Author + information
- Received February 28, 2005
- Revision received May 3, 2005
- Accepted May 15, 2005
- Published online September 6, 2005.
- Victor Aboyans, MD, PhD⁎ (, )
- Philippe Lacroix, MD,
- Annabel Postil, MD,
- Jérôme Guilloux, MD,
- Florence Rollé, MD,
- Elisabeth Cornu, MD and
- Marc Laskar, MD
- ↵⁎Reprint requests and correspondence:
Dr. Victor Aboyans, Department of Thoracic and Cardiovascular Surgery and Vascular Medicine, Dupuytren University Hospital, 2, Ave. Martin Luther King, 87042 Limoges, France
Objectives This study was designed to determine the prevalence of peripheral arterial disease (PAD) in candidates for coronary artery bypass grafting (CABG) and to assess the predictive value of different types of subclinical PAD (peripheral occlusive disease and medial arterial calcification [incompressible ankle arteries]).
Background Observational studies report poor prognosis after CABG in the presence of clinical PAD, but data on subclinical PAD are scarce.
Methods We prospectively enrolled CABG candidates and measured ankle-brachial index (ABI) preoperatively. Patients were divided into four groups: clinical PAD, subclinical PAD (ABI <0.85), incompressible arteries (ABI >1.5), and no PAD. The primary end point was a composite combining death, acute coronary syndrome, stroke or transient ischemic attack (TIA), and coronary or peripheral revascularization. Secondary end points were overall and cardiovascular death, acute coronary syndrome, and stroke or TIA. Statistical analyses were performed using the Cox regression model.
Results We consecutively enrolled 1,022 patients (mean age 66.9 ± 9.2 years). In addition to the 14% with clinical PAD, we detected subclinical PAD in 13% and medial artery calcification in 12%. During an actuarial follow-up of 4.4 years, 81.2% of patients remained event-free. Adverse factors were (p < 0.05) supraventricular arrhythmia (odds ratio [OR] 2.5), ejection fraction <0.40 (OR 2.3), combined valvular surgery (OR 2.5), clinical PAD (OR 3.6), subclinical PAD (OR 3.3), and medial artery calcification (OR 1.9). The latter three factors were also independently predictive for overall and cardiovascular death.
Conclusions Beyond clinical PAD, the measurement of ABI before coronary surgery provides substantial information on long-term postoperative prognosis. To our knowledge, this is the first study highlighting the prognostic role of incompressible ankle arteries in secondary prevention.
Peripheral arterial disease (PAD) is a common condition in patients undergoing coronary artery bypass grafting (CABG). In several surgical series, the coexistence of PAD in these patients is associated with a poorer outcome (1–11). In the majority of these studies, PAD is defined according to a history of peripheral revascularization and/or presence of intermittent claudication. It is well known that beyond clinical PAD, a larger group of people present with subclinical PAD, which is generally defined by a low ankle-brachial index (ABI). Data on the utility of ABI measurement for the risk stratification of patients undergoing CABG are scarce (12,13).
Moreover, in another subgroup, especially in elderly patients (14,15) and/or diabetic patients (16,17) (both conditions being frequent in CABG candidates), the development of medial arterial calcification is generally asymptomatic. This condition can also be assessed by ankle pressure measurement and corresponds to incompressible arteries with an elevated ABI (18). In a recent publication on a general population, a high ABI had the same poor prognostic significance as a low ABI (18). However, it is unknown whether this finding would be of any prognostic interest in the setting of secondary coronary prevention.
In this prospective study, we aimed to determine the ability of ABI to detect candidates for CABG whose further risk of secondary cardiovascular events would be high. We hypothesized that beyond clinical PAD, asymptomatic PAD defined by an abnormally low or high ABI would be a prognostic marker for the long-term postoperative period.
Materials and methods
From August 1998 to January 2002, we prospectively enrolled all candidates consecutively referred to our department for CABG. Only those referred in an emergency without the opportunity to assess peripheral arteries were excluded from the cohort.
Definition of clinical variables
Preoperative data included cardiovascular history and cardiovascular risk factors, as well as data obtained at cardiac catheterization. Patients were considered smokers if they were actively smoking or had discontinued smoking within the two years before surgery. Diabetic patients were defined by a fasting blood glucose at entry >11.1 mmol/l or taking oral antidiabetic drugs and/or insulin. Those with a blood pressure >160/90 mm Hg measured twice before surgery or taking antihypertensive drugs for that purpose were considered as presenting with hypertension. Hypercholesterolemia was defined by a fasting blood cholesterol level at entry >2.40 g/l or taking lipid-lowering agents for that purpose. Renal insufficiency was defined by a preoperative blood creatinine >150 μmol/l. Clinical PAD was defined according to the presence of an intermittent claudication and/or a history of vascular surgery for PAD. The presence of supraventricular arrhythmia was defined as the presence of sustained atrial fibrillation or flutter on the preoperative electrocardiogram. Preoperative data concerned the number of bypasses performed, whether the revascularization was complete, the requirement of a cardiopulmonary bypass or an off-pump surgery, and the performance of any combined surgery (i.e., concomitant valvular or carotid surgery).
Measurement of ABI
To detect subclinical PAD, the ABI was measured by a hand-held Doppler probe and an adequate blood pressure cuff. The systolic pressure of the posterior tibial, dorsalis pedis, and humeral arteries was measured twice at each limb, and the mean of both results was obtained for each artery. The ABI was calculated by dividing the mean of posterior tibial and dorsalis pedis arteries’ systolic pressures by the mean of systolic pressures of both humeral arteries. For each patient, the lowest ABI between both ankles was kept for further analysis. The inter- and intraobserver variability of ABI measurement are reported elsewhere (19). Patients with no history of clinical PAD but with a low ABI (<0.85) were considered as presenting with subclinical PAD. Those without clinical PAD but presenting with at least one incompressible ankle artery (Doppler signal present even when the cuff was inflated >300 mm Hg) and/or an ABI >1.5 were considered as presenting with medial artery calcification.
Follow-up of the study population and definition of end points
Follow-up was performed by telephone contacts with patients’ family physicians as well as a systematic review of hospital charts. The follow-up period began at the day of surgery and ran until March 1, 2004. The primary end point was a composite, corresponding to the occurrence of at least one of these events during the follow-up: cardiovascular death, non-fatal acute coronary syndrome corresponding to any type of myocardial infarction or unstable angina after surgery stay discharge requiring admission to a coronary care unit, secondary coronary revascularization, non-fatal stroke or transient ischemic attack (TIA) requiring admission to a neurology department and defined by a neurologist during the hospital stay, and peripheral vascular surgery. Secondary end points studied were overall mortality, cardiovascular mortality, acute coronary syndrome, and stroke or TIA.
Data are reported as mean ± SD. The Kaplan-Meier survival method was used for the comparison of survival according to the absence or presence of any pattern of PAD using the log-rank test. The univariate and multivariate analyses were performed using a Cox proportional-hazards model. For this purpose, all factors presenting a p value ≤0.25 in the univariate analysis were introduced in the multivariate model. A p value ≤0.05 was considered statistically significant. The software for statistical analysis was Statview 5.0 for Windows (SAS Institute, Cary, North Carolina).
We consecutively included 1,022 patients (186 women and 836 men) with a mean age of 66.9 ± 9.2 years who were referred to our institution for CABG. General data are reported in Table 1.
In addition to the 141 patients (13.8%) presenting with clinical PAD, 135 patients (13.2%) presented with subclinical occlusive PAD with an ABI <0.85 and 124 patients (12.1%) presented with medial artery calcification in at least one ankle artery.
The patients benefited from 3.02 ± 0.98 coronary bypasses. Coronary revascularization was considered complete in 872 cases (85.4%). Off-pump CABG was performed in 136 cases (13.3%). Coronary artery bypass grafting was combined with a valvular surgery in 159 cases (15.6%). In 33 cases (3.2%), carotid surgery was associated with coronary bypass.
The actuarial follow-up period was 4.4 years (range: 0 to 65.1 months). During this period, 191 patients (18.7%) met the primary end point. Details of the events are reported in Table 2.
Primary end point
Table 3displays the univariate and multivariate analyses predicting the occurrence of the primary end point. The following factors were independently predictive of the occurrence of the primary end point during the follow-up period: preoperative supraventricular arrhythmia, an ejection fraction <40%, a concomitant valvular surgery, and the presence of a clinical or subclinical PAD, as well as the presence of peripheral medial artery calcification. Figure 1displays the survival curves according to the absence and presence of any pattern of PAD.
Secondary end points
During the follow-up period, 105 patients (10.3%) died, including 68 cardiovascular deaths (6.7%). The univariate and multivariate analyses for overall and cardiovascular mortality are displayed in Table 4.Female gender, a concomitant valvular surgery, and all types of PAD were independently predictive for both overall and cardiovascular mortality. In addition, age and renal insufficiency were also independent predictive factors for overall mortality, whereas supraventricular arrhythmia and a low ejection fraction were predictive of cardiovascular death.
Fatal and non-fatal acute coronary events occurred in 38 patients (3.7%). Table 5displays the corresponding univariate and multivariate analyses. A supraventricular arrhythmia and a subclinical PAD (ABI <0.85) were the two independent predictive factors.
Fatal and non-fatal stroke or TIA occurred in 61 patients (6.0%). Table 5displays the corresponding univariate and multivariate analyses. A concomitant valvular or carotid surgery and a clinical or subclinical PAD were the independent factors predicting the occurrence of stroke or TIA during follow-up.
In this prospective study, the measurement of ABI turns out to be a major predictive marker in patients after a coronary bypass surgery. This prognostic value is independent of many usual prognostic factors, including the presence of a clinical PAD.
In patients undergoing CABG, clinical PAD is a well-known factor of poor short-term (2–7) and long-term prognosis (1,8–11). Conversely, data on subclinical PAD and its prognostic interest are scarce. In a substudy of the Bypass Angioplasty Revascularization Investigation (BARI), Burek et al. (12) followed up 405 patients undergoing either coronary angioplasty or coronary bypass (171 patients) during three years with an initial measurement of ABI. They defined the presence of PAD according to a low ABI (<0.90). They did not determine any specific subgroup with abnormally high ABI and included those patients in the “no-PAD” group. In their results, those with a low ABI presented with a more severe prognosis at follow-up without providing specific results for CABG. Moreover, the investigators did not make any distinction between clinical or subclinical PAD. For the short-term prognosis, a Finnish study (13) of 178 CABG patients presented low ABI predictive of postoperative myocardial infarction and atrial fibrillation, without any distinction according to clinical status or elevated ABI. To our knowledge, our study is the first providing evidence on the additive predictive power of ABI to the clinical presence of PAD. The presence of a subclinical occlusive PAD is associated with a three-fold excess risk of a subsequent cardiovascular event after CABG. Moreover, the larger size of our cohort enabled us to confirm this hypothesis through several end points. The risk excess was also noted for total and cardiovascular mortality in both conditions. An asymptomatic low ABI was even an independent predictive marker of subsequent acute coronary or cerebrovascular events.
Medial arterial calcification is a condition commonly observed in diabetes (16,17) and end-stage renal disease (20), and its prevalence also rises with age (14,15). The presence of medial artery calcification diagnosed by radiography is a marker of poor cardiovascular prognosis in diabetics (17). An incompressible ankle artery or a high ABI are indirect methods to diagnose this condition (18).
Recently, the Strong Heart Study (18) provided important data on the predictive value of ABI in a general population. After a follow-up of 4,393 American Indians for more than eight years, Resnick et al. (18) clearly evidenced a U-shaped association between ABI and mortality risk. They noted that an elevated ABI was also a significant independent prognostic marker.
In the field of secondary coronary prevention, the prognostic significance of medial artery calcification in peripheral arteries was still unknown. According to our results, the presence of high ABI is a strong prognostic marker, independent of frequently associated conditions that are per se prognostic, namely age (3–5,10), diabetes (4,10,21,22), and renal insufficiency (3–5,23,24). Hence, a high ABI could be an integrative marker reflecting arterial stiffening and calcification after these different risk factors.
Unlike that of other vascular markers, the measurement of ABI is easy and largely available, as it requires small and cheap materials (a hand-held pocket Doppler device and a blood pressure cuff). Therefore, because ABI is of great prognostic value and abnormal ABI is frequent among CABG candidates, we recommend its preoperative measurement for risk assessment by physicians (cardiologists, anesthesiologists, or surgeons) managing these patients.
In this prospective study, the preoperative measurement of ABI presented a strong power to predict further cardiovascular events in patients undergoing CABG. Its predictive ability is independent and additive to the presence of a clinical PAD and other classic predictors. Both low and elevated ABI are of prognostic significance. These results and the large availability of this vascular marker lead us to recommend its measurement in the preoperative risk assessment of every candidate for coronary bypass surgery.
- Abbreviations and Acronyms
- ankle-brachial index
- coronary artery bypass grafting
- peripheral arterial disease
- transient ischemic attack
- Received February 28, 2005.
- Revision received May 3, 2005.
- Accepted May 15, 2005.
- American College of Cardiology Foundation
- Eagle K.A.,
- Rihal C.S.,
- Foster E.D.,
- Mickel M.C.,
- Gersh B.J.,
- The Coronary Artery Surgery Study (CASS) Investigators
- Nashef S.A.M.,
- Roques F.,
- Michel P.,
- Gauducheau E.,
- Lemeshow S.,
- Salamon R.
- Roques F.,
- Gabrielle F.,
- Michel P.,
- de Vincentiis C.,
- David M.,
- Baudet E.
- Staat P.,
- Cucherat M.,
- George M.,
- et al.
- Jones R.H.,
- Hannan E.L.,
- Hammermeister K.E.,
- et al.
- Ivanov J.,
- Weissel R.,
- David T.E.,
- Naylor C.D.
- Burek K.A.,
- Sutton-Tyrrel K.,
- Brooks M.M.,
- et al.
- Loponen P.,
- Taskinen P.,
- Laakkonen E.,
- et al.
- Coni N.,
- Tennison B.,
- Troup M.
- Lehto S.,
- Niskanen L.,
- Suhonen M.,
- Ronnemaa T.,
- Laakso M.
- Resnick H.E.,
- Lindsay R.S.,
- McDermott M.M.,
- et al.
- Cole J.H.,
- Miller J.I. 3rd.,
- Sperling L.S.,
- Weintraub W.S.
- Gupta R.,
- Birnbaum Y.,
- Uretsky B.F.