Does a Completely Accomplished Duplex-based Surveillance Prevent Vein-graft Failure?

Transcription

1 Eur J Vasc Endovasc Surg 18, (1999) Article No. ejvs Does a Completely Accomplished Duplex-based Surveillance Prevent Vein-graft Failure? L. Ihlberg, M. Luther, A. Albäck, I. Kantonen and M. Lepäntalo Division of Vascular Surgery, Department of Surgery, Helsinki University Central Hospital, Helsinki, Finland Objectives: to assess the benefits of duplex-based vein-graft surveillance over clinical surveillance with distal pressure measurements. Design: prospective randomised comparative trial. Material and methods: three hundred and forty-four patients with 362 consecutive infrainguinal vein bypasses were prospectively randomised to a follow-up regime with or without duplex scanning (ABI group and DD group) at 1, 3, 6, 9, and 12 months postoperatively. Results: one hundred and eighty-three grafts were enrolled to the ABI group and 179 to the DD group. The primary assisted patency, secondary patency and limb salvage rates were 67%, 74%, 85% for the ABI group and 67%, 73%, 81% for the DD group. Ninety grafts in the ABI group and 57 in the DD group had surveillance that completely adhered to the protocol. The outcome was also similar for these groups at one year (77%, 87%, 94% and 77%, 83%, 93% respectively), although grafts were revised more frequently in the DD group. Conclusions: intensive surveillance with duplex scanning did not improve the results of any outcome criteria examined. To demonstrate any potential benefit of duplex scanning for vein-graft surveillance a multicentre study with a large number of patients to ensure sufficient power is needed. Key Words: Infrainguinal bypass; Vein graft; Surveillance; Duplex scanning. Introduction a lot of indirect evidence of improved clinical outcome with duplex-based surveillance programmes. 9,11 12 Autogenous vein of good quality is the conduit of However, randomised trials are few in number, small choice for infrainguinal bypass grafting. However, in sample size and their results contradictory. In a good initial success may be threatened by the de- Swedish trial the group with a 3-year intensive followvelopment of intrinsic vein-graft stenosis, which ac- up regime had a 25% better assisted primary patency counts for approximately 60% of graft thromboses. 1,2 rate than the group with clinical follow-up only. 13 In The strategy of periodic non-invasive graft surveillance their trial it was not differentiated whether the positive and elective revision of failing grafts prior to occlusion effect was gained due to more intensive follow-up or has become common practice. 4 6,8,10 This treatment due to the surveillance methods. In our own previous policy is believed to improve clinical results in terms study the utilisation of duplex scanning did not salvage of salvaged bypasses and limbs and also to be conduits or limbs, despite a higher revision rate in the cost-effective, at least during the first postoperative duplex group. 23 As a confounding factor the difficulty year. 7,14,16 was also highlighted of following up these elderly Duplex scanning is the most widely used surpatients with significant co-morbidities. veillance method, as it provides both anatomical and Herein, we report on the results of our prospective haemodynamic data about graft function. It is wellrandomised surveillance trial of 362 infrainguinal veindocumented that the duplex findings of a stenosis graft reconstructions as an extension of our previously correlate well with both the angiographic findings and published preliminary report of 179 patients. 23 Spedirect haemodynamic measurements There is also cifically, this study concentrates on those patients who completely adhered to the surveillance protocol, thus Please address all correspondence to: L. Ihlberg, Division of Vaseliminating the potential bias of patient non-comcular Surgery, Department of Surgery, Helsinki University Central Hospital, P.O. Box 262, FIN HYKS, Finland. pliance on the results /99/ $12.00/ Harcourt Publishers Ltd.

2 396 L. Ihlberg et al. Materials and Methods velocity at the site of the stenosis and V1 is the peak systolic velocity at any other point within 2 cm at All primary infrainguinal bypass autogenous vein the normal adjacent graft). Other intervention criteria grafts performed at the Division of Vascular Surgery, included clinical signs of a failing graft and a decrease Helsinki University Central Hospital, between January in ABI of 0.15 or more compared to the immediate 1991 and December 1995 were included. The indication postoperative value. If a graft was suspected to have for surgical procedure was critical leg ischaemia, disduplex a stenosis or an occlusion on the basis of findings with abling intermittent claudication, popliteal aneurysm examination or on the basis of clinical criteria, or, on some rare occasions, acute ischaemia. The prefied intra-arterial angiography was undertaken. All veri- operative examinations included selective angiomore stenoses with a diameter reduction of 50% or graphy of the affected leg. Patients were routinely were corrected with an open surgical procedure. operated on under epidural anaesthesia. Autologous vein was used as the graft material of choice with surgeon s preference for the in situ technique. Before January 1994, the reconstruction was intraoperatively Statistical analysis assessed with an on-table angiogram and/or con- The analysis for the whole study group was done on an tinuous-wave Doppler. After that point of time, transit- intention-to-treat basis. For patients with completely time flow measurements were used and completed accomplished surveillance, only patients alive, with with angioscopy, as necessary. Vancomycin and ce- no amputations and with grafts open after the first furoxime were given intravenously as an antibiotic postoperative month and complete follow-up exprophylaxis. On the first postoperative day, the ankle aminations up to 1 year were analysed on a treatmentpressure was measured and ankle brachial pressure given basis. The primary assisted and secondary patindex (ABI) calculated. No duplex scans were done ency rates as well as limb salvage rates after the first prior to discharge. postoperative year were calculated by Kaplan Meier Patients were randomised at the time of operation, survival estimates (according to the methods reaccording to the date of birth, into two different sur- commended by the revised version of Ad Hoc Comveillance groups. Patients born on odd days were mittee on Reporting Standards 3 and the data from enrolled into a surveillance programme with clinical these survival estimates were compared with logrank examination and ankle brachial index measurement test. Non-parametric variables were presented as at each visit (ABI group), whereas, for patients born median and interquartile range and the significance of on even days, duplex scans were added (DD group). any difference was calculated with the Mann Whitney Patients underwent a surveillance programme with test. For categorical parameters, comparisons were outpatient visits at 1, 3, 6, 9, and 12 months after the performed by Pearson s Chi-squared analysis and, operation. Only patients alive, with no limbs amless when appropriate, with Fisher s exact test. p Values of putated, and their grafts open at first visit were included than 0.05 were considered significant. in the study. The duplex scans were performed by radiologists or radiologists-in-training under supervision of senior colleagues. An ATL Ultramark 9 colour duplex scanner Results (Advanced Technology Laboratories, Stevenage, U.K.) with 7.5 MHz transducer was used. For deeply located Intention-to-treat analysis grafts a 5 MHz probe was used as necessary. Patients Three hundred and sixty-two primary infrainguinal were routinely examined in the supine position. The bypass reconstructions on 344 patients were performed whole graft, including anastomotic areas, inflow artery during the study period. Three hundred and two and few centimetres of outflow vessel, was serially (83.2%) procedures were due to critical leg ischaemia, scanned, searching for specific structural abnormalities whereas 52 (14.4%) were due to severe claudication, or exceptional flow characteristics in colour-flow im- 5 (1.4%) due to popliteal aneurysm and 3 (0.8%) due ages. Graft flow velocity and blood flow patterns were to acute ischaemia. One hundred and eighty-three evaluated at multiple sites on the course of the bypass. grafts in 175 patients were randomised to the ABI A graft at risk of failure on the basis of duplex group, whereas the respective figures for the DD group scanning was defined as one with low flow (<45 cm/ were 179 and 167. During the first postoperative month s), and/or a focal velocity disturbance, where V2/V1 19 grafts (10%) in the ABI group and 21 (12%) in the ratio is greater than 2.0 (V2 is the the peak systolic DD group became occluded, 13 (7%) and 11 (6%)

3 Does Duplex-based Surveillance Prevent Vein-graft Failure? 397 Table 1. Vascular risk factors and indications for revascularisation. Table 3. Location of graft-related stenosis in angiograms and type of revision. ABI group DD group ABI group DD group Age (years) Sex (M/F) 53/37 33/24 Angiographic findings in failing Diabetes 36 (40.0%) 22 (39.3%) grafts Hypertension 29 (32.2%) 18 (32.7%) Proximal native artery 2 1 Cardiac risk 42 (46.7%) 25 (44.6%) Proximal anastomosis 1 2 Cerebrovascular disease 14 (15.7%) 9 (16.1%) Vein graft 3 8 Indication Distal anastomosis 0 1 Intermittent claudication 18 (19.5%) 10 (17.0%) Distal native artery 0 1 Rest pain 23 (25.0%) 19 (32.2%) Revisions in failing grafts Ischaemic ulcer 35 (38.0%) 22 (37.3%) Patch angioplasty 3 8 Gangrene 14 (15.2%) 7 (11.9%) Vein graft interposition 1 1 Popliteal aneurysm 1 (1.1%) 1 (1.7%) Proximal native artery 2 1 revascularisation Distal vein jump graft 0 2 Table 2. The choice of graft material and the level of outflow artery. ABI group DD group and in the DD group 0.47 (0.26 to 0.68) (NS). The In situ vein immediate postoperative ABI was 0.89 (0.62 to 1.16) Popliteal artery 26 (28.9%) 15 (26.3%) and 0.88 (0.65 to 1.11) in the ABI group and in the DD Crural arteries 26 (28.9%) 15 (26.3%) group respectively (NS). No difference was seen in Pedal arteries 15 (16.7%) 13 (22.8%) the follow-up time for the ABI group and the DD Ex situ vein Popliteal artery 4 (4.4%) 5 (8.8%) group (18.8 months [6.9 to 30.8] and 23.7 months [11.7 Crural arteries 7 (7.8%) 4 (7%) to 35.6] respectively). Pedal arteries 12 (13.3%) 5 (8.8%) patients died, whereas 4 (2%) and 8 (4.5%) legs, respectively, were amputated. Thus, 147 grafts in the Graft stenoses and revision ABI group and 139 grafts in the DD group entered Angiography was programmed, as suggested by the the surveillance period. From these 286 bypasses, 33 intervention criteria, on eight occasions in the ABI were lost to follow-up within a year. The primary group and on 18 occasions in the DD group, including assisted patency rates for the ABI group and the DD all grafts which were subsequently revised (Table 3). group were 67% and 67%, the secondary patency rates Two grafts in the ABI group and 4 grafts in the DD were 74% and 73% and the limb salvage rates were groups turned out to be normal in angiography. Five 85% and 81%, respectively. The differences between of these grafts remained patent during the follow-up the groups were not statistically significant. period. In one graft a stenosis developed 8 months later, with subsequent revision. All angiographically normal grafts had a V2/V1 ratio of less than 3.5. One Analysis of completely accomplished surveillance graft in the DD group with a duplex-suspected stenosis in the body of the graft (V2/V1=4) was found to be Of the 253 bypasses available for one-year follow-up, occluded on angiography 14 days later. After throm- 106 did not attend all the surveillance examinations bolysis, a graft stenosis was identified and corrected according to the study protocol. This left 90 grafts in with patch angioplasty, resulting in subsequent sec- 85 patients in the ABI group and 57 grafts in 57 patients ondary patency. in the DD group that had a completely accomplished Six grafts in the ABI group and 8 grafts in the DD surveillance programme, and a comparison of different group required 6 and 11 revisions, respectively. Three variables was made between these groups. There were grafts required two independent revisions. In one no differences between these groups regarding sex, graft, patch angioplasty of the proximal graft and age, preoperative medical condition, or amount of distal-vein jump graft were done during the same previous vascular interventions (Table 1). Also, the procedure. The type of graft revision is illustrated in indication for, or the type of, the surgical procedure Table 3. The median time of revision was 238 days did not differ between the groups (Table 2). The pre- (range 51 to 350 days). There were no deaths within operative ABI in the ABI group was 0.46 (0.17 to 0.75) 30 days of revision procedures.

4 398 L. Ihlberg et al. Assisted primary patency % 77% Number of patients at risk at different time intervals At discharge 1 month 3 months 6 months 9 months 12 months ABI group DD group Secondary patency rate % 83% Number of patients at risk at different time intervals At discharge 1 month 3 months 6 months 9 months 12 months ABI group DD group Months Months Fig. 1. No difference could be seen in assisted primary patency rates Fig. 2. Secondary patency rates: ABI group ( ); DD group ( ). although revisions were done more frequently in the DD group. ABI group ( ); DD group ( ). Graft occlusion It is well known that graft thrombosis can occur with- out warning signs or symptoms and that attempted salvage of thrombosed grafts produces inferior long- term patency rates. In this respect it seems logical to search for optimal non-invasive methods to identify graft-threatening stenoses and to correct them before thrombosis ensues, in order to improve long-term patency. 21 Few studies have concentrated upon the subject of the natural course of a stenosed graft left untreated, and they are contradictory. In a study by Barnes et al., 22 a drop in ABI considered haemodynamically significant (>0.20) did not correlate with cumulative 5-year patency compared with grafts where ABI remained stable postoperatively. On the contrary, in studies by Mattos et al. 11 and Wilson et al. 12 the patency rates were lower for subgroups of untreated patients known to have stenosis detected by duplex scanning. In these studies, the graft revision was done rather selectively and the decision not to intervene was influenced by both several clinical factors and by the patient s or surgeon s reluctance to undergo an Twenty-four grafts became occluded during the surveillance period, 13 in the ABI group and 11 in the DD group. The median time for occlusion from the operation for the ABI group and the DD group was 181 days (range 44 to 346 days) and 157 days (78 to 365 days) respectively (p=0.927) and the median time from last outpatient visit was 79 days (range 12 to 98 days) and 51 days (range 12 to 93 days), respectively (p=0.178). In the ABI group, none of the revised grafts were occluded during study period. In the DD group, one graft was occluded 85 days after patch angioplasty of distal anastomosis. One graft in the DD group was also found to have low flow (30 cm/s) at a 3 month control visit, but this was neglected with subsequent graft occlusion two months later on. Two grafts in the ABI group and one in the DD group were thrombolysed and the underlying pathology corrected for maintenance of secondary patency. Three grafts in both groups underwent a secondary bypass procedure. For 19 grafts (11 in the ABI group and eight in the DD group), the cause for the occlusion could not be registered. Patency and amputation rates The one-year primary assisted and secondary patency rates calculated from the time of the initial operation for the ABI group were 77% and 87%, respectively, and for the DD group they were 77% and 85%, respectively (Figs 1 and 2). Log-rank tests for comparison of the survival estimates did not show a significant difference between the groups. Also, the leg salvage rates did not differ statistically between the groups (94% and 93%, respectively) (Fig. 3). Leg salvage Months 94% 93% Number of patients at risk at different time intervals At discharge 1 month 3 months 6 months 9 months 12 months ABI group DD group Fig. 3. Leg salvage rates: ABI group ( ); DD group ( ). Discussion

5 Does Duplex-based Surveillance Prevent Vein-graft Failure? 399 operation on an asymptomatic state. This naturally The case for vein-graft surveillance after infrainguinal weakens somewhat the significance of these findings. arterial reconstruction is not proven, espeweakens The rationale of our study design was to allow the cially when it involves technically demanding, subjective, potential of duplex scanning as a screening method to time-consuming and expensive methods used, be detected without other obscuring variables. Indeed, such as duplex scanning. On the basis of our experiences, duplex was very effective in detecting haemodynamically the cost-effectiveness of duplex is quesduplex compromised grafts compared with ABI tionable. There are some promising preliminary results measurements. The revision rate of angiographically of simpler alternatives such as impedance analysis, verified stenoses was 2.5-fold in the duplex group. but they have not gained wider acceptance. So ample This is in accordance with data from other groups, room for new, objective and inexpensive surveillance where the reported sensitivity of ABI measurements methods exists. The considerable workload required in finding stenoses is 38 55% when compared with for a graft surveillance programme and revision policy duplex. 24 However, the reason for graft surveillance is has been highlighted. 29 Shortening the duration of not just to detect stenotic lesions, but to predict which the surveillance period or concentrating on high-risk grafts become occluded without intervention and, fur- groups has also been advocated. 15,30 thermore, to salvage limbs that otherwise are doomed In conclusion, this study failed to show the potential to be amputated. Disappointingly, we were unable to benefit of duplex scanning, in improving one-year show improved clinical outcome from the surveillance patency or limb salvage, of infrainguinal vein bypasses programme with the intention-to-treat analysis of the even in the subset of patients who had complete whole study population, where, despite the case surveillance. This study reinforces the need for a large, sample size, a type II statistical error cannot be ruled randomized multicentre trial of graft surveillance. out. The case was the same when only grafts whose surveillance programme adhered to the study protocol were analysed. To enable duplex to demonstrate its Acknowledgements full potential, the inclusion criteria for this analysis We are grateful to Jukka Ollgren, M.Sc., for his valuable assistance were strict. Patients were deleted from this analysis in the statistical analysis and to Anita Mäkelä, R.N., for technical even if they missed only one surveillance examination. assistance in collecting the data. Leo Ihlberg is funded by a grant Utilising this criterion, 42% of the grafts that were from Finska Läkaresällskapet, which is kindly appreciated. patent at one month were excluded from this analysis. The number of grafts that were actually lost to followup was reasonably low (9%). How these figures compare with other surveillance trials is not known, because References they are not frequently reported. We believe 1Szilagyi DE, Elliot JP, Hageman JH et al. Biologic fate of that future studies should report not only the lost-to- autologous vein implants as arterial substitutes. Ann Surg 1973; follow-up ratio, but also how well the patients actually 178: attended the surveillance programme. 2Mills JI, Roy MF, Taylor SM. The characteristics and anatomic distribution of lesions that cause reversed vein graft failure: A The majority of grafts became occluded without five-year prospective study. J Vasc Surg 1993; 17: warning signs or any obvious cause. The question 3Rutherford RB, Baker JD, Ernst C et al. Recommended stand- of whether a large proportion of haemodynamically ards for reports dealing with lower extremity ischemia: Revised version. J Vasc Surg 1997; 26: compromised grafts was missed and, if so, whether it 4Moody P, Gould DA, Harris PL. Vein graft surveillance improves patency in femoro-popliteal bypass. Eur J Vasc Surg 1990; was due to limitations of the method itself or due to accuracy differences between investigators, remains 4: Harris PL. Vein graft surveillance all part of the service. Br J unanswered. Duplex is believed to be a highly op- Surg 1992; 79: erator-dependent method, but up to what extent this 6Bergamini TM, George SM, Massey HT et al. Intensive sur- is true is actually not known. No data exists on intraimproves long-term graft patency and limb salvage. Ann Surg veillance of femoropopliteal-tibial autogenous vein bypasses and interobserver agreement of duplex scanning for 1995; 221: vein-graft surveillance and this issue should be ad- 7Caps MT, Cantwell-Gab K, Bergelin RO, Strandness DE Jr. dressed in future studies. Nevertheless, an essential Vein graft lesions: time of onset and rate of progression. J Vasc Surg 1995; 22: prerequisite for a duplex-based surveillance pro- 8Colledge J, Beattie DK, Greenhalgh RM, Davies AH. Have gramme to succeed is a skilled operator. The duplex the results of infrainguinal bypass improved with the widespread criteria we used for interventions were probably strict utilisation of postoperative surveillance? Eur J Vasc Endovasc enough, as others have suggested that mild stenoses (V2/V ) can be treated by observation. 25 Surg 1966; 11: Idu MM, Blankenstein JD, de Geier P, Truyen E, Buth J. Impact of a color-flow duplex surveillance program on infrainguinal

6 400 L. Ihlberg et al. vein graft patency: a five-year experience. J Vasc Surg 1992; 15: of a surveillance protocol and elective revision. J Vasc Surg 1989; : Mills JL, Harris EJ, Taylor Jr LM. et al. The importance of 22 Barnes RW, Thompson BW, MacDonald CM et al. Serial noninvasive routine surveillance of distal bypass grafts with duplex scanning: studies do not herald postoperative failure of femororoutine a study of 379 reversed vein grafts. J Vasc Surg 1990; 12: popliteal or femorotibial bypass grafts. Ann Surg 1989; 210: 11 Mattos MA, Bemmelsen PS, Hodgson KJ et al. Does correction of stenoses identified with color duplex scanning improve infra- 23 Ihlberg L, Luther M, Tierala E, Lepäntalo M. The utility of inguinal graft patency? J Vasc Surg 1993; 17: duplex scanning in infrainguinal vein graft surveillance: Results 12 Wilson YG, Davies AH, Currie IC et al. Vein graft stenosis: from a randomised controlled study. Eur J Vasc Endovasc Surg incidence and intervention. Eur J Vasc Endovasc Surg 1996; 11: 1998; 16: Lepäntalo M, Mätzke S, Luther M. Is it necessary to assess 13 Lundell A, Lindblad B, Bergqvist D, Hansen F. Femoropopliteal-crural pressures after infrainguinal bypass? In: Greenhalgh RM, Fowkes graft patency is improved by an intensive sur- FGR, eds. Trials and Tribulations of Vascular Surgery. London: veillance program. A prospective randomised study. J Vasc Surg Saunders, 1996: ; 21: Olojugba DH, McCarthy MJ, Naylor AR, Bell PRF, London 14 Mills JL, Bandyk DF, Gahtan V, Esses GE. The origin of NJM. At what peak velocity ratio value should duplex-detected infrainguinal vein graft stenosis: A prospective study based on infrainguinal vein graft stenoses be revised? Eur J Vasc Endovasc duplex surveillance. J Vasc Surg : Surg 1988; 15: Idu MM, Buth J, Cuypers Ph et al. Economising vein-graft 26 Woodburn KR, Murtagh A, Breslin P et al. Insonation and surveillance programs. Eur J Vasc Endovasc Surg 1998; 15: impedance analysis in graft surveillance. Br J Surg 1995; 82: 16 Taylor PR, Wolfe JHN, Tyrrell MR et al. Graft stenosis: justification for 1 year surveillance. Br J Surg 1990; 77: Arora S, Meier GH, Pedersen H et al. Non-invsive impedance 17 Buth J, Disselhoff B, Sommelin C, Stam L. Color flow duplex analysis: a new non-invasive test for graft surveillance. Cardiovasc criteria for grading stenosis in infrainguinal vein grafts. J Vasc Surg 1995; 3: Surg 1991; 14: Wyatt MG, Muir RM, Tennant WG et al. Impedance analysis 18 Sladen JG, Reid JDS, Cooperberg PL et al. Color flow duplex to identify at risk femorodistal graft. J Vasc Surg 1991; 13: screening of infrainguinal grafts combining low- and high-veload 29 Loftus IM, Reid A, Thompson MM et al. The increasing work- locity criteria. Am J Surg 1989; 158: required to maintain infrainguinal bypass graft patency. 19 Taylor PR, Tyrrell MR, Crofton M et al. Colour flow imaging Eur J Vasc Endovasc Surg 1998; 15: in the detection of femorodistal graft and native artery stenosis: 30 Mercer KG, Scott DJA, Turton EPL, Berridge DC, Weston improved criteria. Eur J Vasc Surg 1992; 6: MJ. Can intraoperative flow measurements identify grafts for 20 Belkin M, Schwartz LB, Donaldson MC, Mannick JA, intensive duplex surveillance? ESVS Abstract Book 1998; 114. Whittemore AD. Hemodynamic impact of vein graft stenoses 31 Jensen LP, Nielsen OM, Schroeder TV. The importance of and their prediction in vascular laboratory. J Vasc Surg 1997; 25: complete follow-up for results after femoro-infrapopliteal sur gery. Eur J Vasc Endovasc Surg 1996; 12: Bandyk DF, Schmitt DD, Seabrook GR, Adams MB, Towne JB. Monitoring functional patency of in situ vein bypasses: the impact Accepted 28 April 1999