Abstract
The relationship between cataract and arthritis remains underexplored, highlighting the need for comprehensive investigation. This study aimed to examine the association between these 2 conditions using data from the National Health and Nutrition Examination Survey and Mendelian randomization (MR) analysis. We utilized National Health and Nutrition Examination Survey data from 1999 to 2008 and applied multiple statistical techniques, including logistic regression, subgroup analysis, and the k-nearest neighbors machine learning algorithm to evaluate associations. For causal inference, we performed MR analysis using inverse variance weighting (a method that combines genetic evidence across variants) to assess causality, with sensitivity analyses (e.g., Steiger filtering) to assess robustness. Arthritis and 11 covariates (e.g., age, gender) differed significantly between cataract and control groups. Logistic regression confirmed arthritis as a risk factor for cataract across adjusted models (odds ratio > 1, P < .05). The k-nearest neighbors model ranked age as the strongest predictor, with arthritis 3rd in predictive importance among 13 variables. MR analysis of 7 arthritis subtypes (including rheumatoid arthritis (RA), osteoarthritis, psoriatic arthritis, gout, lupus-related arthritis, fibromyalgia, and reactive arthritis) revealed a modest but significant causal effect of RA on cataract (odds ratio = 1.025 (1.007-1.044), P < .01). Sensitivity analyses supported robustness. Arthritis, particularly RA, is a novel risk factor for cataract, with implications for early screening and anti-inflammatory strategies in high-risk populations. While the MR effect size is small, this study integrates multi-method evidence (epidemiological, genetic, and machine learning), advancing understanding of systemic inflammation's role in ocular pathology.