Abstract
This cohort study evaluates the risk of developing inflammatory arthritis among patients with hidradenitis suppurativa.
Hidradenitis suppurativa (HS) has been associated with a high prevalence of spondyloarthritis; however, the population-based risk of developing inflammatory arthritis remains unclear.1 This population-based cohort study sought to evaluate the risk of developing inflammatory arthritis among patients with HS compared with matched patients without HS.
Methods
Patients
We used longitudinal claims data from commercially insured patients, including those with Medicaid and Medicare, covering 185 million lives in the United States between January 1, 2003, and January 1, 2016. We identified patients of all ages who received a diagnosis of HS (International Classification of Diseases, Ninth Revision code 705.83 or International Statistical Classification of Diseases and Related Health Problems, Tenth Revision code L73.2), defined as 1 diagnosis by a dermatologist or 3 diagnoses by any health care professional.2 The cohort entry date for the HS group was the first recorded diagnosis date of HS after at least 180 days of continuous enrollment (eFigure in the Supplement). For the non-HS group, we risk-set sampled 2 patients without HS from all plan enrollees who did not have a diagnosis of HS matched on the date the patient with HS entered the cohort. We excluded patients who had less than 180 days of continuous enrollment before cohort entry or had preexisting inflammatory arthritis, including ankylosing spondylitis, psoriatic arthritis, other spondyloarthritis, or rheumatoid arthritis (eAppendix in the Supplement). The Brigham and Women’s Hospital’s Institutional Review Board approved this study. All patient information was deidentified.
Outcomes
We used validated algorithms using hospital discharge diagnoses or physician visit claims to identify the following outcomes: ankylosing spondylitis, psoriatic arthritis, other spondyloarthritis (reactive arthropathy, spinal enthesopathy, sacroiliitis, or unspecified inflammatory spondylopathies), and rheumatoid arthritis (eTable 1 in the Supplement).3,4 Follow-up started the day after cohort entry and lasted until one of the following events occurred: outcome, death, disenrollment, or end of data stream, whichever came first (eTable 3 in the Supplement).
Baseline Patient Characteristics
All the characteristics of the patients were assessed during the 180 days before or on the cohort entry date: age, sex, region, number of outpatient visits, number of unique medications, use of systemic biologic or nonbiologic immunomodulatory agents, comorbidities (ie, psoriasis or inflammatory bowel disease), and combined comorbidity score (eTable 2 in the Supplement).5
Statistical Analysis
Statistical analysis was performed from January 1, 2003, to January 1, 2016. We estimated hazard ratios (HRs) of the outcomes with 95% CIs by fitting Cox proportional hazards regression models, adjusting for all 10 baseline characteristics. To further account for confounding, HRs of developing inflammatory arthritis were computed after 1:1 propensity score matching.6 Patients’ propensity scores were estimated using the multivariable logistic regression that included all aforementioned patient characteristics.
Results
We identified 70 697 patients with HS (mean [SD] age, 36.5 [14.7] years) and 141 412 risk-set sampled patients without HS (mean [SD] age, 38.3 [21.1] years). Before matching, patients with HS were more often female (78.0% vs 52.0%) and had an increased comorbidity burden (Table 1). Median follow-up was 1.5 years (interquartile range, 0.6-3.1 years), with a maximum of 11 years. Both groups were free of arthritis prior to the cohort entry. After propensity score matching, patients with HS had an increased risk for developing ankylosing spondylitis compared with those without HS (incidence rate, 0.60 vs 0.36 per 1000; HR, 1.65 [95% CI, 1.15-2.35]), psoriatic arthritis (incidence rate, 0.84 vs 0.58 per 1000; HR, 1.44 [95% CI, 1.08-1.93]), and rheumatoid arthritis (incidence rate, 4.54 vs 3.86 per 1000; HR, 1.16 [95% CI, 1.03-1.31]) (Table 2). We did not observe an increased risk of other spondyloarthritis among patients with HS vs those without HS (incidence rate, 3.07 vs 3.00 per 1000; HR, 1.02 [95% CI, 0.89-1.17]).
Table 1. Characteristics of Patients Before and After 1:1 Propensity Score Matching.
| Characteristic, 180 d Before Cohort Entry | Propensity Score Matching, No. (%) | |||
|---|---|---|---|---|
| Before | After | |||
| HS (n = 70 697) | Non-HS (n = 141 412) | HS (n = 60 872) | Non-HS (n = 60 872) | |
| Age, y | ||||
| Mean (SD) | 36.5 (14.7) | 38.3 (21.1) | 36.6 (14.7) | 37.6 (20.9) |
| Median (IQR) | 35 (25-47) | 40 (20-55) | 35 (25-47) | 38 (20-54) |
| Sex | ||||
| Male | 15 526 (22.0) | 67 872 (48.0) | 14 916 (24.5) | 14 157 (23.3) |
| Female | 55 171 (78.0) | 73 539 (52.0) | 45 956 (75.5) | 46 715 (76.7) |
| Region | ||||
| Northeast | 13 320 (18.8) | 24 390 (17.2) | 11 276 (18.5) | 11 728 (19.3) |
| North central | 15 047 (21.3) | 32 775 (23.2) | 13 203 (21.7) | 13 085 (21.5) |
| South | 31 635 (44.7) | 54 435 (38.5) | 26 668 (43.8) | 27 150 (44.6) |
| West | 9182 (13.0) | 27 233 (19.3) | 8450 (13.9) | 7587 (12.5) |
| Unknown | 1513 (2.1) | 2578 (1.8) | 1275 (2.1) | 1322 (2.2) |
| Comorbidities | ||||
| Psoriasis | 1438 (2.0) | 427 (0.3) | 695 (1.1) | 422 (0.7) |
| Inflammatory bowel disease | 1028 (1.5) | 488 (0.3) | 548 (0.9) | 425 (0.7) |
| Combined comorbidity score, mean (SD) | 0.15 (0.85) | 0.11 (0.78) | 0.13 (0.82) | 0.12 (0.86) |
| No. of outpatient visits | ||||
| Mean (SD) | 8.34 (8.67) | 4.13 (6.79) | 7.05 (7.21) | 6.91 (8.91) |
| Median (IQR) | 6.00 (3.00-10.00) | 2.00 (0.00-5.00) | 5.00 (3.00-9.00) | 4.00 (2.00-8.00) |
| No. of unique medications | ||||
| Mean (SD) | 5.88 (6.96) | 3.26 (5.45) | 4.98 (6.03) | 5.17 (6.79) |
| Median (IQR) | 4.00 (0.00-9.00) | 1.00 (0.00-5.00) | 3.00 (0.00-7.00) | 3.00 (0.00-8.00) |
| Use of immunomodulatory agent | ||||
| Biologic systemica | 891 (1.3) | 405 (0.3) | 493 (0.8) | 348 (0.6) |
| Nonbiologic systemicb | 1172 (1.7) | 1177 (0.8) | 796 (1.3) | 776 (1.3) |
Abbreviations: HS, hidradenitis suppurativa; IQR, interquartile range.
Biologic agents: adalimumab, etanercept, infliximab, certolizumab, golimumab, ustekinumab, ixekizumab, secukinumab, guselkumab, tofacitinib, rituximab, and tocilizumab.
Nonbiologic systemic agents: mycophenolate, methotrexate, azathioprine, sulfasalazine, cyclosporine, hydroxychloroquine, and leflunomide.
Table 2. Rate of Newly Recorded Inflammatory Arthritis, Before and After 1:1 Propensity Score Matching.
| Unmatched | HS (n = 70 697) | Non-HS (n = 141 412) | HS vs Non-HS | |||||
|---|---|---|---|---|---|---|---|---|
| Person-Time in 1000 Person-Years | No. of Events | Incidence Rate per 1000 Person-Years | Person-Time in 1000 Person-Years | No. of Events | Incidence Rate per 1000 Person-Years | HR (95% CI) | Rate Difference per 1000 Person-Years (95% CI) | |
| Outcome of interest | ||||||||
| Ankylosing spondylitis | 154 135 | 97 | 0.63 | 313 389 | 79 | 0.25 | 2.48 (1.84 to 3.33) | 0.38 (0.24 to 0.51) |
| Psoriatic arthritis | 153 933 | 190 | 1.23 | 313 287 | 112 | 0.36 | 3.41 (2.70 to 4.31) | 0.88 (0.69 to 1.06) |
| Other spondyloarthritisa | 153 237 | 593 | 3.87 | 312 552 | 582 | 1.86 | 2.07 (1.85 to 2.32) | 2.01 (1.66 to 2.35) |
| Rheumatoid arthritis | 152 417 | 829 | 5.44 | 311 844 | 746 | 2.39 | 2.25 (2.04 to 2.48) | 3.05 (2.64 to 3.45) |
| 1:1 Propensity Score Matched | HS (n = 60 872) | Non-HS (n = 60 872) | ||||||
| Outcome of interest | ||||||||
| Ankylosing spondylitis | 133 012 | 80 | 0.60 | 135 345 | 49 | 0.36 | 1.65 (1.15 to 2.35) | 0.24 (0.07 to 0.41) |
| Psoriatic arthritis | 132 928 | 112 | 0.84 | 135 250 | 78 | 0.58 | 1.44 (1.08 to 1.93) | 0.27 (0.06 to 0.47) |
| Other spondyloarthritisa | 132 422 | 407 | 3.07 | 134 755 | 404 | 3.00 | 1.02 (0.89 to 1.17) | 0.08 (–0.34 to 0.49) |
| Rheumatoid arthritis | 131 749 | 598 | 4.54 | 134 310 | 519 | 3.86 | 1.16 (1.03 to 1.31) | 0.67 (0.18 to 1.17) |
Abbreviations: HR, hazard ratio; HS, hidradenitis suppurativa.
Other spondyloarthritis (excluding ankylosing spondylitis): spinal enthesopathy, sacroiliitis, reactive arthritis, unspecified inflammatory spondyloarthritis.
Discussion
This large cohort study provides population-based rates of newly recorded inflammatory arthritis after a diagnosis of HS. We observed increased risks of developing ankylosing spondylitis, psoriatic arthritis, and rheumatoid arthritis among patients with HS compared with those without HS. Generally, the incidence of arthritis was low, resulting in only between 2 and 6 additional cases per 10 000 patients with HS per 1.5 years. Although the data support a systematic association between HS and subsequent newly diagnosed inflammatory joint disease, the low incremental risk is reassuring. Nevertheless, physicians treating patients with HS should be aware of symptoms suggestive of inflammatory arthritis, including morning stiffness and joint pain or swelling. This study cannot prove a causal relationship between HS and inflammatory arthritis, and further work is needed to elucidate the underlying potential shared pathogenesis of these disorders.
eAppendix. Methods
eFigure. Overview of Longitudinal Study Design
eTable 1. International Classification of Disease (ICD) Codes Used for Creating This Study’s Outcome Measures
eTable 2. Components Included in the Combined Comorbidity Score
eTable 3. Follow-up Time and Reasons for Censoring After Propensity Score Matching
References
- 1.Rondags A, Arends S, Wink FR, Horváth B, Spoorenberg A. High prevalence of hidradenitis suppurativa symptoms in axial spondyloarthritis patients: a possible new extra-articular manifestation. Semin Arthritis Rheum. 2019;48(4):611-617. doi: 10.1016/j.semarthrit.2018.03.010 [DOI] [PubMed] [Google Scholar]
- 2.Kim GE, Shlyankevich J, Kimball AB. The validity of the diagnostic code for hidradenitis suppurativa in an electronic database. Br J Dermatol. 2014;171(2):338-342. doi: 10.1111/bjd.13041 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Curtis JR, Harrold LR, Asgari MM, et al. . Diagnostic prevalence of ankylosing spondylitis using computerized health care data, 1996 to 2009: underrecognition in a US health care setting. Perm J. 2016;20(4):15-151. doi: 10.7812/TPP/15-151 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hanly JG, Thompson K, Skedgel C. The use of administrative health care databases to identify patients with rheumatoid arthritis. Open Access Rheumatol. 2015;7:69-75. doi: 10.2147/OARRR.S92630 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Gagne JJ, Glynn RJ, Avorn J, Levin R, Schneeweiss S. A combined comorbidity score predicted mortality in elderly patients better than existing scores. J Clin Epidemiol. 2011;64(7):749-759. doi: 10.1016/j.jclinepi.2010.10.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika. 1983;70:41-55. doi: 10.1093/biomet/70.1.41 [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eAppendix. Methods
eFigure. Overview of Longitudinal Study Design
eTable 1. International Classification of Disease (ICD) Codes Used for Creating This Study’s Outcome Measures
eTable 2. Components Included in the Combined Comorbidity Score
eTable 3. Follow-up Time and Reasons for Censoring After Propensity Score Matching