The effect of obstructive sleep apnea on fatty liver disease may be obscured by alcohol consumption: An ordinal logistic regression analysis

Obstructive sleep apnea (OSA) is closely associated with non-alcoholic fatty liver disease (NAFLD). The current definition of NAFLD cannot exclude the involvement of alcohol consumption in the development of fatty liver disease (FLD), but alcohol can aggravate OSA and participate in steatosis. There is limited evidence on the relationship between OSA and alcohol and its effect on FLD severity.

To determine the effect of OSA on FLD severity based on ordinal responses, and its relationship with alcohol consumption, in order to develop strategies for the prevention and treatment of FLD.

Patients with chief complaints of "snoring" who underwent polysomnography and abdominal ultrasound between January 2015 and October 2022 were selected. A total of 325 cases were divided into three groups according to abdominal ultrasound results: no FLD (n = 66), mild FLD (n = 116), and moderately severe FLD (n = 143) group. Patients were also categorized into alcoholic and nonalcoholic groups. Univariate analysis was used to examine the correlation between OSA and FLD severity. Multivariate ordinal logistic regression analysis was further used to identify the determinants of FLD severity and differences between the alcoholic and nonalcoholic groups.

A higher proportion of moderately severe FLD was observed in the group with an apnea/hypopnea index (AHI) > 30 compared to the AHI<15 group in all participants and in the nonalcoholic population (all p < 0.05). There was no significant difference among these groups in the alcoholic population. Ordinal logistic regression analysis found that in all participants, age [OR = 0.966(0.947-0.986)], BMI [OR = 1.293 (1.205-1.394)], diabetes mellitus [OR = 1.932(1.132-3.343)], hyperlipidemia [OR = 2.432(1.355-4.464)], severe OSA [OR = 2.36(1.315-4.259)] (all p < 0.05) were the independent risk factors for more severe FLD. However, different risk factors applied according to alcohol consumption. In addition to age and BMI, the independent risk factors for the alcoholic group also included diabetes mellitus [OR = 3.323(1.494-7.834)] while in the non-alcoholic group risk factors included hyperlipidemia [OR = 4.094(1.639-11.137)], and severe OSA[OR = 2.956(1.334-6.664)] (all p < 0.05).

Severe OSA is an independent determinant for developing more severe NAFLD in nonalcoholic population, and alcohol consumption may obscure the effect of OSA on the progression of FLD.