowledgments Financial support. Centers for Disease Control and Prevention (U64/ CCU106795, U64/CCU206798, U64/CCU306802, and U64/CCU506831). Potential conflicts of interest. All authors: no conflicts. References 1. Solomon D, Davey D, Kurman R, et al. The 2001 Bethesda System: terminology for reporting results of cervical cytology. JAMA 2002; 287: Cox JT, Lorincz AT, Schiffman MH, Sherman ME, Cullen A, Kurman RJ. Human papillomavirus testing by hybrid capture appears to be useful in triaging with a cytologic diagnosis of atypical squamous cells of undetermined significance. Am J Obstet Gynecol 1995; 172: Kinney WK, Manos MM, Hurley LB, Ransley JE. Where s the highgrade cervical neoplasia? The importance of minimally abnormal papanicolaou diagnoses. Obstet Gynecol 1998; 91: Wright TC Jr, Sun XW, Koulos J. Comparison of management algorithms for the evaluation of with low-grade cytologic abnormalities. Obstet Gynecol 1995; 85: Holcomb K, Abulafia O, Matthews RP, et al. The significance of ASCUS in HIV-positive. Gynecol Oncol 1999; 75: Wright TC Jr, Cox JT, Massad LS, Twiggs LB, Wilkinson EJ Consensus guidelines for the management of cervical cytologic abnormalities. JAMA 2002; 287: Ellerbrock TV, Chiasson MA, Bush TJ, et al. Incidence of cervical squamous intraepithelial lesions in. JAMA 2000; 283: Schuman P, Ohmit SE, Klein RS, et al. Longitudinal study of cervical squamous intraepithelial lesions in human immunodeficiency virus (HIV) seropositive and at-risk HIV-seronegative. J Infect Dis 2003; 188: Smith DK, Warren DL, Vlahov D, et al. Design and baseline participant characteristics of the Human Immunodeficiency Virus Epidemiology Research (HER) Study: a prospective cohort study of human immunodeficiency virus infection in US. Am J Epidemiol 1997; 146: National Cancer Insitute Workshop. The 1988 bethesda system for reporting cervical/vaginal cytological diagnoses. JAMA 1989; 262: Duerr A, Kieke B, Warren D, et al. Human papillomavirus associated cervical cytologic abnormalities among with or at risk of infection with human immunodeficiency virus. Am J Obstet Gynecol 2001; 184: Jamieson DJ, Duerr A, Burk R, et al. Characterization of genital human papillomavirus infection in who have or who are at risk of having HIV infection. Am J Obstet Gynecol 2002; 186: Centers for Disease Control and Prevention Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Recomm Rep 1992; 41(RR-17): Panel on Clinical Practices for Treatment of HIV Infection. Guidelines for the use of antiretroviral agents in adults and adolescents. Department of Health and Human Services and the Henry J. Kaiser Family Foundation. Available at: Accessed 21 April Massad LS, Ahdieh L, Benning L, Minkoff H, Greenblatt RM, Watts H. Evolution of cervical abnormalities among with HIV-1: evidence from surveillance cytology in the s interagency HIV study. J Acquir Immune Defic Syndr 2001; 27: Broder S. The Bethesda System for reporting cervical/vaginal cytologic diagnoses: report of the 1991 Bethesda Workshop. JAMA 1992; 267: HIV/AIDS CID 2006:42 (15 March) 861