Figure 1. Multiple linear regression curves on the association between energy contribution of ultra-processed food and blood pressure.

Shaded areas represent 95% confidence intervals. Fully adjusted for age, body mass index, income (lower median/upper median), education level (≤6 years, 7–12 years, >12 years), occupation (white, pink, blue, and others), region (rural, urban), smoking status (never, ex-, current), exercise (yes, no), total energy intake, history of disease (at least one of dyslipidemia, stroke, myocardial infarction, angina pectoris, diabetes, and kidney failure), and Korean Health Eating Index.