A prospective study of the determinants of vein graft flow velocity: Implications for graft surveillance

Abstract Purpose: Serial monitoring of vein graft peak systolic flow velocity (PSFV) has been endorsed as a technique for vein graft surveillance with low values ( Methods: Factors recorded for each patient included vein graft diameter (VGD), measured outflow resistance (MOR), conduit length, outflow level (popliteal/tibial/pedal), inflow level (femoral/popliteal), systolic blood pressure, cardiac ejection fraction, the presence of a patent plantar arch, and Society for Vascular Surgery/International Society for Cardiovascular Surgery resistance scoring. MOR was measured by occluding graft inflow and infusing saline solution through a proximal graft cannula at 60 cc/min while simultaneously recording the pressure at the distal anastomosis via a separate cannula. MOR was calculated by dividing the resultant pressure by the infusion rate. MORs were expressed in resistance units and were measured before and after the infusion of papaverine (MOR (PAP) ). PSFVs and VGDs were measured 4 to 6 cm from the distal anastomosis 3 weeks after surgery with duplex scanning (60 degree angle with midstream sample volume). Results: PSFVs ranged from 22 to 148 cm/sec and averaged 83.4 ± 4.8 cm/sec. Pedal bypass grafts had significantly lower PSFVs (64 ± 10 vs 89.5 ± 5 cm/sec, p = 0.02) and significantly higher MOR (PAP) s (0.86 ± 0.15 vs 0.51 ± 0.05 resistance units, p = 0.05) than bypasses to the popliteal/tibial level. When subjected to univariate analysis the factors correlating with PSFV were MOR ( r = -0.59, p = 0.0001), MOR (PAP) ( r = -0.69, p = 0.0001) VGD ( r = -0.31, p = 0.06), the Society for Vascular Surgery/International Society for Cardiovascular Surgery score ( r = -0.35, p = 0.04), inflow level ( r = -0.47, p = 0.002), and outflow level ( r = -0.35, p = 0.03). When subjected to multiple regression analysis, only MOR (PAP) ( r 2 = 0.51, p = 0.001) and VGD ( r 2 = 0.14, p =0.001) contributed significantly to the overall model ( r 2 = 0.65, p = 0.0001) withMOR (PAP) ) eliminating the effect of the other variables. The multiple regression model predicts PSFV as follows: PSFV=176 + VGD (mm) ( -11.7) + MOR (PAP) )( -63.4). Conclusions: Clinically successful and hemodynamically normal vein grafts have widely variable, yet predictable flow characteristics that are influenced primarily by outflow resistance and VGD. This wide variability suggests that no single lower threshold value for PSFV is universally applicable in identifying all grafts at risk for failure. Detection of focal areas of flow acceleration within the graft may be more accurate in identifying grafts at risk for failure. (J VASC SURG 1994;19:259-67.)