Current literature indicates that human immunodeficiency virus (HIV)-infected women are at increased risk of high-risk human papillomavirus (HPV) infection and cervical intraepithelial neoplasia (CIN) and seem to be more likely to experience progression to invasive cervical cancer.1–3 Establishing the optimal method of screening this population has remained a source of debate since the early 1990s. The accuracy of Papanicolaou tests in the detection of cervical neoplasia in HIV-infected women was questioned in reports where cervical cytology obtained in infected women, as compared with HIV-negative women, was associated with a significantly higher false-negative rate.4,5 Others, however, found that HIV status exerted minimal to no effect on the sensitivity of cytologic testing.6–10 In 1999, Goldie et al,11 in a cost-effectiveness analysis of cervical cytology among HIV-positive women, found that annual screening with Pap tests after two negative tests obtained 6 months apart (which remains the current standard) was associated with projected life expectancy benefits similar to other preventive measures. In this analysis, the use of colposcopy in the initial screening of HIV-infected women increased costs and provided no additional benefit.
The optimal management of HIV-infected women with abnormal cervical cytology, specifically, women with mildly abnormal cytology (defined as atypical squamous cells of undetermined significance [ASC-US] and low-grade squamous intraepithelial lesions [LSIL]), however, remains unclear. For example, oncogenic HPV testing was initially not recommended in the management of ASC-US in HIV-infected women as compared with seronegative women.12 Although these guidelines were revised in 2006 and now endorse similar management schemes for ASC-US and LSIL irrespective of HIV status,13 disagreement remains over whether HIV-infected women with mild cytologic abnormalities are at similar or increased risk for clinically significant disease as compared with the uninfected population. The purpose of the present study, then, is to compare the prevalence of histologic high-grade disease (CIN 2+) in a cohort of HIV-positive and HIV-negative women with ASC-US and LSIL.
MATERIALS AND METHODS
Beginning in 1999, women seen in the Colposcopy Clinic at Women and Infants’ Hospital in Providence, Rhode Island have been entered into a colposcopic database. This database of 3,013 patients was searched for cases of women with known HIV infection who were referred to colposcopy between August 1999 and May 2006 for the evaluation of mildly abnormal cervical cytology (ASC-US or LSIL). It should be noted that women referred to the colposcopy clinic originate both internally from providers within the hospital and externally from numerous clinical sites throughout Rhode Island, including a number of clinics specifically serving HIV-infected women.
For the purposes of this cross-sectional study, HIV-infected and HIV-negative women were matched by referral cytology (ASC-US compared with LSIL) and for those with ASC-US, by date of referral cytology (ie, before compared with after July 2002 when reflex HPV testing for ASC-US became the standard of care at our institution). The reasons for this are twofold. In the database, HIV-positive women were more likely to be referred with LSIL, and to avoid overestimating the risk of high-grade histologic disease in this population compared with the HIV-negative population, we chose to match HIV-infected and HIV-negative women by cytologic diagnosis. Due to similar concerns regarding overestimation of risk among HIV-infected women, we also matched by referral date of ASC-US cytology, because the addition of oncogenic HPV testing in the setting of ASC-US in July 2002 allowed for the identification of a population of HIV-negative women at greater risk of significant histologic disease. Because only HPV-positive women with ASC-US underwent colposcopy evaluation, we risked underestimating disease in the HIV-negative population if the date of ASC-US cytology was not considered. Inclusion further required available histologic specimens obtained within 12 months of the referral cytology. In cases where an additional Pap test was performed on the day of colposcopic examination, the referral Pap was used for analysis. Patients were excluded if no biopsy was taken, if there was insufficient tissue on biopsy for analysis, or if the interval between the initial referral cytology and the histologic sample exceeded 12 months. Pregnancy was not an exclusion criterion, as long as a histologic specimen was available. Women with atypical squamous cells, cannot exclude high-grade squamous intraepithelial lesions, high-grade squamous intraepithelial lesions, invasion, or any glandular lesion on cervical cytology were excluded. The Women and Infants’ Institutional Review Board approved the use of the database, the medical record, and computerized record review for this study in June 2006.
The database includes information on a number of demographic variables, including not only age and HIV status but also race/ethnicity, parity, smoking history, age at first coitus, number of lifetime sexual partners, as well as histories of prior abnormal cytologic tests or biopsies. Cervical cytologic tests were obtained using fluid-based thin-layer cytology and were generated from PreservCyt (Cytyc Corporation, Marlborough, MA) samples by use of a semiautomated processor (ThinPrep 3000; Cytyc Corporation). Colposcopic examinations were performed by resident physicians and nurse colposcopists under the guidance of attending physicians, and biopsies (including endocervical sampling) were obtained at the discretion of the providers. All histologic specimens were placed in formalin before transport to pathology. Women and Infants’ Hospital cytotechnicians and pathologists evaluated all specimens and classified them according to 2001 Bethesda terminology or the CIN histologic grading system as appropriate. It is of note that for patients with unknown HIV status, HIV testing (using HIV enzyme immunoassay with confirmation by Western blot assay) was offered at the initial colposcopy visit. For HIV-positive women referred from the Miriam Hospital HIV clinic, the most recent CD4(+) lymphocyte count and viral load measurement was also included in the database if obtained within 12 months of the date of the colposcopic examination. At each visit, current use of antiretroviral therapy was also recorded. Each HIV-positive case was then matched with five controls on the basis of cytologic classification on referral (ASC-US, LSIL) and if ASC-US, timing (before or after July 2002) of referral. A stratified random sampling procedure was used to select controls for each case, without replacement.
Descriptive statistics were calculated and compared for cases and controls. The incidence of histologic high-grade disease was calculated for each group, with high-grade disease being defined as CIN 2 or greater. Differences in patient characteristics were compared by the Wilcoxon rank sum test for continuous variables (all variables had P<.05 by the Shapiro-Wilk test of normality) and χ2 for categorical variables. Conditional logistic regression was used for the calculation of odds ratios (ORs) and 95% confidence intervals (CIs) while accounting for the matched study design. In the final model, adjustment was made for significantly different characteristics between the groups (ie, variables with P<.1 in the univariable analysis). In a subanalysis, the HIV-infected group was also stratified by CD4(+) lymphocyte count to evaluate the effect of the level of immunosuppression on the frequency of biopsy-proven high-grade disease. All analyses were performed using SAS 9.1 (SAS Institute Inc., Cary, NC).
An initial sample size calculation assumed a probability of CIN 2 or worse of 30% in the HIV-infected population and 15% in the HIV-negative population. We based this assumption on data from the ASCUS LSIL Triage study, where approximately 17% of the population with ASC-US or LSIL had prevalent CIN 2 or worse, and from prior data showing HIV-positive women with mild cytologic abnormalities to be at significantly increased risk for the detection of high-grade histologic disease.14,15 With α set at 0.05 and β at 0.20, and allowing for five HIV-negative controls per case, we calculated a final required sample size of 450 women (75 HIV-positive and 375 HIV-negative).
Of the 3,013 women in the colposcopy database, 103 were known to be HIV-positive, 72 (70%) of whom had been referred for colposcopic evaluation following a Pap test classified as ASC-US or LSIL. For study purposes, five HIV-negative women were chosen from the population of 1,788 seronegative women initially referred with mildly abnormal cervical cytology and matched by cytologic diagnoses and timing of diagnosis to the 72 HIV-infected women.
Compared with HIV-negative women, HIV-infected women were significantly more likely to be older (median age 37.5 years compared with 23.0 years; P<.001), African-American (26 or 38% compared with 69 or 20%; P=.02), and of higher gravidity (median 3 compared with 1; P<.0001) and parity (median 2 compared with 1; P<.001). HIV-infected women were also more likely to be current smokers (38 or 55% compared with 139 or 39%; P=.01) and to have a history of sexually transmitted infections (45 or 67% compared with 186 or 52%; P=.03). See Table 1 for full details of this initial analysis.
Human immunodeficiency virus–positive women were as likely as HIV-negative women to have CIN 2 or worse on biopsy (11 of 72 or 15.3% compared with 48 of 360 or 13.3% of the HIV-negative women; crude OR 1.17, 95% CI .58–2.39). After adjustment for factors found to differ significantly between the two populations in the initial analysis (ie, age, race/ethnicity, insurance status, gravidity, current smoking and history of sexually transmitted illnesses), the odds of CIN 2 or worse in HIV-infected women with mildly abnormal cervical cytology remained unchanged (adjusted OR 0.88, 95% CI 0.35–2.18; Table 2). In a subsequent analysis of the HIV-positive cohort with known CD4 counts determined within 12 months of their referral cytology, the prevalence of CIN 2 or greater on biopsy did seem to increase with increasing immunosuppression. Among women with CD4(+) lymphocyte counts less than 200 cells/microliter, 4 of 10 (40%) were found to have CIN 2+ compared with 4 of 30 (13%) of women with CD4 counts greater than or equal to 200 cells/microliter (OR 4.3, 95% CI 0.84–22.47). This result, however, lacked statistical significance due to small sample size (Table 3).
Finally, HIV-infected women with CD4 counts less than 200 cells/microliter, compared with those with higher CD4 counts, were more likely to have higher viral loads (median viral load among those with lower CD4 counts was 17,542 copies/mL compared with 101.5 copies/mL among those with higher counts) and to be on some form of antiretroviral therapy (60% compared with 35% among those with CD4 counts of 200 or more). In a separate analysis of the 68 HIV-infected women with documentation of medication at the time of colposcopic evaluation, highly active antiretroviral therapy use was associated with approximately a threefold increase in the risk of CIN 2+ (crude OR 2.75, 95% CI 0.23–33.28). This comparison was based on small numbers and was not statistically significant (full data available on request).
In recently diagnosed HIV-infected women, recommendations for cervical cancer screening include assessment with biannual Pap tests followed by annual screening if both tests are negative. Based on numerous studies documenting the accuracy and cost-effectiveness of cytology in this population, this recommendation remains the standard of care for HIV-positive women. Fewer reports, however, have addressed the optimal management of infected women with mildly abnormal cytologic findings. From initial studies, it seemed that histologic disease may be more frequently found among HIV-positive women, as compared with HIV-negative, with mildly abnormal Pap tests. In one cross-sectional study of HIV-negative and HIV-infected women referred to a colposcopy clinic for the evaluation of abnormal Pap tests, cytologic and histologic characteristics were found to be highly correlated in the HIV-negative population, a finding not seen in the HIV-positive women. Biopsy results revealed that 49% of the HIV-infected women had histologic characteristics more severe than their cytology indicated, compared with 27% of the HIV-negative patients.16 In another study of HIV-infected and uninfected women, all with mild cytologic atypia on Pap test, Wright et al15 found that the HIV-infected women were approximately four times more likely to have CIN on biopsy, thus leading to the authors’ recommendation that all HIV-positive women with mild atypia undergo colposcopy.
Other investigators, however, found no significant effect of HIV infection on either cytologic–histologic discrepancy or on the accuracy of abnormal Pap tests. In a small study designed specifically to address the issue of discrepancy, Boardman et al17 found no significant difference in the rate of discrepancy between cytology and histology for a cohort of more than 600 women, regardless of their HIV status. Similarly, in a recent prospective study following 189 HIV-infected women and 95 uninfected women enrolled in the Baltimore HIV Epidemiology Research study site, Anderson and colleagues10 demonstrated a high level of concordance between cytology and colposcopic and histologic findings. Although the frequency of abnormal histology was similar among women with ASC-US and SIL (the authors combined LSIL and high-grade squamous intraepithelial lesions) regardless of HIV status, the authors did not indicate the degree of CIN found by cytologic diagnosis. In general, the rates of CIN 2 or worse in our study were low, a finding similar to those in an uncontrolled study by Kirby and colleagues18 in which high-grade histologic disease was documented in only 12% of 42 HIV-positive women with ASC-US.
We have documented not only that the level of concordance between mildly abnormal cervical cytology and histology is similarly high between HIV-infected and uninfected women but also that the frequency of high-grade histologic disease is comparable between the two populations and approximates 15%, not the 30% we initially postulated for the HIV-positive women. HIV status alone, then, did not seem to affect the accuracy of mildly abnormal Pap tests. Within the HIV-infected population, worsening immunosuppression (ie, CD4 counts of less than 200 cells/microliter compared with 200+ cells/microliter) did tend to be more frequently associated with the finding of CIN 2 or worse on biopsy. Due to the small sample available and lack of significant differences detected, we are limited in our ability to draw conclusions regarding the role of immunosuppression. In contrast, Anderson et al10 found that abnormal histologic findings did not differ by HIV infection or by CD4 cell count for women with ASC-US or SIL, although lower CD4 counts (less than 500 cells/microliter) did correlate with the finding of abnormal histology among a small group of HIV-positive women with normal cervical cytology.
Given accumulating evidence that mildly abnormal cervical cytology in HIV-positive women can be managed conservatively, we attempted to isolate potential predictive factors (in this case, CD4 counts) associated with increased risk of clinically significant histologic disease. We were, however, limited in drawing conclusions given that CD4 counts were not available on the entire HIV-infected cohort. Certain characteristics of both the HIV-positive and HIV-negative populations may also limit the study’s generalizability. The retrospective nature of this study likewise imposes inherent limitations, such as inconsistent collecting and recording of patient characteristics. In general, the latter was rarely encountered, because the same data collection form was consistently used for all colposcopy patients.
The strengths of our study include the size of the HIV-infected cohort as well as the study design itself (ability to match for cytology of HIV-negative women pulled from a large colposcopy database). The risk of missing a potential difference in the rate of high-grade histologic disease between HIV-infected and uninfected women with mildly abnormal cervical cytology clearly exists. Although the actual sample size and observed frequency of CIN 2 or worse among the HIV-uninfected were close to the expected values, our power to detect the originally specified difference (OR approximately 2.5) remained approximately 80%. The power to detect smaller differences, however, was limited, although the clinical relevance of smaller differences is unclear.
Our finding that mildly abnormal Pap tests are equally accurate in detecting histologic high-grade disease in women who are HIV-positive and HIV-negative and that the rate of CIN 2 or greater is similar between the two groups of women supports cytologic assessment in the infected population and suggests that HIV status alone does not confer an augmented risk of significant histologic disease. Whether the degree of immunosuppression should be considered in the management of HIV-infected women with mildly abnormal cervical cytology requires further investigation.
1. Ellerbrock TV, Chiasson MA, Bush TJ, Sun XW, Sawo D, Brudney K, et al. Incidence of cervical squamous intraepithelial neoplasia in HIV infected women. JAMA 2000;283:1031–7.
2. Hawes SE, Critchlow CW, Faye Niang MA, Diouf MB, Diop A, Touré P, et al. Increased risk of high-grade cervical squamous intraepithelial lesions and cervical cancer among African women with human immunodeficiency virus type 1 and 2 infections. J Infect Dis 2003;188:555–63.
3. Schuman P, Ohmit SE, Klein RS, Duerr A, Cu-Uvin S, Jamieson DJ, et al. Longitudinal study of cervical squamous intraepithelial lesions in human immunodeficiency virus (HIV)–seropositive and at-risk HIV-seronegative women. J Infect Dis 2003;188:128–36.
4. Maiman M, Tarricone N, Vieira J, Suarez J, Serur E, Boyce JG. Colposcopic evaluation of human immunodeficiency virus–seropositive women. Obstet Gynecol 1991;78:84–8.
5. Del Priore G, Maag T, Bhattacharya M, Garcia PM, Till M, Lurain JR. The value of cervical cytology in HIV-infected women. Gynecol Oncol 1995;56:395–8.
6. Branca M, Rossi E, Alderisio M, Migliore G, Morosini PL, Vecchione A, et al. Performance of cytology and colposcopy in diagnosis of cervical intraepithelial neoplasia (CIN) in HIV-positive and HIV-negative women. Cytopathology 2001;12:84–93.
7. Korn AP, Autry M, DeRemer PA, Tan W. Sensitivity of the Papanicolaou smear in human immunodeficiency virus–infected women. Obstet Gynecol 1994;83:401–4.
8. Adachi A, Fleming I, Burk RD, Ho GY, Klein RS. Women with human immunodeficiency virus infection and abnormal Papanicolaou smears: a prospective study of colposcopy and clinical outcome. Obstet Gynecol 1993;81:372–7.
9. Spinillo A, Capuzzo E, Tenti P, De Santolo A, Piazzi G, Iasci A. Adequacy of screening cervical cytology among human immunodeficiency virus–seropositive women. Gynecol Oncol 1998;69:109–13.
10. Anderson JR, Paramsothy P, Heilig C, Jamieson DJ, Shah K, Duerr A, et al. Accuracy of Papanicolaou test among HIV-infected women. Clin Infect Dis 2006;42:562–8.
11. Goldie SJ, Weinstein MC, Kuntz KM, Freedberg KA. The costs, clinical benefits, and cost-effectiveness of screening for cervical cancer in HIV-infected women. Ann Intern Med 1999;130:97–107.
12. Wright TC Jr, Cox JT, Massad LS, Twiggs LB, Wilkinson EJ, ASCCP-Sponsored Consensus Conference. 2001 Consensus Guidelines for the management of women with cervical cytological abnormalities. JAMA 2002;287:2120–9.
13. Wright TC Jr, Massad LS, Dunton CJ, Spitzer M, Wilkinson EJ, Solomon D, et al. 2006 consensus guidelines for the management of women with abnormal cervical cancer screening tests. Am J Obstet Gynecol 2007;197:346–55.
14. Guido R, Schiffman M, Solomon D, Burke L, for the ASCUS LSIL Triage Study (ALTS) group. Postcolposcopy management strategies for women referred with low-grade squamous intraepithelial lesions or human papillomavirus DNA-positive atypical squamous cells of undetermined significance: a two-year prospective study. Am J Obstet Gynecol 2003;188:1401–5.
15. Wright TC Jr, Moscarelli RD, Dole P, Ellerbrock TV, Chiasson MA, Vandevanter N. Significance of mild cytologic atypia in women infected with human immunodeficiency virus. Obstet Gynecol 1996;87:515–9.
16. Fruchter RG, Maiman M, Sillman FH, Camilien L, Webber CA, Kim DS. Characteristics of cervical intraepithelial neoplasia in women infected with human immunodeficiency virus. Am J Obstet Gynecol 1994;171:531–7.
17. Boardman LA, Peipert JF, Cooper AS, Cu-Uvin S, Flanigan T, Raphael SI. Cytologic-histologic discrepancy in human immunodeficiency virus–positive women referred to a colposcopy clinic. Obstet Gynecol 1994;84:1016–20.
18. Kirby TO, Allen ME, Alvarez RD, Hoesley CJ, Huh WK. High-risk human papillomavirus and cervical intraepithelial neoplasia at time of atypical squamous cells of undetermined significant cytologic results in a population with human immunodeficiency virus. J Low Genit Tract Dis 2004;8:298–303.