Abstract
Objective To determine the incidence of breast, cervical, prostate, and colorectal cancer screening in patients with rheumatoid arthritis (RA) vs matched non-RA comparators.
Methods We performed a retrospective, matched cohort study of patients with and without RA living in an 8-county region of southern Minnesota on January 1, 2015. Through review of medical records, patients who fulfilled either the 1987 American College of Rheumatology (ACR) or 2010 ACR/European Alliance of Associations for Rheumatology classification criteria for RA were identified. Patients with RA were matched 1:1 to non-RA comparators on age, sex, and county of residence. Cancer screening was determined from review of the US Preventative Task Force recommendations. Cumulative incidence of cancer screening was estimated accounting for the competing risk of death, and Cox proportional hazard models adjusted for age, smoking, and race assessed for the risk of delay.
Results The study included 1614 patients with RA and 1597 comparators without RA (mean age 63 years, 71% female). At 5-years of follow-up, 51.6% (95% CI 47.9-55.6) of the RA cohort had cervical cancer screening compared to 58.2% (95% CI 54.5-62.2) in the non-RA cohort. After adjusting for age, smoking, and race, RA was associated with decreased cervical cancer screening (adjusted hazard ratio [aHR] 0.83, 95% CI 0.72-0.96). RA was not significantly associated with a decrease in breast (aHR 0.98, 95% CI 0.87-1.10), prostate (aHR 0.99, 95% CI 0.74-1.34), or colorectal (aHR 1.04, 95% CI 0.93-1.16) cancer screening.
Conclusion Women with RA were more likely to experience delayed cervical cancer screening. Increased diligence by healthcare providers to ensure cervical cancer screening in patients with RA is important to reduce the morbidity and mortality seen in these patients.
Rheumatoid arthritis (RA) is a systemic inflammatory autoimmune disease with a profound impact on morbidity and mortality in patients.1 Studies show that patients with RA have an increased mortality risk ranging between 30% and 70% higher than those without RA,2 and a previous study demonstrated that the leading causes of mortality are cardiovascular diseases, cancer, and respiratory diseases.3 Metaanalyses demonstrated that patients with RA have an increased risk of malignancy compared to the general population.4-6 However, colorectal and breast cancer risk are often decreased, and cervical and prostate cancer do not show a consistent trend in risk.4,5
The increased risk of malignancy in patients with RA is partially due to autoimmunity, chronic inflammation, and smoking.7,8 There is a concern that the immunosuppression used in the treatment of RA further increases the risk of malignancies. However, a recent systematic literature review (SLR) including patients with inflammatory arthritis (IA) who were prescribed immunosuppression and had a history of cancer did not find an increased risk of a new primary cancer or cancer reoccurrence in comparison to patients with IA who received a different treatment given their history of cancer.9 This SLR aided informing the 2024 European Alliance of Associations for Rheumatology (EULAR) recommendations for initiation of targeted therapies in patients with both IA and a history of cancer. Notable, themes were the importance of treating active IA even with a history of cancer, and that a history of cancer that is currently in remission should not delay treatment of an IA.10
Previous studies have explored the use of cancer screenings in this population. One previous study found that patients with RA had no difference in cervical, breast, or colorectal cancer screening compared to the general population.11 However, whereas a study from Canada also showed there were no significant differences between RA and non-RA comparators in screening for breast and colorectal cancer, patients with RA had decreased cervical cancer screening.12 Further, a separate study identified that patients with RA undergo less screening for breast and cervical cancer screening.13
Understanding the use of cancer screenings will inform providers whether patients with RA are at an increased risk of delay in cancer screenings. The primary objective of this study was to determine the cumulative incidence of breast, cervical, prostate, and colorectal cancer screening in patients with RA compared to matched non-RA comparators. The second objective of this study was to determine the influence of demographics and health behaviors on cancer screening. We hypothesized that individuals with RA would have a decreased cumulative incidence of cancer screening for breast, cervical, prostate, and colorectal cancers compared to matched non-RA comparators, given that patients with RA can have a suboptimal frequency of visits to their primary care providers and rheumatologists, and previous studies have demonstrated decreased cancer screening.12-14
METHODS
Study design. We conducted a retrospective matched cohort study of patients with and without RA within the Rochester Epidemiology Project (REP). The REP is a large record linkage system that provides the medical records of residents in a 27-county region in Minnesota and Wisconsin who receive care at outpatient offices, urgent care offices, emergency departments, and hospitals within the area. This study included patients from 8 Minnesota counties: Olmsted, Dodge, Mower, Goodhue, Wabasha, Freeborn, Steele, and Waseca. The demographics of patients within the REP are similar to that of the Upper Midwest United States.15
The patients in both cohorts were followed from the cohort entry date (January 1, 2015) to the end of the study period (February 29, 2020), emigration out of the geographical area, or death. Additional analyses followed patients from March 1, 2020, to September 30, 2023, to assess the effect of the coronavirus 2019 (COVID-19) pandemic. For colorectal cancer screening, patients were followed from January 15, 2015, to September 30, 2023, due to the extended length of time recommended between cancer screenings.
Patient selection and matching. Subjects in the RA cohort were identified by nurse abstractors who reviewed the medical records of each potential patient. Those in the RA cohort fulfilled either the 1987 American College of Rheumatology (ACR) classification criteria or the 2010 ACR/EULAR classification criteria for RA.16 Subjects were aged ≥ 18 years and resided in 1 of the 8 Minnesota counties of interest on January 1, 2015. If a subject moved to the 8-county area following the diagnosis of RA, they were included in the RA cohort if prescribed a disease-modifying antirheumatic drug (DMARD) and had a physician diagnosis of RA.
All subjects within the RA cohort were matched 1:1 by sex, age, and county to non-RA comparators who were randomly selected from residents of the 8 counties included in this study. Subjects were subsequently excluded from either cohort if there was diagnosis of the relevant cancer within the previous 5 years, no documented medical history 1 year prior to January 1, 2015, or no documented follow-up. Subjects with a complete bilateral mastectomy were excluded from the breast cancer screening analyses.
Study outcome. The primary outcome was the cumulative incidence of screening for breast, cervical, prostate, and colorectal cancer as determined through electronic retrievals of diagnoses, procedures, laboratory test results, and healthcare services. For each screening analysis, patients with and without RA were included if both their age and sex were within the US Preventative Services Task Force (USPSTF) guidelines.
Screening was defined as completed if it occurred at least once during the follow-up period, with the first date of documented screening considered as the date of completion. From 2016 to 2023, the USPSTF recommended breast cancer screening in women aged 50-74 years.17,18 Thus, breast cancer screening was assessed by women receiving a mammogram between the ages of 50 and 74 years. Cervical cancer screening was assessed with either cytology or high-risk human papillomavirus (hrHPV) testing for women between the ages of 21 and 65 years, and prostate cancer screening was assessed through screening with the prostate-specific antigen (PSA) between the ages of 55 and 69 years based on the USPSTF recommendations.19,20 Given the USPSTF endorses a variety of methods for colorectal cancer screening,21-23 screening was assessed through the gold standard for screening, colonoscopy, for both sexes between the ages of 50 and 75 years, or one of the alternative methods: flexible sigmoidoscopy, computed tomography (CT) colonography, DNA-fecal immunochemical test (FIT) testing, stool FIT testing, or guaiac fecal occult blood test (gFOBT). Colorectal screenings were examined overall and for individual screening types: (1) colorectal imaging (colonoscopy, flexible sigmoidoscopy, or CT colonography); (2) fecal DNA by DNA-FIT; and (3) fecal blood by FIT or gFOBT.
Study covariates and descriptive variables. The demographics and health behavior data were electronically collected from the REP. Race was defined as non-Hispanic White and non-White in the analyses. Non-White was defined as patients who identified as Black, Asian, Hawaiian/Pacific Islander, American Indian, Hispanic, or other/mixed. Education level was defined as completion of any education beyond high school. Area Deprivation Index (ADI) was determined by the participants’ addresses and stratified into 4 quartiles of national percentile rankings, with individuals in the fourth quartile living in the most deprived areas.
Patients who were former or current smokers were categorized as ever smokers. Obesity was defined as a BMI (calculated as weight in kilograms divided by height in meters squared) ≥ 30. The Charlson Comorbidity Index was calculated based on the Deyo adaption using 16 different comorbidity categories,24 with the exclusion of the category for rheumatic disease.
Previous use of DMARDs for the treatment of RA was determined through prescription review of the 5 years prior to the cohort entry date. RA disease features included the age at fulfillment of RA criteria, disease duration, and seropositivity status as determined by a positive rheumatoid factor or anticyclic citrullinated peptide antibody.
Statistical analysis. Descriptive statistics were used to summarize all participants. Aalen-Johansen methods were used to determine the cumulative incidence of cancer screening for both cohorts, accounting for the competing risk of death. Cox proportional hazard models adjusted for age, smoking status, and race (and sex for colon cancer screening) were used to compare the rates of cancer screening. The interactions between RA and various covariates were further investigated. If a covariate had an interaction, adjusted cumulative incidence curves were estimated using inverse probability weighting to ensure comparability between curves. A P value of < 0.05 was considered statistically significant for all analyses. Analyses were performed using SAS version 9.4 (SAS Institute) and R 4.4.1 (R Foundation for Statistical Computing).
RESULTS
Baseline patient characteristics. The study population included 1614 patients with RA and 1597 non-RA comparators. The RA cohort was female-predominant (71.38%) and primarily non-Hispanic White (92.01%), with a mean age of 63.2 (SD 14.4) years. Most patients with RA were either in the first or second quartile of the ADI (12.62% and 41.70%, respectively) and had a median Charlson Comorbidity Index of 1.0 (IQR 0.0-3.0). Approximately 53.97% of the RA cohort had a smoking history. Patients with RA had been diagnosed for a median of 6.5 (median 3.1-13.5) years, with the majority seropositive (70.31%). Patients had commonly used conventional DMARDs (68.15%) and had infrequently used biologic/targeted synthetic DMARDs (18.71%) during the 5 years prior to cohort entry (Table 1).
Baseline characteristics of RA and non-RA cohort.
A small percentage of the patients with RA were previously diagnosed with cervical (0.17%), prostate (2.16%), or colorectal (0.37%) cancer, and were excluded from the respective analyses. There were 1.49% who had either breast cancer or a complete bilateral mastectomy previously and were excluded from breast cancer screening analyses (Table 1).
Breast cancer screening. There were 647 women with RA who met the inclusion criteria for analyses and 644 non-RA comparators. At 5 years of follow-up, a higher proportion of the non-RA cohort had breast cancer screening compared to the RA cohort (85.2%, 95% CI 82.4-88.0 vs 83.4%, 95% CI 80.6-86.4; Figure 1A and Table 2). When adjusted for age, smoking, and race, the difference in breast cancer screening between the cohorts was not statistically significant (adjusted hazard ratio [aHR] 0.98, 95% CI 0.87-1.10; Table 2). Smoking was significant for decreased breast cancer screening in all patients (aHR 0.86, 95% CI 0.77-0.97; Table 3). When comparing those in the 1st quartile with those in the 2nd and 3rd quartiles combined, as determined by the ADI, RA and non-RA patients in the 1st quartile were more likely to have screening (aHR 1.28, 95% CI 1.08-1.52). Both cohorts were 50% more likely to have screening if they had completed education beyond high school (aHR 1.50, 95% CI 1.30-1.73; Table 3).
Five-year cumulative incidence of cancer screening in RA vs matched non-RA comparators for (A) breast cancer screening, (B) cervical cancer screening, and (C) prostate cancer screening. Non-RA comparators were matched on age, self-identified race, and county. HR: hazard ratio; RA: rheumatoid arthritis.
Cumulative incidence of cancer screening in patients with and without RA.
Association between patient characteristics and cancer screening in patients with RA and non-RA.
When investigating the association between the various covariates in RA and breast cancer screening, obesity had an association (Pinteraction = 0.01). The adjusted cumulative incidence at 5 years of breast cancer screening was highest in RA patients with obesity (90.3%, 95% CI 86.8-94.0) and lowest in RA patients without obesity (79.5%, 95% CI 75.3-84.0; Figure 2A).
Adjusted cumulative incidence of completion of cancer screening by year after January 1, 2015, for patients with and without RA according to (A) obesity in breast cancer screening, (B) race in cervical cancer screening, and (C) smoking status in prostate cancer screening. RA: rheumatoid arthritis.
When evaluating the effect of the COVID-19 pandemic, the cumulative incidence of breast cancer screening was 74.7% (95% CI 71.1-78.4) in the RA cohort and 76.4% (95% CI 72.9-80.1) in the non-RA cohort at 2 years of follow-up (Supplementary Table S1, available with the online version of this article).
Cervical cancer screening. There were 666 women with RA and 660 non-RA comparators included in the analyses. At 5 years of follow-up, over half of the RA (51.6%, 95% CI 47.9-55.6) and non-RA (58.2%, 95% CI 54.5-62.2) cohorts had cervical cancer screening (Figure 1B, Table 2). When adjusted for age, smoking, and race, patients with RA were 17% less likely to have cervical cancer screening compared to the non-RA comparators (aHR 0.83, 95% CI 0.72-0.96; Table 2). For both cohorts, smoking (aHR 0.81, 95% CI 0.69−0.94), Charlson Comorbidity Index ≥ 2 (aHR 0.77, 95% CI 0.63-0.94) and living in the lowest resourced areas by the ADI (aHR 0.76, 95% CI 0.59-0.99, compared to 2nd and 3rd quartiles combined) were significantly associated with decreased screening. Completion of any higher education following high school (aHR 1.26, 95% CI 1.02-1.55) led to increased cervical cancer screening (Table 3).
The effect of race on cervical cancer screening differed in the RA and non-RA cohorts (Pinteraction = 0.047). The adjusted cumulative incidence at 5 years was highest in White non-RA patients (60%, 95% CI 56.1-64.2%) and lowest in non-White non-RA patients (45.7%, 95% CI 33.7-62.0). Non-White patients with RA showed a comparable cumulative incidence to White patients with RA (51.8%, 95% CI 41.4-64.9 and 51.5%, 95% CI 47.6-55.8, respectively; Figure 2B).
Both RA and non-RA patients had a similar cumulative incidence of cervical cancer screening during the COVID-19 pandemic (37.6%, 95% CI 33.4-42.3% in RA vs 39.5%, 95% CI 35.2-44.3 in non-RA; Supplementary Table S1, available with the online version of this article).
Prostate cancer screening. There were 174 men with RA and 172 men without RA who met the inclusion criteria. Over half of both the RA and non-RA cohorts had PSA testing during the study period (51.9%, 95% CI 44.9-60.0) and 55.6% (95% CI 48.6-63.7; Figure 1C, Table 2). When adjusted for age, smoking, and race, men with and without RA were similarly likely to have prostate cancer screening (aHR 0.99, 95% CI 0.74-1.34; Table 2).
For both RA and non-RA men, smoking was associated with significantly decreased screening by 40% (aHR 0.60, 95% CI 0.44-0.80). The relationship between RA and smoking was significant (Pinteraction = 0.01). The adjusted cumulative incidence of prostate cancer screening showed that patients with RA who were never smokers had the highest proportion (73.3%, 95% CI 61.7-87.1), and those with a smoking history had the lowest proportion (43.4%, 95% CI 35.1-53.6; Figure 2C).
The cumulative incidence of prostate cancer screening during the COVID-19 pandemic was similar to prior rates in the RA cohort (36.2%, 95% CI 29.0-45.1) and non-RA cohort (46.5%, 95% CI 38.9-55.6; Supplementary Table S1, available with the online version of this article).
Colorectal cancer screening. As the length of time between recommended colorectal screening is up to 10 years, we investigated the time following the COVID-19 pandemic to determine if an extended time of analysis could be used. The years following the COVID-19 pandemic were like years prior to the pandemic (aHR 1.01, 95% CI 0.87-1.17 vs 1.00, 95% CI 0.88-1.13). Thus, we used the extended time frame for the colorectal cancer screening analyses, ultimately including 988 patients with RA and 980 non-RA comparators. Screening via any screening method for colorectal cancer was similar between both cohorts, with > 65% of RA (65.5%, 95% CI 62.5-68.6) and non-RA (65.6%, 95% CI 62.6-68.8) patients having screening by 8.5 years of follow-up (Figure 3A, Table 2). When adjusted for age, smoking, race, and sex, patients with RA were not shown to have an increased risk of screening delay (aHR 1.04, 95% CI 0.93-1.16; Table 2).
Cumulative incidence of cancer screening in RA vs matched non-RA comparators for colon cancer screening by (A) all methods, (B) imaging (ie, colonoscopy, flexible sigmoidoscopy, and CT colonography), (C) fecal DNA by DNA-FIT, and (D) fecal blood by FIT and gFOBT. Non-RA comparators were matched on age, self-identified race, and county. CT: computed tomography; FIT: fecal immunochemical test; gFOBT: guaiac fecal occult blood test; HR: hazard ratio; RA: rheumatoid arthritis.
For both the RA and non-RA cohorts, completion of any higher education (aHR 1.28, 95% CI 1.12-1.46) and belonging to the most resourced ADI quartile when compared to the middle quartiles (aHR 1.26, 95% CI 1.08-1.47) were protective factors in screening (Table 3).
When sensitivity analyses were performed to investigate the specific screening test used, the majority of both the RA and non-RA cohorts had colorectal cancer screening through colorectal imaging (50.2%, 95% CI 47.1-53.5 and 48.5%, 95% CI 45.4-51.9, respectively, at 8.5 years of follow-up; aHR 1.10, 95% CI 0.97-1.25; Figure 3B). A fifth of participants in each cohort had colorectal cancer screening through fecal DNA-FIT testing (RA 18.6%, 95% CI 16.3-21.3% and non-RA 20%, 17.6-22.8; aHR 0.92, 95% CI 0.75-1.13; Figure 3C). Around 10% of both cohorts had screening through fecal blood at least once during the follow-up period (RA 10.9%, 95% CI 9.1-13.1 and non-RA 8.3%, 95% CI 6.7-10.2; aHR 1.26, 95% CI 0.94-1.69; Figure 3D).
DISCUSSION
In this study, we aimed to determine the cumulative incidence of breast, cervical, prostate, and colorectal cancer screening in patients with RA compared to matched non-RA comparators and the effect of demographics and health behaviors on the age-appropriate cancer screenings. We found individuals with RA had decreased cervical cancer screening compared to non-RA comparators, and although we hypothesized that patients with RA would have a significant delay in breast, prostate, and colorectal cancer screenings, this was not shown in our study. The observed decrease in cervical cancer screening in patients with RA is likely multifactorial and due to similar reasons as seen in the general population, such as lack of resources and health insurance.25 Awareness of the risk of delay in cervical cancer screening in the RA population is important because many of the cervical cancer diagnoses and related deaths occur in women who have not been appropriately screened,26 and once diagnosed, the 5-year relative survival for cervical cancer is < 70%.27 With the use of screening through cytology every 3 years or hrHPV testing every 5 years, the number of cervical cancer-related deaths decreases to 0.74 and 0.29 deaths per 1000 women, respectively.19 White female individuals without RA had the highest incidence of cervical cancer screening, contrasting with the lowest screening occurring in non-White female individuals without RA. Non-White and White patients with RA had similar rates of screening. The increased screening seen in non-White patients with RA relative to their non-RA comparators could be due partially to the recommendations for women with RA prescribed immunosuppressants to follow the guidelines for HIV-infected women.28 However, the decreased screening found among White patients with RA relative to White women without RA is puzzling.
RA was not significantly associated with a delay in breast, prostate, or colorectal cancer screening as had been hypothesized. The hypothesized decrease in cancer screening was informed by previous studies finding decreased healthcare utilization by patients with RA over time, with one study showing half of patients with RA not having a primary care visit in over 6 months and a fifth not seeing their rheumatologist despite being prescribed immunosuppressants.14,29
When investigating factors that could increase the risk of delay, obesity was interestingly found to be a protective factor in breast cancer screening in the RA cohort. Obesity is associated with increased health expenses, with the main contributors to cost being outpatient services and medications, according to Medicare data.30 The increased healthcare expense from obesity could partially be explained by increased encounters with primary care providers. As preventative care is primarily led by primary care providers, preventative cancer screening is likely discussed at these visits. However, an interaction between obesity and increased cancer screening was not seen in cervical, prostate, and colorectal cancer screening, so this could represent a spurious result. Previous studies have shown that RA and incident prostate cancer are unrelated31 or only have a modestly increased risk.32 However, a recent metaanalysis investigating the association found those with a greater genetic susceptibility to RA had an increased risk of prostate cancer greater than 36 times,31 and other studies have found an association between smoking and increased mortality due to prostate cancer.33 This is congruent with our findings that patients with both RA and a smoking history were at an increased risk of delay in prostate cancer screening. Together, these studies and our results suggest that prostate cancer screening in this high-risk population may decrease mortality from prostate cancer.
Previous studies have shown that the rates of cancer screening were decreased significantly during the COVID-19 pandemic.34 Although patients with RA were affected by decreased access to care during the pandemic,35 the rates of cancer screening before and during the COVID-19 pandemic were similar among those with and without RA in our study.
Limitations to the study include the predominantly non-Hispanic White and geographically limited population, which could affect generalizability. We did not exclude patients who had a history of a hysterectomy due to being unable to distinguish between a subtotal and total hysterectomy and did not exclude male individuals with a history of a prostatectomy. This could cause a bias against the null hypothesis. Additionally, the study had a limited follow-up period and may not capture the completion or adequacy of all cancer screenings. Our study did not investigate continued compliance with the recommended cancer screenings, nor did it investigate lung cancer screening in patients with RA, which is needed to fully understand the RA population’s use of preventive cancer screenings. This study did not include patients with a recent history of the respective cancer of interest in the analyses. Based on the updated recommendations from EULAR in 2024,10 we likely will see increased use of targeted therapies for RA regardless of a history of cancer. Thus, it will be important to ensure patients with RA with a history of cancer are participating in both the recommended preventive cancer screenings and the surveillance that may be necessary for their specific cancer.10
In summary, patients with RA are at an increased risk of delay in cervical cancer screening compared to non-RA comparators. Patients with both RA and obesity had a decreased risk of delay in breast cancer screening. Conversely, patients with a smoking history and RA had a decrease in prostate cancer screening. Our study suggests that increased awareness of cervical cancer screening by healthcare providers may help ensure patients with RA receive screening.
Footnotes
CONTRIBUTIONS
All authors contributed to the drafting or revising of the article for important intellectual content. Study concept and design: RTB, CAH, ACH, EM, CSC. Acquisition of data: CAH, ACH, CSC. Analysis and interpretation of data: RTB, CAH, CSC. Writing - original draft: RTB, CSC. Writing - review & editing: RTB, CAH, ACH, JMD, VLK, EM, CSC.
FUNDING
This work was funded by grants from the National Institutes of Health (NIH), National Institute of Arthritis, Musculoskeletal and Skin Diseases (R01 AR46849) and National Institute on Aging (R01 AG068192, K24 AG078179). This study used the resources of the Rochester Epidemiology Project (REP) medical records linkage system, which is supported by the National Institute on Aging (AG 058738), by the Mayo Clinic Research Committee, and by fees paid annually by REP users. The content of this article is solely the responsibility of the authors and does not represent the official views of the NIH or the Mayo Clinic.
COMPETING INTERESTS
JMD declares research grant from Pfizer and licensed technology with royalty rights to Remission Medical. The remaining authors declare no conflicts of interest relevant to this article.
ETHICS AND PATIENT CONSENT
This study received Mayo Clinic and Olmsted Medical Center Institutional Review Board approval (IRB # 17-002593 and 017-OMC-17, respectively). Patients who denied use of their medical records for research purposes were excluded.
- Accepted for publication July 14, 2025.
- Copyright © 2025 by the Journal of Rheumatology
This is an Open Access article, which permits use, distribution, and reproduction, without modification, provided the original article is correctly cited and is not used for commercial purposes.
REFERENCES
SUPPLEMENTARY DATA
Supplementary material accompanies the online version of this article.
