Blood flow after peripheral arterial reconstruction. I. Measurements after iliac-femoro-popliteal arterial reconstructions with the electromagnetic flowmeter and implanted flow probes during operation and in the early postoperative periods

Input Impedance of Revascularized Skeletal Muscle, Renal, and Mesenteric Vascular Beds

Input impedance describes the relationship between pressure and flow in a vascular system and, hence, characterizes the outflow bed. The purpose of this investigation was to measure input impedance spectra in vascular reconstructions of skeletal muscle, renal, and mesenteric beds.

Input impedance was measured in 107 vascular reconstructions in 96 patients. Reconstructions were performed at the aortofemoral/aortoiliac (AF, n = 20), femoropopliteal (FP, n=18), femorodistal (FD, n=41), infrapopliteal-inframalleolar (IM, n = 6), renal (REN, n = 16), or mesenteric (MES, n= 6) level. Grafts were constructed from autologous vein in all cases except AF bypasses in which bifurcated woven Dacron grafts were employed. Input impedance was measured intraoperatively after reperfusion. For impedance calculation, simultaneously acquired intraluminal pressure (transducer-tipped pressure catheter) and blood flow (electromagnetic probe) waveforms of ten-second duration were digitized at 200 Hz and subjected to Fourier transformation in near real-time.

AF grafts exhibited the highest blood flow (443 ± 72.8 mL/minute) followed by MES (300 ± 30.4), REN (172 ± 43.9), FP (91.6 ± 20.0), FD (59.3 ± 5.09), and IM grafts (22.4 ± 5.44 mL/minute). A similar (inverse) trend was observed with respect to resistance (Rin), ie, MES≈AF<REN<<FP<FD<<IM(15.7 ± 2.2=23.7 ± 6.5<69.2 ± 12.2 < < 120 ± 31.9 < 137 ± 13.0 < < 337 ± 75.8 x 103 dyne•s•cm-5). As expected, AF grafts exhibited the lowest characteristic impedance (Z0 3.5 ± 0.8 x 103 dyne•s•cm-5). However, the Z0 of REN (20.2 ± 3.7 x 103 dyne•s•cm-5) grafts exceeded that of FP and MES grafts (10.5 ± 1.1 and 12.4 ± 4.0 x103 dyne•s•cm-5) and more closely approximated that of FD (22.2 ± 2.2 x 103 dyne•s•cm-5). The highest Z0 was seen in IM grafts (42.1 ± 15.8 x 103 dyne•s•cm-5). Therefore, the highest ratios of Z0/Rin were observed with MES (0.75 ± 0.23) and REN (0.33 ± 0.04) grafts as compared with the other groups (AF 0.23 ± 0.03, FP 0.12 ± 0.13, FD 0.19 ± 0.02, IM 0.16 ± 0.06).

As expected, Rin of vascular reconstructions follows the general trend of MES ≈ AF < REN < < FP < FD < < IM. However, examination of the high-frequency components of the impedance spectra reveals that Z0 follows a different pattern, AF < FP MES < FD ≈ REN < < IM and the ratio of ZO/Rin is highest in REN and MES vascular beds compared with skeletal muscle beds. Thus, although the REN and MES beds are “privileged” (ie, maximally dilatated with low arteriolar tone), the intrinsic properties of the graft and larger blood vessels are no different than those of a femorodistal bypass.

Pharmacokinetics and pharmacodynamics of intra-graft iloprost in femorodistal bypass surgery

Intra-graft injection of the prostacyclin analogue, iloprost, was performed at the end of femorodistal bypass procedures in 12 patients with severe peripheral arterial occlusive disease. Iloprost plasma levels were measured and compared with changes in haemodynamics. There was a high initial iloprost plasma level (mean 625 pg/ml) which dropped to a mean of 50 pg/ml after 15 min. This correlated with an immediate reduction in systolic blood pressure which had returned to pretreatment levels after 15 min. In contrast, the vascular resistance distal to the graft showed a reduction after 5 min which was maintained for at least 20 min after iloprost injection and the mean blood flow through the graft increased steadily throughout the same period of measurement. The study showed an effect of iloprost on blood pressure which correlated with plasma levels, but the time course of the changes in distal vascular resistance and graft blood flow demonstrated an effect more prolonged than the half-life of iloprost.

Measurement of vascular input impedance in infrainguinal vein grafts

The purpose of this investigation was to measure vascular input impedance in infrainguinal vein grafts and assess the importance of clinical and hemodynamic parameters in predicting graft patency. Fifty-seven patients undergoing infrainguinal vein bypass grafting for limb salvage (n = 40) or claudication (n = 17) were prospectively studied. At the time of revascularization, simultaneously acquired intraluminal pressure and blood flow waveforms were digitized at 200 Hz and subjected to Fourier transformation in near real-time. Input impedance was calculated at baseline (immediately after unclamping) and after stimulation with either papaverine or completion arteriography. Resistance (Rin) was calculated as mean pressure divided by mean blood flow (Q). Characteristic impedance (Z0) was calculated as the mean of harmonics 3-10. Intraoperative data acquisition required approximately 5 min, utilized the completion angiography cannula already in place, and was uncomplicated in all patients. Stimulation with either papaverine or arteriography resulted in increased Q (72 +/- 7 to 146 +/- 11 ml/min, p < 0.001), decreased Rin (126 +/- 13 to 52 +/- 4 x 10(3) dyne.s.cm-5, p < 0.001), and slightly decreased Z0 (18 +/- 2 to 15 +/- 1, p = 0.002). After a mean follow-up of 20 months, the 2-year primary patency, secondary patency, limb salvage, and survival rates were 61 +/- 8%, 74 +/- 7%, 76 +/- 6%, and 86 +/- 6%, respectively. Primary patency was not associated with any of the clinical variables studied including age, sex, smoking history, history of previous vascular surgery, hypertension, coronary artery disease, diabetes mellitus, creatinine, indication for revascularization (claudication versus limb salvage), anesthesia (general versus regional), or level of distal anastomosis (popliteal versus infrapopliteal). Furthermore, there was no association between primary patency and baseline Q, baseline Rin, or stimulated Z0. However, using univariate analysis, patency was positively associated with decreased stimulated Rin (p = 0.002), elevated stimulated Q (p = 0.006), and decreased baseline Z0 (p = 0.02). Multiple regression analysis identified stimulated Rin as the only independent predictor of primary patency (p = 0.002). Stimulated Rin > or = 50 x 10(3) dyne.s.cm-5 was 71% sensitive and 65% specific for graft failure. It is concluded that 1) vascular input impedance can be simply and reliably measured in the operating room, and 2) elevated stimulated Rin is an independent predictor of primary patency.

Iloprost improves femoro-distal graft flow after a single bolus injection

A double-blind, randomised, placebo-controlled trial was conducted to study the effect of the stable prostacyclin analogue iloprost on femoro-distal graft blood flow. After completing femoro-distal reconstruction, 3000 ng of iloprost or placebo was injected into the graft over 2 min. Graft blood flow, measured by electromagnetic flowmetry, increased by a mean (range) of 94% (12 to 192%) in patients receiving iloprost (n = 15) compared to 6% (-34 to 53%) in controls (n = 16; p less than 0.0001, t-test). Increased graft flow, measured by duplex ultrasound, was maintained in the iloprost group over a 7 day period postoperatively (F = 5.2, p = 0.03; analysis of variance) and remained higher at 7 days (p = 0.007, t-test). Iloprost produces an immediate, sustained increase in graft blood flow after femoro-distal reconstruction and may therefore be of benefit in reducing the incidence of early graft failure.

In vitro adherence and kinetics studies of adult human endothelial cell seeded polytetrafluoroethylene and gelatin impregnated Dacron grafts

Lining the luminal surface of small diameter vascular prostheses with living endothelial cells reduces thrombogenicity, decreases infection and improves patency. In vitro adherence and kinetics studies of adult human endothelial cell seeded Polytetrafluoroethylene (ePTFE) and Gelatin impregnated Dacron (Gelseal) were performed. Endothelial cell adherence on ePTFE and Gelseal coated with collagen IV, fibronectin and preclot matrices was compared. Untreated ePTFE and Gelseal were also used. Ten graft segments in each group coated with these matrices were incubated with radio-labelled adult human endothelial cells for 30, 60 and 90 min. Labelled endothelial cells seeded in supra-confluent densities on fibronectin coated ePTFE and Gelseal grafts were used for kinetic studies. Resultant endothelial cell monolayers were then exposed to varying shear stress at flow rates of 200 and 300 ml/min in an artificial flow circuit. Endothelial cell attachment to untreated grafts was poor and a suitable matrix significantly improved adherence with fibronectin and preclot but less so with collagen. A 30 min incubation was sufficient for optimum cell attachment. Cell adherence to ePTFE was significantly better than Gelseal. Scanning electron micrographs (SEM) of ePTFE showed preferential attachment to the nodes whilst on Gelseal, cells conformed to Dacron fibres at different levels and directions. Rapidly formed endothelial cell monolayers on ePTFE and Gelseal grafts resisted shear stress of flow with significant cell retention at 2 h. There was patchy coverage of both grafts with evidence of bridging of gaps between individual fibres in Gelseal.

Effects of shear stress on endothelial cell monolayers on expanded polytetrafluoroethylene (ePTFE) grafts using preclot and fibronectin matrices

Animal studies have shown that endothelial seeding of vascular prosthetic grafts reduces thrombogenicity and improves their patency. However, for endothelial seeding to be of clinical benefit in humans, it must withstand shear stress of blood flow. Endothelial cells labelled with Indium-111-oxine were seeded in supra-confluent densities on preclot or fibronectin coated ePTFE graft segments over a period of 90 min. These grafts with rapidly formed endothelial cell monolayers were then exposed to varying shear stresses up to a flow rate of 300 ml/min, using tissue culture medium in an artificial flow circuit. Grafts coated with preclot matrix showed 2 h cell retentions of 82.4 +/- 6.8% at 25 ml/min, 79.9 +/- 8.2% at 100 ml/min, 75.4 +/- 9.5% at 200 ml/min and 58.3 +/- 15.5% at 300 ml/min whilst those for the fibronectin matrix were 57.8 +/- 9.9%, 55.2 +/- 13.3%, 55.4 +/- 12.9% and 56.5 +/- 15.2% respectively. Overall the preclot matrix was found to be better than fibronectin (P less than 0.001). Light and scanning electron microscopy revealed well-formed endothelial cell monolayers retained on preclot matrix up to a flow rate of 200 ml/min whereas uncovered patches were seen at 300 ml/min and at all flow rates on fibronectin matrix.

Operative assessment of femorodistal bypass grafts using a new Doppler flowmeter

Operative measurements using a newly developed Doppler flowmeter were made on 47 in situ femorodistal bypass grafts performed for critical ischaemia. Graft blood flow, peripheral resistance and impedance were measured before and after the injection of 15 mg papaverine down the graft. The proportion of retrograde blood flow at the distal anastomosis was also measured. The success of a graft at 1 month was defined by patency on duplex ultrasound scanning, a rise in the Doppler ankle pressures and an improvement in the clinical state of the limb. By these criteria the 1-month primary success rate was 35 out of 47 (74 per cent). Graft blood flow was of little value in predicting graft outcome but, combined with segmental pressures, accurately located six of eight technical problems which were successfully corrected. After papaverine, the peripheral resistance of all the successful grafts fell below one peripheral resistance unit (PRU). No single measurement was able to completely distinguish between success and failure, although discriminant analysis of all measurements was able to predict success with a sensitivity of 100 per cent and specificity of 97 per cent. A simpler combination of a peripheral resistance after papaverine less than 1 PRU and a retrograde blood flow less than 33 per cent predicted success, with a sensitivity of 97 per cent and specificity 83 per cent. These criteria have been validated on a further 26 grafts, achieving a sensitivity of 88 per cent and a specificity of 75 per cent. Simple operative measurements, using a newly developed Doppler flowmeter, enable the detection of technical problems and the accurate prediction of the early outcome of femorodistal grafting.

The predictive value of blood flow velocity in femoro-distal bypass

The predictive value of peroperative blood flow measurements has been evaluated in a prospective study of femoro-distal bypass grafts. Seventy grafts placed in patients with critical ischaemia were assessed with a minimum follow-up of 2 years. Thirty-one grafts were performed using autogenous saphenous vein (ASV), 27 using a polytetrafluoroethylene prosthesis (PTFE) and in 12 human umbilical vein (HUV) was used. Six grafts failed early (within 1 month of operation) and 19 failed late, giving an overall patency rate of 60% at 5 years. ASV grafts had a significantly better patency rate than either prosthetic material. Analysis of flow rates measured before and after regional vasodilatation revealed no significant differences between types of graft or between grafts that subsequently failed and those remaining patent. Flow velocity prior to vasodilatation was significantly higher in patent grafts (mean 13.6) compared to those failing early (mean 6.3), but not compared to those failing late (mean 9.4). Flow velocities were also significantly higher in ASV grafts than prosthetic grafts. Flow velocity measured after vasodilatation was significantly higher in grafts remaining patent (mean 28.4) than those which failed late (mean 19.2). It is concluded that flow velocity is of value in the prediction of graft outcome. Graft velocity after vasodilatation of less than 20 cm/s. Predicted failure in 19 of 25 grafts that occluded. We are currently evaluating the selective use of anticoagulants in femoropopliteal grafts with a maximal flow velocity below 20 cm/s.

Negative effects of dihydroergotamine (Orstanorm) on central haemodynamics and aorto-coronary bypass blood flow. An intraoperative study

Dihydroergotamine (DHE), Orstanorm, because of its strong constrictor action on capacitance vessels, is used in the treatment of hypotension caused by orthostatism or spinal or epidural anaesthesia. Lately Orstanorm has also been used in combination with heparin as prophylaxis against postoperative thromboembolism. In the present study, Orstanorm (0.01 mg/kg b.w.) was given intravenously to 20 patients after coronary bypass surgery. Coronary bypass blood flow, heart rate, systemic mean and systolic blood pressures, right and left atrial pressures, pulmonary artery pressure and cardiac output were measured. Stroke volume, cardiac index, stroke index, systemic and regional myocardial vascular resistance and cardiac work index were then calculated. The results showed that despite increased filling pressures there was no rise in cardiac output, and despite increased cardiac work the bypass flow significantly decreased. The significant increase in regional myocardial vascular resistance found after administration of DHE may explain the absence of expected increase of cardiac output and coronary bypass flow.

The effect of O-(beta-hydroxyethyl)-rutosides on central haemodynamics during and after aortocoronary bypass surgery

Clinical use of O-(beta-hydroxyethyl)-rutosides (HR) has been mainly in chronic venous insufficiency, but to some extent also in arterial insufficiency. We investigated the effect of HR on the central circulation in a group of patients undergoing aortocoronary bypass surgery. Measurements were made during operation and on the first postoperative day, both before and after intravenous injection of 15 ml 10% HR (1.5 g). None of the obtained or calculated values for cardiac output, central blood pressures, coronary blood flow, heart rate, stroke volume and systemic, pulmonary and regional myocardial vascular resistance showed significant change following HR injection intraoperatively or postoperatively. We conclude that this intravenous dosage of HR has no adverse effect on central haemodynamics in patients with advanced coronary disease. This suggests that the drug can also be safely used in patients with peripheral arterial disease, as its beneficial effect on nutritional skin blood flow in these patients seems to be due to action on the microrheology of the blood and not to peripheral vasodilatation.

Axillofemoral and femorofemoral grafts: a 6-year experience with emphasis on the relationship of peroperative flow measurement to graft survival

Over a 6-year period 64 axillofemoral bypass and femorofemoral crossover grafts have been performed in 58 patients, most of whom were considered unfit for intra-abdominal surgery. Indications were peripheral ischaemia in 78 per cent and disabling claudication in 22 per cent. The limb salvage rate at 3 years was 75 per cent. No claudicants lost limbs, but only one-third of patients presenting with forefoot gangrene or ulceration avoided amputation. Most patients presenting with ischaemic symptoms at rest had associated femoropopliteal and distal disease, confirmed by the ankle pressure index measurements, and this influenced graft patency. Although the cumulative patency at 3 years for all grafts combined was 57 per cent with similar patencies for both the axillofemoral and femorofemoral grafts, early occlusion was more common in axillofemoral grafts and this may be reduced in bifemoral grafts by the increased flow rate in the vertical limbs. Peroperative electromagnetic flowmeter measurements were made after reconstruction on 55 femoral arteries in 46 of the patients and graft flow velocities were derived from these measurements. Comparison between velocities from those grafts remaining patent and those subsequently occluding showed a high incidence of occlusion in grafts with a maximal velocity after distal vasodilatation of less than 8 cm/s. Graft occlusion after the first postoperative month was more commonly associated with other factors such as continued smoking, severity of distal disease and perigraft infection.

Surgical management of obliterative disease of the brachiocephalic trunk. Experience from 24 cases

Transsternal endarterectomy was performed in 24 patients for symptomatic obliterative disease of the brachiocephalic trunk in the period 1961-1981. In 10 of the patients the vessel was occluded, and in 14 there were various degrees of stenosis. Reversal of vertebral artery flow was found in 15 patients. Subclavian steal-carotid recovery phenomenon was present in six patients, and one patient had reversed flow in both the right vertebral and the common carotid artery. Only 10 of the 24 patients were free from obliterative changes in other extracranial arteries. After the reconstruction, the blood flow in the brachiocephalic trunk increased from 370 to 610 ml/min in patients with stenosis, and from 0 to 692 ml/min in patients with occlusion. The flow rate changed in the right vertebral artery from -58 to + 119 ml/min and in the right common carotid artery from 116 to 340 ml/min. Two patients died in the early postoperative period (8.5%) and six died during the follow-up period. Follow-up was possible in 21 patients. The result of endarterectomy was assessed as cure in 14, improvement in two and failure in three patients. Stroke occurred in two patients.

Peroperative assessment in arterial surgery--flow determination or angiography?

Complete data from flow measurement and intraoperative angiography were obtained in 70 femoro-popliteal vein by pass procedures. The results were compared in respect to their predictability of early outcome after arterial reconstruction. A correct prediction was given in 65 of 70 cases (92 per cent) by angiography. The corresponding figure for flowmeter determination was 52 of 70 (74 per cent). Angiography was found to be preferable to flow determination especially when low flow values were registered.

Blood flow in small arteries after end-to-end and end-in-end anastomoses: an experimental quantitative comparison

Microvascular anastomoses were performed under standardized conditions on the central artery of the ear in 23 rabbits. Fifteen end-to-end and 13 end-in-end anastomoses were compared. No anticoagulants or vasodilating agents, other than the local application of lidocaine, were used. Blood flow was measured by electromagnetic flowmetry before anastomosis and after anastomosis at intervals of 30 minutes for 4 hours and at 1 and 3 days. The validity of the electromagnetic flowmetry method was established in a control group of 7 rabbits. The mean volume of blood flow after end-to-end anastomosis was 93% of the preoperative value (P greater than 0.05). The corresponding mean volume of blood flow after end-to-end anastomosis was 47% of the preoperative value (P less than 0.001). The difference in flow between end-to-end and end-in-end anastomosis was highly significant (P less than 0.001). Microangiography demonstrated marked stenosis in the end-in-end anastomoses but not in the end-to-end anastomoses.

Intra-operative graft blood flow related to failure of femoro-popliteal bypass grafts

The prognostic significance of intra-operative blood flow, as measured by electromagnetic flowmetry, was investigated in 127 limbs, which were operated on with a reversed femoropopliteal saphenous vein bypass graft because of symptomatic atherosclerotic occlusion of the superficial femoral artery. Thromboses occurring in the first postoperative month are defined as early failures and thereafter as late failures. There were three early graft failures. The basal and augmented flow rates during pharmacological vasodilation of these grafts were less than half those of the grafts remaining patent. Eight additional late graft occlusions occurred. Whereas the basal blood flow of these grafts did not differ significantly from that of the grafts which remained patent, the maximal flow rate was significantly lower (p less than 0.05). The overall incidence of graft failure at basal and maximal flow rates of 100 ml/min or less and 150 ml/min or less, respectively, was 35%, while, at higher flow rates, this incidence was reduced to 5% (p less than 0.001). This study emphasizes that intra-operative femoropopliteal vein graft flow provides prognostic indications of graft failure.

A continuous wave Doppler velocimeter for monitoring blood flow in the popliteal artery, compared with venous occlusion plethysmography of the calf

A transcutaneous Doppler velocimeter has been used for monitoring changes in blood flow in the popliteal artery during and after exercise of the calf muscles on a calf-ergometer. The instrument and the positioning of the probe are described. The validity of the Doppler measurements has been assessed by comparing results after exercise and after 5 min arterial occlusion with venous occlusion plethysmography. For 10 healthy volunteers calibration lines were found which are curved. This can be ascribed partly to alinearity of the Doppler instrument and partly to changes of the diameter of the popliteal artery. Displacement of the probe with respect to the artery, which influences the measured velocity signal, can be detected and to a certain extent corrected by taking into account the intensity of the Doppler signal. The largest deviation of a data point from the corresponding calibration line varies for the ten subjects between 90 and 170 Hz in the high flow range (mean Doppler frequency around 1000 Hz) and between 20 and 60 Hz in the low flow range (mean Doppler frequency below 200 Hz).

LEg venous oxygen saturation in the evaluation of intra-operative blood flow during arterial reconstructive surgery

In twenty-four patients, undergoing a femoro-popliteal saphenous vein bypass graft for symptomatic atherosclerotic occlusion of the superficial femoral artery, oxygen saturation values for the femoral and popliteal veins were compared to the directly measured blood flows in the common femoral artery and in the bypass graft, respectively. Blood flow and venous oxygen saturation increased significantly after transfusion of 900 ml of blood. Pharmacological vasodilation caused a significant increase in blood flow both before and after transfusion, whereas the changes in venous oxygen saturation were significant only before blood transfusion. A close statistical relationship was found between initial femoral venous oxygen saturation and initial blood flow in the common femoral artery as well as between initial popliteal venous oxygen saturation and initial byapss blood flow. However, especially at low saturation values, the evaluation of blood flow was very uncertain. Whole leg and lower leg oxygen uptakes were not altered by intraoperative changes in blood volume. It is concluded that blood flow in the common femoral artery and the bypass graft can be roughly estimated from analysis of oxygen saturation in the femoral and popliteal veins, respectively. Furthermore, by determining leg venous oxygen saturation both before and after flow augmentation, induced by pharmacological vasodilation, a conception of the load on the vascular system may be obtained.

Haemodynamic studies in renal artery stenosis

With the aid of an electromagnetic flowmeter and a mechano-electrical pressure transducer, peroperative renal artery blood flows and pressures in the aorta and the renal artery distal to the stenosis were measured in 48 patients with renovascular hypertension before and after renal artery reconstruction. Calculations of pressure difference across the stenosis and resistances over the stenosis and the renal parenchyma were made. There was a significant blood flow increase through the renal artery after arterial reconstruction in all patients, irrespective as to whether they were normotensive, improved or failures postoperatively. The pressure difference across the stenosis after reconstruction was eliminated in all 3 groups of patients. The highest resistances over both the stenosis and the renal parenchyma before arterial reconstruction were found in the normotensive group, but there was no significant variation between the 3 groups. After reconstruction, a significant decrease in resistance over the renal parenchyma was found in the normotensive and the improved group of patients. The resistance of the stenosis was higher than over the renal parenchyma only in the postoperatively normotensive patients. This indicates that the removal of a resistance over the renal artery stenosis, that is higher than that over the renal parenchyma, gives the most favourable long-term results concerning normalization of the blood pressure.

Intra-operative haemodynamic findings and their prognostic significance in femoropopliteal reversed saphenous vein graft bypass operations

Post-reconstructive graft blood flow (basal flow') was measured by means of an electromagnetic flowmeter in 72 consecutive femoropopliteal reversed saphenous vein grafts performed for occlusive atherosclerosis causing leg ischaemia. In the last 48 operations, post-reconstructive pre- and post-graft intra-arterial pressures were recorded, and graft blood flow after papavarine-induced vasodilation ('maximal flow') was measured. Mean 'basal flow' was 132 ml/min, mean 'maximal flow' 285 ml/min and mean flow increment after papavarin 136%. Mean 'basal flow', mean 'maximal flow' and mean flow increment after papavarine were higher in the claudication group than in the gangrene group, and higher in cases having good angiographic run-off than in those having poor. Mean common femoral and popliteal pressures were 85 and 80 mmHg, with no differences between different groups. Mean post-reconstructive graft pressure gradient was 5.1 mm. Mean peripheral resistance was 769 milliprus, being higher in the gangrene group and in cases having poor run-off. The prognostic significance of these findings was evaluated by analysing patency in all patients living 3 months ('early') and 1 year ('late') after operation. Cases having a 'basal flow' less than 60 ml/min, a 'maximal flow' less than 200 ml/min or a flow increment after papavarine less than 100% had significantly poorer 'early' patency (p less than 0.01) and 'late' patency (p less than 0.05). Femoral and popliteal pressures, graft pressure gradients and peripheral resistance did not influence patency significantly.