Risk Factors for Diagnosis of Psoriatic Arthritis, Psoriasis, Rheumatoid Arthritis, and Ankylosing Spondylitis: A Set of Parallel Case-control Studies

Elana Meer, Telma Thrastardottir, Xingmei Wang, Maureen Dubreuil, Yong Chen, Joel M. Gelfand, Þorvarður Jón Löve, Alexis OgdieThe Journal of Rheumatology 2021 · DOI

Abstract

Shared and different risk factors for psoriatic arthritis (PsA), psoriasis (PsO), rheumatoid arthritis (RA), and ankylosing spondylitis (AS) were identified. Statin use was inversely associated with all four conditions.

Introduction

Inflammatory arthritis (IA) and psoriasis are chronic, debilitating conditions affecting approximately 4% of the adult population. Early diagnosis is critical; for many patients, irreversible joint damage begins within the first year of symptom onset. Identifying clinical "red flags" or risk factors before a formal diagnosis could enable earlier intervention and significantly improve long-term patient outcomes.

However, these diseases often share overlapping symptoms and genetic backgrounds. The challenge lies in distinguishing between risk factors unique to a specific disease and those that serve as general markers of systemic inflammation. This study systematically maps these associations to facilitate the development of better screening algorithms within Electronic Medical Records (EMRs).

Methodology

Researchers conducted four massive, parallel case-control studies using data from The Health Improvement Network (THIN), a longitudinal database of primary care records in the United Kingdom (1994–2015). The study included 7,594 PsA cases, 111,375 PsO cases, 28,341 RA cases, and 3,253 AS cases, matched against over 1.5 million controls based on age, sex, general practice, and year of diagnosis.

To ensure the analysis focused on pre-diagnosis risk factors, the team required at least 12 months of medical history and assessed exposures up to the diagnosis date. Multivariable logistic regression with automated stepwise selection was used to identify the most significant predictors from over 100 potential covariates, including comorbidities, infections, and medications.

Key Findings

The study quantified risk using Odds Ratios (OR), where an OR > 1 indicates increased risk and < 1 indicates decreased risk.

The "Statin Shield" and Medication Effects

One of the most consistent findings was that statin use was associated with a 40–50% lower likelihood of being diagnosed with any of the four diseases. This suggests that the pleiotropic anti-inflammatory effects of statins might inhibit the development of these conditions.

Interestingly, diabetes also showed a negative association. This is likely not a protective effect of the disease itself, but rather a result of common diabetes medications, such as metformin, which possess potent anti-inflammatory properties that may mask or delay the onset of arthritis.

Table 2: Risk factors and Odds Ratios (OR) for PsA, PsO, RA, and AS. Blue highlights indicate protective factors (OR < 1), while yellow highlights indicate increased risk factors (OR > 1).

Disease-Specific Triggers

The results reinforced the "distal trigger" theory for inflammatory diseases:

  • Infections: Pharyngitis and skin infections were specific precursors for PsA, while gastrointestinal infections were linked to AS.
  • Physical Trauma: Trauma to joints and bones significantly predicted PsA and RA, supporting the "deep Koebner phenomenon," where physical injury triggers an inflammatory response in susceptible individuals.
  • Ankylosing Spondylitis (AS): Showed the most dramatic associations, with uveitis (eye inflammation) having a staggering OR of 37.97 and Inflammatory Bowel Disease (IBD) an OR of 5.46.
  • Lifestyle Factors: Smoking was a universal risk factor, particularly for RA and PsO. Obesity was strongly linked to the "psoriatic spectrum" (PsA OR 1.64; PsO OR 1.27).
Table 3: Summary of specific risk factors identified for each condition.

Discussion and Limitations

While the study provides a comprehensive map of risk factors, several limitations exist. As a case-control study using EMR data, establishing causality is difficult. Some factors, like uveitis, might be early symptoms rather than causes (protopathic bias). Additionally, EMR data accuracy depends on clinician coding.

The ultimate goal is to transition from identifying individual factors to creating integrated predictive risk scores. By combining these variables, researchers hope to build EMR tools that automatically flag high-risk patients for early rheumatology referrals, potentially preventing irreversible damage.

Summary updated

Abstract

Shared and different risk factors for PsA, PsO, RA, and AS were identified. Statin use was inversely associated with all 4 conditions.
BibTeX
@article{meer2021risk,
  author = {Elana Meer and Telma Thrastardottir and Xingmei Wang and Maureen Dubreuil and Yong Chen and Joel M. Gelfand and Þorvarður Jón Löve and Alexis Ogdie},
  title = {{Risk Factors for Diagnosis of Psoriatic Arthritis, Psoriasis, Rheumatoid Arthritis, and Ankylosing Spondylitis: A Set of Parallel Case-control Studies}},
  journal = {The Journal of Rheumatology},
  year = {2021},
  doi = {10.3899/jrheum.210006},
}
Figure