0523 Evaluating Symptom-Based Screening Tools for Obstructive Sleep Apnea in Emergency Responders

Abstract Introduction Fire Department of the City of New York (FDNY) emergency responders have a high prevalence of obstructive sleep apnea (OSA), yet the performance of oft-used subjective symptom instruments in this population is not fully understood. Clarifying relationships among the Epworth Sleepiness Scale (ESS), Berlin Questionnaire (Berlin), and physiologic severity (apnea-hypopnea-index, AHI) may improve OSA screening among responders referred for polysomnography. Methods We analyzed FDNY emergency responders who performed clinically indicated diagnostic polysomnography and completed ESS and Berlin. We categorized ESS as normal (0-10) or high (10). We summarized Berlin risk as the number of positive domains (0-3) and dichotomized as low (0-1) or high (2-3). We combined dichotomized ESS and Berlin measures creating four symptom phenotypes - low Berlin/low ESS, low Berlin/high ESS, high Berlin/low ESS, and high Berlin/high ESS. We compared AHI across ESS, Berlin, and phenotype using log-binomial regression to estimate risk ratios (RRs) for moderately severe OSA (AHI≥15), adjusting for age. Results Among 1,863 responders performing polysomnograms, the median AHI was 14.3 (IQR=5.7–28.5). Both age and BMI increased with AHI severity. Most participants had normal ESS scores (1,213; 63.1%), despite substantial Berlin symptom burden. 1,262 (65.6%) screened positive on ≥2 domains with phenotypes high Berlin/low ESS (713; 38%), high Berlin/high ESS (N=531; 29%), low Berlin/low ESS (N=462; 25%), and low Berlin/high ESS (N=157; 8%). AHI increased across Berlin domain counts (median 8.2, 10.4, 15.6, and 17.2 for 0, 1, 2, and 3 domains, respectively) but varied minimally across ESS categories. In adjusted models, symptom phenotype predicted OSA. Compared with low Berlin/low ESS, risk of OSA were higher for high Berlin/high ESS (RR=1.51; 95%CI=1.30-1.75) and for high Berlin/low ESS (RR=1.63; 95%CI=1.42-1.87), while low Berlin/high ESS was not associated (RR=1.15; 95%CI=0.91-1.45). Conclusion In this clinically selected WTC-exposed cohort referred for polysomnography, Berlin symptom burden, but not ESS sleepiness, closely reflected physiologic OSA severity. The high Berlin/low ESS group showed that many responders had substantial OSA risk despite minimal subjective sleepiness. Support (if any)