Association of oxidative balance score with cardiovascular disease and all-cause and cardiovascular mortality in American adults with type 2 diabetes: data from the National Health and Nutrition examination survey 1999-2018.

Background: Oxidative stress has an important role in type 2 diabetes (T2D). Oxidative balance score (OBS) is an emerging assessment of dietary and lifestyle oxidative balance. We aimed to explore the association of OBS with cardiovascular disease (CVD) and all-cause and CVD mortality in the T2D population through NHANES 1999-2018. Methods: OBS integrated 16 dietary components and 4 lifestyle components. T2D was diagnosed according to the American Diabetes Association criteria. Multivariate logistic regression and multivariate Cox proportional hazards regression analyses were used to explore the association of OBS with CVD and mortality in T2D, respectively. Results: 3801 adult T2D participants were included. In fully adjusted models, OBS, dietary OBS, and lifestyle OBS were all negatively associated with the prevalence of CVD (odds ratios of 0.98, 0.98, and 0.85, respectively). Higher OBS and lifestyle OBS (p for trend 0.016 and <0.001, respectively) rather than dietary OBS (p for trend = 0.06) were associated with significantly lower odds of CVD. Higher OBS, dietary OBS, and lifestyle OBS were all negatively associated with all-cause mortality (hazard ratios [HR] of 0.98, 0.98, and 0.92, respectively; p for trend of 0.002, 0.009, and 0.035, respectively). Higher OBS and dietary OBS were negatively associated with CVD mortality (HR 0.96 and 0.95, respectively; p for trend both <0.001), whereas lifestyle OBS was not. Restricted cubic spline analysis suggested that most associations were linear. Stratified analyses showed that these associations were influenced by some demographic variables and disease status. Conclusions: Adherence to higher OBS was associated with reduced CVD prevalence and mortality risk in T2D. Antioxidant diet and lifestyle had more significant associations with mortality and CVD prevalence, respectively. However, as these findings are merely associations and do not allow causal inferences to be drawn, future validation in high-quality randomized controlled trials is needed.