You are here

Article Text

Article menu
Osteoarthritis
Increased risk of osteoarthritis in patients with atopic disease
  1. Matthew C Baker1,
  2. Khushboo Sheth1,2,
  3. Rong Lu3,
  4. Di Lu3,
  5. Ericka P von Kaeppler4,
  6. Archana Bhat5,
  7. David T Felson6,
  8. William H Robinson1,7
  1. 1 Division of Immunology and Rheumatology, Department of Medicine, Stanford University, Stanford, California, USA
  2. 2 Chinook Therapeutics Inc, Berkeley, California, USA
  3. 3 Quantitative Sciences Unit, Division of Biomedical Informatics Research, Department of Medicine, Stanford University, Stanford, California, USA
  4. 4 Institute for Stem Cell Biology and Regenerative Medicine, Stanford University, Stanford, California, USA
  5. 5 Research Informatics Center, Stanford University, Stanford, California, USA
  6. 6 Section of Rheumatology, Boston University School of Medicine, Boston, Massachusetts, USA
  7. 7 Division of Rheumatology, Palo Alto VA Medical Center, Palo Alto, California, USA
  1. Correspondence to Dr Matthew C Baker, Division of Immunology and Rheumatology, Department of Medicine, Stanford University, Stanford, CA 94304, USA; mbake13{at}stanford.edu

Abstract

Objectives To determine the incidence of osteoarthrits (OA) in patients with atopic disease compared with matched non-exposed patients.

Methods We conducted a retrospective cohort study with propensity score matching using claims data from Optum’s de-identified Clinformatics Data Mart (CDM) (January 2003 to June 2019) and electronic health record data from the Stanford Research Repository (STARR) (January 2010 to December 2020). We included adult patients without pre-existing OA or inflammatory arthritis who were exposed to atopic disease or who were non-exposed. The primary outcome was the development of incident OA.

Results In Optum CDM, we identified 117 346 exposed patients with asthma or atopic dermatitis (mean age 52 years; 60% female) and 1 247 196 non-exposed patients (mean age 50 years; 48% female). After propensity score matching (n=1 09 899 per group), OA incidence was higher in patients with asthma or atopic dermatitis (26.9 per 1000 person-years) compared with non-exposed patients (19.1 per 1000 person-years), with an adjusted odds ratio (aOR) of 1.58 (95% CI 1.55 to 1.62) for developing OA. This effect was even more pronounced in patients with both asthma and atopic dermatitis compared with non-exposed patients (aOR=2.15; 95% CI 1.93 to 2.39) and in patients with asthma compared with patients with chronic obstructive pulmonary disease (aOR=1.83; 95% CI 1.73 to 1.95). We replicated our results in an independent dataset (STARR), which provided the added richness of body mass index data. The aOR of developing OA in patients with asthma or atopic dermatitis versus non-exposed patients in STARR was 1.42 (95% CI 1.36 to 1.48).

Conclusions This study demonstrates an increased incidence of OA in patients with atopic disease. Future interventional studies may consider targeting allergic pathways for the prevention or treatment of OA.

  • osteoarthritis
  • epidemiology
  • inflammation
  • osteoarthritis, knee

Data availability statement

No data are available.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

http://dx.doi.org/10.1136/ard-2022-223640

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • Osteoarthritis (OA) is a common chronic disabling disease for which few effective treatments exist. Mounting evidence suggests that mast cell activation and allergic pathways may play key roles in OA pathogenesis. If so, OA occurrence may be increased in patients with atopic diseases; however, this currently remains unknown.

    WHAT THIS STUDY ADDS

    • In this propensity score matched study using administrative health claims data, we found a significantly increased incidence of OA in patients with asthma or atopic dermatitis or the combination of both asthma and atopic dermatitis. These results were confirmed in an academic institutional cohort.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • This study demonstrates an association between atopic disease and the development of OA; patients may benefit from the use of treatments that inhibit mast cells and allergic cytokines to treat or prevent OA.

    Introduction

    Osteoarthritis (OA) is the most common form of arthritis, with a lifetime risk of developing symptomatic knee OA as high as 44.7%.1 2 Osteoarthritis is associated with considerable morbidity—it is one of the major contributors to the global years lived with disability and the principal cause of lower extremity disability in older adults.3 Despite the high prevalence, substantial economic burden and debilitating impact of OA, the current therapeutic armamentarium is limited and focuses on symptom management. To date, there are no effective disease-modifying drugs that halt, slow or reverse the progression of OA.4

    Accumulating evidence implicates chronic, low-grade inflammation as a key driver of OA pathogenesis.5 We previously demonstrated increased activation and degranulation of mast cells in human OA synovium,6 and others demonstrated a correlation between synovial mast cell numbers and synovitis scores.7 In addition, we reported that IgE-mediated activation of mast cells and the mast cell mediator tryptase play key roles in driving the pathogenesis of OA in mice.6 Single nucleotide polymorphisms in interleukin 4 (IL-4)-associated genes, which are associated with type 2 inflammation and atopic disease, are also associated with OA.8–11 Together, these findings suggest that atopic disease might predispose or contribute to the development of OA in humans.

    The aim of this study was to investigate whether patients with atopic disease have an increased risk of developing OA. To address this question, we used a nationwide insurance claims database and electronic health records from a large academic institution to determine the risk of developing OA in patients with asthma and/or atopic dermatitis.

    Methods

    This study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for cohort studies and was approved by the Stanford University institutional review board (IRB-57463). There was no patient or public involvement in this study.

    Data source

    We used Optum’s de-identified Clinformatics Data Mart Database (CDM), a database derived from large, adjudicated claims data, from January 1, 2003 to June 30, 2019, and electronic health records from the Stanford Research Repository (STARR) from January 1, 2010 to December 31, 2020.12

    Study populations

    Optum CDM cohort

    The Optum CDM study population included individuals aged 18 years or older with at least 7 years of continuous enrolment in Optum CDM. Patients were excluded if they had a diagnosis of OA or atopic disease in the first 2 years of enrolment (defined by one or more International Classification of Disease (ICD) code 9 or 10 for OA or atopic disease) or any history of inflammatory arthritis based on one or more ICD-9 or ICD−10 code (online supplemental table 1). The exclusion of patients with prior OA or atopic disease was done to ensure the order of events: that patients did not have pre-existing OA, that patients in the exposed group then developed incident atopic disease, whereas patients in the non-exposed group did not develop atopic disease, and that the outcome of incident OA could then be assessed in both groups. It was necessary to include only patients who developed atopic disease after 2 years of enrolment, as including patients with atopic disease in their first year of enrolment would not allow for a lookback period of 2 years before this exposure, which was needed to exclude pre-existing OA and to ensure that we captured confounder information as comprehensively as possible.

    Supplemental material

    The exposed group included patients with a diagnosis of incident atopic disease, specifically asthma and/or atopic dermatitis. Asthma was defined as two or more ICD-9 or ICD-10 codes for asthma separated by 7 days or more, and a prescription for an inhaler after the earliest ICD diagnosis date.13–16 Atopic dermatitis was defined as two or more ICD-9 or ICD-10 codes for atopic dermatitis separated by 7 days or more, and in addition, at least two visits to a dermatologist and at least one prescription for an emollient, topical corticosteroid or topical calcineurin inhibitor.17

    The non-asthma/non-atopic dermatitis group (called the non-exposed group) was defined as patients without an ICD-9 or ICD-10 code for asthma or atopic dermatitis. Patients with only one ICD-9 or ICD-10 code for asthma or atopic dermatitis, and therefore not meeting the criteria for exposed or non-exposed were excluded from the analysis. For the secondary analysis in the Optum CDM cohort, patients with chronic obstructive pulmonary disease (COPD) were compared with patients with asthma. COPD was chosen as a comparator given the similarities with asthma—both are pulmonary diseases that are associated with increased healthcare use—however, COPD is not mediated by allergic pathology. Patients with COPD were defined as having two or more ICD-9 or ICD-10 codes for COPD separated by 7 days or more and a prescription for an inhaler (β-adrenergic agent, anticholinergic or inhaled glucocorticoid) after the earliest ICD diagnosis date for COPD.

    STARR cohort

    The study population for the STARR cohort included individuals from the Stanford Research Repository, aged 18 years or older, with at least 5 years continuous enrolment; 5 years of continuous follow-up rather than 7 years was chosen for this cohort in order to maximise the sample size in this smaller dataset. Patients were excluded if they had a diagnosis of OA or atopic disease in the first 2 years of enrolment (defined by one or more ICD-9 or ICD-10 code for OA or atopic disease) or any history of inflammatory arthritis based on one or more ICD-9 or ICD-10 code (online supplemental table 1). The same definitions as for the Optum CDM cohort for exposed and non-exposed were used.

    For both cohorts, we extracted data on age, sex, race/ethnicity, duration of follow-up and outpatient visit frequency. We applied R package ‘comorbidity’ to calculate the Charlson Comorbidity Score, using all ICD codes within the 5- year window around the index date for each patient.18 19 For the exposed group, the duration of follow-up started from the diagnosis date of atopic disease. For the non-exposed group, in order to deal with immortal time bias (time during which the patient cannot experience the outcome), we individually matched non-exposed patients with exposed patients using prescription time-distribution matching.20 This method generates an index date for the non-exposed patients by randomly sampling with replacement from the empirical distribution of the duration of immortal time from the exposed group. Thus, the index date for exposed patients was the date of diagnosis of atopic disease and the index date for non-exposed patients was the eligibility start date plus a randomly assigned duration of immortal time. This method helps to balance immortal time between the exposed and non-exposed groups.

    Outcome ascertainment

    The primary outcome was the development of OA, which was defined as two or more ICD-9 or ICD-10 codes for OA separated by 7 days or more. The use of two or more codes for OA has a positive predictive value of 0.85, sensitivity of 0.25, and specificity of 0.91 for the presence of knee, hand or hip OA,21 22 using the American College of Rheumatology classification criteria for OA as the standard,23 and these criteria have been used in previous OA studies.24–26 As a secondary outcome, we also analysed OA subtypes, including OA of the knee, hip and hands. Patients were followed up until the eligibility end date of the last insurance plan (ie, if there were gaps in coverage, their follow-up time included only the time they were actually in the dataset, and follow-up time ended when the last period of observation in the dataset ended).

    Statistical analysis

    Optum CDM cohort

    Baseline characteristics of patients with atopic disease and non-exposed patients were described. In the Optum CDM cohort, we conducted 1:1 propensity score matched analyses of (1) patients with asthma or atopic dermatitis versus non-exposed patients, (2) patients with both asthma and atopic dermatitis versus non-exposed patients and (3) patients with asthma versus patients with COPD. For all propensity score matched analyses, patients were matched on age, sex, race/ethnicity, education level, Charlson Comorbidity Score, duration of follow-up and outpatient visit frequency.18 Subjects who could not be matched were excluded. Standardised mean differences (SMDs) were calculated using the ‘stddiff’ package in R.27 Standardised mean differences were used to describe the magnitude of difference in characteristics between cohorts. An SMD≥0.20 was interpreted as meaningful. The magnitude of effect was considered small if SMD was ≥0.20 to <0.50, medium if SMD was ≥0.50 to <0.80 and large if SMD was ≥0.80. Logistic regression was used to estimate the propensity score. Complete case analysis was performed, as the level of missing variables was low (<1% for all continuous covariates). All missing categorical variables were coded as ‘unknown.’

    For the propensity score matched Optum CDM cohort, the relationship between atopic disease and the development of OA was evaluated using logistic regression. In our multivariable propensity score matched model, we also adjusted for age, sex, race/ethnicity, Charlson Comorbidity Score, education, duration of follow-up and frequency of outpatient visits. Raw incidence rates with 95% confidence intervals were calculated for each comparison group using the R package ‘epiR’, both before and after propensity score matching. Calculating the raw incidence rate using propensity score matched cohorts provided adjusted incidence rates (adjusted for all variables used in matching).

    STARR cohort

    To test the robustness of our results, we performed a second analysis using the STARR dataset, which included information on body mass index (BMI). For this analysis, we did not use propensity score matching owing to the smaller sample sizes. In order to assess the relationship between atopic disease and the development of OA, we calculated raw incidence rates as above and performed multivariable logistic regression, adjusting for age, sex, race/ethnicity, BMI, Charlson Comorbidity Score, duration of follow-up and frequency of outpatient visits. BMI was extracted at the index date and was present for roughly 97% of patients (otherwise listed as ‘unknown’ if missing).

    For both cohorts, all statistical analyses were conducted using R software, version 4.0.28 All 95% CIs and p values were based on two-sided hypothesis tests, where a p value of <0.05 was considered statistically significant.

    Results

    Optum CDM cohort before propensity score matching

    In the Optum CDM cohort, there were a total of 1 364 542 patients meeting inclusion and exclusion criteria, including 117 346 exposed patients with asthma or atopic dermatitis and 1 247 196 non-exposed patients (online supplemental figure 1). Before propensity score matching, for patients with asthma or atopic dermatitis compared with non-exposed patients, the mean age was 52 vs 50 years, the percentage female was 60% vs 48%, the mean Charlson Comorbidity Score was 1.6 vs 0.9, and the mean number of outpatient visits per year was 6.7 vs 6.0 (table 1).

    View this table:
    Table 1

    Baseline characteristics of patients in the Optum Clinformatics Data Mart cohort

    Before propensity score matching, the incidence rate of developing OA per 1000 person-years was 27.1 (95% CI 26.8 to 27.4) for patients with asthma or atopic dermatitis, 34.8 (95% CI 33.0 to 36.7) for patients with both asthma and atopic dermatitis, and 14.2 (95% CI 14.1 to 14.3) for non-exposed patients over 5 years or more of follow-up (table 2). Compared with non-exposed patients, after adjusting for age, sex, race/ethnicity, Charlson Comorbidity Score, education, duration of follow-up and frequency of outpatient visits, the aOR of developing OA in patients with asthma or atopic dermatitis was 1.76 (95% CI 1.73 to 1.79, p<0.001), and it was 2.41 (95% CI 2.25 to 2.59, p<0.001) in patients with both asthma and atopic dermatitis (table 2).

    View this table:
    Table 2

    Incidence and odds ratio of developing osteoarthritis (OA) in exposed patients with atopic disease versus non-exposed patients in Optum Clinformatics Data Mart.

    Optum CDM propensity score matched cohorts

    Asthma or atopic dermatitis versus non-exposed patients

    Patients with asthma or atopic dermatitis (n=109 899) were propensity score matched with non-exposed patients (n=109 899). The mean age in both cohorts was 52 years, with 59% female, a mean Charlson Comorbidity Score of 1.6, a mean number of outpatient visits per year of 7 and a mean follow-up time of 8 years (online supplemental table 2). The incidence rate of OA was 26.9 (95% CI 26.6 to 27.3) per 1000 person-years for patients with asthma or atopic dermatitis compared with 19.1 (95% CI 18.9 to 19.4) per 1000 person-years for non-exposed patients, with an aOR of developing OA in patients with asthma or atopic dermatitis compared with non-exposed patients of 1.58 (95% CI 1.55 to 1.62, p<0.001) (figure 1). For knee, hip and hand OA subtypes, the incidence rates of OA per 1000 person-years in patients with asthma or atopic dermatitis compared with non-exposed patients were 10.9 vs 7.5, 3.3 vs 2.3 and 2.3 vs 1.3 (online supplemental table 3).

    Figure 1
    Figure 1

    Risk of developing osteoarthritis in Optum Clinformatics Data Mart propensity score matched cohorts. COPD, chronic obstructive pulmonary disease.

    Both asthma and atopic dermatitis versus non-exposed patients

    After propensity score matching, patients with both asthma and atopic dermatitis (n=4325) and non-exposed patients (n=4325) were similar in age (53 years), sex (67% female), with a mean Charlson Comorbidity Score of 2, a mean number of outpatient visits per year of 8 and a mean follow-up time of 9 years (online supplemental table 4). The incidence rate of OA was 35.0 (95% CI 33.2 to 36.9) per 1000 person-years for patients with both asthma and atopic dermatitis compared with 21.5 (95% CI 20.0 to 22.9) per 1000 person-years for non-exposed patients. The aOR of developing OA in patients with both asthma and atopic dermatitis compared with non-exposed patients was 2.15 (95% CI 1.93 to 2.39, p<0.001) (figure 1). For knee, hip and hand OA subtypes, the incidence rates of OA per 1000 person-years in patients with both asthma and atopic dermatitis compared with non-exposed patients were 15.3 vs 9.1, 4.2 vs 3.0 and 3.9 vs 2.1 (online supplemental table 3).

    Asthma versus patients with COPD

    As a secondary analysis using the Optum CDM dataset, patients with asthma only (n=11 820) were compared with patients with COPD and no history of atopic disease (n=11 820). After propensity score matching, in both groups the average age was 65 years, 53% of patients were female, the mean Charlson Comorbidity Score was 3.4, the mean number of outpatients visits per year was 8, and the mean follow-up time was 8 years (online supplemental table 5). The incidence rate of OA was 41.2 (95% CI 39.9 to 42.5) per 1000 person-years for patients with asthma compared with 27.1 (95% CI 26.1 to 28.2) per 1000 person-years for patients with COPD. The aOR of developing OA was 1.83 (95% CI 1.73 to 1.95, p<0.001) for patients with asthma compared with patients with COPD (figure 1). The higher incidence rate of OA in patients with asthma compared with COPD was seen across knee, hip, and hand OA subtypes (16.3 vs 7.8, 5.1 vs 3.3 and 2.8 vs 1.6 respectively, all per 1000 person-years) (online supplemental table 3).

    STARR cohort

    The STARR cohort comprised 114 427 patients, including 43 728 exposed patients with asthma or atopic dermatitis and 70 699 non-exposed patients without atopic disease (online supplemental figure 2). Compared with non-exposed patients, those with asthma or atopic dermatitis were younger (mean age 51 vs 53 years), were more likely to be female (64% vs 61%), had a higher mean Charlson Comorbidity Score (0.9 vs 0.8), had fewer outpatient visits per year (3.9 vs 4.1) and had a slightly higher mean BMI (27.3 vs 27.0) (table 3).

    View this table:
    Table 3

    Baseline characteristics of patients in the STARR cohort

    The raw incidence rate for OA per 1000 person-years was 18.6 (95% CI 18.2 to 19.1) for patients with asthma or atopic dermatitis, 17.5 (95% CI 16.6 to 18.5) for patients with both asthma and atopic dermatitis, and 11.4 (95% CI 11.2 to 11.7) for non-exposed patients (table 4). After adjusting for age, sex, race/ethnicity, Charlson Comorbidity Score, duration of follow-up and frequency of outpatient visits, compared with non-exposed patients, the aOR of developing OA in patients with asthma or atopic dermatitis was 1.47 (95% CI 1.42 to 1.53, p<0.001), and the aOR of developing OA in patients with both asthma and atopic dermatitis was 1.25 (95% CI 1.16 to 1.34, p<0.001). After additionally adjusting for BMI, compared with non-exposed patients, the aOR of developing OA in patients with asthma or atopic dermatitis was 1.42 (95% CI 1.36 to 1.48, p<0.001), and the aOR of developing OA in patients with both asthma and atopic dermatitis was 1.19 (95% CI 1.11 to 1.28, p<0.001) (table 4).

    View this table:
    Table 4

    Incidence and odds ratio of developing osteoarthritis in exposed patients with atopic disease versus non-exposed patients in STARR

    Discussion

    In our study, compared with non-exposed patients, patients with asthma or atopic dermatitis had a 58% increased risk of developing OA in a nationwide insurance claims dataset and 42% increased risk of developing OA in a large academic institutional dataset, after adjusting for important covariables. Furthermore, in the Optum CDM cohort, after propensity score matching, patients with asthma had an 83% increased risk of developing OA compared with patients with COPD, and patients with both asthma and atopic dermatitis had 115% increased risk of developing OA compared with non-exposed patients. Notably, the risk of developing OA in patients with atopic disease was attenuated in the STARR cohort after adjusting for BMI compared with the Optum CDM cohort, but overall trends between the two cohorts were consistent. This is the first cohort study to our knowledge to demonstrate an association between atopic disease and the risk of developing OA.

    The association between atopic disease and OA support mechanistic studies that implicate a number of allergic pathways in the pathogenesis of OA: (1) activated and degranulating mast cells are statistically increased in the synovium of patients with OA,6 and there is a non-significant trend between synovial mast cell numbers and OA severity7; (2) mast cell activation and the release of tryptase mediates synovitis and joint tissue breakdown in OA6 29 30; (3) genomic analyses have identified associations between IL-4 and IL-4 receptor gene polymorphisms and the development of OA.8–10 31 Furthermore, increased mast cell numbers and elevated tryptase have also been implicated in asthma pathogenesis.32 Taken together, we hypothesise that pathways involved in atopic disease may contribute to the development of OA.

    In support of our results, prior work has shown an increased prevalence of OA in patients with asthma compared with patients with COPD and control patients.33 This prior study was limited by a small sample size, which included 425 patients with asthma (135 with OA) and 1131 patients with COPD (201 with OA). The cross-sectional nature of this prior study limited its ability to draw conclusions about the order of disease accrual and therefore whether atopic disease directly contributed to the development of OA. The findings of this prior study are supportive of our results; however, we extend the analysis to include the use of two independent, much larger cohorts, which include patients with atopic dermatitis in addition to asthma. Furthermore, the design of our study allows us to identify patients in whom the diagnosis of atopic disease definitively comes before the diagnosis of OA.

    The strengths of our study include a large sample size and validation of the results in an independent cohort. We were able to adjust for age, sex, Charlson Comorbidity Score, and outpatient visit frequency, which are all potential confounders of the relationship between atopic disease and OA. In the STARR cohort, we were able to further adjust for BMI, which confirmed the results in Optum CDM. Using COPD as a respiratory comparator disease to asthma, we demonstrated a strong association between asthma and the development of OA. Rheumatoid arthritis, which has a distinct pathophysiology from OA, has been shown to be associated with both asthma and COPD.34 Our study demonstrated a strong association between asthma and OA, but not between COPD and OA, further supporting the hypothesis that type 2 immune responses specifically contribute to the risk of developing OA, not pulmonary disease in general.

    Our study has several limitations. First, this is a retrospective study using claims data, and there may be residual or unmeasured confounders. In order to address this, we used propensity score matching to balance the covariates, and we adjusted for baseline characteristics, including age, sex, race/ethnicity, Charlson Comorbidity Score, outpatient visit frequency and the duration of follow-up. However, despite these efforts, the fact that the STARR cohort demonstrated that the risk of OA was attenuated after adjusting for BMI, given that asthma and atopic dermatitis are obesity related, supports the possibility of additional confounders.

    Second, in the Optum CDM dataset, we did not have data on important variables such as BMI, history of trauma to the joints or level of physical activity. We conducted a second analysis using electronic health record data in which we were able to adjust for BMI, which yielded similar trends. In addition, in the Optum CDM cohort for which we lacked BMI data, we found an increased risk of developing hand OA in patients with atopic disease compared with controls, and this manifestation of OA is not strongly or consistently associated with BMI.

    Third, our use of ICD codes might have resulted in misclassification of exposures and outcomes. We believe this is likely to be non-differential between the two groups. Patients with atopic disease had a higher mean Charlson Comorbidity Score and a higher number of yearly outpatient visits compared with non-exposed patients, which could have biased the results in either direction. If patients with atopic disease accrued more ICD codes due to more frequent visits, they might have been more likely to collect a diagnosis code for OA, biasing the results in favour of the hypothesis that atopic disease contributes to OA. If patients with atopic disease were sicker and therefore had additional non-OA ICD codes used to justify visits, OA might not have been coded, biasing the results towards the null. Further, we used a case finding approach for OA that is highly specific but not sensitive, and therefore might have missed incident OA cases. We assume that those missed were not differentially those with allergic or non-allergic conditions.

    Fourth, we did not have data regarding the severity of atopic disease, the severity of OA or any information about commonly used over-the-counter treatments for atopic disease (ie, antihistamines) or OA (ie, non-steroidal anti-inflammatory drugs (NSAIDs)). However, we were able to adjust for outpatient visit frequency and the Charlson Comorbidity Score, which may serve as proxies for disease severity. Despite this, if there is differential NSAID use in patients with atopic disease (ie, some consider NSAIDs contraindicated in asthma), OA severity might be higher in the atopic disease group and account for some of the differences seen.

    Fifth, it is difficult to separate the effect of atopic disease from the potential medication used to treat atopic disease, and thus we cannot definitively conclude that the association we observe is due to atopic disease itself.

    Sixth, we selected for patients with incident atopic disease to define a 2-year period before the first diagnosis of atopic disease in which we could rule out pre-existing OA. By excluding prevalent atopic disease at the time of enrolment in the dataset, we might have selected patients who developed atopic disease later in life, thus affecting generalisability. However, this provided confidence that the exposure of atopic disease came before the outcome of OA. In addition, we believe that excluding patients with longstanding atopic disease likely led to an underestimation of the association between atopic disease and the development of OA.

    Lastly, our study assumes that a patient has developed incident OA at the time of receiving an ICD code for OA, which might not be the case. It is possible that our outcome more accurately assesses when a patient has developed symptomatic OA to a large enough degree that the clinician acknowledges it through ICD coding.

    In conclusion, patients with atopic disease have an increased risk of developing OA compared with the general population. The association between atopic disease and OA is supported by recent observations that mast cells and type II cytokines may play important roles in the pathogenesis of OA broadly, not just in patients with atopic disease. Our findings provide further support for the concept that allergic pathways may contribute to the development of OA. If this is indeed true, non-atopic patients may also benefit from the use of treatments that inhibit mast cells and allergic cytokines to treat or prevent OA.

    Data availability statement

    No data are available.

    Ethics statements

    Patient consent for publication

    Acknowledgments

    We thank John and Jacque Jarve for their generous support of rheumatology clinical research at Stanford University. Optum CDM data for this project were accessed using the Stanford Center for Population Health Sciences Data Core. The PHS Data Core is supported by a National Institutes of Health National Center for Advancing Translational Science Clinical and Translational Science Award (UL1TR003142) and by Internal Stanford funding. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

    References

    1. ↵
      2. Zhang Y ,
      3. Jordan JM
      . Epidemiology of osteoarthritis. Clin Geriatr Med 2010;26:355–69. doi:10.1016/j.cger.2010.03.001
      OpenUrlCrossRefPubMed
    2. ↵
      2. Murphy L ,
      3. Schwartz TA ,
      4. Helmick CG , et al
      . Lifetime risk of symptomatic knee osteoarthritis. Arthritis Rheum 2008;59:1207–13. doi:10.1002/art.24021
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      2. GDaIIaP C
      . Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet 2018;392:1789–858. doi:10.1016/S0140-6736(18)32279-7
      OpenUrlCrossRefPubMed
    4. ↵
      2. Hunter DJ ,
      3. Bierma-Zeinstra S
      . Osteoarthritis. Lancet 2019;393:1745–59. doi:10.1016/S0140-6736(19)30417-9
      OpenUrlCrossRefPubMed
    5. ↵
      2. Robinson WH ,
      3. Lepus CM ,
      4. Wang Q , et al
      . Low-grade inflammation as a key mediator of the pathogenesis of osteoarthritis. Nat Rev Rheumatol 2016;12:580–92. doi:10.1038/nrrheum.2016.136
      OpenUrlPubMed
    6. ↵
      2. Wang Q ,
      3. Lepus CM ,
      4. Raghu H , et al
      . IgE-mediated mast cell activation promotes inflammation and cartilage destruction in osteoarthritis. Elife 2019;8:e39905. doi:10.7554/eLife.39905
    7. ↵
      2. de Lange-Brokaar BJE ,
      3. Kloppenburg M ,
      4. Andersen SN , et al
      . Characterization of synovial mast cells in knee osteoarthritis: association with clinical parameters. Osteoarthritis Cartilage 2016;24:664–71. doi:10.1016/j.joca.2015.11.011
      OpenUrl
    8. ↵
      2. Vargiolu M ,
      3. Silvestri T ,
      4. Bonora E , et al
      . Interleukin-4/interleukin-4 receptor gene polymorphisms in hand osteoarthritis. Osteoarthritis Cartilage 2010;18:810–6. doi:10.1016/j.joca.2010.02.005
      OpenUrlPubMed
    9. ↵
      2. Rogoveanu OC ,
      3. Calina D ,
      4. Cucu MG , et al
      . Association of cytokine gene polymorphisms with osteoarthritis susceptibility. Exp Ther Med 2018;16:2659–64. doi:10.3892/etm.2018.6477
      OpenUrl
    10. ↵
      2. Forster T ,
      3. Chapman K ,
      4. Loughlin J
      . Common variants within the interleukin 4 receptor alpha gene (IL4R) are associated with susceptibility to osteoarthritis. Hum Genet 2004;114:391–5. doi:10.1007/s00439-004-1083-0
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      2. McCormick SM ,
      3. Heller NM
      . Commentary: IL-4 and IL-13 receptors and signaling. Cytokine 2015;75:38–50. doi:10.1016/j.cyto.2015.05.023
      OpenUrlCrossRefPubMed
    12. ↵
      1. Stanford Center for Population Health Sciences
      . Optum SES. 2020. Available: https://doi.org/10.57761/a5hg-b442
    13. ↵
      2. Nissen F ,
      3. Quint JK ,
      4. Wilkinson S , et al
      . Validation of asthma recording in electronic health records: protocol for a systematic review. BMJ Open 2017;7:e014694. doi:10.1136/bmjopen-2016-014694
    14. ↵
      2. Rubak S ,
      3. Høst A ,
      4. Christensen LB , et al
      . Validity of asthma diagnoses and patterns of anti-asthmatic drug use in a cohort of 2053 Danish children. Health Sci Rep 2018;1:e77. doi:10.1002/hsr2.77
    15. ↵
      2. Quint JK ,
      3. Müllerova H ,
      4. DiSantostefano RL , et al
      . Validation of chronic obstructive pulmonary disease recording in the Clinical Practice Research Datalink (CPRD-GOLD). BMJ Open 2014;4:e005540. doi:10.1136/bmjopen-2014-005540
    16. ↵
      2. Nissen F ,
      3. Quint JK ,
      4. Wilkinson S , et al
      . Validation of asthma recording in electronic health records: a systematic review. Clin Epidemiol 2017;9:643–56. doi:10.2147/CLEP.S143718
      OpenUrlCrossRefPubMed
    17. ↵
      2. Dizon MP ,
      3. Yu AM ,
      4. Singh RK , et al
      . Systematic review of atopic dermatitis disease definition in studies using routinely collected health data. Br J Dermatol 2018;178:1280–7. doi:10.1111/bjd.16340
      OpenUrl
    18. ↵
      2. Quan H ,
      3. Li B ,
      4. Couris CM , et al
      . Updating and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge Abstracts using data from 6 countries. Am J Epidemiol 2011;173:676–82. doi:10.1093/aje/kwq433
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      2. Gasparini A
      . Comorbidity: an R package for computing comorbidity scores. JOSS 2018;3:648. doi:10.21105/joss.00648
      OpenUrl
    20. ↵
      2. Zhou Z ,
      3. Rahme E ,
      4. Abrahamowicz M , et al
      . Survival bias associated with time-to-treatment initiation in drug effectiveness evaluation: a comparison of methods. Am J Epidemiol 2005;162:1016–23. doi:10.1093/aje/kwi307
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      2. Rahman MM ,
      3. Kopec JA ,
      4. Goldsmith CH , et al
      . Validation of administrative osteoarthritis diagnosis using a clinical and radiological population-based cohort. Int J Rheumatol 2016;2016:6475318. doi:10.1155/2016/6475318
    22. ↵
      2. Shrestha S ,
      3. Dave AJ ,
      4. Losina E , et al
      . Diagnostic accuracy of administrative data algorithms in the diagnosis of osteoarthritis: a systematic review. BMC Med Inform Decis Mak 2016;16:82. doi:10.1186/s12911-016-0319-y
    23. ↵
      2. Altman R ,
      3. Asch E ,
      4. Bloch D , et al
      . Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and therapeutic committee of the American Rhemuatism Association. Arthritis Rheum 1986;29:1039–49. doi:10.1002/art.1780290816
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      2. Marshall DA ,
      3. Vanderby S ,
      4. Barnabe C , et al
      . Estimating the burden of osteoarthritis to plan for the future. Arthritis Care Res (Hoboken) 2015;67:1379–86. doi:10.1002/acr.22612
      OpenUrl
    25. ↵
      2. Atiquzzaman M ,
      3. Karim ME ,
      4. Kopec J , et al
      . Role of nonsteroidal antiinflammatory drugs in the association between osteoarthritis and cardiovascular diseases: a longitudinal study. Arthritis Rheumatol 2019;71:1835–43. doi:10.1002/art.41027
      OpenUrl
    26. ↵
      2. Baker MC ,
      3. Weng Y ,
      4. Robinson WH , et al
      . Reduction in osteoarthritis risk after treatment with ticagrelor compared to clopidogrel: a propensity score-matching analysis. Arthritis Rheumatol 2020;72:1829–35. doi:10.1002/art.41412
      OpenUrl
    27. ↵
      2. Du Z
      . Stddiff: calculate the standardized difference for numeric, binary and category variables.: R package version 3.1. 2022.
    28. ↵
      1. R Core Team
      . R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing, 2022.
    29. ↵
      2. Buckley MG ,
      3. Gallagher PJ ,
      4. Walls AF
      . Mast cell subpopulations in the synovial tissue of patients with osteoarthritis: selective increase in numbers of tryptase-positive, chymase-negative mast cells. J Pathol 1998;186:67–74. doi:10.1002/(SICI)1096-9896(199809)186:1<67::AID-PATH132>3.0.CO;2-D
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      2. Ioan-Facsinay A
      . Initiating pain in osteoarthritis (OA): is it the mast cell? Osteoarthritis Cartilage 2018;26:1–3. doi:10.1016/j.joca.2017.10.005
      OpenUrl
    31. ↵
      2. Yigit S ,
      3. Inanir A ,
      4. Tekcan A , et al
      . Significant association of interleukin-4 gene intron 3 VNTR polymorphism with susceptibility to knee osteoarthritis. Gene 2014;537:6–9. doi:10.1016/j.gene.2013.12.060
      OpenUrl
    32. ↵
      2. Zhang MQ ,
      3. Timmerman H
      . Mast cell tryptase and asthma. Mediators Inflamm 1997;6:311–7. doi:10.1080/09629359791433
      OpenUrlCrossRefPubMed
    33. ↵
      2. Koo HK ,
      3. Song P ,
      4. Lee JH
      . Novel association between asthma and osteoarthritis: a nationwide health and nutrition examination survey. BMC Pulm Med 2021;21:59. doi:10.1186/s12890-021-01425-6
    34. ↵
      2. Ford JA ,
      3. Liu X ,
      4. Chu SH , et al
      . Asthma, chronic obstructive pulmonary disease, and subsequent risk for incident rheumatoid arthritis among women: a prospective cohort study. Arthritis Rheumatol 2020;72:704–13. doi:10.1002/art.41194
      OpenUrl

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Handling editor Josef S Smolen

    • Twitter @mbake013

    • MCB and KS contributed equally.

    • Presented at Preliminary data from this study were previously presented as a poster at the 2021 American College of Rheumatology (ACR) Meeting: Baker M, Sheth K, Lu R, Lu D, von Kaeppler E, Bhat A, Felson D, Robinson W. Osteoarthritis risk Is increased in patients with atopic disease [abstract]. Arthritis Rheumatol 2021;73 (suppl 10). https://acrabstracts.org/abstract/osteoarthritis-risk-is-increased-in-patients-with-atopic-disease/. Accessed 20 January 2023.

    • Contributors RL and DL performed cohort selection and statistical analysis of the Optum data (Stanford Center for Population Health Sciences, 2020). Optum SES (v3.0). Redivis. (dataset) https://doi.org/10.57761/a5hg-b442) in the study, and AB had full access to the STARR data in the study. These authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. All authors were involved in the planning, analysis and writing of this paper. MCB accepts full responsibility for the work and/or the conduct of the study, had access to the data and controlled the decision to publish.

    • Funding MCB received support for this work from the NIH R25 AI 147369. WR received support for this work from the Department of Veterans Affairs Merit Awards I01RX002689 and I01CX002523, and Department of Defense Award W81XWH-18-1-0590. As part of the Stanford Center for Clinical and Translational Research and Education, RL and DL received support from UL1-TR003142. DF received support for this work from NIH P30 072571.

    • Competing interests None declared.

    • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.