Is Peripheral Artery Disease an Independent Predictor of Isolated Coronary Artery Bypass Outcome?

HLC3087_proof ■ 4 March 2020 ■ 1/9

Heart, Lung and Circulation (2020) -, -–1443-9506/20/$36.00 https://doi.org/10.1016/j.hlc.2020.01.013

ORIGINAL ARTICLE

Is Peripheral Artery Disease an Independent Predictor of Isolated Coronary Artery Bypass Outcome? Massimo Bonacchi, MD a,*,1, Orlando Parise, MSc a,b,1, Francesco Matteucci, MD a,b, Cecilia Tetta, MD b, Amalia Ioanna Moula, BSc b, Linda Renata Micali, BSc b, Aleksander Dokollari, MD c, Marco De Matino, MD a, Guido Sani, MD a,d, Andrea Grasso, MD a, Edvin Prifti, MD e, Sandro Gelsomino, MD, PhD a,b,1 a

Cardiac Surgery Unit, Department of Experimental and Clinical Medicine, University of Florence, Firenze, Italy Cardiovascular Research Institute Maastricht – CARIM, Maastricht University Medical Centre, The Netherlands c Cardiac Surgery, St. Michael Hospital, Toronto, ON, Canada d Cardiac Surgery, Department of Medical Biotechnologies, University of Siena, Siena, Italy e Division of Cardiac Surgery, University Hospital Center of Tirana, Albania b

Received 31 May 2019; received in revised form 14 December 2019; accepted 11 January 2020; online published-ahead-of-print xxx

Aim

The aim was to use a propensity score-based analysis to determine the impact of peripheral artery disease (PAD) on early outcomes after coronary artery bypass surgery grafting (CABG) in patients with PAD.

Method

We conducted a multicentre retrospective analysis of 11,311 consecutive patients who underwent CABG between 1997 and 2017. Patients with previous or concomitant vascular surgery were excluded. The main endpoints were death, stroke, and limb ischaemia requiring percutaneous or surgical revascularisation. Subgroup analyses were performed to test the interaction of PAD with concomitant factors.

Results

There was no difference in mortality in patients with and without PAD (p=0.06 and p=0.179, respectively). Patients with PAD had a greater incidence of stroke (p=0.04), acute kidney disease (p=0.003), and limb ischaemia requiring interventions (p,0.001) than those without PAD. The use of off-pump or no-touch aortic techniques did not influence the effect of PAD on the outcomes. Early mortality rate increased in patients with PAD when associated with long cardiopulmonary bypass, cross-clamp times (both p,0.001), and postoperative low cardiac output (p=0.01).

Conclusions

The presence of PAD is associated, independently of other factors, with greater incidence of stroke, acute kidney disease, and limb ischaemia following CABG, irrespective of the technique employed. Operative mortality was greater in patients with PAD only when associated with long cardiopulmonary bypass and aortic cross-clamp times, and low cardiac output.

Keywords

Coronary artery bypass  Peripheral arterial disease  Peripheral vascular disease

*Corresponding author at: Cardiac Surgery Unit – Department of Experimental and Clinical Medicine, University of Florence, Largo Brambilla, 3 - 50134 Firenze, Italy; Email: mbonacchi@unifi.it 1

Contributed equally to this work.

Ó 2020 Australian and New Zealand Society of Cardiac and Thoracic Surgeons (ANZSCTS) and the Cardiac Society of Australia and New Zealand (CSANZ). Published by Elsevier B.V. All rights reserved.

Please cite this article in press as: Bonacchi M, et al. Is Peripheral Artery Disease an Independent Predictor of Isolated Coronary Artery Bypass Outcome? Heart, Lung and Circulation (2020), https://doi.org/10.1016/j.hlc.2020.01.013

HLC3087_proof ■ 4 March 2020 ■ 2/9

2

M. Bonacchi et al.

Introduction The prevalence of peripheral artery disease (PAD) in patients with coronary artery disease (CAD) is 40% [1] whereas the prevalence of CAD in patients undergoing peripheral arterial surgery is 78% [2], and is influenced by age [3]. Moreover, the presence of PAD in patients undergoing coronary artery bypass grafting (CABG) is associated with poorer outcome [4–8].The aetiology of both diseases is multifactorial (e.g., atherosclerosis, smoking, diabetes, hypertension, and dyslipidaemia) and the relevant factors may influence the outcomes of CABG [9]. PAD is a risk factor in patients undergoing percutaneous coronary intervention. These patients with coronary artery disease and PAD have lower procedural success and higher rate of in-hospital major cardiovascular complications, including higher blood loss requiring transfusion, after PCI with stent implantation. The impact of PAD on early outcomes in patients undergoing CABG is incompletely understood. Using a propensity score (PS)-based analysis, we investigated the predictive

value of PAD in early post-CABG outcomes and assessed whether its predictive ability is independent or linked to the concurrent risk factors in this study.

Methods The need for ethical committee approval was waived according to national laws regulating observational retrospective studies (Italian law no. 11960, released on 13 July 2004). In total, 11,311 patients who underwent isolated CABG between 1997 and 2017 by three consultant authors (MB, EP, GS) at their respective centres were included in the study. Patients who underwent previous or concomitant vascular surgery were excluded. The patients were divided into two groups based on the presence or absence of PAD. Tables 1 and 2 show baseline and intraoperative data.

Definitions and Endpoints Peripheral artery disease was defined an extracardiac arteriopathy including one or more of the following:

Table 1 Patient characteristics. Overall (n=11,311) Age (yr)

68.00 (61.00–74.00)

No PAD (n=9,350)

PAD (n=1,961)

P

ASMD

67.00 (60.00–74.00)

71.00 (65.00–76.00)

,0.001

0.49a

Sex (male)

9,154 (80.9)

7,575 (81.0)

1,579 (80.5)

0.634

Hypertension

7,047 (62.3)

5,489 (58.7)

1,558 (79.4)

,0.001

0.01 0.51a

Smoke Hypercholesterolaemia

4,485 (39.7) 5,513 (48.7)

3,403 (36.4) 4,225 (45.2)

1,082 (55.2) 1,288 (65.7)

,0.001 ,0.001

0.38a 0.43a

Diabetes

3,270 (28.9)

2,480 (26.5)

790 (40.3)

,0.001

0.28a

Insulin-dependent diabetes

1,070 (9.5)

712 (7.6)

358 (18.3)

,0.001

0.28a

COPD

738 (6.5)

533 (5.7)

205 (10.5)

,0.001

0.16

AKD

472 (4.2)

275 (2.9)

197 (10.0)

,0.001

0.24a

99 (0.9)

56 (0.6)

43 (2.2)

,0.001

0.21a

Dialysis

1,877 (16.6)

1,028 (11.0)

849 (43.3)

,0.001

0.3a

Stroke history TIA history

199 (1.8) 345 (3.1)

91 (1.0) 200 (2.1)

108 (5.5) 145 (7.4)

,0.001 ,0.001

0.21a 0.21a

Pulmonary hypertension

579 (5.1)

423 (4.5)

156 (8.0)

,0.001

0.13

IABP

252 (2.2)

213 (2.3)

39 (2.0)

0.481

–0.02

NYHA 3

2,079 (18.4)

1,574 (16.8)

505 (25.8)

LVEF (%)

55.00 (45.00–60.00)

55.00 (45.00–60.00)

LVEF ,35%

1,344 (11.9)

LCOS

697 (7.4)

51.00 (43.00–60.00) 687 (35.0)

,0.001

0.21a

,0.001

0.25a

,0.001

0.3a

3,677 (32.5)

2,935 (31.4)

742 (37.8)

,0.001

0.13

LAD stenosis Two vessels disease

10,908 (96.4) 3,280 (29.0)

9,078 (97.1) 2,727 (29.2)

1,830 (93.3) 553 (28.2)

,0.001 0.407

0.15 0.02

Three vessels disease

7,945 (70.2)

6,551 (70.1)

1,394 (71.1)

0.383

0.02

Previous PTCA

4,779 (42.3)

4,041 (43.2)

738 (37.6)

,0.001

0.12

Antiplatelet agents ,5 d

3,810 (33.7)

3,221 (34.4)

589 (30.0)

,0.001

0.1

LMCA stenosis

Data are median (interquartile range) or n (%). Abbreviations: PAD, peripheral arterial disease; ASMD, absolute standardised mean difference; COPD, chronic obstructive pulmonary disease; AKD, acute kidney disease (creatinine .2 mg/dL); LCOS, low cardiac output syndrome; TIA, transient ischaemic attack; IABP, intra-aortic balloon pump; NYHA, New York Heart Association Score; LVEF, left ventricular ejection fraction; LMCA, left main coronary artery; LAD, left anterior descending coronary artery; PTCA, percutaneous transluminal coronary angioplasty. Unbalanced variables (ASMD .0.20).

a

Please cite this article in press as: Bonacchi M, et al. Is Peripheral Artery Disease an Independent Predictor of Isolated Coronary Artery Bypass Outcome? Heart, Lung and Circulation (2020), https://doi.org/10.1016/j.hlc.2020.01.013

HLC3087_proof ■ 4 March 2020 ■ 3/9

3

CABG Outcome in PAD Patients

Table 2 Operative data. Overall (n=11,311)

No PAD (n=9,350)

PAD (n=1,961)

P

CPB time (min)

80.00 (58.68–92.00)

79.00 (56.00–92.00)

87.00 (68.00–94.00)

,0.001

CC time (min)

55.00 (40.00–69.32)

55.00 (39.00–69.00)

55.00 (43.00–70.00)

0.188

Surgery

5,723 (50.6)

4,484 (48.0)

1,239 (63.2)

Urgent

735 (6.5)

630 (6.7)

105 (5.4)

4,853 (42.9) 374 (3.3)

4,236 (45.3) 78 (0.8)

617 (31.5) 296 (15.1)

On-pump

9,120 (80.6)

7,598 (81.3)

1,522 (77.6)

ON-TC

2,882 (31.6)

2,358 (31.0)

524 (34.0)

0.17

ON-SC

6,238 (68.4)

5,240 (69.0)

998 (65.6)

,0.001

Emerging Elective

OPCAB

,0.001

2,191 (19.4)

1,752 (18.7)

439 (22.4)

OPCAB-NT

839 (38.3)

605 (34.5)

234 (53.3)

,0.001

OPCAB-SC

1,352 (61.7)

1,147 (65.5)

205 (46.7)

0.02

11,293 (99.8) 2,488 (22.0)

9,343 (99.9) 2,050 (21.9)

1,950 (99.4) 438 (22.3)

0.87 0.82

LITA RITA RA Distal anastomosis Distal anastomosis/pat

905 (8.0) 30,473

748 (8.0) 25,178

157 (8.0)

.0.9

5,295

2.69

2.69

2.7

.0.9

1.3

1.3

1.3

.0.9

Arterial anastomosis/pat Data are median (interquartile range) or n (%).

Abbreviations: PAD, peripheral arterial disease; CPB, cardiopulmonary bypass; CC, cross-clamp; TC, total clamp; SC, side clamp; OPCAB, off-pump coronary artery bypass; NT, no touch; LITA, left internal thoracic artery; RITA, right internal thoracic artery; RA, radial artery.

claudication, abdominal aorta occlusion or .75% stenosis, limb arteries occlusion or .75% stenosis, planned intervention on the abdominal aorta, and/or limb arteries [10]. Low cardiac output syndrome (LCOS) was diagnosed if the patient required inotropic medication (dopamine, dobutamine, milrinone, or epinephrine) to maintain a systolic blood pressure .90 mmHg and cardiac output .2.2 L/min/m2 for at least 30 minutes in the intensive care unit, after optimising the preload, the afterload, electrolyte, and blood gas abnormalities [11]. Prolonged cardiopulmonary bypass (CPB) and aortic cross-clamp (CC) times were defined as durations .120 minutes and .90 minutes, respectively [12,13]. The main endpoints were as follows: (1) mortality, defined as mortality of any cause, occurring within 30 days after surgery inside or outside the hospital or after 30 days during the same hospitalisation period following surgery; (2) stroke defined as any acute neurological deficit lasting .24 hours, of a cerebrovascular aetiology, confirmed on brain imaging (head computed tomography or magnetic resonance imaging) and clinically diagnosed by a neurologist within 30 days of surgery; (3) renal failure-defined postoperative creatinine level of .200 mmol/L or as a condition requiring dialysis; and (4) postoperative limb ischaemia requiring surgical and/ or endovascular revascularisation procedures.

Statistics A generalised boosted model (GBM) was employed for PS estimation for multiple treatments [14]. The average treatment effect on the treated (ATT) was selected as the causal

effect estimand. Inverse probability of treatment weighting was employed to reduce confounding due to observed variables [15]. Positivity assumption was assessed by testing the overlap in the true distribution of the observed pretreatment characteristics. The relationship of potential unmeasured confounders was tested within treatment groups. Propensity score weights were estimated fitting separate GBMs for each group. The Kolmogorov–Smirnov statistical test was employed as a stopping rule to select the optimal iteration of the GBM for use in estimating the PS weights. Balance of each group toward the other groups was checked using either diagnostic plots or tables. An optimise plot, which is a graphical display of the balance criteria as a function of the GBM iteration, was employed. Furthermore, overlapping was tested using boxplots of the PS scores for the treatment; a substantial overlap of the box plots with great separation of the plots was typically considered insufficient. Moreover, we plotted the absolute standardised mean differences (ASMDs) between the treatment groups on the pretreatment covariates, before and after weighting. Finally, we used quantile-quantile (Q-Q) plots to check the balance further. In these plots, the quantiles of the observed p-values were plotted against the quantiles of the uniform distribution (45-degree line). Ideally, the p-values from independent tests in which the null hypothesis is true will have a uniform distribution In contrast, p-values severely deviating below the diagonal area suggests lack of balance, while p-values equal to or deviating above the diagonal area suggest that balance might have been achieved.

Please cite this article in press as: Bonacchi M, et al. Is Peripheral Artery Disease an Independent Predictor of Isolated Coronary Artery Bypass Outcome? Heart, Lung and Circulation (2020), https://doi.org/10.1016/j.hlc.2020.01.013

HLC3087_proof ■ 4 March 2020 ■ 4/9

4

For bias statistics, ASMDs of ,0.20 were considered small. We estimated the pairwise ATT based on the difference in the estimated population means. We implemented the model by pooling data from the three groups and estimating the ATT and population means using weighted linear regression. Subgroup analyses were conducted to analyse the interaction of PAD with other potential co-factors in predicting outcomes. As the observations are over 20 years, the analysis was corrected by the year of operation treated as a categorical variable. R software version 3.5.2 (R Foundation for Statistical Computing, Vienna, Austria) with the TWANG and SURVEY packages [16] was used for the analysis.

Results Balance by Plots We obtained a good balance with 8,000 iterations (Figure 1A). Figure 1(B) compares the distributions of the estimated PSs across the treatments: a substantial overlap was obtained. In addition, Figure 1(C) provides comparisons of the ASMDs between the treatment groups on the pretreatment covariates, before and after weighting. In all cases, the weighted ASMD values f were ,0.20. Finally, The Q-Q plot in Figure 1(D) shows that, before weighting, most of the values were below the 45-degree line. After weighting, good balance was achieved with the p-values equal to or above the diagonal area.

Balance Table Table 1 shows for each preoperative variable the unweighted and weighted ATT means for each group and the unweighted overall population mean and standard deviation. Before weighting, the standard difference (SD) was .0.20 in 14 of the pretreatment covariates; after weighting, SDs were generally attenuated with no value at .0.20. The combination of the balance table with the balance graphs allowed us to conclude that the groups were sufficiently similar to support causal estimation of the treatment estimand.

M. Bonacchi et al.

Moreover, the incidence of stroke was slightly, although still significantly, higher (1.4%, 95% CI, 0.02–2.8; p=0.04) in patients with PAD. Additionally, when we weighted the sample, acute kidney disease (AKD) (3.0%, 95% CI, 1.0–5.1; p=0.03) and postoperative acute limb ischaemia (10%, 95% CI, 9.3–11.5; p,0.001) were higher in patients with PAD than in those without PAD.

Subanalysis by CABG Technique Employed Figure 3 shows the interactions between PAD and CABG in relation to the main outcomes. There was no difference in mortality rate (Figure 3A) between on-pump and off-pump techniques (off-pump coronary artery bypass [OPCAB]; p=0.08), between the on-pump total clamp and on-pump side clamp techniques (p=0.07), and between OPCAB no touch and OPCAB side-clamp techniques (p=0.6). Similarly, no difference was observed for stroke (Figure 3B; on-pump vs OPCAB, p=0.5; on-pump total clamp vs on-pump side clamp, p=0.1; OPCAB no touch vs OPCAB side clamp, p=0.4), AKD (Figure 3C; onpump vs OPCAB, p=0.5; on-pump total clamp vs on-pump side clamp, p=0.3; OPCAB no touch vs OPCAB side clamp, p=0.2), and limb ischaemia (Figure 3D; on-pump vs OPCAB, p=0.1; on-pump total clamp vs on-pump side clamp, p=0.2; OPCAB no touch vs OPCAB side clamp, p=0.4).

Subanalysis by Intra and Postoperative Factors Figure 4 shows the interactions between PAD and the other intra and postoperative factors related to the main outcomes. The presence of postoperative LCOS significantly increased the mortality in patients with PAD (Figure 4A, p=0.01) Similarly, long CPB and aortic CC times (both p,0.001) strengthened the impact of PAD as a risk factor increasing the mortality rate; however, the use of intra-aortic balloon pump in the postoperative period did not have any significant influence (p=0.4). The impact of PAD on stroke (Figure 4B), AKD (Figure 4C), and limb ischaemia (Figure 4C) was not influenced by these factors (all p.0.05)

Early Outcomes Early outcomes are shown in Table 3. In the unweighted population, apart from LCOS requiring inotropes and atrial fibrillation, patients with PAD experienced a significantly higher incidence of postoperative complications, including the main end points of mortality, stroke, and limb ischaemia (all p,0.001).

Causal Effects Estimation The estimated incidence of the study endpoints in patients with or without PAD are shown in Figure 2. The morality rate was not significantly higher in patients with PAD compared with those without PAD in the weighted sample (1.6%, 95% confidence interval [CI], –0.6 to 4.0; p=0.06).

Discussion A propensity-based analysis of 11,311 patients was performed, which, in contrast to the methods used in most published studies on the topic [10,17–20], enabled unbiased estimation of the true influence of PAD on post-CABG outcomes, enhancing excellent balance in perioperative variables between groups. The aim of the study was to assess whether, in a large patient population, PAD was an independent predictor of outcomes and to test the false hypothesis that only the association of PAD with concomitant atherosclerotic factors in CAD is responsible for a higher incidence of complications after CABG.

Please cite this article in press as: Bonacchi M, et al. Is Peripheral Artery Disease an Independent Predictor of Isolated Coronary Artery Bypass Outcome? Heart, Lung and Circulation (2020), https://doi.org/10.1016/j.hlc.2020.01.013

HLC3087_proof ■ 4 March 2020 ■ 5/9

5

CABG Outcome in PAD Patients

Figure 1 (A) The optimise plot. The scales of the Kolmogorov–Smirnov (KS) statistic presented show a very good balance starting from 1,000 iterations of the model. (B) Overlap assessment. Each panel presents boxplots by treatment group of the estimated propensity score for one of the treatments (see text). (C) Standardised effect size plot for assessing the balance of pretreatment variables before and after weighting. The statistically significant difference is indicated by the solid circle. The red lines represent pretreatment covariates for which the pairwise absolute standardised mean differences (ASMDs) increase after weighting. The weighted ASMD values were, in all cases, ,0.20. (D) Quantile-quantile (Q-Q) plot. Weighted values showed a good balance with p-values running at or above the diagonal. Abbreviation: AAT, average treatment effect on the treated.

We found that: (1) the presence of PAD was associated with a higher incidence of stroke, acute limb ischaemia requiring revascularisation, and AKD; (2) patients with PAD did not have a greater mortality rate than those without PAD; (3) the employment of the OPCAB or no-touch aortic techniques did not influence the effect of PAD on outcomes; (4) the early mortality rate increased with PAD when arteriopathy was associated with long CPB and CC times, and postoperative LCOS requiring the use of inotropes.

Cerebral embolism and carotid stenosis are the most important causes of stroke and they increase the risk of mortality after CABG [21,22]. Both factors are strictly related to aortic atherosclerosis and linked to aortic manipulation (aortic clamp, side-biting clamp for proximal anastomoses, aortic cannulation), which may lead to embolisation of atheromatous debris [23]. The use of the OPCAB and notouch techniques has been advocated, in an attempt to reduce the incidence of these complications [23–25];

Please cite this article in press as: Bonacchi M, et al. Is Peripheral Artery Disease an Independent Predictor of Isolated Coronary Artery Bypass Outcome? Heart, Lung and Circulation (2020), https://doi.org/10.1016/j.hlc.2020.01.013

HLC3087_proof ■ 4 March 2020 ■ 6/9

6

M. Bonacchi et al.

Table 3 Early postoperative outcomes. Overall (n=11,311) Mortality Cardiac mortality

No PAD (n=9,350)

PAD (n=1,961)

P

131 (1.1)

87 (0.9)

44 (2.2)

,0.001

75 (0.7)

46 (0.5)

29 (1.5)

,0.001

4,915 (43.4)

3,927 (42.0)

988 (50.4)

,0.001

Re-exploration for bleeding

162 (1.4)

121 (1.3)

41 (2.1)

,0.001

Mediastinal Infection LCOS requiring inotropic support

164 (1.4) 285 (2.5)

118 (1.3) 224 (2.4)

46 (2.3) 61 (3.1)

0.002 0.07

Stroke

187 (1.6)

133 (1.4)

54 (2.7)

,0.001

IABP

285 (2.5)

280 (3.0)

5 (0.2)

,0.001

AKD

407 (3.6)

233 (2.5)

174 (8.9)

,0.001

Postoperative dialysis

301 (2.7)

168 (1.8)

133 (6.8)

,0.001

AF

702 (6.2)

561 (6.0)

141 (7.2)

0.053

Acute limb ischaemia

221 (1.9)

84 (0.9)

137 (7.0)

,0.001

Transfusions

Ventilation time (hr) ICU LoS (d) In-hospital LoS (d)

9 (4–1,236) 2 (0–37) 14 (1–287)

8 (4–844) 2 (0–26)

15 (4–1,236) 3 (0–37)

14 (1–222)

15 (16–1,287)

,0.001 0.02 0.006

Data are median (interquartile range) or n (%). Abbreviations: PAD, peripheral artery disease; LCOS, low cardiac output syndrome; IABP, intra-aortic balloon pump; AKD, acute kidney disease (creatinine .2 mg/dL); AF, atrial fibrillation; ICU, intensive care unit; LoS, length of stay.

Figure 2 Early outcomes (mortality, stroke, acute kidney disease [AKD] and acute limb ischaemia) in patients with or without peripheral artery disease (PAD). *Significance between the two groups.

Please cite this article in press as: Bonacchi M, et al. Is Peripheral Artery Disease an Independent Predictor of Isolated Coronary Artery Bypass Outcome? Heart, Lung and Circulation (2020), https://doi.org/10.1016/j.hlc.2020.01.013

HLC3087_proof ■ 4 March 2020 ■ 7/9

7

CABG Outcome in PAD Patients

Figure 3 Subgroup analysis of the occurrence of outcomes by surgical technique (on-pump [ON] vs offpump coronary artery bypass [OPCAB]), on-pump total clamp [ON-TC] vs. on-pump side clamp ([N-SC], OPCAB no touch vs OPCAB side clamp). When one of the co-factors examined is present (on-pump, ON-TC, and OFF-SC) and the odds ratios are equal, lower, or not significantly different from when these co-factors are absent, there is no interaction between PAD and the covariate examined and peripheral artery disease is an independent predictor. (A) Mortality. (B) Stroke. (C) Acute kidney disease (AKD). (D) Limb ischaemia.

however, based on our data, aortic manipulation does not seem to strengthen the influence of PAD on stroke. Therefore, this complication seems to be related to the diffuse atherosclerotic disease rather than to the surgical techniques. This is confirmed on the basis of the results of the comparisons between on-pump and OPCAB, between on-pump total clamp and on-pump side clamp, and between OPCAB notouch and OPCAB side clamp in the weighted sample. Hypoperfusion may be a co-factor, impairing the clearance of micro-emboli [26], while it can be worsened by carotid stenosis [25]. However, the incidence of carotid stenosis was comparable in the balanced population and, therefore, it cannot be a significant co-factor. LCOS significantly strengthened the influence of PAD on mortality but had no effect on stroke. Based on our observation, we believe that clinicians should pay more attention to these subjects in relation to the timing and decision-making regarding PAD treatment. The choice of vascular surgery before or concomitant with CABG should be patient-tailored and taken into account with the

Figure 4 Subgroup analysis of the occurrence of outcomes by intra and postoperative factors (low cardiac output [LCO], long cardiopulmonary bypass [CPB] time, long aortic cross-clamp [CC] time, use of postoperative intra-aortic balloon pump [IABP]). When one of the co-factors examined is present (on-pump, onpump total clamp, and off-pump side clamp) and the odds ratios are equal, lower, or not significantly different from when these co-factors are absent, there is no interaction between PAD and the covariate examined and peripheral artery disease is an independent predictor. (A) Mortality. (B) Stroke. (C) Acute kidney disease (AKD). (D) Acute limb ischaemia. *Significance between the two groups.

patient’s clinical conditions. The approach also varies depending on the site of the critical arterial disease. The management of patients with carotid disease undergoing CABG is still controversial. Therapeutic options include CABG alone, staged or synchronous carotid endarterectomy and CABG, and staged or synchronous carotid artery stenting. In our opinion, the non-negligible incidence of stroke in our PAD population (3.4%, n=392/11,331), suggests an inadequate evaluation for carotid revascularisation before CABG. Among the surgical options, carotid stenting followed by CABG [27,28] seems to guarantee better results than carotid endarterectomy followed by CABG [29,30]. Controversies arise in patients with severe asymptomatic or moderate carotid stenosis for whom a wait-and-see approach

Please cite this article in press as: Bonacchi M, et al. Is Peripheral Artery Disease an Independent Predictor of Isolated Coronary Artery Bypass Outcome? Heart, Lung and Circulation (2020), https://doi.org/10.1016/j.hlc.2020.01.013

HLC3087_proof ■ 4 March 2020 ■ 8/9

8

seems to be more appropriate [28,31]. Based on our data, the presence of advanced atherosclerotic disease with PAD and significant carotid stenosis, even if not associated with previous stroke or ischaemic attack is a risk factor. Therefore, we agree with authors supporting treatment interventions before CABG to prevent perioperative stroke [27]. In addition, our results indicate that the risk of postoperative AKD is greatly increased in patients with PAD (p=0.003). Chronic kidney disease is a common complication of CABG and it is associated with adverse outcomes [32,33]. The previously described pathogenic mechanisms include CPB-induced haemolysis with an increase in plasma-free haemoglobin levels, CPB non-pulsatile flow, reperfusion injury, oxidative stress, and inflammation [34]. Renal artery stenosis is not uncommon in CAD and may be progressive in certain patients [35,36]. As indicated earlier, PS weighting allowed a balance across the possible variables that might influence the occurrence of postoperative AKD (age, sex, comorbid conditions, medication intake, type of surgery, duration of surgery, chronic kidney disease). Therefore, after appropriate screening for renal artery stenosis and evaluation of the impact of CABG on postoperative renal function, renal artery angioplasty, or stenting prior to cardiac surgery seem to be indicated in patients with significant renal and CADs. The use of OPCAB, with or without the no-touch and single-clamping techniques did not affect the incidence of AKD in patients with PAD. Remarkably, PAD had a strong impact on postoperative acute limb ischaemia requiring percutaneous/surgical revascularisation (p,0.001). The association between CABG and the subsequent development of critical limb ischaemia is a high-risk factor for postoperative adverse outcomes [37]; further, a decreased ankle–brachial index [38,39] and incompressible ankle arteries [38] adversely affect these outcomes. In patients with PAD, collateral blood flow to the lower extremities can be provided by the internal thoracic arteries (ITAs), inferior mesenteric artery, or lumbar arteries. Specifically, ITAs may provide the main collateral supply to the ipsilateral iliac artery via the inferior epigastric artery and superior epigastric artery (Winslow pathway) [40]. They represent one of the largest branches of the so-called parietal collateral pathway, which also involves the intercostal, subcostal, lumbar, and middle sacral branches of the aorta. Therefore, the use of the left ITA, right ITA, or both for coronary revascularisation in patients with PAD may lead to serious vascular complications, by breaking a potential major collateral pathway [41]. For this reason, some authors recommend an accurate pre-CABG evaluation of the arterial tree and avoiding the use of the ipsilateral ITA, in patients with aorto-iliac occlusion and an open external iliac artery and common femoral artery [41]. Nonetheless, this is not supported by our findings as the estimated incidence of postoperative acute limb ischaemia requiring surgical/ percutaneous revascularisation was higher in the PAD group not only in the weighted, but also in the unweighted population showing left ITA and right ITA graft in the two

M. Bonacchi et al.

groups. Nonetheless, we believe that selective ITA territory angiography is advisable in PAD and in the absence of femoral pulse, ITA harvesting should be avoided. In contrast with aortic and renal atherosclerotic involvement, prophylactic lower extremity revascularisations in patients with PAD concomitant with CABG are not justified [42]. A strict follow-up is advisable and delayed lowerextremity revascularisation performed in stable patients after successful CABG seems more reasonable, enabling a safer, completer and extensive revascularisation. To confirm or deny this approach we are collecting data to compare long term-results of prophylactic and post-CABG lower extremity revascularisations. In our experience, patients with PAD had comparable estimated mortality rate (p=0.06) with those without PAD. This is in accordance with data [4,19,20] reporting that PAD is an independent risk factor only for late but not early mortality [17,19]. Nonetheless, the mortality rate in PAD increased significantly when LCOS occurred or when CPB and CC times were prolonged. Hence our findings confirm those reported by Kaul et al. [43], who reported that, after risk factor adjustment, patients with PAD had mortality rates twice as high as those of patients without PAD. The strength of this paper was in the use of matched data to give credibility and evidence to some of clinical preconceptions or expectations, and to highlight that the risk lies in the presence of the PVD and cannot be overcome by the operative technique. Further research is needed to draw final conclusions on this matter.

Limitations This study had a retrospective, observational nature, which limits the generalisability of our findings.

Conclusions The presence of PAD independently increased the incidence of postoperative stroke, AKD, and limb ischaemia requiring revascularisation, irrespective of the technique employed. The mortality rate was higher in patients with PAD only when arteriopathy was associated with postoperative LCOS and long CPB and aortic CC times. Our findings need to be confirmed by further, larger studies.

Conflicts of Interest There are no conflicts of interest to disclose.

Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Please cite this article in press as: Bonacchi M, et al. Is Peripheral Artery Disease an Independent Predictor of Isolated Coronary Artery Bypass Outcome? Heart, Lung and Circulation (2020), https://doi.org/10.1016/j.hlc.2020.01.013

HLC3087_proof ■ 4 March 2020 ■ 9/9

CABG Outcome in PAD Patients

Acknowledgements We thank Professor Bobby Yanagawa for the English revision of the manuscript.

References [1] Dormandy J, Heeck L, Vig S. Lower-extremity arteriosclerosis as a reflection of a systemic process: implications for concomitant coronary and carotid disease. Semin Vasc Surg 1999;12:118–22. [2] Gersh BJ, Rihal CS, Rooke TW, Ballard DJ. Evaluation and management of patients with both peripheral vascular and coronary artery disease. J Am Coll Cardiol 1991;18:203–14. [3] Luther M, Lepantalo M. Femorotibial reconstructions for chronic critical leg ischaemia: influence on outcome by diabetes, gender and age. Eur J Vasc Endovasc Surg 1997;13:569–77. [4] O’Rourke DJ, Quinton HB, Piper W, Hernandez F, Morton J, Hettleman B, et al., Northern New England Cardiovascular Disease Study Group. Survival in patients with peripheral vascular disease after percutaneous coronary intervention and coronary artery bypass graft surgery. Ann Thorac Surg 2004;78:466–70. [5] Sutton-Tyrrell K, Rihal C, Sellers MA, Burek K, Trudel J, Roubin G, et al. Long-term prognostic value of clinically evident noncoronary vascular disease in patients undergoing coronary revascularization in the bypass angioplasty revascularization investigation (BARI). Am J Cardiol 1998;81:375–81. [6] Van Straten AH, Firanescu C, Soliman Hamad MA, Tan ME, ter Woorst JF, Martens EJ, et al. Peripheral vascular disease as a predictor of survival after coronary artery bypass grafting: comparison with a matched general population. Ann Thorac Surg 2010;89:414–20. [7] Eagle KA, Rihal CS, Foster ED, Mickel MC, Gersh BJ, The Coronary Artery Surgery Study (CASS) Investigators. Long-term survival in patients with coronary artery disease: importance of peripheral vascular disease. J Am Coll Cardiol 1994;23:1091–5. [8] Kilo J, Czerny M, Zimpfer D, Gorlitzer M, Wolner E, Grimm M. Predictors of perioperative mortality after coronary bypass grafting in the elderly. Thorac Cardiovasc Surg 2003;51:33–7. [9] Jang SY, Ju EY, Cho SI, Lee SW, Kim DK. Comparison of cardiovascular risk factors for peripheral artery disease and coronary artery disease in the korean population. Korean Circ J 2013;43:316–28. [10] Popovic B, Arnould MA, Selton-Suty C, Latarche C, Angioi M, Carteaux JP, et al. Comparison of two-year outcomes in patients undergoing isolated coronary artery bypass grafting with and without peripheral artery disease. Am J Cardiol 2009;104:1377–82. [11] Algarni KD, Maganti M, Yau TM. Predictors of low cardiac output syndrome after isolated coronary artery bypass surgery: trends over 20 years. Ann Thorac Surg 2011;92:1678–84. [12] Madhavan S, Chan SP, Tan WC, Eng J, Li B, Luo HD, et al. Cardiopulmonary bypass time: every minute counts. J Cardiovasc Surg (Torino) 2018;59:274–81. [13] Doenst T, Borger MA, Weisel RD, Yau TM, Maganti M, Rao V. Relation between aortic cross-clamp time and mortality–not as straightforward as expected. Eur J Cardiothorac Surg 2008;33:660–5. [14] McCaffrey DF, Griffin BA, Almirall D, Slaughter ME, Ramchand R, Burgette LF. A tutorial on propensity score estimation for multiple treatments using generalized boosted models. Stat Med 2013;32:3388–414. [15] Robins JM, Hernán MÁ, Brumback B. Marginal structural models and causal inference in epidemiology. Epidemiology 2000;11:550–60. [16] Lumley T. Analysis of complex survey samples. J Stat Soft 2004;9:19. [17] Van Straten AH, Firanescu C, Soliman Hamad MA, Tan ME, ter Woorst JF, Martens EJ, et al. Peripheral vascular disease as a predictor of survival after coronary artery bypass grafting: comparison with a matched general population. Ann Thorac Surg 2010;89:414–20. [18] Collison T, Smith JM, Engel AM. Peripheral vascular disease and outcomes following coronary artery bypass graft surgery. Arch Surg 2006;141:1214–8. [19] Harskamp RE, Alexander JH, Schulte PJ, Jones WS, Williams JB, Mack MJ, et al. Impact of extracardiac vascular disease on vein graft failure and outcomes after coronary artery bypass surgery. Ann Thorac Surg 2014;97:824–30. [20] Chu D, Bakaeen FG, Wang XL, Dao TK, LeMaire SA, Coselli JS, et al. The impact of peripheral vascular disease on long-term survival after coronary artery bypass graft surgery. Ann Thorac Surg 2008;86:1175–80.

9

[21] Dacey LJ, Likosky DS, Leavitt BJ, Lahey SJ, Quinn RD, Hernandez Jr F, et al., Northern New England Cardiovascular Disease Study Group. Perioperative stroke and long-term survival after coronary bypass graft surgery. Ann Thorac Surg 2005;79:532–6. [22] Tarakji KG, Sabik JF, Bhudia SK, Batizy LH, Blackstone EH. Temporal onset, risk factors, and outcomes associated with stroke after coronary artery bypass grafting. JAMA 2011;305:381–90. [23] Palmerini T, Savini C, Di Eusanio M. Risks of stroke after coronary artery bypass graft - recent insights and perspectives. Interv Card 2014;9:77–83. [24] Karthik S, Musleh G, Grayson AD, Keenan DJ, Pullan DM, Dihmis WC, et al. Coronary surgery in patients with peripheral vascular disease: effect of avoiding cardiopulmonary bypass. Ann Thorac Surg 2004;77:1245–9. [25] Li Z, Fan G, Zheng X, Gong X, Chen T, Liu X, et al. Risk factors and clinical significance of acute kidney injury after on-pump or off-pump coronary artery bypass grafting: a propensity score-matched study. Interact Cardiovasc Thorac Surg 2019;28:893–9. [26] Caplan LR, Hennerici M. Impaired clearance of emboli (washout) is an important link between hypoperfusion, embolism, and ischemic stroke. Arch Neurol 1998;55:1475–82. [27] Paraskevas KI, Nduwayo S, Saratzis AN, Naylor AR. Carotid stenting prior to coronary bypass surgery: an updated systematic review and meta-analysis. Eur J Vasc Endovasc Surg 2017;53:309–19. [28] Mahmoudi M, Hill PC, Xue Z, Torguson R, Ali G, Boyce SW, et al. Patients with severe asymptomatic carotid artery stenosis do not have a higher risk of stroke and mortality after coronary artery bypass surgery. Stroke 2011;42:2801–5. [29] Shishehbor MH, Venkatachalam S, Sun Z, Rajeswaran J, Kapadia SR, Bajzer C, et al. A direct comparison of early and late outcomes with three approaches to carotid revascularization and open heart surgery. J Am Coll Cardiol 2013;62:1948–56. [30] Biller J, Feinberg WM, Castaldo JE, Whittemore AD, Harbaugh RE, Dempsey RJ, et al. Guidelines for carotid endarterectomy: a statement for healthcare professionals from a special writing group of the Stroke Council, American Heart Association. Circulation 1998;97:501–9. [31] Masabni K, Raza S, Blackstone EH, Gornik HL, Sabik 3rd JF. Does preoperative carotid stenosis screening reduce perioperative stroke in patients undergoing coronary artery bypass grafting? J Thorac Cardiovasc Surg 2015;149:1253–60. [32] Lassnigg A, Schmidlin D, Mouhieddine M, Bachmann LM, Druml W, Bauer P, et al. Minimal changes of serum creatinine predict prognosis in patients after cardiothoracic surgery: a prospective cohort study. J Am Soc Nephrol 2004;15:1597. [33] Rydén L, Ahnve S, Bell M, Hammar N, Ivert T, Holzmann MJ. Acute kidney injury following coronary artery bypass grafting: early mortality and postoperative complications. Scand Cardiovasc J 2012;46:114–20. [34] O’Neal JB, Shaw AD, Billings FT. Acute kidney injury following cardiac surgery: current understanding and future directions. Crit Care 2016;20:187. [35] Crowley JJ, Santos RM, Peter RH, Puma JA, Schwab SJ, Phillips HR, et al. Progression of renal artery stenosis in patients undergoing cardiac catheterization. Am Heart J 1998;136:913–8. [36] Gross CM, Kramer J, Waigand J, Luft FC, Dietz R. Relation between arteriosclerosis in the coronary and renal arteries. Am J Cardiol 1997;80:1478–81. [37] Biancari F, Kangasniemi OP, Mahar MA, Ylonen K. Need for late lower limb revascularization and major amputation after coronary artery bypass surgery. Eur J Vasc Endovasc Surg 2008;35:596–602. [38] Aboyans V, Lacroix P, Postil A, Guilloux J, Rollé F, Cornu E, et al. Subclinical peripheral arterial disease and incompressible ankle arteries are both long-term prognostic factors in patients undergoing coronary artery bypass grafting. J Am Coll Cardiol 2005;46:815–20. [39] Burek KA, Sutton-Tyrrell K, Brooks MM, Naydeck B, Keller N, Sellers MA, et al. Prognostic importance of lower extremity arterial disease in patients undergoing coronary revascularization in the bypass angioplasty revascularization investigation (BARI). J Am Coll Cardiol 1999;34:716–21. [40] Chait A. The internal mammary artery: an overlooked collateral pathway to the leg. Radiology 1976;121:621–4. [41] De Heer LM, Buijsrogge MP, Lahpor JR. Acute limb ischemia after internal thoracic artery harvesting: case report. J Thorac Cardiovasc Surg 2011;141:e5–7. [42] Danczyk RC, Mitchell EL, Petersen BD, Edwards J, Liem TK, Landry GJ, et al. Outcomes of open operation for aortoiliac occlusive disease after failed endovascular therapy. Arch Surg 2012;147:841–5. [43] Kaul TK, Fields BL, Wyatt DA, Jones CR, Kahn DR. Surgical management in patients with coexistent coronary and cerebrovascular disease: longterm results. Chest 1994;106:1349–57.

Please cite this article in press as: Bonacchi M, et al. Is Peripheral Artery Disease an Independent Predictor of Isolated Coronary Artery Bypass Outcome? Heart, Lung and Circulation (2020), https://doi.org/10.1016/j.hlc.2020.01.013