Comparison of angioscopy and angiography for monitoring infrainguinal bypass vein grafts: Results of a prospective randomized trial☆

Article Outline

• Abstract
• Patients and methods
• Results
• Discussion
• Discussion
• References
• Copyright

Abstract 

Purpose: This study was designed to determine whether, in primary infrainguinal bypass grafts in which only saphenous vein is used as the graft conduit, routine monitoring with intraoperative angioscopy can improve early graft patency as compared with standard monitoring with intraoperative completion angiography; and to delineate the advantages and disadvantages of these two modalities and their respective roles for the routine monitoring of the infrainguinal bypass graft. Methods: A total of 293 patients undergoing primary saphenous vein infrainguinal bypass grafting were prospectively randomized and monitored with either completion angioscopy or completion angiography. Clinical parameters, indications for operation, graft anatomy, and configuration were evenly matched in both groups. Forty-three bypasses were excluded from the study after randomization, including 12 veins randomized to angiogram, deemed inferior, and prepared with angioscopy. Results: In the 250 bypass grafts (angioscopy 128, angiography 122) there were 39 interventions (conduit, 29; anastomosis, 8; distal artery, 2), 32 with angioscopy and 7 with angiography (p < 0.0001). Twelve (4.8%) of the 250 grafts failed in less than 30 days, four (3.1%) of 128 in the angioscopy group and eight (6.6%) of 122 in the angiography group (p = 0.11 by one-sided hypothesis test). Conclusion: Although no statistical improvement in the proportions of failures in primary saphenous vein bypass grafts routinely monitored with completion angioscopy rather than the standard completion angiogram was demonstrated, the study delineates a trend that favors completion angioscopy for routine vein graft monitoring and demonstrates the advantages of angioscopy in preparing the optimal vein conduit. (J VASC SURG 1993;17:382-98.)

The patency of autogenous vein bypass grafts in the lower extremity has continued to improve despite their more distal extension to the infrageniculate vessels and even into the foot.¹, ², ³, ⁴, ⁵ The highest failure rate continues to occur within the first 30 postoperative days. Most of these failures may be attributed to errors in technique, poor quality of the vein conduit, insufficient distal runoff, or, rarely, unsuspected coagulopathies. Direct and indirect techniques to detect and avoid such problems continue to be the subject of intense investigation.⁶, ⁷, ⁸ Once the bypass graft is established, the failure rate is lower and is more dependent on the nature of the graft and the biologic healing process of the vein and vein graft-artery interface.

Since its introduction during the past two decades, routine completion intraoperative angiography has been advocated as the method of choice for monitoring infrainguinal bypass operations.³, ⁵, ⁹, ¹⁰, ¹¹ Although most reported studies of completion angiography during bypass operations are retrospective or anecdotal, the angiogram is still generally regarded as the gold standard for monitoring these bypass grafts.

In previous retrospective studies we and others have shown that intraoperative angioscopy is a safe and effective alternative or complementary monitoring technique during lower extremity revascularization procedures, with many advantages over the traditional intraoperative angiogram.¹², ¹³, ¹⁴, ¹⁵ Not only does it allow the detection of technical errors, but it also allows their immediate localization and correction. It is particularly useful in the preparation of the autogenous vein conduit, for which it accurately detects unsuspected intraluminal pathologic conditions that might cause the graft to fail,¹⁶, ¹⁷ it allows accurate and complete valvulotomy in the in situ and nonreversed vein grafts,¹⁸, ¹⁹, ²⁰ and it assists in the selection of optimal-quality vein segments when a composite autogenous vein graft is necessary.²¹ However, successful application of angioscopy requires an investment in new instrumentation and a commitment to overcome the initial learning curve necessary to the acquisition of new technical and interpretive skills.

The purpose of this study was to determine in a prospective and randomized fashion whether, in primary infrainguinal bypass grafts in which only saphenous vein was used as the graft conduit, routine monitoring with intraoperative angioscopy can improve the early graft patency as compared with standard monitoring with intraoperative completion angiography and to delineate the advantages and disadvantages of these two modalities and their respective roles for the routine monitoring of the infrainguinal bypass graft.

Back to Article Outline

Patients and methods 

Between February 1, 1990, and October 30, 1991, 293 patients undergoing primary infrainguinal bypass grafting with the greater saphenous vein from either limb were prospectively randomized into two groups, in which either intraoperative angiography or angioscopy was used to monitor the bypass procedure at its completion. Randomization was performed with the use of a table of random numbers provided by the Rand Corporation (Freepress Publishers, Glencoe, Ill.).

All patients undergoing repeat bypass operation, graft revision, or bypass with conduit other than greater saphenous vein and those with a serum creatinine level greater than 2.0 mg/dl or a history of contrast allergy were excluded from participation in the randomization process for the study. Once randomized and enrolled in the study, exclusion from the study was possible during operation if adequate saphenous vein was unavailable for the bypass procedure and conduit other than saphenous vein was used for the bypass graft; if the operating surgeon decided that no completion monitoring study should be performed or that a completion study other than the study to which the patient had been randomized was clinically indicated; or, finally, if either angioscopy or angiography was logistically unavailable and could not be performed. The study was approved by the New England Deaconess Hospital institutional review board on human studies, and informed consent was obtained from all patients enrolled in the study.

All data relevant to this study were collected according to a fixed protocol and data form, and the data were entered into a dedicated data base. Clinical details of the two groups were recorded, including incidence of preoperative cardiac events, preoperative evaluation of the severity of each patient's underlying illnesses as defined by the American Society of Anesthesiologists (ASA) classification of physical status, details of the operative procedures including operating room time, method of anesthesia, and all fluids administered during operation, including those used for angiography and angioscopy. Intraoperative angiography was performed in standardized fashion through a small Mark's needle inserted via a tributary into the graft with 20 to 40 ml Renografin 60 (Squibb, New Brunswick, N.J.) contrast media with proximal occlusion and no delay in exposure. Roentgenograms were taken in a single plane and repeated as deemed necessary by the operating surgeon. Our technique for intraoperative angioscopy performed by a separate angioscopy team with the use of a foot-controlled, dedicated irrigation pump (Angiopump; Olympus Corp., Lake Success, N.Y.) has been previously described.²² The irrigation fluid was a balanced saline solution (Plasmalyte; Baxter Pharmaceutical, Hookset, N.H.) with no added heparin, papaverine, or other medications. The vein graft, the distal anastomosis, and the distal artery were routinely examined during these completion procedures. The proximal anastomosis or proximal artery was only visualized at the request of the surgeon or when an inflow problem was noted.

Interpretation of the angiography studies was made by the surgical team and interpretation of the angioscopy studies by both the surgical and angioscopy teams. The operating surgeon was responsible for all clinical decisions. Details of these decisions, including all abnormal findings from either completion study, were noted in the standard data form. Abnormal findings that were observed but not corrected were recorded as noninterventions. Any finding that resulted in a surgical manipulation was recorded as an intervention. Graft patency during the first 30 postoperative days was the primary endpoint for this study. Patency was defined as a palpable graft pulse or a distal artery pulse, a Doppler pressure ankle/brachial index of more than 0.15 greater than the preoperative level, or a patent graft on a duplex scan. The clinical state of the distal extremity was not used as a criterion for graft patency. Follow-up data were obtained by a review of the hospital records, discharge summaries, and office notes of the individual participating surgeons. Given the advantage of angioscopy over angiography in the visualization of conduit defects, a one-sided hypothesis test was used in the comparison of the proportions with graft failures during the first 30 postoperative days. A Fisher's exact test was used in the comparison of the angioscopy and angiography subgroups for the percents with graft failure during the first 30 postoperative days. Life-table statistical analysis was done with criteria recommended by the Ad Hoc Committee on Reporting Standards, Society of Vascular Surgery/North American Chapter, International Society of Cardiovascular Surgery.²³

Back to Article Outline

Results 

Of 293 patients prospectively randomized to monitoring of the completed infrainguinal bypass graft with either completion angioscopy or completion angiography, 250 were included in this study (Table I). No patient enrolled in the study underwent more than one primary infrainguinal bypass grafting procedure. Forty-three (14.7%) of 293 patients randomized preoperatively and enrolled were subsequently excluded from the study. Table II summarizes details of the 43 bypasses excluded from study 1, including the reason for their exclusion after initial prospective randomization. The clinical data of the patients in the two groups were similar to those of the 250 patients included in the study (Table I) as regards age, sex distribution, preoperative ASA grading, incidence of diabetes mellitus, prior myocardial infarction, indication for operation, and anatomy of bypass and confirms the lack of selection bias in the study.

Table I. Clinical details of 250 patients included in study

Table II. Details of the 43 bypasses excluded after initial randomization of the 293 patients

NRV, Nonreversed; ISV, in situ vein; A/K, above knee; B/K, below knee; RV, reversed.

Among the 19 excluded bypasses randomized to angioscopy (Table II), six were excluded because adequate saphenous vein was unavailable (in two grafts the final conduit was arm vein; in three grafts polytetrafluoroethylene [PTFE]; and in one graft a composite PTFE-saphenous vein graft). A completion angiogram was performed instead of angioscopy in five bypasses (lack of access for the angioscope in two, surgeon preference in one, unavailability of angioscope in one, and an unsuccessful completion angioscopic procedure in one). In the remaining eight exclusions, no completion study was performed (lack of access for the angioscope in five, surgeon preference in two, unavailability of angioscope in one). In 13 of 16 vein conduits (two cephalic arm veins and 14 saphenous veins), the angioscope was used in vein conduit preparation, including valvulotomy.

Among the 24 excluded bypasses randomized to completion angiography (Table II), eight were excluded because the vein conduit appeared to be of inferior quality (on the basis of clinical criteria of vein size [less than 2.5 mm external diameter], external sclerosis or stenosis, and angioscopic criteria with regions of thrombosis and recanalization seen as webs or bands within the lumen of the vein, sclerotic and stenotic segments, or adherent intraluminal thrombus) and the surgeon preferred to prepare the vein conduit with the aid of the angioscope; eight were excluded because of lack of adequate saphenous vein (arm vein was used in three, PTFE in three, and composite saphenous-arm vein in two); five because of surgical preference not to perform completion angiography (four reversed and one nonreversed vein configuration); two where angioscopic inspection was performed after “blind” valvulotomy (one for a valvulotomy (one for a valvulotome injury in an inferior vein and another where poor flow was noted after “blind” valvulotomy); and one where the completion angiography procedure was unsuccessful and completion angioscopy performed. Vein conduits of inferior quality (exclusion number labeled with an asterisk in Table II) were identified by the surgeon in 16 of 21 vein grafts. In 15 of 16 of these inferior-quality veins, angioscopy was used to prepare the conduit (saphenous vein, 10; arm vein, 3; composite saphenous vein, 2), including optimal graft segment selection and valvulotomy. The one inferior vein conduit not prepared with angioscopy was used in the reversed configuration (exclusion number 40, Table II). In the five vein grafts in which the angioscope was not used for conduit preparation, four veins were used in the reversed and one in the nonreversed configuration. Clinical details of the 250 eligible patients (128 in the angioscopy group and 122 in the arteriography group) are shown in Table I. The randomized groups were similar in age, male/female ratio, preoperative ASA clinical evaluation for anesthetic risk, incidence of diabetes mellitus, and history of prior myocardial infarction. The indications for operation, methods of anesthesia, and total fluid volume administered intraoperatively were similar in the two groups.

The overall mean operating time, from skin incision to skin closure, for the various bypass procedures was not significantly different in the two groups, except for the longer and more distal bypasses, which were mostly in the in situ configuration, in which the use of the angioscope shortened the mean operating time by 17 minutes for the femorotibial and 35 minutes for the femoropedal bypass grafts. In the shorter bypasses from the popliteal artery distally, most of the bypasses were in the reversed or nonreversed vein configuration (Table I).

Patient characteristics were comparable in the two completion study groups. Limb salvage was the indication for operation in 235 (94.0%) of the 250 bypass procedures and claudication in 15 (6.0%) of 250. Bypass grafts to the infrageniculate arteries distal to the popliteal artery accounted for 185 (74.0%) of the 250 bypass grafts performed. The anatomy of the various bypass grafts and their vein grafts configurations (reversed, n = 30 and 41; nonreversed, n = 46 and 35; and in situ, n = 46 and 44; composite segments of saphenous vein, n = 6 and 2; for completion angioscopy and completion angiography, respectively) was similar in both study groups (Table III).

Table III. The anatomy and vein graft configuration of the 250 infrainguinal bypass grafts and their early (less than 30-day) primary patency

RV, Reversed; NRV, nonreversed; ISV, in situ vein.

Clinically relevant findings during the completion monitoring procedures of the 250 bypasses and the subsequent clinical decisions for nonintervention or intervention are listed in Table IV. In 14 completion studies in the nonintervention group, 14 abnormal findings were noted, 13 with completion angioscopy and 1 with the completion angiogram. One of these 14 bypasses failed within 30 days.

Table IV. Relevant findings and clinical decisions during the completion monitoring procedures of 250 infrainguinal bypass grafts for nonintervention and intervention

A narrowed vein segment in the distal vein graft (Fig. 1, A) was the single abnormal angiographic finding.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1. 

  A, Completion angiogram shows narrowed segment (between arrows) of femoropopliteal (below-knee) nonreversed saphenous vein bypass graft. B, Completion angiogram shows filling defect in region of distal anastomosis of femorodorsalis pedis in situ saphenous vein bypass graft. C, Completion angiogram shows filling defect in anastomosis of femoroposterior tibial reversed saphenous vein bypass graft (see text and Table VI, patient 4).

Of the 13 abnormal angioscopic findings, five were related to the vein conduit and eight to the anastomosis. Multiple separate areas of localized valvulotome injury to the intimal surface were noted in three vein grafts, and a localized area of nonobstructing organized mural thrombus and a short segment of vein conduit with a calcified, thickened intimal surface were noted in one of two other vein conduits. One anastomosis to a plantar artery was noted to have an extremely small apex and outflow tract, but had no other obvious technical errors (Fig. 2, A).

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 2. 

  A, Anastomosis to plantar artery with extremely small apex and outflow tract, and small volume of fresh thrombus, with no obvious technical errors. B, Completely occluded anastomosis to dorsalis pedis artery. C, Revision and attempted correction of anastomosis shown in B resulted in distorted anastomosis that occluded in immediate postoperative period (see text and Table VI, patient 8).

No correction was performed inasmuch as there was no alternate anastomotic site possible in the distal runoff artery. The graft occluded 2 days after operation (Table V, patient 1). Small nonocclusive suture-line flaps were seen in four anastomoses and an irregular suture line and small subintimal hematoma on the floor of the anastomosis in two other anastomoses.

Table VI. Details of eight infrainguinal bypass grafts that failed within 30 days to distal arteries other than the plantar arteries (n = 239)

ISV, in situ vein; B/K, below knee; RV, reversed.

Table V. Details of the 11 bypass grafts to the plantar arteries

NRV, Nionreversed; ISV, in situ vein; RV, reversed; ADV, angioscopically directed valvulotomy; BVL, “blind” valvulotomy

In the 36 completion studies in the intervention group, 39 abnormal findings led to therapeutic interventions. In the completion angioscopy group there were 32 interventions and in the completion angiogram group only seven interventions (p < 0.0001, by Fisher's exact test). After intervention, three grafts failed, two in the angioscopy group and one in the angiography group.

Of the 32 abnormal findings after angioscopy, 28 were related to the vein conduit, 3 to the anastomosis, and 1 to the distal runoff artery. Nine residual competent valves were detected after “blind” valvulotomy for in situ vein preparation, and 13 unligated tributaries were detected and subsequently ligated. In four vein grafts, three recanalized segments and one sclerotic vein segment were detected. In one of the recanalized vein segments the “webs” seen on angioscopy were cut under angioscopic direction. In the remaining three vein grafts, the diseased segment was excised. In two vein grafts significant volumes of intraluminal debris, fresh thrombus, and fibrinous material, were detected and removed with the aid of a valvulotome and high-flow saline solution irrigation under angioscopic direction. One graft failed because of a technical problem at the site of the venovenostomy done after complete division of the vein conduit during valvulotomy (Table V, patient 3). Three anastomoses were explored and revised after technical errors were detected. One anastomosis showed a markedly irregular suture line, a second showed a flap across the apex of the anastomosis, and in a third the anastomosis was noted to be completely occluded (Fig. 2, B). Attempted correction of the occluded anastomosis resulted in a distorted, abnormal, slitlike anastomosis (Fig. 2, C). Further correction was not undertaken. This graft occluded in the immediate postoperative period (Table VI, patient 8). The only abnormal finding in a distal artery was a partially occluding intimal or fibrin flap just beyond the anastomosis, which was disrupted by the repeated passage of the angioscope and high-flow saline irrigation.

Of the seven angiographic findings that resulted in revision, five were related to the distal anastomosis, one to the distal artery, and one to the vein conduit. One anastomosis was occluded, the second stenotic, and the third incorrectly sited in a distal branch of the dorsalis pedis artery. The two technically inadequate anastomoses were explored and revised and the third resited into the proximal dorsalis pedis artery. Two other anastomoses were explored because of filling defects detected on the completion angiogram (Fig. 1, B and C). Neither exploration found a cause for the defect. In one of these anastomoses (Fig. 1, C), which was explored and the distal artery probed with a dilator, the repeat intraoperative angiogram was normal, but the graft failed 2 days after operation (Table VI, patient 4).

Early graft patency (less than 30 days) is shown in Table III and Fig. 3,A through C. Four of 128 grafts failed in the completion angioscopy group and eight of 122 grafts failed in the angiography group, for 30-day patency rates of 96.9% and 93.4%, respectively (p = 0.11, by one-sided hypothesis test).

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 3. 

  Life-table analysis to 1 month comparing proportions of primary graft failure. A, Comparison of angioscopy and angiography groups. Difference at 1 month is not statistically significant (p = 0.1). B, Comparison of 11 bypasses to plantar arteries and remaining 239 bypasses in study. At 1 month difference is statistically significant (p = 0.04). C, Comparison of angioscopy and angiography groups with 11 plantar arteries excluded. Difference at 1 month is statistically significant (p = 0.03).

On the basis of the anatomy of the bypass graft alone, no statistically significant improvement in the 30-day patency rate could be demonstrated for completion angioscopy over completion angiography. However, the 30-day patency rate of the 11 bypasses to the plantar arteries of 65% was significantly lower than the 30-day patency rate of 96.6% for all the other 239 bypasses: p = 0.04 (Table III and Fig. 3, B). Furthermore, none of the four failed bypass grafts to the plantar arteries were subsequently salvaged, whereas six (75.0%) of the remaining eight failed grafts were salvaged and were patent at 30 days (Table V, Table VI). The only two bypass grafts that could not be salvaged demonstrated poor runoff on the preoperative or intraoperative angiogram. The overall 30-day patency rates for each of the other subgroups were as follows: femoropopliteal (above-knee) bypass 100%, femoropopliteal (below-knee) bypass 97.6%, femorotibial bypass 91.7% (all failures occurred in the posterior tibial bypass grafts), femorodorsalis pedis 100%, popliteotibial 96.2%, and popliteodorsalis pedis 93.9%.

Four of the 12 early graft failures of the 250 bypass grafts occurred in the bypass grafts to the plantar arteries (Table V). Poor runoff, demonstrated on both the preoperative angiogram and the intraoperative completion studies, probably contributed to failure of three of the four grafts that failed within 30 days. Exclusion of the 11 bypasses to the plantar arteries from the study group analysis results in an overall 30-day patency rate for the 122 grafts monitored with completion angioscopy of 99.2% (one of 122 failed grafts) as compared with 94.0% (seven of 117 failed grafts) for the 117 grafts monitored with completion angiography. This difference in early graft patency reaches statistical significance (p = 0.01, Fisher's exact test).

Details of the 8 of 239 bypass grafts that failed within 30 days are shown in Table VI. Only one of the seven failed grafts in the completion angiogram group was examined by angioscopy at the time of reoperation and two separate residual competent valves were detected and subsequently cut under angioscopic direction. The probable causes of graft failure in the other bypass grafts, as listed in Table IV, were arrived at by a consensus of the operating team. Five of the seven were thought to be related to problems within the vein conduit, and in the remaining two grafts, runoff problems were thought to be the primary cause of failure. None of the conduit problems were detected on the completion angiogram at the time of operation by the surgeon and radiologist or on retrospective review of the original completion angiograms for this study by authors (A.M. and E.J.M.). In the completion angioscopy group, although the defect was detected, it was not adequately corrected (Fig. 2, C) and the graft failed. The graft was salvaged with reoperation. Graft thrombectomy was performed and the distal anastomosis resited from the dorsalis pedis artery to the distal peroneal artery.

Four (9.3%) of 43 bypasses in the exclusion group failed within 30 days (Table II). None of these failed bypasses were subsequently salvaged.

The median follow-up time for the entire study group was approximately 5 months. Almost all patients were followed up for at least 3 months. Only six (2%) of 293 patients, three in the completion angiography group, one in the completion angioscopy group, and two in the exclusion group, were lost to follow-up. In the angiography group, studies 111, 133, and 148 were lost to follow-up with patent grafts at 10, 12, and 18 days, respectively. In the angioscopy group, study 82 was lost to follow-up with a patent graft at 15 days. In the exclusion group, studies 102 and 169 were lost to follow-up with patent grafts at 21 and 12 days, respectively. Life-table analysis considering primary patency for the first month of the study is shown in Fig. 3, A through C, for all 250 prospectively randomized bypasses with comparison of completion angioscopy and completion angiography by the anatomy of the graft and for the subgroup of 239 bypasses with the 11 bypasses to the plantar arteries excluded.

Overall 30-day mortality was 6 (2.0%) of 293 for all bypasses initially enrolled, 5 (2.0%) of 250 in the study group, and 1 (2.9%) of 43 for the exclusions. All grafts were patent at time of death. There were three deaths in the angioscopy group: study 10, after a postoperative myocardial infarction on day 30; study 80, after a massive cerebrovascular accident on day 25; and study 144, after an acute ventricular arrhythmia on day 5. In the angiography group, studies 137 and 32 both died of ischemic heart disease and congestive heart failure on postoperative day 2 and 28, respectively. Study 260, in the exclusion group, died on postoperative day 4, after a postoperative myocardial infarction (Table II).

Back to Article Outline

Discussion 

The results of our prospectively randomized study of 250 primary saphenous vein infrainguinal bypass grafts do not demonstrate any statistically significant improvement in the early (< 30 days) patency of infrainguinal bypass grafts routinely monitored with completion angioscopy rather than the standard completion angiogram. Detailed analysis of the study delineated a trend that favors completion angioscopy for the routine monitoring of these primary autogenous vein bypass grafts and reveals a clinical bias introduced into the study of favoring angioscopy for vein preparation whenever the quality of the vein conduit was in question. This bias was clearly demonstrated in the increasing number of bypasses initially enrolled in the study, and subsequently excluded, as the trial progressed. Finally, in the modern practice of infrainguinal bypass surgery, in which the number of the more distal bypassed with almost exclusive use of autogenous vein, often in the situ and nonreversed configurations, is dramatically increasing,⁵, ²⁴, ²⁵ the results of our study clearly demonstrate the advantages of angioscopy in the preparation of the optimal vein conduit.

The most striking difference between the two monitoring modalities is illustrated in the total number of surgical interventions performed after completion monitoring studies in 250 randomized bypass grafts, 32 with completion angioscopy and 7 with completion angiography (p < 0.0001, Table IV). Most of these interventions were associated with vein conduit preparation, 28 with completion angioscopy and only one with completion angiography (p < 0.0001). The number of interventions in the distal anastomosis and runoff artery was not significantly different for either monitoring procedure, but unlike results for the completion angiogram, there were no false positive findings with completion angioscopy. In the nonintervention group, many subtle findings were noted on angioscopy. None of these findings was associated with an increased early failure rate. The interpretation of such findings with regard to graft patency in the long term remains to be evaluated. Most conduit interventions in the completion angioscopy group were related to the in situ and nonreversed vein techniques, for which the incidence of residual competent valve leaflets after “blind” valvulotomy was approximately 20%, little different from the incidence in our previous series.¹², ¹³, ¹⁹ In the angiography group, 40% (two of five) of abnormalities detected in the anastomosis and that resulted in intervention and revision of the anastomosis were false positives (Fig. 1, B and C), for an incidence of false positives for the 122 bypasses in this group of 1.6%. Angioscopic examination of the anastomosis and distal artery in 128 bypass grafts was accurate with no false positives, but in two anastomoses, although technical problems were recognized (Fig. 2, A through C) they were not corrected, with subsequent graft failure. No false-positive findings from angioscopic examination resulted in unnecessary interventions in the distal anastomosis or distal artery. This experience is similar to that of others who showed the angioscope to be both very specific and highly sensitive when compared with the intraoperative angiogram in the same bypass grafts, in blinded and prospectively randomized studies.²⁰, ²⁶ Since the completion angiogram has been routinely applied to infrainguinal bypass grafting during the past two decades, despite the increasing frequency of bypass grafts to the more distal arteries, there has been a progressive reduction in the number of technical errors detected during these studies, from as high as 25% in the early studies to between 2% and 6% in the more recent reports.³, ⁵, ⁹, ¹⁰, ¹¹, ²⁶, ²⁷ As suggested by Liebman et al.,¹¹ this is probably because of an overall improvement in surgical techniques with the use of magnification, improved lighting, and modern instrumentation and sutures.

Most of the errors detected with routine intraoperative angiography during infrainguinal bypass operation in these studies were related to technical problems at the anastomosis and runoff artery rather than to intrinsic defects of the conduit itself. Many of the studies, however, included synthetic or biologic conduits other than saphenous vein. In a previous series of 259 infrainguinal bypass grafts monitored with routine intraoperative angioscopy, 124 clinical or surgical “decisions” were made.¹³ Only 6% of these decisions were related to anastomotic or runoff artery errors. Most of the surgical decisions were pertinent to the vein conduit and especially to the unique demands of in situ and nonreversed vein grafts. Nevertheless, 12% of vein conduit decisions were findings of unsuspected intraluminal pathologic conditions, such as segments of recanalized vein and vin stenosis, or problems with tunneling and torsion of the vein conduit.

Preliminary data analysis early on in the study had suggested that, provided the ratios of early failure demonstrated between the two monitoring techniques were maintained, study cohorts of approximately 120 grafts in each group would be required to demonstrate an improvement of statistical significance of angioscopy over angiography. However, two factors were not appreciated: the high early failure rate of the 11 bypasses to the plantar arteries, 35% as compared with a mean of 3.4% for all the remaining 239 bypasses in the study group (Table III, Fig. 3, B), and the increasing number of exclusions as the study progressed. Review of the reasons for failure of the plantar bypasses (Table V) suggests problems of runoff as being the most important cause of failure, a problem that neither of the monitoring techniques could alter to improve outcome. Review of study numbers of the exclusions in Table II shows the number of exclusions increased almost fourfold in those patients enrolled in the latter half of the study. Presuming that the same ratios of early failures would continue as in this study, a sample size of approximately 600 bypasses, more than twice as large as our present study group, would be required to demonstrate a statistically significant difference between the two monitoring modalities. Such a study would require the initial randomization of more than 700 bypass grafts, a study too large and too long for any single institution.

Close examination of the 43 bypass grafts in the exclusion group is instructive. The overall early primary graft failure rate in the exclusion group was 9.3%, as opposed to 4.8% for the 250 grafts included in the study, and attests to the restricted nature of the study cohort, which consisted only of primary bypass grafts with autogenous saphenous vein, excluding all reoperations, use of conduit other than saphenous vein, and patients with creatinine levels higher than 2 mg/dl. Although completion studies were only performed in approximately two thirds of the 43 bypasses (12 of 19 in the angioscopy group and 15 of 24 in the angiography group), the angioscope was used in conduit preparation in 78% (29 of 37) of the bypasses in which the conduit was vein, including arm vein (13 of 16 cases in those originally randomized to angioscopy and 16 of 21 of those randomized to angiography).

In the exclusions originally randomized to completion angiography, 12 saphenous vein conduits were deemed inferior on surgical exposure or after preliminary preparation by the surgeon. These bypasses were excluded from the study to optimize their preparation, with angioscopically directed valvulotomy, choice of optimal quality vein segments for use in a composite graft, or both of these procedures. This safety net for studies randomized to angiography may have altered the proportions of early failures in the study group by eliminating the advantages of angioscopy over angiography, namely, complete valvulotomy with minimal intimal and vein wall injury¹⁹ and increased sensitivity in the detection of unsuspected endoluminal pathologic conditions, which if detected and either corrected or the vein segment excluded improves the quality of the vein conduit. We have shown that in the arm vein conduit used for infrainguinal bypasses there is an unusually high, approximately 70%, incidence of unsuspected intraluminal pathologic conditions. Detection and correction or exclusion of diseased vein segments and selection of optimal-quality segments of vein can upgrade the quality of the vein conduit with improved early arm vein graft patency.¹⁷, ²¹

Analysis of 239 bypasses, with exclusion of the 11 bypass grafts to the plantar arteries, shows only a single early failure in 122 bypass grafts in the completion angioscopy group as opposed to seven early failures among 117 bypass grafts monitored with completion angiography (Table III). The statistically significant difference (p = 0.01, Fisher's exact test) is even more impressive when cognizance is given to the fact that the recognized technical error (Fig. 2, B and C) that led to the only failure in the angioscopy group was irreparable. This emphasizes the importance of optimal conduit preparation even for primary bypass grafts with greater saphenous vein and demonstrates the distinct advantages and efficacy of completion angioscopy over the completion angiogram.

Criticisms other than those of the efficacy of angioscopy itself, leveled against the routine use of angioscopy in modern vascular surgery, have been many. They include unwarranted expense with un-wieldy modern electronic equipment, increased cost and operating room time, difficulty in acquiring sufficient clinical angioscopic skills or interpreting the endoscopic findings, dangers of injury to the delicate grafts or native vessels with the introduction of various endoscopic instrumentation, and dangers of fluid overload associated with the high flow rates and large volumes of irrigation fluids necessary to obtain consistently good studies in the unique high-risk patient population undergoing lower extremity re-vascularization. Systematic examination of each of these factors in this and previous studies has shown that with good understanding of the technology and techniques, persistence, and dedicated application, none of these predictions have materialized.¹², ¹³, ¹⁷, ¹⁹, ²⁸ Modern angioscopic and angioscopic systems are continually undergoing improvements that make their routine use both easier and safer. Endoluminal trauma from the procedure is extremely infrequent and the addition of the necessary irrigation fluid volumes does not increase the morbidity or mortality of these procedures.²⁹ The overall operating time is not increased with the routine use of completion angioscopy. In this study, in the longer and more distal bypasses in which the use of the in situ bypass was more common, the operating time was reduced, which made angioscopy a more cost-effective procedure. In addition, with recent routine introduction of the techniques of minimally invasive surgery into the operating room, most of the necessary electronic equipment is not unique to angioscopy and is available as standard operating room equipment.

Our study suggests that despite the small numbers of defects detected and corrected with the use of both angioscopy and angiography in primary bypass grafts, completion monitoring for lower extremity infrainguinal bypass is worthwhile, particularly in cases in which failure of bypass grafting in patients with a threatened limb may result in limb loss or repeated operations with increasing morbidity and medical costs. When the quality of the vein conduit is questionable or when vein other than the greater saphenous vein, such as arm vein, is the only available conduit, angioscopic preparation may be essential to ensure the best quality conduit possible. Finally, we consider angioscopy and angiography to be complementary procedures. Although our study shows clearly that angioscopy is more effective in autogenous vein preparation and more specific and sensitive in detecting defects in the anastomosis and distal artery, the completion angiogram still remains useful in delineating the runoff vasculature of the bypass graft when this has not been well visualized on the preoperative angiogram.

Back to Article Outline

Discussion 

Dr. Samuel S. Ahn (Los Angeles, Calif.). I congratulate the authors on this important study. They set out to determine in a prospective randomized fashion whether routine angioscopy during infrainguinal bypass procedures improves early graft patency and to delineate the advantages and disadvantages of angioscopy versus angiography. Various investigators, including ourselves, have demonstrated that angioscopy detects numerous abnormalities and that this often alters the operative management. But, so far, no one has shown that these angioscopic-based interventions improve the surgical results to the benefit of the patient. The authors conclude from their study that angioscopy is safe and cost-effective, is more effective than angiography in monitoring infrainguinal vein bypass grafts, and results in significantly improved early graft patency. Unfortunately, I must disagree.

There are five major flaws in the design, conduct, and data analysis of this study. First of all, the study design allowed 43 patients to be excluded from data analysis after previous study enrollment and randomization assignment. As you know, such exclusion after the fact is not valid inasmuch as it introduces investigator bias into the study; and, indeed, it did. Most of the exclusions were made because of the surgeon's preference not to follow the randomization assignment or because of the investigator's inability to complete a satisfactory angioscopic examination.

Second, the authors did not stratify the randomization for the different bypass procedures that were subsequently analyzed separately. This oversight led to the subgroup analysis that excluded the plantar artery group to erroneously show a statistically significant difference between the angioscopy and angiography groups at 30 days. In fact, when one stratifies the groups for the subgroup analysis, there is no difference between angioscopy and angiography in any of the subgroups except for the pedal artery group, which showed a trend favoring angiography, not angioscopy.

Third, it is erroneous to use a one-sided Fisher's exact test to compare patency results on the basis of the assumption that the angioscope visualizes conduit defects better. The issue here is not whether angioscopy visualizes better than angiography, but whether the patency result is better. A one-sided test in this study would assume ahead of time that the patency results of the angioscopy group are better than those of the other group, which clearly is not necessarily correct. In a prospective, randomized study, when one does not know which group will achieve a better result, a two-sided test is mandatory.

Fourth, after finding no statistically significant difference in patency at 3 months, the authors claimed that the angioscopy group had better patency at 1 month by separately analyzing the patency curve at the 1-month follow-up point. This represents post hoc analysis, which is statistically invalid. Separate analysis of the other follow-up points would clearly show no statistically significant difference, which would suggest that the statistical difference seen at the 1-month follow-up point occurred by random chance. Just for this reason, statisticians insist that we compare the entire patency curve and not just a point in the curve.

Fifth, the 1- to 3-month follow-up period is too short to adequately address the main purpose and conclusion of the study. To determine whether angioscopy improves early graft patency requires at least a 6-month and preferably a 2-year follow-up period for infrainguinal bypass procedures, which are often reported in terms of 3 to 5 years of follow-up. To conclude that angioscopy results in a significantly improved patency on the basis of only a 3-month follow-up is erroneous and misleading, inasmuch as we have no idea what the patency curves would do later. In fact, it is quite conceivable that the patency curves could cross and show that the angioscopy group ultimately does worse because of angioscopy-induced injury leading to intimal hyperplasia.

Despite these serious shortcomings, I believe the authors have brought to the Society some valuable information. Their vast experience has shown us that angioscopy can be performed safely and can be useful in preparation of the vein as a conduit. Their data suggest that angioscopy may be faster and easier than angiography for long femoropedal bypasses, but more difficult for the shorter and more delicate popliteopedal bypasses. Also, they have shown us that angioscopy, even in the hands of a dedicated angioscopy team, is currently less dependable than angiography by a factor of 10.

Finally, they have shown us that we need better criteria for interpreting the angioscopic findings. Despite the significantly greater incidence of intervention in the angioscopic group, the 3-month patencies for the angioscopy and angiography groups are similar. Clearly, the currently used criteria for intervention are too sensitive and could lead to unnecessary or meddlesome intervention.

In closing, I congratulate the authors for bringing to the Society their extensive angioscopy data base and for attempting to answer some fundamental questions regarding this new technology. However, this study is marred by multiple major flaws and the authors' conclusions must be seriously questioned. My analysis of the data leads to the following conclusions: (1) improved short-term patency was not achieved with the use of angioscopy; (2) angioscopy was performed safely and was useful in preparation of a vein conduit; (3) angioscopy facilitated the long femoropedal bypass but impeded the short popliteopedal bypass; (4) angioscopy was less dependable than angiography; and (5) angioscopic evaluation in its current form may be too sensitive.

I have three questions. First, what are your current recommendations for intervention when the angioscopic findings are abnormal? Second, were the saphenous vein exposures similar in the angioscopy and angiography groups? Third, were the wound complication rates for the two groups different?

John Mehigan has reduced the wound complication rate of his in situ bypass procedures from 20% to 5% by using angioscopy to help cut the valves and identify side branches through just the proximal and distal arterial incisions and to thus avoid long incisions. Perhaps the most appropriate indication for angioscopy is to reduce incision length and wound complications.

Dr. Arnold Miller. Let me first respond to your specific questions and then I will reply to the global criticisms you have made.

First of all, criteria for intervention with both monitoring modalities were clinical and based on past experience that has been well documented. Each of the findings that led to an intervention and the interventions themselves were detailed in the paper that was sent to you for review in advance of the meeting.

All saphenous veins were exposed in a long continuous incision. Surgical techniques were similar for all participating surgeons. The incidence of wound complications was not part of the study but in available in our patient data base.

From the tone of your five points in regard to the design and conduct of the study, it is apparent that you have failed to appreciate the essence of the study. To put the study in clinical perspective, our goal was to compare the efficacy of two monitoring techniques of bypass surgery in detecting technical errors that may cause graft failure within 30 days of operation. Our study was limited to primary bypass operations with use only of autogenous saphenous vein in which the early (<30-day) failure rate is extremely low (approximately 5%). In previous studies we had already shown that preparation of poor-quality vein with angioscopy significantly improves the early patency over the use of the completion angiogram. In this study we eliminated this clear-cut advantage of angioscopy.

Although our study does not demonstrate a statistically significant difference between the completion monitoring techniques in the early patency in this selected group of patients, it does show a clear-cut trend favoring the angioscope as the preferable intraoperative monitoring tool. We openly have shared the reasons for the inability to reach statistical significance in the patient group study. Our study shows the real-life clinical difficulties and pitfalls of conducting such a study.

The prospective randomization for the study resulted in well-balanced groups. However, adherence to the preoperative randomization was at the discretion of each of the participating surgeons. These colleagues made all clinical decisions in selecting the best management for the individual patient and in maintaining ethical responsibility to their patients. As clearly pointed out and emphasized in the paper, as the study progressed, all the surgeons preferred to prepare the vein conduit with the angioscope whenever they believed the quality of the vein was in question. This created a preference for the angioscope over the completion angiogram.

In the literature and again as demonstrated in this study, there is no statistical difference in the early patency of grafts proximal or distal to the popliteal artery at the knee, so there is little justification for such a stratification. The only grafts whose patency was vastly different were the 11 bypasses to the plantar arteries in which the patency rate was only 65% at 30 days. This reflected a poor runoff situation rather than a failure of either monitoring technique to detect technical errors. By chance, more failures in this group occurred in the angioscopy group. Without this group, the benefit of angioscopy would have been larger.

As regards the statistical analysis, the application of both the one- or two-sided Fisher's analysis is statistically legitimate. Angioscopy clearly increases the visualization of intraluminal defects as compared with angiography, the so-called gold standard, particularly visualization of the endoluminal status of the vein conduit. The hypothesis rested with the one-sided Fisher's test was whether this benefit of angioscopy could demonstrate a statistical improvement in the 30-day patency. It would not be reasonable to expect that this increased visualization could result in poorer graft survival at 30 days.

This brings me to the primary study endpoint, which was primary graft patency at 30 days, in which failure is generally accepted as relating to correctable technical deficits. Later failure is multifactorial and complex and clearly beyond the scope of this study.

Careful analysis of the data shows that most of the longer grafts were performed with the conduit in the in situ or nonreversed configuration. Use of the angioscope in the preparation and monitoring of these grafts actually shortened the mean operating time by approximately 30 minutes as compared with use of “blind” valvulotomy with a completion angiogram. In the shorter grafts, in which the reversed vein configuration was used more frequently, especially in the angiography group, the opposite held true.

I am uncertain where you draw your conclusion that angioscopy was less dependable in our hands than angiography for monitoring bypass operations. The most striking revelation of the study was the paucity of findings that required intervention after completion angiogram. There were only seven findings that led to interventions in the 122 bypasses monitored with completion angiography and two of those findings were false positives, which resulted in unnecessary explorations of the anastomosis. In contrast, in the angioscopy group, there were 32 findings in 128 bypass grafts that led to interventions, with no false positive interventions. Furthermore, in the angioscopy group, technical errors were detected and recognized in two anastomoses, but because of clinical circumstances (elaborated on in the paper) they were not corrected and caused graft failure. This, of course, was not factored into the statistical analysis. Our interpretation of the study shows angioscopy, in our hands, to be more sensitive and specific and thus more dependable than completion angiography in monitoring infrainguinal bypass grafts.

Finally, for the record, each of your criticisms was addressed in detail in the manuscript sent for your review. Your comments deflect from the more important study issues and conclusions. However, your comments do underscore the difficulties of conducting a controlled clinical trial of two interventions in which a clear preference develops for one of the interventions.

Dr. J. Dennis Baker (Los Angeles, Calif.). Today's presentation is trying to encourage us to add angioscopy to our routine for dealing with lower extremity reconstructions. Because we are being forced into contracted packages for our services, what is the cost per angioscopic examination that we are going to have to absorb as part of the cost of doing the operation?

Dr. Miller. Dr. Baker, that is an important question particularly with current concerns about health costs. Unfortunately, the comparative cost of these two monitoring modalities was not part of the study. However, from our results I have no hesitation in recommending the angioscope for the evaluation of the vein conduit whenever there is any concern as to vein quality. Reoperation in these patients is extremely costly both in dollars and associated morbidity.

Back to Article Outline

References 

1. DeLaurentis D, Friedman P. Arterial reconstruction about and below the knee: another look. Am J Surg. 1971;121:392–397
   • View In Article
   • MEDLINE
   • CrossRef
2. Veith F, Moss C, Daly V, et al.  New approaches in limb salvage by extended extraanatomic bypasses and prosthetic reconstructions to foot arteries. Surgery. 1978;84:764–774
   • View In Article
   • MEDLINE
3. Taylor L, Edwards J, Phinney E, Porter J. Reversed vein bypass to infrapoplitcal arteries. Ann Surg. 1987;205:90–97
   • View In Article
   • MEDLINE
   • CrossRef
4. Ascer E, Veith F, Gupta S. Bypasses to plantar arteries and other tibial branches: an extended approach to limb salvage. J Vasc Surg. 1988;8:434–441
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (668 KB)
   • CrossRef
5. Donaldson M, Mannick J, Whittemore A. Causes of primary graft failure after in situ saphenous vein bypass grafting. J Vasc Surg. 1992;15:113–120
   • View In Article
   • Abstract
   • Full Text
   • CrossRef
6. O'Donnell T, Cossman D, Callow A. Noninvasive intraoperative monitoring: a prospective study comparing Doppler systolic occlusion pressure and segmental plethysmography. Am J Surg. 1978;135:539–546
   • View In Article
   • MEDLINE
   • CrossRef
7. Bandyk D, Cato R, Towne J. A low flow velocity predicts failure of femoropopliteal and femorotibial bypass grafts. Surgery. 1985;4:799–809
   • View In Article
8. Ascer E, Veith F, White-Flores S, Morin L, Gupta S, Lesser M. Intraoperative outflow resistance as a predictor of late patency of femoropopliteal and infrapopliteal arterial bypasses. J Vasc Surg. 1987;5:820–827
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (667 KB)
   • CrossRef
9. Plecha F, Pories W. Intraoperative angiography in the immediate assessment of arterial reconstruction. Arch Surg. 1972;105:902–907
   • View In Article
   • MEDLINE
10. Dardik H, Ibrahim I, Koslow A, Dardik I. Evaluation of intraoperative arteriography as a routine for vascular recconstructions. Surg Gynccol Obstet. 1978;147:853–858
    • View In Article
11. Liebman P, Menzoian J, Mannick J, Lowney B, LoGerfo F. Intraoperative arteriography in femoropopliteal and femorotibial bypass grafts. Arch Surg. 1981;116:1019–1021
    • View In Article
    • MEDLINE
12. Miller A, Campbell D, Gibbons G, et al.  Routine intraoperative angioscopy in lower extremity revascularization. Arch Surg. 1989;124:604–608
    • View In Article
    • MEDLINE
13. Miller A, Stonebridge P, Jepsen S, et al.  Continued experience with intraoperative angioscopy for monitoring infrainguinal bypass grafting. Surgery. 1991;109:286–293
    • View In Article
    • MEDLINE
14. Grundfest W, Litvack F, Glick D, et al.  Intraoperative decisions based on angioscopy in peripheral vascular surgery. Circulation. 1985;78(suppl I):I-13–I-17
    • View In Article
15. Mehigan J, Olcott C. Videoangioscopy as an alternative to intraoperative arteriography. Am J Surg. 1986;152:139–145
    • View In Article
    • MEDLINE
    • CrossRef
16. Miller A, Jepsen S, Stonebridge P, et al.  New angioscopic findings in graft failure after infra-inguinal bypass grafting. Arch Surg. 1990;125:749–755
    • View In Article
    • MEDLINE
17. Stonebridge P, Miller A, Tsoukas A, et al.  Angioscopy of arm vein infrainguinal bypass grafts. Ann Vasc Surg. 1991;5:170–175
    • View In Article
    • Abstract
    • Full-Text PDF (863 KB)
    • CrossRef
18. Mehigan J. Angioscopic preparation of the in situ saphenous vein for arterial bypass: technical considerations. In:  White GH,  White RA editor. Angioscopy, vascular and coronary applications. Chicago: Year Book Medical; 1989;p. 72–75
    • View In Article
19. Miller A, Stonebridge P, Tsoukas A, et al.  Angioscopically directed valvulotomy: a new valvulotome and technique. J Vasc Surg. 1991;13:813–821
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (3057 KB)
    • CrossRef
20. Gilbertson J, Walsh D, Zwolak R, et al.  A blinded comparison of arteriography, angioscopy, and duplex scanning in the intraoperative evaluation of in situ saphenous vein bypass grafts. J Vasc Surg. 1992;15:121–129
    • View In Article
    • Abstract
    • Full Text
    • CrossRef
21. Marcaccio E, Miller A, Tannenbaum G, et al.  Can angioscopically directed intervention upgrade arm vein quality and improve early graft patency?. [Abstract] J Vasc Surg. 1992;16:480
    • View In Article
    • CrossRef
22. Miller A, Jepsen S. Angioscopy in arterial surgery. In:  Bergan JJ,  Yao JST editor. Techniques in arterial surgery. Philadelphia: WB Saunders; 1990;p. 409–416
    • View In Article
23. Rutherford R, Flannigan DP, Gupta S, et al.  Suggested standards for reports dealing with lower extremity ischemia. J Vasc Surg. 1986;4:80–94
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (1276 KB)
    • CrossRef
24. Leather R, Shah D, Chang B, Kaufman J. Resurrection of the in situ vein bypass: 1000 cases later. Ann Surg. 1988;208:435–442
    • View In Article
    • MEDLINE
    • CrossRef
25. Taylor L, Edwards J, Porter J. Present status of reversed vein bypass: five-year results of a modern series. J Vasc Surg. 1990;11:207–215
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (935 KB)
    • CrossRef
26. Baxter B, Rizzo R, Flinn W, et al.  A comparative study of intraoperative angioscopy and completion arteriography following femorodistal bypass. Arch Surg. 1990;125:997–1002
    • View In Article
    • MEDLINE
27. Stept L, Flinn W, McCarthy WI, Bartlett S, Bergan J, Yao JS. Technical defects as a cause of early graft failure after femorodistal bypass. Arch Surg. 1987;22:599–604
    • View In Article
28. Miller A, Lipson W, Isaacsohn J, Schoen F, Lees R. Intraoperative angioscopy: principles of irrigation and description of a new dedicated irrigation pump. Am Heart J. 1989;118:391–399
    • View In Article
    • MEDLINE
    • CrossRef
29. Kwolek C, Miller A, Stonebridge P, et al.  Safety of saline irrigation for angioscopy: results of a prospective randomized trial. Ann Vasc Surg. 1992;6:62–68
    • View In Article
    • Abstract
    • Full-Text PDF (704 KB)
    • CrossRef