- Email: [email protected]
Current practice of first-line treatment strategies in patients with critical limb ischemia
Current practice of first-line treatment strategies in patients with critical limb ischemia Theodosios Bisdas, MD,a Matthias Borowski, PhD,b and Giovanni Torsello, MD,a for the First-Line Treatments in Patients With Critical Limb Ischemia (CRITISCH) Collaborators, Muenster, Germany Objective: Critical limb ischemia (CLI) is growing in global prevalence and is associated with high rates of limb loss and mortality. However, a relevant gap of evidence about the most optimal treatment strategy still exists. The aim of this study of the prospective, multicenter First-Line Treatments in Patients With Critical Limb Ischemia (CRITISCH) registry was to assess the current practice of all first-line treatments strategies in CLI patients in German vascular centers. Methods: Between January 2013 and September 2014, five first-line treatment strategiesdendovascular revascularization (ER), bypass surgery (BS), femoral/profundal artery patchplasty (FAP), conservative treatment, and primary amputationdwere determined among CLI patients in 27 vascular tertiary centers. The main composite end point was major amputation or death, or both, during the hospital stay. Secondary outcomes were hemodynamic failure, major adverse cardiovascular and cerebral events, and reintervention. Univariate logistic models were additionally built to preselect possible risk factors for either event, which were then used as candidates for a multivariate logistic model. Results: The study included 1200 consecutive patients. First-line treatment of choice was ER in 642 patients (53.4%), BS in 284 (23.7%), FAP in 126 (10.5%), conservative treatment in 118 (9.8%), and primary amputation in 30 (2.5%). The composite end point was met in 24 patients (4%) after ER, in 17 (6%) after BS, in 8 (6%) after FAP, and in 9 (8%) after conservative treatment (P [ .172). The highest rate of in-hospital death was observed after primary amputation (10%) and of hemodynamic failure after conservative treatment (91%). Major adverse cardiovascular and cerebral events developed in 4% of patients after ER, in 5% after BS, in 6% after FAP, in 5% after conservative treatment, and in 13% after primary amputation. The reintervention rate was 8%, 14%, 6%, 5%, and 3% in each group, respectively. In the multivariate regression model, coronary artery disease (odds ratio [OR], 2.96; 95% confidence interval [CI], 1.42-6.17) and previous myocardial infarction (PMI) <6 months (OR, 3.67, 95% CI, 1.51-8.88) were identified as risk factors for the composite end point. Risk factors for amputation were dialysis (OR, 3.31, 95% CI, 1.44-7.58) and PMI (OR, 3.26, 95% CI, 1.238.36) and for death, BS compared with ER (OR, 3.32; 95% CI, 1.10-10.0), renal insufficiency without dialysis (OR, 6.34; 95% CI, 1.71-23.5), and PMI (OR, 7.41; 95% CI, 2.11-26.0). Conclusions: The CRITISCH registry revealed ER as the most common first-line approach in CLI patients. Coronary artery disease and PMI <6 months were independent risk factors for the composite end point. Special attention should be also paid to CLI patients with renal insufficiency, with or without dialysis, and those undergoing BS. (J Vasc Surg 2015;-:1-9.)
Critical limb ischemia (CLI) is the most severe type of peripheral arterial vascular disease and remains a substantial cause of death and health costs: the 6-month mortality rate accounts for 20%, and the inpatient hospital treatment averages V20,000 at 1 year.1 In the PAD Awareness, Risk and Treatment: New Resources for Survival (PARTNERS) From the Clinic for Vascular and Endovascular Surgery, University Clinic of Muenster, and Department of Vascular Surgery, St. Franziskus Hospitala; and the Institute of Biostatistics and Clinical Research, University of Muenster.b A list of the First-Line Treatments in Patients With Critical Limb Ischemia (CRITISCH) collaborators is provided in the Appendix. The German Society of Vascular Surgery and Medicine (Deutsche Gesellschaft für Gefäßchirurgie and Gefäßmedizin) sponsored this study. Clinical trial registration number NCT01877252 (clinicaltrials.gov). Author conflict of interest: none. Additional material for this article may be found online at www.jvascsurg.org. Reprint requests: Theodosios Bisdas, MD, PhD, Clinic for Vascular and Endovascular Surgery, University Clinic of Muenster, Albert-Schweitzer Campus 1, 48144, Muenster, Germany (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.2015.04.441
study, which included patients aged 50 to 70 years with history of smoking or diabetes, the overall proportion of CLI patients was 29%.2 Considering that the population ages and the metabolic syndrome is growing in global prevalence, the clinical and socioeconomic effect of the disease will be magnified in the near future.1-3 Nevertheless, there is still a relevant lack of evidence for the most optimal treatment strategy in CLI. This can be explained due to the involvement of different disciplines,4 the lack of high-level evidence,3 and the rapid material evolution in the endovascular era.3,4 The first results of Bypass versus Angioplasty in Severe Ischaemia of the Leg (BASIL-2) and Best Endovascular vs Best Surgical Therapy in Patients With Critical Limb Ischemia (BEST-CLI) trials are anticipated after 2018, and whether their results will cover all aspects of the disease remains controversial.5 At present, the durability of the endovascular approach in CLI patients is a matter of debate, and a current systematic review revealed that the endovascular approach is not inferior to bypass surgery (BS) for limb salvage in those patients.6 Hence, decision making depends mostly on the physician’s expertise, and little is known about which approach accounts as the first-line treatment strategy in CLI patients among the vascular centers. 1
2 Bisdas et al
JOURNAL OF VASCULAR SURGERY --- 2015
Fig 1. Schematic overview of recruitment procedure in First-Line Treatments in Patients With Critical Limb Ischemia (CRITISCH) registry. ICU, Intensive care unit.
To inform this debate, we assessed the current practice in Germany, in the framework of the real-world, multicenter, prospective First-Line Treatments in Patients With Critical Limb Ischemia (CRITISCH) registry. In particular, we analyzed physicians’ preferences and described the respective outcomes of each approach. In this report we present the early results of the CRITISCH registry with the focus on strategy-specific differences and the detection of risk factors for poor in-hospital outcomes. METHODS The study was performed in accordance with the principles in the Declaration of Helsinki. Ethical approval was obtained from all participating centers before patients were recruited. All patients gave written informed consent. Study design and patients. A summary of the recruitment procedure is illustrated in Fig 1, and Supplementary Table I (online only) provides an overview of the study assessment requirements. The inclusion criterion in CRITISCH registry was the diagnosis of CLI lasting >2 weeks. To avoid any bias for possible previous vascular interventions, only patients with new-onset of CLI at the time of presentation and not these with ongoing symptoms after the last previous vascular intervention at the index limb were included. CLI was defined as an ankle-brachial index (ABI) #0.40 or pain at rest, or both, with or without on-going degrees of tissue loss in
the presence of peripheral artery disease (Rutherford classification stages 4-6).7 Only one leg per patient was assessed. The study excluded patients with acute limbthreatening ischemia (embolic or thrombotic), isolated iliac interventions, bone fractures at the index leg, nonatherosclerotic disease (eg, arteritis), and documented hypercoagulable status (Fig 1). The type of treatment was left exclusively to the discretion of the treating physician (best treatment strategy). The different first-line treatment strategies were classified into five groups: Group I included patients undergoing all types or techniques of solely endovascular revascularization (ER). ER was considered any intervention where a percutaneous endovascular technique alone was used. In case of technical failure to cross the lesions, the first-line treatment remained the endovascular approach, and any further procedure by means of BS was considered as a reintervention. Group II consisted of patients undergoing BS using all possible types of conduit. Group III included patients undergoing only femoral/ profundal artery patchplasty (FAP), with or without a concomitant distal endovascular intervention, by means of a hybrid procedure. Group IV included patients treated conservatively. Group V were those undergoing a primary major amputation (PMA).
JOURNAL OF VASCULAR SURGERY Volume -, Number -
End points. All end points are in accordance to the suggested objective performance goals and clinical trial design for evaluating catheter-based treatment for CLI.7 The primary end point of the CRITISCH registry is amputation-free survival (AFS), defined as the time until an above-ankle amputation (PMA) of the index limb or death (any cause), or both. However, such a time-to-event analysis was not possible in this phase of the study because we analyzed only the in-hospital outcomes. Thus, the primary end point was in-hospital amputation or death, or both. Secondary end points of the CRITISCH study are occurrence of a procedure-related reintervention, any major adverse cardiovascular and cerebral event (MACCE), defined as myocardial infarction, stroke, or death, and hemodynamic failure. Hemodynamic failure was defined as first occurrence of any of the following events: 1. Major amputation, 2. A secondary procedure to maintain vascular patency in the index limb, 3. Failure to increase the ABI by at least 0.15 postprocedure compared with the baseline value, 4. Decrease in the ABI by 0.15, 5. Duplex ultrasound imaging demonstrating occlusion of the graft or treated vessel, 6. Duplex ultrasound imaging demonstrating critical graft stenosis (peak systolic velocity >300 cm/s, end diastolic velocity >20 cm/s, or velocity ratio >3.0), or 7. An angiogram demonstrating occlusion of the graft or of any treated vessel, or >50% stenosis in the presence of recurrent clinical symptoms.7 Sampling and power calculation. Our main assumption was that an equal number of patients would undergo ER or BS among the centers and that equal AFS would be observed between treatment groups I and II. From these hypotheses we aimed to prove the noninferiority of intervention group I vs II in the composite end point. Noninferiority was defined by a hazard ratio <1.33 (noninferiority bound ¼ 1.33). The sample size calculation yielded a target cohort size of 1200 patients (power, 80%; level of significance, 5%) in a 24-month recruitment period (started in January 2013), followed by a 24-month follow-up period. In this study design, the exact length of the follow-up period cannot be predetermined, but will be event-driven due to the adaptive-sequential study design.8 Randomization or blinding was not feasible in this study because best treatment was performed as established at each of the 27 vascular centers that participated in this registry. The registry was validated via an audit at participating centers to monitor how accurately all patients and procedures were documented. All centers were compelled to recruit consecutive CLI patients; otherwise, the reason patients were missing was clarified. The recruitment of the target cohort was completed in September 2014.
Bisdas et al 3
Statistical analysis. Descriptive statistical analysis of demographic and periprocedural characteristics was done using the median and the interquartile range as well as absolute and relative frequencies. Inferential statistics, in particular P values, were intended to be exploratory (hypotheses-generating), not confirmatory, and therefore were interpreted by Fisher sense, representing the metric weight of evidence. Hence, an adjustment for multiple testing was not applied, and P values <.05 are consequently considered noticeable instead of significant. Fisher exact and c2 tests were used to assess the relationship between two categoric variables. Differences in quantitative variables between patients groups were evaluated by the Kruskal-Wallis test. The composite end point analysis excluded patients undergoing PMA. Univariate logistic models were built to preselect possible risk factors for each in-hospital end point. The preselected variables were then used as candidates for the multivariate logistic model to weight their influence on the outcomes, and model building was accomplished by means of backward variable selection based on Akaike information criterion. To avoid any statistical bias in the analysis of possible risk factors due to the small samples of groups III, IV, and V, only ER and BS were included in the multiple regression analysis. Similarly, an additional univariate and multivariate analysis was performed to identify which patient characteristics could predict BS as the first-line treatment. Statistical computations were performed using R 3.1.0 software (The R Foundation for Statistical Computing, http://www.r-project.org/ foundation/). RESULTS First-line treatment strategies. Patient characteristics and demographics are reported in Table I. The first-line treatment of choice was ER (group I) in 642 patients (53.5%), followed by BS (group II) in 284 (23.7%), FAP (group III) in 126 (10.5%), conservative treatment (group IV) in 118 (9.8%), and PMA (group V) in 30 patients (2.5%). In group III, 28 patients (2% of the study cohort, 22% of the group cohort) received an additional distal endovascular intervention, and in group V, 23 patients (77%) received above-knee and seven (23%) below-knee amputation. Table II summarizes symptoms and anatomic characteristics in each treatment group. Of note, 24% of patients in group I, 26% in group III, 43% in group IV, and 30% in group V could not be classified by the TransAtlantic Inter-Society Consensus (TASC) classes due to the very distal localization of the disease (Fig 2). Supplementary Tables II and III (online only) summarize the procedural characteristics and the discharge data of the study cohorts, respectively. The multivariate logistic regression model (Table III) suggests that the implementation of open BS was noticeably increased in patients with normal renal function (odds ratio [OR], 3.01), current tobacco use (OR, 2.32), previous vascular intervention (OR, 1.41), TASC C (OR, 17.1) or D lesions (OR, 33.4), and three runoff vessels (OR, 3.13).
JOURNAL OF VASCULAR SURGERY --- 2015
4 Bisdas et al
Table I. Patient characteristics and demographics Group I, ER
Variablea Total No. Age, median (IQR) years Males Angina/CAD PMI <6 monthsd Normal renal function Renal insufficiencye Without dialysis (1530 kg/m2) Previous vascular intervention Previous stroke/TIA Current tobacco use
Group II, SB
642 75 405 298 25 326
(53.5) (14.75) (63) (46) (4) (51)
284 73 192 118 17 185
(23.7) (14) (68) (42) (6) (65)
251 65 310 93 251 73 94
(39) (10) (48) (14) (39) (11) (15)
86 13 136 40 139 31 89
(30) (5) (48) (14) (49) (11) (31)
Group III, FAP 126 72 78 60 2 86 29 11 43 19 40 17 31
Group IV, Conservative
(10.5) (14.75) (62) (48) (2) (68) (23) (9) (34) (15) (32) (13) (25)
Group V, PMA
118 76 74 54 5 62
(9.8) (17) (63) (46) (4) (53)
30 72.5 19 15 3 15
46 10 54 15 57 14 16
(39) (8) (46) (13) (48) (12) (14)
12 3 19 4 13 9 6
P value
(2.5) (12.5) (63) (50) (10) (50)
<.001b <.001c .705 .648 .148 <.001
(40) (10) (63) (13) (43) (30) (20)
.002 .096 .016 .986 .004 .04 <.001
BMI, Body mass index; BS, bypass surgery; CAD, coronary artery disease; eGFR, estimated glomerular filtration rate; ER, endovascular repair; FAP, femoral/ profundal artery patchplasty; IQR, interquartile range; PMA, primary major amputation; PMI, previous myocardial infarction; TIA, transient ischemic attack. a Data are shown as number (%) unless indicated otherwise. b Calculated by c2 test. c Calculated by Kruskal-Wallis test. d Any one of the following criteria meets the diagnosis for PMI: previously documented MI, pathologic Q waves, with or without symptoms in the absence of nonischemic causes, imaging evidence of a region of loss of viable myocardium that is thinned and fails to contract, in the absence of a nonischemic cause. e Calculated as mL/min/1.73 m2. Renal insufficiency or chronic kidney disease was defined as abnormalities of kidney structure or function, present for >3 months, with implications for health.
Table II. Anatomic characteristics and symptomatology in patients for each treatment groupa Group I
Variables Rutherford classification Stage 4 Stage 5 Stage 6 TASC classification TASC A TASC B TASC C TASC D Not applicable Runoff vessels None 1 2 3
Group IV
Group V
BS, FAP, No. (%) No. (%)
Conservative, No. (%)
PMA, No. (%)
I vs II
I vs III
I vs IV
I vs V
II vs III
II vs IV
279 141 (51) 72 (26) 47 (17) 643 375 (58) 140 (22) 52 (8) 278 126 (45) 72 (26) 27 (10)
17 (6) 73 (11) 28 (10)
2 (1) 3 (0.4) 25 (9)
.28 .01 .06
<.001 .001 .63
.07 .54 .32
.06 <.001 <.001
.02 .45 .45
.02 .80 .80
86 65 (76) 2 (2) 203 153 (75) 17 (8) 146 78 (53) 44 (30) 495 192 (39) 198 (40) 270 155 (57) 22 (8)
No.
ER, No. (%)
Group II
Group III
P values
14 15 19 45 33
(16) (7) (13) (9) (12)
5 15 4 43 51
(6) (7) (3) (9) (19)
0 3 1 17 9
(0) (2) (1) (3) (3)
<.001 <.001 .17 <.001 <.001
.75 .04 .002 .008 .38 .003 .17 .13 .65 <.001
272 173 (64) 31 (11) 18 512 277 (54) 134 (26) 43 251 114 (45) 76 (30) 35 165 78 (47) 43 (26) 30
(7) (8) (14) (18)
32 49 24 13
(12) (10) (10) (8)
18 9 2 1
(7) (2) (1) (1)
<.001 .28 .003 .25
.002 1 .06 .76 .13 .52 .001 .88
.1 .08 .24 .004 .52 <.001 .19 .14 .24
II vs V
III vs IV
III vs IV
.02 <.001 .001 .37 <.001 .76 <.001 <.001 <.001 .76 <.001 <.001
<.001 .03 1 .06 .07 .05 .04 .42 .85 1 1 <.001 .09 .002 .13 <.001 <.001 .15 1 .04 <.001 <.001 .001 .007 .65 .33 .02 .90 .04
IV vs V
<.001 <.001 .32 .08 .21 .01 .34 .09
.02 .24 .18 .01
<.001 .83 .02 .01
.58 1 1 .06 .22 .001 .30 .11 .30
BS, Bypass surgery; ER, endovascular repair; FAP, femoral/profundal artery patchplasty; PMA, primary major amputation; TASC, TransAtlantic Inter-Society Consensus. a The P values of c2 tests compare the treatment groups regarding the distribution of Rutherford and TASC classes and number of runoff vessels.
End points. The absolute and relative frequencies of the investigated in-hospital end points are listed in Table IV. No statistically noticeable difference between the unmatched groups (I-IV) regarding the composite end point was observed (P ¼ .172). Group II revealed the highest in-hospital reintervention rate (14%), and PMA showed the
highest in-hospital mortality (10%). As expected, group IV showed the highest rate of hemodynamic failure (91%). Risk factors. Supplementary Table IV (online only) summarizes the results of the univariate logistic regression model of all documented risk factors for the six in-hospital end points. The coefficients of the resulting multivariate
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Bisdas et al 5
Fig 2. Angiography of the right lower limb in a diabetic patient at Rutherford class 5; The TransAtlantic Inter-Society Consensus (TASC) classification was not applicable in this patient. ATA, Anterior tibial artery; DFA, deep femoral artery; PA, popliteal artery; SFA, superficial femoral artery; TFT, tibiofibular trunk.
logistic regression models are listed in Fig 3 and Supplementary Table V (online only). According to this model, the risk of the in-hospital composite end point was 2.96-times and 3.67-times statistically noticeable higher for patients with coronary artery disease (CAD) and previous myocardial infarction (PMI) <6 months, respectively, compared with their respective counterparts. Risk factors for amputation were dialysis (OR, 3.31) and PMI <6 months (OR, 3.26), whereas in-hospital death was affected by BS vs ER (OR, 3.32), renal insufficiency without dialysis (OR, 6.34), and PMI <6 months (OR, 7.41). Renal insufficiency, with (OR, 1.86) or without dialysis (OR, 1.55), current tobacco use (OR, 0.63), and previous intervention in the index limb (OR, 1.42) were also independent risk factors for hemodynamic failure. Finally, patients undergoing BS compared with ER carried a higher risk for reintervention (OR, 1.70). DISCUSSION In our registry, 54% of CLI patients among the participating centers underwent primarily an ER attempt, 24% underwent BS as primary approach, and 11% received FAP. Finally, 13% of CLI patients were not considered as appropriate candidates for an ER or BS approach and did not receive any vascular intervention (groups IV and V), a proportion that is much lower than previously reported from Medicare and insurance data sets.9
In-hospital mortality and major amputation were 1% and 3% in group I and 3% and 4% in group II, respectively. Our registry showed similar in-hospital mortality rates compared with the PRoject of Ex-Vivo vein graft ENgineering via Transfection III (PREVENT III) and BASIL trials and a significant decrease especially after ER (1%) compared with a previous report originating from the nationwide data of all CLI-associated hospitalizations in Germany from 2005 to 2009, where the crude overall inhospital mortality amounted to 8% in CLI patients.3,10,11 The analysis of the comorbidities and the anatomic characteristics in each group showed that ER was the most common first-line treatment regardless severity of ischemia, TASC lesion, or number of run-off vessels (Table II). In the ER group (n ¼ 642), 375 patients (58%) were in Rutherford stage 5, showed a comparable contribution between TASC B (153; 24%) and TASC D (188; 29%) lesions, and had one-vessel runoff status (277; 43%). However, the multivariate regression analysis revealed that the probability of receiving BS due to CLI was statistically noticeable higher in patients with normal renal function (3-fold), current tobacco use (2-fold), previous vascular intervention (1.4-fold), TASC C (17-fold) or D (33-fold) lesions and three runoff vessels (3-fold). In the third group of FAP (n ¼ 126), comparable proportions of patients were in Rutherford stages 4 (47; 37%) and 5 (52; 41%), 45 patients (36%) were treated for TASC D lesions,
JOURNAL OF VASCULAR SURGERY --- 2015
6 Bisdas et al
Table III. Results of the univariate and multivariate logistic regression models for the probability of receiving an open bypass surgery (BS) Separate univariate models Clinical variable
OR (95% CI)
Age in years Male gender Diabetes mellitus Normal renal function vs Renal insufficiency without dialysis Dialysis Current tobacco use Obesity CAD PMI <6 months Previous vascular intervention TASC B vs TASC A TASC C vs TASC A TASC D vs TASC A Nonclassifiable vs TASC A 3 vs no runoff vessels 3 vs 1 runoff vessel 3 vs 2 runoff vessels
P value
0.98 (0.97-0.99) 1.23 (0.93-1.63) 1.06 (0.81-1.38) 1.49 2.59 2.39 0.98 0.81 1.60 1.47 3.84 18.1 28.4 3.73 2.74 0.99 0.81
Multivariate model OR (95% CI)
P value
1.35 (0.97-1.89) 3.01 (1.55-5.83) 2.32 (1.62-3.33)
.075 .001 <.001
1.41 3.39 17.1 33.4 4.32 3.13 0.94 0.73
.027 .110 <.001 <.001 .053 <.001 .799 .228
.001 .148 .684
(1.11-1.99) (1.41-4.75) (1.76-3.24) (0.67-1.44) (0.62-1.07) (0.88-2.91) (1.13-1.93) (0.87-17.0) (4.27-76.9) (6.90-117) (0.86-16.2) (1.64-4.57) (0.67-1.48) (0.52-1.26)
.008 .002 <.001 .927 .134 .120 .005 .076 <.001 <.001 .079 <.001 .977 .352
(1.04-1.91) (0.76-15.2) (3.99-73.6) (8.05-139) (0.98-19.0) (1.77-5.56) (0.59-1.50) (0.44-1.21)
CAD, Coronary artery disease; CI, confidence interval; OR, odds ratio; PMI, previous myocardial infarction; TASC, TransAtlantic Inter-Society Consensus.
Table IV. In-hospital end points of the First-Line Treatments in Patients With Critical Limb Ischemia (CRITISCH) registry
End point Composite end point Amputation Death Hemodynamic failure MACCE Reintervention Type of reintervention Endovascular Open surgery Minor amputation
Group I
Group II
Group III
Group IV
Group V
ER (n ¼ 642), No. (%)
BS (n ¼ 284), No. (%)
FAP (n ¼ 126), No. (%)
Conservative (n ¼ 118), No. (%)
PMA (n ¼ 30), No. (%)
P value
e e 3 (10) e 4 (13) 1 (3)
.172 .67 .003 <.001 .097 .015
24 20 6 81 23 50
(4) (3) (1) (13) (4) (8)
32 (64) 18 (36) 80 (12)
17 10 8 24 15 33
(6) (4) (3) (8) (5) (14)a
6 (9) 30 (91) 39 (14)
8 5 4 11 8 11
(6) (4) (3) (9) (6) (9)
5 (45) 6 (55) 7 (6)
9 6 4 107 6 6
(8) (5) (3) (91) (5) (5)
1 (17) 5 (83) 5 (4)
0 (0) 1 (100) e
BS, Bypass surgery; ER, endovascular repair; FAP, femoral/profundal artery patchplasty; MACCE, major adverse cardiovascular and cerebral event; PMA, primary major amputation. a Indications: bypass stenosis/need for extension in 10, bypass thrombectomy in 9, wound infection in 9, and hematoma in 5.
and 108 (86%) had more than one runoff vessel. Conservative treatment (n ¼ 118) was applied mostly in patients with Rutherford stage 5 (73; 62%), very distal occlusive disease, which could not be classified due to the TASC classification (51; 43%; Fig 2) and one runoff vessel (49; 42%). Finally, 25 of 30 (83%) undergoing PMA were in Rutherford stage 6 and had mostly TASC D lesions (17; 57%) and no run-off vessels (18; 60%). The multiple regression analysis revealed that patients who received a BS had an elevated risk of in-hospital death (OR, 3.32) compared with ER. The procedure-associated morbidity (periprocedural complications: 26%, MACCE: 6%; Supplementary Table III, online only) and the
increased rates of in-hospital reintervention (14%) after BS might explain this outcome. A similar trend could also be confirmed in a meta-analysis of ER vs BS of femoropopliteal arterial disease: the incidence of 30-day morbidity was 33% in the surgical arm and 17% in the endovascular intervention arm (P ¼ .007).12 In the BASIL trial, BS was also associated with a significantly higher rate of early morbidity (57%) than the rate associated with angioplasty (41%).3 Furthermore, we found that BS also carried a 1.70-times higher risk for in-hospital intervention compared with the ER. However, it remains noteworthy that 49% (129 of 284) of patients treated by BS had previous failed interventions, compared with 39% (n ¼ 251) in
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Bisdas et al 7
Bypass surgery
Composite
Coronary artery desease
endpoint
Previous vascular intervention
Renal insuff. w/o dialysis Renal insuff. with dialysis
Endpoint
Coronary artery desease
amputation
Previous myocardial infarction Previous vascular intervention
Bypass surgery
Endpoint
Renal insuff. w/o dialysis
mortality
Renal insuff. with dialysis Previous myocardial infarction
0.5
1.0
2.0
5.0
10.0
20.0
Odds Ratios and 95% Confidence Intervals Fig 3. Odds ratios (ORs; diamonds) and 95% confidence intervals (CIs; horizontal lines) are shown of the risk factors of all in-hospital end points. The ORs and CIs were deduced from multivariate logistic regression models.
the ER group. In this context, we were not able to assess the type of previous interventions performed in each group and any endovascular intervention in the bypass group might have influenced the outcomes. Patients undergoing PMA showed the highest inhospital mortality rates (10%). This outcome admonishes the necessity of revascularization attempt in CLI patients before any major amputation and fuels the current recommendation, that for limb salvage, revascularization is indicated whenever technically feasible.9,13 Of note, the real-world frequency of PMA is not well addressed in the literature. An estimated 10% to 40% of CLI patients require PMA due to overwhelming infection or unreconstructable vascular disease.1 The incidence of PMA was much lower in our registry, and only 3% of the patients received this as the first-line treatment. Taylor et al14 reported that compared with young healthy patients, patients with limited preoperative ambulatory ability, age $70 years, dementia, endstage renal disease, and advanced CAD perform poorly after PMA and cannot wear a leg prothesis. In our registry, the median age of patients undergoing PMA was 73 years, CAD was present in 50%, 10% were on dialysis, and a previous stroke had occurred in 30% (Table I). These comorbidities are surely relevant contributors to all-cause mortality in this group of patients. We are not able to deliver any data on postoperative functional use of leg
prosthesis because such outcomes were not documented in the registry. Of note, 33% of our patients undergoing PMA had no ambulatory ability at discharge (Supplementary Table III, online only). In our study, gender, diabetes mellitus, obesity, TASC classification, and run-off status did not influence any multivariate model. In contrast, renal insufficiency and PMI <6 months were identified as crucial risk factors for poor in-hospital outcomes (Fig 3). Coexistence of CAD showed 2.96-times and 2.75-times higher risk for the composite end point and MACCE, respectively. Hobbs et al15 observed that more than one-third of CLI patients sustain a perioperative myocardial injury, many of which are not clinically apparent. Several studies have also reported that CLI patients are suboptimally treated with lipid-lowering and antiplatelet medications. Statins were used in 46% of CLI patients in the Coronary Artery Revascularization Prophylaxis (CARP) and PREVENT III trials, but this rate was much lower (36%) in the BASIL trial.3,10,16 In our registry, more than the half of the patients regardless treatment strategy were receiving statin therapy, and most received at least acetylsalicylic acid (Supplementary Table III, online only) in conformance with the current guidelines on management of patients undergoing noncardiac surgery.17 This seems to be contradictory to previously published data from 2009 to 2011 that originated from the largest public
JOURNAL OF VASCULAR SURGERY --- 2015
8 Bisdas et al
health insurance in Germany (41,882 patients).11 These data, however, represent all German centers dealing with CLI patients, whereas only of centers of excellence participated in CRITISCH registry. Finally, another interesting result of the registry was that impaired renal function without dialysis was crucial risk factor for in-hospital death (OR, 6.34), hemodynamic failure (OR, 1.55), and MACCE (OR, 2.09). Few studies have investigated any correlation of impaired GFR to poor outcomes after treatment for CLI, especially after ER. Here, Ortmann et al18 observed that CLI patients at all levels of renal impairment benefit from revascularization (endovascular or surgical technique) compared with medical therapy alone. Thus, they suggested that a revascularization attempt should not be withheld in CLI patients with renal disease.18 This analysis has some limitations that are inherent to all registries. The exact selection criteria for each treatment modality that were, principally, left to the discretion of the responsible physician could not be assessed. In any case, all treating physicians provided the best treatment strategy based on their judgment. The design of this study was nonrandomized, and there was a relevant heterogeneity among the groups. However, any discrepancies between comorbidities, symptoms, and anatomy (eg, TASC classification) could be weighted and balanced by the multivariate regression analysis. Inclusion of patients was based only on Rutherford stages including ABI (due to its widespread use). No implementation of transcutaneous partial pressure of oxygen (compared with clinical trials) was done. This may cause bias, especially in diabetic patients, as well as by rating the severity of leg ischemia. In this context, we were also not able to identify any infection-associated amputation with, however, functional vascular reconstruction. The registry was designed to include all consecutive patients based on the duration of their symptoms and the bedside clinical examination by board-certified vascular surgical or medical specialists. No provision was made for a clinical event committee. The responsible physician assessed and reported the angiographic and ultrasound findings and whether an end point was met. This registry represents the current practice in 27 leading vascular centers and includes any available new endovascular technology applied during the recruitment period; however, these results may not represent all vascular centers of the country. Despite implementation of TASC classification, additional information about specific characteristics of the lesions was not recorded. Similarly, we were not able to deliver any details (diameter, length, etc) about the devices used. Any comparison of the in-hospital outcomes between different therapeutic options within the same group (eg, plain angioplasty vs primary stenting) was outside of the topics at this phase of the study. Our initial assumption that an equal number of patients would undergo ER or BS could not be confirmed. Hence, the test of the primary efficacy end point AFS probably has reduced power. Finally, we want to point out again that all
of the P values are to be interpreted again in an explorative sense and that the results need to be confirmed in prospective confirmatory studies. CONCLUSIONS The CRITISCH registry revealed that ER is at present the first-line treatment strategy of choice (54%) in CLI patients among the participating German vascular centers. BS received one in four patients with CLI in this registry, but compared with ER was associated with higher risk for in-hospital death and reintervention. However, these were only in-hospital outcomes, and the 2-year results are awaited. Main risk factors for the composite end point of in-hospital amputation or death, or both, were CAD and PMI <6 months. Patients with PMI <6 months or renal insufficiency without dialysis carry the highest risk for poor in-hospital outcomes. AUTHOR CONTRIBUTIONS Conception and design: TB, GT Analysis and interpretation: TB, MB, GT Data collection: TB, MB, GT Writing the article: TB, MB, GT Critical revision of the article: TB, MB, GT Final approval of the article: TB, MB, GT Statistical analysis: MB Obtained funding: GT Overall responsibility: GT REFERENCES 1. Norgren L, Hiatt WR, Dormandy JA, Nehler MR, Harris KA, Fowkes FG; on behalf of the TASC II Working Group. Intersociety consensus for the management of peripheral arterial disease. J Vasc Surg 2007;45:5-67. 2. Hirsch AT, Criqui MH, Treat-Jacobson D, Regensteiner JG, Creager MA, Olin JW, et al. Peripheral arterial disease detection, awareness, and treatment in primary care. JAMA 2001;286:1317-24. 3. Adam DJ, Beard JD, Cleveland T, Bell J, Bradbury AW, Forbes JF, et al. Bypass versus angioplasty in severe ischaemia of the leg (BASIL): multicentre, randomised controlled trial. Lancet 2005;366:1925-34. 4. Diehm N, Baumgartner I, Jaff M, Do DD, Minar E, Schmidlii J, et al. A call for uniform reporting standards in studies assessing endovascular treatment for chronic ischemia of lower limb arteries. Eur Heart J 2007;28:798-805. 5. Farber A, Rosenfield K, Menard M. The BEST-CLI trial: a multidisciplinary effort to assess which therapy is best for patients with critical limb ischemia. Tech Vasc Interv Radiol 2014;17:221-4. 6. Jones WS, Dolor RJ, Hasselblad V, Vemulapalli S, Subherwal S, Schmit K, et al. Comparative effectiveness of endovascular and surgical revascularization for patients with peripheral artery disease and critical limb ischemia: systematic review of revascularization in critical limb ischemia. Am Heart J 2014;167:489-98. 7. Conte MS, Geraghty PJ, Bradbury AW, Hevelone ND, Lipsitz SR, Moneta GL, et al. Suggested objective performance goals and clinical trial design for evaluating catheter-based treatment of critical limb ischemia. J Vasc Surg 2009;50:1462-73. 8. Bisdas T, Stachmann A, Weiss K, Borowski M, Grundmann RT, Torsello G; for the CRITISCH collaborators. German national registry for first-line treatment strategies in patients with critical limb ischemia (CRITISCH registry): study design and current status. Gefaesschirurgie 2014;19:135-40. 9. Allie DE, Hebert CJ, Lirtzman MD, Wyatt CH, Keller VA, Khan MH, et al. Critical limb ischemia: a global epidemic. A critical analysis of
JOURNAL OF VASCULAR SURGERY Volume -, Number -
10.
11.
12.
13.
14.
15.
16.
17.
18.
current treatment unmasks the clinical and economic costs of CLI. EuroIntervention 2005;1:75-84. Conte MS, Bandyk DF, Clowes AW, Moneta GL, Seely L, Lorenz TJ, et al. Results of PREVENT III: a multicentre, randomized trial of edifoligide for the prevention of vein graft failure in lower extremity bypass surgery. J Vasc Surg 2006;43:742-51. Malyar N, Fürstenberg T, Wellmann J, Meyborg M, Lüders F, Gebauer K, et al. Recent trends in morbidity and in-hospital outcomes of in-patients with peripheral arterial disease: a nationwide populationbased analysis. Eur Heart J 2013;34:2706-14. Antoniou GA, Chalmers N, Georgiadis GS, Lazarides MK, Antoniou SA, Serracino-Inglott F, et al. A meta-analysis of endovascular versus surgical reconstruction of femoropopliteal arterial disease. J Vasc Surg 2013;57:242-53. Anderson JL, Adams CD, Antman EM, Bridges CR, Califf RM, Casey DE, et al. 2011 ACCF/AHA focused update incorporated into the ACC/AHA 2007 guidelines for the management of patients with unstable angina/noneST-elevation myocardial infarction: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. Circulation 2011;123:426-579. Taylor SM, Kalbaugh CA, Blackhurst DW, Hamontree SE, Cull DL, Messich HS, et al. Preoperative clinical factors predict postoperative functional outcomes after major lower limb amputation: an analysis of 553 consecutive patients. J Vasc Surg 2005;42:227-35. Hobbs SD, Yapanis M, Burns PJ, Wilmink AB, Bradbury AW, Adam DJ. Perioperative myocardial injury in patients undergoing surgery for critical limb ischemia. Eur J Vasc Endovasc Surg 2005;29:301-4. Raghunathan A, Rapp JH, Littooy F, Santilli S, Krupski WC, Ward HB, et al. Postoperative outcomes fro patients undergoing elective revascularization for critical limb ischemia and intermittent claudication: a subanalysis of the Coronary Artery Revascularization Prophylaxis (CARP) trial. J Vasc Surg 2006;43:1175-82. Fleisher LA, Fleischmann KE, Auerbach AD, Barnason SA, Beckman JA, Bozkurt B, et al. 2014 ACC/AHA guideline on perioperative cardiovascular evaluation and management of patients undergoing noncardiac surgery: executive summary. Circulation 2014;130:2215-45. Ortmann J, Gahl B, Diehm N, Dick F, Traupe T, Baumgartner I. Survival benefits of revascularization in patients with critical limb ischemia and renal insufficiency. J Vasc Surg 2012;56:737-45.
Bisdas et al 9
APPENDIX The CRITISCH collaborators are: Farzin Adili, MD (Darmstadt), Kai Balzer, MD (Bonn), Thomas Betz, MD (Regensburg), Arend Billing, MD (Offenbach), Dittmar Böckler, MD (Heidelberg), Daniel Brixner, MD (Offenbach), Sebastian E. Debus, MD (Hamburg), Konstantinos P. Donas (Münster), Hans-Henning Eckstein, MD (Munich), Hans-Joachim Florek, MD (Freital), Asimakis Gkremoutis, MD (Frankfurt), Reinhardt Grundmann, MD, Thomas Hupp, MD (Stuttgart), Tobias Keck, MD (Lübeck), Joachim Gerß, PhD (Münster), Wojciech, Klonek, MD (Cloppenburg), Werner Lang, MD (Erlangen), Ute Ludwig (Munich), Björn May (Freital), Alexander Meyer, MD (Erlangen), Bernhard Mühling, MD (Biberach), Alexander Oberhuber, MD (Düsseldorf), Holger Reinecke, MD (Münster), Christian Reinhold, MD (Mönchengladbach), Ralf-Gerhard Ritter, MD (Bielefeld), Hubert Schelzig, MD (Düsseldorf), Christian Schlensack, MD (Tübingen), Thomas Schmitz-Rixen, MD (Frankfurt), Karl-Ludwig Schulte, MD (Berlin), Matthias Spohn, MD (Bamberg), Konstantinos Stavroulakis, MD (Münster), Markus Steinbauer, MD (Regensburg), Martin Storck, MD (Karlsruhe), Matthias Trede, MD (Bremen), Barbara Weis-Müller, MD (Mönchengladbach), Heiner Wenk (Bremen), Thomas Zeller, MD (Bad Kronzingen), and Alexander Zimmermann, MD (Munich).
Submitted Feb 24, 2015; accepted Apr 28, 2015.
Additional material for this article may be found online at www.jvascsurg.org.
JOURNAL OF VASCULAR SURGERY --- 2015
9.e1 Bisdas et al
Supplementary Table I (online only). Trial assessment requirements Variable Demography Medical history Physical exam Comorbidities ABI Rutherford staging Inclusion evaluation Informed consent Medicationa Duplex ultrasound Angiography Procedure details Hospital FU visits Telephone FU
Baseline X X X X X X X X Xb X
Procedure
Discharge
6 months
12 months
24 months
Unscheduled
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X Xc
X X Xc
X X Xc
X X
X X
X X
X
X X X
ABI, Ankle-brachial index; FU, follow-up. a Aspirin, clopidogrel, prasugrel, other anticoagulants, statins. b In case of patients without preoperative diagnostic angiography (group II-IV), the findings of the magnetic resonance or computed tomography angiography were used. c Angiography during surveillance was performed only in case of reintervention or according to the physician’s discretion.
Supplementary Table II (online only). Procedural characteristics in the first three intervention groups
Group Group I (n ¼ 642) ER
Femoral vessels (n ¼ 347)
Popliteal vessels (n ¼ 272)
Characteristic
Characteristic
No. (%)
No. (%)
Tibial vessels (n ¼ 368) Characteristic
No. (%)
PTA only 125 (36.0) PTA only 104 (38.2) PTA only 259 (70.4) PTA þ stent 141 (40.6) PTA þ stent 90 (33.1) PTA þ stent 24 (6.5) DCB 63 (18.2) DCB 53 (19.5) DCB 36 (9.8) DCS 6 (1.7) DCS 2 (0.7) DCS 12 (3.3) Other 38 (11.0) Other 32 (11.8) Other 22 (6.0) Failed to cross 19 (5.5) Failed to cross 17 (6.3) Failed to cross 33 (9.0) Group II Leg vein 151 (53.2) (n ¼ 284) BS Arm vein 11 (3.9) Combined leg 3 (1.1) and arm vein Dacron 32 (11.3) PTFE 71 (25) Other 22 (7.7) Group III Common FAP 51 (40.1) (n ¼ 126) FAP Deep FAP 9 (7.1) Common and 66 (52.4) deep FAP Hybrid procedurea 28 (22.2) Patchplasty only 98 (77.8)
Previous bypass (n ¼ 26) Characteristic
No. (%)
PTA only 8 (30.8) PTA þ stent 3 (11.5) DCB 2 (7.7) DCS 2 (7.7) Other 12 (46.2) Failed to cross 1 (3.8)
BS, Bypass surgery; DCB, drug-coated balloon; DCS, drug-coated stent, ER, endovascular repair; FAP, femoral/profundal artery patchplasty; PTA, percutaneous transluminal angioplasty, PTFE, polytetrafluoroethylene. a These patients were not included in the ER or BS groups.
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Bisdas et al 9.e2
Supplementary Table III (online only). Discharge data of the study groups Group I, ER
Variablea In-hospital stay, median (IQR) days Periprocedural complications Hematoma Wound infection Lymph fistula/lymphocele Pneumonia Dialysis Compartment syndrome Deep vein thrombosis Acute coronary event Stroke Discharge Home Other institution Medication at discharge Statins Acetylsalicylic acid Vitamin K antagonists Clopidogrel Prasugrel New anticoagulants
7 58 16 26
(15) (9) (2) (4) 0 7 (1) 3 (<1) 0 0 18 (3) 3 (<1)
Group II, BS
Group III, FAP
(10) (26) (6) (11) (7) (<1) (<1) (2) 0 7 (2) 1 (<1)
13 (11) 23 (18) 0 6 (5) 12 (10) 2 (2) 1 (1) 0 0 4 (3) 1 (1)
563 (88) 72 (11)
216 (76) 58 (20)
100 (79) 22 (17)
359 490 126 340 4 93
176 207 94 58 1 48
(56) (76) (20) (53) (1) (14)
15 73 16 31 21 1 1 5
(62) (73) (33) (20) (0) (17)
74 98 18 44 0 21
(59) (78) (14) (35) (0) (17)
Group IV, Conservative
Group V, PMA
P value
10 14 3 7
16 10 1 8
(11) (33) (3) (27) 0 1 (3) 0 0 0 1 (3) 0
<.001
99 (84) 14 (12)
17 (57) 10 (33)
<.001 <.001
72 82 19 27 0 27
16 22 2 6 0 13
.446 .443 <.001 <.001 .766 <.001
(21) (12) (3) (6) 0 3 (3) 0 0 0 2 (2) 0
(61) (69) (16) (23) (0) (23)
(53) (73) (7) (20) (0) (43)
BS, Bypass surgery; ER, endovascular repair; FAP, femoral/profundal artery patchplasty; PMA, primary major amputation. a Data are shown as number (%) unless indicated otherwise.
Supplementary Table IV (online only). Results of the univariate logistic regression models for each in-hospital end point, given by odds ratios (ORs) and referring 95% confidence intervals (CIs) of all possible risk factors Variable BS vs ER Age in years Male gender Renal insufficiency without dialysis vs normal renal function Dialysis vs normal renal function Diabetes mellitus Current tobacco use Obesity CAD PMI <6 months Previous vascular intervention TASC B vs TASC A TASC C vs TASC A TASC D vs TASC A Nonclassifiable vs TASC A No vs 3 runoff vessels 1 vs 3 runoff vessels 2 vs 3 runoff vessels
Composite end point 1.64 1.02 1.77 1.92
(0.87-3.10) (0.99-1.05) (0.98-3.21) (1.09-3.40)
Amputation 1.14 0.97 2.81 1.15
(0.52-2.46) (0.97-1.03) (1.24-6.40) (0.56-2.38)
Death 3.07 1.05 1.20 5.93
(1.06-8.94) (1.00-1.09) (0.51-2.80) (2.19-16.1)
Hemodynamic failure 0.64 1.01 0.98 1.62
(0.40-1.03) (0.99-1.02) (0.73-1.31) (1.21-2.18)
MACCE 1.50 1.04 1.57 2.54
(0.77-2.92) (1.01-1.07) (0.86-2.87) (1.44-4.46)
Reintervention 1.88 1.00 0.98 0.61
(1.21-2.94) (0.98-1.01) (0.65-1.48) (0.39-0.96)
3.42 (1.61-7.25)
3.96 (1.77-8.84)
2.68 (0.51-14.0)
2.00 (1.25-3.21)
0.94 (0.28-3.22)
0.52 (0.22-1.24)
1.36 0.60 1.35 2.06 4.41 1.48
(0.81-2.28) (0.28-1.29) (0.69-2.64) (1.21-3.50) (2.04-9.53) (0.88-2.47)
1.63 0.56 1.03 2.14 4.15 2.03
(0.87-3.07) (0.22-1.44) (0.43-2.49) (1.12-4.08) (1.66-10.4) (1.08-3.81)
1.24 0.77 1.93 2.17 6.00 1.10
(0.56-2.73) (0.26-2.28) (0.76-4.91) (0.95-4.95) (2.16-16.7) (0.50-2.45)
1.23 0.57 1.13 1.13 1.88 1.40
(0.94-1.63) (0.39-0.85) (0.76-1.66) (0.86-1.49) (1.03-3.41) (1.06-1.85)
1.33 0.77 1.16 3.82 8.57 0.97
(0.78-2.27) (0.37-1.60) (0.56-2.41) (2.06-7.08) (4.26-17.2) (0.57-1.68)
0.81 1.13 0.90 1.02 0.85 1.67
(0.54-1.21) (0.70-1.84) (0.50-1.61) (0.68-1.51) (0.30-2.39) (1.12-2.49)
0.62 1.19 1.38 0.87
(0.17-2.27) (0.35-4.07) (0.48-4.02) (0.27-2.81)
0.84 1.19 1.05 0.84
(0.21-3.45) (0.29-4.87) (0.30-3.65) (0.22-3.26)
0.42 1.78 2.68 1.28
(0.03-6.81) (0.18-17.4) (0.35-20.6) (0.14-11.6)
0.73 0.62 1.16 2.01
(0.37-1.43) (0.30-1.28) (0.65-2.08) (1.10-3.67)
0.62 0.88 1.33 0.71
(0.17-2.27) (0.24-3.21) (0.46-3.89) (0.21-2.36)
0.97 0.74 1.54 0.71
(0.38-2.44) (0.27-2.07) (0.68-3.50) (0.28-1.79)
2.40 (0.88-6.56) 1.85 (0.70-4.87) 1.19 (0.39-3.62)
1.86 (0.59-5.86) 1.55 (0.52-4.63) 0.99 (0.27-3.55)
5.61 (0.70-44.7) 3.27 (0.42-25.7) 3.33 (0.39-28.8)
3.13 (1.83-5.34) 1.98 (1.19-3.31) 1.42 (0.80-2.52)
1.66 (0.63-4.32) 1.36 (0.55-3.38) 0.99 (0.34-2.82)
1.70 (0.85-3.40) 1.17 (0.60-2.27) 1.10 (0.52-2.32)
BS, Bypass surgery; CAD, coronary artery disease; ER, endovascular repair; MACCE, major adverse cardiovascular and cerebral event; PMI, previous myocardial infarction; TASC, TransAtlantic Inter-Society Consensus.
JOURNAL OF VASCULAR SURGERY --- 2015
9.e3 Bisdas et al
Supplementary Table V (online only). Coefficients of multivariate logistic regression modelsa Coefficients Intercept BS vs ER Age in years Male gender Renal insufficiency without dialysis vs normal renal function Dialysis vs normal renal function Diabetes mellitus Current tobacco use Obesity CAD PMI <6 months Previous vascular intervention TASC B vs TASC A TASC C vs TASC A TASC D vs TASC A Non-classifiable vs TASC A No vs 3 runoff vessels 1 vs 3 runoff vessels 2 vs 3 runoff vessels
Composite end point
Amputation
Death
Hemodynamic failure
MACCE
Reintervention
4.01 0.50
4.16
5.98 1.20 *
1.61
4.11
2.41 0.53
*
* 0.05
1.85
0.44
0.74
0.48
*
1.20
14.06
0.62
0.41
0.58
*
0.47 1.09 1.30
0.41 1.18 0.61
2.00
* 0.35 * * * * * * *
1.01 1.68 0.38
BS, Bypass surgery; CAD, coronary artery disease; ER, endovascular repair; MACCE, major adverse cardiovascular and cerebral event; PMI, previous myocardial infarction; TASC, TransAtlantic Inter-Society Consensus. a Asterisks mark risk factors that have been detected by univariate logistic regression but not by multivariate logistic regression. The multivariate logistic regression model for the composite end point, for instance, is given by log[P/(1 P)] ¼ 4.01 þ 0.50 BS þ 1.09 CAD þ 1.30 PMI, where P ¼ P(composite end point) is the probability of an in-hospital amputation or death, or both. The model suggests that the chance of an amputation or death during in-hospital stay was 1.64 (exp[0.50]) times higher for patients who received BS compared with those who underwent ER. Similarly, the chance was estimated to be 2.96 (exp[1.09]) and 3.67 (exp[1.30]) times higher for patients with CAD and PMI than for their respective counterpart. Through the coefficients, odds ratios (ORs) of each end point for any combination of covariates can be obtained. For instance, considering the composite end point, the ORs of a patient with BS, CAD, and PMI vs a patient without them is estimated to be exp(0.5 þ 1.09 þ 1.3) ¼ 17.99.
Critical limb ischemia (CLI) is the most severe type of peripheral arterial vascular disease and remains a substantial cause of death and health costs: the 6-month mortality rate accounts for 20%, and the inpatient hospital treatment averages V20,000 at 1 year.1 In the PAD Awareness, Risk and Treatment: New Resources for Survival (PARTNERS) From the Clinic for Vascular and Endovascular Surgery, University Clinic of Muenster, and Department of Vascular Surgery, St. Franziskus Hospitala; and the Institute of Biostatistics and Clinical Research, University of Muenster.b A list of the First-Line Treatments in Patients With Critical Limb Ischemia (CRITISCH) collaborators is provided in the Appendix. The German Society of Vascular Surgery and Medicine (Deutsche Gesellschaft für Gefäßchirurgie and Gefäßmedizin) sponsored this study. Clinical trial registration number NCT01877252 (clinicaltrials.gov). Author conflict of interest: none. Additional material for this article may be found online at www.jvascsurg.org. Reprint requests: Theodosios Bisdas, MD, PhD, Clinic for Vascular and Endovascular Surgery, University Clinic of Muenster, Albert-Schweitzer Campus 1, 48144, Muenster, Germany (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.2015.04.441
study, which included patients aged 50 to 70 years with history of smoking or diabetes, the overall proportion of CLI patients was 29%.2 Considering that the population ages and the metabolic syndrome is growing in global prevalence, the clinical and socioeconomic effect of the disease will be magnified in the near future.1-3 Nevertheless, there is still a relevant lack of evidence for the most optimal treatment strategy in CLI. This can be explained due to the involvement of different disciplines,4 the lack of high-level evidence,3 and the rapid material evolution in the endovascular era.3,4 The first results of Bypass versus Angioplasty in Severe Ischaemia of the Leg (BASIL-2) and Best Endovascular vs Best Surgical Therapy in Patients With Critical Limb Ischemia (BEST-CLI) trials are anticipated after 2018, and whether their results will cover all aspects of the disease remains controversial.5 At present, the durability of the endovascular approach in CLI patients is a matter of debate, and a current systematic review revealed that the endovascular approach is not inferior to bypass surgery (BS) for limb salvage in those patients.6 Hence, decision making depends mostly on the physician’s expertise, and little is known about which approach accounts as the first-line treatment strategy in CLI patients among the vascular centers. 1
2 Bisdas et al
JOURNAL OF VASCULAR SURGERY --- 2015
Fig 1. Schematic overview of recruitment procedure in First-Line Treatments in Patients With Critical Limb Ischemia (CRITISCH) registry. ICU, Intensive care unit.
To inform this debate, we assessed the current practice in Germany, in the framework of the real-world, multicenter, prospective First-Line Treatments in Patients With Critical Limb Ischemia (CRITISCH) registry. In particular, we analyzed physicians’ preferences and described the respective outcomes of each approach. In this report we present the early results of the CRITISCH registry with the focus on strategy-specific differences and the detection of risk factors for poor in-hospital outcomes. METHODS The study was performed in accordance with the principles in the Declaration of Helsinki. Ethical approval was obtained from all participating centers before patients were recruited. All patients gave written informed consent. Study design and patients. A summary of the recruitment procedure is illustrated in Fig 1, and Supplementary Table I (online only) provides an overview of the study assessment requirements. The inclusion criterion in CRITISCH registry was the diagnosis of CLI lasting >2 weeks. To avoid any bias for possible previous vascular interventions, only patients with new-onset of CLI at the time of presentation and not these with ongoing symptoms after the last previous vascular intervention at the index limb were included. CLI was defined as an ankle-brachial index (ABI) #0.40 or pain at rest, or both, with or without on-going degrees of tissue loss in
the presence of peripheral artery disease (Rutherford classification stages 4-6).7 Only one leg per patient was assessed. The study excluded patients with acute limbthreatening ischemia (embolic or thrombotic), isolated iliac interventions, bone fractures at the index leg, nonatherosclerotic disease (eg, arteritis), and documented hypercoagulable status (Fig 1). The type of treatment was left exclusively to the discretion of the treating physician (best treatment strategy). The different first-line treatment strategies were classified into five groups: Group I included patients undergoing all types or techniques of solely endovascular revascularization (ER). ER was considered any intervention where a percutaneous endovascular technique alone was used. In case of technical failure to cross the lesions, the first-line treatment remained the endovascular approach, and any further procedure by means of BS was considered as a reintervention. Group II consisted of patients undergoing BS using all possible types of conduit. Group III included patients undergoing only femoral/ profundal artery patchplasty (FAP), with or without a concomitant distal endovascular intervention, by means of a hybrid procedure. Group IV included patients treated conservatively. Group V were those undergoing a primary major amputation (PMA).
JOURNAL OF VASCULAR SURGERY Volume -, Number -
End points. All end points are in accordance to the suggested objective performance goals and clinical trial design for evaluating catheter-based treatment for CLI.7 The primary end point of the CRITISCH registry is amputation-free survival (AFS), defined as the time until an above-ankle amputation (PMA) of the index limb or death (any cause), or both. However, such a time-to-event analysis was not possible in this phase of the study because we analyzed only the in-hospital outcomes. Thus, the primary end point was in-hospital amputation or death, or both. Secondary end points of the CRITISCH study are occurrence of a procedure-related reintervention, any major adverse cardiovascular and cerebral event (MACCE), defined as myocardial infarction, stroke, or death, and hemodynamic failure. Hemodynamic failure was defined as first occurrence of any of the following events: 1. Major amputation, 2. A secondary procedure to maintain vascular patency in the index limb, 3. Failure to increase the ABI by at least 0.15 postprocedure compared with the baseline value, 4. Decrease in the ABI by 0.15, 5. Duplex ultrasound imaging demonstrating occlusion of the graft or treated vessel, 6. Duplex ultrasound imaging demonstrating critical graft stenosis (peak systolic velocity >300 cm/s, end diastolic velocity >20 cm/s, or velocity ratio >3.0), or 7. An angiogram demonstrating occlusion of the graft or of any treated vessel, or >50% stenosis in the presence of recurrent clinical symptoms.7 Sampling and power calculation. Our main assumption was that an equal number of patients would undergo ER or BS among the centers and that equal AFS would be observed between treatment groups I and II. From these hypotheses we aimed to prove the noninferiority of intervention group I vs II in the composite end point. Noninferiority was defined by a hazard ratio <1.33 (noninferiority bound ¼ 1.33). The sample size calculation yielded a target cohort size of 1200 patients (power, 80%; level of significance, 5%) in a 24-month recruitment period (started in January 2013), followed by a 24-month follow-up period. In this study design, the exact length of the follow-up period cannot be predetermined, but will be event-driven due to the adaptive-sequential study design.8 Randomization or blinding was not feasible in this study because best treatment was performed as established at each of the 27 vascular centers that participated in this registry. The registry was validated via an audit at participating centers to monitor how accurately all patients and procedures were documented. All centers were compelled to recruit consecutive CLI patients; otherwise, the reason patients were missing was clarified. The recruitment of the target cohort was completed in September 2014.
Bisdas et al 3
Statistical analysis. Descriptive statistical analysis of demographic and periprocedural characteristics was done using the median and the interquartile range as well as absolute and relative frequencies. Inferential statistics, in particular P values, were intended to be exploratory (hypotheses-generating), not confirmatory, and therefore were interpreted by Fisher sense, representing the metric weight of evidence. Hence, an adjustment for multiple testing was not applied, and P values <.05 are consequently considered noticeable instead of significant. Fisher exact and c2 tests were used to assess the relationship between two categoric variables. Differences in quantitative variables between patients groups were evaluated by the Kruskal-Wallis test. The composite end point analysis excluded patients undergoing PMA. Univariate logistic models were built to preselect possible risk factors for each in-hospital end point. The preselected variables were then used as candidates for the multivariate logistic model to weight their influence on the outcomes, and model building was accomplished by means of backward variable selection based on Akaike information criterion. To avoid any statistical bias in the analysis of possible risk factors due to the small samples of groups III, IV, and V, only ER and BS were included in the multiple regression analysis. Similarly, an additional univariate and multivariate analysis was performed to identify which patient characteristics could predict BS as the first-line treatment. Statistical computations were performed using R 3.1.0 software (The R Foundation for Statistical Computing, http://www.r-project.org/ foundation/). RESULTS First-line treatment strategies. Patient characteristics and demographics are reported in Table I. The first-line treatment of choice was ER (group I) in 642 patients (53.5%), followed by BS (group II) in 284 (23.7%), FAP (group III) in 126 (10.5%), conservative treatment (group IV) in 118 (9.8%), and PMA (group V) in 30 patients (2.5%). In group III, 28 patients (2% of the study cohort, 22% of the group cohort) received an additional distal endovascular intervention, and in group V, 23 patients (77%) received above-knee and seven (23%) below-knee amputation. Table II summarizes symptoms and anatomic characteristics in each treatment group. Of note, 24% of patients in group I, 26% in group III, 43% in group IV, and 30% in group V could not be classified by the TransAtlantic Inter-Society Consensus (TASC) classes due to the very distal localization of the disease (Fig 2). Supplementary Tables II and III (online only) summarize the procedural characteristics and the discharge data of the study cohorts, respectively. The multivariate logistic regression model (Table III) suggests that the implementation of open BS was noticeably increased in patients with normal renal function (odds ratio [OR], 3.01), current tobacco use (OR, 2.32), previous vascular intervention (OR, 1.41), TASC C (OR, 17.1) or D lesions (OR, 33.4), and three runoff vessels (OR, 3.13).
JOURNAL OF VASCULAR SURGERY --- 2015
4 Bisdas et al
Table I. Patient characteristics and demographics Group I, ER
Variablea Total No. Age, median (IQR) years Males Angina/CAD PMI <6 monthsd Normal renal function Renal insufficiencye Without dialysis (15
Group II, SB
642 75 405 298 25 326
(53.5) (14.75) (63) (46) (4) (51)
284 73 192 118 17 185
(23.7) (14) (68) (42) (6) (65)
251 65 310 93 251 73 94
(39) (10) (48) (14) (39) (11) (15)
86 13 136 40 139 31 89
(30) (5) (48) (14) (49) (11) (31)
Group III, FAP 126 72 78 60 2 86 29 11 43 19 40 17 31
Group IV, Conservative
(10.5) (14.75) (62) (48) (2) (68) (23) (9) (34) (15) (32) (13) (25)
Group V, PMA
118 76 74 54 5 62
(9.8) (17) (63) (46) (4) (53)
30 72.5 19 15 3 15
46 10 54 15 57 14 16
(39) (8) (46) (13) (48) (12) (14)
12 3 19 4 13 9 6
P value
(2.5) (12.5) (63) (50) (10) (50)
<.001b <.001c .705 .648 .148 <.001
(40) (10) (63) (13) (43) (30) (20)
.002 .096 .016 .986 .004 .04 <.001
BMI, Body mass index; BS, bypass surgery; CAD, coronary artery disease; eGFR, estimated glomerular filtration rate; ER, endovascular repair; FAP, femoral/ profundal artery patchplasty; IQR, interquartile range; PMA, primary major amputation; PMI, previous myocardial infarction; TIA, transient ischemic attack. a Data are shown as number (%) unless indicated otherwise. b Calculated by c2 test. c Calculated by Kruskal-Wallis test. d Any one of the following criteria meets the diagnosis for PMI: previously documented MI, pathologic Q waves, with or without symptoms in the absence of nonischemic causes, imaging evidence of a region of loss of viable myocardium that is thinned and fails to contract, in the absence of a nonischemic cause. e Calculated as mL/min/1.73 m2. Renal insufficiency or chronic kidney disease was defined as abnormalities of kidney structure or function, present for >3 months, with implications for health.
Table II. Anatomic characteristics and symptomatology in patients for each treatment groupa Group I
Variables Rutherford classification Stage 4 Stage 5 Stage 6 TASC classification TASC A TASC B TASC C TASC D Not applicable Runoff vessels None 1 2 3
Group IV
Group V
BS, FAP, No. (%) No. (%)
Conservative, No. (%)
PMA, No. (%)
I vs II
I vs III
I vs IV
I vs V
II vs III
II vs IV
279 141 (51) 72 (26) 47 (17) 643 375 (58) 140 (22) 52 (8) 278 126 (45) 72 (26) 27 (10)
17 (6) 73 (11) 28 (10)
2 (1) 3 (0.4) 25 (9)
.28 .01 .06
<.001 .001 .63
.07 .54 .32
.06 <.001 <.001
.02 .45 .45
.02 .80 .80
86 65 (76) 2 (2) 203 153 (75) 17 (8) 146 78 (53) 44 (30) 495 192 (39) 198 (40) 270 155 (57) 22 (8)
No.
ER, No. (%)
Group II
Group III
P values
14 15 19 45 33
(16) (7) (13) (9) (12)
5 15 4 43 51
(6) (7) (3) (9) (19)
0 3 1 17 9
(0) (2) (1) (3) (3)
<.001 <.001 .17 <.001 <.001
.75 .04 .002 .008 .38 .003 .17 .13 .65 <.001
272 173 (64) 31 (11) 18 512 277 (54) 134 (26) 43 251 114 (45) 76 (30) 35 165 78 (47) 43 (26) 30
(7) (8) (14) (18)
32 49 24 13
(12) (10) (10) (8)
18 9 2 1
(7) (2) (1) (1)
<.001 .28 .003 .25
.002 1 .06 .76 .13 .52 .001 .88
.1 .08 .24 .004 .52 <.001 .19 .14 .24
II vs V
III vs IV
III vs IV
.02 <.001 .001 .37 <.001 .76 <.001 <.001 <.001 .76 <.001 <.001
<.001 .03 1 .06 .07 .05 .04 .42 .85 1 1 <.001 .09 .002 .13 <.001 <.001 .15 1 .04 <.001 <.001 .001 .007 .65 .33 .02 .90 .04
IV vs V
<.001 <.001 .32 .08 .21 .01 .34 .09
.02 .24 .18 .01
<.001 .83 .02 .01
.58 1 1 .06 .22 .001 .30 .11 .30
BS, Bypass surgery; ER, endovascular repair; FAP, femoral/profundal artery patchplasty; PMA, primary major amputation; TASC, TransAtlantic Inter-Society Consensus. a The P values of c2 tests compare the treatment groups regarding the distribution of Rutherford and TASC classes and number of runoff vessels.
End points. The absolute and relative frequencies of the investigated in-hospital end points are listed in Table IV. No statistically noticeable difference between the unmatched groups (I-IV) regarding the composite end point was observed (P ¼ .172). Group II revealed the highest in-hospital reintervention rate (14%), and PMA showed the
highest in-hospital mortality (10%). As expected, group IV showed the highest rate of hemodynamic failure (91%). Risk factors. Supplementary Table IV (online only) summarizes the results of the univariate logistic regression model of all documented risk factors for the six in-hospital end points. The coefficients of the resulting multivariate
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Bisdas et al 5
Fig 2. Angiography of the right lower limb in a diabetic patient at Rutherford class 5; The TransAtlantic Inter-Society Consensus (TASC) classification was not applicable in this patient. ATA, Anterior tibial artery; DFA, deep femoral artery; PA, popliteal artery; SFA, superficial femoral artery; TFT, tibiofibular trunk.
logistic regression models are listed in Fig 3 and Supplementary Table V (online only). According to this model, the risk of the in-hospital composite end point was 2.96-times and 3.67-times statistically noticeable higher for patients with coronary artery disease (CAD) and previous myocardial infarction (PMI) <6 months, respectively, compared with their respective counterparts. Risk factors for amputation were dialysis (OR, 3.31) and PMI <6 months (OR, 3.26), whereas in-hospital death was affected by BS vs ER (OR, 3.32), renal insufficiency without dialysis (OR, 6.34), and PMI <6 months (OR, 7.41). Renal insufficiency, with (OR, 1.86) or without dialysis (OR, 1.55), current tobacco use (OR, 0.63), and previous intervention in the index limb (OR, 1.42) were also independent risk factors for hemodynamic failure. Finally, patients undergoing BS compared with ER carried a higher risk for reintervention (OR, 1.70). DISCUSSION In our registry, 54% of CLI patients among the participating centers underwent primarily an ER attempt, 24% underwent BS as primary approach, and 11% received FAP. Finally, 13% of CLI patients were not considered as appropriate candidates for an ER or BS approach and did not receive any vascular intervention (groups IV and V), a proportion that is much lower than previously reported from Medicare and insurance data sets.9
In-hospital mortality and major amputation were 1% and 3% in group I and 3% and 4% in group II, respectively. Our registry showed similar in-hospital mortality rates compared with the PRoject of Ex-Vivo vein graft ENgineering via Transfection III (PREVENT III) and BASIL trials and a significant decrease especially after ER (1%) compared with a previous report originating from the nationwide data of all CLI-associated hospitalizations in Germany from 2005 to 2009, where the crude overall inhospital mortality amounted to 8% in CLI patients.3,10,11 The analysis of the comorbidities and the anatomic characteristics in each group showed that ER was the most common first-line treatment regardless severity of ischemia, TASC lesion, or number of run-off vessels (Table II). In the ER group (n ¼ 642), 375 patients (58%) were in Rutherford stage 5, showed a comparable contribution between TASC B (153; 24%) and TASC D (188; 29%) lesions, and had one-vessel runoff status (277; 43%). However, the multivariate regression analysis revealed that the probability of receiving BS due to CLI was statistically noticeable higher in patients with normal renal function (3-fold), current tobacco use (2-fold), previous vascular intervention (1.4-fold), TASC C (17-fold) or D (33-fold) lesions and three runoff vessels (3-fold). In the third group of FAP (n ¼ 126), comparable proportions of patients were in Rutherford stages 4 (47; 37%) and 5 (52; 41%), 45 patients (36%) were treated for TASC D lesions,
JOURNAL OF VASCULAR SURGERY --- 2015
6 Bisdas et al
Table III. Results of the univariate and multivariate logistic regression models for the probability of receiving an open bypass surgery (BS) Separate univariate models Clinical variable
OR (95% CI)
Age in years Male gender Diabetes mellitus Normal renal function vs Renal insufficiency without dialysis Dialysis Current tobacco use Obesity CAD PMI <6 months Previous vascular intervention TASC B vs TASC A TASC C vs TASC A TASC D vs TASC A Nonclassifiable vs TASC A 3 vs no runoff vessels 3 vs 1 runoff vessel 3 vs 2 runoff vessels
P value
0.98 (0.97-0.99) 1.23 (0.93-1.63) 1.06 (0.81-1.38) 1.49 2.59 2.39 0.98 0.81 1.60 1.47 3.84 18.1 28.4 3.73 2.74 0.99 0.81
Multivariate model OR (95% CI)
P value
1.35 (0.97-1.89) 3.01 (1.55-5.83) 2.32 (1.62-3.33)
.075 .001 <.001
1.41 3.39 17.1 33.4 4.32 3.13 0.94 0.73
.027 .110 <.001 <.001 .053 <.001 .799 .228
.001 .148 .684
(1.11-1.99) (1.41-4.75) (1.76-3.24) (0.67-1.44) (0.62-1.07) (0.88-2.91) (1.13-1.93) (0.87-17.0) (4.27-76.9) (6.90-117) (0.86-16.2) (1.64-4.57) (0.67-1.48) (0.52-1.26)
.008 .002 <.001 .927 .134 .120 .005 .076 <.001 <.001 .079 <.001 .977 .352
(1.04-1.91) (0.76-15.2) (3.99-73.6) (8.05-139) (0.98-19.0) (1.77-5.56) (0.59-1.50) (0.44-1.21)
CAD, Coronary artery disease; CI, confidence interval; OR, odds ratio; PMI, previous myocardial infarction; TASC, TransAtlantic Inter-Society Consensus.
Table IV. In-hospital end points of the First-Line Treatments in Patients With Critical Limb Ischemia (CRITISCH) registry
End point Composite end point Amputation Death Hemodynamic failure MACCE Reintervention Type of reintervention Endovascular Open surgery Minor amputation
Group I
Group II
Group III
Group IV
Group V
ER (n ¼ 642), No. (%)
BS (n ¼ 284), No. (%)
FAP (n ¼ 126), No. (%)
Conservative (n ¼ 118), No. (%)
PMA (n ¼ 30), No. (%)
P value
e e 3 (10) e 4 (13) 1 (3)
.172 .67 .003 <.001 .097 .015
24 20 6 81 23 50
(4) (3) (1) (13) (4) (8)
32 (64) 18 (36) 80 (12)
17 10 8 24 15 33
(6) (4) (3) (8) (5) (14)a
6 (9) 30 (91) 39 (14)
8 5 4 11 8 11
(6) (4) (3) (9) (6) (9)
5 (45) 6 (55) 7 (6)
9 6 4 107 6 6
(8) (5) (3) (91) (5) (5)
1 (17) 5 (83) 5 (4)
0 (0) 1 (100) e
BS, Bypass surgery; ER, endovascular repair; FAP, femoral/profundal artery patchplasty; MACCE, major adverse cardiovascular and cerebral event; PMA, primary major amputation. a Indications: bypass stenosis/need for extension in 10, bypass thrombectomy in 9, wound infection in 9, and hematoma in 5.
and 108 (86%) had more than one runoff vessel. Conservative treatment (n ¼ 118) was applied mostly in patients with Rutherford stage 5 (73; 62%), very distal occlusive disease, which could not be classified due to the TASC classification (51; 43%; Fig 2) and one runoff vessel (49; 42%). Finally, 25 of 30 (83%) undergoing PMA were in Rutherford stage 6 and had mostly TASC D lesions (17; 57%) and no run-off vessels (18; 60%). The multiple regression analysis revealed that patients who received a BS had an elevated risk of in-hospital death (OR, 3.32) compared with ER. The procedure-associated morbidity (periprocedural complications: 26%, MACCE: 6%; Supplementary Table III, online only) and the
increased rates of in-hospital reintervention (14%) after BS might explain this outcome. A similar trend could also be confirmed in a meta-analysis of ER vs BS of femoropopliteal arterial disease: the incidence of 30-day morbidity was 33% in the surgical arm and 17% in the endovascular intervention arm (P ¼ .007).12 In the BASIL trial, BS was also associated with a significantly higher rate of early morbidity (57%) than the rate associated with angioplasty (41%).3 Furthermore, we found that BS also carried a 1.70-times higher risk for in-hospital intervention compared with the ER. However, it remains noteworthy that 49% (129 of 284) of patients treated by BS had previous failed interventions, compared with 39% (n ¼ 251) in
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Bisdas et al 7
Bypass surgery
Composite
Coronary artery desease
endpoint
Previous vascular intervention
Renal insuff. w/o dialysis Renal insuff. with dialysis
Endpoint
Coronary artery desease
amputation
Previous myocardial infarction Previous vascular intervention
Bypass surgery
Endpoint
Renal insuff. w/o dialysis
mortality
Renal insuff. with dialysis Previous myocardial infarction
0.5
1.0
2.0
5.0
10.0
20.0
Odds Ratios and 95% Confidence Intervals Fig 3. Odds ratios (ORs; diamonds) and 95% confidence intervals (CIs; horizontal lines) are shown of the risk factors of all in-hospital end points. The ORs and CIs were deduced from multivariate logistic regression models.
the ER group. In this context, we were not able to assess the type of previous interventions performed in each group and any endovascular intervention in the bypass group might have influenced the outcomes. Patients undergoing PMA showed the highest inhospital mortality rates (10%). This outcome admonishes the necessity of revascularization attempt in CLI patients before any major amputation and fuels the current recommendation, that for limb salvage, revascularization is indicated whenever technically feasible.9,13 Of note, the real-world frequency of PMA is not well addressed in the literature. An estimated 10% to 40% of CLI patients require PMA due to overwhelming infection or unreconstructable vascular disease.1 The incidence of PMA was much lower in our registry, and only 3% of the patients received this as the first-line treatment. Taylor et al14 reported that compared with young healthy patients, patients with limited preoperative ambulatory ability, age $70 years, dementia, endstage renal disease, and advanced CAD perform poorly after PMA and cannot wear a leg prothesis. In our registry, the median age of patients undergoing PMA was 73 years, CAD was present in 50%, 10% were on dialysis, and a previous stroke had occurred in 30% (Table I). These comorbidities are surely relevant contributors to all-cause mortality in this group of patients. We are not able to deliver any data on postoperative functional use of leg
prosthesis because such outcomes were not documented in the registry. Of note, 33% of our patients undergoing PMA had no ambulatory ability at discharge (Supplementary Table III, online only). In our study, gender, diabetes mellitus, obesity, TASC classification, and run-off status did not influence any multivariate model. In contrast, renal insufficiency and PMI <6 months were identified as crucial risk factors for poor in-hospital outcomes (Fig 3). Coexistence of CAD showed 2.96-times and 2.75-times higher risk for the composite end point and MACCE, respectively. Hobbs et al15 observed that more than one-third of CLI patients sustain a perioperative myocardial injury, many of which are not clinically apparent. Several studies have also reported that CLI patients are suboptimally treated with lipid-lowering and antiplatelet medications. Statins were used in 46% of CLI patients in the Coronary Artery Revascularization Prophylaxis (CARP) and PREVENT III trials, but this rate was much lower (36%) in the BASIL trial.3,10,16 In our registry, more than the half of the patients regardless treatment strategy were receiving statin therapy, and most received at least acetylsalicylic acid (Supplementary Table III, online only) in conformance with the current guidelines on management of patients undergoing noncardiac surgery.17 This seems to be contradictory to previously published data from 2009 to 2011 that originated from the largest public
JOURNAL OF VASCULAR SURGERY --- 2015
8 Bisdas et al
health insurance in Germany (41,882 patients).11 These data, however, represent all German centers dealing with CLI patients, whereas only of centers of excellence participated in CRITISCH registry. Finally, another interesting result of the registry was that impaired renal function without dialysis was crucial risk factor for in-hospital death (OR, 6.34), hemodynamic failure (OR, 1.55), and MACCE (OR, 2.09). Few studies have investigated any correlation of impaired GFR to poor outcomes after treatment for CLI, especially after ER. Here, Ortmann et al18 observed that CLI patients at all levels of renal impairment benefit from revascularization (endovascular or surgical technique) compared with medical therapy alone. Thus, they suggested that a revascularization attempt should not be withheld in CLI patients with renal disease.18 This analysis has some limitations that are inherent to all registries. The exact selection criteria for each treatment modality that were, principally, left to the discretion of the responsible physician could not be assessed. In any case, all treating physicians provided the best treatment strategy based on their judgment. The design of this study was nonrandomized, and there was a relevant heterogeneity among the groups. However, any discrepancies between comorbidities, symptoms, and anatomy (eg, TASC classification) could be weighted and balanced by the multivariate regression analysis. Inclusion of patients was based only on Rutherford stages including ABI (due to its widespread use). No implementation of transcutaneous partial pressure of oxygen (compared with clinical trials) was done. This may cause bias, especially in diabetic patients, as well as by rating the severity of leg ischemia. In this context, we were also not able to identify any infection-associated amputation with, however, functional vascular reconstruction. The registry was designed to include all consecutive patients based on the duration of their symptoms and the bedside clinical examination by board-certified vascular surgical or medical specialists. No provision was made for a clinical event committee. The responsible physician assessed and reported the angiographic and ultrasound findings and whether an end point was met. This registry represents the current practice in 27 leading vascular centers and includes any available new endovascular technology applied during the recruitment period; however, these results may not represent all vascular centers of the country. Despite implementation of TASC classification, additional information about specific characteristics of the lesions was not recorded. Similarly, we were not able to deliver any details (diameter, length, etc) about the devices used. Any comparison of the in-hospital outcomes between different therapeutic options within the same group (eg, plain angioplasty vs primary stenting) was outside of the topics at this phase of the study. Our initial assumption that an equal number of patients would undergo ER or BS could not be confirmed. Hence, the test of the primary efficacy end point AFS probably has reduced power. Finally, we want to point out again that all
of the P values are to be interpreted again in an explorative sense and that the results need to be confirmed in prospective confirmatory studies. CONCLUSIONS The CRITISCH registry revealed that ER is at present the first-line treatment strategy of choice (54%) in CLI patients among the participating German vascular centers. BS received one in four patients with CLI in this registry, but compared with ER was associated with higher risk for in-hospital death and reintervention. However, these were only in-hospital outcomes, and the 2-year results are awaited. Main risk factors for the composite end point of in-hospital amputation or death, or both, were CAD and PMI <6 months. Patients with PMI <6 months or renal insufficiency without dialysis carry the highest risk for poor in-hospital outcomes. AUTHOR CONTRIBUTIONS Conception and design: TB, GT Analysis and interpretation: TB, MB, GT Data collection: TB, MB, GT Writing the article: TB, MB, GT Critical revision of the article: TB, MB, GT Final approval of the article: TB, MB, GT Statistical analysis: MB Obtained funding: GT Overall responsibility: GT REFERENCES 1. Norgren L, Hiatt WR, Dormandy JA, Nehler MR, Harris KA, Fowkes FG; on behalf of the TASC II Working Group. Intersociety consensus for the management of peripheral arterial disease. J Vasc Surg 2007;45:5-67. 2. Hirsch AT, Criqui MH, Treat-Jacobson D, Regensteiner JG, Creager MA, Olin JW, et al. Peripheral arterial disease detection, awareness, and treatment in primary care. JAMA 2001;286:1317-24. 3. Adam DJ, Beard JD, Cleveland T, Bell J, Bradbury AW, Forbes JF, et al. Bypass versus angioplasty in severe ischaemia of the leg (BASIL): multicentre, randomised controlled trial. Lancet 2005;366:1925-34. 4. Diehm N, Baumgartner I, Jaff M, Do DD, Minar E, Schmidlii J, et al. A call for uniform reporting standards in studies assessing endovascular treatment for chronic ischemia of lower limb arteries. Eur Heart J 2007;28:798-805. 5. Farber A, Rosenfield K, Menard M. The BEST-CLI trial: a multidisciplinary effort to assess which therapy is best for patients with critical limb ischemia. Tech Vasc Interv Radiol 2014;17:221-4. 6. Jones WS, Dolor RJ, Hasselblad V, Vemulapalli S, Subherwal S, Schmit K, et al. Comparative effectiveness of endovascular and surgical revascularization for patients with peripheral artery disease and critical limb ischemia: systematic review of revascularization in critical limb ischemia. Am Heart J 2014;167:489-98. 7. Conte MS, Geraghty PJ, Bradbury AW, Hevelone ND, Lipsitz SR, Moneta GL, et al. Suggested objective performance goals and clinical trial design for evaluating catheter-based treatment of critical limb ischemia. J Vasc Surg 2009;50:1462-73. 8. Bisdas T, Stachmann A, Weiss K, Borowski M, Grundmann RT, Torsello G; for the CRITISCH collaborators. German national registry for first-line treatment strategies in patients with critical limb ischemia (CRITISCH registry): study design and current status. Gefaesschirurgie 2014;19:135-40. 9. Allie DE, Hebert CJ, Lirtzman MD, Wyatt CH, Keller VA, Khan MH, et al. Critical limb ischemia: a global epidemic. A critical analysis of
JOURNAL OF VASCULAR SURGERY Volume -, Number -
10.
11.
12.
13.
14.
15.
16.
17.
18.
current treatment unmasks the clinical and economic costs of CLI. EuroIntervention 2005;1:75-84. Conte MS, Bandyk DF, Clowes AW, Moneta GL, Seely L, Lorenz TJ, et al. Results of PREVENT III: a multicentre, randomized trial of edifoligide for the prevention of vein graft failure in lower extremity bypass surgery. J Vasc Surg 2006;43:742-51. Malyar N, Fürstenberg T, Wellmann J, Meyborg M, Lüders F, Gebauer K, et al. Recent trends in morbidity and in-hospital outcomes of in-patients with peripheral arterial disease: a nationwide populationbased analysis. Eur Heart J 2013;34:2706-14. Antoniou GA, Chalmers N, Georgiadis GS, Lazarides MK, Antoniou SA, Serracino-Inglott F, et al. A meta-analysis of endovascular versus surgical reconstruction of femoropopliteal arterial disease. J Vasc Surg 2013;57:242-53. Anderson JL, Adams CD, Antman EM, Bridges CR, Califf RM, Casey DE, et al. 2011 ACCF/AHA focused update incorporated into the ACC/AHA 2007 guidelines for the management of patients with unstable angina/noneST-elevation myocardial infarction: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. Circulation 2011;123:426-579. Taylor SM, Kalbaugh CA, Blackhurst DW, Hamontree SE, Cull DL, Messich HS, et al. Preoperative clinical factors predict postoperative functional outcomes after major lower limb amputation: an analysis of 553 consecutive patients. J Vasc Surg 2005;42:227-35. Hobbs SD, Yapanis M, Burns PJ, Wilmink AB, Bradbury AW, Adam DJ. Perioperative myocardial injury in patients undergoing surgery for critical limb ischemia. Eur J Vasc Endovasc Surg 2005;29:301-4. Raghunathan A, Rapp JH, Littooy F, Santilli S, Krupski WC, Ward HB, et al. Postoperative outcomes fro patients undergoing elective revascularization for critical limb ischemia and intermittent claudication: a subanalysis of the Coronary Artery Revascularization Prophylaxis (CARP) trial. J Vasc Surg 2006;43:1175-82. Fleisher LA, Fleischmann KE, Auerbach AD, Barnason SA, Beckman JA, Bozkurt B, et al. 2014 ACC/AHA guideline on perioperative cardiovascular evaluation and management of patients undergoing noncardiac surgery: executive summary. Circulation 2014;130:2215-45. Ortmann J, Gahl B, Diehm N, Dick F, Traupe T, Baumgartner I. Survival benefits of revascularization in patients with critical limb ischemia and renal insufficiency. J Vasc Surg 2012;56:737-45.
Bisdas et al 9
APPENDIX The CRITISCH collaborators are: Farzin Adili, MD (Darmstadt), Kai Balzer, MD (Bonn), Thomas Betz, MD (Regensburg), Arend Billing, MD (Offenbach), Dittmar Böckler, MD (Heidelberg), Daniel Brixner, MD (Offenbach), Sebastian E. Debus, MD (Hamburg), Konstantinos P. Donas (Münster), Hans-Henning Eckstein, MD (Munich), Hans-Joachim Florek, MD (Freital), Asimakis Gkremoutis, MD (Frankfurt), Reinhardt Grundmann, MD, Thomas Hupp, MD (Stuttgart), Tobias Keck, MD (Lübeck), Joachim Gerß, PhD (Münster), Wojciech, Klonek, MD (Cloppenburg), Werner Lang, MD (Erlangen), Ute Ludwig (Munich), Björn May (Freital), Alexander Meyer, MD (Erlangen), Bernhard Mühling, MD (Biberach), Alexander Oberhuber, MD (Düsseldorf), Holger Reinecke, MD (Münster), Christian Reinhold, MD (Mönchengladbach), Ralf-Gerhard Ritter, MD (Bielefeld), Hubert Schelzig, MD (Düsseldorf), Christian Schlensack, MD (Tübingen), Thomas Schmitz-Rixen, MD (Frankfurt), Karl-Ludwig Schulte, MD (Berlin), Matthias Spohn, MD (Bamberg), Konstantinos Stavroulakis, MD (Münster), Markus Steinbauer, MD (Regensburg), Martin Storck, MD (Karlsruhe), Matthias Trede, MD (Bremen), Barbara Weis-Müller, MD (Mönchengladbach), Heiner Wenk (Bremen), Thomas Zeller, MD (Bad Kronzingen), and Alexander Zimmermann, MD (Munich).
Submitted Feb 24, 2015; accepted Apr 28, 2015.
Additional material for this article may be found online at www.jvascsurg.org.
JOURNAL OF VASCULAR SURGERY --- 2015
9.e1 Bisdas et al
Supplementary Table I (online only). Trial assessment requirements Variable Demography Medical history Physical exam Comorbidities ABI Rutherford staging Inclusion evaluation Informed consent Medicationa Duplex ultrasound Angiography Procedure details Hospital FU visits Telephone FU
Baseline X X X X X X X X Xb X
Procedure
Discharge
6 months
12 months
24 months
Unscheduled
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X Xc
X X Xc
X X Xc
X X
X X
X X
X
X X X
ABI, Ankle-brachial index; FU, follow-up. a Aspirin, clopidogrel, prasugrel, other anticoagulants, statins. b In case of patients without preoperative diagnostic angiography (group II-IV), the findings of the magnetic resonance or computed tomography angiography were used. c Angiography during surveillance was performed only in case of reintervention or according to the physician’s discretion.
Supplementary Table II (online only). Procedural characteristics in the first three intervention groups
Group Group I (n ¼ 642) ER
Femoral vessels (n ¼ 347)
Popliteal vessels (n ¼ 272)
Characteristic
Characteristic
No. (%)
No. (%)
Tibial vessels (n ¼ 368) Characteristic
No. (%)
PTA only 125 (36.0) PTA only 104 (38.2) PTA only 259 (70.4) PTA þ stent 141 (40.6) PTA þ stent 90 (33.1) PTA þ stent 24 (6.5) DCB 63 (18.2) DCB 53 (19.5) DCB 36 (9.8) DCS 6 (1.7) DCS 2 (0.7) DCS 12 (3.3) Other 38 (11.0) Other 32 (11.8) Other 22 (6.0) Failed to cross 19 (5.5) Failed to cross 17 (6.3) Failed to cross 33 (9.0) Group II Leg vein 151 (53.2) (n ¼ 284) BS Arm vein 11 (3.9) Combined leg 3 (1.1) and arm vein Dacron 32 (11.3) PTFE 71 (25) Other 22 (7.7) Group III Common FAP 51 (40.1) (n ¼ 126) FAP Deep FAP 9 (7.1) Common and 66 (52.4) deep FAP Hybrid procedurea 28 (22.2) Patchplasty only 98 (77.8)
Previous bypass (n ¼ 26) Characteristic
No. (%)
PTA only 8 (30.8) PTA þ stent 3 (11.5) DCB 2 (7.7) DCS 2 (7.7) Other 12 (46.2) Failed to cross 1 (3.8)
BS, Bypass surgery; DCB, drug-coated balloon; DCS, drug-coated stent, ER, endovascular repair; FAP, femoral/profundal artery patchplasty; PTA, percutaneous transluminal angioplasty, PTFE, polytetrafluoroethylene. a These patients were not included in the ER or BS groups.
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Bisdas et al 9.e2
Supplementary Table III (online only). Discharge data of the study groups Group I, ER
Variablea In-hospital stay, median (IQR) days Periprocedural complications Hematoma Wound infection Lymph fistula/lymphocele Pneumonia Dialysis Compartment syndrome Deep vein thrombosis Acute coronary event Stroke Discharge Home Other institution Medication at discharge Statins Acetylsalicylic acid Vitamin K antagonists Clopidogrel Prasugrel New anticoagulants
7 58 16 26
(15) (9) (2) (4) 0 7 (1) 3 (<1) 0 0 18 (3) 3 (<1)
Group II, BS
Group III, FAP
(10) (26) (6) (11) (7) (<1) (<1) (2) 0 7 (2) 1 (<1)
13 (11) 23 (18) 0 6 (5) 12 (10) 2 (2) 1 (1) 0 0 4 (3) 1 (1)
563 (88) 72 (11)
216 (76) 58 (20)
100 (79) 22 (17)
359 490 126 340 4 93
176 207 94 58 1 48
(56) (76) (20) (53) (1) (14)
15 73 16 31 21 1 1 5
(62) (73) (33) (20) (0) (17)
74 98 18 44 0 21
(59) (78) (14) (35) (0) (17)
Group IV, Conservative
Group V, PMA
P value
10 14 3 7
16 10 1 8
(11) (33) (3) (27) 0 1 (3) 0 0 0 1 (3) 0
<.001
99 (84) 14 (12)
17 (57) 10 (33)
<.001 <.001
72 82 19 27 0 27
16 22 2 6 0 13
.446 .443 <.001 <.001 .766 <.001
(21) (12) (3) (6) 0 3 (3) 0 0 0 2 (2) 0
(61) (69) (16) (23) (0) (23)
(53) (73) (7) (20) (0) (43)
BS, Bypass surgery; ER, endovascular repair; FAP, femoral/profundal artery patchplasty; PMA, primary major amputation. a Data are shown as number (%) unless indicated otherwise.
Supplementary Table IV (online only). Results of the univariate logistic regression models for each in-hospital end point, given by odds ratios (ORs) and referring 95% confidence intervals (CIs) of all possible risk factors Variable BS vs ER Age in years Male gender Renal insufficiency without dialysis vs normal renal function Dialysis vs normal renal function Diabetes mellitus Current tobacco use Obesity CAD PMI <6 months Previous vascular intervention TASC B vs TASC A TASC C vs TASC A TASC D vs TASC A Nonclassifiable vs TASC A No vs 3 runoff vessels 1 vs 3 runoff vessels 2 vs 3 runoff vessels
Composite end point 1.64 1.02 1.77 1.92
(0.87-3.10) (0.99-1.05) (0.98-3.21) (1.09-3.40)
Amputation 1.14 0.97 2.81 1.15
(0.52-2.46) (0.97-1.03) (1.24-6.40) (0.56-2.38)
Death 3.07 1.05 1.20 5.93
(1.06-8.94) (1.00-1.09) (0.51-2.80) (2.19-16.1)
Hemodynamic failure 0.64 1.01 0.98 1.62
(0.40-1.03) (0.99-1.02) (0.73-1.31) (1.21-2.18)
MACCE 1.50 1.04 1.57 2.54
(0.77-2.92) (1.01-1.07) (0.86-2.87) (1.44-4.46)
Reintervention 1.88 1.00 0.98 0.61
(1.21-2.94) (0.98-1.01) (0.65-1.48) (0.39-0.96)
3.42 (1.61-7.25)
3.96 (1.77-8.84)
2.68 (0.51-14.0)
2.00 (1.25-3.21)
0.94 (0.28-3.22)
0.52 (0.22-1.24)
1.36 0.60 1.35 2.06 4.41 1.48
(0.81-2.28) (0.28-1.29) (0.69-2.64) (1.21-3.50) (2.04-9.53) (0.88-2.47)
1.63 0.56 1.03 2.14 4.15 2.03
(0.87-3.07) (0.22-1.44) (0.43-2.49) (1.12-4.08) (1.66-10.4) (1.08-3.81)
1.24 0.77 1.93 2.17 6.00 1.10
(0.56-2.73) (0.26-2.28) (0.76-4.91) (0.95-4.95) (2.16-16.7) (0.50-2.45)
1.23 0.57 1.13 1.13 1.88 1.40
(0.94-1.63) (0.39-0.85) (0.76-1.66) (0.86-1.49) (1.03-3.41) (1.06-1.85)
1.33 0.77 1.16 3.82 8.57 0.97
(0.78-2.27) (0.37-1.60) (0.56-2.41) (2.06-7.08) (4.26-17.2) (0.57-1.68)
0.81 1.13 0.90 1.02 0.85 1.67
(0.54-1.21) (0.70-1.84) (0.50-1.61) (0.68-1.51) (0.30-2.39) (1.12-2.49)
0.62 1.19 1.38 0.87
(0.17-2.27) (0.35-4.07) (0.48-4.02) (0.27-2.81)
0.84 1.19 1.05 0.84
(0.21-3.45) (0.29-4.87) (0.30-3.65) (0.22-3.26)
0.42 1.78 2.68 1.28
(0.03-6.81) (0.18-17.4) (0.35-20.6) (0.14-11.6)
0.73 0.62 1.16 2.01
(0.37-1.43) (0.30-1.28) (0.65-2.08) (1.10-3.67)
0.62 0.88 1.33 0.71
(0.17-2.27) (0.24-3.21) (0.46-3.89) (0.21-2.36)
0.97 0.74 1.54 0.71
(0.38-2.44) (0.27-2.07) (0.68-3.50) (0.28-1.79)
2.40 (0.88-6.56) 1.85 (0.70-4.87) 1.19 (0.39-3.62)
1.86 (0.59-5.86) 1.55 (0.52-4.63) 0.99 (0.27-3.55)
5.61 (0.70-44.7) 3.27 (0.42-25.7) 3.33 (0.39-28.8)
3.13 (1.83-5.34) 1.98 (1.19-3.31) 1.42 (0.80-2.52)
1.66 (0.63-4.32) 1.36 (0.55-3.38) 0.99 (0.34-2.82)
1.70 (0.85-3.40) 1.17 (0.60-2.27) 1.10 (0.52-2.32)
BS, Bypass surgery; CAD, coronary artery disease; ER, endovascular repair; MACCE, major adverse cardiovascular and cerebral event; PMI, previous myocardial infarction; TASC, TransAtlantic Inter-Society Consensus.
JOURNAL OF VASCULAR SURGERY --- 2015
9.e3 Bisdas et al
Supplementary Table V (online only). Coefficients of multivariate logistic regression modelsa Coefficients Intercept BS vs ER Age in years Male gender Renal insufficiency without dialysis vs normal renal function Dialysis vs normal renal function Diabetes mellitus Current tobacco use Obesity CAD PMI <6 months Previous vascular intervention TASC B vs TASC A TASC C vs TASC A TASC D vs TASC A Non-classifiable vs TASC A No vs 3 runoff vessels 1 vs 3 runoff vessels 2 vs 3 runoff vessels
Composite end point
Amputation
Death
Hemodynamic failure
MACCE
Reintervention
4.01 0.50
4.16
5.98 1.20 *
1.61
4.11
2.41 0.53
*
* 0.05
1.85
0.44
0.74
0.48
*
1.20
14.06
0.62
0.41
0.58
*
0.47 1.09 1.30
0.41 1.18 0.61
2.00
* 0.35 * * * * * * *
1.01 1.68 0.38
BS, Bypass surgery; CAD, coronary artery disease; ER, endovascular repair; MACCE, major adverse cardiovascular and cerebral event; PMI, previous myocardial infarction; TASC, TransAtlantic Inter-Society Consensus. a Asterisks mark risk factors that have been detected by univariate logistic regression but not by multivariate logistic regression. The multivariate logistic regression model for the composite end point, for instance, is given by log[P/(1 P)] ¼ 4.01 þ 0.50 BS þ 1.09 CAD þ 1.30 PMI, where P ¼ P(composite end point) is the probability of an in-hospital amputation or death, or both. The model suggests that the chance of an amputation or death during in-hospital stay was 1.64 (exp[0.50]) times higher for patients who received BS compared with those who underwent ER. Similarly, the chance was estimated to be 2.96 (exp[1.09]) and 3.67 (exp[1.30]) times higher for patients with CAD and PMI than for their respective counterpart. Through the coefficients, odds ratios (ORs) of each end point for any combination of covariates can be obtained. For instance, considering the composite end point, the ORs of a patient with BS, CAD, and PMI vs a patient without them is estimated to be exp(0.5 þ 1.09 þ 1.3) ¼ 17.99.