Patient-Reported Sleep vs Objective Sleep and Physical Activity Measures in Ankylosing Spondylitis: A Pre-Post Study

Overview

Eighteen patients with active ankylosing spondylitis were evaluated at initiation of anti-TNF-α therapy and again after 4 months. Subjective sleep quality was assessed using the Pittsburgh Sleep Quality Index (PSQI). Objective sleep parameters were measured by overnight polysomnography, and physical activity was assessed using a multi-sensor wearable accelerometer. Clinical outcomes included BASDAI/BASFI, ASQoL, and HADS. This was a prospective, single-arm pre-post pilot study

Objective: This prospective, single-arm pre-post pilot study characterized patient-reported and objective sleep and physical activity measures in patients with active ankylosing spondylitis (AS) who were planned to start anti-TNF-α therapy as part of routine clinical care despite use of at least two NSAIDs. The primary purpose was to evaluate sleep and physical activity in relation to disease activity periods, rather than to test a prespecified treatment effect. Design and Participants: Eighteen adults fulfilling the modified New York criteria (mNYC) and classified as active based on BASDAI were enrolled. Assessments were performed at baseline (prior to starting anti-TNF-α therapy) and repeated at 4 months. Assessments: At both time points, participants completed patient-reported questionnaires and underwent objective measurements. Subjective sleep quality was measured using the Pittsburgh Sleep Quality Index (PSQI). Objective sleep was assessed by overnight polysomnography (PSG), and physical activity was assessed using a multi-sensor wearable accelerometer (SenseWear Armband, BodyMedia Inc., Pittsburgh, PA, USA). Clinical and psychosocial measures included BASDAI, BASFI, BASMI, ASQoL, and HADS (anxiety and depression). Given the exploratory aim and limited sample size, anti-TNF agent subtypes were not specified in the registry record. Primary Outcome (PSG): The primary outcome measure was sleep efficiency (%) derived from PSG, assessed at baseline and at 4 months. Additional PSG parameters included time in bed, sleep period time, total sleep time, stage durations (N1, N2, N3, and REM), arousal index, periodic limb movement index, apnea events, and apnea-hypopnea index (AHI). Wearable-derived parameters included total energy expenditure (kcal/day), total step count (steps/hour), total sedentary time (min/hour), sleep duration (min/hour), and physical activity duration (min/hour). Statistical Analysis: Distribution of continuous variables was assessed using the Shapiro-Wilk test. Non-normally distributed data were summarized as median (Q1-Q3). Pre-post comparisons were performed using the Wilcoxon signed-rank test. Exploratory Spearman rank correlations examined associations between changes in PSQI subscales and changes in BASDAI, BASFI, BASMI, ASQoL, HADS, and wearable-derived activity parameters. Multiple comparisons were controlled within hypothesis families using the Benjamini-Hochberg false discovery rate (FDR) procedure (FDR-adjusted p\<0.05). For sleep efficiency, the Hodges-Lehmann pseudo-median difference and 95% confidence interval were calculated in R; primary analyses were performed in SPSS (v24).