Defining utility and predicting outcome of cadaveric lower extremity bypass grafts in patients with critical limb ischemia

Abstract

Objective

Despite poor long-term patency, acceptable limb salvage has been reported with cryopreserved saphenous vein bypass (CVB) for various indications. However, utility of CVB in patients with critical limb ischemia (CLI) remains undefined. The purpose of this analysis was to determine the role of CVB in CLI patients and to identify predictors of successful outcomes.

Methods

A retrospective review of all lower extremity bypass (LEB) procedures at a single institution was completed, and CVB in CLI patients were further analyzed. The primary end point was amputation-free survival. Secondary end points included primary patency and limb salvage. Life tables were used to estimate occurrence of end points. Cox regression analysis was used to determine predictors of limb salvage.

Results

From 2000 to 2012, 1059 patients underwent LEB for various indications, of whom 81 received CVB for either ischemic rest pain or tissue loss. Mean age (±standard deviation) was 66 ± 10 years (male, 51%; diabetes, 51%; hemodialysis dependence, 12%), and 73% (n=59) had history of failed ipsilateral LEB or endovascular intervention. None had sufficient autogenous conduit for even composite vein bypass. Infrainguinal CVB (infrapopliteal target, 96%; n=78) was completed for multiple indications including Rutherford class 4 (42%; n =34), class 5 (40%; n =32), and class 6 (18%; n =15). Eleven (14%) had CLI and concomitant graft infection (n =8) or acute on chronic ischemia (n =3). Intraoperative adjuncts (eg, profundaplasty, suprainguinal stent or bypass) were completed in 49% (n =40) of cases. Complications occurred in 36% (n =29), with 30-day mortality of 4% (n =3). Median follow-up for CLI patients was 11.8 (interquartile range, 0.4–28.4) months with corresponding 1- and 3-year actuarial estimated survival (±standard error mean) of 84% ±4% and 62% ±6%. Primary patency of CVB for CLI was 27% ±6% and 17% ±6% at 1 and 3 years, respectively. Amputation-free survival was 43% ±6% and 23% ±6% at 1 and 3 years, respectively, and significantly higher for rest pain (59% ±9%, 36% ±10%) compared with tissue loss (31% ±7%, 14% ±7%; log-rank, P = .04). Freedom from major amputation after CVB for CLI was 57% ±6% and 43% ±7% at 1 and 3 years. Multivariable predictors of limb salvage for the CVB CLI cohort included postoperative warfarin (hazard ratio [HR], 0.4; 95% confidence interval [CI], 0.2–0.8), dyslipidemia (HR, 0.4; 95% CI, 0.2–0.9), and rest pain (HR, 0.4; 95% CI, 0.2–0.9). Predictors of major amputation included graft infection (HR, 3.1; 95% CI, 1.1–9.0).

Conclusions

In CLI patients with no autologous conduit and prior failed infrainguinal bypass, CVB outcomes are disappointing. CVB performs best in patients with rest pain, particularly those who can be anticoagulated with warfarin. However, it may be an acceptable option in patients with minor tissue loss or concurrent graft infection, but consideration should be weighed against the known natural history of nonrevascularized CLI and nonbiologic conduit alternatives, given potential cost implications.

The most important determinate of lower extremity bypass (LEB) outcomes in patients with critical limbischemia (CLI) is availability of autologous vein.^1–3 However, up to 30% of CLI patients present with inadequate or unavailable autogenous conduit.^4–6 In this scenario, a variety of alternative conduit choices exist, including prosthetic and nonautogenous biologic grafts. In 1955, Linton described the first clinical use of a homologous saphenous vein allograft,⁷ and reports in the 1970s highlighted the utility of cryopreserved saphenous vein bypass (CVB) for infrainguinal arterial reconstruction.⁸ Since that time, cryopreserved allografts have been implanted for numerous indications, the most common being limb-threatening ischemia and infection.^9–11

Several series have documented dismal 1-year primary patency (~25%–30%) of CVB in CLI patients, tempering enthusiasm for these conduits because prosthetic grafts have similar or better outcomes at relatively reduced cost.¹² Despite these results, acceptable 2-year limb salvage rates (42%–78%) for CVB in CLI have been reported, suggesting that these grafts are worthy of consideration in selected patients.¹³ Moreover, patients with overt lower extremity infection and no usable autogenous conduit may be a specific subpopulation in which CVB offers an advantage over prosthetic conduits, given their relative resistance to infection.¹¹ In addition, few studies characterize CVB outcomes in the context of recently proposed objective performance goals for CLI therapies.^14,15

Because of the lack of rigorous comparative effectiveness data, the utility of CVB in CLI patients remains undefined. The purpose of this analysis was to determine the role of CVB in CLI patients and to identify predictors of successful outcomes.

METHODS

Database and subjects

This study was approved by the Institutional Review Board at the University of Florida (IRB #201300025). A waiver of informed consent was granted because all collected data pre-existed in medical records and no study related interventions or subject contact occurred. Therefore, the rights and welfare of these subjects was not adversely affected. All patients who underwent infrainguinal bypass from 2000 to 2012 for any indication were retrospectively reviewed. Subjects receiving cryopreserved allografts were further analyzed. Patients receiving autogenous vein or prosthetic bypass were excluded from analysis. Patient demographics, prior vascular surgery history, noninvasive vascular laboratory testing, postoperative reintervention or reoperation, minor or major amputation, and use of statins, antiplatelets, and anticoagulants were recorded. Comorbidities and complications were recorded and defined on the basis of recommended reporting standards.¹⁶

Clinical practice

All subjects underwent preoperative digital subtraction angiography with bilateral upper and lower extremity vein mapping. Patients had multilevel occlusive disease for which endovascular salvage was deemed unfeasible by the treating surgeon. CVB was offered to patients only if they had insufficient autogenous conduit even for composite vein bypass. Our group’s philosophy on infrainguinal bypass, vein mapping assessment, and arteriogram interpretation has been previously reported. ^17,18 The hierarchy for conduit selection was ipsilateral greater saphenous vein > contralateral saphenous vein (if ankle-brachial index >0.6 in that limb), arm vein ≫ nonautogenous biologic or prosthetic grafts. Prosthetic bypass with a vein cuff¹⁹ was an alternative conduit used in a minority of subjects with inadequate autogenous vein, primarily those with good runoff and low risk of infection. Need for intraprocedural adjuncts, such as concomitant suprainguinal stent or bypass, femoral reconstruction or bypass, and distal anastomotic vein patch or digital/forefoot amputation, was left to the operating surgeon’s discretion.

Cadaveric saphenous vein was obtained at a cost of $7000 to $7500 per vessel. Grafts were stored in a proprietary tissue cryoprotectant solution containing Dulbecco’s Modified Eagle’s Medium with dimethyl sulfoxide, fetal bovine serum, and chondroitin sulfate at or below −135°C that did not change during the study interval (Cryograft; CryoLife, Kennesaw, Ga). Donor characteristics and storage times were verified by direct communication with the manufacturer. Grafts were thawed and prepared per instructions for use. Single-segment grafts were preferred; however, in selected cases, composite grafts were constructed with arterial or venous cadaveric products if any section had poor integrity (eg, thin-walled or aneurysmal sections, synechia, webs, or sclerotic segments).

Implantation technique

Grafts were dilated with a heparin/papaverine hydrochloride solution (60 mg papaverine hydrochloride, 5000 units heparin, and 4°C Plasmalyte). Early in the study interval, routine ABO/Rh blood group matching was performed; however, evolving literature bias^9,13,20 and conduit availability led to decreased prioritization of matching over the duration of this review. Patients were heparinized (80–100 units/kg initial bolus), and intraprocedural activated clotting time was generally maintained ≥250 seconds during reconstruction. On completion of the distal anastomosis, adequacy of the bypass was determined by angiography and the presence of graftdependent pedal pulses or Doppler signals. Selective use of distal anastomotic adjuncts (eg, vein patch¹⁹) was based on the surgeon’s preference, particularly for small (<2 mm) or diseased tibial targets requiring local endarterectomy. Protamine (1 mg/100 units heparin) was routinely administered to reverse the heparin effect at case completion.

Postoperative management

Nearly all patients were treated with postoperative antiplatelet therapy (aspirin unless a confirmed allergy existed), and as practice evolved, patients were also increasingly prescribed statin therapy. Anticoagulation with warfarin sodium (goal international normalized ratio of 2–3) was initiated at the operating surgeon’s discretion on the basis of the assessment of the likelihood of graft failure, determined by distal target artery quality and completion arteriography. No immunomodulatory therapy was used to augment graft patency.

Postoperative surveillance included ankle-brachial indices performed at 1 month, then at 3-month intervals in the first year, and then semiannually. Most patients also underwent duplex ultrasound examination according to the same schedule. CVB grafts were considered at risk for thrombosis if a 3.5× step-up in peak systolic velocity was discovered on postoperative duplex ultrasound examination, particularly when accompanied by a 0.15 reduction in ankle-brachial index. Type and timing of reintervention were left to the operating surgeon’s discretion. Patency from the date of initial operation to the date of thrombosis was determined by graft imaging (duplex ultrasound or arteriography) or the first documented onset of ischemic symptoms or reintervention. Thrombosis was verified by duplex ultrasound if no flow within the graft was detected by clinical examination.

Definitions

Minor amputations included digital and transmetatarsal resections; major amputation was defined as above the ankle. The degree of chronic limb ischemia was classified by the Rutherford scale.¹⁶ Objective performance goals proposed by the Society for Vascular Surgery^14,21 were selectively applied to better understand utility of CVB. Specifically, amputation-free survival (AFS) and major adverse limb events (MALEs) were analyzed.

To study AFS, our primary outcome was major amputation ipsilateral to the index cadaveric saphenous vein bypass or death from any cause, whichever occurred first. A second ipsilateral CVB was analyzed as a MALE and loss of patency. Contralateral CVBs were not analyzed. Therefore, 112 ipsilateral CVB allografts in 112 patients were included in this analysis. Each patient’s study period began on the index CVB date and continued until the outcome occurred or until August 1, 2013. Mortality events were verified with the Social Security Index Masterfile. Limb salvage was the time from CVB to ipsilateral major amputation. Primary patency was defined as time from implant free from graft thrombosis and reintervention. MALEs were recorded if the index CVB limb underwent major amputation or any major vascular reintervention (thrombectomy, thrombolysis, new bypass graft, jump/interposition graft revision).¹⁴

End points and statistics

The primary end point was AFS. Secondary end points included primary patency, limb salvage, and MALEs. Categorical variables were summarized with frequencies and percentages. Normally distributed continuous variables were analyzed with means and standard deviations; nonparametric data were examined by medians with interquartile range (IQR). Comparisons of characteristics in subgroup analysis were performed with Fisher exact test, two-sample t-test, Wilcoxon rank sum text, and analysis of variance, when appropriate.

Patency, patient survival, limb salvage, AFS, and freedom from MALEs were estimated by Kaplan-Meier life-table methodology and compared between subgroups by the log-rank test. Cox regression was used to develop a model to predict major amputation. Variables found on univariate analysis to have P < .2 were included in the multivariable model, which was refined with stepwise backward regression and log-likelihood ratio testing. All statistical analysis was performed with STATA 11 software (StataCorp, College Station, Tex). A P < .05 was considered significant.

RESULTS

Patient cohort

From 2000 to 2012, 1059 patients underwent LEB, of whom 112 (11%) had a total of 127 CVB grafts implanted. This cohort had a mean age (±standard deviation) of 66 ± 10 years, and 54% (n = 60) were male. Notably, 47% (n = 53) were diabetic and 9% (n = 10) were hemodialysis dependent. Additional details of patient demographics and comorbidities of the entire cohort as well as of the CLI patients are highlighted in Table I.

Table I.

Demographics and comorbidities

Characteristic	All (N = 112), No. (%)	CLI (n = 81), No. (%)
Age, years, mean ± SD	66 ± 10	66 ± 10
Male	60 (54)	41 (51)
Body mass index, kg/m2, mean ± SD	26 ± 6	25 ± 5
Comorbidities
Hypertension	106 (95)	77 (95)
Tobacco use (past or current)	81 (72)	56 (69)
Coronary artery disease	68 (61)	51 (63)
Dyslipidemia	77 (69)	47 (71)
Diabetes	53 (47)	41 (51)
Chronic obstructive pulmonary disease	31 (28)	26 (32)
Congestive heart failure	29 (26)	22 (27)
Cerebrovascular occlusive disease	25 (22)	18 (22)
Renal insufficiency (eGFR <50)	29 (26)	25 (31)
Dialysis dependent	10 (9)	10 (12)
Preoperative medications
Aspirin	97 (87)	73 (90)
Clopidogrel	24 (22)	19 (23)
Statin	62 (56)	47 (58)
Warfarin	28 (25)	22 (27)

CLI, Critical limb ischemia; eGFR, estimated glomerular filtration rate; SD, standard deviation.

A majority of patients (65%; n = 73) had a history of failed ipsilateral, infrainguinal leg bypass, and 72% (n = 81) had a history of failed leg bypass or endovascular intervention. A significant proportion (52%) of the patients had a preoperative American Society of Anesthesiologists score of 4, and 37% (n = 41) of the cases were performed urgently or emergently. Among all CVB patients (n = 112), Rutherford classification was class 3 in 3% (n = 3, all with graft infection), class 4 in 30% (n = 34), class 5 in 29% (n = 32), class 6 in 13% (n = 15), and nonapplicable in 25% (n = 28; 13 with acute ischemia, 12 with graft infection, and three with both). Of the 81 CLI patients, 11 (14%) had concomitant graft infection (n = 8) or acute on chronic limb ischemia (n = 3).

CLI patients

In focusing only on CLI patients (n = 81), mean (±standard deviation) preoperative ankle-brachial index was 0.27 ± 0.18, and median toe brachial index was 0 (IQR, 0–0.1). Intraoperative adjuncts were completed in 49% (n = 40) of cases. Operative details and specific donor characteristics of the various allografts used in CLI patients are detailed in Table II (and the Supplementary Table, for the entire cohort). Notably, valves were lysed in 36% (n = 29) of cases, and the common femoral artery was the most common site of inflow (94%); an infrapopliteal target artery was used in 96% of cases. ABO/Rh-matched grafts were implanted in 55% (n = 41) of cases, and most of the allografts came from younger, male donors with mean preservation time of 196 ± 143 days.

Table II.

Operative details and donor characteristics in critical limb ischemia (CLI) patients

Operative details	N = 81, No. (%)
Bypass configuration
Proximal anastomosis
External iliac artery	1 (1)
Common femoral artery	76 (94)
Superficial femoral artery	3 (4)
Profunda femoris artery	1 (1)
Distal anastomosis
Above-knee popliteal artery	3 (4)
Below-knee popliteal artery	7 (9)
Tibial artery	70 (86)
Tibioperoneal trunk	2 (2)
Posterior tibial artery	28 (35)
Anterior tibial artery	21 (26)
Peroneal artery	19 (23)
Dorsalis pedis artery	1 (1)
Conduit configuration
Reversed greater saphenous vein	43 (53)
Nonreversed greater saphenous vein	29 (36)
Composite saphenous vein	7 (9)
Composite vein/artery	2 (4)
Adjuncts (any)	40 (49)
Inflow procedure	28 (34)
Distal anastomotic adjunct	5 (6)
Donor characteristics	N = 73
Age, years ± SD	38 ± 13
Male	99% (72)
Preservation time, days ± SD	196 ± 143
ABO/Rh matched	41% (55)

SD, Standard deviation.

CLI patient outcomes

The 30-day mortality in CLI patients was 4% (n = 3), and complications occurred in 37% of cases (n = 29). The most frequent complication was wound morbidity in 22% (n = 18). Table III demonstrates the outcomes of all CLI patients undergoing CVB. Median follow-up time for CLI patients was 11.8 (IQR, 0.4–28.4) months, with corresponding 1- and 3-year actuarial estimated survival (±standard error mean) of 84% ± 4% and 62% ± 6% (Fig 1). Primary patency of CVB for CLI was 27%±6% and 17%±6% at 1 and 3 years, respectively (Fig 2, A). There were 24 patients (30%) who underwent revision of a failing or failed CVB. The 1- and 3-year primary-assisted patency was 39% ± 6% and 18% ± 7%, respectively (Fig 2, B). The types of reinterventions that were performed to remediate the CVB grafts are listed in Table IV. In the 70% of patients who underwent postoperative duplex ultrasound surveillance after CVB, reinterventions were found to be more frequent compared with those who did not (32% vs 15%; P = .06); however, no improvement in primary-assisted patency was noted (P = 1).

Table III.

Outcomes after cadaveric vein lower extremity bypass (LEB) in critical limb ischemia (CLI) patients

Outcome	N = 81
Thirty-day mortality	3 (4)
In-hospital mortality	3 (4)
Median length of stay, days (IQR)	7 (5–11)
Any complication	29 (37)
Wound/bleeding	18 (22)
Pulmonary	6 (7)
Cardiac	5 (6)
Gastrointestinal	5 (6)
Renal	4 (5)
Neurologic	3 (4)
Genitourinary	3 (4)
Postoperative medications
Aspirin	76 (94)
Statin	57 (70)
Warfarin	50 (62)
Clopidogrel or dipyridamole	22 (27)

IQR, Interquartile range.

Fig. 1.

This Kaplan-Meier curve depicts the overall survival for patients receiving cadaveric vein bypass for a critical limb ischemia (CLI) indication. Actuarial estimated survival at 1 and 3 years is 84% (95% confidence interval [CI], 74–90) and 62% (95% CI, 49–72), respectively. All displayed time intervals have <10% standard error of the mean.

Fig. 2.

A, Primary patency of cadaveric vein bypass in critical limb ischemia (CLI) patients was 27% (95% confidence interval [CI], 16–39) and 17% (95%CI, 7–30) at 1 and 3 years, respectively. B, The primary-assisted patency after cadaveric vein bypass for a CLI indication was not significantly different from the primary patency despite that a majority of patients underwent routine postoperative duplex ultrasound surveillance and had a reintervention threshold of a peak systolic velocity step-up of 3.5 × or drop in ankle brachial index ≥0.15. All displayed time intervals have <10% standard error of the mean.

Table IV.

Procedures performed to remediate failing or thrombosed cadaveric bypass in critical limb ischemia (CLI) patients

Procedure type	No.
Balloon angioplasty ± stent for critical stenosis	6
Redo bypass	5
Open graft revision	5
Ligation	3
Thrombectomy ± graft revision	2
Thrombolysis	1
Excision of aneurysmal degeneration	1

A total of 47 grafts (58%) thrombosed during the follow-up interval, and 36 CLI patients underwent major amputation (above knee; n = 31; 38%). Median time from initial graft thrombosis to major amputation was 0.6 (IQR, 0.1–3.1) months. Notably, eight of 36 patients underwent major amputation without experiencing graft failure. Nineteen patients experienced CVB thrombosis and did not subsequently undergo major amputation. The indications for this subset of patients included rest pain (n = 8), Rutherford class 5 (n = 4), Rutherford class 6 (n = 4), and mixed indications (n = 3). Most (n = 13) did not undergo a concomitant inflow procedure at the time of CVB.

AFS was 43% ± 6% and 23% ± 6% at 1 and 3 years, respectively, and significantly higher for rest pain (59% ± 9%, 36% ± 10%) compared with tissue loss (31% ± 7%, 14% ± 7%; log-rank, P=.04) (Fig 3), in contrast to primary patency and survival, which did not differ by indication. Freedom from major amputation after CVB for CLI was 57% ± 6% and 43% ± 7% at 1 and 3 years, respectively (Fig 4, A), and significantly higher for rest pain patients compared with those with tissue loss (Fig 4, B). Freedom from MALEs for all CLI patients was 47% ± 7% at 1 year and 25% ± 7% at 3 years (Fig 5, A) and again higher for a rest pain indication compared with tissue loss (Fig 5, B).

Fig. 3.

Amputation-free survival (AFS) was significantly different (log-rank, P = .04) between patients undergoing cadaveric vein bypass for a rest pain indication and subjects with tissue loss. All displayed time intervals have <10% standard error of the mean.

Fig. 4.

A, Freedom from major (above the ankle) amputation in all critical limb ischemia (CLI) patients is demonstrated on this graph. B, Similar to amputation-free survival (AFS), freedom from major amputation was significantly different (log-rank, P = .01) between rest pain and tissue loss patients. Notably, a significant proportion of patients in the rest pain group underwent concomitant femoral arterial or iliac inflow reconstruction at the time of their cadaveric vein bypass, which likely contributed to these findings. All displayed time intervals have <10% standard error of the mean.

Fig. 5.

A, Overall, estimated freedom from major adverse limb events (MALEs) after cadaveric vein bypass for a critical limb ischemia (CLI) indication was 47% (95% confidence interval [CI], 28–53) and 25% (95% CI, 12–39) at 1 and 3 years, respectively. B, Not surprisingly, primarily owing to significant differences in rates of major amputation after cadaveric vein bypass revascularization, significant differences in freedom from MALEs exist between rest pain and tissue loss patients (P = .004). All displayed time intervals have <10% standard error of the mean.

Multivariable predictors for limb salvage in CLI patients included postoperative warfarin (hazard ratio [HR], 0.4; 95% confidence interval [CI], 0.2–0.8), dyslipidemia (HR, 0.4; 95% CI, 0.2–0.9), and rest pain (HR, 0.4; 95% CI, 0.2–0.9). Predictors of major amputation included graft infection (HR, 3.1; 95% CI, 1.1–9.0).

DISCUSSION

This study analyzed the results of patients undergoing CVB for CLI and confirmed that these conduits have poor patency. Rates of MALEs were significant in this cohort of patients with insufficient autogenous conduit in which a majority had a history of prior failed ipsilateral infrainguinal revascularization. Despite the dismal patency, acceptable AFS was observed for the subset of patients with rest pain. Notably, in addition to an indication of rest pain, preoperative dyslipidemia and postoperative warfarin therapy independently predicted improved limb salvage, whereas operations performed for graft infection predicted the highest risk of subsequent ipsilateral major amputation. To our knowledge, this series represents one of the first to report multiple patient and periprocedural factors that are independent predictors of limb salvage after CVB, rather than patency, as well as to apply selected objective performance goals in an attempt to understand clinical efficacy. These data provide further insight into the utility of CVB in CLI patients and designate which patients may derive the most benefit from this revascularization strategy.

The merits of revascularization are well established, even in patients presenting with a failed ipsilateral LEB and CLI.^4,5,19,22 Autogenous vein is the preferred conduit for infrainguinal reconstruction, especially because many CLI patients have a disease pattern that is not amenable to endovascular therapy and require crural bypass for adequate revascularization.^16,21,23 CLI patients with insufficient autogenous vein require nonautogenous biologic or synthetic conduits, which have mixed, often poor, results.^{6,11,12,19,24} These outcomes have led some authors to question whether use of such conduits is futile and to advocate for earlier consideration of primary amputation.²⁵ Given that reported outcomes of CVB are frequently from high-risk CLI cohorts,^8,9,11,26 the inherent selection bias makes it challenging to determine which clinical scenarios, if any, are best treated with this conduit.

Despite discouraging CVB graft patency, acceptable limb salvage rates of 50% to 75% at 1 to 3 years have been reported.^9,13,26 Indeed, comparable rates were observed in our series of Rutherford class 4 patients. Historically, metrics of success after LEB have focused on graft patency, limb salvage, and death. More recently, outcomes focusing on freedom from MALEs, AFS, and functional outcome have been suggested as more fitting criteria for evaluating CLI therapies.^21,27,28 By use of this framework, outcomes of CVBs are sobering because 1-year freedom from MALE and AFS in our experience was 47% and 43%, respectively.

At first glance, these results appear dismal, although most of these patients were likely facing major amputation. Nearly 75% of the patients in this series presented with CLI or history of failed prior LEB, and none had sufficient autogenous conduit even for composite arm vein bypass. This subset of patients has a 60% to 80% risk of major amputation at 1 year, a 20% chance of death within 6 months, and a 15% risk of contralateral amputation within 2 years.²³ The rationale for attempting infrageniculate revascularization in this group of high-risk patients reflects the aggressive limb salvage philosophy within our group. An important factor that likely formed our bias in favor of crural bypass over primary amputation is the notably young mean age of the patients in our series at 66 ± 10 years compared with others studies.^9,29,30

Several factors have been reported to influence CVB results. Immunomodulation, ABO/Rh matching, degree of tissue loss, and certain patient-level factors like diabetes and renal failure have been variably linked to outcomes after CVB.^9,13,26 However, most reports have focused on understanding factors predicting CVB graft patency. Because graft patency is not the best indicator of clinical success,³¹ we chose to analyze the performance of CVB with respect to more patient-centered outcomes. We focused on limb salvage, as this was the goal of therapy. Limb salvage was found to be positively affected by a preoperative diagnosis of dyslipidemia. The seemingly protective effect of hyperlipidemia (HR, 0.4) has been reported in other series³¹ and may be a proxy for statin use or a patient cohort with improved access to health care.

Interestingly, postoperative warfarin anticoagulation also predicted improved limb salvage, which intuitively can be hypothesized to augment graft patency. However, primary patency was not significantly improved by warfarin therapy, which is consistent with other reports on CVB outcomes.^9,13 Patients in our practice managed with anticoagulation after LEB are typically younger, have better functional status with corresponding lower fall risk, and have sufficient compliance to adhere to anticoagulation surveillance, which could confound these outcomes. Presentation with rest pain was also predictive of improved AFS and could potentially be explained by two factors. First, the natural history of rest pain with or without successful revascularization is better than that for ischemic tissue loss.²³ Second, rest pain patients in this series were more likely to undergo a concomitant inflow procedure compared with tissue loss patients (50% vs 23%; P = .02), and this likely improved the eventual outcome of the bypassed limb.

A noteworthy distinction about this analysis is the incorporation of more donor- and procedure-specific information compared with other CVB series.^9,11,26 Donor age, gender, ABO/Rh characteristics, and preservation time were all evaluated but did not significantly affect outcomes. Distal anastomotic adjuncts, such as vein cuffs or arteriovenous fistulas, were seldom employed because they have not consistently been proved to improve results of cadaveric allograft or prosthetic crural bypasses.^9,13,32,33 Although the study was not designed to determine the utility of duplex ultrasound surveillance, it is notable that 70% of patients had postoperative duplex imaging. Patients undergoing surveillance seemed more likely to also undergo reinterventions (32% vs 15%; P = .06), but no improvement in primary-assisted patency was noted between patients evaluated with duplex ultrasound and those without surveillance (P = 1). Similar to polytetrafluoroethylene (PTFE) grafts, routine surveillance does not appear to reliably predict impending CVB failure, unlike that for autogenous bypasses.^13,34,35 In addition, secondary procedures for failing or failed CVB grafts were not common in this series, and their success rate did not significantly affect AFS (P = .3).

Given these marginal outcomes with CVB, logical comparisons for consideration are PTFE bypass, endovascular therapy, and primary amputation in CLI patients without autogenous vein. A meta-analysis by Albers et al³⁶ reported the 1- and 5-year primary patency rates of below-knee PTFE grafts to be 59% and 31%, respectively. However, these results may not reflect real-world experience, especially given the inherent publication bias that exists with these studies.³² Similarly, PTFE bypass limb salvage rates are comparable to or better than those of most CVB series,^9,13,36 but direct comparison between PTFE and CVB is difficult because prosthetic bypasses are unlikely to be used in patients with poor tibial runoff or overt infection. This point highlights a potential advantage of CVBs for use within infected fields when autogenous vein is unavailable because prosthetic conduits have yielded inferior results in this setting and are prone to infection in the presence of infected ulcers.³³ Another important consideration is that patients in need of a crural bypass often have femoral/popliteal artery occlusion precluding successful endovascular therapy, further limiting options in these patients with advanced disease.³⁷ Moreover, the negative impact of major amputation in CLI patients has been well documented,³⁸ making primary amputation also a poor option.

Treatment strategies for CLI have become increasingly complex owing to the advent of multiple endoluminal, open, and hybrid revascularization techniques. Accordingly, we concur with Taylor et al,³¹ who advocate for identification of patient cohorts at risk of failure regardless of treatment modality in the CLI patient population. Our current practice surrounding this unique subset of patients encompasses a thoughtful strategy that reflects a philosophy similar to that of Nehler et al,²⁵ which considers the patient’s ambulatory status, comorbidities, conduit status, degree of tissue loss, presence of infection, and complexity of the required revascularization.

In a good-risk, ambulatory patient with minimal pedal tissue loss, suitable tibial outflow, and no autogenous conduit, an aggressive attempt at revascularization with alternative conduits is attempted. However, in a nonambulatory patient with multiple advanced comorbidities or extensive tissue loss, primary amputation is typically considered. The results of this analysis suggest that CVB can be performed with efficacy similar to reported outcomes of alternative nonbiologic prosthetic conduits; however, costs are significant.¹² Bearing this in mind, younger, good-risk patients with active extremity infection and no usable autogenous vein may be the optimal group in which to attempt CVB despite the higher risk of major amputation.

The results of this analysis must be considered within the limitations of its retrospective, single-center study design. Small patient numbers and lack of direct, prospective comparison to alternative strategies, such as prosthetic bypass or primary amputation, make it difficult to fully define the utility of CVB in CLI. However, tremendous bias in selecting therapy exists for these heterogeneous groups of patients, which makes any comparison challenging to interpret. Importantly, no functional outcome data were provided to better define success or failure of CVB outside of limb loss. Further, data regarding return of rest pain symptoms and time to wound healing in patients with CVB failure were poorly documented. Finally, the natural history of medically managed CLI patients presenting after prior failed lower extremity intervention has been reported previously, but no similar group was available in this analysis for comparison.

CONCLUSIONS

In CLI patients with no autologous conduit and prior failed infrainguinal bypass, the clinical outcomes of CVB are disappointing. CVB performs best in patients with rest pain, particularly those who can be anticoagulated postoperatively with warfarin. However, it is also an acceptable option in patients with tissue loss or graft infection.