- Email: [email protected]
Monitoring functional patency of in situ saphenous vein bypasses: The impact of a surveillance protocol and elective revision
Monitoring functional patcncy of in situ saphcnous vein bypasses: The impact of a surveillance protocol and elective revision Dennis F. Bandyk, M D , David D. Schmitt, M D , Gary R. Seabrook, M D , M a r k B. Adams, M D , and Jonathan B. Towne, M D , Milwaukee, Wis. Implementation of a protocol that monitored in situ saphenous vein bypass hemodynamics for low-flow states provided insight into the pathophysiologic characteristics and time course of graft failure. From 1981 to 1988, 250 in situ bypasses to popliteal (n = 83) or tibial (n = 167) arteries were performed in 231 patients. Indications for operation included critical limb ischemia in 232 cases (93%), popliteal aneurysm in 11 cases (4%), and disabling claudication in seven cases (3%). Arterial pressure measurements, continuous-wave Doppler spectral analysis, and duplex ultrasonography were used to assess patency, detect hemodynamic changes indicative of graft stenosis, and localize anatomic hemodynamic changes indicative of graft stenosis. Seventy grafts with correctable anatomic lesions (retained venous valves, graft stenosis, arteriovenous fistula, native vessel atherosclerosis) that decreased graft blood flow or ankle arterial pressure or both were identified. Correction of vein conduit or anastomotic lesions comprised 73 (77%) of the 95 revisions performed. Vein-patch angioplasty of a stenosis was the most common secondary operation performed. Graft revision was highest in the perioperative period (10% at 30 days), decreased to 7% per 6-month interval until 18 months, and was 3% per year thereafter. The primary patency rate of grafts not identified to have a correctable lesion was 86% at 4 years, a level similar to the secondary patency of 81% for grafts requiring one or multiple revisions. The surveillance protocol identified grafts with correctable lesions before thrombosis thereby permitting elective revision of patent grafts. Hemodynamic studies confirmed that a frequent mechanism of late failure of grafts was the development of a low-flow state produced by lesions not amenable to revision. (J VASC SURG 1989;9:286-96.)
The steady decline with time in the patency of vascular grafts mandates a protocol of postoperative surveillance to identify grafts at risk for thrombosis. The correction of lesions before graft thrombosis can have a significant impact on long-term patency and is particularly important for vein grafts because most will not maintain patency after thrombectomy'alone or with vein-patch angioplasty. 1-s Since we adopted the in situ bypass grafting technique in 1981, 6 we instituted a vascular laboratory protocol that monitored hemodynamics of the revascularized limb (arterial systolic pressure) and the in situ venous conduit (blood flow velocity). Initially the goal of hemodynamic monitoring was to assess technical adequacy From the Department of Surgery, The Medical Collegeof Wisconsin, and the SurgicalService,VeteransAdministrationMedical Center. Presented at the Thirty-sixth Scientific Meeting of the North American Chapter, International Society for Cardiovascular Surgery, Chicago, Ill., June 14-15, 1988. Reprint requests:DennisF. BandylqMD, Departmentof Surgery, MCMC, 8700 W. WisconsinAve., Milwaukee,WI 53226. 286
of the bypass, but the detection of vein graft stenosis and other lesions that increase the risk of sudden graft thrombosis became equally important. 7 Implementation of this protocol resulted in a progressive de.& cline in primary graft patency (patency uninterruptcia by revision) because of the revision of patent grafts identified to have hemodynarnicaUy important lesions. 5 Serial evaluation of vein bypass hemodynamics has proved more accurate than limb blood pressure measurements in both predicting the initial technical success and identifying deterioration in functional patency at a time when developing occlusive lesions are easily managed by operative revision or percutaneous transluminal angioplasty (PTA). s'9 In this report we detail our experience with a protocol, which is based on hemodynamic testing performed in the noninvasive peripheral vascular laboratory, that monitored the functional patency of 250 consecutive in situ saphenous vein bypasses. Graft surveillance has improved our understanding of the mechanisms of in situ vein graft failure; these mechanisms include (1) the detection of residual
Volume 9 Number 2 February 1989
Monitoring functional patency of vein bypasses 2 8 7
Table I. Distribution of distal anastomosis for 250 in situ saphenous vein arterial bypasses Sites of distal ana~omosis No. Popliteal artery Above-knee Below-knee Isolated segment Tibial-peroneal mank Posterior tibial artery Peroneal artery Anterior tibial artery Dorsalis pedis artery Sequential bypass Isolated popliteal-posterior tibial Posterior tibial-dorsalis pedis Peroneal-anterior tibial Peroneal-posterior tibial
1 82 4 3 59 55 38 3 1 2 1 1
:chnical defects related to the in situ grafting technique, (2) the location and morphologic characteristics of graft stenosis, a term denoting acquired myointimal or atherosclerotic lesions in the venous conduit, anastomotic sites, or adjacent native arteries, and (3) graft thrombosis caused by thromboembolism or hypercoagulable states. The impact of a surveillance protocol can be assessed by comparing the secondary patency of revised grafts with the primary patency of grafts not identified to have a correctable abnormality after operation. Graft failure may occur despite monitoring because of failure of the surveillance protocol, failure of graft revision, or identification of lesions not amenable to reconstructive or restorative procedures. The observations and hemodynamic data reported in this study provide expected results up to 4 years after operation for in situ ~aphenous vein grafts anastomosed to the popliteal or tibial arteries. MATERIAL AND METHODS From January 1981 to December 1987 we performed 163 femorotibial, 83 femoropopliteal, and four femoropopliteal (isolated segment) in situ saphenous vein bypasses for lower-extremity occlusive or aneurysmal disease in 231 patients. The study included 190 men and 41 women whose ages ranged from 42 to 92 years (mean 65 years). The incidence of tobacco use was 82%; hypertension, 63%; and heart disease defined by history of angina pectoris, myocardial infarction, or congestive heart failure, 60%. One hundred twenty-two bypasses were done in 118 patients with diabetes, 96 of whom required insulin for control of hyperglycemia. Indications for bypass grafting included critical limb ischemia (n = 232, 93%), popliteal artery an-
Table II. Requirements for calculation of Doppler-derived blood flow velocity from in situ saphenous vein bypasses Normal laminar flow pattern Accurate measurement of Doppler beam angle (0) Vessel segment with constant diameter No arteriovenous fistula distal to recording site Measurement of Doppler frequency shift (Fs) Doppler equation * *Blood flow velocity (cm/sec) = C Fs/2 Fo cos (0), where C indicates speed of sound in tissue (1.54 × 104 cm/sec), Fo is beam_ frequency of Doppler probe, and 0 is Doppler beam angle (cos 0 = 0.5).
curysm (n = 11, 4%), and life-style-limiting claudication (n = 7, 3%). In all but seven limbs the proximal anastomosis was to the common femoral artery. Other sites of graft origin included aortoprofunda femoris prosthetic graft limbs (n = 3), superficial femoral artery (n = 1), or popliteal (n = 3) artery. The choice of site of distal anastomosis was based on angiographic criteria of thc least-diseased artery with patency to the ankle or pedal arch. Table I lists the sites of distal anastomosis for the 250 in situ grafts. Five patients required short (less than 15 cm long) sections of reversed saphenous or cephalic vein to rcplace sclerotic vein segments or obtain sufficient graft length for anastomosis to the outflow artery. The technique of in situ bypass grafting has not been modified from our earlier reports, s,6 Intraoperative pulsed-Doppler spectral analysis remains an important component of the procedure and was used to assess both technical adequacy and graft hemodynamics. 7 Doppler flowmeter signal analysis was used to identify patent vein side branches for ligation, to assess valve and anastomotic sites for residual flow disturbances produced by intact cusp(s) or suture stenosis, respectively, and to calculate blood flow velocity in the distal graft segment (typically the smallest-diameter segment of a tapered saphenous vein left in situ). Listed in Table II are the requirements for accurate calculation of blood flow velocity. Blood flow velocity was calculated under resting basal conditions approximately 20 minutes after restoration of flow. Ifa low peak systolic blood flow velocity (<40 cm/sec) was measured in the graft segment of smallest diameter, particularly if associated with absent forward diastolic flow (high peripheral vascular resistance), a thorough search for technical error was carried out. In addition, papaverine hydrochloride (30 rag) was injected into the graft before the arteriogram was completed to eliminate any component ofoutttow resistance caused by vasospasm. Persistent
Journal of VASCULAR SURGERY
288 Bandyk et al.
Table Ill. Life-table analysis: Secondary patency data of 83 femoropopliteal and 163 femorotibial/4 femoropopliteal (isolated segment) in situ saphenous vein bypasses Interval (mo)
No. grafts at r~k
No. grafts failing
No. grafts not observed during period Death
Withdrawn ~
Interval patency rate (%)
Cumulative patency rate (%)
100 99 100 100 96 98 96 97 100 92
100 99 99 99 95 93 89 86 86 86
--
--
96 97 96 96 100 100 100 98 100 100 --
96 93 89 86 86 86 86 84 84 84 --
Femorop~liteMgrafts 0-1 1-3 3-6 6-12 12-18 18-24 24-30 30-36 36-42 42-48 >48t
83 81 75 65 60 47 41 34 21 13 9
0 1 0 0 2 1 1 1 0 0
2 3 3 1 3 1 1 1 2 1
0 2 7 4 8 4 5 11 6 2
i
0
i
5 1 14 7 4 4 3 3 4 2 1
1 5 4 15 5 18 7 12 7 5 3
Femorotibial/ isolated popliteal artery segment 0-1 1-3 3-6 6-12 12-18 18-24 24-30 30-36 36-42 42-48 >48t
167 154 144 120 94 85 63 53 37 26 19
7 4 6 4 0 0 0 l 0 0 3
+Reasons for withdrawal of grafts were duration and loss to follow-up. t Standard error of cumulative patency greater than 9%.
low graft flow after exclusion of technical error and vasospasm prompted sequential bypass to a second outflow artery if anatomically possible. The intraoperative hemodynamic studies served as a baseline for comparison with postoperative studies. Graft surveillance protocol. All patients underwent preoperative and serial postoperative noninvasive hemodynamic testing (intervals of 1 and 7 days, 6 weeks, and 3 months). In situ bypass hemodynamics were evaluated by a combination of noninvasive vascular laboratory methods: (1) measurement of resting limb arterial pressure by the transcutaneous Doppler ultrasonographic flow detection technique, (2) transcutaneous, continuouswave (CW) Doppler spectral analysis (5 MHz probe frequency, 45-degree Doppler angle to the skin) of graft blood flow patterns, and (3) duplex ultrasonography. The anlde-brachial systolic pressure index (ABI) was calculated for each limb by dividing the ankle systolic pressure by the higher of two simultaneously obtained brachial artery pressures. Duplex scanning and CW Doppler velocimetry were added to the postoperative surveillance protocol in 1983. Duplex scanning was used to map the entire bypass graft for flow abnormality before patient dis-
charge, and to locate the site of graft stenoses when a low graft flow state was identified by CW Doppler flow analysis (peak systolic flow velocity less than 45 cm/sec or decrease greater than 30 cm/sec compared to earlier outpatient evaluation). We have found that duplex examination of grafts that developed a low blood flow velocity can localize an occlusive lesion in most (85%) patients, and can categorize the severity of stenosis based on velocity spectra changes at and distal to the lesion, s'9 In general the development of graft stenosis can be predicted from changes in configuration and magnitude of the graft velocity waveform recorded by either a transcutaneous CW Doppler flowmeter or duplex scanner. Graft revision, if technically feasible, was recommended when either duplex scanning or angiography or both identified a greater than 50% diameterreducing lesion. Data analysis. Graft patency was determined by physical examination, decrease in ABI, and duplex scanning. An interval decrease in ABI greater than 0.2 was used as a criterion for arteriography or duplex scanning to assess graft patency. Primary and secondary graft patency rates were determined by the life-table method. Graft revisions, such as ligation of
Yolumr 9
Number2 1-~1.____
Monitoring functional patency of vein bypasses 289
1 0 ~ 0
(83) (75) (60) . . . . 100" __~__, • : -" -" . z' .tqll T ,~41) 7 " 3 4 ) 16~;~t'~'o'"'~":
80-
(144,
-
- P:
I
-J i(21)
;;at ;;;;t/;;;Ti;'3/'I
I
(13)
86% ,,o,o
r
PERCENT
60' PATENT
40'
20'
|
6
,
i
i
12
=
18
MONTHS
•
|
24 AFTER
,
,
30
,
n
36
,
,
42
,
=
48
SURGERY
Fig. 1. Secondary patency for 250 in situ saphenous vein bypasses calculated by life-table method. Closed circles indicate patency for 83 femoropopliteal bypasses, and open circles indicate patency of i63 femorotibial and four femoropopliteal (isolated segment) bypasses. Difference in patency was statistically significant by Wilcoxon test (p < 0.03). pressure-reducing arteriovenous (AV) fistulas, incision of residual valve leaflets, or PTA of in situ vein or adjacent native arterial stenoses, excluded the graft from primary patency. If graft patency was restored by thrombectomy, thrombolysis, or revision of a patent bypass, the graft was listed under secondary patency. Patency rates relative to the outflow artery (popliteal versus tibial) and the requirement for revision were compared for statistical difference by means of the generalized Wilcoxon test. Arterial pressure and blood flow velocity data are expressed as mean _+ SD. RESULTS Seven patients died within 30 days of operation, ,or an operative mortality rate of 3%. Five patients died of myocardial infarction, one patient died of hemorrhagic shock because of a displaced ligature, and one patient died at home of undetermined cause. The cumulative secondary patency data to 4 years after operation for the 83 femoropopliteal, 163 femorotibial, and four femoropopliteal (isolated segment) in situ bypasses are listed in Table III and presented graphically in Fig. 1. Grafts to isolated popliteal segments and sequential bypasses were included with the femorotibial graft results because of their similar hemodynamics. Secondary patency of femoropopliteal bypasses was higher (Wilcoxon test, p < 0.028) compared to bypasses to tibial arteries and isolated popliteal artery segments. This was due to a higher femoropopliteal graft patency during the first 2 postoperative years. Despite a steady decline
in primary patency to 63% and 61% at 4 )rears for femoropopliteal and femorotibial bypasses, respectively, secondary patency was in excess of 80% for both graft configurations. The cumulative survival of patients with patent grafts was 82% at 1 year, 77% at 2 years, 74% at 3 years, and 65% at 4 years. Seventy grafts were identified to have correctable anatomic lesions and underwent a total of 95 revisions. No lesion was identified in three grafts that had thrombectomy alone. Thirteen bypasses had multiple revisions with one graft revised four times over a period of 3 years. Correction of vein conduit or anastomotic abnormalities comprised 79% of revisions, whereas correction of atherosclerotic lesions identified in the iliac or infrapopliteal arteries accounted for 10% and 11% of revisions, respectively (Table IV). The mean time interval from primary operation to graft revision ranged from 8 months for lesions involving the in situ bypass (conduit and anastomoses), to 10 months for iliac artery stenosis, to 16 months for outflow tract disease progression. Most (18/21) graft lesions corrected within 30 days of operation (retained valve leaflet, residual AV fistula, anastomotic suture stenosis ) were identified and assessed for hemodynamic significance by duplex scanning. Correction of these lesions increased distal graft blood flow velocity and limb arterial pressure. Occlusive and aneurysmal lesions corrected more than 1 month after operation were identified by a combination of physical examination, noninvasive hemodynamic testing, and arteriography. The development of graft stenosis accounted for 64 (90%)
}'oarna/of 290
VA~CUL~
Bandyk et al.
Table IV. Site and morphologic characteristics of 92 lesions revised after in situ saphenous vein bypass grafting Site and morphologic characteristics of lesion
No.
Vein conduit Valve site stenosis Arteriovenous fisttda Retained valve leaflet Graft entrapment Graft torsion Graft aneurysm Unexplained thrombosis Total
20 9 5 4 2 1 6 47
Anastomotic sites Stricture of myointimal hyperplasia/atherosclerosis Suture stenosis Anastomotic pseudoaneurysm Total
19 6 1 26
Native vessel atherosclerotic stenosis Iliac artery Popliteal artery Tibial arteries Total
9 1 9 19
of 71 late graft revisions. Transformation of the graft velocity waveform from a triphasic to a monophasic or biphasic configuration coupled with a decrease in peak systolic blood flow velocity reliably predicted a remote occlusive lesion. Hemodynamic data were available for comparison from 56 in situ bypasses with diagnosed graft stenosis. Peak systolic velocity in the distal graft segment ranged from 12 to 50 cm/sec (mean 32 + 8 cm/sec), a level decreased from 75 + 16 cm/sec measured after the primary operation. The decrease in ABI ranged from 0.1 to 0.55 (mean 0.24 ___ 0.12). In 19 (34%) of the 56 limbs, resting ABi did not identify the presence of graft or native artery stenosis. Patients with stenosed but patent in situ bypasses showed no symptoms at the time of diagnosis. Graft revision rate (calculated by life-table method) was highest in the perioperative period (10% at 30 days), decreased to 7% per 6-month interval until 18 months after operation, and was 3% per year thereafter. Overall, 70 (28%) of the 250 bypasses were revised after the primary operation. Ten bypasses (10%) were thrombosed at the time of graft revision. Graft thrombectomy alone restored patency in only one graft. Three bypasses were salvaged with thrombectomy and revision of the outflow tract by sequential bypass grafting, anastomosis to another outflow artery, or vein-patch angioplasty. Six grafts were deemed failures and replaced by a polytetra-
Table V. Ninety,five secondary procedures performed on 70 in situ saphenous vein bypasses to restore normal functional patency Time t~fter primary operation
Operative procedure
<1 mo
Valve leaflet incision Thrombectomy alone Revision of anastomosis
5 3~ 2
Ligation of AV fistula Correction graft torsion/entrapment Vein-patch angioplasty Sequential vein bypass
6 2 5 1
Interposition vein bypass Autologous vein PTFE PTLA Total
>1 mo
3 4 28 15 5 5
__ 24
11
71
AV, arteriovenous fistula; PTLA, percutaneous transluminal balloon dilatation; PTFE, expanded polytetrafluoroethylene. nOne bypass patent after thrombectomy.
fluoroethylene conduit. Thrombolytic therapy restored patency in three grafts before revision for vein conduit stenosis (n = 2) or aneurysm (n = 1). A variety of secondary procedures were performed to restore bypass anatomy and functional patency to normal (Table V). Vein-patch angioplasty of a stenosed valve or anastomotic site was the most common secondary reconstructive procedure (33 of 70 procedures). Transluminal balloon dilatation of a vein conduit stenosis was successful in only two of four cases. Graft thrombosis occurred 1 week after dilatation of a high-grade stenosis loated in the dist~l~ segment of a femoral posterior-tibial bypass. Despit~e graft thrombectomy and vein-patch angioplasty, patency could not be restored. In contrast, all PTAs of native artery stenoses remained patent and improved graft and limb hemodynamics. Two secondary operative procedures performed on patent bypasses failed within 90 days. Late failure of direct surgical repairs involved repeat stenosis of vein-patched anastomotic sites (n = 4) or occlusion of cephalic vein sequential grafts (n = 3) performed for anastomotic myointimal hyperplasia. Revision of occlusive lesions increased mean peak systolic flow velocity in the distal graft from 33 cm/sec to 77 +_ 17 cm/sec (p < 0,0001, Mann-Whitney U test), and ABI increased to 0.93 --- 0.1. As shown in Fig. 2 and Tabl e VI, the secondary patency of 70 revised in situ bypasses up to 4 years
Volume 9 Number 2 February 1989
Monitoring functional patency of vein bypasses 291
Table YI. Life-laDle analyaia: Patcncy data of 280 in situ saphenous vein bypasses relative to requirement for graft revision after operation Interval
(too)
No. grafts not observed during period
No. grafts at risk
No. grafts failing
Death
180 169 155 130 106 90 67 55 38 22 18
4 3 2 4 1 1 1 0 0 0 1
3 2 4 0 1 0 0 2 0 0 3
Withdrawn +
Interval patency rate (%)
Cumulative patency rate (%)
6 4 15 6 6 5 3 3 6 1 1
1 7 8 14 9 17 8 14 10 3 2
98 98 99 97 99 99 98 100 100 100 --
98 96 95 92 91 89 86 86 86 86 --
1 0 2 2 1 0 1 1 0 2 0
0 0 3 5 4 5 4 9 3 4 2
96 97 94 100 98 100 100 94 100 100 --
96 93 88 88 86 86 86 81 81 8I --
No r~is~n 0-1 1-3 3-6 6-12 12-18 18-24 24-30 30-36 36-42 42-48 >48t
Singleor multipkgraft~viswn(s) 0-1 1-3 3-6 6-12 12-18 18-24 24-30 30-36 36-42 42-48 >48#
70 66 64 55 48 42 37 32 20 I7 11
+Reasons for withdrawal o f grafts were duration and loss to follow-up. j-Standard error of cumulative patency greater than >9%.
after operation was similar to the primary patency of 180 bypasses not identified to have a correctable abnormality, 81% versus 86%, respectively. Of the 32 in situ bypass failures recorded during patient follow-up, seven occurred within 30 days of operation. Three early graft failures occurred despite repeated thrombectomies and operative revision. ,¢iechanisms of early graft failure included low flow in two nonrevised grafts, associated hypercoagulable state (n = 2), or trauma caused by valve incision with formation platelet aggregation and thrombus (n = 3). Late (greater than 1 month) failure occurred in 12 revised and 13 nonrevised in sire bypasses. Hemodynamic studies confirmed the most common mechanism of late failure was the development of a low-flow state not amenable to revision (Table VII). Serial evaluations of these grafts showed a decrease in blood flow velocity compared to initial postoperative levels but retention of a triphasic veiocity waveform configuration signifying increased outflow resistance (Fig. 3). Angiographic evaluation showed a diseased outflow tract with atherosclerotic disease progression compared with pre- or intraoperative angiograms (Fig. 4). Peak systolic velocity
Table VII. Mechanisms involved in the late (> 1 mo) failure of 25 in situ saphenous vein bypasses Mechanism offailure
Low-flow state Prior graft revision N o revision
Missed graft stenosis Incomplete graft surveillance + Thromboembolism PTA failure
Infection#
No. 8 3 3 6 1 1 3
+Four grafts failed before 1983. tAll grafts patent at time of amputation for leg/foot infection.
in these grafts was typically less than 35 cm/scc (range 18 t o 40 cm/sec), and graft thrombosis occurred within 3 to 9 months of identifying the lowflow state. In three patients with unexpected graft failure, review of vascular laboratory studies indicated a graft stenosis was missed because of erroneous interpretation. Six of the 25 late graft failures occurred in patients not observed at regular intervals
Journalof VA~ULAK SURGERY
292 Bandyk et al.
100'
[..~o) (1~5)I(130)T ,'~.i"~'"':'-.-
0o.
:
(8,) •
""'
(90)
_
T
-
-, (so
T
r(5 ) I
i
I
i
I 86% 81%
PERCENT 60" PATENT 40"
20
|
|
6
|
i
|
|
i
,
|
i
|
|
12 18 24 30 36 MONTHS AFTER SURGERY
,
|
42
|
|
48
Fig. 2. Life-table analysis. Cumulative patency for 70 revised (open circles) and 180 nonrevised (closed circles) in situ saphenous vein bypasses. Number at each time interval after operation indicates number of grafts at risk. Difference in patency not statisticallysignificant by Wilcoxon test (p = 0.1). in the vascular laboratory or before 1983 when graft surveillance was by physical examination and limb pressure measurements. During the last 3 years of the study, unexpected failure of grafts judged to have normal hemodynamics occurred in only one of the 100 to 120 patients routinely studied by vascular laboratory personnel. This patient had previously undergone a vein-patch angioplasty of a valve site and probably experienced a repeat stenosis at this location. DISCUSSION Routine hemodynamic surveillance of in situ bypasses confirmed the most common mechanism of graft failure was the development of low-flow state (low blood flow velocity), which increased the likelihood of unexpected graft thrombosis. Occlusive lesions in the vein conduit, anastomotic sites, or inflow and outflow vessels tended to develop without symptoms, but bypass graft hemodynamics were altered sufficiently that detection before graft thrombosis was possible. Other conditions such as hypercoagulable states (antithrombin III, protein C and S deficiencies, heparin-induced platelet aggregation, abnormal plasminogen), graft or anastomotic aneurysm, or thromboembolism can also result in graft thrombosis. These uncommon mechanisms of graft failure should be suspected when graft thrombosis occurs unexpectedly and earlier (within three months) graft surveillance showed normal hemodynamics. In situ bypass surveillance was best per-
formed with direct physiologic testing methods, such as CW Doppler velocity waveform analysis or duplex scanning. Doppler-derived ankle systolic pressure measurements were not predictive of graft flow and were less sensitive than duplex scanning in the detection of graft stenosis. The importance of routine vascular laboratory follow-up after femoropopliteal vein bypass grafting has been emphasized in multiple reports. 1,3-s,1° Less than one fourth of the patients will have unequivocal evidence of graft stenosis based on decreased pedal/graft pressure pulse wave or recurrence of limb ischemia.1 The absence of symptoms in our patients who had graft stenosis was not surprising since lesions were identified early, before gra~ thrombosis, and ABI had decreased only moderateiy from initial postoperative levels to a mean of 0.62. Although the development of a low-flow state in the in situ vein conduit predicted the presence of an occlusive lesion, graft revision by direct surgical intervention or PTA was not always possible. Atherosclerotic disease progression involving the outflow arteries, typically immediately adjacent to anastomotic sites, and failure of secondary graft procedures were the most common causes of late graft failure. One fourth (6/25) of the late graft failures had outflow tracts judged not amenable to sequential bypass grafting or PTA for the purpose of increasing graft blood flow velocity. Serial hemodynamic testing predicted failure in these bypasses with an acquired lowflow state. Blood flow velocity in the graft segment of smallest diameter was typically less than 35~
Volume 9 Number 2 February 1989
Monitoring functional patency of vdn bypasses 293
POSTOPERAT|VE TiME iNTERVALI
m
C
Fig. 3. Development of low graft blood flow velocity after operation. Serial velocity spectra recorded from disal segment of popliteal-posterior tibial in situ saphenous vein graft with a duplex scanner. Hyperemic flow pattern (forward flow throughout the pulse cycle) present during first week after operation evolved to triphasic configuration at 3 months. Decrease in peak systolic flow velocity noted 14 months after operation was due to progression of disease and occlusion of retrograde flow in the posterior tibial artery (see Fig. 4). rm/scc, and the presence of a triphasic or monophasic velocity waveform configuration (absent diastolic forward flow) signified high outflow resistance. The thrombotic threshold velocity, n a minimum blood flow velocity below which thrombus formation occurs, of in situ bypasses is variable and depends on the time interval after operation, thrombogenicity of the graft flow surface, coagulation status of the patient, and configuration of the velocity waveform. Although the detection of low blood flow velocity in the graft was helpful in the recognition of technical errors and graft stenosis, no absolute minimum level has been identified that predicts graft failure. 12A peak systolic flow velocity less than 45 cm/sec was measured in 96% of grafts with correctable lesions. Graft blood flow vdocity in this range should prompt an angiographic or duplex examination to identify a cor:ectable lesion. When a low-flow state in a graft was
Fig. 4. Angiograms of popliteal-posterior tibial in situ saphenous vein bypass with diseased outflow artery. Arrow indicates site of distal anastomosis. found not to be amenable to revision, prolonged patency of the bypass was uncommon. Typically the grafts remained patent for 3 to 9 months. Thc efficacy of using oral anticoagulant therapy for the purpose of prolonging patency in these grafts with diseased outflow tracts is unknown. A rccent report involving femoropopliteal polytetrafluoroethylene grafts, a bypass with a lower than average flow rate, suggested the incidence of unexpected thrombosis can be decreased by long-term sodium warfarin (Coumadin) anticoagulation. 13 An important observation of this study was the similar 4-year cumulative patency rates for nonreviscd (86%) and revised (81%) in situ bypasses. Revision of stenotic lesions returned functional graft patency to normal levels; graft blood flow velocity increased, the velocity waveform reverted to a triphasic configuration, and limb arterial pressure returned to normal. The elective revision of patent bypasses, even before the onset of ischemic symptoms,
294
Journal of VASCULAR SURGERY
Bandyk et al.
minimizes the likelihood that thrombectomy will be a necessary part of a graft salvage operation. Our results of secondary operations on patent grafts were close to the 5-year patency of 90% reported by Whittemore et al.2 for reversed saphenous vein grafts repaired before thrombosis. We believe the surveillance plan and diagnostic criteria used for in situ bypass follow-up is also applicable for monitoring functional patency of reversed saphenous vein grafts. A correctable lesion was identified in most grafts with abnormal hemodynarnics. Repair of vein conduit or anastomotic lesions comprised 79% of the secondary reconstructive or restorative procedures performed. Vein-patch angioplasty of stenosis was the most common reconstructive procedure, and repeat stenosis of this repair was uncommon. Interposition grafting was used to correct long strictures (>4 cm) of the venous conduit or for recurrent graft stenosis. Use of large-diameter (>6 ram) cephalic veins was a sequential bypass was associated with low blood flow velocity (range 20 to 45 cm/sec) in this conduit after operation, a level that may have contributed to the failure of this vein when used in secondary operations. Percutaneous transluminal balloon dilatation of native artery stenoses was a durable restorative procedure, but treatment of vein graft stenoses was not uniformly successful, particularly when performed within 6 months of the primary operation. The development of graft stenosis within the first year after operation was frequently associated with minor technical imperfections in Gypass construction with valve and anastomotic sites having residual flow abnormality. Based on the time course of graft revision and failure, hemodynamic surveillance should be intensive in the immediate postoperative period and continue at frequent intervals for 2 years, at which time the incidence of graft stenosis decreases. The ability of duplex scanning to localize an anatomic abnormality in 85% of patients with suspected graft stenosis, coupled with its expense and comfort advantages compared to angiography, make it the preferred method for postoperative surveillance. Lesions judged to reduce diameter by greater than 50% should be considered for revision. Despite intensive anatomic and hemodynamic monitoring of in situ bypasses, graft failure still occurs. Graft thromboembolism, misinterpretation of vascular laboratory studies, and lack of patient compliance with the outpatient appointment schedule or recommendation to undergo angiography and graft revision contribute to a small but irreducible incidence of graft failure. Failure of graft revision and identification of lesions not amenable to surgical intervention remain unsolved problems contributing
to late vein graft failure. Improvement in these areas will require a better understanding o f the pathologic nature of artery wall healing, atherosclerotic disease progression, and myointimal hyperplasia. Patient assessment at 3-month intervals after discharge from the hospital appears adequate to detect the occasional malignant progression ofmyointimal lesions at valve and anastomotic sites. The graft surveillance protocol can be implemented by vascular laboratory personnel, and only when abnormalities are identified is a physician-directed evaluation required. This approach does require the vascular technologist to have a thorough knowledge of bypass graft anatomy in each patient, and an understanding of graft and limb hemodynamics after arterial bypass grafting. The goal of noninvasive hemodynamic testing is to provide the physician with reliable guidelines to aid decision making regarding technical adequacy and functional patency of the arterial bypass. Duplex scanning w~,~ blood flow velocity calculation has the capability of providing the necessary anatomic and physiologic data to achieve this goal. REFERENCES
1. Berkowitz HD, Hobbs CL, Roberts B, et al. Value of routine vascular laboratory studies to identify vein graft stenosis. Surgery 1981;90:971-9. 2. Whittemore AD, Clowes AW, Couch NP, et ai. Secondary femoropopliteal reconstruction. Ann Surg 1981;193:35-42. 3. Sladen JD, Gilmour JL. Vein graft stenosis. Characteristics and effect of treatrnent. Am J Surg 1981;141:549-53. 4. Cohen JR, Mannick IA, Couch NP, Whittemore AD. Recognition and management of impending vein graft failure. Arch Surg 1986;121:758-9. 5. Bandyk DF, Kaebnick HW, Stewart GW, Towne JB. Durability of in situ saphenous vein bypass: a comparison of primary and secondary patency. J VAsc SV•G 1987;2: 256-68. 6. Levine AW, Bandyk DF, Bonier PH, Towne JB. Less~,@ learned in adopting the in situ saphenous vein bypass. J VAsc SURG 1985;1:145-53. 7. Bandyk DF, Jorgensen RA, Towne JB. Intraoperative assessment of in situ saphenous vein arterial grafts using pulsed Doppler special analysis.Arch Surg 1986;121:292-9. 8. Bandyk DF, Kaebnick HW, Bergamini TM, et al. Hemodynamics of in situ saphenous vein arterial bypass. Arch Surg 1988; 123:477-82. 9. BandykDF. Monitoring functional patency of vasculargrafts. Sere Vasc Surg 1988;1:40-50. 10. Turnipseed WD, Acher CW. Postoperative surveillance: an effective means of detecting correctable lesions that threaten graft patency. Arch Surg 1985;120:324-8. 11. Sauvage LR, Walker MW, Berger KG, et al. Current arterial prosthesis. Arch Surg 1979;114:687-91. 12. BandykDF, Cato RF, Towne JB. A low blood flow velocity predicts failure of femoropopliteal and femorotibial bypass grafts. Surgery 1985;98:799-809. 13. Ffinn WR, Kohrer MJ, Yao JST, et al. Improved longterm patency of inftagenicular polytetrafluoroethylenegrafts.' J VASCSURG1988;7:685-90.
Volume 9 Number 2 February 1989
DISCUSSION
Dr. R o b e r t P. Leather (Albany, N.Y.). I compliment the presenter, Dr. Schmitt, on an elegant presentation and also compliment his fellow authors on a thorough analysis of a very careful and complete postoperative surveillance program on their patients in whom they have done in situ bypasses. There is a wealth o f information in the manuscript, and I highly recommend it for review when it is published. Our experience with virtually identical postoperative surveillance protocol in a similar patient population is basically the same, with two notable exceptions. One exception is the incidence o f revisions in this series, 70 out of 250, or 28% o f the conduits, versus 123 conduits threatened out o f 1066 in our series, for a gross incidence o f 11.5%. The other exception is that the revised bypass in our experience does not enjoy the same patency rate as the in situ conduit that did not require revision. : : There is a remarkably similar distribution o f distal anastomoses to tibial arteries for every popliteal artery. However, there is a difference in the distribution of tibial vessels used. The most common, 45% o f tibial bypasses , go to the peroneal artery, and in our experience the posterior tibial artery is the least frequently used. As has already been pointed out by many authors, the highest frequency o f revision occurs in the first two 6month periods and then drops to a less than 1% incidence by 2 years. Eighty percent o f the revisions are in the conduit itself, but the frequency is related to size. When size alone is observed, those grafts o f 3 mm or less have an incidence o f stenosis development that is double that of veins that are 4 mm or greater. Therefore, I ask the authors if they have any data on the distribution o f these stenoses, specifically the distal or proximal mobilized segment as compared to the undisturbed truly in situ portion. Also, I would like to know if they find any correlation between bypass diameter and the frequency o f these stenoses. Another correlation with size is that the development of a proximal stenosis averaged 17 months, and as one went further down the vein, the average development of a stenosis in the distal mobilized segment was 6 months. The other exception is that if one takes all those bypasses that have no revisions, the primary patency at 5 years is 79%. However, when one analyzes the revised bypasses, excluding the residual fistulas that are not bypass threatening, the patency at 5 years is 61%. This difference is statistically significant. The importance o f a proper surveillance program is very clearly shown when one uses the now conventional primary and secondary patency, as recommended by Dr. Rutherford's committee, with a patency improvement of approximately 15% by the application o f a proper surveillance program. We have not found the peak systolic velocity to be an accurate cutoff indicator because of the wide variation in diameters of bypass conduits. What we have found to correlate best with continued patency is the reactive hyperemic
Monitoring functional patency of vein bypasses
295
flow, that is, a reactive hyperemia i n d u c e d by a cuff compression for a period of 3 minutes. If this flow at least doubles then it is highly unlikely that a bypass-threatening lesion is developing. I thank the Society for the privilege and the invitation to discuss this paper. Dr. A n t h o n y D. W h i t t e m o r e (Boston, Mass.). As Dr. Leather mentioned, the manuscript presents a wealth of information that evokes a series o f questions far too numerous for a limited discussion. However, this series of in situ grafts, which were prospectively evaluated with duplex imaging, is one of the largest that I am aware of, and serves to underscore the importance o f identifying vein graft stenosis before actual thrombosis. We have been aware of this necessity for some time and have, for the most part, relied on recurrent ischemic symptoms or a significant decrease in ankle pressures to monitor graft patency. This present study further documents the fact that recurrent symptoms or changes in ankle pressures cannot be relied on to identify significantly stenotic lesions in nearly one third o f such patients. I have three questions. Prevention o f graft thrombosis is o f critical importance, and monitoring with a duplex scanner seems effective to this end. I ask Dr. Schmitt whether we should strictly adhere to his protocol with regard to the frequency of follow-up examinations. I ask this in light of the fact that whereas noninvasive vascular evaluation is increasingly reliable and applicable to more and more clinical situations, the equipment involved is of unconscionable expense, approaching $200,000 per unit, and many vascular laboratories are significantly overburdened with the multiplicity o f studies now required. Second, since your revision rate is higher when compared with our experience as well as Dr. Leather's, is it possible that the imaging techniques may, in fact, be too sensitive? Did you feel that you perhaps revised some grafts that might not have required it? Finally, whereas individual criteria continue to vary, we have begun to distinguish more effectively between primary and secondary graft patency rates. Since the secondary patency rates achieved with stenotic but patent grafts are dramatically discrepant from those applicable to thrombosed grafts, do you think it appropriate to fury_her subdivide the secondary patency curves? The revision of a thrombosed graft differs markedly from the management of a stenotic but patent graft, particularly with regard to the use of thrombolytic therapy. Both situations have, at present, very different anticipated outcomes, and it seems inappropriate to combine the two situations into one statistic. Dr. John Wolfe (London, England). As you know~ we have also had a considerable interest in this area, and one of the criticisms leveled against us is that a surveillance program o f this type is totally impractical in a busy clinical practice. Therefore we have tried to define our indications for doing various tests. We found that duplex-measured graft flow is extremely
296
Bandyk et al.
unreliable. We have also found that an isolated measurement of graft velocity is not specific or sensitive, and I would like to ask you whether you have looked at ankle brachial indices alone in these patients with very low graft velocities or asked them carefully about their symptoms. H o w many of these patients would have been picked up as having grafts at risk without the use of duplex? We have concentrated on the nonhemodynamically significant stenoses because we think that the other can be picked up by simple means. By measuring the graft velocity at one point and at an adjacent point 2 cm away, even when the graft flow remains high, we can pick up a nonhemodynamically significant stenosis and then watch the progression o f that stenosis. This ratio is unaltered by cardiac output or peripheral resistance and therefore is reproducible from one to the next. Therefore our policy has been to look at these grafts very carefully over the first 6 months and again at a year, and we have found that have all graft stenoses been picked up within 6 to 9 months of the graft being inserted. H o w many of your stenoses developed after that period? My final point is on semantics. Dr. Rutherford went to great efforts to introduce a term that we could all use, critical ischemia. This term was originally developed in the British Journal of Surgery by Crawford Jamieson's ad hoc committee. This term has already been abused. Limb salvage has become a worthless term because people are using the term limb salvage for legs that would remain viable without intervention. I suggest to you that in any group of patients there must be some patients who do not fit the term critical ischemia despite the fact that they have rest pain (i.e., either gangrenous or an ankle pressure of less than 40 mm Hg). There should be three groups in any study o f this type. There should be the patients with claudication, the patients with rest pain, and the patients with critical ischemia. Dr. Schmitt (dosing). I thank the discussants for their comments. In particular, I would like to thank Dr. Leather
Journal of VASCULAR SURGERY
and the Albany group for their contributions to the research and development of the in situ saphenous vein bypass technique. Dr. Leather asked us about our patency rates. For femoropopliteal bypasses our primary and secondary patencies are 63% and 86%, respectively, and for femorotibial bypasses at 4 years they are 61% and 84%, respectively. We have not established the critical thrombotic velocity as an absolute value, but the concept does have importance in identifying a threatened or failing graft. Once a velocity of 40 to 45 cm/sec is obtained, the grafts should be studied to identify a correctable cause for the decreasing velocity. The graft will not necessarily thrombose in the immediate future, but we have found these grafts usually fail within 3 to 9 months. We have no experience at this time with reactive hyperemia as a test to predict graft survival. Dr. Whittemore asked about our surveillance protocol. It consists of assessing the graft by (1) transcutaneous CW Doppler ultrasonic flow detection to obtain resting limb arterial pressures and (2) CW Doppler spectral analysis of graft blood flow patterns. Those patients who have a lo,~a flow state or a significant change from an earlier examination then undergo duplex scanning of the entire graft. Our revision rate may be high compared to some other studies, but we revise a graft, including ligation of an AV fistula, only if it is a hemodynamically significant lesion resulting in a low-flow state or a lesion that produces a 50% reduction in diameter. Dr. Wolfe asked about the reproducibility of graft velocity. We have found that by consistently measuring the velocities in the area of the graft of the smallest diameter, accurate, reproducible results can be obtained, identifying the development of a low-flow state. Patients we define as having critical ischemia are those that have rest pain on a nightly basis, nonhealing ulcers on their feet, or those who have gangrene.
The steady decline with time in the patency of vascular grafts mandates a protocol of postoperative surveillance to identify grafts at risk for thrombosis. The correction of lesions before graft thrombosis can have a significant impact on long-term patency and is particularly important for vein grafts because most will not maintain patency after thrombectomy'alone or with vein-patch angioplasty. 1-s Since we adopted the in situ bypass grafting technique in 1981, 6 we instituted a vascular laboratory protocol that monitored hemodynamics of the revascularized limb (arterial systolic pressure) and the in situ venous conduit (blood flow velocity). Initially the goal of hemodynamic monitoring was to assess technical adequacy From the Department of Surgery, The Medical Collegeof Wisconsin, and the SurgicalService,VeteransAdministrationMedical Center. Presented at the Thirty-sixth Scientific Meeting of the North American Chapter, International Society for Cardiovascular Surgery, Chicago, Ill., June 14-15, 1988. Reprint requests:DennisF. BandylqMD, Departmentof Surgery, MCMC, 8700 W. WisconsinAve., Milwaukee,WI 53226. 286
of the bypass, but the detection of vein graft stenosis and other lesions that increase the risk of sudden graft thrombosis became equally important. 7 Implementation of this protocol resulted in a progressive de.& cline in primary graft patency (patency uninterruptcia by revision) because of the revision of patent grafts identified to have hemodynarnicaUy important lesions. 5 Serial evaluation of vein bypass hemodynamics has proved more accurate than limb blood pressure measurements in both predicting the initial technical success and identifying deterioration in functional patency at a time when developing occlusive lesions are easily managed by operative revision or percutaneous transluminal angioplasty (PTA). s'9 In this report we detail our experience with a protocol, which is based on hemodynamic testing performed in the noninvasive peripheral vascular laboratory, that monitored the functional patency of 250 consecutive in situ saphenous vein bypasses. Graft surveillance has improved our understanding of the mechanisms of in situ vein graft failure; these mechanisms include (1) the detection of residual
Volume 9 Number 2 February 1989
Monitoring functional patency of vein bypasses 2 8 7
Table I. Distribution of distal anastomosis for 250 in situ saphenous vein arterial bypasses Sites of distal ana~omosis No. Popliteal artery Above-knee Below-knee Isolated segment Tibial-peroneal mank Posterior tibial artery Peroneal artery Anterior tibial artery Dorsalis pedis artery Sequential bypass Isolated popliteal-posterior tibial Posterior tibial-dorsalis pedis Peroneal-anterior tibial Peroneal-posterior tibial
1 82 4 3 59 55 38 3 1 2 1 1
:chnical defects related to the in situ grafting technique, (2) the location and morphologic characteristics of graft stenosis, a term denoting acquired myointimal or atherosclerotic lesions in the venous conduit, anastomotic sites, or adjacent native arteries, and (3) graft thrombosis caused by thromboembolism or hypercoagulable states. The impact of a surveillance protocol can be assessed by comparing the secondary patency of revised grafts with the primary patency of grafts not identified to have a correctable abnormality after operation. Graft failure may occur despite monitoring because of failure of the surveillance protocol, failure of graft revision, or identification of lesions not amenable to reconstructive or restorative procedures. The observations and hemodynamic data reported in this study provide expected results up to 4 years after operation for in situ ~aphenous vein grafts anastomosed to the popliteal or tibial arteries. MATERIAL AND METHODS From January 1981 to December 1987 we performed 163 femorotibial, 83 femoropopliteal, and four femoropopliteal (isolated segment) in situ saphenous vein bypasses for lower-extremity occlusive or aneurysmal disease in 231 patients. The study included 190 men and 41 women whose ages ranged from 42 to 92 years (mean 65 years). The incidence of tobacco use was 82%; hypertension, 63%; and heart disease defined by history of angina pectoris, myocardial infarction, or congestive heart failure, 60%. One hundred twenty-two bypasses were done in 118 patients with diabetes, 96 of whom required insulin for control of hyperglycemia. Indications for bypass grafting included critical limb ischemia (n = 232, 93%), popliteal artery an-
Table II. Requirements for calculation of Doppler-derived blood flow velocity from in situ saphenous vein bypasses Normal laminar flow pattern Accurate measurement of Doppler beam angle (0) Vessel segment with constant diameter No arteriovenous fistula distal to recording site Measurement of Doppler frequency shift (Fs) Doppler equation * *Blood flow velocity (cm/sec) = C Fs/2 Fo cos (0), where C indicates speed of sound in tissue (1.54 × 104 cm/sec), Fo is beam_ frequency of Doppler probe, and 0 is Doppler beam angle (cos 0 = 0.5).
curysm (n = 11, 4%), and life-style-limiting claudication (n = 7, 3%). In all but seven limbs the proximal anastomosis was to the common femoral artery. Other sites of graft origin included aortoprofunda femoris prosthetic graft limbs (n = 3), superficial femoral artery (n = 1), or popliteal (n = 3) artery. The choice of site of distal anastomosis was based on angiographic criteria of thc least-diseased artery with patency to the ankle or pedal arch. Table I lists the sites of distal anastomosis for the 250 in situ grafts. Five patients required short (less than 15 cm long) sections of reversed saphenous or cephalic vein to rcplace sclerotic vein segments or obtain sufficient graft length for anastomosis to the outflow artery. The technique of in situ bypass grafting has not been modified from our earlier reports, s,6 Intraoperative pulsed-Doppler spectral analysis remains an important component of the procedure and was used to assess both technical adequacy and graft hemodynamics. 7 Doppler flowmeter signal analysis was used to identify patent vein side branches for ligation, to assess valve and anastomotic sites for residual flow disturbances produced by intact cusp(s) or suture stenosis, respectively, and to calculate blood flow velocity in the distal graft segment (typically the smallest-diameter segment of a tapered saphenous vein left in situ). Listed in Table II are the requirements for accurate calculation of blood flow velocity. Blood flow velocity was calculated under resting basal conditions approximately 20 minutes after restoration of flow. Ifa low peak systolic blood flow velocity (<40 cm/sec) was measured in the graft segment of smallest diameter, particularly if associated with absent forward diastolic flow (high peripheral vascular resistance), a thorough search for technical error was carried out. In addition, papaverine hydrochloride (30 rag) was injected into the graft before the arteriogram was completed to eliminate any component ofoutttow resistance caused by vasospasm. Persistent
Journal of VASCULAR SURGERY
288 Bandyk et al.
Table Ill. Life-table analysis: Secondary patency data of 83 femoropopliteal and 163 femorotibial/4 femoropopliteal (isolated segment) in situ saphenous vein bypasses Interval (mo)
No. grafts at r~k
No. grafts failing
No. grafts not observed during period Death
Withdrawn ~
Interval patency rate (%)
Cumulative patency rate (%)
100 99 100 100 96 98 96 97 100 92
100 99 99 99 95 93 89 86 86 86
--
--
96 97 96 96 100 100 100 98 100 100 --
96 93 89 86 86 86 86 84 84 84 --
Femorop~liteMgrafts 0-1 1-3 3-6 6-12 12-18 18-24 24-30 30-36 36-42 42-48 >48t
83 81 75 65 60 47 41 34 21 13 9
0 1 0 0 2 1 1 1 0 0
2 3 3 1 3 1 1 1 2 1
0 2 7 4 8 4 5 11 6 2
i
0
i
5 1 14 7 4 4 3 3 4 2 1
1 5 4 15 5 18 7 12 7 5 3
Femorotibial/ isolated popliteal artery segment 0-1 1-3 3-6 6-12 12-18 18-24 24-30 30-36 36-42 42-48 >48t
167 154 144 120 94 85 63 53 37 26 19
7 4 6 4 0 0 0 l 0 0 3
+Reasons for withdrawal of grafts were duration and loss to follow-up. t Standard error of cumulative patency greater than 9%.
low graft flow after exclusion of technical error and vasospasm prompted sequential bypass to a second outflow artery if anatomically possible. The intraoperative hemodynamic studies served as a baseline for comparison with postoperative studies. Graft surveillance protocol. All patients underwent preoperative and serial postoperative noninvasive hemodynamic testing (intervals of 1 and 7 days, 6 weeks, and 3 months). In situ bypass hemodynamics were evaluated by a combination of noninvasive vascular laboratory methods: (1) measurement of resting limb arterial pressure by the transcutaneous Doppler ultrasonographic flow detection technique, (2) transcutaneous, continuouswave (CW) Doppler spectral analysis (5 MHz probe frequency, 45-degree Doppler angle to the skin) of graft blood flow patterns, and (3) duplex ultrasonography. The anlde-brachial systolic pressure index (ABI) was calculated for each limb by dividing the ankle systolic pressure by the higher of two simultaneously obtained brachial artery pressures. Duplex scanning and CW Doppler velocimetry were added to the postoperative surveillance protocol in 1983. Duplex scanning was used to map the entire bypass graft for flow abnormality before patient dis-
charge, and to locate the site of graft stenoses when a low graft flow state was identified by CW Doppler flow analysis (peak systolic flow velocity less than 45 cm/sec or decrease greater than 30 cm/sec compared to earlier outpatient evaluation). We have found that duplex examination of grafts that developed a low blood flow velocity can localize an occlusive lesion in most (85%) patients, and can categorize the severity of stenosis based on velocity spectra changes at and distal to the lesion, s'9 In general the development of graft stenosis can be predicted from changes in configuration and magnitude of the graft velocity waveform recorded by either a transcutaneous CW Doppler flowmeter or duplex scanner. Graft revision, if technically feasible, was recommended when either duplex scanning or angiography or both identified a greater than 50% diameterreducing lesion. Data analysis. Graft patency was determined by physical examination, decrease in ABI, and duplex scanning. An interval decrease in ABI greater than 0.2 was used as a criterion for arteriography or duplex scanning to assess graft patency. Primary and secondary graft patency rates were determined by the life-table method. Graft revisions, such as ligation of
Yolumr 9
Number2 1-~1.____
Monitoring functional patency of vein bypasses 289
1 0 ~ 0
(83) (75) (60) . . . . 100" __~__, • : -" -" . z' .tqll T ,~41) 7 " 3 4 ) 16~;~t'~'o'"'~":
80-
(144,
-
- P:
I
-J i(21)
;;at ;;;;t/;;;Ti;'3/'I
I
(13)
86% ,,o,o
r
PERCENT
60' PATENT
40'
20'
|
6
,
i
i
12
=
18
MONTHS
•
|
24 AFTER
,
,
30
,
n
36
,
,
42
,
=
48
SURGERY
Fig. 1. Secondary patency for 250 in situ saphenous vein bypasses calculated by life-table method. Closed circles indicate patency for 83 femoropopliteal bypasses, and open circles indicate patency of i63 femorotibial and four femoropopliteal (isolated segment) bypasses. Difference in patency was statistically significant by Wilcoxon test (p < 0.03). pressure-reducing arteriovenous (AV) fistulas, incision of residual valve leaflets, or PTA of in situ vein or adjacent native arterial stenoses, excluded the graft from primary patency. If graft patency was restored by thrombectomy, thrombolysis, or revision of a patent bypass, the graft was listed under secondary patency. Patency rates relative to the outflow artery (popliteal versus tibial) and the requirement for revision were compared for statistical difference by means of the generalized Wilcoxon test. Arterial pressure and blood flow velocity data are expressed as mean _+ SD. RESULTS Seven patients died within 30 days of operation, ,or an operative mortality rate of 3%. Five patients died of myocardial infarction, one patient died of hemorrhagic shock because of a displaced ligature, and one patient died at home of undetermined cause. The cumulative secondary patency data to 4 years after operation for the 83 femoropopliteal, 163 femorotibial, and four femoropopliteal (isolated segment) in situ bypasses are listed in Table III and presented graphically in Fig. 1. Grafts to isolated popliteal segments and sequential bypasses were included with the femorotibial graft results because of their similar hemodynamics. Secondary patency of femoropopliteal bypasses was higher (Wilcoxon test, p < 0.028) compared to bypasses to tibial arteries and isolated popliteal artery segments. This was due to a higher femoropopliteal graft patency during the first 2 postoperative years. Despite a steady decline
in primary patency to 63% and 61% at 4 )rears for femoropopliteal and femorotibial bypasses, respectively, secondary patency was in excess of 80% for both graft configurations. The cumulative survival of patients with patent grafts was 82% at 1 year, 77% at 2 years, 74% at 3 years, and 65% at 4 years. Seventy grafts were identified to have correctable anatomic lesions and underwent a total of 95 revisions. No lesion was identified in three grafts that had thrombectomy alone. Thirteen bypasses had multiple revisions with one graft revised four times over a period of 3 years. Correction of vein conduit or anastomotic abnormalities comprised 79% of revisions, whereas correction of atherosclerotic lesions identified in the iliac or infrapopliteal arteries accounted for 10% and 11% of revisions, respectively (Table IV). The mean time interval from primary operation to graft revision ranged from 8 months for lesions involving the in situ bypass (conduit and anastomoses), to 10 months for iliac artery stenosis, to 16 months for outflow tract disease progression. Most (18/21) graft lesions corrected within 30 days of operation (retained valve leaflet, residual AV fistula, anastomotic suture stenosis ) were identified and assessed for hemodynamic significance by duplex scanning. Correction of these lesions increased distal graft blood flow velocity and limb arterial pressure. Occlusive and aneurysmal lesions corrected more than 1 month after operation were identified by a combination of physical examination, noninvasive hemodynamic testing, and arteriography. The development of graft stenosis accounted for 64 (90%)
}'oarna/of 290
VA~CUL~
Bandyk et al.
Table IV. Site and morphologic characteristics of 92 lesions revised after in situ saphenous vein bypass grafting Site and morphologic characteristics of lesion
No.
Vein conduit Valve site stenosis Arteriovenous fisttda Retained valve leaflet Graft entrapment Graft torsion Graft aneurysm Unexplained thrombosis Total
20 9 5 4 2 1 6 47
Anastomotic sites Stricture of myointimal hyperplasia/atherosclerosis Suture stenosis Anastomotic pseudoaneurysm Total
19 6 1 26
Native vessel atherosclerotic stenosis Iliac artery Popliteal artery Tibial arteries Total
9 1 9 19
of 71 late graft revisions. Transformation of the graft velocity waveform from a triphasic to a monophasic or biphasic configuration coupled with a decrease in peak systolic blood flow velocity reliably predicted a remote occlusive lesion. Hemodynamic data were available for comparison from 56 in situ bypasses with diagnosed graft stenosis. Peak systolic velocity in the distal graft segment ranged from 12 to 50 cm/sec (mean 32 + 8 cm/sec), a level decreased from 75 + 16 cm/sec measured after the primary operation. The decrease in ABI ranged from 0.1 to 0.55 (mean 0.24 ___ 0.12). In 19 (34%) of the 56 limbs, resting ABi did not identify the presence of graft or native artery stenosis. Patients with stenosed but patent in situ bypasses showed no symptoms at the time of diagnosis. Graft revision rate (calculated by life-table method) was highest in the perioperative period (10% at 30 days), decreased to 7% per 6-month interval until 18 months after operation, and was 3% per year thereafter. Overall, 70 (28%) of the 250 bypasses were revised after the primary operation. Ten bypasses (10%) were thrombosed at the time of graft revision. Graft thrombectomy alone restored patency in only one graft. Three bypasses were salvaged with thrombectomy and revision of the outflow tract by sequential bypass grafting, anastomosis to another outflow artery, or vein-patch angioplasty. Six grafts were deemed failures and replaced by a polytetra-
Table V. Ninety,five secondary procedures performed on 70 in situ saphenous vein bypasses to restore normal functional patency Time t~fter primary operation
Operative procedure
<1 mo
Valve leaflet incision Thrombectomy alone Revision of anastomosis
5 3~ 2
Ligation of AV fistula Correction graft torsion/entrapment Vein-patch angioplasty Sequential vein bypass
6 2 5 1
Interposition vein bypass Autologous vein PTFE PTLA Total
>1 mo
3 4 28 15 5 5
__ 24
11
71
AV, arteriovenous fistula; PTLA, percutaneous transluminal balloon dilatation; PTFE, expanded polytetrafluoroethylene. nOne bypass patent after thrombectomy.
fluoroethylene conduit. Thrombolytic therapy restored patency in three grafts before revision for vein conduit stenosis (n = 2) or aneurysm (n = 1). A variety of secondary procedures were performed to restore bypass anatomy and functional patency to normal (Table V). Vein-patch angioplasty of a stenosed valve or anastomotic site was the most common secondary reconstructive procedure (33 of 70 procedures). Transluminal balloon dilatation of a vein conduit stenosis was successful in only two of four cases. Graft thrombosis occurred 1 week after dilatation of a high-grade stenosis loated in the dist~l~ segment of a femoral posterior-tibial bypass. Despit~e graft thrombectomy and vein-patch angioplasty, patency could not be restored. In contrast, all PTAs of native artery stenoses remained patent and improved graft and limb hemodynamics. Two secondary operative procedures performed on patent bypasses failed within 90 days. Late failure of direct surgical repairs involved repeat stenosis of vein-patched anastomotic sites (n = 4) or occlusion of cephalic vein sequential grafts (n = 3) performed for anastomotic myointimal hyperplasia. Revision of occlusive lesions increased mean peak systolic flow velocity in the distal graft from 33 cm/sec to 77 +_ 17 cm/sec (p < 0,0001, Mann-Whitney U test), and ABI increased to 0.93 --- 0.1. As shown in Fig. 2 and Tabl e VI, the secondary patency of 70 revised in situ bypasses up to 4 years
Volume 9 Number 2 February 1989
Monitoring functional patency of vein bypasses 291
Table YI. Life-laDle analyaia: Patcncy data of 280 in situ saphenous vein bypasses relative to requirement for graft revision after operation Interval
(too)
No. grafts not observed during period
No. grafts at risk
No. grafts failing
Death
180 169 155 130 106 90 67 55 38 22 18
4 3 2 4 1 1 1 0 0 0 1
3 2 4 0 1 0 0 2 0 0 3
Withdrawn +
Interval patency rate (%)
Cumulative patency rate (%)
6 4 15 6 6 5 3 3 6 1 1
1 7 8 14 9 17 8 14 10 3 2
98 98 99 97 99 99 98 100 100 100 --
98 96 95 92 91 89 86 86 86 86 --
1 0 2 2 1 0 1 1 0 2 0
0 0 3 5 4 5 4 9 3 4 2
96 97 94 100 98 100 100 94 100 100 --
96 93 88 88 86 86 86 81 81 8I --
No r~is~n 0-1 1-3 3-6 6-12 12-18 18-24 24-30 30-36 36-42 42-48 >48t
Singleor multipkgraft~viswn(s) 0-1 1-3 3-6 6-12 12-18 18-24 24-30 30-36 36-42 42-48 >48#
70 66 64 55 48 42 37 32 20 I7 11
+Reasons for withdrawal o f grafts were duration and loss to follow-up. j-Standard error of cumulative patency greater than >9%.
after operation was similar to the primary patency of 180 bypasses not identified to have a correctable abnormality, 81% versus 86%, respectively. Of the 32 in situ bypass failures recorded during patient follow-up, seven occurred within 30 days of operation. Three early graft failures occurred despite repeated thrombectomies and operative revision. ,¢iechanisms of early graft failure included low flow in two nonrevised grafts, associated hypercoagulable state (n = 2), or trauma caused by valve incision with formation platelet aggregation and thrombus (n = 3). Late (greater than 1 month) failure occurred in 12 revised and 13 nonrevised in sire bypasses. Hemodynamic studies confirmed the most common mechanism of late failure was the development of a low-flow state not amenable to revision (Table VII). Serial evaluations of these grafts showed a decrease in blood flow velocity compared to initial postoperative levels but retention of a triphasic veiocity waveform configuration signifying increased outflow resistance (Fig. 3). Angiographic evaluation showed a diseased outflow tract with atherosclerotic disease progression compared with pre- or intraoperative angiograms (Fig. 4). Peak systolic velocity
Table VII. Mechanisms involved in the late (> 1 mo) failure of 25 in situ saphenous vein bypasses Mechanism offailure
Low-flow state Prior graft revision N o revision
Missed graft stenosis Incomplete graft surveillance + Thromboembolism PTA failure
Infection#
No. 8 3 3 6 1 1 3
+Four grafts failed before 1983. tAll grafts patent at time of amputation for leg/foot infection.
in these grafts was typically less than 35 cm/scc (range 18 t o 40 cm/sec), and graft thrombosis occurred within 3 to 9 months of identifying the lowflow state. In three patients with unexpected graft failure, review of vascular laboratory studies indicated a graft stenosis was missed because of erroneous interpretation. Six of the 25 late graft failures occurred in patients not observed at regular intervals
Journalof VA~ULAK SURGERY
292 Bandyk et al.
100'
[..~o) (1~5)I(130)T ,'~.i"~'"':'-.-
0o.
:
(8,) •
""'
(90)
_
T
-
-, (so
T
r(5 ) I
i
I
i
I 86% 81%
PERCENT 60" PATENT 40"
20
|
|
6
|
i
|
|
i
,
|
i
|
|
12 18 24 30 36 MONTHS AFTER SURGERY
,
|
42
|
|
48
Fig. 2. Life-table analysis. Cumulative patency for 70 revised (open circles) and 180 nonrevised (closed circles) in situ saphenous vein bypasses. Number at each time interval after operation indicates number of grafts at risk. Difference in patency not statisticallysignificant by Wilcoxon test (p = 0.1). in the vascular laboratory or before 1983 when graft surveillance was by physical examination and limb pressure measurements. During the last 3 years of the study, unexpected failure of grafts judged to have normal hemodynamics occurred in only one of the 100 to 120 patients routinely studied by vascular laboratory personnel. This patient had previously undergone a vein-patch angioplasty of a valve site and probably experienced a repeat stenosis at this location. DISCUSSION Routine hemodynamic surveillance of in situ bypasses confirmed the most common mechanism of graft failure was the development of low-flow state (low blood flow velocity), which increased the likelihood of unexpected graft thrombosis. Occlusive lesions in the vein conduit, anastomotic sites, or inflow and outflow vessels tended to develop without symptoms, but bypass graft hemodynamics were altered sufficiently that detection before graft thrombosis was possible. Other conditions such as hypercoagulable states (antithrombin III, protein C and S deficiencies, heparin-induced platelet aggregation, abnormal plasminogen), graft or anastomotic aneurysm, or thromboembolism can also result in graft thrombosis. These uncommon mechanisms of graft failure should be suspected when graft thrombosis occurs unexpectedly and earlier (within three months) graft surveillance showed normal hemodynamics. In situ bypass surveillance was best per-
formed with direct physiologic testing methods, such as CW Doppler velocity waveform analysis or duplex scanning. Doppler-derived ankle systolic pressure measurements were not predictive of graft flow and were less sensitive than duplex scanning in the detection of graft stenosis. The importance of routine vascular laboratory follow-up after femoropopliteal vein bypass grafting has been emphasized in multiple reports. 1,3-s,1° Less than one fourth of the patients will have unequivocal evidence of graft stenosis based on decreased pedal/graft pressure pulse wave or recurrence of limb ischemia.1 The absence of symptoms in our patients who had graft stenosis was not surprising since lesions were identified early, before gra~ thrombosis, and ABI had decreased only moderateiy from initial postoperative levels to a mean of 0.62. Although the development of a low-flow state in the in situ vein conduit predicted the presence of an occlusive lesion, graft revision by direct surgical intervention or PTA was not always possible. Atherosclerotic disease progression involving the outflow arteries, typically immediately adjacent to anastomotic sites, and failure of secondary graft procedures were the most common causes of late graft failure. One fourth (6/25) of the late graft failures had outflow tracts judged not amenable to sequential bypass grafting or PTA for the purpose of increasing graft blood flow velocity. Serial hemodynamic testing predicted failure in these bypasses with an acquired lowflow state. Blood flow velocity in the graft segment of smallest diameter was typically less than 35~
Volume 9 Number 2 February 1989
Monitoring functional patency of vdn bypasses 293
POSTOPERAT|VE TiME iNTERVALI
m
C
Fig. 3. Development of low graft blood flow velocity after operation. Serial velocity spectra recorded from disal segment of popliteal-posterior tibial in situ saphenous vein graft with a duplex scanner. Hyperemic flow pattern (forward flow throughout the pulse cycle) present during first week after operation evolved to triphasic configuration at 3 months. Decrease in peak systolic flow velocity noted 14 months after operation was due to progression of disease and occlusion of retrograde flow in the posterior tibial artery (see Fig. 4). rm/scc, and the presence of a triphasic or monophasic velocity waveform configuration (absent diastolic forward flow) signified high outflow resistance. The thrombotic threshold velocity, n a minimum blood flow velocity below which thrombus formation occurs, of in situ bypasses is variable and depends on the time interval after operation, thrombogenicity of the graft flow surface, coagulation status of the patient, and configuration of the velocity waveform. Although the detection of low blood flow velocity in the graft was helpful in the recognition of technical errors and graft stenosis, no absolute minimum level has been identified that predicts graft failure. 12A peak systolic flow velocity less than 45 cm/sec was measured in 96% of grafts with correctable lesions. Graft blood flow vdocity in this range should prompt an angiographic or duplex examination to identify a cor:ectable lesion. When a low-flow state in a graft was
Fig. 4. Angiograms of popliteal-posterior tibial in situ saphenous vein bypass with diseased outflow artery. Arrow indicates site of distal anastomosis. found not to be amenable to revision, prolonged patency of the bypass was uncommon. Typically the grafts remained patent for 3 to 9 months. Thc efficacy of using oral anticoagulant therapy for the purpose of prolonging patency in these grafts with diseased outflow tracts is unknown. A rccent report involving femoropopliteal polytetrafluoroethylene grafts, a bypass with a lower than average flow rate, suggested the incidence of unexpected thrombosis can be decreased by long-term sodium warfarin (Coumadin) anticoagulation. 13 An important observation of this study was the similar 4-year cumulative patency rates for nonreviscd (86%) and revised (81%) in situ bypasses. Revision of stenotic lesions returned functional graft patency to normal levels; graft blood flow velocity increased, the velocity waveform reverted to a triphasic configuration, and limb arterial pressure returned to normal. The elective revision of patent bypasses, even before the onset of ischemic symptoms,
294
Journal of VASCULAR SURGERY
Bandyk et al.
minimizes the likelihood that thrombectomy will be a necessary part of a graft salvage operation. Our results of secondary operations on patent grafts were close to the 5-year patency of 90% reported by Whittemore et al.2 for reversed saphenous vein grafts repaired before thrombosis. We believe the surveillance plan and diagnostic criteria used for in situ bypass follow-up is also applicable for monitoring functional patency of reversed saphenous vein grafts. A correctable lesion was identified in most grafts with abnormal hemodynarnics. Repair of vein conduit or anastomotic lesions comprised 79% of the secondary reconstructive or restorative procedures performed. Vein-patch angioplasty of stenosis was the most common reconstructive procedure, and repeat stenosis of this repair was uncommon. Interposition grafting was used to correct long strictures (>4 cm) of the venous conduit or for recurrent graft stenosis. Use of large-diameter (>6 ram) cephalic veins was a sequential bypass was associated with low blood flow velocity (range 20 to 45 cm/sec) in this conduit after operation, a level that may have contributed to the failure of this vein when used in secondary operations. Percutaneous transluminal balloon dilatation of native artery stenoses was a durable restorative procedure, but treatment of vein graft stenoses was not uniformly successful, particularly when performed within 6 months of the primary operation. The development of graft stenosis within the first year after operation was frequently associated with minor technical imperfections in Gypass construction with valve and anastomotic sites having residual flow abnormality. Based on the time course of graft revision and failure, hemodynamic surveillance should be intensive in the immediate postoperative period and continue at frequent intervals for 2 years, at which time the incidence of graft stenosis decreases. The ability of duplex scanning to localize an anatomic abnormality in 85% of patients with suspected graft stenosis, coupled with its expense and comfort advantages compared to angiography, make it the preferred method for postoperative surveillance. Lesions judged to reduce diameter by greater than 50% should be considered for revision. Despite intensive anatomic and hemodynamic monitoring of in situ bypasses, graft failure still occurs. Graft thromboembolism, misinterpretation of vascular laboratory studies, and lack of patient compliance with the outpatient appointment schedule or recommendation to undergo angiography and graft revision contribute to a small but irreducible incidence of graft failure. Failure of graft revision and identification of lesions not amenable to surgical intervention remain unsolved problems contributing
to late vein graft failure. Improvement in these areas will require a better understanding o f the pathologic nature of artery wall healing, atherosclerotic disease progression, and myointimal hyperplasia. Patient assessment at 3-month intervals after discharge from the hospital appears adequate to detect the occasional malignant progression ofmyointimal lesions at valve and anastomotic sites. The graft surveillance protocol can be implemented by vascular laboratory personnel, and only when abnormalities are identified is a physician-directed evaluation required. This approach does require the vascular technologist to have a thorough knowledge of bypass graft anatomy in each patient, and an understanding of graft and limb hemodynamics after arterial bypass grafting. The goal of noninvasive hemodynamic testing is to provide the physician with reliable guidelines to aid decision making regarding technical adequacy and functional patency of the arterial bypass. Duplex scanning w~,~ blood flow velocity calculation has the capability of providing the necessary anatomic and physiologic data to achieve this goal. REFERENCES
1. Berkowitz HD, Hobbs CL, Roberts B, et al. Value of routine vascular laboratory studies to identify vein graft stenosis. Surgery 1981;90:971-9. 2. Whittemore AD, Clowes AW, Couch NP, et ai. Secondary femoropopliteal reconstruction. Ann Surg 1981;193:35-42. 3. Sladen JD, Gilmour JL. Vein graft stenosis. Characteristics and effect of treatrnent. Am J Surg 1981;141:549-53. 4. Cohen JR, Mannick IA, Couch NP, Whittemore AD. Recognition and management of impending vein graft failure. Arch Surg 1986;121:758-9. 5. Bandyk DF, Kaebnick HW, Stewart GW, Towne JB. Durability of in situ saphenous vein bypass: a comparison of primary and secondary patency. J VAsc SV•G 1987;2: 256-68. 6. Levine AW, Bandyk DF, Bonier PH, Towne JB. Less~,@ learned in adopting the in situ saphenous vein bypass. J VAsc SURG 1985;1:145-53. 7. Bandyk DF, Jorgensen RA, Towne JB. Intraoperative assessment of in situ saphenous vein arterial grafts using pulsed Doppler special analysis.Arch Surg 1986;121:292-9. 8. Bandyk DF, Kaebnick HW, Bergamini TM, et al. Hemodynamics of in situ saphenous vein arterial bypass. Arch Surg 1988; 123:477-82. 9. BandykDF. Monitoring functional patency of vasculargrafts. Sere Vasc Surg 1988;1:40-50. 10. Turnipseed WD, Acher CW. Postoperative surveillance: an effective means of detecting correctable lesions that threaten graft patency. Arch Surg 1985;120:324-8. 11. Sauvage LR, Walker MW, Berger KG, et al. Current arterial prosthesis. Arch Surg 1979;114:687-91. 12. BandykDF, Cato RF, Towne JB. A low blood flow velocity predicts failure of femoropopliteal and femorotibial bypass grafts. Surgery 1985;98:799-809. 13. Ffinn WR, Kohrer MJ, Yao JST, et al. Improved longterm patency of inftagenicular polytetrafluoroethylenegrafts.' J VASCSURG1988;7:685-90.
Volume 9 Number 2 February 1989
DISCUSSION
Dr. R o b e r t P. Leather (Albany, N.Y.). I compliment the presenter, Dr. Schmitt, on an elegant presentation and also compliment his fellow authors on a thorough analysis of a very careful and complete postoperative surveillance program on their patients in whom they have done in situ bypasses. There is a wealth o f information in the manuscript, and I highly recommend it for review when it is published. Our experience with virtually identical postoperative surveillance protocol in a similar patient population is basically the same, with two notable exceptions. One exception is the incidence o f revisions in this series, 70 out of 250, or 28% o f the conduits, versus 123 conduits threatened out o f 1066 in our series, for a gross incidence o f 11.5%. The other exception is that the revised bypass in our experience does not enjoy the same patency rate as the in situ conduit that did not require revision. : : There is a remarkably similar distribution o f distal anastomoses to tibial arteries for every popliteal artery. However, there is a difference in the distribution of tibial vessels used. The most common, 45% o f tibial bypasses , go to the peroneal artery, and in our experience the posterior tibial artery is the least frequently used. As has already been pointed out by many authors, the highest frequency o f revision occurs in the first two 6month periods and then drops to a less than 1% incidence by 2 years. Eighty percent o f the revisions are in the conduit itself, but the frequency is related to size. When size alone is observed, those grafts o f 3 mm or less have an incidence o f stenosis development that is double that of veins that are 4 mm or greater. Therefore, I ask the authors if they have any data on the distribution o f these stenoses, specifically the distal or proximal mobilized segment as compared to the undisturbed truly in situ portion. Also, I would like to know if they find any correlation between bypass diameter and the frequency o f these stenoses. Another correlation with size is that the development of a proximal stenosis averaged 17 months, and as one went further down the vein, the average development of a stenosis in the distal mobilized segment was 6 months. The other exception is that if one takes all those bypasses that have no revisions, the primary patency at 5 years is 79%. However, when one analyzes the revised bypasses, excluding the residual fistulas that are not bypass threatening, the patency at 5 years is 61%. This difference is statistically significant. The importance o f a proper surveillance program is very clearly shown when one uses the now conventional primary and secondary patency, as recommended by Dr. Rutherford's committee, with a patency improvement of approximately 15% by the application o f a proper surveillance program. We have not found the peak systolic velocity to be an accurate cutoff indicator because of the wide variation in diameters of bypass conduits. What we have found to correlate best with continued patency is the reactive hyperemic
Monitoring functional patency of vein bypasses
295
flow, that is, a reactive hyperemia i n d u c e d by a cuff compression for a period of 3 minutes. If this flow at least doubles then it is highly unlikely that a bypass-threatening lesion is developing. I thank the Society for the privilege and the invitation to discuss this paper. Dr. A n t h o n y D. W h i t t e m o r e (Boston, Mass.). As Dr. Leather mentioned, the manuscript presents a wealth of information that evokes a series o f questions far too numerous for a limited discussion. However, this series of in situ grafts, which were prospectively evaluated with duplex imaging, is one of the largest that I am aware of, and serves to underscore the importance o f identifying vein graft stenosis before actual thrombosis. We have been aware of this necessity for some time and have, for the most part, relied on recurrent ischemic symptoms or a significant decrease in ankle pressures to monitor graft patency. This present study further documents the fact that recurrent symptoms or changes in ankle pressures cannot be relied on to identify significantly stenotic lesions in nearly one third o f such patients. I have three questions. Prevention o f graft thrombosis is o f critical importance, and monitoring with a duplex scanner seems effective to this end. I ask Dr. Schmitt whether we should strictly adhere to his protocol with regard to the frequency of follow-up examinations. I ask this in light of the fact that whereas noninvasive vascular evaluation is increasingly reliable and applicable to more and more clinical situations, the equipment involved is of unconscionable expense, approaching $200,000 per unit, and many vascular laboratories are significantly overburdened with the multiplicity o f studies now required. Second, since your revision rate is higher when compared with our experience as well as Dr. Leather's, is it possible that the imaging techniques may, in fact, be too sensitive? Did you feel that you perhaps revised some grafts that might not have required it? Finally, whereas individual criteria continue to vary, we have begun to distinguish more effectively between primary and secondary graft patency rates. Since the secondary patency rates achieved with stenotic but patent grafts are dramatically discrepant from those applicable to thrombosed grafts, do you think it appropriate to fury_her subdivide the secondary patency curves? The revision of a thrombosed graft differs markedly from the management of a stenotic but patent graft, particularly with regard to the use of thrombolytic therapy. Both situations have, at present, very different anticipated outcomes, and it seems inappropriate to combine the two situations into one statistic. Dr. John Wolfe (London, England). As you know~ we have also had a considerable interest in this area, and one of the criticisms leveled against us is that a surveillance program o f this type is totally impractical in a busy clinical practice. Therefore we have tried to define our indications for doing various tests. We found that duplex-measured graft flow is extremely
296
Bandyk et al.
unreliable. We have also found that an isolated measurement of graft velocity is not specific or sensitive, and I would like to ask you whether you have looked at ankle brachial indices alone in these patients with very low graft velocities or asked them carefully about their symptoms. H o w many of these patients would have been picked up as having grafts at risk without the use of duplex? We have concentrated on the nonhemodynamically significant stenoses because we think that the other can be picked up by simple means. By measuring the graft velocity at one point and at an adjacent point 2 cm away, even when the graft flow remains high, we can pick up a nonhemodynamically significant stenosis and then watch the progression o f that stenosis. This ratio is unaltered by cardiac output or peripheral resistance and therefore is reproducible from one to the next. Therefore our policy has been to look at these grafts very carefully over the first 6 months and again at a year, and we have found that have all graft stenoses been picked up within 6 to 9 months of the graft being inserted. H o w many of your stenoses developed after that period? My final point is on semantics. Dr. Rutherford went to great efforts to introduce a term that we could all use, critical ischemia. This term was originally developed in the British Journal of Surgery by Crawford Jamieson's ad hoc committee. This term has already been abused. Limb salvage has become a worthless term because people are using the term limb salvage for legs that would remain viable without intervention. I suggest to you that in any group of patients there must be some patients who do not fit the term critical ischemia despite the fact that they have rest pain (i.e., either gangrenous or an ankle pressure of less than 40 mm Hg). There should be three groups in any study o f this type. There should be the patients with claudication, the patients with rest pain, and the patients with critical ischemia. Dr. Schmitt (dosing). I thank the discussants for their comments. In particular, I would like to thank Dr. Leather
Journal of VASCULAR SURGERY
and the Albany group for their contributions to the research and development of the in situ saphenous vein bypass technique. Dr. Leather asked us about our patency rates. For femoropopliteal bypasses our primary and secondary patencies are 63% and 86%, respectively, and for femorotibial bypasses at 4 years they are 61% and 84%, respectively. We have not established the critical thrombotic velocity as an absolute value, but the concept does have importance in identifying a threatened or failing graft. Once a velocity of 40 to 45 cm/sec is obtained, the grafts should be studied to identify a correctable cause for the decreasing velocity. The graft will not necessarily thrombose in the immediate future, but we have found these grafts usually fail within 3 to 9 months. We have no experience at this time with reactive hyperemia as a test to predict graft survival. Dr. Whittemore asked about our surveillance protocol. It consists of assessing the graft by (1) transcutaneous CW Doppler ultrasonic flow detection to obtain resting limb arterial pressures and (2) CW Doppler spectral analysis of graft blood flow patterns. Those patients who have a lo,~a flow state or a significant change from an earlier examination then undergo duplex scanning of the entire graft. Our revision rate may be high compared to some other studies, but we revise a graft, including ligation of an AV fistula, only if it is a hemodynamically significant lesion resulting in a low-flow state or a lesion that produces a 50% reduction in diameter. Dr. Wolfe asked about the reproducibility of graft velocity. We have found that by consistently measuring the velocities in the area of the graft of the smallest diameter, accurate, reproducible results can be obtained, identifying the development of a low-flow state. Patients we define as having critical ischemia are those that have rest pain on a nightly basis, nonhealing ulcers on their feet, or those who have gangrene.