- Email: [email protected]
Outcomes of completion imaging for lower extremity bypass in the Vascular Quality Initiative
From the Western Vascular Society
Outcomes of completion imaging for lower extremity bypass in the Vascular Quality Initiative Karen Woo, MD, Owen P. Palmer, MD, Fred A. Weaver, MD, and Vincent L. Rowe, MD, for the Society for Vascular Surgery Vascular Quality Initiative, Los Angeles, Calif Objective: The objective of this study was to determine the association of intraoperative completion imaging (CI) for lower extremity vein bypass to a below-knee target with primary patency in the Vascular Quality Initiative. Methods: The Vascular Quality Initiative database was queried from January 2003 to October 2013 for lower extremity bypass (LEB) procedures that were elective, had an indication of occlusive disease, used a single-segment greater saphenous vein conduit, and had a below-knee target. LEBs with inflow arteries above the knee and below the knee were included. LEBs with concomitant endovascular procedures were excluded. CI was defined as completion angiography, completion duplex ultrasound, or both. The end points were primary patency at discharge and at 1 year. Multivariable analysis was performed controlling for patient demographics, comorbidities, bypass characteristics, and center. Results: Of 14,284 LEBs that were performed during the study period, 3147 satisfied the inclusion and exclusion criteria. Of 1457 (46%) that underwent CI, 287 (20%) underwent duplex ultrasound, 1116 (77%) underwent angiography, and 54 (3.7%) underwent both duplex ultrasound and angiography. There were more patients in the CI group with a history of smoking and a bypass graft crossing the knee. There was no difference in primary patency at discharge between the two groups (CI, 93.2% vs no CI, 93.8%; P [ .52). Of the patients who underwent CI, the discharge primary patency was 95.1% for completion duplex ultrasound vs 92.8% for completion angiography (P [ .17). On multivariable analysis, there was no significant association of CI with discharge primary patency (P [ .69). The 1-year primary patency was 63% in the CI group vs 68% in the no CI group (P [ .051). The 1-year primary patency was 60% for the duplex ultrasound group vs 65% for the angiography group (P [ .61). On multivariable analysis, there was no significant association of CI with 1-year primary patency (P [ .69). Conclusions: In electively performed LEBs using single-segment saphenous vein to a below-knee target artery for occlusive disease, CI does not influence primary graft patency at discharge or at 1 year. (J Vasc Surg 2015;-:1-5.)
Peripheral arterial disease (PAD) affects 10.7% of Americans annually.1 Despite the recent increase in the use of endovascular intervention for the treatment of PAD, lower extremity bypass (LEB) remains an important management option in patients with symptomatic PAD.2 Postoperative surveillance duplex ultrasound imaging of LEB vein grafts with appropriate intervention for correction of lesions has been widely implemented as routine practice to increase cumulative patency of grafts.3,4 More controversial is the use of intraoperative completion imaging (CI) to evaluate the technical adequacy of the bypass graft. Completion angiography was advocated as early as 1968 for the identification of correctable technical defects.5 Since then, other authors have also supported the routine use of completion angiography.6 More recently, investigators have described the association of abnormal completion From the Division of Vascular Surgery and Endovascular Therapy, Keck School of Medicine, University of Southern California. Author conflict of interest: none. Presented at Scientific Session I of the Twenty-ninth Annual Meeting of the Western Vascular Society, Coronado, Calif, September 20-23, 2014. Reprint requests: Karen Woo, MD, 1520 San Pablo St, Ste 4300, Los Angeles, CA 90033 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2015 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvs.2015.03.037
duplex ultrasound findings with increased risk of early graft failure.7,8 At the same time, a single-institution study has demonstrated that completion angiography did not improve rates of early graft occlusion.9 Furthermore, data from the Project of Ex Vivo graft Engineering via Transfection III (PREVENT III) trial, which included 1404 patients, demonstrated no association between CI and short- and long-term primary patency.10 As such, the objective of this study was to determine the association of intraoperative CI with primary patency of LEB in a large observational database. METHODS The Vascular Quality Initiative (VQI) is a collaborative of regional quality groups collecting and analyzing data in an effort to improve patient care. A detailed description of the VQI database can be found at http://www. vascularqualityinitiative.org. Information about regional quality groups, participating centers, and participating surgeons is available at http://www.vascularqualityinitiative. org. The LEB module of the VQI database was queried from January 2003 through October 2013. Only elective LEBs performed with a single-segment greater saphenous vein conduit for occlusive disease to a below-knee target vessel were included. LEBs with a concomitant peripheral vascular intervention were excluded, as were procedures with missing data regarding conduit, target vessel, 1
JOURNAL OF VASCULAR SURGERY --- 2015
2 Woo et al
indication for bypass, and CI. CI was defined as intraoperative angiography, duplex ultrasound, or both. Grafts were classified as crossing the knee or not crossing the knee. The grafts that did not cross the knee were grafts with both the proximal and distal target artery below the knee. The primary outcomes were primary patency at discharge and primary patency at 1 year. Statistical analysis was performed with Statistical Analysis System 9.4 (SAS Institute, Cary, NC). As two primary outcomes are being evaluated, a P value # .025 was considered statistically significant. Results for continuous variables are presented as mean 6 standard deviation. Two-group comparisons of continuous variables were assessed by the independent samples t-test after confirmation of normal distribution of continuous variables. Group comparisons of categorical variables were assessed by c2 test. Two multivariable models were constructed to evaluate the association of CI with discharge primary patency and 1year primary patency. The variables in Table were chosen as covariates for the multivariable models on the basis of clinical judgment of factors that may be associated with patency. The interactions between all covariates and CI were tested sequentially, and none was found to be significant. Thus, no interaction terms were included in the models. Generalized estimating equations logistic regression was used to evaluate the association of CI with discharge primary patency. All covariates in Table were included in the multivariable model, and the analysis was adjusted for center by taking center as a repeated factor in the generalized estimating equations model and assuming an exchangeable working correlation. A multivariable Cox proportional hazards model was used to assess the association of CI with 1-year primary patency. Only patients with 1-year follow-up data available were included in the analysis of 1-year primary patency. All covariates in Table were included in the multivariable proportional hazards model, and the analysis was adjusted for center by using a robust sandwich covariance matrix estimate to account for the dependence of observations clustered on center. The proportional hazards assumption was tested by introducing a time-dependent variable (CI*time) into the model.11 Although the sample sizes for this study are predetermined by the number of cases available in the VQI database, a sample size analysis was performed. For discharge primary patency, to have an 80% power of detecting a difference between 95% and 92% at a significance level of 2.5% (two tailed), 1279 patients would be required in each group. For 1-year primary patency, to have an 80% power of detecting a difference between 70% and 65% at a significance level of 2.5% (two tailed), 1652 patients would be required in each group. This study was approved by the VQI Patient Safety Organization Research Advisory Committee. Data are collected under the auspices of the Society for Vascular Surgery Patient Safety Organization, which does not require patient consent. For research purposes, the data are de-identified and are exempt from Institutional Review Board review.
Table. Characteristics of patients with completion imaging (CI) and of those without CI
Mean age, years Male White race Hypertension Diabetes Dialysis dependence Coronary artery disease History of smoking ASA class 1 or 2 3 or 4 Indication Claudication Rest pain/tissue loss Previous ipsilateral LEB Previous ipsilateral lower extremity endovascular intervention Tibial/pedal target artery Graft crossing the knee Bypass orientation Reversed In situ Nonreversed, transposed Concomitant femoral endarterectomy Concomitant suprainguinal bypass Discharge antiplatelet Discharge statin
CI (n ¼ 1457)
No CI (n ¼ 1690)
66.6 1031 (71) 1238 (85) 1294 (89) 742 (51) 97 (6.7) 458 (31) 1217 (84)
67.2 1165 (69) 1393 (82) 1489 (88) 874 (52) 108 (6.4) 478 (28) 1356 (80)
160 (11) 1297 (89)
173 (10) 1517 (90)
377 1080 154 336
437 1253 185 423
P .12 .25 .05 .57 .65 .77 .05 .02 .54 .64
(26) (74) (11) (23)
812 (56) 1363 (94) 477 650 330 498
(33) (45) (23) (35)
(26) (74) (11) (25)
929 (55) 1536 (91) 848 414 428 413
(50) (25) (25) (25)
.73 .2 .67 .005 <.001
<.001
12 (1)
20 (1)
.57
1216 (84) 1083 (79)
1422 (84) 1295 (78)
.85 .41
ASA, American Society of Anesthesiologists; LEB, lower extremity bypass. Data are presented as number (%).
RESULTS Of the 14,284 LEBs in the VQI database, 3147 LEBs met the inclusion and exclusion criteria for the study (Fig 1). These LEBs were performed by 447 surgeons at 124 centers in 15 regional quality groups of the VQI. Of the included patients, 1457 (46%) underwent CI and 1690 (54%) did not. For patients undergoing CI, angiography was used in 1116 (77%), duplex ultrasound in 287 (20%), and both angiography and duplex ultrasound in 54 (3.7%). There were more smokers and more patients with a graft crossing the knee in the CI group (Table). More patients in the CI group were taking an antiplatelet agent preoperatively than in the no CI group (79% vs 76%; P ¼ .02), but there was no difference in the prevalence of patients discharged on an antiplatelet agent (Table). There was no difference in the prevalence of patients taking a statin preoperatively between the two groups (CI, 1045 [72%] vs no CI, 1188 [70%]; P ¼ .4). There was also no difference in the prevalence of those discharged on a statin (Table). More LEBs were performed in the in situ orientation in the CI group, and more were performed in the reversed orientation in the no CI group (Table). Mean length of stay after LEB was 6.3 days in the CI group vs 5.7 days
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Woo et al 3
Fig 2. One-year primary patency. CI, Completion imaging; SE, standard error.
association of CI with primary patency at 1 year (P ¼ .69). We found no evidence that the proportional hazard assumption with respect to CI was not met (P ¼ .70). DISCUSSION
Fig 1. Consolidated Standards of Reporting Trials (CONSORT) diagram of included and excluded cases.
in the no CI group (P ¼ .02). There was no difference between the two groups with respect to the incidence of anticoagulation at discharge (CI, 327 [23%] vs no CI, 346 [21%]; P ¼ .16). One-year follow-up data were available for 59% of the LEBs, 984 in the CI group and 873 in the no CI group, with a mean follow-up of 351 days. The primary patency at discharge was 93.2% for the CI group and 93.8% for the group that received no CI (P ¼ .52). For CI patients who had completion angiography, the primary discharge patency was 92.8%; whereas for those who underwent completion duplex ultrasound, the primary discharge patency was 95.1% (P ¼ .17). After controlling for all of the covariates in Table as well as for center, there was no significant association of CI with discharge primary patency (P ¼ .69). The primary patency at 1 year (Fig 2) was 63% for the CI group and 68% for the no CI group (P ¼ .05). The 1-year primary patency was 65% for the completion angiography group vs 60% for the completion duplex ultrasound group (P ¼ .61). After controlling for all of the covariates in Table as well as for center, there was no significant
When Renwick described the use of completion angiography during LEB in 1968, his manuscript concluded that “operative arteriography is a simple, valuable, and safe procedure, and that no femoral reconstruction is complete until it has been carried out.”5 These sentiments are emphasized four decades later in the eighth edition of Rutherford’s Vascular Surgery, the textbook widely accepted in the vascular surgery community to be the authority on the practice of vascular surgery.12 These strong recommendations exist despite insufficient and conflicting evidence in the literature. Proponents of CI assert that CI identifies technical graft issues that are not detected by pulse palpation or continuous-wave Doppler examination.6 For proponents of CI, CI detects these technical graft issues, leading to improved early primary patency and thus presumably long-term patency as well. In a single-institution study of 116 crural bypasses, a low distal graft end-diastolic velocity of <5 cm/s increased the risk of loss of primary patency with a hazard ratio of 3.3.7 Similarly, other authors have demonstrated that a distal graft end-diastolic velocity of <8 cm/s predicted early graft thrombosis with 76% sensitivity and 75% specificity.8 However, one small single-institution study of 93 LEBs demonstrated no difference in early graft occlusion between those that underwent completion angiography and those that did not.9 Data from the PREVENT III study, which included 1404 patients, demonstrated similar findings, although the type of completion study was not described.10 As such, we sought to investigate the influence of CI on primary patency of LEBs in the VQI. We chose primary patency as the primary outcome as we believe that primary assisted and secondary patency are largely related to diligence of surveillance imaging and aggressiveness in treating identified lesions. As the objective of CI is to detect technical errors and fixed valves at the time of bypass creation, we
JOURNAL OF VASCULAR SURGERY --- 2015
4 Woo et al
believe that primary patency is the best reflection of the association of CI with patency. The VQI database is an ideal setting in which to conduct this study as it includes academic centers, community centers, and nonacademic teaching hospitals, which provides a broad view of patient populations as well as practice patterns. A fundamental issue regarding the efficacy of CI is the definition of a finding that requires revision. The interpretation of angiograms can be subjective as to what is a flow-limiting lesion. In the literature, the rate of abnormal findings on CI varies from 8% to 15%.6,13 Likewise, when duplex ultrasound is used, the definition of a hemodynamically significant lesion varies from absolute peak systolic velocity of >300 cm/s and a velocity ratio of >5 to enddiastolic velocity of <5 or <8.7,8,13 In our previous work with the VQI database, we found that in LEB patients who had concomitant peripheral vascular intervention during their LEB, a tibial or pedal target artery or disadvantaged vein was more likely to undergo CI.14 These findings appear to reflect the bias of surgeons toward performing CI when factors exist that are known to be associated with vein graft failure.10 On the basis of these findings, to minimize selection bias and confounding factors associated with vein graft failure, we imposed the inclusion criteria of single-segment saphenous vein and below-knee target vessel as well as the exclusion criterion of concomitant peripheral vascular intervention. By selecting out only single-segment saphenous vein grafts, it may seem, on the surface, that we selected for a group of bypasses that required CI the least. However, as demonstrated in the literature, there is debate about the efficacy of CI for LEB, and further investigation is indicated. We chose not to select for “disadvantaged” bypasses in the VQI data set as the disadvantaged bypasses were heterogeneous with multiple segments of different types of vein, composite bypasses, various target vessels, various indications, and various levels of urgency. We believe that the heterogeneity would introduce potentially unmeasured confounders of patency that would not reflect the true association with CI. The findings of this analysis of the national VQI data set are consistent with a recent report from the Vascular Study Group of New England (VSGNE), a regional quality group of the VQI.15 The VSGNE study encompassed 2032 LEBs performed by 48 surgeons and included all conduit types, target arteries, and urgencies. No difference in discharge primary patency or 1-year primary patency was found for LEBs performed with and without CI. The current study includes a more uniform group of LEBs with the same results. The VQI does not collect data regarding whether the surgeon routinely or selectively performs CI. It is conceivable that the incidence of CI is a reflection of the surgeon’s instinct that there is a technical problem with the bypass. In an effort to address this issue, the VSGNE study also categorized surgeons into those who performed CI routinely or selectively, which was not done in this study. Nevertheless, there was no difference in the VSGNE data in discharge primary patency or 1-year primary patency between routine imagers and selective imagers.
One of the major limitations of this study is that the findings of CI, when it is performed, are not recorded in the VQI. Similarly, the interventions performed for abnormal CI findings are not recorded. Thus, it is unknown what the rate of significant findings on CI is and whether these findings spurred intraoperative intervention. Had these data been collected, it would have potentially enabled analysis of outcomes in grafts with abnormal findings on CI that were not corrected or residual abnormal findings after correction. There are several other limitations of this study. VQI data are self-reported and subject to the biases and errors of self-reporting. Ideally, a randomized controlled study would be performed to definitively answer the question of the efficacy of CI for LEB. However, given the cost and scope associated with such a study, it is unlikely that it will be performed. The follow-up rate at 59% is also less than ideal. Follow-up data were entered on only 984 patients in the CI group and 873 patients in the no CI group; based on the sample size analysis to detect a 5% difference in 1-year primary patency, these numbers are insufficient. Thus, it is possible that the lack of significant difference in 1-year primary patency between the two groups is related to small sample size. Ultimately, this is not a randomized controlled study and is vulnerable to surgeon bias and unmeasured patient and bypass graft bias. Whereas the results of this study did not demonstrate a patency benefit of CI, this does not indicate that CI does not have value. CI remains an important educational tool. Performance of CI gives the surgeon immediate feedback on the technical adequacy of the operation performed. Completion duplex ultrasound, in particular, allows the trainee to gain experience in performing duplex ultrasound scanning and interpreting duplex ultrasound images. CONCLUSIONS In electively performed LEBs using single-segment saphenous vein to a below-knee target artery for occlusive disease, CI does not influence primary graft patency at discharge or at 1 year. However, CI remains a valuable educational tool that provides immediate feedback to the surgeon. CI may provide a greater patency benefit for bypasses performed with disadvantaged conduit. Future studies should focus on the outcomes of CI in the disadvantaged conduit population. AUTHOR CONTRIBUTIONS Conception and design: KW, OP, FW, VR Analysis and interpretation: KW Data collection: KW Writing the article: KW, OP Critical revision of the article: KW, OP, FW, VR Final approval of the article: KW, OP, FW, VR Statistical analysis: KW Obtained funding: Not applicable Overall responsibility: KW
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Woo et al 5
REFERENCES 1. Nehler MR, Duval S, Diao L, Annex BH, Hiatt WR, Rogers K, et al. Epidemiology of peripheral arterial disease and critical limb ischemia in an insured national population. J Vasc Surg 2014;60: 686-95. 2. Goodney PP, Beck AW, Nagle J, Welch HG, Zwolak RM. National trends in lower extremity bypass surgery, endovascular interventions and major amputations. J Vasc Surg 2009;50:54-60. 3. Bandyk DF, Schmitt DD, Seabrook GR, Adams MB, Towne JB. Monitoring functional patency of in situ saphenous vein bypasses: the impact of a surveillance protocol and elective revision. J Vasc Surg 1989;9:286-96. 4. Idu MM, Blankenstein JD, de Gier P, Truyen E, Buth J. Impact of a color-flow duplex surveillance program on infrainguinal graft patency: a five-year experience. J Vasc Surg 1993;17:42-53. 5. Renwick S, Royle JP, Martin P. Operative angiography after femoropopliteal arterial reconstruction: its influence on early failure rate. Br J Surg 1968;55:134-6. 6. Mills JL, Fujitani RM, Taylor SM. Contribution of routine intraoperative completion arteriography to early infrainguinal bypass patency. Am J Surg 1992;164:506-11. 7. Scali ST, Beck AW, Nolan BW, Stone DH, De Martino RR, Chang CK, et al. Completion duplex ultrasound predicts early graft thrombosis after crural bypass in patients with critical limb ischemia. J Vasc Surg 2011;54:1006-10. 8. Rzucidlo EM, Walsh DB, Powell RJ, Zwolak RM, Fillinger MF, Schermerhorn ML, et al. Prediction of early graft failure with
9. 10.
11.
12.
13.
14.
15.
intraoperative completion duplex ultrasound scan. J Vasc Surg 2002;36:975-81. Dalman RL, Harris EJ, Zarins CK. Is completion arteriography mandatory after reversed-vein bypass grafting? J Vasc Surg 1996;23:637-44. Schanzer A, Hvelone N, Owens CD, Belkin M, Bandyk DF, Clowes AW, et al. Technical factors affecting autogenous vein graft failure: observations from a large multicenter trial. J Vasc Surg 2007;46:1180-90. Harrell FE, Lee KL. Verifying assumptions of the Cox proportional hazards model. In: Proceedings of the Eleventh Annual SAS Users Group International Conference. Cary, NC: SAS Institute Inc; 1986. p. 823-8. Mills JL. Infrainguinal disease: surgical treatment. In: Cronenwett JL, Johnston KW, editors. Rutherford’s vascular surgery. 8th ed. Philadelphia: Elsevier Saunders; 2014. p. 1758-81. Johnson BL, Bandyk DF, Back MR, Avino AJ, Roth SM. Intraoperative duplex monitoring of infrainguinal vein bypass procedures. J Vasc Surg 2000;31:678-90. Woo K, Palmer OP, Weaver FA, Rowe VL. Use of completion imaging during infrainguinal bypass in the Vascular Quality Initiative. J Vasc Surg 2015. [Epub ahead of print]. Tan TW, Rybin D, Kalish JA, Doros G, Hamburg N, Schanzer A, et al. Routine use of completion imaging after infrainguinal bypass is not associated with higher bypass graft patency. J Vasc Surg 2014;60: 678-85.
Submitted Dec 3, 2014; accepted Mar 17, 2015.
Outcomes of completion imaging for lower extremity bypass in the Vascular Quality Initiative Karen Woo, MD, Owen P. Palmer, MD, Fred A. Weaver, MD, and Vincent L. Rowe, MD, for the Society for Vascular Surgery Vascular Quality Initiative, Los Angeles, Calif Objective: The objective of this study was to determine the association of intraoperative completion imaging (CI) for lower extremity vein bypass to a below-knee target with primary patency in the Vascular Quality Initiative. Methods: The Vascular Quality Initiative database was queried from January 2003 to October 2013 for lower extremity bypass (LEB) procedures that were elective, had an indication of occlusive disease, used a single-segment greater saphenous vein conduit, and had a below-knee target. LEBs with inflow arteries above the knee and below the knee were included. LEBs with concomitant endovascular procedures were excluded. CI was defined as completion angiography, completion duplex ultrasound, or both. The end points were primary patency at discharge and at 1 year. Multivariable analysis was performed controlling for patient demographics, comorbidities, bypass characteristics, and center. Results: Of 14,284 LEBs that were performed during the study period, 3147 satisfied the inclusion and exclusion criteria. Of 1457 (46%) that underwent CI, 287 (20%) underwent duplex ultrasound, 1116 (77%) underwent angiography, and 54 (3.7%) underwent both duplex ultrasound and angiography. There were more patients in the CI group with a history of smoking and a bypass graft crossing the knee. There was no difference in primary patency at discharge between the two groups (CI, 93.2% vs no CI, 93.8%; P [ .52). Of the patients who underwent CI, the discharge primary patency was 95.1% for completion duplex ultrasound vs 92.8% for completion angiography (P [ .17). On multivariable analysis, there was no significant association of CI with discharge primary patency (P [ .69). The 1-year primary patency was 63% in the CI group vs 68% in the no CI group (P [ .051). The 1-year primary patency was 60% for the duplex ultrasound group vs 65% for the angiography group (P [ .61). On multivariable analysis, there was no significant association of CI with 1-year primary patency (P [ .69). Conclusions: In electively performed LEBs using single-segment saphenous vein to a below-knee target artery for occlusive disease, CI does not influence primary graft patency at discharge or at 1 year. (J Vasc Surg 2015;-:1-5.)
Peripheral arterial disease (PAD) affects 10.7% of Americans annually.1 Despite the recent increase in the use of endovascular intervention for the treatment of PAD, lower extremity bypass (LEB) remains an important management option in patients with symptomatic PAD.2 Postoperative surveillance duplex ultrasound imaging of LEB vein grafts with appropriate intervention for correction of lesions has been widely implemented as routine practice to increase cumulative patency of grafts.3,4 More controversial is the use of intraoperative completion imaging (CI) to evaluate the technical adequacy of the bypass graft. Completion angiography was advocated as early as 1968 for the identification of correctable technical defects.5 Since then, other authors have also supported the routine use of completion angiography.6 More recently, investigators have described the association of abnormal completion From the Division of Vascular Surgery and Endovascular Therapy, Keck School of Medicine, University of Southern California. Author conflict of interest: none. Presented at Scientific Session I of the Twenty-ninth Annual Meeting of the Western Vascular Society, Coronado, Calif, September 20-23, 2014. Reprint requests: Karen Woo, MD, 1520 San Pablo St, Ste 4300, Los Angeles, CA 90033 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2015 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvs.2015.03.037
duplex ultrasound findings with increased risk of early graft failure.7,8 At the same time, a single-institution study has demonstrated that completion angiography did not improve rates of early graft occlusion.9 Furthermore, data from the Project of Ex Vivo graft Engineering via Transfection III (PREVENT III) trial, which included 1404 patients, demonstrated no association between CI and short- and long-term primary patency.10 As such, the objective of this study was to determine the association of intraoperative CI with primary patency of LEB in a large observational database. METHODS The Vascular Quality Initiative (VQI) is a collaborative of regional quality groups collecting and analyzing data in an effort to improve patient care. A detailed description of the VQI database can be found at http://www. vascularqualityinitiative.org. Information about regional quality groups, participating centers, and participating surgeons is available at http://www.vascularqualityinitiative. org. The LEB module of the VQI database was queried from January 2003 through October 2013. Only elective LEBs performed with a single-segment greater saphenous vein conduit for occlusive disease to a below-knee target vessel were included. LEBs with a concomitant peripheral vascular intervention were excluded, as were procedures with missing data regarding conduit, target vessel, 1
JOURNAL OF VASCULAR SURGERY --- 2015
2 Woo et al
indication for bypass, and CI. CI was defined as intraoperative angiography, duplex ultrasound, or both. Grafts were classified as crossing the knee or not crossing the knee. The grafts that did not cross the knee were grafts with both the proximal and distal target artery below the knee. The primary outcomes were primary patency at discharge and primary patency at 1 year. Statistical analysis was performed with Statistical Analysis System 9.4 (SAS Institute, Cary, NC). As two primary outcomes are being evaluated, a P value # .025 was considered statistically significant. Results for continuous variables are presented as mean 6 standard deviation. Two-group comparisons of continuous variables were assessed by the independent samples t-test after confirmation of normal distribution of continuous variables. Group comparisons of categorical variables were assessed by c2 test. Two multivariable models were constructed to evaluate the association of CI with discharge primary patency and 1year primary patency. The variables in Table were chosen as covariates for the multivariable models on the basis of clinical judgment of factors that may be associated with patency. The interactions between all covariates and CI were tested sequentially, and none was found to be significant. Thus, no interaction terms were included in the models. Generalized estimating equations logistic regression was used to evaluate the association of CI with discharge primary patency. All covariates in Table were included in the multivariable model, and the analysis was adjusted for center by taking center as a repeated factor in the generalized estimating equations model and assuming an exchangeable working correlation. A multivariable Cox proportional hazards model was used to assess the association of CI with 1-year primary patency. Only patients with 1-year follow-up data available were included in the analysis of 1-year primary patency. All covariates in Table were included in the multivariable proportional hazards model, and the analysis was adjusted for center by using a robust sandwich covariance matrix estimate to account for the dependence of observations clustered on center. The proportional hazards assumption was tested by introducing a time-dependent variable (CI*time) into the model.11 Although the sample sizes for this study are predetermined by the number of cases available in the VQI database, a sample size analysis was performed. For discharge primary patency, to have an 80% power of detecting a difference between 95% and 92% at a significance level of 2.5% (two tailed), 1279 patients would be required in each group. For 1-year primary patency, to have an 80% power of detecting a difference between 70% and 65% at a significance level of 2.5% (two tailed), 1652 patients would be required in each group. This study was approved by the VQI Patient Safety Organization Research Advisory Committee. Data are collected under the auspices of the Society for Vascular Surgery Patient Safety Organization, which does not require patient consent. For research purposes, the data are de-identified and are exempt from Institutional Review Board review.
Table. Characteristics of patients with completion imaging (CI) and of those without CI
Mean age, years Male White race Hypertension Diabetes Dialysis dependence Coronary artery disease History of smoking ASA class 1 or 2 3 or 4 Indication Claudication Rest pain/tissue loss Previous ipsilateral LEB Previous ipsilateral lower extremity endovascular intervention Tibial/pedal target artery Graft crossing the knee Bypass orientation Reversed In situ Nonreversed, transposed Concomitant femoral endarterectomy Concomitant suprainguinal bypass Discharge antiplatelet Discharge statin
CI (n ¼ 1457)
No CI (n ¼ 1690)
66.6 1031 (71) 1238 (85) 1294 (89) 742 (51) 97 (6.7) 458 (31) 1217 (84)
67.2 1165 (69) 1393 (82) 1489 (88) 874 (52) 108 (6.4) 478 (28) 1356 (80)
160 (11) 1297 (89)
173 (10) 1517 (90)
377 1080 154 336
437 1253 185 423
P .12 .25 .05 .57 .65 .77 .05 .02 .54 .64
(26) (74) (11) (23)
812 (56) 1363 (94) 477 650 330 498
(33) (45) (23) (35)
(26) (74) (11) (25)
929 (55) 1536 (91) 848 414 428 413
(50) (25) (25) (25)
.73 .2 .67 .005 <.001
<.001
12 (1)
20 (1)
.57
1216 (84) 1083 (79)
1422 (84) 1295 (78)
.85 .41
ASA, American Society of Anesthesiologists; LEB, lower extremity bypass. Data are presented as number (%).
RESULTS Of the 14,284 LEBs in the VQI database, 3147 LEBs met the inclusion and exclusion criteria for the study (Fig 1). These LEBs were performed by 447 surgeons at 124 centers in 15 regional quality groups of the VQI. Of the included patients, 1457 (46%) underwent CI and 1690 (54%) did not. For patients undergoing CI, angiography was used in 1116 (77%), duplex ultrasound in 287 (20%), and both angiography and duplex ultrasound in 54 (3.7%). There were more smokers and more patients with a graft crossing the knee in the CI group (Table). More patients in the CI group were taking an antiplatelet agent preoperatively than in the no CI group (79% vs 76%; P ¼ .02), but there was no difference in the prevalence of patients discharged on an antiplatelet agent (Table). There was no difference in the prevalence of patients taking a statin preoperatively between the two groups (CI, 1045 [72%] vs no CI, 1188 [70%]; P ¼ .4). There was also no difference in the prevalence of those discharged on a statin (Table). More LEBs were performed in the in situ orientation in the CI group, and more were performed in the reversed orientation in the no CI group (Table). Mean length of stay after LEB was 6.3 days in the CI group vs 5.7 days
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Woo et al 3
Fig 2. One-year primary patency. CI, Completion imaging; SE, standard error.
association of CI with primary patency at 1 year (P ¼ .69). We found no evidence that the proportional hazard assumption with respect to CI was not met (P ¼ .70). DISCUSSION
Fig 1. Consolidated Standards of Reporting Trials (CONSORT) diagram of included and excluded cases.
in the no CI group (P ¼ .02). There was no difference between the two groups with respect to the incidence of anticoagulation at discharge (CI, 327 [23%] vs no CI, 346 [21%]; P ¼ .16). One-year follow-up data were available for 59% of the LEBs, 984 in the CI group and 873 in the no CI group, with a mean follow-up of 351 days. The primary patency at discharge was 93.2% for the CI group and 93.8% for the group that received no CI (P ¼ .52). For CI patients who had completion angiography, the primary discharge patency was 92.8%; whereas for those who underwent completion duplex ultrasound, the primary discharge patency was 95.1% (P ¼ .17). After controlling for all of the covariates in Table as well as for center, there was no significant association of CI with discharge primary patency (P ¼ .69). The primary patency at 1 year (Fig 2) was 63% for the CI group and 68% for the no CI group (P ¼ .05). The 1-year primary patency was 65% for the completion angiography group vs 60% for the completion duplex ultrasound group (P ¼ .61). After controlling for all of the covariates in Table as well as for center, there was no significant
When Renwick described the use of completion angiography during LEB in 1968, his manuscript concluded that “operative arteriography is a simple, valuable, and safe procedure, and that no femoral reconstruction is complete until it has been carried out.”5 These sentiments are emphasized four decades later in the eighth edition of Rutherford’s Vascular Surgery, the textbook widely accepted in the vascular surgery community to be the authority on the practice of vascular surgery.12 These strong recommendations exist despite insufficient and conflicting evidence in the literature. Proponents of CI assert that CI identifies technical graft issues that are not detected by pulse palpation or continuous-wave Doppler examination.6 For proponents of CI, CI detects these technical graft issues, leading to improved early primary patency and thus presumably long-term patency as well. In a single-institution study of 116 crural bypasses, a low distal graft end-diastolic velocity of <5 cm/s increased the risk of loss of primary patency with a hazard ratio of 3.3.7 Similarly, other authors have demonstrated that a distal graft end-diastolic velocity of <8 cm/s predicted early graft thrombosis with 76% sensitivity and 75% specificity.8 However, one small single-institution study of 93 LEBs demonstrated no difference in early graft occlusion between those that underwent completion angiography and those that did not.9 Data from the PREVENT III study, which included 1404 patients, demonstrated similar findings, although the type of completion study was not described.10 As such, we sought to investigate the influence of CI on primary patency of LEBs in the VQI. We chose primary patency as the primary outcome as we believe that primary assisted and secondary patency are largely related to diligence of surveillance imaging and aggressiveness in treating identified lesions. As the objective of CI is to detect technical errors and fixed valves at the time of bypass creation, we
JOURNAL OF VASCULAR SURGERY --- 2015
4 Woo et al
believe that primary patency is the best reflection of the association of CI with patency. The VQI database is an ideal setting in which to conduct this study as it includes academic centers, community centers, and nonacademic teaching hospitals, which provides a broad view of patient populations as well as practice patterns. A fundamental issue regarding the efficacy of CI is the definition of a finding that requires revision. The interpretation of angiograms can be subjective as to what is a flow-limiting lesion. In the literature, the rate of abnormal findings on CI varies from 8% to 15%.6,13 Likewise, when duplex ultrasound is used, the definition of a hemodynamically significant lesion varies from absolute peak systolic velocity of >300 cm/s and a velocity ratio of >5 to enddiastolic velocity of <5 or <8.7,8,13 In our previous work with the VQI database, we found that in LEB patients who had concomitant peripheral vascular intervention during their LEB, a tibial or pedal target artery or disadvantaged vein was more likely to undergo CI.14 These findings appear to reflect the bias of surgeons toward performing CI when factors exist that are known to be associated with vein graft failure.10 On the basis of these findings, to minimize selection bias and confounding factors associated with vein graft failure, we imposed the inclusion criteria of single-segment saphenous vein and below-knee target vessel as well as the exclusion criterion of concomitant peripheral vascular intervention. By selecting out only single-segment saphenous vein grafts, it may seem, on the surface, that we selected for a group of bypasses that required CI the least. However, as demonstrated in the literature, there is debate about the efficacy of CI for LEB, and further investigation is indicated. We chose not to select for “disadvantaged” bypasses in the VQI data set as the disadvantaged bypasses were heterogeneous with multiple segments of different types of vein, composite bypasses, various target vessels, various indications, and various levels of urgency. We believe that the heterogeneity would introduce potentially unmeasured confounders of patency that would not reflect the true association with CI. The findings of this analysis of the national VQI data set are consistent with a recent report from the Vascular Study Group of New England (VSGNE), a regional quality group of the VQI.15 The VSGNE study encompassed 2032 LEBs performed by 48 surgeons and included all conduit types, target arteries, and urgencies. No difference in discharge primary patency or 1-year primary patency was found for LEBs performed with and without CI. The current study includes a more uniform group of LEBs with the same results. The VQI does not collect data regarding whether the surgeon routinely or selectively performs CI. It is conceivable that the incidence of CI is a reflection of the surgeon’s instinct that there is a technical problem with the bypass. In an effort to address this issue, the VSGNE study also categorized surgeons into those who performed CI routinely or selectively, which was not done in this study. Nevertheless, there was no difference in the VSGNE data in discharge primary patency or 1-year primary patency between routine imagers and selective imagers.
One of the major limitations of this study is that the findings of CI, when it is performed, are not recorded in the VQI. Similarly, the interventions performed for abnormal CI findings are not recorded. Thus, it is unknown what the rate of significant findings on CI is and whether these findings spurred intraoperative intervention. Had these data been collected, it would have potentially enabled analysis of outcomes in grafts with abnormal findings on CI that were not corrected or residual abnormal findings after correction. There are several other limitations of this study. VQI data are self-reported and subject to the biases and errors of self-reporting. Ideally, a randomized controlled study would be performed to definitively answer the question of the efficacy of CI for LEB. However, given the cost and scope associated with such a study, it is unlikely that it will be performed. The follow-up rate at 59% is also less than ideal. Follow-up data were entered on only 984 patients in the CI group and 873 patients in the no CI group; based on the sample size analysis to detect a 5% difference in 1-year primary patency, these numbers are insufficient. Thus, it is possible that the lack of significant difference in 1-year primary patency between the two groups is related to small sample size. Ultimately, this is not a randomized controlled study and is vulnerable to surgeon bias and unmeasured patient and bypass graft bias. Whereas the results of this study did not demonstrate a patency benefit of CI, this does not indicate that CI does not have value. CI remains an important educational tool. Performance of CI gives the surgeon immediate feedback on the technical adequacy of the operation performed. Completion duplex ultrasound, in particular, allows the trainee to gain experience in performing duplex ultrasound scanning and interpreting duplex ultrasound images. CONCLUSIONS In electively performed LEBs using single-segment saphenous vein to a below-knee target artery for occlusive disease, CI does not influence primary graft patency at discharge or at 1 year. However, CI remains a valuable educational tool that provides immediate feedback to the surgeon. CI may provide a greater patency benefit for bypasses performed with disadvantaged conduit. Future studies should focus on the outcomes of CI in the disadvantaged conduit population. AUTHOR CONTRIBUTIONS Conception and design: KW, OP, FW, VR Analysis and interpretation: KW Data collection: KW Writing the article: KW, OP Critical revision of the article: KW, OP, FW, VR Final approval of the article: KW, OP, FW, VR Statistical analysis: KW Obtained funding: Not applicable Overall responsibility: KW
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Woo et al 5
REFERENCES 1. Nehler MR, Duval S, Diao L, Annex BH, Hiatt WR, Rogers K, et al. Epidemiology of peripheral arterial disease and critical limb ischemia in an insured national population. J Vasc Surg 2014;60: 686-95. 2. Goodney PP, Beck AW, Nagle J, Welch HG, Zwolak RM. National trends in lower extremity bypass surgery, endovascular interventions and major amputations. J Vasc Surg 2009;50:54-60. 3. Bandyk DF, Schmitt DD, Seabrook GR, Adams MB, Towne JB. Monitoring functional patency of in situ saphenous vein bypasses: the impact of a surveillance protocol and elective revision. J Vasc Surg 1989;9:286-96. 4. Idu MM, Blankenstein JD, de Gier P, Truyen E, Buth J. Impact of a color-flow duplex surveillance program on infrainguinal graft patency: a five-year experience. J Vasc Surg 1993;17:42-53. 5. Renwick S, Royle JP, Martin P. Operative angiography after femoropopliteal arterial reconstruction: its influence on early failure rate. Br J Surg 1968;55:134-6. 6. Mills JL, Fujitani RM, Taylor SM. Contribution of routine intraoperative completion arteriography to early infrainguinal bypass patency. Am J Surg 1992;164:506-11. 7. Scali ST, Beck AW, Nolan BW, Stone DH, De Martino RR, Chang CK, et al. Completion duplex ultrasound predicts early graft thrombosis after crural bypass in patients with critical limb ischemia. J Vasc Surg 2011;54:1006-10. 8. Rzucidlo EM, Walsh DB, Powell RJ, Zwolak RM, Fillinger MF, Schermerhorn ML, et al. Prediction of early graft failure with
9. 10.
11.
12.
13.
14.
15.
intraoperative completion duplex ultrasound scan. J Vasc Surg 2002;36:975-81. Dalman RL, Harris EJ, Zarins CK. Is completion arteriography mandatory after reversed-vein bypass grafting? J Vasc Surg 1996;23:637-44. Schanzer A, Hvelone N, Owens CD, Belkin M, Bandyk DF, Clowes AW, et al. Technical factors affecting autogenous vein graft failure: observations from a large multicenter trial. J Vasc Surg 2007;46:1180-90. Harrell FE, Lee KL. Verifying assumptions of the Cox proportional hazards model. In: Proceedings of the Eleventh Annual SAS Users Group International Conference. Cary, NC: SAS Institute Inc; 1986. p. 823-8. Mills JL. Infrainguinal disease: surgical treatment. In: Cronenwett JL, Johnston KW, editors. Rutherford’s vascular surgery. 8th ed. Philadelphia: Elsevier Saunders; 2014. p. 1758-81. Johnson BL, Bandyk DF, Back MR, Avino AJ, Roth SM. Intraoperative duplex monitoring of infrainguinal vein bypass procedures. J Vasc Surg 2000;31:678-90. Woo K, Palmer OP, Weaver FA, Rowe VL. Use of completion imaging during infrainguinal bypass in the Vascular Quality Initiative. J Vasc Surg 2015. [Epub ahead of print]. Tan TW, Rybin D, Kalish JA, Doros G, Hamburg N, Schanzer A, et al. Routine use of completion imaging after infrainguinal bypass is not associated with higher bypass graft patency. J Vasc Surg 2014;60: 678-85.
Submitted Dec 3, 2014; accepted Mar 17, 2015.