OBJECTIVE: External mesh support of vein grafts has been shown to mitigate the formation of intimal hyperplasia. The aim of the present study was to address the issue of optimal mesh size in a nonhuman primate model that mimics the dimensional mismatch typically encountered between clinical vein grafts and their target arteries. METHODS: The effect of mesh size on intimal hyperplasia and endothelial preservation was assessed in bilateral femoral interposition grafts in Chacma baboons (n(Sigma) = 32/n = 8 per mesh size). No mesh support (group I) was compared with external nitinol meshes at three different sizes: loose fitting (group II), 25% diameter constricting (group III), and 50% diameter constricting (group IV). Mesh sizes were seen not only in isolation but also against the background of anastomotic size mismatch at implantation, expressed as quotient of cross-sectional area of host artery to vein graft (Q(C)). RESULTS: Significant amounts of intimal hyperplasia were found in group I (Q(C) median 0.20; intimal hyperplasia 6 weeks = 1.63 +/- 0.34 mm(2); intimal hyperplasia 12 weeks = 1.73 +/- 0.5 mm(2)) and group II (Q(C) median 0.25; intimal hyperplasia 6 weeks = 1.96 +/- 1.64 mm(2); intimal hyperplasia 12 weeks = 2.88 +/- 1.69 mm(2)). In contrast, group III (Q(C) median 0.45; intimal hyperplasia 6 weeks = 0.08 +/- 0.13 mm(2); intimal hyperplasia 12 weeks = 0.18 +/- 0.32 mm(2)) and IV (Q(C) median 1.16; intimal hyperplasia 6 weeks = 0.02 +/- 0.03 mm(2); intimal hyperplasia 12 weeks = 0.11 +/- 0.10 mm(2)) showed dramatically suppressed intimal hyperplasia (P < .01) at both time points. Endothelial integrity was only preserved in group IV (P < .05). There were no significant differences in vascularization and inflammation in either interlayer or intergroup comparisons. CONCLUSION: By using an animal model that addressed the clinical phenomenon of diameter discrepancy between vein graft and bypassed artery, we could demonstrate that suppression of intimal hyperplasia required constrictive mesh sizes.