Abstract:Abstract: Objective To explore the relationship between serum homocysteine (Hcy) and brain-derived neurotrophic factor (BDNF) levels in severe obstructive sleep apnea-hypopnea syndrome (OSAHS) and cognitive function. Method Sixty patients with severe OSAHS were randomly selected to evaluate their cognitive function with Montreal Cognitive Assessment Scale (MoCA). According to the evaluation results, they were divided into cognitive impairment group (18 cases) and cognitive normal group (42 cases), and 30 healthy subjects were used as the control group. The levels of Hcy and BDNF in serum were measured by ELISA. Results There were significant differences in MOCA total score, visual space and executive ability, attention, abstraction and delayed recall between cognitive impairment group and cognitive normal group and control group (P < 0.01), but there were no significant differences in naming, language ability and orientation score (P > 0.05). The level of serum Hcy in cognitive impairment group (34.12 ± 2.85) were higher than that in cognitive normal group (30.88 ± 2.10), while the level of serum BDNF (9.00 ± 1.67) were lower than that in cognitive normal group (11.64 ± 1.73). The lowest arterial oxygen saturation (LSaO2%) and apnea ventilation index (AHI) in cognitive impairment group were different from those in cognitive normal group (P < 0.01). In cognitive impairment group, serum Hcy was negatively correlated with the total score of MOCA (r = -0.880, P = 0.000), LSaO2% (r = -0.595, P = 0.009) and BDNF (r = -0.818, P = 0.000), and positively correlated with AHI (r = 0.681, P = 0.002); Serum BDNF were positively correlated with the total score of MOCA (r = 0.751, P = 0.000) and LSaO2% (r = 0.521, P = 0.026), and negatively correlated with AHI (r = -0.553, P = 0.017). The areas under the ROC curve predicted by serum Hcy and BDNF levels were 0.902 (95% CI: 0.824, 0.979) and (95% CI: 0.868, 0.985), respectively. Conclusion The changes of serum Hcy and BDNF levels in patients with OSAHS are related to the formation of cognitive impairment, and can be used as biomarkers for the early diagnosis and prediction of cognitive impairment.