2018 Volume 82 Issue 1 Pages 267-274
Background: Most patients with critical limb ischemia (CLI) exhibit severe comorbidities accompanied by frailty. This study assessed and risk-stratified mortality after infrainguinal bypass (IB) in CLI and investigated the effects of frailty.
Methods and Results: The study retrospectively reviewed 107 consecutive CLI patients who had undergone de novo IB due to atherosclerotic disease. Data regarding patient age, comorbidities, laboratory data, and functional status were collected; functional status was evaluated using the Barthel index (BI) and nutritional status was evaluated using albumin concentrations and body mass index (BMI). Mean (±SD) BI and BMI were 75±16 and 22±4 kg/m2, respectively. BI (hazard ratio [HR] 0.96; 95% confidence interval [CI] 0.94–0.99, P=0.004), BMI (HR 0.85; 95% CI 0.75–0.95, P=0.003), atrial fibrillation (AF; HR 5.31; 95% CI 2.12–13.30, P<0.001), and ejection fraction (EF; HR 0.94; 95% CI 0.91–0.98, P=0.003) were independent predictors of mortality. Patients were divided into 2 groups based on BI (BI >75, n=71; and BI <70, n=36). Survival after IB was significantly lower for the lower BI group (P<0.001, log-rank test). After propensity score matching, post-IB survival remained significantly lower in the lower BI group (P=0.02).
Conclusions: BI, BMI, AF, and EF were independently associated with all-cause mortality after IB for CLI. BI and BMI may be useful in identifying and optimizing treatment for high-risk frail patients.
Critical limb ischemia (CLI) is a limb-threatening condition that requires revascularization to relieve ischemic pain, heal ischemic ulcers, prevent limb loss, and preserve ambulatory status and quality of life. Surgical bypass using an autogenous vein is the standard treatment for CLI, and numerous studies have documented graft- and limb-related outcomes.1,2 The application of endovascular therapy (EVT) for CLI has recently increased; this technique is particularly attractive for patients who are at increased risk for surgery. Some studies have advocated for the preferential use of EVT as an initial approach in CLI.3,4 However, the first-line treatment for the revascularization of CLI remains contentious. The TransAtlantic Inter-Society Consensus (TASC) recommends that the profile and extent of arterial lesions should be used to select EVT or bypass surgery,5 whereas the Bypass vs. Angioplasty in Severe Ischaemia of the Leg (BASIL) trial group suggested that patients with a life expectancy >2 years and good vein quality are better served using autogenous vein bypass surgery as the initial treatment.6
Frailty is defined as a biological syndrome of decreased reserve and resistance to stressors resulting from cumulative declines across multiple physiological systems, which causes vulnerability to adverse outcomes.7 Some studies have validated multiple assessments to identify frail patients based on the recognition of numerous physiological domains (e.g., comorbidity, physical function, nutrition, cognition, geriatric syndrome, and social vulnerability), but there is no single gold standard measure of frailty.8 CLI patients generally exhibit some severe comorbidities and poor functional status because they cannot engage in sufficient exercise due to the ischemic symptoms (e.g., rest pain, tissue loss, or gangrene) or cardiopulmonary dysfunction. Their physiological reserves are decreased in many cases, and many patients in the CLI population experience concomitant frailty.
Bypass surgery is generally a more invasive treatment than EVT, and careful patient selection for surgery is essential. Therefore, it is important for vascular surgeons to recognize the degree of a patient’s frailty and determine risk stratification prior to surgical revascularization. However, little information is available on the relationship between the prognosis of CLI patients and their frailty. Therefore, in the present study we investigated the effects of frailty and other risk factors that affect mortality after infrainguinal bypass (IB) in CLI patients at a single institute in Japan. We used the Barthel frailty index as an indicator of activities of daily living (ADL) to measure frailty.9 This score is valid and reliable for the assessment of disability in stroke patients.10 The aim of the present study was to determine whether the Barthel index (BI) is useful for risk stratification after bypass surgery.
This clinical investigation compared preoperative medical risk factors, including frailty, with the outcomes of 107 consecutive CLI patients who underwent de novo IB in the Division of Vascular Surgery, Department of Surgery, Nagoya University Graduate School of Medicine from January 2007 to October 2016. One limb was included per enrolled patient. The prevalence of bilateral CLI was 10% (11/107). In case of bilateral CLI, the first treated limb was enrolled in the study. Patients whose surgeries were attributable to acute arterial occlusion (n=4) or non-atherosclerotic diseases (n=14) and the 1 patient who underwent primary major amputation were excluded.
Data collection was performed using a prospectively collected database and scrutinized retrospectively via reviews of the patient records. This study was conducted in accordance with the mandates of the Declaration of Helsinki. The Nagoya University School of Medicine Institutional Review Board approved the study, and all patients provided written informed consent prior to surgery and data collection.
Revascularization Procedure and Follow-upAll CLI patients in our institution were generally considered for revascularization. The authors performed angioplasty and surgical bypass. Results of laboratory tests, cardiovascular function (peripheral angiography, electrocardiography, coronary angiography, and echocardiography), respiratory function, and other parameters were examined routinely. The treatment of all patients was discussed among vascular surgeons, and the best revascularization strategy was decided on after consideration of each patient’s comorbidities and ADL, the availability of a suitable venous conduit, the anatomical features of arterial disease, and the severity of ischemic wounds. In many cases, CLI patients exhibit extensive atherosclerosis and long occlusion. Therefore, bypass surgery is preferred as a first-line treatment, especially for infrapopliteal lesions. We generally perform EVT in TASCII A and B and bypass surgery in TASCII C and D for the revascularization of femoropopliteal lesions.11 We did not perform bypass in patients with severe dementia or those who were long-term bedridden because of neurogenic deficits, broad infection above the ankle, or refusal of revascularization.
The technical details of the bypass procedure have been published previously.12 Briefly, a distal anastomotic site was selected based on the optimal run-off vessel to the ischemic wounds. The greater saphenous vein was used as the first-choice conduit after preoperative duplex scan imaging. The lesser saphenous vein, an arm vein, was used if the greater saphenous vein was not accessible, and a prosthetic graft was used in cases of femoropopliteal (above the knee) bypass. Angiography was performed after completion of the anastomoses to check for graft stenosis and anastomotic site problems. Moreover, graft flow was assessed with ultrasonography using a linear transducer. Measurements were obtained using a transit time flowmeter (300-Series Flowmeter; Transonic Systems, Ithaca, NY, USA). Heparin was infused for 24 h, and prostaglandin E1 was infused for several days postoperatively. All patients received antiplatelet drugs. Generally, the ankle-brachial index (ABI) and skin perfusion pressure (SPP) were measured nearly 1 week after bypass grafting.
Routine follow-up consisted of monthly postoperative visits for 3 months, followed by visits at 3-monthly intervals for 2 years and at 6-monthly intervals thereafter. Routine surveillance included ABI, SPP, and duplex ultrasound (US). Duplex scanning of the graft was the primary method for detecting occult graft lesions. A focal increase in graft velocity (peak systolic velocity >300 cm/s) was considered significant and followed by arteriography and revision when appropriate.
DefinitionsCLI was defined as tissue loss or gangrene and rest pain lasting >2 weeks, in accordance with TASC guidelines.5 CLI was associated with ankle pressure (<50 mmHg for rest pain, <70 mmHg for tissue loss) or toe pressure (<30 mmHg for rest pain; <50 mmHg for tissue loss). SPP was measured when these measurements could not be obtained due to intractable rest pain or a non-compressible artery secondary to severe calcification, and an SPP <40 mmHg was defined as indicative of a critically ischemic limb.13
Coronary artery disease (CAD) was defined as stable angina with documented CAD, a history of any revascularization of the coronary arteries, or previous myocardial infarction. Atrial fibrillation (AF) was defined as chronic AF and was diagnosed using electrocardiography upon admission. Cerebrovascular disease (CVD) was defined as a history of stroke and/or cerebral hemorrhage. Diabetes mellitus (DM) was diagnosed in patients who were taking hypoglycemic drugs or self-injected insulin. A current smoker was defined as anyone with a history of smoking within 1 month prior to bypass surgery.
BIThe BI is a 10-item scale that assesses a patient’s ability to feed, groom, bathe, use the toilet, dress, walk, transfer, and climb stairs, as well as fecal incontinence and urinary incontinence (Table 1).9 The BI is calculated by adding 5, 10, or 15 points for the presence of each variable (final score 0–100 points). In the present study, the BI was calculated by physical therapists or nurses, who were blind to the study, as an indicator of ADL upon admission, prior to revascularization.
| Bowels | Transfer | |
| 0=incontinent | 0=unable, no balance while sitting | |
| 5=occasional accident | 5=major help, can sit | |
| 10=continent | 10=minor help | |
| Bladder | 15=independent | |
| 0=incontinent, or catheterized and unable to manage | Mobility | |
| 5=occasional accident (maximum once per 24 h) | 0=immobile | |
| 10=continent (for >7 days) | 5=wheelchair independent, including corners etc. | |
| Toilet use | 10=walks with help of 1 person | |
| 0=dependent | 15=independent (but may use any aid, e.g., stick) | |
| 5=needs some help, but can do something alone | Dressing | |
| 10=independent | 0=dependent | |
| Feeding | 5=needs help, but can do about half unaided | |
| 0=unable | 10=independent (including buttons, zips etc.) | |
| 5=needs help cutting, spreading butter etc. | Grooming | |
| 10=independent (food provided within reach) | 0=needs help with personal care | |
| Stairs | 5=independent (face, hair, teeth, shaving) | |
| 0=unable | Bathing | |
| 5=needs help (verbal, physical, carrying aid) | 0=dependent | |
| 10=independent up and down | 5=independent (or in shower) |
BI, barthel index.
All-cause mortality after IB surgery was set as the primary endpoint. The present study also investigated whether the BI was an independent predictor of mortality after IB. The following variables were collected for analysis: age, comorbidities (hypertension, hyperlipidemia, DM, CAD, AF, CVD, hemodialysis-dependent renal failure, current smoker), laboratory data (white blood cell count, hematocrit, albumin and C-reactive protein concentrations, ejection fraction determined via US, volume capacity, and forced expiratory volume in 1 s [FEV1], determined via spirometry), medication, and lower limb status (Rutherford classification). All variables, including the laboratory data, were obtained prior to surgery.
Statistical AnalysisData are expressed as the mean±SD for continuous variables or as percentages for dichotomous variables. Overall survival (i.e., freedom from all-cause mortality) was estimated using Kaplan-Meier curves. Cox multivariate regression analysis was used to determine the predictors of mortality. Any covariates with P<0.05 in the univariate analysis were assessed using multivariate analysis.
Patients were divided into 2 groups based on mean BI (i.e., those with a BI >75 [n=71] and those with a BI <70 [n=36]) to clarify the effects of the BI. Continuous variables were compared between groups using unpaired t-tests. The Chi-squared test was used to compare proportions between 2 groups. Propensity score matching methods were used to minimize background differences between these 2 groups and to better evaluate the effect of the BI. A logistic regression analysis model was created to estimate the likelihood of mortality in the higher and lower BI groups. The covariates entered into the model included factors that were identified as significantly different. The propensity score was calculated according to the logistic regression model and used for 1:1 matching according to a <0.1 difference in the propensity score between 2 groups. The overall survival of the 2 groups was analyzed using Kaplan-Meier curves and compared using a log-rank test.
In all analyses, P<0.05 was considered significant. Statistical analyses were performed using the SPSS version 24 (IBM Corp., Armonk, NY, USA).
Table 2 lists overall patient characteristics. Mean patient age was 69±9 years, 72 patients (67%) had DM, 62 (58%) had CAD, and 43 (40%) were dependent on hemodialysis (HD). The mean BI was 75±16, and the mean body mass index (BMI) was 22±4 kg/m2.
| Patient status | |
| Age (years) | 69±9 |
| No. men (%) | 70 (65) |
| BI | 75±16 |
| BMI (kg/m2) | 22±4 |
| Comorbidities | |
| Hypertension | 84 (79) |
| Hyperlipidemia | 54 (51) |
| DM | 72 (67) |
| CAD | 62 (58) |
| AF | 20 (19) |
| Cerebrovascular disease | 26 (24) |
| Hemodialysis | 43 (40) |
| Current smoker | 25 (23) |
| Laboratory data | |
| WBC count (/μL) | 7,600±2,900 |
| Hematocrit (%) | 33.2±5.3 |
| Albumin (g/dL) | 3.2±0.6 |
| CRP (mg/dL) | 3.0±4.4 |
| EF (%) | 62.8±10.0 |
| VC (%) | 93±21 |
| FEV1 (%) | 74±13 |
| Medication use | |
| Cilostazol | 48 (45) |
| Statin | 38 (36) |
| β-blocker | 26 (24) |
| Lower limb status | |
| Rutherford classification | |
| 4 | 12 (11) |
| 5 | 70 (65) |
| 6 | 25 (24) |
| ABI before bypass | 0.42±0.31 |
| SPP before bypass (mmHg) | 18±8 |
Data are presented as n (%) or as the mean±SD. ABI, ankle-brachial index; AF, atrial fibrillation; BI, barthel index; BMI, body mass index; CAD, coronary artery disease; CRP, C-reactive protein; DM, diabetes mellitus; EF, ejection fraction; FEV1, forced expiratory volume in 1 second; SPP, skin perfusion pressure; VC, volume capacity; WBC, white blood cell count.
Table 3 summarizes the details of the bypass surgery. In all, 22 limbs (21%), of which 14 had aortoiliac lesions and 8 had femoropopliteal lesions, also underwent inflow revascularization via EVT at the time of surgery. The outflow target vessels were crural in 45 limbs (42%) and inframalleolar in 36 (34%). A single vein graft was used in 83 bypasses (78%), and prosthetic grafts were used in all femoropopliteal (above the knee) bypasses. The postoperative ABI and SPP were 0.89±0.21 and 49±21 mmHg, respectively. The details of distal bypass are summarized in Table 3.
| Conduit | |
| Prosthetic graft | 6 (5) |
| Non-reversed vein graft | 46 (43) |
| Reversed vein graft | 21 (20) |
| In situ vein graft | 16 (15) |
| Spliced vein graft | 18 (17) |
| Inflow | |
| Common femoral artery | 47 (44) |
| Superficial femoral artery | 11 (10) |
| Deep femoral artery | 7 (7) |
| Popliteal artery (AK) | 17 (16) |
| Popliteal artery (BK) | 24 (22) |
| Tibioperoneal trunk | 1 (1) |
| Outflow | |
| Popliteal artery (AK) | 10 (9) |
| Popliteal artery (BK) | 16 (15) |
| Anterior tibial artery | 15 (14) |
| Posterior tibial artery | 21 (20) |
| Peroneal artery | 7 (7) |
| Dorsal pedis artery | 23 (21) |
| Tarsal artery | 3 (3) |
| Plantar artery | 10 (9) |
| Tibioperoneal trunk | 1 (1) |
| Sural artery | 1 (1) |
| Concomitant procedures | |
| Minor amputation | 54 (50) |
| Angioplasty of inflow vessel | 22 (21) |
| Thromboendarterectomy | 9 (8) |
| Operation time (min) | 301±73 |
| Blood loss (mL) | 268±218 |
| Mean flow of graft (mL/min) | 61±53 |
Data are presented as n (%) or as the mean±SD. AK, above the knee; BK, below the knee.
The median duration of follow-up was 707 days (range 10–2,613 days). One patient died within 30 days after de novo bypass grafting. This patient was a 92-year-old woman who received steroids for bullous pemphigoid and died suddenly due to rupture of the fragile venous graft on postoperative Day 5.
Overall survival rates at 1 and 3 years were 87.0% and 62.6%, respectively (Figure 1). Thirty-eight patients died during the follow-up period because of cardiovascular disease (n=14), respiratory disease (n=6), stroke (n=5), sepsis (n=4), bowel ischemia (n=3), graft rupture (n=1), malignancy (n=1), gastrointestinal tract perforation (n=1), or unknown causes (n=3). Univariate regression analysis revealed that age, the BI, BMI, AF, HD, serum albumin concentrations, volume capacity, and EF were independent predictors of mortality (P<0.05 for all). Multivariate analysis of these confirmed the BI (hazard ratio [HR] 0.96; 95% confidence interval [CI] 0.94–0.99; P=0.004), BMI (HR 0.85; 95% CI 0.75–0.95; P=0.003), AF (HR 5.31; 95% CI 2.12–13.30; P<0.001), and EF (HR 0.94; 95% CI 0.91–0.98; P=0.003) as predictors of mortality (Table 4).
Overall survival rates after infrainguinal bypass.
| Predictors | Univariate analysis | Multivariate analysis | ||
|---|---|---|---|---|
| HR (95% CI) | P value | HR (95% CI) | P value | |
| Age | 1.04 (1.004–1.08) | 0.03 | 1.01 (0.96–1.07) | 0.68 |
| Male sex | 1.26 (0.61–2.60) | 0.54 | ||
| BI (per 1-point increase) | 0.95 (0.93–0.97) | <0.001 | 0.96 (0.94–0.99) | 0.004 |
| BMI | 0.85 (0.77–0.94) | 0.002 | 0.85 (0.75–0.95) | 0.003 |
| Hypertension | 1.73 (0.67–4.45) | 0.26 | ||
| Hyperlipidemia | 0.82 (0.43–1.56) | 0.54 | ||
| DM | 0.68 (0.34–1.37) | 0.29 | ||
| CAD | 0.92 (0.48–1.77) | 0.80 | ||
| AF | 3.58 (1.74–7.38) | 0.001 | 5.31 (2.12–13.30) | <0.001 |
| Cerebrovascular disease | 1.15 (0.57–2.35) | 0.70 | ||
| Hemodialysis | 2.33 (1.22–4.44) | 0.01 | 1.77 (0.79–3.93) | 0.16 |
| Current smoker | 0.63 (0.29–1.38) | 0.25 | ||
| WBC count | 0.95 (0.84–1.08) | 0.44 | ||
| Hematocrit | 0.98 (0.92–1.05) | 0.58 | ||
| Albumin | 0.48 (0.26–0.88) | 0.02 | 1.16 (0.54–2.47) | 0.71 |
| CRP | 1.02 (0.95–1.10) | 0.56 | ||
| EF | 0.95 (0.92–0.98) | 0.001 | 0.94 (0.91–0.98) | 0.003 |
| VC | 0.98 (0.97–0.99) | 0.03 | 0.99 (0.97–1.01) | 0.33 |
| FEV1 | 0.99 (0.97–1.02) | 0.64 | ||
| Cilostazol use | 0.77 (0.40–1.49) | 0.44 | ||
| Statin use | 0.86 (0.44–1.67) | 0.65 | ||
| β-blocker use | 1.67 (0.81–3.44) | 0.16 | ||
| Rest pain | 0.70 (0.25–1.98) | 0.50 | ||
| Rutherford 6 | 1.90 (0.90–4.01) | 0.09 | ||
CI, confidence interval; HR, hazard ratio. Other abbreviations as in Table 2.
Groups with lower and higher BI differed significantly in terms of age (mean 73 vs. 68 years, respectively; P=0.01), BMI (mean 20.6 vs. 22.4 kg/m2, respectively; P=0.04), incidence of HD (56% vs. 32%, respectively; P=0.02) and EF (mean 59.0% vs. 64.7%, respectively; P=0.007). These factors were used to construct the propensity score model. None of the factors measured, including the aforementioned factors, was significantly different following propensity score matching (Table 5).
| Unmatched cohort | Propensity score-matched cohort | |||||
|---|---|---|---|---|---|---|
| Lower BI (n=36) |
Higher BI (n=71) |
P value | Lower BI (n=24) |
Higher BI (n=24) |
P value | |
| Age (years) | 73±8 | 68±10 | 0.01 | 71±7 | 70±12 | 0.64 |
| No. men (%) | 22 (61) | 48 (68) | 0.53 | 14 (58) | 19 (79) | 0.21 |
| BMI (kg/m2) | 20.6±4.2 | 22.4±3.9 | 0.04 | 21.9±4.4 | 22.1±3.4 | 0.79 |
| Hypertension | 28 (78) | 56 (79) | 1.00 | 19 (79) | 19 (79) | 1.00 |
| Hyperlipidemia | 18 (50) | 36 (51) | 1.00 | 14 (58) | 11 (46) | 0.56 |
| DM | 24 (67) | 48 (67) | 1.00 | 16 (67) | 15 (63) | 1.00 |
| CAD | 24 (67) | 38 (54) | 0.22 | 16 (67) | 16 (67) | 1.00 |
| AF | 10 (28) | 10 (14) | 0.12 | 8 (33) | 5 (21) | 0.52 |
| Cerebrovascular disease | 9 (25) | 17 (24) | 1.00 | 7 (29) | 6 (25) | 1.00 |
| Hemodialysis | 20 (56) | 23 (32) | 0.02 | 10 (42) | 9 (38) | 1.00 |
| Current smoker | 7 (19) | 18 (25) | 0.63 | 6 (25) | 2 (8) | 0.25 |
| WBC (/μL) | 8,100±3,300 | 7,500±2,800 | 0.38 | 8,100±2,800 | 7,300±2,000 | 0.25 |
| Hematocrit (%) | 32.1±6.0 | 33.8±4.9 | 0.13 | 32.9±6.0 | 34.6±5.1 | 0.29 |
| Albumin (g/dL) | 3.1±0.6 | 3.3±0.6 | 0.21 | 3.2±0.7 | 3.2±0.6 | 0.86 |
| CRP (mg/dL) | 3.5±5.3 | 2.8±3.6 | 0.44 | 3.0±3.7 | 3.5±3.9 | 0.70 |
| EF (%) | 59.0±9.5 | 64.7±9.9 | 0.007 | 60.6±9.3 | 61.7±9.8 | 0.72 |
| VC (%) | 90.4±20.4 | 94.6±21.9 | 0.33 | 91.7±21.7 | 87.5±25.9 | 0.55 |
| FEV1 (%) | 76.5±10.8 | 73.3±14.4 | 0.20 | 76.5±9.7 | 70.5±19.6 | 0.19 |
| Medication use | ||||||
| Cilostazol | 16 (44) | 32 (45) | 1.00 | 11 (46) | 12 (50) | 1.00 |
| Statin | 10 (28) | 28 (39) | 0.29 | 6 (25) | 7 (29) | 1.00 |
| β-blocker | 11 (31) | 15 (21) | 0.34 | 8 (33) | 6 (25) | 0.75 |
| Rutherford classification | ||||||
| 4 | 4 (11) | 8 (11) | 4 (17) | 3 (13) | ||
| 5 | 19 (53) | 51 (72) | 10 (42) | 15 (63) | ||
| 6 | 13 (36) | 12 (17) | 0.08 | 10 (42) | 6 (25) | 0.34 |
Data are presented as n (%) or as the mean±SD. Abbreviations as in Tables 2,4.
Figure 2A shows that overall survival rates 1 and 3 years after IB were significantly lower in the group with the lower BI (69% and 32%, respectively) than in the group with the higher BI (96% and 77%, respectively; P<0.001). A paired log-rank test after propensity score matching also revealed that the group with the lower BI had a significantly lower rate of survival than the group with the higher BI (P=0.02; Figure 2B).
(A) Comparison of overall survival rates between groups with a lower (<70, n=36) and higher (>75, n=71) Barthel index (BI). Mortality after infrainguinal bypass was higher in the group of patients with a lower BI (P<0.001). (B) Comparison of overall survival between groups with a lower and higher BI after propensity score matching. Propensity score matching extracted a total of 24 pairs. The paired log-rank test demonstrated that the group with a lower BI had a significantly lower overall survival rate than the group with a higher BI (P=0.02).
In the present study, we evaluated clinical predictors of all-cause mortality, including ADL and nutrition status, in CLI patients treated with IB. We demonstrated that the predictors of mortality were AF, EF, BMI, and the BI. Predictive scoring models for CLI patients were developed in the Finnvasc,14 Prevent III,15 and BASIL16 trials. However, nutritional variables, such as serum albumin level or BMI, were not systematically examined in the Finnvasc and Prevent III studies, although the BASIL scoring model identified being underweight as one of the variables. In addition, these studies did not examine patient frailty. The present study evaluated numerous factors, including nutritional factors (e.g., BMI and serum albumin concentrations) and functional factors (the BI) to determine the predictors of clinical prognosis. To the best of our knowledge, the present study is the first report that the BI, a tool for scoring frailty, is associated with the mortality of CLI patients undergoing IB.
Frailty is a syndrome that reflects decreased physiological reserves and an accumulation of comorbid conditions. Many tools are available to measure frailty using some key criteria (e.g., exercise capacity, muscle strength, and nutritional status).17 Many evaluations of exercise capacity assess gait speed or the ability to climb stairs.18 However, some CLI patients experience difficulty in walking or climbing. Recent studies have demonstrated that the proportion of CLI patients with a non-ambulatory status, defined as needing a wheelchair or being bedridden, ranged from 30% to 40%.12,19–21 These studies reported that non-ambulatory status was an independent predictor of mortality or amputation-free survival after revascularization in CLI patients. However, whether CLI patients were non-ambulatory was unclear in the previous studies,12,19–21 because some patients who were defined as non-ambulatory generally moved around using a wheelchair, but were able to walk a short distance with a cane to use the toilet. Because CLI patients are a heterogeneous population in terms of the severity of comorbidities, ischemic wounds, anatomical features, and functional capacity, a more detailed classification system is needed for these patients. The BI consists of 10 items based on “in-house” ADL, and is widely used and ease to apply. In the present study, we used this useful index to evaluate the degree of frailty of patients with CLI. Previous studies have demonstrated that the BI score is valid in stroke patients and is associated with their clinical prognosis.10,22 Recent single-center studies have demonstrated the usefulness of the BI score to estimate prognosis for elderly patients with acute coronary syndrome, acquired pneumonia, and on admission to an acute geriatric ward.23–25 However, whether the functional status assessed by this score affects the mortality of CLI patients after revascularization has not been assessed previously.
Age was not a predictor of mortality in the present study. Some recent studies reported that very elderly patients who underwent surgical bypass had good clinical outcomes.26,27 Indeed, in actual clinical settings, selection bias may occur during the selection by surgeons of patients who should undergo bypass surgery. It is necessary to at least recognize the heterogeneity between biological and chronological age when treating elderly patients. We believe that the BI is a suitable objective tool for evaluation of CLI patients.
In addition, BMI was a predictor of mortality in CLI patents after bypass surgery in the present study. Nutritional status affects BMI, which has been shown to provide a valid assessment of the clinical prognosis of CLI patients.28,29 The authors of those studies reported an inverse association between prognosis and BMI in CLI patients, which is consistent with the results of the present study.
The present study demonstrated that AF was also associated with mortality. AF has not been previously reported as a predictor of clinical outcomes in CLI patients after bypass surgery. Whether AF has a similar effect on the outcome of patients with peripheral artery disease (PAD) as PAD has on patients with AF is less well known. However, a multicenter registry reported that the prevalence of AF among patients with PAD was much higher than that in the general population and that AF is a predictor of poor 2-year cardiovascular outcomes in patients with symptomatic PAD.30,31 Moreover, a prospective, single-center study in Taiwan revealed that AF was a predictor of mortality and amputation in CLI patients who underwent EVT.32 These findings may be associated with left ventricular dysfunction and thromboembolic complications, which results in cardiovascular mortality.30 In the present study, 11 of the 20 patients with AF died during the study period: 3 died of heart failure, 3 died from an embolic event, and 5 died as a result of other causes.
The present study has several limitations. First, this study was a retrospective, single-center study. We compensated for the high probability of bias in many of the data assignments by using propensity-matched scores. Moreover, the cohort excluded patients with CLI who underwent EVT or conservative therapy. Second, the choices of procedure and medications, which may have affected the clinical outcome, depended on the surgeon. Third, the assessment of frailty, such as nutritional (BMI) or functional (BI) status, was performed on admission before surgical bypass, and thus changes in frailty over time (e.g., before the onset of CLI, after wound healing) were not taken into account; the relationship between changes in frailty status and outcomes must be elucidated in future studies. Moreover, larger multicenter studies are warranted to corroborate the findings of the present study.
The BI, BMI, AF, and EF were independently associated with all-cause mortality after IB in CLI patients. The BI and BMI may be integrated into the preoperative risk assessment of vascular patients to identify and optimize treatment for high-risk frail patients and direct them toward less invasive interventions. Consideration of frailty may also improve patient counseling on operative risk and ensure that patients benefit from a more personalized treatment strategy.
None declared.
