Prevalence and risk factors for heparin-bonded expanded polytetrafluoroethylene vascular graft infection after infrainguinal femoropopliteal bypasses

Prevalence and risk factors for heparin-bonded expanded polytetrafluoroethylene vascular graft infection after infrainguinal femoropopliteal bypasses Gabriele Piffaretti, MD, PhD,a Walter Dorigo, MD,b Paolo Ottavi, MD,c Raffaele Pulli, MD,d Ruth L. Bush, MD,e Patrizio Castelli, MD,b and Carlo Pratesi, MD,a on behalf of the PROPATEN Italian Registry Group, Varese, Florence, Terni, and Bari, Italy; and Houston, Tex

ABSTRACT Background: To analyze the prevalence and predictors of prosthetic vascular graft infection (PVGI) in a multicenter registry. Methods: This registry-based, multicenter study retrospectively evaluated PVGI that developed after infrainguinal revascularization performed with a heparin-bonded expanded polytetrafluoroethylene graft that was used in 1400 interventions between 2002 and 2016. A prosthetic graft with infection was defined as direct involvement of the graft with positive bacterial cultures of graft or perigraft material, intraoperative gross purulence or failure of graft incorporation, or exposed graft in an infected wound. Results: Critical limb ischemia (CLI) was the main indication for bypass (n ¼ 915 [65%]). The median duration of follow-up was 29 months (range, 1-168 months; interquartile range, 12-60 months). A total of 33 heparin-bonded expanded polytetrafluoroethylene grafts (2.3%) became infected; the median time to occurrence was 5 months (range, 1-54 months; interquartile range; 2.00-13.25 months). Freedom from PVGI at 1 year was 98% (standard error, 0.4; 95% confidence interval [CI], 97.2-98.9), and 97% (standard error, 0.6; 95% CI, 95.6-98.0) at 5 years. The multivariate model identified CLI (P ¼ .042; hazard ratio, 0.39; 95% CI, 0.164-0.969) to be independently associated with PVGI. In-hospital mortality of PVGI treatment was 12% (n ¼ 4/33). Freedom from major amputation was significantly different between patients with PVGI and those who did not experience this complication (at 1 year, 67.0% vs 88.5%; Log-rank c2 ¼ 22.5; P ¼ .001). Conclusions: In our “real-world” multicenter experience the prevalence of PVGI after infrainguinal femoropopliteal bypasses was relatively low at 2.3%, but still associated with significant mortality and limb loss. CLI was the only significant predictor of PVGI. This conclusion is reasonable; however, more comprehensive data are required to confirm these findings, because the presence of ischemic ulcers or gangrene was not predictive of PVGI. (J Vasc Surg 2019;-:1-9.) Keywords: Femoropopliteal bypass; Prosthetic vascular graft infection; Heparin-bonded ePTFE graft

By 1965, the prominent surgeon H. Haimovici had already recognized that infection of a prosthetic vascular graft was a major complication of catastrophic proportions.1 Despite optimization of preoperative protocols, improvement of technical skills, and antibiotic prophylaxis, prosthetic vascular graft infection (PVGI) after

From the Vascular Surgery, Department of Medicine and Surgery, Circolo University Teaching Hospital, University of Insubria School of Medicine, Varesea; the Vascular Surgery, Department of Cardiothoracic and Vascular Surgery, Careggi University Teaching Hospital, University of Florence School of Medicine, Florenceb; the Vascular Surgery, Cardiothoracic and Vascular Department, Santa Maria Hospital, Ternic; the Vascular Surgery, Department of Cardiothoracic Surgery, University of Bari School of Medicine, Barid; and the University of Houston, College of Medicine, Houston.e Author conflict of interest: none. Additional material for this article may be found online at www.jvascsurg.org. Correspondence: Gabriele Piffaretti, MD, PhD, Vascular Surgery, Department of Medicine and Surgery, Circolo University Teaching Hospital, University of Insubria School of Medicine, Via Guicciardini, 9, 21100 Varese, Italy (e-mail: [email protected]).

infrainguinal revascularization is still a limb-threatening complication. The prevalence of PVGI is reported in the range of 3.8% to 29.3% in single-center experiences, with a resulting limb loss rate of up to 60%.2-7 Identifying significant predictors of PVGI may improve the stratification of perioperative risk and lead to better operative planning. At present, most recent studies have focused on predictors for surgical site infection (SSI).8-12 Very few studies specifically analyzed the risk factors for PVGI after infrainguinal revascularization with prosthetic femoropopliteal bypass graft, and most report a single-center experience with few cases.3,7,13 The aim of this study was to analyze the prevalence of PVGI after femoropopliteal bypasses performed with a bioactive heparin-bonded expanded polytetrafluoroethylene (HB-ePTFE) graft. Further, potential predictors of PVGI are investigated, as are the results of its ensuing treatment in a large cohort of patients with peripheral arterial obstructive disease from a multicenter registry.

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 2019 by the Society for Vascular Surgery. Published by Elsevier Inc. https://doi.org/10.1016/j.jvs.2019.03.023

METHODS Patients population, study design, and cohort of patients. The Italian Propaten Registry is a multicenter, financially unsupported, single-device registry that 1

2

Journal of Vascular Surgery

Piffaretti et al

---

enrolled 1916 patients who underwent infrainguinal bypasses over a 14-year period ending in March 2016 (Table I). During the study period, 1400 procedures were performed using the Gore Propaten graft (W. L. Gore & Associates, Flagstaff, Ariz). In the same period of time, the same centers performed 516 infrainguinal bypasses with autologous saphenous vein (ASV), including 509 belowknee femoropopliteal and tibial bypasses, all for revascularization in patients with critical limb ischemia (CLI). The registry involves seven Italian centers, each of which has 800 or more overall beds and performs more than 150 cases per year of infrainguinal revascularization procedures. Beginning in June 2002, a HB-ePTFE graft was used in selected patients undergoing infrainguinal bypasses. Methods for graft selection, insertion, data collection, procedural details, and follow-up have been already elsewhere.14,15 Data concerning these interventions were prospectively recorded in the certified institutional registry at each participating center. Starting from January 2008, the data from each center were merged and maintained at the coordinator center, tracking the main anatomic, clinical, diagnostic, and technical variables. From then on, all data were prospectively collected in the final merged database. This database also contains perioperative (<30 days) results and all relevant clinical and diagnostic data collected during follow-up. The registry was approved by local ethical committee of each center. All subjects gave informed consent to the treatment of their personal data. The reliability of the data contained in the registry was certified by an external independent commission (Castalia Group, ICT and Quality Management; Aosta, Italy) in two different sessions (2009 and 2013). A retrospective analysis of the registry was performed and data concerning the subgroup of patients who experienced a PVGI during follow-up were collected. For the final analysis in the present study, the end of study period was March 1, 2016. Index intervention: Patient preparation, surgical details, and follow-up protocol. At the index intervention, patient skin shaving was done with electric razors. All patients received antibiotic prophylaxis starting 60 minutes before skin incision. Generally, in elective cases with no tissue loss, prophylaxis was short-term cefazolin (2 g twice daily; Cefamezin, Pfizer, Milan, Italy). In the presence of an infected ulcer or gangrene at the time of the index intervention, generally we proceeded with bypass intervention after an initial toe amputation and/or incision to decompress the infection and facilitate drainage from all affected anatomic spaces.16-18 Subsequent formal toe or forefoot amputations were delayed 4 to 10 days after bypass to maximize tissue reperfusion and allow the clear demarcation of marginal areas. At each center, followup was always performed at 1 and 12 months and on

2019

ARTICLE HIGHLIGHTS d

d

d

Type of Research: Multicenter, retrospective, registry-based cohort study Key Findings: Infrainguinal revascularization using heparin-bonded expanded polytetrafluoroethylene bypass graft in 1400 patients resulted in a prosthetic vascular graft infection (PVGI) rate of 2.3%, with a perioperative mortality of 12% and freedom from PVGI-related limb loss of 56% at 1 year. Take Home Message: Femoropopliteal bypasses using a heparin-bonded expanded polytetrafluoroethylene bypass graft have an acceptably low PVGI, but are still associated with significant mortality and risk of limb loss.

a yearly basis thereafter. Follow-up visits consisted of clinical examination, ankle-brachial index measurements, and duplex ultrasound examination. During ultrasound examinations, the patency of the bypass and the status of the inflow and outflow arteries were assessed. If symptoms suggested a recurrent femoropopliteal obstruction, computed tomography angiography and/or digital subtraction angiography were used to confirm and/or eventually guide treatment. PVGI management. All patients with a clinical suspicion of PVGI were evaluated with full panel blood tests, including inflammatory markers, bloodstream and urinary tract cultures, and computed tomography angiography. At admission, broad-spectrum antibiotics were generally started using an association of glycopeptide (Vancotex, Pharmatex, Milano, Italy) and penicillin/beta-lactamase inhibitors (Textazo, Pharmatex). Thereafter, they were replaced by other selective antibiotics on the basis of microbiological findings. An infectious disease physician specialist evaluated each patient at admission and regularly during the entire hospitalization to optimize the type, dosage, and duration of antibiotic therapy. In all cases, the infected graft was totally excised and periprosthetic tissues were aggressively debrided to obtain macroscopically normal tissues. All specimens were cultured for aerobes, anaerobes, and fungi. Generally, redo revascularization was performed in a two-stage fashion: after prosthetic graft removal, redo revascularization was performed with the ASV if available. A second-line alternative was a an extra-anatomic prosthetic graft implantation because cryopreserved human allografts were not always available at each center. Primary amputation was indicated for patients with a deemed unsalvageable limb owing to the absence of a targetable limb vessel.18 Postoperatively, the duration of antibiotic therapy varied according to joint clinical evaluation with an infectious disease physician specialist.

Journal of Vascular Surgery Volume

-,

Number

Piffaretti et al

3

-

Table I. Clinical indication and type of bypass of the 1916 patients enrolled in the Propaten registry Propaten (n ¼ 1400)

AGSV (n ¼ 516)

IC

450 (23.5)

42 (2.2)

CLI

950 (49.6)

474 (24.7)

AK

362 (18.9)

6 (0.3)

BK

755 (39.4)

162 (8.4)

Tibial

283 (14.8)

348 (18.2)

P value <.001

Clinical stage

<.001

Anatomic level

AGSV, Autologous great saphenous vein; AK, above-the-knee; BK, below-the-knee; CLI, critical limb ischemia; IC, intermittent claudication. Values are presented as number (%).

Definitions and outcomes. Comorbidities, risk factors, and clinical status were defined as previously described.19,20 PVGI was defined as direct involvement of the prosthetic graft/underlying artery with positive bacterial cultures of graft or perigraft material, intraoperative gross purulence or failure of graft incorporation, or an exposed graft in an infected wound. The depth of infection and degree of graft involvement was defined accordingly to Samson’s modified classification.21 Considering time of appearance, infection was defined early when it occurred less than 4 months after the index intervention and late when it was detected more than 4 months after the intervention.18,22,23 Duration of intervention was classified according to the cutoff risk defined by the National Nosocomial Infection Surveillance system for vascular surgery.5,24 In this analysis, we calculated PVGI prevalence and PVGI-related major amputation and mortality rate. The follow-up index (FUI) for late survival in the study group was defined as the ratio between the investigated follow-up period, and the theoretically possible follow-up period up to December 2016.25 Statistical analysis. Clinical data were recorded prospectively and tabulated in Microsoft Excel (Microsoft Corp, Redmond, Wash) database; statistical analysis was performed by means of SPSS 24.0 for Windows (IBM SPSS Inc., Chicago, Ill). Categorical variables were presented using frequencies and percentages, continuous variables were presented with mean 6 standard deviation or median and ranges, on the basis of data distribution. Continuous variables were analyzed with the c2 test and Fisher’s exact test, when necessary. Independent samples Student’s t-test was used for continuous variables, and the Wilcoxon signed-rank test was used to evaluate the difference in ankle-brachial index measurement before and after intervention. Follow-up data were analyzed by life-table analysis (Kaplan-Meier test). A univariate analysis to identify potential significant predictors of PVGI was performed with Kaplan-Meier

survival estimates 6 standard error (SE), and the logrank test for each covariate. Associations that yielded a P value of less than .20 on univariate screen were then included in a forward Cox regression analysis. The strength of the association of variables with PVGI was estimated by calculating the hazard ratio and 95% confidence intervals (Cis; significance criteria of 0.25 for entry and 0.05 for removal). All reported P values were 2two sided; a P value of less than .05 was considered significant.

RESULTS General data and surgical details. Overall, 1400 infrainguinal bypasses was performed with HB-ePTFE, of which 309 (22%) were for a failed previous ipsilateral open or endovascular femoropopliteal intervention. CLI was present in 915 patients (65%), and 485 (35%) had life-limiting intermittent claudication (IC). Groin incision was performed in 1294 patients (92%) and the common femoral artery was the most frequently used inflow vessel for the bypass (n ¼ 1274 [91%]). Distal anastomosis was performed on the above-the-knee (AK) popliteal artery in 364 patients (26%), on the below-the-knee (BK) popliteal artery in 753 (54%), and on a tibial vessel in 283 (20%). In patients with tissue loss, defined as the presence of ischemic ulcers or gangrene (n ¼ 469 [33.5%]), location of the distal anastomosis was performed on a BK vessel in 381 patients (81%). Adjunctive procedures at the distal anastomosis were performed in 243 patients (17%) in the form of vein cuff (n ¼ 150) or patch (n ¼ 93). Duration of intervention was less than 3 hours in 1064 patients (76%), 3 to 5 hours in 211 (15%), and more than 5 hours in 125 (9%). The duration of the intervention was significantly different when stratified for the type of bypass (AK, 137 6 63 minutes vs BK, 165 6 77 minutes vs tibial, 224 6 119 minutes; P < .001). Demographic data, comorbidities, and risk factors did not differ significantly between those who developed PVGI and those who did not experience such complication (Table II). Early outcomes, survival, and bypass outcome after the indexed intervention. Early mortality (<30 days) was 2% (n ¼ 27). The ankle-brachial index improved significantly after revascularization (preoperative, 0.36 6 0.27 [interquartile range (IQR), 0.03-0.50] vs postoperative, 0.79 6 0.72 [IQR, 0.65-1.0]; P < .001). Early thromboses (<30 days) occurred in 66 patients (5%). There was no significant difference between patients who developed PVGI and those who did not experience such complication in terms of site of distal anastomosis (AK vs BK vs tibial; odds ratio [OR], 2.17, P ¼ .337), duration of intervention was not significantly different (<3 vs 3-5 hours vs >5 hours, OR, 1.76 [P ¼ .414]), hospitalization (<5 days vs 5-10 days vs >10 days, OR, 2.43 [P ¼ .296]), and early graft thrombosis (6% vs 5%; OR, 0.12, P ¼ .670). All patients who survived the operation entered the follow-up: the

4

Journal of Vascular Surgery

Piffaretti et al

---

Table II. Demographic data, comorbidities, and risk factors Infected (n ¼ 33)

Variable

Noninfected (n ¼ 1367)

P OR value

Demographic Gender

0.93 .489

Male

17

1089

Female

6

278

72 6 10

73 6 9

Age

0.74 .720

Comorbidities Fem-pop reintervention

9 (27)

301 (22) 3 (42)

0.52 .523

Diabetes

17 (51.5)

Active smoking

16 (48)

875 (64)

3.53 .067

1.16

Hypertension

27 (82)

1217 (89)

0.49 .448

Dyslipidemia

26 (79)

861 (63)

3.29 .097

Ischemic heart disease

15 (45)

547 (40)

0.34 .593

Chronic renal failurea

1 (3)

191 (14)

3.26 .074

Hemodialysis

0 (0)

41 (3)

1.02 .622

CLI

26 (79)

875 (64)

2.90 .099

Tissue loss

15 (45)

451 (33)

2.17

11 (33)

492 (36)

0.26 .845

Poor run-off

Table III. Follow-up mortality (entire cohort; N ¼ 1400) Cause of death

No. (%)

Cardiac

143 (10)

Cancer

45 (3)

Sepsis

19 (1.5)

Respiratory

15 (1.0)

Stroke

9 (0.6)

Aorta related

3 (0.2)

Other

73 (5)

Unknown

40 (3)

.290

Risk factors

b

2019

.190

CLI, Critical limb ischemia; Fem-pop, femoropopliteal; OR, odds ratio; Tx, treatment. Values are presented as mean 6 standard deviation or number (%) unless otherwise indicated. a Creatinine >2 g/dL. b One or fewer tibial vessels.

median duration of follow-up was 29 months (range, 1-168 months; mean, 36 months; IQR, 12-60 months): 1389 patients (99%) had at least one postoperative follow-up evaluation. The median cumulative FUI for survival was 0.75 6 0.25. During the follow-up, 347 deaths (25%) were recorded. Causes of death are reported in Table III. Estimated survival was 92% (SE, 0.008; 95% CI, 90-93) at 1 year, 68.5% (SE, 0.02; 95% CI, 65-72) at 5 years, and 44% (SE, 0.03; 95% CI, 38-50) at 10 years (Fig 1). Survival was not significantly impaired by PVGI (c2 ¼ 0.311; log-rank P < .577). Primary graft patency rates were 78% at 1 year (SE, 0.01), 50.5% at 5 years (SE, 0.02), and 39.5% at 10 years (SE, 0.03); the corresponding figures in terms of limb preservation were 88% (SE, 0.02), 77% (SE, 0.02), and 69% (SE, 0.03), respectively. Either survival (log-rank: c2 ¼ 42.5, P < .001) or limb loss (log-rank: c2 ¼ 7.1, P ¼ .008) was significantly worse in patients with CLI if compared with patients presenting with IC. Infection. During the study period, a total of 33 HBePTFE grafts (2.3%) became infected. Early infection was detected in 16 cases (48%), of which 7 (21%) developed within the first postoperative month. The median time of appearance was 5 months (range, 1-54 months;

IQR; 2.00-13.25 months). Accordingly to Samson’s classification of infection depth and degree of graft involvement, type 3 (n ¼ 20), type 4 (n ¼ 4), and type 5 (n ¼ 9) cases were observed. Eleven infected grafts (33%) eventually thrombosed; the cumulative HB-ePTFE graft thrombosis was not significantly higher in infected grafts when compared with noninfected grafts (33% vs 38%; P ¼ .593). Freedom from PVGI at 1 year was 98% (SE, 0.4; 95% CI, 97.2-98.9) and 97% (SE, 0.6; 95% CI, 95.6-98.0) at 5 years (Fig 2). The univariate screen identified several variables associated with PVGI (Table IV), but at the multivariate model only CLI was the only predictor for PVGI development. Bacteriology. Microorganisms were identified in all cases: in 28 cases (86%), two or more microorganisms were isolated. Among microorganisms, methicillinresistant Staphylococcus aureus was isolated in 23 cases (71%), followed by Escherichia coli in 19 (57%), and Pseudomonas aeruginosa in 9 (28%); fungi were isolated in only one case (5%; Candida albicans). When systemic sepsis was present, we observed a perfect correlation between graft and blood cultures. PVGI treatment and outcomes. The infected graft was explanted in all patients. Eleven patients (33%) underwent redo BK revascularization with autologous GSV (n ¼ 6) or cryopreserved human allograft (n ¼ 5). Fifteen patients (45%) did not undergo revascularization, either because they were asymptomatic (n ¼ 11) or owing to the absence of adequate distal target vessel (n ¼ 4, profundaplasty). Seven patients (21%) underwent major amputation (AK, n ¼ 6; BK, n ¼ 1) because the limb was not salvageable secondary to the extent of infection. Early (<30 days) mortality of operative management of PVGI was 12% (n ¼ 4/33), with causes of death being sepsis (n ¼ 3) and acute coronary syndrome (n ¼ 1). After treatment, 5 additional patients (15%; asymptomatic, n ¼ 3; redo bypass, n ¼ 2) required major amputation, because of severe ischemia (n ¼ 3) or ascending lower limb infection (n ¼ 2). Freedom from major amputation was significantly different between patients with PVGI and those who did not experience

Journal of Vascular Surgery Volume

-,

Number

Piffaretti et al

5

-

100

Overall survival, (%)

80

60

40

20

0 0

12

24

36

48

60

72

84

96

108

120

Time from intervention, (months)

1400

1077

814

629

474

349

237

169

122

77

51

98

92

85.5

80

74

68.5

62.5

60

54

50

44

S.E.

0.04

0.8

1.0

1.3

1.5

1.7

1.9

2.1

2.4

2.7

3.2

95%CI

97-98

90-93

83-87

77-82

71-77

65-72

59-66

56-64

49 -59

45 -53

38 -50

No. at risk Survival, (%)

Fig 1. Kaplan-Meier estimate of survival for the entire cohort. CI, Confidence interval; S.E., standard error.

this complication (Fig 3). In particular, PVGI-related freedom from major amputation was 56% (SE, 0.9; 95% CI, 39-72) at 1 year.

DISCUSSION Although randomized clinical trials represent the benchmark method to perform research studies, prospective registries remain a valid alternative research method to obtain prompt real-world data on the safety, efficacy, and durability of a specific treatment with newer techniques or for rare pathologies.26 As such, in the present analysis, we evaluated the risk factors and outcomes of PVGI in a large cohort of patients undergoing infrainguinal femoropopliteal bypass using an HB-ePTFE graft. We consider ASV the material of choice for infrainguinal bypass in the BK and tibial vessels, particularly in patients with CLI. Although one can wonder at the high number of prosthetic grafts, these data warrant some comments. First, in the same period of time, 516 infrainguinal ASV bypasses were performed in the same centers, and the number of BK performed for CLI with HB-ePTFE was similar to that of ASV bypasses (521 and 509 interventions, respectively). Second, the use of alternative autologous conduits over contemporary prosthetic conduits in patients with CLI has been already shown that it may not offer a significant patency

advantage in midterm follow-up, either with arm veins or biologic conduits.27-30 Last, in the AK setting for patients with IC, a large proportion of interventions are performed using a prosthetic graft even in the presence of an available ASV, not only in Italy but also in other European countries.27,31 There is no doubt that PVGI is one of the most dreaded complications after infrainguinal bypasses. A recent analysis of the American College of Surgeons National Quality Improvement Program showed that infection was the most frequent cause of readmission after vascular surgery, especially after lower extremity bypasses, with occurrence rates as high as 40%.10 Unfortunately, PVGI has been specifically analyzed in numerically small single-center series, with reported prevalence rates was reported up to 29.3%, 49% of which were early (<6 months) PVGIs.4,5 In our cohort, the PVGI rate was 2.3%. As occurred in other studies, we acknowledge that follow-up is of utmost importance to evaluate this type of complication. However, although lower than previously reported, the catchment area of the participating centers is such that, for this type of complication, PVGIs are more easily referred back to the operating centers, and thus have better detection during the study period.5 Moreover, to minimize the impact of patients lost to follow-up after a certain period of time on late outcomes,

6

Journal of Vascular Surgery

Piffaretti et al

---

2019

Freedom from PVGI, (%)

100

80

60

40

20

0 0

12

24

36

48

60

72

84

96

108

120

Time from intervention, (months)

No. at risk

1400

1070

813

628

467

342

231

166

120

75

49

Survival, (%)

100

98.2

97.7

97.2

97.2

97

97

97

97

97

97

0

0.04

0.04

0.05

0.05

0.06

0.06

0.06

0.06

0.06

0.06

97- 99.

97-98

96 -98

96 -98

96-98

96 -98

96 -98

96 -98

96 -98

96 -98

S.E. 95%CI

Fig 2. Kaplan-Meier estimate of freedom from prosthetic vascular graft infection (PVGI). CI, Confidence interval; S.E., standard error.

we reported the value of the FUI in our study group. The FUI describes the follow-up completeness at a given study end date (specifically at March 1, 2016) as the ratio between the investigated and the potential follow-up period. It is a marker of accuracy of survival estimates. In the present study, our FUI was 0.75, which, even if not excellent, is a satisfactory value. Finally, it is noteworthy that 87% of patients had a follow-up of at least 6 months after prosthetic graft placement, which should have allowed the detection of PVGI in most cases. Very few studies have sought to identify predictors of PVGI.3,7,32 Although CLI was a reasonable predictor for PVGI, it could be considered a too generalized marker of PVGI.3 Although the incidence of rest pain among all patients treated is higher than previously reported for those presenting with CLI, when we stratified the analyses to assess whether tissue loss could be even more predictive of PVGIs, Rutherford’s 5 and 6 grades were not independent predictors for this complication.4-8 These data find support in previous papers. A nonstatistical trend of an increased risk of PVGI in the presence of tissue loss or skin ulcers at the time of prosthetic vascular graft placement has been reported in other studies as well.5-7 Further, although it may be prudent to avoid the use a prosthetic graft in the setting of tissue

loss, the ASV is not always available or suitable.13 In our experience, there was no increased risk of infection in the presence of tissue loss or gangrene. In our opinion, this is of relevance because it is a result of a series from multiple hospitals, with similar characteristics in terms of volume of patients, as well as expertise and type of procedures performed. As surgeons, we follow up with our patients soon after the operation with a thorough method of care that includes diligent wound management and surveillance.3,6,7,13Among those studies that described the predictors of PVGI, wound infection has been indicated to be a significant event, potentially associated with such a prosthetic graft complication.3,7,33 Although our study does not confirm it, this was not surprising: wound infection has been inconsistently reported in literature as a predisposing factor for PVGI in several other series.3,4,7 Additionally, a correlation between wound infection and subsequent PVGI has not been specifically mentioned in studies reporting on SSI.8-10 Previous studies have shown that PVGI may lead to high rate of limb loss. In particular, a recent large dataset analysis demonstrated a significant correlation of SSI (ie, including PVGI) and major amputation.9 Similarly, Siracuse et al7 reported that freedom from PVGI-related

Journal of Vascular Surgery Volume

-,

Number

Piffaretti et al

Table IV. Univariate analysis for prosthetic vascular graft infection (PVGI) during follow-up Variable

Infection rate, %

Age >80 y

Log-rank,

c

2

2.1

No

2.3

Gender

Table IV. Continued.

0.14

Infection rate, %

Variable

P value

.008

Yes

.927

.905

<5

3.5

5-10

0.8

>10

2.5

2.3

Yes

6.4

Female

2.1

No

2.2

3.74

Yes

3.9

No

2.0

Diabetes Yes

2.9

No

1.9

Active smoking Yes

1.7

No

3.3

Hypertension

1.83

.176

5.76

.016

2.2

No

3.1

IHD

.738

Yes

2.6

No

2.0

CRF

2.75

Yes

0.5

No

2.6

CLI 2.9

No

1.2

Tissue loss Yes

3.2

No

1.8

Run-off (1 vessel) Yes

2.0

No

2.0

Period of study 2002-2006

2.1

2007-2011

1.7

2012-2017

3.4

Distal anastomosis AK

1.6

BK

2.1

Tibial

3.5

Duration, hours <3

2.2

3-5

3.1

>5

5.3

.414

.390

.097

4.40

Yes

c2

Wound infection Yes

9.1

No

2.2

P value

3.50

.061

1.81

.179

AK, Above the knee; BK, below the knee; CLI, critical limb ischemia; CRF, chronic renal failure (creatinine >2 mg/dL); IHD, ischemic heart disease; LOS, length of stay.

100

.666

Yes

.154

.036

Freedom from major amputation, (%)

Previous treatment

Log-rank,

Perioperative thrombectomy

Male

LOS, days

7

-

80

60

40

20 no PVGI PVGI

0 0

Log-rank χ2 = 22.5, P = 0.001 12

24

36

Time from intervention, (months)

3.77

.052

0.02

.887

4.69

.193

2.99

.224

2.76

.251

4.53

.104

(Continued)

Fig 3. Kaplan-Meier estimate of freedom from major amputation in patients with or without prosthetic vascular graft infection (PVGI).

major amputation rate was 71% at 1 year. Our experience confirms this finding. Although lower than reported, PVGI was significantly associated with major amputation, no matter the method of treatment. The 61% freedom from major amputation was significantly lower when compared with that of patients who did not develop this complication. In particular, we observed a 56% freedom from PVGI-related major amputation rate. These data are confirmed by the fact that most patients had no reconstructive options at the time graft removal was required, or the affected limb was not salvageable because of the infection extent. Last, our 30-day mortality rate of 12% is another point of concern after PVGI. This finding is comparable with prior reports: Pounds et al4 reported that 11% of patients died as a result of PVGI occurrence. In a larger multicenter dataset, Gupta et al34 found that the mortality rate for unplanned readmission after infrainguinal bypass was significantly

8

Journal of Vascular Surgery

Piffaretti et al

---

higher in patients with infection-related readmission. All these data still support the idea that PVGI is one of the most feared complications in vascular surgery. Moreover, because sepsis was the dominant cause of postoperative death despite PVGI eradication, we believe operative risk optimization would be the most effective method to further decrease the occurrence this life-threatening complication. Limitations. This analysis has some limitations. It is a registry-based study, statistical analysis was retrospective, and the criteria for patient selection and postoperative treatment varied among the different centers. We acknowledge that a major criticism is also the limited follow-up, which poses a substantial risk of selection bias. Nevertheless, the incomplete follow-up was a described issue in other experiences. Although it is difficult to solve this problem, our follow-up was performed consistently beyond 2 years.

4.

5.

6.

7.

8.

9.

CONCLUSIONS In our large, real-world, multicenter experience, although the prevalence was relatively low at 2.3%, PVGI remained as a major concern after infrainguinal prosthetic bypass. At 1 year, the mortality rate was still high at 12%; freedom from major amputation was significantly lower in patients with such a complication. CLI was the only significant predictor of PVGI; although it is a reasonable marker, it should be used as a more comprehensive measure because the presence of ischemic ulcers or gangrene was not predictive of PVGI. We recommend meticulous care of wounds after prosthetic bypass, including surgical sites and close followup of patients after discharge.

10.

11.

12.

13.

14.

AUTHOR CONTRIBUTIONS Conception and design: GP, WD, RP Analysis and interpretation: GP, WD, RP, RB Data collection: GP, WD, PO, RP, PC, CP Writing the article: GP, WD, PO, RP, PC, CP, RB Critical revision of the article: GP, WD, PO, RP, PC, CP, RB Final approval of the article: GP, WD, PO, RP, PC, CP, RB Statistical analysis: GP, WD, RP Obtained funding: Not applicable Overall responsibility: GP GP and WD contributed equally to this article and share co-first authorship.

15.

16.

17.

REFERENCES 1. Hoffert PW, Gensler S, Haimovici H. Infection complicating arterial grafts: personal experience with 12 cases and review of the literature. Arch Surg 1965;90:427-35. 2. Mertens RA, O’Hara PJ, Hertzer NR, Krajewski LP, Beven EG. Surgical management of infrainguinal arterial prosthetic graft infections: review of a thirty-five-year experience. J Vasc Surg 1995;21:782-90. 3. Chang JK, Calligaro KD, Ryan S, Runyan D, Dougherty MJ, Stern JJ. Risk factors associated with infection of lower

18.

19.

2019

extremity revascularization: analysis of 365 procedures performed at a teaching hospital. Ann Vasc Surg 2003;17: 91-6. Pounds LL, Montes-Walters M, Mayhall CG, Falk PS, Sanderson E, Hunter GC, et al. A changing pattern of infection after major vascular reconstructions. Vasc Endovascular Surg 2005;39:511-7. Antonios VS, Noel AA, Steckelberg JM, Wilson WR, Mandrekar JN, Harmsen WS, et al. Prosthetic vascular graft infection: a risk factor analysis using a case-control study. J Infect 2006;53:49-55. Brothers TE, Robison JG, Elliott BM. Predictors of prosthetic graft infection after infrainguinal bypass. J Am Coll Surg 2009;208:557-61. Siracuse JJ, Nandivada P, Giles KA, Hamdan AD, Wyers MC, Chaikof EL, et al. Prosthetic graft infections involving the femoral artery. J Vasc Surg 2013;57:700-5. Wiseman JT, Fernandes-Taylor S, Barnes ML, Saunders RS, Saha S, Havlena J, et al. Predictors of surgical site infection after hospital discharge in patients undergoing major vascular surgery. J Vasc Surg 2015;62:1023-31. Davis FM, Sutzko DC, Grey SF, Mansour MA, Jain KM, Nypaver TJ, et al. Predictors of surgical site infection after open lower extremity revascularization. J Vasc Surg 2017;65: 1769-78. Hicks CW, Bronsert M, Hammermeister KE, Henderson WG, Gibula DR, Black JH 3rd, et al. Operative variables are better predictors of post discharge infections and unplanned readmissions in vascular surgery patients than patient characteristics. J Vasc Surg 2017;65:1130-41. Nguyen BN, Neville RF, Abugideiri M, Amdur R, Sidawy AN. The effect of graft configuration on 30-day failure of infrapopliteal bypasses. J Vasc Surg 2014;59:1003-8. Moreira CC, Leung AD, Farber A, Rybin D, Doros G, Siracuse JJ, et al. Alternative conduit for infrageniculate bypass in patients with critical limb ischemia. J Vasc Surg 2016;64:131-9. Kolakowski S Jr, Dougherty MJ, Calligaro KD. Does the timing of reoperation influence the risk of graft infection? J Vasc Surg 2007;45:60-4. Pulli R, Dorigo W, Castelli P, Dorrucci V, Ferilli F, De Blasis G, et al; Propaten Italian Registry Group. Midterm results from a multicenter registry on the treatment of infrainguinal critical limb ischemia using a heparin-bonded ePTFE graft. J Vasc Surg 2010;51:1167-77. Dorigo W, Pulli R, Piffaretti G, Castelli P, Griselli F, Dorrucci V, et al. Results from an Italian multicentric registry comparing heparin-bonded ePTFE graft and autologous saphenous vein in below-knee femoro-popliteal bypasses. J Cardiovasc Surg (Torino) 2012;53:187-94. Andros G, Lavery LA. Diabetic foot ulcers. In: Cronenwett JL, Johnston W, editors. Rutherford’s vascular surgery. 7th ed. Philadelphia: Saunders Elsevier; 2010. p. 1735-46. Schaper NC, Andros G, Apelqvist J, Bakker K, Lammer J, Lepantalo M, et al. Diagnosis and treatment of peripheral arterial disease in diabetic patients with a foot ulcer. A progress report of the International Working Group on the Diabetic Foot. Diabetes Metab Res Rev 2012;28(Suppl 1):218-24. Back MR. Local complications: graft infection. In: Cronenwett JL, Johnston W, editors. Rutherford’s vascular surgery. 7th ed. Philadelphia: Saunders Elsevier; 2010. p. 643-61. Dorigo W, Piffaretti G, Pulli R, Castelli P, Pratesi C; PROPATEN Italian Registry Group. A multicenter predictive score for amputation-free survival for patients operated on with an heparin-bonded ePTFE graft for critical limb ischemia. World J Surg 2017;41:306-13.

Journal of Vascular Surgery Volume

-,

Number

Piffaretti et al

9

-

20. Norgren L, Hiatt WR, Dormandy JA, Nehler MR, Harris KA, Fowkes FG; TASC II Working Group. Inter-Society Consensus for the Management of Peripheral Arterial Disease (TASC II). Eur J Vasc Endovasc Surg 2007;33(Suppl 1):S1-75. 21. Samson RH, Veith FJ, Janko GS, Gupta SK, Scher LA. A modified classification and approach to the management of infections involving peripheral arterial prosthetic grafts. J Vasc Surg 1988;8:147-53. 22. Inui T, Bandyk DF. Vascular surgical site infection: risk factors and preventive measures. Semin Vasc Surg 2015;28:201-7. 23. Gaynes RP, Culver DH, Horan TC, Edwards JR, Richards C, Tolson JS. Surgical site infection (SSI) rates in the United States, 1992-1998: the National Nosocomial Infections Surveillance System basic SSI risk index. Clin Infect Dis 2001;33(Suppl 2):S69-77. 24. von Allmen RS, Weiss S, Tevaearai HT, Kuemmerli C, Tinner C, Carrel TP, et al. Completeness of follow-up determines validity of study findings: results of a prospective repeated measures cohort study. PLoS One 2015;10:e0140817. 25. Lees T, Troëng T, Thomson IA, Menyhei G, Simo G, Beiles B, et al. International variations in infrainguinal bypass surgery a VASCUNET report. Eur J Vasc Endovasc Surg 2012;44: 185-92. 26. Nordon IM, Hinchliffe RJ, Holt PJ, Thompson MM, Loftus IM. Should the role of EVAR be re-evaluated in light of the 10 year results of EVAR-1? J Cardiovasc Surg (Torino) 2011;52: 179-87. 27. Peinetti F, Bellandi G, Cappelli A, Castagnola M, Dorigo W, Gargiulo M, et al. Patologia ostruttiva cronica aorto-iliaca e delle arterie degli arti inferiori. It J Vasc Endovasc Surg 2015;22(Suppl 2):25-68. 28. Farber A, Major K, Wagner WH, Cohen JL, Cossman DV, Lauterbach SR, et al. Cryopreserved saphenous vein

29.

30.

31.

32.

33.

34.

allografts in infrainguinal revascularization: analysis of 240 grafts. J Vasc Surg 2003;38:15-21. Albertini JN, Barral X, Branchereau A, Favre JP, Guidicelli H, Magne JL, et al. Long-term results of arterial allograft belowknee bypass grafts for limb salvage: a retrospective multicenter study. J Vasc Surg 2000;31:426-35. Avgerinos ED, Sachdev U, Naddaf A, Doucet DR, Mohapatra A, Leers SA, et al. Autologous alternative veins may not provide better outcomes than prosthetic conduits for below-knee bypass when great saphenous vein is unavailable. J Vasc Surg 2015;62:385-91. Lees T, Troëng T, Thomson IA, Menyhei G, Simo G, Beiles B, et al. International variations in infrainguinal bypass surgery a VASCUNET report. Eur J Vasc Endovasc Surg 2012;44:185-92. Schanzer A, Mega J, Meadows J, Samson RH, Bandyk DF, Conte MS. Risk stratification in critical limb ischemia: derivation and validation of a model to predict amputation-free survival using multicenter surgical outcomes data. J Vasc Surg 2008;48:1464-71. Tan TW, Kalish JA, Hamburg NM, Rybin D, Doros G, Eberhardt RT, et al. Shorter duration of femoral-popliteal bypass is associated with decreased surgical site infection and shorter hospital length of stay. J Am Coll Surg 2012;215: 512-8. Gupta PK, Fernandes-Taylor S, Ramanan B, Engelbert TL, Kent KC. Unplanned readmissions after vascular surgery. J Vasc Surg 2014;59:473-82.

Submitted Nov 5, 2018; accepted Mar 10, 2019.

Additional material for this article may be found online at www.jvascsurg.org.

9.e1

Piffaretti et al

Journal of Vascular Surgery ---

APPENDIX (online only). Collaborators. Alessandro Alessi Innocenti, Elena Giacomelli, Aaron Fargion Vascular Surgery, Department of Cardiothoracic and Vascular Surgery, Careggi University Teaching Hospital, University of Florence School of Medicine, Florence Giovanni De Blasis, Luciano Scalisi Vascular and Endovascular Surgery e “SS. Filippo e Nicola” Hospital Vincenzo Monaca, Giuseppe Battaglia Vascular Surgery e “Vittorio Emanuele” Ferrarotto Hospital

2019

Enrico Vecchiati, Giovanni Casali Vascular Surgery e “S. Maria Nuova” Hospital Fiore Ferilli, Raimondo Micheli, Francesco Grasselli, Paolo Bonanno Vascular Surgery e “S. Maria” Hospital Marco Franchin, Matteo Tozzi, Nicola Rivolta, Massimo Ferrario, Marco Franchin, Matteo Tozzi, Nicola Rivolta, Massimo Ferrario, Maria Cristina Cervarolo, Gaddiel Mozzetta, Emma Nahal Vascular Surgery, Department of Medicine and Surgery, Circolo University Teaching Hospital, University of Insubria School of Medicine, Varese, Italy