0473 Homocysteine as a Predictor of Apnea-hypopnea Index in Obstructive Sleep Apnea: Longitudinal Epidemiological Study

Introduction Obstructive sleep apnea (OSA) affects nearly 1 billion people globally, and has established links with cardiovascular and neurocognitive complications. Although it has some limitations, the apnea-hypopnea index (AHI) is commonly used to gauge OSA severity and therapeutic response. Our objective was to investigate homocysteine (Hcy) levels as a predictor of AHI values and as a risk factor for OSA. Methods This is a prospective longitudinal cohort study with 8 years follow-up based on a subsample of the São Paulo Epidemiological Sleep Study (EPISONO). Data from polysomnography and blood analysis were used, data from the AHI and plasma Hcy, cobalamin and folic acid levels were extracted from dataset. The generalized estimation equation was used to estimate the predictive value of Hcy levels, and logistic regression was performed to estimate the risk. Results Our findings showed that Hcy was a predictor for an increased AHI, and AHI increased over time. Individuals with plasma Hcy levels ≥15 µmol/L experienced an average AHI increase of 7.43 events/hour ([beta coefficient] β=7.43; 95%CI: 2.73 to 12.13) over time, compared to those with plasma levels < 10 µmol/L. A similar trend was apparent in those with plasma Hcy levels between 10≥ to < 15 µmol/L, who had an AHI increase with an average beta coefficient of 3.20 events/hour (95%CI:1.01 to 5.39) compared to those with plasma Hcy levels < 10µmol/L. Our results revealed that increased plasma Hcy levels could be considered a risk factor for the development of OSA. Those with higher plasma Hcy levels have a greater risk of developing OSA (odds ratio=1.098 95%CI: 1.022 to 1.181). Conclusion In conclusion, the results suggest that plasma Hcy levels can predict the severity of OSA, highlighting their association with AHI. Support (if any) This study was supported by the Associação Fundo de Incentivo à Pesquisa (AFIP), São Paulo, Brazil. MLA and VDA are recipients of CNPq fellowships. MLA is supported by the São Paulo Research Foundation (FAPESP #2020/13467-8).