Abstract
Objective Women with systemic lupus erythematosus (SLE) are vulnerable to cervical dysplasia due to the persistence of human papillomavirus (HPV) infection. The objective of this cross-sectional retrospective study was to investigate the prevalence of cervical cancer screening according to the American Society for Colposcopy and Cervical Pathology (ASCCP) SLE-specific cervical cancer screening guidelines. We also aimed to identify SLE-specific determinants associated with ASCCP adherence.
Methods Women aged 21 to 64 years enrolled in our institutional SLE registry were included in the study. The electronic medical record was manually reviewed to determine whether the patient was up to date on screening and which organizational guideline was used, in addition to other clinical variables. Multivariable logistic regression was used to estimate adjusted odds ratios (ORs) for ASCCP-congruent screening for each baseline characteristic.
Results This study included 118 women with SLE; 38% were up to date per ASCCP guidelines, 16% were up to date per non-ASCCP guidelines, and 46% were overdue for screening. Having a gynecologist and being actively treated with immunosuppressant therapies were both associated with an increased odds of being up to date per the ASCCP guidelines, while Hispanic ethnicity was associated with reduced odds.
Conclusion Only half of the women with SLE in our study had guideline-congruent cervical cancer screening. Current immunosuppression exposure, rather than SLE disease activity, was associated with an increased odds of being up to date according to ASCCP guidelines. This study suggests the need for increased awareness and consensus among interdisciplinary providers regarding SLE-specific cervical cancer screening.
Cervical cancer is the fourth most common cancer among women and has an estimated mortality rate of 90%.1 Human papillomavirus (HPV), a major driver of invasive cervical cancer, is a common infection that is particularly problematic for immunocompromised individuals due to its persistence in this population, leading to increased risk of malignant transformation. Because HPV is prevalent in the United States and is strongly linked to the development of invasive cervical cancer, it is important that immunocompromised individuals be screened regularly.
HPV persistence and malignant transformation in patients with systemic lupus erythematosus (SLE) has been well studied, and it is now accepted that SLE is an independent risk factor for the development of cervical intraepithelial neoplasia (CIN).2-5 While SLE has not been found to be associated with increased incidence of invasive cervical cancer, this observation may be a result of regular screening practices and early treatment of CIN lesions when discovered.6 Importantly, SLE affects young women and disease severity is higher in ethnic minorities, which is particularly concerning since ethnic minorities are also most at risk for cervical cancer. As of the 2020 US cancer statistics report produced by the Centers for Disease Control and Prevention (CDC), Hispanic and Black women ages 35 to 59 years are the most likely demographic group to be diagnosed with cervical cancer.7
Despite the consensus regarding the importance of routine cervical cancer screening, discordant organizational recommendations providing different cervical cancer screening schedules exist (Figure). Of these organizations, the American Society of Colposcopy and Cervical Pathology (ASCCP) provides the only SLE-specific guidelines for cervical cancer screening schedules. These recommendations, which are derived from recommendations for women who are HIV positive, are more proactive, supporting annual cytology testing after sexual debut or starting at age 21 years, whichever is earlier, then extending interval cytology or co-testing (cytology and genotyping) screening to every 3 years after 3 consecutive negative cytology tests.8
Several studies suggest that women with rheumatic diseases do not receive optimal cancer screening tests and other preventative medical services; this is problematic because this patient population is routinely exposed to immunosuppression that is associated with an increased risk of virally induced cancers.9-12 A variety of factors might explain this; patients with SLE often have several specialty physicians as part of their care team, yet the rheumatologist may serve in a primary care role and may be unaware of SLE-specific published guidelines, specifically on cervical cancer screening. Gynecologists and primary care providers (PCPs) routinely perform cervical cancer screening yet may overlook or lack familiarity in recognizing SLE when risk-stratifying patients.
To investigate the cervical cancer screening practice patterns in our institution, we performed a cross-sectional study using our SLE registry program to identify the adherence to the ASCCP schedule. We also aimed to identify disease-specific determinants associated with ASCCP adherence compared to those overdue for screening or to those screened according to other organizational guidelines intended for the general population.
METHODS
Patient selection. This was an exploratory cross-sectional study using available clinical data from patients enrolled in the University of Washington (UW) SLE biorepository program. This program is a combined longitudinal biorepository and registry, collecting patient serum samples for translational research, accompanied by clinical data on disease characteristics, activity, and treatment at the time of sample collection. Inclusion criteria were women fulfilling either the 2012 Systemic Lupus International Collaborating Clinics (SLICC) criteria or the 2019 European Alliance of Associations for Rheumatology/American College of Rheumatology (EULAR/ACR) SLICC criteria for SLE, aged between 21 and 64 years, and rheumatologic care within our university system over a 5-year period (from January 1, 2016, to January 1, 2021). Exclusion criteria were patients with SLE overlap syndromes and those with an incomplete documented Systemic Lupus Erythematosus Disease Activity Index 2000 (SLEDAI-2K) score, an incomplete SLICC/ACR Damage Index score (SDI), or missing data for > 3 variables. Ethical approval was granted by the UW Institutional Review Board (approval no. 00003007).
Variable extraction. The following variables were extracted from the biorepository database: patient demographics (age, race/ethnicity, educational status, insurance status), smoking status, SLE disease characteristics (date of diagnosis, 2012 or 2019 EULAR/ACR SLICC criteria met, SDI score at time of visit), comorbidities (diabetes, cardiovascular disease [CVD], kidney disease), and current and any documented prior medication exposure during the patient lifetime. Conventional disease-modifying antirheumatic drugs (cDMARDs) were defined by the use of methotrexate (MTX), leflunomide, or sulfasalazine; immunosuppressive (IS) medications were defined by the use of azathioprine (AZA), mycophenolate mofetil (MMF), or cyclophosphamide (CYC); and biologic medications were defined by the use of belimumab, rituximab (RTX), or abatacept (ABA).
Chart review of the electronic medical record by 2 independent reviewers was performed to extract additional variables regarding healthcare factors (establishment of PCP and/or gynecologist, duration of follow-up with current rheumatologist, and gender of rheumatologist) and gynecologic history (sexual activity, personal history of cervical cancer, personal history of cervical dysplasia, HPV vaccination status, prior history of sexually-transmitted infection [STI], current HIV status, and contraceptive use at time of visit). All available cytology, HPV genotype testing, colposcopy reports, and cervical tissue pathology reports were extracted and reviewed. Patients were categorized based on the organization guideline that best matched their cervical cancer screening schedule: ASCCP, US Preventative Services Task Force (USPSTF), American College of Obstetrics and Gynecology (ACOG), or the American Cancer Society (ACS; Figure). We also used any available narrative from provider notes to assist in the categorization of organizational guideline screening per patient.
Statistical analysis. Prevalence of baseline patient characteristics among guideline strata was summarized using descriptive statistics: frequency and percentage for categorical variables, and median and IQR for continuous variables. For statistical analyses of the association between baseline patient characteristics and degree of guideline congruence, the categories ACOG and Overdue were combined into a single category relabeled “Non-ASCCP,” signifying that the ASCCP guidelines were not met (patients who met the USPSTF and ACS criteria by definition also met ACOG criteria and had thus been included in the ACOG group; Figure). Univariable logistic regression was used to estimate the crude odds ratio (OR) for ASCCP-congruent screening for each baseline characteristic, with 95% CIs and significance tests of the null hypothesis of no association (ie, OR 1). Univariable results, combined with clinical judgment, were used to identify a set of characteristics for multivariable modeling and included Hispanic ethnicity, gynecologist (yes/no), history of nephritis (yes/no), SLEDAI-2K total score, current corticosteroid (CS) use (yes/no), and current IS and/or biologic use (yes/no). Multivariable logistic regression was used to estimate adjusted ORs for ASCCP-congruent screening for each baseline characteristic, with 95% CIs and significance tests of the null hypothesis of no association (ie, OR 1). P values < 0.05 were considered statistically significant. All analyses were conducted in Stata/SE Version 16 for Windows (StataCorp).
RESULTS
A total of 130 women with SLE met our eligibility criteria. There were 12 women for whom screening status could not be determined from the data available and who were excluded, resulting in a final sample of 118 women. The median age was 37.1 years (IQR 16) and 47% self-reported race as non-White.
Of those who met the eligibility criteria, 95% met the 2019 EULAR/ACR SLICC classification criteria for SLE and 5% met the 2012 SLICC classification criteria for SLE. Mean SLEDAI-2K score was 5.5 (SD 5.3). Roughly one-third (36%) of the cohort had a history of biopsy-proven nephritis. Ninety percent were on hydroxychloroquine therapy, 44% were on oral CS (equivalent to prednisone ≥ 5 mg daily), and 5% were receiving biologic therapy (RTX or belimumab). MMF (39%) and MTX (20%) were the most common CS-sparing agents used in these patients.
The majority of patients had a PCP (94%); only a third (35%) had a gynecology provider. Fifty-eight percent of women had been followed by the same rheumatologist for ≥ 1 year, and there was 50% gender congruency between the patient and rheumatology provider.
Table 1 shows the baseline patient and disease factors stratified by guideline-congruent cervical cancer screening. Thirty-eight percent of the cohort was up to date on cervical cancer screening per ASCCP guidelines. Sixteen percent were up to date per non-ASCCP guidelines for the general population, and 46% were overdue according to any organizational guideline.
Patients without a gynecologist were more likely to be overdue for screening according to any guideline vs screened per ASCCP or ACOG guidelines (P < 0.001; Table 1). Additionally, patients of Hispanic ethnicity tended to be overdue for screening (35%, P = 0.04). Median SLEDAI-2K and SDI scores were similar among the groups; however, a higher proportion (67%) of women screened per ASCCP guidelines were on IS therapy compared to cDMARDs (16%) and biologics (7%). Women who were screened per ASCCP guidelines also had a higher proportion of having a gynecologist (51%), being a current or former tobacco smoker (31%), having had known CVD (20%), and having had documented completion of the HPV vaccination series (73%).
Univariable analyses showed that Hispanic ethnicity was associated with a reduced odds (OR 0.34, 95% CI 0.12-0.91) of being screened per ASCCP guidelines (Table 2). Having a gynecologist (OR 3.19, 95% CI 1.44-7.00) and being on CS or IS therapy was associated with an increased odds (CS OR 2.13, 95% CI 1.00-4.53; IS therapy OR 3.03, 95% CI 1.40-6.60) of being screened per ASCCP guidelines.
Results of the multivariable analyses are shown in Table 3. Hispanic ethnicity was associated with a reduced odds (OR 0.34, 95% CI 0.11-1.01) of being screened per ASCCP guidelines. Having a gynecologist was associated with 3-fold higher odds (OR 3.26, 95% CI 1.30-8.15) of being screened per ASCCP guidelines. Being on IS and/or biologic therapy was associated with 3-fold higher odds (OR 3.31, 95% CI 1.34-8.17) of being screened per ASCCP guidelines. SLE disease activity, damage indices, and history of renal involvement were not found to be associated with ASCCP guideline adherence.
DISCUSSION
We observed that among patients with SLE enrolled in our registry program, only half (54%) were up to date on cervical cancer screening at the time of their last rheumatology visit according to any organizational guideline, and only 38% were up to date per ASCCP SLE-specific guidelines. These outcomes reflect published findings that suggest women with SLE are not receiving adequate cervical cancer screening.9-12 Additionally, of those screened, roughly a third (29%) of patients with SLE were screened per non-ASCCP organizational guidelines meant for the general population.
The reduced odds of cervical cancer screening among patients of Hispanic ethnicity are alarming, particularly given data that suggest increased severity and mortality due to SLE factors alone in this ethnic group.13,14 This finding was not influenced by insurance status or highest educational level in the multivariate analysis. Additional measures of socioeconomic factors, such as employment status or number of dependents, were not available and thus could not be accounted for in our analyses.
In this cohort, current use of IS and biologic medications (namely CYC, MMF, AZA, RTX, belimumab, and ABA) was associated with 3-fold higher odds of being up to date per ASCCP guidelines, thus suggesting that immunosuppression, rather than disease activity or characteristics, may be prompting providers, largely gynecologists, to use ASCCP guidelines for patients with SLE.
We had hypothesized that patients with high disease activity as measured by SLEDAI-2K scores might have an increased odds of more aggressive cervical cancer screening, yet this was not found to be the case in this study. In fact, SLE disease activity, damage indices, and history of renal involvement were not found to be associated with ASCCP guideline adherence. The mean SLEDAI-2K and SDI scores of the cohort correlate with low-to-moderate disease activity and little irreversible damage from SLE disease and treatment. It is possible that these scores were suppressed due to concomitant IS therapy.
There is a dearth of publications that examine cervical cancer screening practices among patients with SLE on immunosuppresion. One recent study by Bruera et al used a large claims database over a 13-year period, showing that cervical cancer screening rates were higher in women with SLE (73.3%) compared to a general population cohort (58.5%) and a diabetes mellitus cohort (56.2%).10 The use of claims data is limited to patients with private employer-sponsored insurance; however, the study still found that over 25% of patients with SLE had not had a diagnosis or procedure code for cytology or HPV testing within 1 year prior to and 2 years post SLE claim. CS and IS use were associated with a reduced odds of having had cytology or HPV testing at time of SLE claim. This contrasts to our findings, which show an increased likelihood of screening frequency in patients on IS therapy.
The ASCCP SLE-specific guidelines were formally published in 2019, yet we found that patients followed screening schedules that aligned with these guidelines even prior to 2019. We suspect that the 2012 ASCCP guidelines, which were the first to provide risk-stratified recommendations outlining more frequent screening for immunosuppressed individuals (defined by history of solid organ transplant and HIV-positive status), continued to influence gynecology providers. This would correspond well with the finding that the likelihood of ASCCP guideline adherence was increased with current IS use, rather than SLE disease-specific factors. By using a study period from 2016 to 2021, it is possible that our data might mispresent low screening rates of ASCCP. However, the majority of screening status data points were taken from visits 2019 and onward (ASCCP 33/45 [73%]; non-ASCCP 47/73 [64%]).
Limitations of this study include the small sample size as well as the potential for selection bias as a result of using a well-curated cohort. These patients with SLE tend to be engaged in rheumatologic care, with visit frequency averaging every 3 months. We did not extract other variables regarding socioeconomic status beyond education and insurance status. We did not include pregnancy history in the analyses, which could influence the exposure to obstetrics/gynecology care. While we did include data regarding prior medication exposure, we did not calculate total lifetime dose exposures to cDMARD, IS, or biologic agents, acknowledging that cumulative exposure to immunomodulatory agents likely increases the risk for cervical dysplasia. Since this was an exploratory study, we did not compare this SLE cohort with a control cohort (general population) or a cohort with an analogous immunosuppressive risk (ie, solid organ transplant or HIV). Such studies examining screening practice patterns among a larger, more diverse cohort of patients with SLE should be performed to reiterate these findings and identify barriers to screening in this patient population.
Strengths of this study include the granularity of the data available regarding SLE disease factors in women with consistent rheumatologic care as well as access to SLEDAI-2K and SDI scores, which provide an individualized context of disease activity and burden when considering cervical cancer screening status. Manual chart review to determine screening status allowed for accurate documentation of which organizational screening pattern was met. Since a variety of potential outcomes are possible within a patient’s lifetime (ie, transition to dysplasia, need for colposcopy, subsequent HPV genotype testing, resolution or progression of dysplasia), the measured outcome (which organizational screening pattern was met) likely could be collected only in this manner.
This study demonstrates the inadequate rate of disease-specific cervical cancer screening, even in an academic setting, and suggests that it is the immunosuppression exposure, rather than disease activity markers, that is associated with an increased odds of being screened per ASCCP guidelines. This relatively poor rate of adherence to screening guidelines is particularly important given the recent approval and adoption of primary HPV genotype screening, which is increasingly promoted in the general population but inadequate for patients with SLE. These findings suggest the need for interventions to increase awareness and access to collaborative care among rheumatologists, PCPs, and gynecologists regarding the ASCCP SLE-specific screening guideline schedule to ultimately enhance the quality of care of patients.
Footnotes
This work was supported by funding from the Rheumatology Research Foundation to Dr. Namrata Singh and the National Center for Advancing Translational Sciences of the National Institutes of Health (NIH; award number UL1 TR002319). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
The authors declare no conflicts of interest relevant to this article.
- Accepted for publication August 2, 2022.
- Copyright © 2022 by the Journal of Rheumatology