Background: Anastomotic neointimal hyperplasia plays a significant role in the late failure of infrainguinal prosthetic arterial bypass grafts. Previous work from our laboratory revealed that placing a stent across an end-to-end arterio-arterial anastomosis resulted in an increase in the luminal area as well as in the intimal thickness (IT) at the anastomotic level. This study was designed to evaluate the effects on neointimal hyperplasia when a stent is placed across an end-to-side polytetrafluoroethylene (PTFE) graft arterial anastomosis.
Methods: A canine model of an end-to-side anastomosis was developed using a 12 x 6 mm polytetrafluoroethylene aortobi-iliac graft. A self-expanding stainless steel Wallstent was placed across one randomly selected distal anastomosis leaving the opposite side as a control. Dogs were sacrificed at 4 and 12 weeks. At sacrifice, the graft and intact anastomoses were pressure-perfusion fixed with glutaraldehyde. Sections of each distal graft, anastomosis, and recipient artery were obtained for analysis. Computer images of each section were digitized to determine the luminal area and the mean IT. The data were analyzed statistically using univariate repeated measures of analysis of variance.
Results: One animal died prior to early sacrifice. Eight of 10 graft limbs remained patent at sacrifice. Of the 2 limbs that occluded, one was stented and one was nonstented. At 4 weeks, stented graft limbs had significantly greater IT at the proximal stent level (mean difference between control and stented sides 0.163 mm +/- 0.054, P = 0.01). Stented and nonstented anastomoses had similar luminal area and IT at other levels where sections were taken. At 12 weeks, control limbs had significantly greater IT at the anastomotic level compared to the 4-week measurements (mean difference 12 weeks versus 4 weeks 0.185 mm +/- 0.06, P = 0.006). In the stented limbs, IT at the anastomotic level had stabilized and was not significantly thicker than at 4 weeks. The control limbs had greater IT at the anastomotic level than the stented limbs (mean difference between controls and stented sides at 12 weeks 0.091 mm +/- 0.044, P = 0.06). At the proximal end of the stent, IT progressed significantly between the 4th and 12th weeks (mean difference 12 weeks versus 4 weeks 0.155 mm +/- 0.06, P = 0.02). The IT at the proximal end of the stent at 12 weeks was significantly greater than the IT at a comparable level in the controls (mean difference stent versus control 0.132 mm +/- 0.05, P = 0.04). The luminal area in the control limbs was significantly greater than in the stented anastomoses at levels corresponding to either end of the stent (mean difference at proximal end 4.163 mm2 +/- 1.633, P = 0.01; mean difference at distal end 7.192 mm2 +/- 1.633, P = 0.0005). However, there was no difference in luminal area at the anastomotic level.
Conclusion: We conclude that the presence of an intraluminal stent alters the siting and degree of anastomotic neointimal hyperplasia in a canine model of an end-to-side anastomosis resulting in translocation of the intimal hyperplastic response to the proximal graft stent interface in a magnitude similar to that which would normally be found at the anastomosis.