Secondary Logo

Journal Logo

Clinical Science

Invasive Cervical Cancer Risk Among HIV-Infected Women: A North American Multicohort Collaboration Prospective Study

Abraham, Alison G. PhD*; D’Souza, Gypsyamber PhD*; Jing, Yuezhou MS*; Gange, Stephen J. PhD*; Sterling, Timothy R. MD; Silverberg, Michael J. PhD§; Saag, Michael S. MD||; Rourke, Sean B. PhD; Rachlis, Anita MD; Napravnik, Sonia PhD#; Moore, Richard D. MD**; Klein, Marina B. MD††; Kitahata, Mari M. MD‡‡; Kirk, Gregory D. MD, PhD*; Hogg, Robert S. PhD§§; Hessol, Nancy A. MSPH||||; Goedert, James J. MD¶¶; John Gill, M. MB##; Gebo, Kelly A. MD**; Eron, Joseph J. MD#; Engels, Eric A. MD¶¶; Dubrow, Robert MD, PhD***; Crane, Heidi M. MD‡‡; Brooks, John T. MD†††; Bosch, Ronald J. PhD‡‡‡; Strickler, Howard D. MDfor the North American AIDS Cohort Collaboration on Research and Design of IeDEA

Author Information
JAIDS Journal of Acquired Immune Deficiency Syndromes: April 1st, 2013 - Volume 62 - Issue 4 - p 405-413
doi: 10.1097/QAI.0b013e31828177d7
  • Free



Human papillomavirus (HPV), a common sexually transmitted virus, is a necessary cause of invasive cervical cancer (ICC). Although the vast majority of cervical HPV infections clear or become undetectable, these infections persist in a subset of women. HIV-infected women are significantly more likely than HIV-uninfected women to have incident and persistent HPV cervical infections1 and to develop incident precancers such as squamous intraepithelial lesions (SIL)1–4 including high-grade SIL (HSIL).5,6 Among HIV-infected women, the incidence of HPV infection and SIL increases with lower CD4+ T-cell count (CD4).7,8 These collective findings strongly support a dose–response relationship between host immune status and the risk of early and intermediate stages of HPV-related tumorigenesis.1,9,10

There are few data, however, regarding the influence of immunodeficiency on the risk of incident ICC.11 Few prospective studies of HIV-infected women have had sufficient size to evaluate ICC as an outcome. Though ICC was included as an AIDS-defining event in the 1993 case definition, the evidence for inclusion came from studies of cervical dysplasia rates among HIV-infected women.11,12 Inferences regarding the risk of ICC in HIV-infected women have been based primarily on evidence from studies linking HIV/AIDS diagnosis with cancer registries. These studies have reported several fold greater incidence of ICC among women with HIV/AIDS compared with the general population.13–16 Linkage studies, however, lack detailed prospective data to assess the temporality between host immunity and cancer risk. To our knowledge, only 1 prospective cohort study examined the association between time-updated current CD4 and ICC.17 In this study, based in the French Hospital Database on HIV cohort, Guiguet et al reported a significant association of CD4 with risk of ICC.

The current study is the first multicohort prospective investigation of the relationship between HIV infection, immunosuppression, and incident ICC in North America. Using data from 18 collaborating cohorts of the North American AIDS Cohort Collaboration on Research and Design (NA-ACCORD), we examined rates of incident ICC based on cases ascertained through a rigorous standardized validation procedure. The association between CD4 and ICC risk was assessed prospectively to characterize the relevant periods of immunosuppression in relation to ICC risk.


Study Population and Design

Cases of ICC were identified from 18 prospective cohorts collaborating in the NA-ACCORD.18 The NA-ACCORD represents more than 60 clinical sites and uses standardized methods of data collection. Briefly, each contributing cohort has developed standardized cohort-specific methods of data collection. At scheduled intervals, these cohorts submit data regarding enrolled participants' demographic characteristics, dates of prescribed antiretrovirals, dates and results of laboratory tests including HIV-1 RNA viral load and CD4, dates of clinical diagnoses, and vital status. These data are transferred securely to the NA-ACCORD central Data Management Core, where they undergo quality control for completeness and accuracy before they are combined into harmonized data files. Quality control included instituting measures to reduce the probability that an individual was participating in more than 1 clinical cohort. The human subject activities of the NA-ACCORD and of each of the participating cohort studies have been reviewed and approved by their respective local institutional review boards. HIV-infected women from these cohorts contributed follow-up time to the analysis from January 1, 1996, or study entry until the earliest of ICC diagnosis, loss to follow-up, death, or cohort-specific end of follow-up (December 31, 2010, for most cohorts). Three cohorts (Kaiser Permanente Northern California, Women's Interagency Health Study, and AIDS Linked to the IntraVenous Experience Study) also contributed data from HIV-uninfected women.

Case Validation

Cases of ICC were initially identified by each cohort through chart review, linkage to a formal cancer registry or diagnostic codes. For this study, each case was individually reviewed using a standardized abstraction survey that included histologic confirmation of cancer, date of diagnosis, and source of cancer confirmation (medical records, pathology reports, and/or cancer registry records); only cases that had clear documentation of a histologic diagnosis of ICC were included. This approach emphasized specificity over sensitivity because it is well established that estimates of association between exposure (eg, host immune status) and disease are more influenced by specificity when the outcome is rare (as is the case for ICC).19 This is a particular concern for ICC because precancerous cervical lesions, including carcinoma in situ, are much more common than invasive cancer, and record misclassification can readily occur. For incident case analyses, we limited cases to those women diagnosed 6 or more months after cohort enrollment to reduce the likelihood of including prevalent but undiagnosed cases; that is, the starting time at risk for these analyses began 6 months after study enrollment. A sensitivity analysis was done that limited cases to those women diagnosed 18 or more months after cohort enrollment to assess the impact on estimates of the possible inclusion of a prevalent ICC case.

Screening History

A supplementary survey collected information from each cohort on the completeness of Papanicolaou (Pap) test history records and likelihood of women receiving screening or treatment outside of the study care facility. The survey also collected abstracted Pap screening history and colposcopy results for each case.

Using Pap screening, colposcopy, biopsy, and surgery records, we classified each incident case of ICC as associated with one of the following screening histories: (1) no known prior Pap screening within the past 5 years, (2) Pap screening within the past 5 years without detection of disease, or (3) a high-grade abnormal cytology detected within the past 5 years with no/insufficient treatment of disease.

Statistical Methods

Using all validated ICC cases (prevalent and incident), trends in ICC risk were assessed graphically by plotting the cumulative incidence of ICC as a function of age, stratified by HIV serostatus, baseline CD4 (categorized as <200, 200–349, and ≥350 cells/μL), and baseline HIV-1 RNA viral load (categorized as <4000, 4000–99,999, and ≥100,000 copies/mL). Crude ICC incidence rates were calculated by age strata (<40, 40–49, and ≥50 years; time updated) and compared with general population data from Surveillance, Epidemiology, and End Results (SEER)20 with a χ2 test. Standardized incidence ratios (SIRs) by age strata were estimated to assess risk in HIV-infected women relative to the general population. To test whether HIV-infected women were diagnosed with ICC at younger ages compared with the general population, the observed distribution of age at diagnosis among the HIV-infected women was compared with the expected age-at-diagnosis distribution of age at diagnosis in the general population following the method of Sheils et al.21

Linear mixed-effects Poisson regression was used to model the incidence of ICC. The model included time-updated age and CD4 count with HIV-infected women treated as the reference category. We separately assessed CD4 measured at the time of outcome ascertainment or diagnosis (±6 months), at 18 months before the time of outcome ascertainment or diagnosis (±6 months), and at entry into the study (which approximates nadir CD4 in many of the clinical cohorts). Differences in the incidence of ICC by cohort were accounted for using a random intercept in the model. Separate models assessed the effects of HIV RNA level (categorized as previously specified with HIV-uninfected as a reference) and calendar period (1996–2001 and 2002–2010) adjusted for age.

In addition to the incidence analysis using the full prospective data, we conducted 2 nested case–control studies to characterize and contrast the changes in CD4 before diagnosis in those who did and did not develop incident ICC. In the first study, which looked prospectively at CD4 patterns following effective antiretroviral therapy (ART) initiation, we included only incident cases of ICC among women who initiated ART during observation and prior to the ICC diagnosis date. We defined ART as a regimen containing at least three drugs, including a protease inhibitor, a non-nucleoside reverse transcriptase inhibitor, an entry inhibitor or an integrase inhibitor (new agents), or three nucleoside reverse-transcriptase inhibitors, including abacavir or tenofovir. Controls were individually matched to cases on the following factors: the date of ART initiation (±6 months), duration of follow-up after ART initiation (±6 months), CD4 category (≤200, 200–350, 350–500, and >500 cells/μL), age within ±2 years, and study cohort. All controls who met the matching criteria were included and each case had at least 1 matched control. Using these data, the post-ART (prediagnosis) CD4 patterns were compared between cases and controls using piecewise linear mixed-effects models with spline terms (broken lines) at 1 year and 3 years after initiation and random intercepts for case–control clusters and individual repeated measurements. Cases and controls contributed CD4 measurements from ART initiation to the time of diagnosis (or equivalent follow-up for controls). Models were only fit to the first 5 years of data after ART initiation due to sparse data beyond 5 years post-ART.

In the second nested case–control study, which looked retrospectively at CD4 patterns leading up to ICC diagnosis, all incident ICC cases regardless of ART history were matched individually to controls at the time of cancer diagnosis using the following variables: age (±3 months), cohort, current ART use (yes/no), and exact calendar year. Specifically, CD4 changes over time were compared between cases and controls for more than the 5 years before diagnosis using piecewise linear mixed-effects models with random intercepts for case–control clusters and individual repeated measurements.


A total of 13,690 HIV-infected women without a prior diagnosis of ICC contributed 66,249 person-years (pys) to this analysis. In addition, there were 12,021 HIV-uninfected women from 3 cohorts who contributed 70,815 pys. A cohort-level description of the cases and pys contributed is provided in the Appendix Table 1. The median follow-up time was 4.5 years [interquartile range (IQR) = 1.5–8.3 years] for HIV-infected women and 5.0 years (IQR = 2.3–10.0) for HIV-uninfected women. At enrollment, HIV-infected and -uninfected women both had a median age of 37 years, though HIV-infected women were more likely to have enrolled later (P < 0.001) and be of black race (P < 0.001). The median baseline CD4 in HIV-infected women was 342 cells per microliter, and the prevalence of current or prior ART use was low (29%), although 71% initiated therapy during follow-up (Table 1).

Baseline Characteristics of HIV-Infected Women Diagnosed With ICC in NA-ACCORD 1996–2010, Stratified Into Incident Cases (Detected >6 Months After Enrollment) and Prevalent (Detected Before Enrollment or Within First 6 Months of Enrollment). Characteristics of All HIV-Infected Women in the Study, HIV-uninfected Women From a 3 Cohorts and Women Diagnosed With ICC in the General Population (SEER) Are Shown for Comparison

Case Validation

There were 119 initially identified HIV-infected ICC cases and after review, a total of 67 ICC cases were validated. Most importantly, of the 30 that were considered potential incident cases, 17 were validated, whereas 9 were found to be cancer in situ, 3 were HSIL (cervical precancer), and 1 case had insufficient records for an ICC diagnosis to be validated—results that emphasize the importance of case validation given the known risk of record misclassification of cervical precancerous lesions as ICC (see Methods). Among the 17 confirmed HIV-infected ICC cases, the median time to diagnosis was 3.0 years (IQR: 2.2–5.9). Among the HIV-uninfected women, 4 ICC cases were confirmed (with diagnoses at 3.1, 4.0, 4.9, and 8.2 years).

Comparison of HIV-Infected Women With the General Population

SEER data were used as reference to estimate expected numbers of ICC cases and estimate SIRs by age strata. Overall, the SIR contrasting ICC incidence rates in HIV-infected NA-ACCORD women to expected rates based on SEER was 4.1 [95% confidence interval (CI): 2.3 to 6.6]. When stratified by age group, the SIR was 4.0 (95% CI: 1.3 to 9.3) for women <40 years and 5.0 (95% CI: 2.3 to 9.6) for women 40–49 years, whereas it was 2.3 (95% CI: 0.3 to 8.3) for women 50 and older. A comparison of the distributions of age at diagnosis between the observed cases in HIV-infected women and expected cases in the general population was not significant (P = 0.098).

Screening Data

Six of the 17 incident ICC cases (35%) in our study, each from a separate cohort, had no known Pap tests within the 5 years before diagnosis. However, 2 of these cases came from cohorts that reported women may seek screening outside the cohort and that information may not be captured. Five (29%) of the women who developed ICC had a history of recent Pap screening without detection of disease. This included 1 case with a normal Pap test 7 months before diagnosis, 1 case whose last Pap test was insufficient/uninterpretable, 2 with atypical squamous cells of undetermined significance, and 1 with low-grade SIL, all within a year of diagnosis. Six cases (35%) had HSIL detected through Pap screening an average of 3 years before their ICC diagnosis (range 2–4 years) without subsequent treatment. Four of these 6 women had prior HSIL detected through Pap had colposcopy, but none of the 6 received treatment. For 2 of these cases, notes were made in the medical files indicating that the patient had been referred for surgical excision but had missed scheduled appointments.

Cumulative ICC Risk Trends Among All Cases

The differences in lifetime ICC risk by HIV serostatus are shown in Figure 1A. There was an increasing disparity with increasing age in the cumulative incidence of ICC experienced by HIV-infected women compared with HIV-uninfected women or the general US female population, represented by SEER. The trends in ICC risk by CD4 in Figure 1B show that the risk groups differentiate by approximately age 45 and indicate the highest risk is among HIV-infected women with a CD4 <200 cells per microlite, whereas the lowest risk is among HIV-uninfected women in the participating cohorts. No clear differences in ICC risk could be readily discerned between HIV RNA strata; HIV-uninfected women continued to show the lowest risk of ICC.

Nonparametric estimation of the cumulative incidence of cervical cancer among all validated cases (prevalent and incident). A, Cumulative incidence of ICC per 100,000 pys, by time-updated age, in HIV-infected compared with HIV-uninfected women in NA-ACCORD and compared with the general US population sampled by SEER. B, Cumulative ICC per 100,000 pys by time-updated age, baseline HIV status, and CD4 cell count. C, Cumulative incidence of ICC per 100,000 pys by time-updated age, baseline HIV status, and HIV viral load.

Prospective Analysis

The crude ICC incidence rate among HIV-infected women was 26 per 100,000 pys (95% CI: 16 to 41). When stratified by age, the incidence rate was lower among HIV-infected women <40 years of age than those aged 40–49 years (18 vs 39 per 100,000 pys, P = 0.005), but similar to the rate among women ≥50 years and older (16 per 100,000 pys). The overall age-standardized incidence of ICC was higher among HIV-infected compared with -uninfected women in the study (16 vs 5 per 100,000, P = 0.03), although this difference was primarily in those younger than 50 years.

When stratified by CD4, women with lower CD4 had significantly higher ICC incidence (Ptrend = 0.003) regardless of whether CD4 was assessed at baseline, 18 months before the outcome assessment, or at the time of outcome assessment (Table 2). Although ICC incidence decreased with higher CD4, the rate among HIV-infected women with CD4 ≥350 cells per microliter was more than twice the rate of HIV-uninfected women (14 vs 6 per 100,000; Table 2). Figure 1B shows the increased cumulative ICC incidence among immunosuppressed women. ICC risk was associated with higher HIV viral load, although the association was weaker than that observed for CD4 cell count; no trend over calendar period was noted (results not shown).

Crude Incidence of ICC Among HIV-Infected Women in NA-ACCORD by CD4 Cell Count at Entry Into Study Cohorts (Baseline), at 18 Months Before Outcome Ascertainment or ICC Diagnosis, and at the Time of Outcome Ascertainment or ICC Diagnosis

In age-adjusted Poisson regression models with HIV-uninfected women as the referent, the incidence rate ratio of ICC was increased by 2.3, 3.0, and 7.7 times among HIV-infected women with baseline CD4 ≥350, 200–349, and <200 cells per microliter, respectively (Table 3). Results were similar whether CD4 at the time of ICC diagnosis or 18 months before diagnosis were considered.

Results From the Poisson Regression Model Assessing the Association of CD4 T-cell Count (CD4) With the Rate of Incident Cervical Cancer Adjusting for Age. The Effect of CD4 Was Evaluated Separately Using Measurements Obtained at 3 Time Points: Baseline CD4, CD4 at 18 Months Before Outcome Ascertainment or ICC Diagnosis, and CD4 at the Time of Outcome Ascertainment or ICC Diagnosis

Nested Case–Control Analyses

Among the 17 incident cases of ICC, 9 cases occurred in women who initiated ART during follow-up. Cases were diagnosed up to 6 years after ART initiation and the median time between ART initiation and diagnosis was 1.3 years. These 9 cases were individually matched to 103 controls from among the 7463 HIV-infected women without an ICC diagnosis who initiated ART therapy during follow-up. The nonparametric smoothed fit of the mean CD4 trajectories after ART initiation and associated 95% CIs are shown in Figure 2A. The estimates from the linear mixed model indicate nonsignificant differences in the slope of the CD4 trajectory in the first year after ART initiation (8 cells/μL per month gain for controls vs 6 cells/μL per month gain for cases) and from 1 to 3 years after ART initiation (6 cells/μL per month gain for controls vs 3 cells/μL per month gain for cases). However, by 3 years after ART initiation, a deviation in the case and control trajectories is noted from the nonparametric fit to the data, with cases exhibiting falling CD4 on average compared with controls (Fig. 2A); albeit, the observed smoothed trajectory shown in Figure 2A could not be modeled due to sparse data because most cases were diagnosed before 3 years after ART initiation. Furthermore, the suggested decline is consistent with the second matched analysis using all 17 cases, which indicated falling CD4 levels in cases before diagnosis (Figure 2B). From the linear mixed model, the slope in the controls was estimated to be flat (∼0 cells/μL per month) compared with a −2 cells per microliter per month estimated slope in the cases resulting in a substantial CD4 discrepancy at the time of diagnosis of −185 cells per microliter (P = 0.01).

CD4 T-cell count mean trajectories and 95% CIs over time among incident ICC cases (dashed) and matched controls (solid) from NA-ACCORD cohorts. A, Trajectories of cases and controls matched at ART initiation on CD4 T-cell count, age, cohort, year of ART initiation, and time subject followed after ART initiation. B, Trajectories of cases and controls matched at ICC diagnosis on ART use, age, and calendar year.

Sensitivity Analyses

In the main analyses, we limited cases to those women diagnosed 6 or more months after cohort enrollment to reduce the likelihood of including prevalent but undiagnosed cases. However, there is still the possibility that prevalent cases were included in the analysis. Therefore, we reran the analysis using the 14 women diagnosed 18 months or more after cohort enrollment. Results were consistent in terms of the magnitude and direction of effect estimates compared with the main analysis. The estimated relative rates for the CD4 strata <200, 200–350, and >350 using CD4 lagged 18 months and excluding cases during the first 18 months were 9.5, 5.3, and 2.0, respectively, compared with 9.2, 5.1, and 1.9, respectively, from the primary analysis that excluded cases during the first 6 months (Table 3).


In this multicohort analysis of incident ICC, HIV-infected women had significantly higher risk of incident ICC than HIV-uninfected women, and the risk of ICC increased significantly with diminishing immune status as measured by CD4 count. Furthermore, women who developed ICC after initiation of ART were characterized by declining CD4 counts before diagnosis, which was not observed in noncases. The increased burden of ICC may persist in HIV-infected women even at higher CD4 counts because HIV-infected women with CD4 counts ≥350 cells per microliter still experienced significantly higher rates of ICC than the general population or HIV-uninfected women in these cohorts. Nevertheless, these results suggest that the use of ART to maintain CD4 above 350 cells per microliter may reduce ICC risk.

As in the general population where ICC risk increases with age, we noted a doubling of the incidence of ICC in HIV-infected women comparing those 40–49 years to those younger than 40 years. However, the rate of ICC decreased in HIV-infected women 50 years and older. After accounting for the different age structure in the HIV-infected and general populations, the SIRs suggested an elevated ICC incidence among HIV-infected women in the younger age groups relative to the general population, though the difference in median age at diagnosis was not significant, likely due to the small sample size. Competing risks that preferentially remove HIV-infected women at highest risk of ICC from the risk set, which are not accounted for in the present analysis, may explain the noted differences in SIRs between age strata. Alternatively, cervical screening at earlier stages of ICC may result in diagnosis of ICC at earlier ages. Cancer staging information was not available for this study, and thus, we could not access whether HIV-infected women tended to be diagnosed at earlier disease stages. No other studies that we are aware of have examined the question of age-at-ICC diagnosis among HIV-infected women, and further research is warranted to determine whether HIV-infected ICC cases are more likely to occur at younger ages relative to the general population. Addressing this question could also shed light on the etiologic underpinnings of the increased risk of ICC among HIV-infected women by implicating a more rapid progression of cervical disease as a result of immunosuppression.

Given that regular screening and treatment for precancerous lesions are thought to prevent most ICC cases,22–24 it is notable that 6 incident cases had no evidence of Pap screening within the 5 years before diagnosis. Some of these women may not have been engaged in regular care, or some may have been receiving HIV specialty care only, with gynecologic screening provided elsewhere. HIV-uninfected women in our study had a relatively low rate of ICC—lower than that observed in the general (SEER) population. Most of the HIV-uninfected women in our study were women in a primary care setting, and their low cancer rate may speak to the effectiveness of regular cervical cancer screening for preventing ICC.

Although Pap smear results are specific, they have only moderate sensitivity and false-negative results occur—a concern that is partially alleviated through repeated testing.25 Only 1 case was identified that had a Pap test within 2 years of diagnosis indicating normal cytology. The majority of cases (9 of 17) had recent Paps indicating atypical squamous cells of undetermined significance (n = 2), low-grade SIL (n = 1), or HSIL (n = 6), representing potentially preventable disease if colposcopy and appropriate therapy had occurred. Barriers to colposcopy and treatment after abnormal Pap have been extensively studied,26–33 although only a few studies have focused on HIV-infected women.34,35

This study benefited from extensive follow-up of a large population of HIV-infected women enrolled in formal cohort studies with longitudinal information on CD4, screening, and treatment. The results were strengthened by a rigorous validation process for the cases assuring high specificity of the case definition. Although our study had fewer cases than the study by Guiguet et al,17 the difference between the final validated cases and those initially identified in the current study is consistent with other reports of the high degree of misclassification that can result from relying on less rigorous case definitions such as diagnostic codes like ICD10. The reported ICC incidence in the HIV-infected women in Guiguet et al17 was notably higher than that observed in our study (6900 vs 26 per 100,000), despite a higher median current CD4 cell count than cases in our study (287 vs 178 cells/μl). However, their estimate included cases diagnosed within the first 6 months of follow-up.

Overall, the data from this large prospective study of ICC in HIV-infected women suggest that maintaining CD4 at higher counts could lower ICC risk. Although prevention of ICC may not alone provide an adequate indication for the early initiation of ART,36 our data provide important information regarding the impact of host immunity on ICC in HIV-infected women because they continue to live longer through ART. Cervical cancer screening is important for preventing progression to ICC,37 and further research is warranted on the barriers among HIV-infected women to seeking treatment after abnormal Paps.


NA-ACCORD participating cohorts and representatives (*indicates contributed data to this study): *AIDS Link to the IntraVenous Experience (ALIVE; G.K.), *Adult AIDS Clinical Trials Group Longitudinal Linked Randomized Trials (ALLRT; C. Benson, R.B., and A. Collier), *HIV Research Network (HIVRN; K.G.), *HAART Observational Medical Evaluation and Research (HOMER; R.H.), *HIV Outpatient Study (HOPS; J.T.B.), *Johns Hopkins HIV Clinical Cohort (JHHCC; R.D.M.), *Kaiser Permanente Northern California (M. Horberg and M.S.), *Multicenter Hemophilia Cohort Study–II (MHCS-II; J.J.G.) and Multicenter AIDS Cohort Study (MACS; L. Jacobson), *Montreal Chest Institute Immunodeficiency Service Cohort (M.K.), *Ontario HIV Treatment Network Cohort Study (OHTN; S.R. and A.R.), *John T. Carey Special Immunology Unit Patient Care and Research Database, Case Western Reserve University (PCRD; B. Rodriguez) and Polaris HIV Seroconversion Study (L. Calzavara), *Southern Alberta Clinic Cohort (M.J.G.). Studies of the Consequences of the Protease Inhibitor Era (SCOPE; S. Deeks and J. Martin), *University of Alabama at Birmingham Clinic Cohort (M. Saag), *University of North Carolina, Chapel Hill HIV Clinic Cohort (UCHCC; J.J.E. and S.N.), *University of Washington HIV Cohort (M.K.), *Veterans Aging Cohort Study (VACS; A. Justice), *Vanderbilt-Meharry CFAR Cohort (T.R.S.), *Women's Interagency HIV Study (K. Anastos and S.J.G). Executive Committee: R. Moore, M. Saag, S. Gange, M. Kitahata, R. McKaig (NIH), and A. Freeman; Epidemiology/Biostatistics Core: A. Abraham, K. Althoff, E. Golub, Y. Jing, B. Lau, S. Modur, and J. Zhang; Data Management Core: S. Van Rompaey, E. Webster, and B. Simon; and Administrative Core: C. Lent.


Table 1 ICC Incidence Among the HIV-Infected and -Uninfected Women by Contributing Cohort


1. Strickler HD, Burk RD, Fazzari M, et al.. Natural history and possible reactivation of human papillomavirus in human immunodeficiency virus-positive women. J Natl Cancer Inst. 2005;97:577–586.
2. Ahdieh L, Munoz A, Vlahov D, et al.. Cervical neoplasia and repeated positivity of human papillomavirus infection in human immunodeficiency virus-seropositive and -seronegative women. Am J Epidemiol. 2000;151:1148–1157.
3. Ahdieh L, Klein RS, Burk R, et al.. Prevalence, incidence, and type-specific persistence of human papillomavirus in human immunodeficiency virus (HIV)-positive and HIV-negative women. J Infect Dis. 2001;184:682–690.
4. Sun XW, Kuhn L, Ellerbrock TV, et al.. Human papillomavirus infection in women infected with the human immunodeficiency virus. N Engl J Med. 1997;337:1343–1349.
5. Massad LS, Ahdieh L, Benning L, et al.. Evolution of cervical abnormalities among women with HIV-1: evidence from surveillance cytology in the women's interagency HIV study. J Acquir Immune Defic Syndr. 2001;27:432–442.
6. Six C, Heard I, Bergeron C, et al.. Comparative prevalence, incidence and short-term prognosis of cervical squamous intraepithelial lesions amongst HIV-positive and HIV-negative women. AIDS. 1998;12:1047–1056.
7. Harris TG, Burk RD, Palefsky JM, et al.. Incidence of cervical squamous intraepithelial lesions associated with HIV serostatus, CD4 cell counts, and human papillomavirus test results. JAMA. 2005;293:1471–1476.
8. Strickler HD, Burk RD, Fazzari M, et al.. Natural history and possible reactivation of human papillomavirus in human immunodeficiency virus-positive women. J Natl Cancer Inst. 2005;97:577–586.
9. Harris TG, Burk RD, Palefsky JM, et al.. Incidence of cervical squamous intraepithelial lesions associated with HIV serostatus, CD4 cell counts, and human papillomavirus test results. JAMA. 2005;293:1471–1476.
10. Strickler HD, Palefsky JM, Shah KV, et al.. Human papillomavirus type 16 and immune status in human immunodeficiency virus-seropositive women. J Natl Cancer Inst. 2003;95:1062–1071.
11. Strickler HD. Does HIV/AIDS have a biological impact on the risk of human papillomavirus-related cancers? J Natl Cancer Inst. 2009;101:1103–1105.
12. Castro KG, Ward JW, Slutsker L, et al.. Revised Classification System for HIV Infection and Expanded Surveillance Case Definition for AIDS Among Adolescents and Adults. Atlanta, GA: Centers for Disease Control and Prevention; 1993. Morbidity and Mortality Weekly Report June 17, 2012.
13. Biggar RJ, Chaturvedi AK, Goedert JJ, et al.. AIDS-related cancer and severity of immunosuppression in persons with AIDS. J Natl Cancer Inst. 2007;99:962–972.
14. Chaturvedi AK, Madeleine MM, Biggar RJ, et al.. Risk of human papillomavirus-associated cancers among persons with AIDS. J Natl Cancer Inst. 2009;101:1120–1130.
15. Clifford GM, Polesel J, Rickenbach M, et al.. Cancer risk in the Swiss HIV Cohort Study: associations with immunodeficiency, smoking, and highly active antiretroviral therapy. J Natl Cancer Inst. 2005;97:425–432.
16. Patel P, Hanson DL, Sullivan PS, et al.. Incidence of types of cancer among HIV-infected persons compared with the general population in the United States, 1992-2003. Ann Intern Med. 2008;148:728–736.
17. Guiguet M, Boue F, Cadranel J, et al.. Effect of immunodeficiency, HIV viral load, and antiretroviral therapy on the risk of individual malignancies (FHDH-ANRS CO4): a prospective cohort study. Lancet Oncol. 2009;10:1152–1159.
18. Gange SJ, Kitahata MM, Saag MS, et al.. Cohort profile: the North American AIDS cohort Collaboration on research and Design (NA-ACCORD). Int J Epidemiol. 2007;36:294–301.
19. Franco EL, Rohan TE. Assessing Epidemiologic Relations and the Role of Measurement Error. Cancer Precursors: Epidemiology, Detection and Prevention. New York: Springer-Verlag; 2002.
20. Surveillance Epidemiology and End Result (SEER) Program, SEER*Stat Database. Age-Adjusted SEER Incidence Rates, Cancer of the Cervix Uteri (Invasive)—SEER 13 Regs Research Data, Nov 2010 Sub (1992–2008) With Katrina/Rita Adjustment, Linked to County Attributes, Total US, 1969-2009 Counties National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2011 based on November 2010 submission. 2010. Available at: Accessed March 31, 2012.
21. Shiels MS, Pfeiffer RM, Engels EA. Age at cancer diagnosis among persons with AIDS in the United States. Ann Intern Med. 2010;153:452–460.
22. Kaplan JE, Benson C, Holmes KH, et al.. Guidelines for prevention and treatment of opportunistic infections in HIV-infected adults and adolescents: recommendations from CDC, the National Institutes of Health, and the HIV Medicine Association of the Infectious Diseases Society of America. MMWR Recomm Rep. 2009;58:1–207.
23. Saslow D, Runowicz CD, Solomon D, et al.. American Cancer Society Guideline for the early detection of cervical neoplasia and cancer. J Low Genit Tract Dis. 2003;7:67–86.
24. U.S. Preventive Services Task Force. Screening for Cervical Cancer. 2012. Rockville, MD: U.S. Preventive Services Task Force. Available at: Accessed March 1, 2012.
25. National Cancer Institute. PDQ® Cervical Cancer Screening. Bethesda, MD: National Cancer Institute. Available at: Accessed February 5, 2012. 2012.
26. Eggleston KS, Coker AL, Das IP, et al.. Understanding barriers for adherence to follow-up care for abnormal pap tests. J Womens Health (Larchmt). 2007;16:311–330.
27. Battaglia TA, Santana MC, Bak S, et al.. Predictors of timely follow-up after abnormal cancer screening among women seeking care at urban community health centers. Cancer. 2010;116:913–921.
28. Kuo TM, Benard VB, Berkman ND, et al.. Timing of colposcopy after cervical cytologic abnormalities. Obstet Gynecol. 2010;115:629–636.
29. Benard VB, Lawson HW, Eheman CR, et al.. Adherence to guidelines for follow-up of low-grade cytologic abnormalities among medically underserved women. Obstet Gynecol. 2005;105:1323–1328.
30. Coker AL, Eggleston KS, Meyer TE, et al.. What predicts adherence to follow-up recommendations for abnormal Pap tests among older women? Gynecol Oncol. 2007;105:74–80.
31. McKee MD, Lurio J, Marantz P, et al.. Barriers to follow-up of abnormal Papanicolaou smears in an urban community health center. Arch Fam Med. 1999;8:129–134.
32. Melnikow J, Chan BK, Stewart GK. Do follow-up recommendations for abnormal Papanicolaou smears influence patient adherence? Arch Fam Med. 1999;8:510–514.
33. Peterson NB, Han J, Freund KM. Inadequate follow-up for abnormal Pap smears in an urban population. J Natl Med Assoc. 2003;95:825–832.
34. Cejtin HE, Komaroff E, Massad LS, et al.. Adherence to colposcopy among women with HIV infection. J Acquir Immune Defic Syndr. 1999;22:247–252.
35. Baranoski AS, Stier EA. Factors associated with time to colposcopy after abnormal pap testing in HIV-infected women. J Womens Health (Larchmt). 2012;21:418–424.
36. Minkoff H, Zhong Y, Burk RD, et al.. Influence of adherent and effective antiretroviral therapy use on human papillomavirus infection and squamous intraepithelial lesions in human immunodeficiency virus-positive women. J Infect Dis. 2010;201:681–690.
37. Massad LS, Fazzari MJ, Anastos K, et al.. Outcomes after treatment of cervical intraepithelial neoplasia among women with HIV. J Low Genit Tract Dis. 2007;11:90–97.

human papillomavirus; HIV-infection; invasive cervical cancer; immunosuppression

© 2013 Lippincott Williams & Wilkins, Inc.