Individual Regression Modeling of Spinal Mobility Measurements in Long-Term Ankylosing Spondylitis: In-Depth Analyses With Comparison to Norm Data.

Abstract

Normal age-related decline and temporary restrictions in mobility complicate the understanding of spinal mobility deterioration over time in patients with ankylosing spondylitis (AS). In this study, we aimed to determine whether spinal mobility deterioration occurred linearly in patients with AS. We also aimed to compare patterns of change with corresponding age-related normal values and analyze variations in temporary fluctuations in mobility measurements over time. We included 111 men and 30 women (median age 20.9 years at symptom onset), who were followed for a median of 34 years since symptom onset. This inclusion resulted in 9,697 spinal mobility measurements for analysis. Individual linear regression models for development of lateral spinal flexion (LSF), the 10-cm Schober test (ST10), chest expansion (CE), and cervical rotation (CR) were analyzed and compared with normal age-related decline over time. The median values for the constants of all measurements were significantly lower than the norm data. However, LSF, ST10, and CE followed a yearly linear decline comparable to the norm data, whereas CR declined approximately twice as fast as expected from the norm data (beta median -0.62° [25th-75th percentile -1.16, -0.22] and -0.35° [25th-75th percentile -0.35, -0.35], respectively). Temporary fluctuations in LSF and CE were significantly higher during the early phase of the disease, with decreasing residuals over time. Based on median constants of our data, mobility restrictions related to AS seem to mainly occur during the first years of disease, indicating a narrow window of opportunity for prevention.