Advanced atherosclerotic disease increases the risk of stent thrombosis after drug-eluting stent (DES) implantation. We aimed to determine if an abnormal ankle–brachial index (ABI) value as a surrogate of atherosclerotic disease and vascular inflammation provides information on 1-year risk of cardiovascular events after DES implantation. A prospective cohort of 1,437 consecutive patients undergoing DES implantation from January through April 2008 in 26 Spanish hospitals was examined. ABI was calculated by Doppler in a standardized manner. Patients were followed to 12 months after the percutaneous coronary intervention to determine total and cardiovascular mortality, stroke, nonfatal acute coronary syndrome (ACS), and new revascularizations. Association of an abnormal ABI value (i.e., ≤0.9 or ≥1.4) with outcomes was assessed by conventional logistic regression and by propensity-score analysis. Patients with abnormal ABI values (n = 582, 40.5%) in general had higher global cardiovascular risk, the reason for DES implantation was more often ACS, and had a higher rate of complications during admission (heart failure or stroke or major hemorrhage 11.3% vs 5.3%, p <0.001). An abnormal ABI value was independently associated with 1-year total mortality (odds ratio 2.23, 95% confidence interval 1.13 to 4.4) and cardiovascular mortality (odds ratio 2.06, 95% confidence interval 1.04 to 4.22). No independent association was found between an abnormal ABI value and 1-year nonfatal ACS, stroke, and new revascularizations. In conclusion, although an abnormal ABI value was associated with fatal outcomes in patients receiving DESs, no association was found with nonfatal ACS and new revascularizations. A clear relation between abnormal ABI and surrogates of DES thrombosis could not be established.
In the Adherence to Treatment of Coronary Patients after a Catheterization with Drug-Eluting Stent (DES) Implantation (ACDC) study, the hypothesis was advanced that an abnormal ankle–brachial index (ABI) value as a surrogate of atherosclerotic disease burden would lead to a higher risk of DES thrombosis and thus more cardiovascular events. The aim of the present study was to determine in the population of the ACDC study if the ABI provides information on the risk of short-term (1-year) cardiovascular events or mortality independently of conventional cardiovascular risk factors after DES implantation.
Methods
The method of the ACDC study has been reported elsewhere. In brief, it is a prospective cohort of consecutive patients who underwent ≥1 DES implantation in 29 Spanish hospitals from January 28 through April 28, 2008. There were no exclusion criteria. Specifically designated and trained field researchers entered data into an electronic database during patient admission including variables about sociodemographic data, cardiovascular risk factors, noncardiac and cardiac histories, and date of admission. All study variables had standard definitions that were discussed and accepted by field researchers in 3 workshops conducted before the study. To test the hypothesis of the present study, investigators were trained to perform ABI measurements during admission according to American Heart Association recommendations. The ABI value for each leg was calculated using a standardized Doppler ultrasonic device (Bidop ES-100V3, Hadeco, St. Louis, Missouri). ABI values ≤0.9 or ≥1.4 were considered abnormal. All patients who signed an informed consent were followed 3, 6, 9, and 12 months after the percutaneous coronary intervention by a specifically trained professional team (Proyecta’m Company, Barcelona, Spain). The telephone interview was performed according to a structured questionnaire on vital status, other medical problems, new admissions, drug therapy, and changes in drug therapy. When patients reported a new admission, the clinical record was retrieved for classification of the outcomes of interest by the main investigator team. Outcomes considered were acute coronary syndrome (ACS), stroke, coronary revascularization, total mortality, and cardiovascular mortality. ACS, stroke, and new coronary revascularization were considered when these events were clearly reported in the clinical records of subsequent admissions. Cardiovascular death was defined as death secondary to ACS, heart failure, sudden death, or unexplained death (i.e., death clearly not secondary to noncardiovascular causes such as cancer, sepsis, trauma, etc.). Consecutive inclusion was surveyed by visiting all participating centers during the inclusion period and reviewing the log of percutaneous coronary interventions, which was checked against the list of included patients. Quality of data during admission was ensured before the end of the inclusion period by analysis of a random sample of 10% to 15% of clinical records performed by a member of the research team. When >5 disagreements were detected, researchers were asked to review the corresponding variables in all included patients. For the present analysis, only the 26 centers in which >70% of patients had undergone ABI measurement were included (3 centers were excluded).
Continuous variables are summarized as mean ± SD or median (interquartile range) where appropriate and categorical data as percentage. The independent association of an abnormal ABI value with the outcomes of interest was assessed by conventional logistic regression and by propensity-score analysis. In conventional logistic regression, potential predictive variables were selected by clinical plausibility and their association with the outcome in bivariate analysis (selecting all variables with a p value <0.2). The final model was obtained using a backward stepwise method with a threshold for exit set at a p value >0.20. We forced the variable “abnormal ABI” into the best model to obtain the independent association of ABI to the outcomes of interest. Variables for adjustment in the final model for the outcomes mortality and cardiovascular mortality were age, previous heart failure, previous stroke, renal failure, and number of treated vessels. Variables for the outcome ACS were gender, age, previous heart failure, renal failure, previous myocardial infarction, hypertension, number of treated vessels, number of stents implanted, and complications during admission. Variables for the outcome revascularization were gender, history of peripheral vascular disease, chronic pulmonary disease, number of treated vessels, number of stents implanted, off-label indication, and total vessel length stented.
The propensity score, which represents the probability that a patient had an ABI measurement, was computed using extensive nonparsimonious logistic regression modeling with the following covariates: age, gender, hypertension, diabetes, peripheral vascular disease, hypercholesterolemia, smoking status, atrial fibrillation, chronic pulmonary disease, renal failure, previous stroke, previous myocardial infarction, previous heart failure, previous coronary intervention, previous cardiac surgery, Killip class, reason for admission (including type of ACS), participation in a clinical trial, pharmacologic treatment, number of vessels with significant stenosis, left main coronary artery disease, number of stents implanted, ejection fraction (including a category for missing ejection fraction), and complications during admission (heart failure, cardiogenic shock). The resulting propensity score was then used for adjustment of the association of an abnormal ABI value with the outcome events by including the 2 variables in a logistic regression model as independent variables. In addition, we performed a propensity-score matched-paired analysis. We matched each subject with a pathologic ABI value to the closest available subject with a normal ABI value based on the estimated propensity score using a greedy-matching algorithm. Pairs with >0.2 SD (logit-Propensity Score) distance were excluded. A logistic regression model to assess the impact of an abnormal ABI value on outcomes was then estimated using generalized estimating equation methods with a robust estimation of variance to incorporate the matched-pairs design.
As a sensitivity analysis to assess whether exclusion of patients with missing ABI values could have introduced some bias to the final results, we imputed values of ABI in patients with incomplete or missing data for ABI calculations using multiple-imputation methods. We defined an extensive nonparsimonious multiple regression model to impute missing values. Predictive variables included in the model were sociodemographic variables, risk factors, cardiovascular history, co-morbid conditions, diagnosis at admission, complications during admission, and complications during follow-up. Twenty datasets were imputed. Those variables likely to be associated with missing ABI and observed ABI values were included in the imputation model. We assumed that missing ABI values were not associated with unobserved variables, i.e., that these values were missing at random.
Results
After exclusion of 3 centers with poor compliance with the protocol, there were 1,762 patients included in the study. Baseline ABI values were missing in 82 patients and 18 died during admission. Two hundred twenty-five were then lost to follow-up. Therefore, the study cohort for the present analysis consisted of 1,437 patients ( Figure 1 ), 582 with abnormal ABI values and 855 with normal ABI values for whom complete follow-up was available.
Baseline sociodemographic and clinical characteristics are listed in Table 1 for all patients and for those with normal and abnormal ABI values. Table 2 presents features during admission and at discharge. Patients with abnormal ABI values were older and had higher prevalences of diabetes mellitus, cardiovascular history of heart failure, stroke, peripheral artery disease, atrial fibrillation, and chronic renal failure. Also, they had been hospitalized because of ACS more often and, as expected, were included in clinical trials during admission less frequently than patients with normal ABI values ( Table 2 ). However, variables associated with anatomic severity or progression of coronary artery disease (number of diseased vessels or left main coronary artery disease) were similarly distributed in the 2 groups, as were procedure characteristics (number of lesions treated, use of IIb/IIIa inhibitors), number of off-label indications, or procedural complications. Seven patients died during admission, 5 with abnormal and 2 with normal ABI values (p = 0.1). In general, there were more complications during admission (heart failure, cardiogenic shock, stroke, major hemorrhage) in those patients with abnormal ABI values, although only the former reached statistical significance.
| All Patients | Abnormal ABI Value | Normal ABI Value | p Value | |
|---|---|---|---|---|
| (n = 1,437) | (n = 582) | (n = 855) | ||
| Sociodemographic and cardiovascular risk factors | ||||
| Age (years), mean ± SD | 64.2 ± 11.2 | 66 ± 11.5 | 62.9 ± 10.9 | <0.001 |
| Women | 305 (21.2%) | 134 (23%) | 171 (20%) | 0.169 |
| Active smoker | 398 (27.7%) | 159 (27.3%) | 239 (28%) | 0.792 |
| Hypercholesterolemia | 874 (60.8%) | 348 (59.8%) | 526 (61.5%) | 0.51 |
| Hypertension | 958 (66.7%) | 400 (68.7%) | 558 (65.3%) | 0.171 |
| Diabetes mellitus | 533 (37.1%) | 235 (40.4%) | 298 (34.9%) | 0.033 |
| Cardiovascular history | ||||
| Peripheral artery disease | 179 (12.5%) | 117 (20.1%) | 62 (7.3%) | <0.001 |
| Stroke | 80 (5.6%) | 43 (7.4%) | 37 (4.3%) | 0.013 |
| Heart failure | 85 (5.9%) | 48 (8.3%) | 37 (4.3%) | 0.002 |
| Pacemaker | 17 (1.2%) | 7 (1.2%) | 10 (1.2%) | 0.954 |
| Implantable cardiac defibrillator | 6 (0.4%) | 2 (0.3%) | 4 (0.5%) | 0.720 |
| Valvular prosthesis | 7 (0.5%) | 4 (0.7%) | 3 (0.4%) | 0.369 |
| Atrial fibrillation | 63 (4.4%) | 34 (5.8%) | 29 (3.4%) | 0.026 |
| Previous acute myocardial infarction | 423 (29.4%) | 183 (31.4%) | 240 (28.1%) | 0.168 |
| Previous percutaneous transluminal coronary angioplasty | 403 (28%) | 175 (20.1%) | 228 (26.7%) | 0.159 |
| Previous coronary artery bypass grafting | 114 (7.9%) | 51 (8.8%) | 63 (7.3%) | 0.337 |
| Co-morbid conditions | ||||
| Chronic obstructive pulmonary disease | 146 (10.2%) | 69 (11.5%) | 79 (9.2%) | 0.162 |
| Chronic renal failure | 105 (7.3%) | 54 (9.3%) | 51 (6%) | 0.018 |
| Previous major hemorrhage | 34 (2.4%) | 18 (3.1%) | 16 (1.9%) | 0.29 |
| All Patients | Abnormal ABI Value | Normal ABI Value | p Value | |
|---|---|---|---|---|
| (n = 1,437) | (n = 582) | (n = 855) | ||
| Diagnosis at admission | <0.001 | |||
| ST-segment elevation acute coronary syndrome | 220 (15.3%) | 99 (17.1%) | 121 (14.2%) | |
| Non–ST-segment elevation acute coronary syndrome | 608 (42.4%) | 250 (43.1%) | 358 (41.9%) | |
| Coronary disease, elective admission | 470 (32.8%) | 155 (26.7%) | 315 (36.8%) | |
| Other | 137 (9.6%) | 76 (13.1%) | 61 (7.1%) | |
| Patients included in clinical trial | 200 (13.9%) | 62 (10.7%) | 138 (16.1%) | 0.003 |
| Index procedure characteristics | 0.519 | |||
| Number of diseased vessels at index procedure | ||||
| 1 | 726 (50.5%) | 281 (48.3%) | 445 (52.1%) | |
| 2 | 423 (29.4%) | 177 (30.4%) | 246 (28.8%) | |
| 3 | 191 (13.3%) | 107 (12.5%) | 84 (14.4%) | |
| Left main coronary artery disease | 116 (8.1%) | 49 (8.4%) | 67 (7.8%) | 0.69 |
| Total number of lesions treated | 0.688 | |||
| 1 | 961 (66.9%) | 389 (66.8%) | 572 (66.9%) | |
| 2 | 350 (24.4%) | 146 (25.1%) | 204 (23.9%) | |
| ≥3 | 126 (8.8%) | 47 (8.1%) | 79 (9.2%) | |
| IIb/III inhibitors given during procedure | 301 (21%) | 120 (20.6%) | 181 (21.2%) | 0.801 |
| Total number of stents | 0.378 | |||
| 1 | 671 (46.7%) | 273 (46.9%) | 398 (46.6%) | |
| 2 | 413 (28.7%) | 176 (30.2%) | 237 (27.7%) | |
| ≥3 | 353 (24.6%) | 133 (22.9%) | 220 (25.7%) | |
| Total vessel length stented, mean ± SD | 37.4 ± 24.7 | 36.3 ± 23.8 | 38.2 ± 25.2 | 0.076 |
| Off-label indication | 913 (63.5%) | 357 (61.3%) | 556 (65%) | 0.154 |
| Angiographic success | 1,424 (99.2%) | 579 (99.7%) | 845 (99%) | 0.13 |
| Procedural clinical complications ⁎ | 36 (2.5%) | 21 (3.6%) | 15 (1.8%) | 0.027 |
| Ejection fraction (valid, n = 1,094) | <0.001 | |||
| <30% | 34 (3.1%) | 17 (3.7%) | 17 (2.7%) | |
| 30%–44% | 146 (13.4%) | 68 (14.6%) | 78 (12.4%) | |
| 45%–55% | 264 (24.1%) | 142 (30.5%) | 122 (19.4%) | |
| >55% | 650 (59.4%) | 239 (51.3%) | 411 (65.5%) | |
| Complications during admission | ||||
| Heart failure | 86 (6%) | 56 (9.6%) | 30 (3.5%) | <0.001 |
| Cardiogenic shock | 14 (1%) | 9 (1.6%) | 5 (0.6%) | 0.068 |
| Stroke | 6 (0.4%) | 4 (0.7%) | 2 (0.2%) | 0.191 |
| Major hemorrhage | 8 (0.6%) | 5 (0.9%) | 3 (0.4%) | 0.204 |
| Any of above | 110 (7.7%) | 65 (11.2%) | 45 (5.3%) | <0.001 |
| Hemoglobin at discharge, mean ± SD | 13.4 ± 1.7 | 13.1 ± 1.7 | 13.6 ± 1.6 | <0.001 |
| Treatment at discharge | ||||
| β Blockers | 1,075 (74.8%) | 437 (75.1%) | 638 (74.6%) | 0.842 |
| Calcium channel blockers | 36 (2.5%) | 25 (4.3%) | 11 (1.3%) | <0.001 |
| Angiotensin-converting enzyme inhibitors | 730 (50.8%) | 329 (56.5%) | 401 (46.9%) | <0.001 |
| Statins | 1,218 (84.8%) | 500 (85.9%) | 718 (84%) | 0.317 |
| Diuretics | 210 (14.6%) | 106 (18.2%) | 104 (12.2%) | 0.001 |
| Insulin | 164 (11.4%) | 84 (14.4%) | 80 (9.4%) | 0.003 |
| Oral antidiabetic drugs | 300 (20.9%) | 128 (22%) | 172 (20.1%) | 0.39 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
