Abstract
Objective
Autoimmune rheumatic diseases, including rheumatoid arthritis (RA), have recently been associated with an increased risk of cancer, particularly head and neck cancer (HNC). This study aimed to assess the risk of HNCs in patients with RA, focusing on the impact of serological status and specific HNC sites.
Methods
A retrospective cohort study was conducted using data from the Korean National Health Insurance Service. A total of 40 990 patients with RA were divided into seropositive and seronegative RA groups, matched with 204 950 non-RA controls. Cox proportional hazards models were used to calculate adjusted HRs (aHRs) for HNCs, adjusting for important covariates such as smoking, alcohol consumption and comorbidities.
Results
Patients with RA had an increased risk of overall HNCs (aHR: 1.53, 95% CI 1.12 to 2.09) compared with controls, with a significantly higher risk for pharyngeal cancer (aHR: 2.20, 95% CI 1.33 to 3.64). However, RA showed no significant association with oral, salivary or laryngeal cancer. The analysis stratified by seropositivity status revealed no significant difference in the risk of HNCs between seropositive and seronegative RA groups.
Conclusion
Patients with RA have an increased risk of developing overall HNCs, particularly pharyngeal cancer. However, no significant association was found for other HNC types. Further research is warranted to better understand the potential link between RA and increased pharyngeal cancer risk.
Keywords: Rheumatoid Arthritis, Epidemiology, Arthritis
WHAT IS ALREADY KNOWN ON THIS TOPIC
Recent studies have suggested that autoimmune rheumatic diseases (ARDs), including rheumatoid arthritis (RA), are associated with an increased risk of cancer, particularly head and neck cancer (HNC).
WHAT THIS STUDY ADDS
Using large-scale Korean cohort data, this study demonstrated that patients with RA have an increased risk of developing overall HNCs (adjusted HR (aHR): 1.53) compared with non-RA controls.
Specifically, patients with RA showed a significantly higher risk for pharyngeal cancer (aHR: 2.20), while no significant association was found with oral, salivary or laryngeal cancer.
The analysis also revealed no significant difference in the risk of HNCs between seropositive and seronegative RA groups.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This study highlights the need for further research to clarify the pathophysiological mechanisms by which RA selectively increases the risk of specific HNC types (pharyngeal cancer).
Introduction
In recent years, the role of autoimmune rheumatic diseases (ARDs) as a contributing factor to cancer development has gained attention. A complex relationship between these diseases has been proposed, whereby both underlying chronic inflammation and immunosuppressive treatments commonly used in ARDs may increase the risk of cancer.1 For example, rheumatoid arthritis (RA), the prototype of ARD, has been demonstrated to be associated with several cancers, including lung,2 kidney3 and haematological cancer.1 4
Existing research on head and neck cancers (HNCs) in patients with RA is limited. One of the most comprehensive studies is a Taiwanese study using data of 23 644 patients with RA from the National Health Insurance.5 That study reported an elevated risk of nasopharynx, sinus and ear cancers (22 cases, standardised incidence ratio (SIR): 1.39, 95% CI 1.34 to 1.45) and a modest increase in risk for oropharynx and larynx cancers (31 cases, SIR: 1.06, 95% CI 1.01 to 1.10).5 However, it was limited by the absence of adjustment for key covariates, which might have influenced results of that study, particularly lifestyle factors such as smoking and alcohol consumption. Moreover, that study combined multiple cancer types, integrating nasopharynx, sinus and ear cancers into one category. It also combined oropharynx and larynx cancers into a single category rather than analysing them separately.5
The extant literature on HNCs in patients with RA is constrained by two factors: a relatively small sample size and the paucity of cancer cases. One study conducted by Infante-Rivard et al6 in Canada included 2 412 patients with RA and revealed no statistically significant increase in the risk of HNC (SIR: 0.95), with only a few cases observed in the oral cavity. Cibere et al7 in Canada also reported an elevated risk for oropharyngeal cancer (SIR: 3.00) among 862 patients with RA, although the number of cases was low. In Japan, Yamada et al8 conducted an analysis of 7566 patients with RA, noting a few cases of oral, pharyngeal and laryngeal cancers (SIR: 0.93, online supplemental table S1). These gaps, specifically the lack of large-scale studies and robust case numbers, highlight the need for further research to better assess the potential risk of HNCs in patients with RA.
Additionally, previous studies that employed SIR calculations did not adjust for other risk factors that might influence cancer risk, which might have resulted in confounding of their results.5,8 Moreover, existing studies did not differentiate between seropositive rheumatoid arthritis (SPRA) and seronegative rheumatoid arthritis (SNRA), two subtypes of RA that might carry distinct cancer risk profiles.
To further investigate these uncertainties, we conducted a retrospective cohort study examining the risk of HNC in patients with RA compared with matched controls. This study focused on specific HNC sites, including oral, pharyngeal, laryngeal and salivary regions, and adjusted for critical confounders. Additionally, the potential influence of serological status on cancer risk was investigated. We hypothesised that patients with RA would exhibit an elevated risk of developing specific and overall HNCs compared with the general population, with variations based on serological status.
Methods
Data source
This study used data from the Korean National Health Insurance Service (KNHIS), the primary government-run health insurer in South Korea.9 The Korean National Health Insurance provides comprehensive universal health insurance coverage for 97% of the Korean population.9 Additionally, the KNHIS oversees Medical Aid that supports the remaining 3% of the population in the lowest income bracket through government financial subsidies.10
The KNHIS database is extensive, containing detailed information such as diagnostic codes, medical procedures and prescription records for the entire South Korean population.9 Diagnostic codes are based on the International Classification of Diseases, 10th Revision (ICD-10). Prescriptions are categorised using both Korean Drug and Anatomical Therapeutic Chemical codes.9 The KNHIS collects a wide range of health-related data, including demographic details, health behaviours, anthropometric measurements and laboratory results from routine national health examinations.10 As a result, this dataset is highly valuable for clinical and public health research.9 10
Identification of RA and serological status
To identify patients with newly diagnosed SPRA and SNRA, operational definitions were applied in a manner consistent with previous research.2,411 SPRA cases were identified using claims data that included the ICD-10 diagnosis code M05, the Rare and Intractable Disease (RID) registration code V223 and records of disease-modifying anti-rheumatic drug (DMARD) prescriptions (online supplemental table S2). The RID programme administered by the NHIS offers financial support for patients diagnosed with certain conditions. To be included in the RID programme, patients with SPRA must meet specific criteria, including testing positive for rheumatoid factor or anti-citrullinated protein antibodies, alongside documented confirmation from a physician that the patient fulfils the American College of Rheumatology/European Alliance of Associations for Rheumatology classification criteria for RA.11
Newly diagnosed patients with SNRA were identified through the ICD-10 code M06 (excluding M06.1 and M06.4) combined with DMARD prescriptions for at least 270 days to confirm the diagnosis. The index date was based on the RID registration date for SPRA and the initial diagnosis date for SNRA. This multifaceted approach, incorporating specific ICD-10 codes, RID registration and sustained DMARD use, ensures diagnostic validity and minimises the inclusion of transient inflammatory arthritis cases.11
Study population
The initial cohort consisted of patients aged 40 years or more who were newly diagnosed with RA between 1 January 2010 and 31 December 2017. An age cut-off of 40 was selected based on the incidence of RA and HNC, as well as the eligibility of the national health examination.10 16 17 A total of 92 336 patients with newly diagnosed RA without concurrent diagnosis of another rheumatological disease (online supplemental table S3) were included in this study. Of these, 66 493 and 25 843 were classified as having SPRA and SNRA, respectively. The initial control group included 1 162 700 individuals who were NHIS beneficiaries. None of whom had claims records with ICD-10 codes related to RA or other rheumatological conditions before 1 January 2010.
To ensure inclusion of potential confounding data from national health examinations, only those who had attended a health examination within 2 years prior to the index date were included. This resulted in a sample size of 54 910 patients with RA (39 050 with SPRA and 15 860 with SNRA) and 641 341 controls. Individuals with missing data were excluded, resulting in 1333 exclusions from the SPRA group and 592 exclusions from the SNRA group. Individuals with a cancer diagnosis before the index date, including 1937 with SPRA and 965 with SNRA, were also excluded. To minimise over-detection of disease during the RA diagnostic process, a 1-year lag period was introduced and patients diagnosed with HNCs or who died within 1 year of the index date were excluded (n=399 for SPRA and n=18 for SNRA). The same inclusion and exclusion criteria described above were also applied to non-RA controls. Each patient with RA was then matched with five controls based on age and sex, resulting in a population-based matched cohort of 40 990 patients with RA and 204 950 non-RA controls (figure 1).
Figure 1. Flow diagram of the study population. Patients with RA were matched by sex and age at a 1:5 ratio with subjects without RA. N, number; RA, rheumatoid arthritis; SNRA, seronegative rheumatoid arthritis; SPRA, seropositive rheumatoid arthritis.
Outcomes and follow-up
The main outcome of this study was the occurrence of newly diagnosed HNCs. These cancers were identified using the following ICD-10 codes: C00-C06 for oral cavity cancers, C07-C08 for salivary gland cancers, C09-C13 for pharyngeal cancers, C32 for laryngeal cancer and C14, C30-C31 for cancers in other related areas (online supplemental table S4). To validate cancer diagnoses within the claims data, both diagnosis codes and codes for special copayment reduction applicable to cancer patients were used.17 As the majority of cancer patients in Korea are registered in this programme, data on HNC diagnosis used in the present study were sufficiently reliable.17 Follow-up started 1 year after the index date to allow for a 1-year lag. It continued until the earliest occurrence of a cancer diagnosis, death from any cause, emigration or 31 December 2020, whichever came first.
Variables
Comorbidities were identified using a combination of diagnostic codes and clinical measures. Hypertension was identified by ICD-10 codes I10-I13 and I15, or if participants were taking antihypertensive medication or had a blood pressure reading of ≥140/90 mm Hg at the national health examination. Diabetes mellitus was identified using ICD-10 codes E11-E14 in combination with the use of antidiabetic medication or a fasting glucose level ≥126 mg/dL. Dyslipidaemia was identified by ICD-10 code E78, combined with the use of lipid-lowering drugs or a total cholesterol level ≥240 mg/dL. Chronic kidney disease was defined by diagnostic codes (ICD-10 codes N18 and N19) or a glomerular filtration rate less than 60 mL/min/1.73 m² estimated using the Modification of Diet in Renal Disease formula.
Smoking status was categorised into five groups: never-smokers, ex-smokers with less than 20 pack-years (PY), ex-smokers with 20 or more PY, current smokers with less than 20 PY and current smokers with 20 or more PY. Alcohol consumption was classified as non-drinker or drinker (>0 g/day). Regular exercise was defined as moderate activity ≥3 days per week or vigorous physical activity ≥5 days per week.18 Income level was determined by insurance premiums based on household income, with the lowest income group defined as the bottom 25% of beneficiaries and Medicaid recipients. Body mass index (BMI) was calculated by dividing weight in kilograms by height in metres squared (kg/m²).
Statistical analysis
This study compared patients with RA with a non-RA population using various statistical methods. Continuous variables were analysed using t-tests, while categorical variables were assessed using χ2 tests. Kaplan-Meier plots were used to evaluate the cumulative incidence of HNCs in patients with RA compared with the non-RA group. Cox proportional hazards models were used to calculate HRs and 95% CIs for the risk of developing HNCs after adjusting for various covariates.
Three adjustment models were used. Model 1 provided unadjusted estimates. Model 2 was adjusted for factors such as age, sex, smoking status,19 alcohol consumption,20 physical activity, income level, BMI and diabetes mellitus.21 Model 3 was further adjusted for hypertension, dyslipidaemia and chronic kidney disease in addition to variables in model 2. In Model 4, we applied Fine-Gray competing risk regression to account for competing events when analysing the incidence of HNC. Death from any cause was treated as a competing event, while emigration was treated as a censoring event.
All statistical analyses were performed using SAS software (V.9.4). Statistical significance was considered when two-tailed p values were less than 0.05.
Data availability
Data from the KNHIS can be accessed through their official website (http://nhiss.nhis.or.kr). To gain access, researchers must submit a study proposal, which must first be approved by their institutional review board. The proposal is then reviewed by the KNHIS review committee before access is granted.
Results
Baseline characteristics
This study included 40 990 individuals with newly diagnosed RA. Among them, 30 730 (75.0%) were women and 29 062 (70.9%) had SPRA. Patients with RA were slightly more likely to be current or former smokers, with a greater number identified as current (6.2% vs 5.3%) or former (4.6% vs 3.7%) smokers having 20 or more PY compared with those in the non-RA group. Additionally, patients with RA were less likely to engage in regular exercise (18.4% vs 21.0%). Although the prevalence of diabetes was similar, patients with RA had higher rates of hypertension (39.9% vs 36.7%) and chronic kidney disease (CKD) (7.6% vs 6.1%) than patients with non-RA. However, patients with RA had fewer individuals who were classified as obese (BMI ≥25 kg/m²) (30.4% vs 33.8%).
Patients with SPRA were more likely to be current smokers with 20 or more PY compared with patients with SNRA (6.9% vs 4.6%). They were also more likely to be non-drinkers (76.2% vs 73.8%). However, they were less likely to exercise regularly (18.0% vs 19.6%). In terms of comorbidities, patients with SPRA were less likely to be classified as obese (29.2% vs 33.4%). They had a lower prevalence of diabetes, hypertension, dyslipidaemia and CKD (table 1).
Table 1. Baseline characteristics of the study population.
| Total | RA | Serologic status of RA | |||||
|---|---|---|---|---|---|---|---|
| No | Yes | P value | Seropositive | Seronegative | P value | ||
| N | 245 940 | 204 950 | 40 990 | 29 062 | 11 928 | ||
| Sex, n (%) | |||||||
| Male | 61 560 (25.0) | 51 300 (25.0) | 10 260 (25.0) | 7046 (24.2) | 3214 (27.0) | ||
| Female | 184 380 (75.0) | 153 650 (75.0) | 30 730 (75.0) | ||||
| Age (years), n (%) | |||||||
| 40–59 | 143 346 (58.3) | 119 455 (58.3) | 23 891 (58.3) | 16 634 (57.2) | 7257 (60.8) | ||
| ≥60 | 102 594 (41.7) | 85 495 (41.7) | 17 099 (41.7) | 12 428 (42.8) | 4671 (39.2) | ||
| Mean±SD | 57.7±9.4 | 57.7±9.4 | 57.7±9.4 | 58.0±9.4 | 57.1±9.5 | ||
| Income level, lowest 25%, n (%) | 54 982 (22.4) | 45 711 (22.3) | 9271 (22.6) | 0.163 | 6587 (22.7) | 2684 (22.5) | 0.719 |
| Smoking, n (%) | <0.001 | <0.001 | |||||
| Never | 195 591 (79.5) | 163 623 (79.8) | 31 968 (78.0) | 22 710 (78.1) | 9258 (77.6) | ||
| Ex-smoker, <20 PY | 14 249 (5.8) | 11 796 (5.8) | 2453 (6.0) | 1554 (5.4) | 899 (7.5) | ||
| Ex-smoker, ≥20 PY | 9589 (3.9) | 7692 (3.7) | 1897 (4.6) | 1298 (4.5) | 599 (5.0) | ||
| Current, <20 PY | 13 119 (5.3) | 10 985 (5.4) | 2134 (5.2) | 1509 (5.2) | 625 (5.2) | ||
| Current, ≥20 PY | 13 392 (5.5) | 10 854 (5.3) | 2538 (6.2) | 1991 (6.9) | 547 (4.6) | ||
| Non-drinker, n (%) | 174 098 (70.8) | 143 166 (69.9) | 30 932 (75.5) | <0.001 | 22 132 (76.2) | 8800 (73.8) | <0.001 |
| Regular exercise, n (%) | 50 659 (20.6) | 43 101 (21.0) | 7558 (18.4) | <0.001 | 5220 (18.0) | 2338 (19.6) | <0.001 |
| SBP (mm Hg), mean±SD | 123.3±15.3 | 123.4±15.3 | 122.4±15.2 | <0.001 | 122.4±15.2 | 122.3±15.2 | 0.605 |
| DBP (mm Hg), mean±SD | 76.1±10.0 | 76.3±10.0 | 75.6±9.9 | <0.001 | 75.7±9.9 | 75.5±9.8 | 0.286 |
| Comorbidities | |||||||
| BMI≥25 kg/m2, n (%) | 81 652 (33.2) | 69 190 (33.8) | 12 462 (30.4) | <0.001 | 8473 (29.2) | 3989 (33.4) | <0.001 |
| Diabetes, n (%) | 30 844 (12.5) | 25 727 (12.6) | 5117 (12.5) | 0.699 | 3526 (12.1) | 1591 (13.3) | 0.001 |
| Hypertension, n (%) | 91 329 (37.1) | 75 233 (36.7) | 16 096 (39.9) | <0.001 | 10 991 (37.8) | 5105 (42.8) | <0.001 |
| Dyslipidaemia, n (%) | 80 709 (32.8) | 67 102 (32.7) | 13 607 (33.2) | 0.073 | 9226 (31.8) | 4381 (36.7) | <0.001 |
| CKD, n (%) | 15 694 (6.4) | 12 570 (6.1) | 3124 (7.6) | <0.001 | 2068 (7.1) | 1056 (8.9) | <0.001 |
BMI, body mass index; CKD, chronic kidney disease; DBP, diastolic blood pressure; N, number; PY, pack-year; RA, rheumatoid arthritis; SBP, systolic blood pressure.
Risk of HNC in patients with RA
The mean follow-up after the 1-year lag period was 5.5±2.2 years. In this population, 53 cases of HNCs were observed in patients with RA compared with 175 cases in patients with non-RA (table 2). Figure 2 shows the incidence probability of HNC in patients with RA compared with controls. The adjusted HR (aHR) for HNCs in patients with RA was 1.53 (95% CI 1.12 to 2.09). Among RA subtypes, SNRA was associated with increased HNC risk (aHR: 2.00, 95% CI 1.24 to 3.22). However, the risk for HNC did not differ significantly between those with SPRA and those with SNRA (the reference) (aHR: 0.65, 95% CI 0.37 to 1.15) (table 2).
Table 2. HR and 95% CI for cancer incidence in patients with rheumatoid arthritis (RA).
| N | Event (n) | Follow-up duration (person-year) |
IR per 1000 | HR (95% CI) Model 1 |
HR (95% CI) Model 2 |
HR (95% CI) Model 3 |
HR (95% CI) Model 4 |
|
|---|---|---|---|---|---|---|---|---|
| Head and neck cancer | ||||||||
| Non-RA | 204 950 | 175 | 1 129 409.6 | 0.16 | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) |
| RA | 40 990 | 53 | 222 495.0 | 0.24 | 1.54 (1.13 to 2.09) | 1.54 (1.13 to 2.09) | 1.53 (1.12 to 2.09) | 1.48 (1.09 to 2.02) |
| SNRA | 11 928 | 19 | 64 382.0 | 0.30 | 1.91 (1.19 to 3.07) | 2.01 (1.25 to 3.23) | 2.00 (1.24 to 3.22) | 1.98 (1.23 to 3.17) |
| SPRA | 29 062 | 34 | 158 113.1 | 0.22 | 1.39 (0.96 to 2.01) | 1.36 (0.94 to 1.96) | 1.35 (0.94 to 1.96) | 1.30 (0.90 to 1.88) |
| SNRA | 11 928 | 19 | 64 382.0 | 0.30 | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) |
| SPRA | 29 062 | 34 | 158 113.1 | 0.22 | 0.73 (0.41 to 1.27) | 0.65 (0.37 to 1.15) | 0.65 (0.37 to 1.15) | 0.64 (0.37 to 1.11) |
| Oral cancer | ||||||||
| Non-RA | 204 950 | 68 | 1 129 705.8 | 0.06 | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) |
| RA | 40 990 | 14 | 222 585.9 | 0.06 | 1.05 (0.59 to 1.86) | 1.03 (0.58 to 1.84) | 1.02 (0.54 to 1.82) | 0.99 (0.55 to 1.78) |
| SNRA | 11 928 | 5 | 64 420.9 | 0.08 | 1.30 (0.52 to 3.21) | 1.32 (0.53 to 3.29) | 1.28 (0.52 to 3.19) | 1.27 (0.51 to 3.14) |
| SPRA | 29 062 | 9 | 158 165.0 | 0.06 | 0.95 (0.47 to 1.90) | 0.92 (0.46 to 1.85) | 0.91 (0.45 to 1.84) | 0.88 (0.44 to 1.79) |
| SNRA | 11 928 | 5 | 64 420.9 | 0.08 | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) |
| SPRA | 29 062 | 9 | 158 165.0 | 0.06 | 0.74 (0.25 to 2.20) | 0.75 (0.25 to 2.25) | 0.74 (0.24 to 2.22) | 0.71 (0.26 to 1.99) |
| Salivary cancer | ||||||||
| Non-RA | 204 950 | 21 | 1 129 816.1 | 0.02 | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) |
| RA | 40 990 | 5 | 222 605.2 | 0.02 | 1.21 (0.46 to 3.21) | 1.20 (0.45 to 3.19) | 1.22 (0.46 to 3.24) | 1.18 (0.45 to 3.12) |
| SNRA | 11 928 | 0 | 64 423.2 | 0.00 | – | – | – | – |
| SPRA | 29 062 | 5 | 158 182.0 | 0.03 | 1.70 (0.64 to 4.50) | 1.66 (0.62 to 4.42) | 1.67 (0.63 to 4.46) | 1.62 (0.61 to 4.28) |
| SNRA | 11 928 | 0 | 64 423.2 | 0.00 | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) |
| SPRA | 29 062 | 5 | 158 182.0 | 0.03 | – | – | – | – |
| Pharynx cancer | ||||||||
| Non-RA | 204 950 | 51 | 1 129 737.6 | 0.05 | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) |
| RA | 40 990 | 22 | 222 564.0 | 0.10 | 2.19 (1.33 to 3.62) | 2.20 (1.33 to 3.64) | 2.20 (1.33 to 3.64) | 2.14 (1.30 to 3.53) |
| SNRA | 11 928 | 9 | 64 399.5 | 0.14 | 3.12 (1.53 to 6.33) | 3.23 (1.58 to 6.57) | 3.27 (1.60 to 6.68) | 3.23 (1.60 to 6.56) |
| SPRA | 29 062 | 13 | 158 164.5 | 0.08 | 1.82 (0.99 to 3.35) | 1.80 (1.97 to 3.32) | 1.79 (0.97 to 3.31) | 1.73 (0.94 to 3.17) |
| SNRA | 11 928 | 9 | 64 399.5 | 0.14 | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) |
| SPRA | 29 062 | 13 | 158 164.5 | 0.08 | 0.58 (0.25 to 1.35) | 0.51 (0.22 to 1.21) | 0.52 (0.22 to 1.22) | 0.51 (0.22 to 1.15) |
| Larynx cancer | ||||||||
| Non-RA | 204 950 | 43 | 1 129 785.3 | 0.04 | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) |
| RA | 40 990 | 11 | 222 586.9 | 0.05 | 1.30 (0.67 to 2.52) | 1.30 (0.67 to 2.52) | 1.30 (0.67 to 2.53) | 1.23 (0.64 to 2.36) |
| SNRA | 11 928 | 2 | 64 418.2 | 0.03 | 0.82 (0.20 to 3.38) | 0.96 (0.23 to 3.95) | 0.96 (0.23 to 3.97) | 0.94 (0.23 to 3.92) |
| SPRA | 29 062 | 9 | 158 168.6 | 0.06 | 1.50 (0.73 to 3.07) | 1.41 (0.68 to 2.90) | 1.41 (0.68 to 2.90) | 1.31 (0.65 to 2.67) |
| SNRA | 11 928 | 2 | 64 418.2 | 0.03 | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) |
| SPRA | 29 062 | 9 | 158 168.6 | 0.06 | 1.81 (0.39 to 8.40) | 1.35 (0.29 to 6.34) | 1.35 (0.29 to 6.37) | 1.24 (0.26 to 5.93) |
| Others | ||||||||
| Non-RA | 204 950 | 20 | 1 129 805.7 | 0.02 | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) |
| RA | 40 990 | 7 | 222 593.4 | 0.03 | 1.77 (0.75 to 4.19) | 1.69 (0.71 to 4.02) | 1.67 (0.70 to 3.97) | 1.62 (0.67 to 3.94) |
| SNRA | 11 928 | 3 | 64 413.0 | 0.05 | 2.60 (0.77 to 8.76) | 2.69 (0.80 to 9.08) | 2.67 (0.79 to 9.03) | 2.65 (0.79 to 8.91) |
| SPRA | 29 062 | 4 | 158 180.3 | 0.03 | 1.43 (0.49 to 4.18) | 1.32 (0.45 to 3.88) | 1.30 (0.44 to 3.83) | 1.25 (0.41 to 3.82) |
| SNRA | 11 928 | 3 | 64 413.0 | 0.05 | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) | 1 (Ref.) |
| SPRA | 29 062 | 4 | 158 180.3 | 0.03 | 0.55 (0.12 to 2.45) | 0.51 (0.11 to 2.32) | 0.51 (0.11 to 2.32) | 0.50 (0.10 to 2.37) |
Model 1: unadjusted. Model 2: adjusted for age, sex, smoking status with five categories, alcohol drinking, regular physical activity, income level, BMI and diabetes. Model 3: adjusted for hypertension, dyslipidaemia, CKD and variables in model 2. Model 4: adjusted for the variables in model 3, with death from any cause treated as a competing risk using Fine-Gray regression.
BMI, body mass index; CKD, chronic kidney disease; IR, incidence rate; N, number; SNRA, seronegative rheumatoid arthritis; SPRA, seropositive rheumatoid arthritis.
Figure 2. Kaplan-Meier curves illustrating the incidence probability of head and neck cancer in overall patients with RA (left) and in patients with SPRA and SNRA (right) compared with the control group. RA, rheumatoid arthritis; SNRA, seronegative rheumatoid arthritis; SPRA, seropositive rheumatoid arthritis.
Oral cancer
There were 14 cases of oral cancer in patients with RA compared with 68 cases in patients with non-RA (table 2). For oral cancer, patients with RA did not have a significantly higher risk compared with the non-RA population (aHR: 1.02, 95% CI 0.54 to 1.82). The risk for oral cancer did not differ between those with SPRA and those with SNRA (the reference) (aHR: 0.74, 95% CI 0.24 to 2.22).
Salivary cancer
There were 5 cases of salivary cancer in patients with RA compared with 21 cases in patients with non-RA (table 2). For salivary cancer, patients with RA had no significantly different risk compared with non-RA population (aHR: 1.22, 95% CI 0.46 to 3.24) (table 2). Those with SNRA had no case of salivary cancer. Thus, the association between serological status and salivary cancer risk could not be determined.
Pharyngeal cancer
There were 22 cases of pharyngeal cancer in patients with RA compared with 51 cases in the non-RA population (table 2). Patients with RA had a higher risk of developing pharyngeal cancer than those without RA (aHR: 2.20, 95% CI 1.33 to 3.64). The aHR for pharyngeal cancer was 1.79 (95% CI 0.97 to 3.31) for patients with SPRA and 3.27 (95% CI 1.60 to 6.68) for patients with SNRA compared with the non-RA group. However, the risk difference between those with SPRA and those with SNRA was not significantly different (SPRA vs SNRA, aHR: 0.52, 95% CI 0.22 to 1.22).
Larynx cancer
The aHR for patients with RA was 1.30 (95% CI 0.67 to 2.53), suggesting a non-significant increase in risk compared with non-RA. Among RA subtypes, SPRA showed a higher aHR of 1.35 (95% CI 0.29 to 6.37) compared with SNRA, although such a difference was not statistically significant (table 2).
Competing risk analysis for HNC incidence
To account for the potential confounding effect of death as a competing risk, we performed Fine-Gray competing risk regression analysis (model 4). The results were consistent with the conventional Cox proportional hazards models (model 3). The significantly increased risk of HNC in the RA group persisted after accounting for death as a competing event (sHR 1.48, 95% CI 1.09 to 2.02). Specifically, the robust association between RA and pharyngeal cancer remained statistically significant (sHR 2.14, 95% CI 1.30 to 3.53), while the associations for oral, salivary and laryngeal cancers remained non-significant, consistent with the primary analyses.
Discussion
To the best of our knowledge, this is the first study assessing the risk of HNC in patients with RA by specific HNC sites, taking RA serostatus into account. This study, involving more than 40 000 newly diagnosed patients with RA, represents a study with the largest sample size to date studying the risk of HNCs in patients with RA. However, due to the rarity of outcomes, the total number of HNC cases in RA was only 53. Our results showed a 1.53-fold increased risk of HNC in patients with RA compared with those without RA. By specific cancer sites, patients with RA had a particularly higher risk of pharyngeal cancer (greater than two times) than those without RA. The risk of oral, salivary and laryngeal cancers in the RA group was not significantly higher than that in those without RA.
RA and cancer might share key immunological and pathogenic factors that have yet to be fully identified. For example, chronic inflammation associated with RA can lead to a prolonged state of immune activation, which might impede the immune system’s ability to detect and destroy tumours, potentially contributing to cancer development.22 This persistent inflammation might also contribute to gene rearrangements and mutations.22 Furthermore, autoimmune characteristics of RA, along with the use of immunosuppressive treatments, could further influence cancer risk in patients with RA.22
Patients with RA are at an elevated risk of infections due to disturbances in both innate and adaptive immune systems.23 Neutropenia, a common occurrence in severe RA or in individuals undergoing immunosuppressive treatments, along with the formation of pathological immune complexes, can impair neutrophil function and bacterial defence.23 Furthermore, a constricted T-cell receptor repertoire and reduced thymic output can result in an inability of the adaptive immune system to function effectively, while genetic factors such as TRAF1/C5 polymorphisms contribute to an increased susceptibility to infection in patients with RA.24
Associations of human papillomavirus (HPV)25 and Epstein-Barr virus (EBV)26 with HNCs have been established. Individuals with RA might be more susceptible to persistent viral infections,27 which can lead to cancer development, particularly when the immune system is compromised. Our study found a higher risk of pharyngeal cancer in patients with RA than in patients with non-RA. It is worth noting that both HPV and EBV have been identified as significant factors contributing to the high incidence of pharyngeal cancer. Previous studies have indicated an increase in HPV-related and EBV-related oropharyngeal squamous cell carcinoma, reflecting a shift in the epidemiology of HNCs.28 Given persistent immune stimulation and chronic immune activation associated with RA, as well as the EBV-induced polyclonal lymphocyte proliferation, individuals with RA might have an increased risk of malignant transformation, potentially elevating the likelihood of HNCs.
In our study, patients with RA did not have a significantly increased risk of oral cancer, salivary cancer or laryngeal cancer compared with non-RA participants. However, it is important to note that the relatively small number of cases for these specific sites (oral, salivary and laryngeal cancers) may have limited the statistical power to detect a significant association. The low event counts resulted in wider CIs, which might explain the lack of statistical significance. A Taiwanese study reported a marginal increase in the risk of oropharynx and larynx cancer (SIR 1.24), although that finding did not adjust for smoking or alcohol use.5 These conflicting findings highlight the need for further studies to clarify the association between RA and the risk of site-specific HNCs and to explore the underlying mechanisms.
Strengths of our study lie in several key aspects. First, our rigorous criteria employed to identify patients with RA, distinguishing between SPRA and SNRA, helped mitigate the risk of overestimating RA cases and enhanced the reliability of our findings. Second, the use of a large cohort enhanced the statistical power of our study and analysis of specific HNC sites. Third, our comprehensive evaluation of HNC risk accounted for essential confounding factors using national health examination data and claims records, minimising potential biases and enhancing the validity of our results. Furthermore, our findings remained robust in the competing risk analysis using Fine-Gray regression, which accounted for death as a competing event for HNC incidence. This consistency suggests that the observed relationship between RA and HNC is not driven by survival bias, strengthening the validity of our conclusions.
This study also has several limitations. First, despite the large total cohort size, the absolute number of HNC events (n=53) was relatively small. This rarity of outcomes, particularly when stratified by specific cancer site or RA serostatus, led to wide CIs and may have limited our ability to reach statistical significance for specific subtypes. Second, the nature of the administrative claims database provided by the NHIS precluded consideration of granulated clinical factors such as disease activity and cumulative medication exposure (eg, specific DMARDs, steroids or NSAIDs). These factors are potential residual confounders that could influence infection-related cancer risk in patients with RA. Third, study participants were restricted to those who underwent health screening examinations, which might have introduced selection bias. Consequently, our cohort may represent a relatively healthier population with more engagement in healthy lifestyles (eg, higher rates of smoking cessation or regular medical check-ups) than the general population.29 This could potentially lead to an underestimation of the absolute risk of HNC, although the internal comparison between RA and non-RA groups remains valid as both were subject to the same inclusion criteria. Finally, this study was conducted using a database from a single Asian country. Thus, results should be generalised with caution to other populations.
Conclusions
In conclusion, patients with RA had an increased risk of HNCs compared with matched controls, particularly for pharyngeal cancer. However, RA showed no significant association with cancers of the oral cavity, salivary glands or larynx. Further studies are warranted to explore pathophysiological mechanisms for the association between RA and HNC, particularly pharyngeal cancer.
Supplementary material
Acknowledgements
This research was supported by the Institution of Quality of Life in Cancer funded by Samsung Fire & Marine Insurance.
Footnotes
Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Provenance and peer review: Not commissioned; externally peer reviewed.
Patient consent for publication: Not applicable.
Data availability free text: The data from the Korean National Health Insurance Service (KNHIS) is accessible through their website (http://nhiss.nhis.or.kr). Researchers must submit a study proposal approved by their affiliated institutional review board, which is then reviewed by the KNHIS review committee before granting access.
Ethics approval: The study protocol was approved by the Institutional Review Board (IRB) of Samsung Medical Center (IRB No. 2022-06-141). The need for written informed consent from patients was waived by the IRB given the public and anonymised nature of data and adherence to confidentiality guidelines.
Data availability statement
Data may be obtained from a third party and are not publicly available.
References
- 1.Simon TA, Thompson A, Gandhi KK, et al. Incidence of malignancy in adult patients with rheumatoid arthritis: a meta-analysis. Arthritis Res Ther. 2015;17:212. doi: 10.1186/s13075-015-0728-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Cho MH, Cho JH, Eun Y, et al. Rheumatoid Arthritis and Risk of Lung Cancer: A Nationwide Cohort Study. J Thorac Oncol. 2024;19:216–26. doi: 10.1016/j.jtho.2023.10.006. [DOI] [PubMed] [Google Scholar]
- 3.Lee D, Jeon KH, Jung J, et al. Urological cancer risk in patients with rheumatoid arthritis compared to matched controls: a nationwide cohort study. Rheumatology (Oxford) 2025;64:3458–67. doi: 10.1093/rheumatology/keaf103. [DOI] [PubMed] [Google Scholar]
- 4.Cho IY, Han K, Jung JH, et al. Rheumatoid arthritis and the risk of hematologic malignancies: a nationwide cohort study. Leukemia. 2025;39:755–9. doi: 10.1038/s41375-024-02477-3. [DOI] [PubMed] [Google Scholar]
- 5.Chen YJ, Chang YT, Wang CB, et al. The risk of cancer in patients with rheumatoid arthritis: a nationwide cohort study in Taiwan. Arthritis Rheum. 2011;63:352–8. doi: 10.1002/art.30134. [DOI] [PubMed] [Google Scholar]
- 6.Infante-Rivard C, Rivard G-E, Derome F, et al. A retrospective cohort study of cancer incidence among patients treated with radiosynoviorthesis. Haemophilia. 2012;18:805–9. doi: 10.1111/j.1365-2516.2012.02802.x. [DOI] [PubMed] [Google Scholar]
- 7.Cibere J, Sibley J, Haga M. Rheumatoid arthritis and the risk of malignancy. Arthritis Rheum. 1997;40:1580–6. doi: 10.1002/art.1780400906. [DOI] [PubMed] [Google Scholar]
- 8.Yamada T, Nakajima A, Inoue E, et al. Incidence of malignancy in Japanese patients with rheumatoid arthritis. Rheumatol Int. 2011;31:1487–92. doi: 10.1007/s00296-010-1524-0. [DOI] [PubMed] [Google Scholar]
- 9.Cheol Seong S, Kim Y-Y, Khang Y-H, et al. Data Resource Profile: The National Health Information Database of the National Health Insurance Service in South Korea. Int J Epidemiol. 2017;46:799–800. doi: 10.1093/ije/dyw253. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Shin DW, Cho J, Park JH, et al. National General Health Screening Program in Korea: history, current status, and future direction. Precis Future Med . 2022;6:9–31. doi: 10.23838/pfm.2021.00135. [DOI] [Google Scholar]
- 11.Cho SK, Sung YK, Choi CB, et al. Development of an algorithm for identifying rheumatoid arthritis in the Korean National Health Insurance claims database. Rheumatol Int. 2013;33:2985–92. doi: 10.1007/s00296-013-2833-x. [DOI] [PubMed] [Google Scholar]
- 12.Jeon KH, Han K, Jung J, et al. Rheumatoid Arthritis and Risk of Depression in South Korea. JAMA Netw Open . 2024;7:e241139. doi: 10.1001/jamanetworkopen.2024.1139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Kang J, Eun Y, Jang W, et al. Rheumatoid Arthritis and Risk of Parkinson Disease in Korea. JAMA Neurol. 2023;80:634–41. doi: 10.1001/jamaneurol.2023.0932. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Cho MH, Jung J, Kim TH, et al. Rheumatoid arthritis and risk of thyroid cancer: a nationwide cohort study. Arthritis Care Res (Hoboken) 2021;73:1516–23. doi: 10.1002/acr.24377. [DOI] [PubMed] [Google Scholar]
- 15.Jeon KH, Jung J, Cho MH, et al. RA and risk of colorectal cancer: a nationwide cohort study. Rheumatology (Sunnyvale) 2026;65 doi: 10.1093/rheumatology/keaf451. [DOI] [PubMed] [Google Scholar]
- 16.van Delft ETAM, Jamal M, den Braanker H, et al. A systematic review on time trend incidence of rheumatoid arthritis in outpatient rheumatology clinics. Front Med (Lausanne) 2022;9:933884. doi: 10.3389/fmed.2022.933884. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Yang MS, Park M, Back JH, et al. Validation of Cancer Diagnosis Based on the National Health Insurance Service Database versus the National Cancer Registry Database in Korea. Cancer Res Treat . 2022;54:352–61. doi: 10.4143/crt.2021.044. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Pescatello LS. ACSM’s guidelines for exercise testing and prescription. 9 edn. Philadelphia: Lippincott Williams & Wilkins; 2014. [Google Scholar]
- 19.Freedman ND, Abnet CC, Leitzmann MF, et al. Prospective investigation of the cigarette smoking-head and neck cancer association by sex. Cancer. 2007;110:1593–601. doi: 10.1002/cncr.22957. [DOI] [PubMed] [Google Scholar]
- 20.Koo HY, Han K, Shin DW, et al. Alcohol Drinking Pattern and Risk of Head and Neck Cancer: A Nationwide Cohort Study. Int J Environ Res Public Health. 2021;18:11204. doi: 10.3390/ijerph182111204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Ujpál M, Matos O, Bíbok G, et al. Diabetes and oral tumors in Hungary: epidemiological correlations. Diabetes Care. 2004;27:770–4. doi: 10.2337/diacare.27.3.770. [DOI] [PubMed] [Google Scholar]
- 22.Egiziano G, Bernatsky S, Shah AA. Cancer and autoimmunity: Harnessing longitudinal cohorts to probe the link. Best Pract Res Clin Rheumatol. 2016;30:53–62. doi: 10.1016/j.berh.2016.03.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Starkebaum G. Chronic neutropenia associated with autoimmune disease. Semin Hematol. 2002;39:121–7. doi: 10.1053/shem.2002.31918. [DOI] [PubMed] [Google Scholar]
- 24.Listing J, Gerhold K, Zink A. The risk of infections associated with rheumatoid arthritis, with its comorbidity and treatment. Rheumatology (Oxford) 2013;52:53–61. doi: 10.1093/rheumatology/kes305. [DOI] [PubMed] [Google Scholar]
- 25.Joseph AW, D’Souza G. Epidemiology of human papillomavirus-related head and neck cancer. Otolaryngol Clin North Am. 2012;45:739–64. doi: 10.1016/j.otc.2012.04.003. [DOI] [PubMed] [Google Scholar]
- 26.Al-Saeed R, Al-Rawi NH. Evidence of Epstein-Barr Virus Association with Head and Neck Cancers: A Review. J Can Dent Assoc. 2016;82:g2. [PubMed] [Google Scholar]
- 27.Kudaeva FM, Speechley MR, Pope JE. A systematic review of viral exposures as a risk for rheumatoid arthritis. Semin Arthritis Rheum. 2019;48:587–96. doi: 10.1016/j.semarthrit.2018.03.011. [DOI] [PubMed] [Google Scholar]
- 28.Samara P, Athanasopoulos M, Mastronikolis S, et al. The Role of Oncogenic Viruses in Head and Neck Cancers: Epidemiology, Pathogenesis, and Advancements in Detection Methods. Microorganisms. 2024;12:1482. doi: 10.3390/microorganisms12071482. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Lee H, Cho J, Shin DW, et al. Association of cardiovascular health screening with mortality, clinical outcomes, and health care cost: A nationwide cohort study. Prev Med. 2015;70:19–25. doi: 10.1016/j.ypmed.2014.11.007. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data from the KNHIS can be accessed through their official website (http://nhiss.nhis.or.kr). To gain access, researchers must submit a study proposal, which must first be approved by their institutional review board. The proposal is then reviewed by the KNHIS review committee before access is granted.
Data may be obtained from a third party and are not publicly available.