0628 Distribution of Cardiometabolic Complications in U. S. Veterans Before and After Sleep Testing

Abstract Introduction Obstructive sleep apnea (OSA) is a prevalent condition among Veterans affecting 24% and ranking among the top five most common diagnoses in the Veterans Health Administration (VHA). Epidemiologic studies demonstrate a strong association between OSA and cardiometabolic disease. This study describes the temporal distribution of common cardiometabolic comorbidities in relationship to the first sleep test within VHA. Methods This study is a sub-analysis of data from a larger project leveraging Artificial Intelligence (AI) to predict which Veterans with OSA are most likely to develop cardiometabolic comorbidities. Data from 24 million Veteran records collected between 1999 and 2022 were analyzed using Current Procedural Terminology for diagnostic sleep testing, and International Classification of Disease 9 and 10 codes for OSA and cardiometabolic comorbidities. Prevalence of comorbidities was documented within two years prior to the first sleep study (T0) and for up to 22 years afterwards. A preliminary logistic regression model was used to predict the likelihood of developing an additional comorbidity after T0. Results The analysis identified 1.3 million Veterans who underwent sleep testing within VHA, with 939,954 having at least one comorbidity. Within two years prior to T0 the following comorbidities were observed: hypertension 54%, diabetes mellitus 29%, coronary artery disease 22%, heart failure 8.5%, stroke 8.9%, and atrial fibrillation 7.4%. After T0 the number of Veterans with these comorbidities increased: hypertension 60%, diabetes mellitus 42%, coronary artery disease 33%, heart failure 18%, stroke 16%, and atrial fibrillation 15%. At T0 44% had at least two comorbidities, 21% had at least three comorbidities, 9% had at least four comorbidities. An initial logistic regression predictive model has an AUROC of 0.74 with OSA having an odds ratio of 2.13 for developing a new comorbidity after T0. Conclusion These findings suggest many Veterans have already developed significant cardiometabolic comorbidities by the time of OSA diagnoses, with additional diagnoses occurring subsequently. This may reflect increased healthcare engagement leading to the diagnosis of comorbidities or potentially indicate a causal relationship. Additional research is needed to better define the relationship between OSA and incident comorbidities. Support (if any) Million Veteran Program MVP063.