Background and aimsApolipoprotein C-III (apoC-III) proteoform composition show distinct relationships with plasma lipids and cardiovascular risk. The present study tested whether apoC-III proteoforms are associated with risk of peripheral artery disease (PAD). MethodsApoC-III proteoforms, i.e., native (C-III0a), and glycosylated with zero (C-III0b), one (C-III1) or two (C-III2) sialic acids, were measured by mass spectrometry immunoassay on 5,734 Multi-Ethnic Study of Atherosclerosis participants who were subsequently followed for clinical PAD over 17 years. Ankle-brachial index (ABI) was also assessed at baseline and then 3 and 10 years later in 4,830 participants. ResultsHigher baseline C-III0b/C-III1 and lower baseline C-III2/C-III1 were associated with slower decline in ABI (follow-up adjusted for baseline) over time, independently of cardiometabolic risk factors, and plasma triglycerides and HDL cholesterol levels (estimated differences per 1 SD were 0.31% for both, p <0.01). The associations between C-III2/C-III1 and changes in ABI were stronger in men (-1.21% vs. -0.27% in women), and in Black and Chinese participants (-0.83% and -0.86% vs. 0.12% in White). Higher C-III0b/C-III1 was associated with a trend for lower risk of PAD (HR=0.84 [95%CI: 0.67-1.04]) that became stronger after excluding participants on lipid-lowering medications (0.73 [95%CI: 0.57-0.94]). Neither change in ABI nor clinical PAD were related to total apoC-III levels. ConclusionsWe found associations of apoC-III proteoform composition with changes in ABI that were independent of other risk factors, including plasma lipids. Our data further support unique properties of apoC-III proteoforms in modulating vascular health that go beyond total apoC-III levels.