The impact of polyvascular disease on long-term outcome in percutaneous coronary intervention patients

Authors


Abstract

Background

Previous studies demonstrated the prognostic importance of concomitant polyvascular disease in patients with coronary artery disease (CAD). However, the significance of the number of diseased vascular territories and subclinical disease is unknown.

Materials and methods

The number of diseased vascular territories was evaluated in 2299 percutaneous coronary intervention (PCI) patients. Vascular disease was defined by documented atherosclerotic disease, either diagnosed in the medical history (clinical) or at the standardized cardiovascular screening (subclinical). The following territories were evaluated: cerebrovascular disease, peripheral arterial disease, abdominal aortic aneurysm and vascular renal disease. The outcome measures were all-cause mortality, cardiovascular mortality and a composite cardiovascular endpoint (myocardial infarction, stroke, cardiovascular mortality). Patients with monovascular disease (CAD) served as the reference category. Hazard ratios (HRs) were adjusted for baseline characteristics.

Results

Mean follow-up was 7·3 years. The HRs (95% confidence interval) for patients with two diseased territories compared to monovascular disease were for all-cause mortality 1·60 (1·14–2·25), cardiovascular mortality 2·13 (1·29–3·50) and the combined cardiovascular endpoint 1·66 (1·20–2·31). Moreover, the HRs (95% confidence intervals) for patients with more than two diseased territories compared to monovascular disease were for all-cause mortality 3·81 (2·45–5·92), cardiovascular mortality 4·40 (2·32–8·35) and the combined cardiovascular endpoint 2·75 (1·69–4·47). The HRs of patients with subclinical disease were comparable to the HRs of patients with clinical disease.

Conclusions

In patients undergoing PCI, the presence of subclinical and clinical polyvascular disease is associated with an increased long-term mortality and morbidity. Moreover, the outcome is highly influenced by the number of diseased territories.

Introduction

Patients with atherosclerotic disease have an increased risk of concomitant arterial disease in other vascular territories, because atherosclerosis is a progressive and generalized process [1]. It has been shown that the prognosis of patients with symptomatic polyvascular disease is impaired in comparison with patients with only one atherosclerotic vascular bed. The OPUS-TIMI 16 study demonstrated that in patients with coronary artery disease (CAD) who presented with an acute coronary syndrome (ACS), the presence of prior clinical atherosclerotic cardiovascular disease (cerebrovascular disease [CVD] and peripheral arterial disease [PAD]) was associated with a worse 10 months outcome [2]. In the Dynamic Registry, clinical atherosclerotic disease on top of CAD was an independent predictor of both in-hospital cardiovascular events, and death or myocardial infarction at 1 year after PCI [3]. Also long-term survival is impaired in patients undergoing coronary revascularization with concomitant PAD as compared to patients without PAD [4, 5]. Although the previously mentioned studies have shown the prognostic importance of polyvascular disease in a population with known CAD, the prognostic significance of the number of diseased atherosclerotic vascular territories was not studied. Moreover, the importance of subclinical concomitant atherosclerotic disease in all vascular territories in patients undergoing PCI was not considered. Therefore, we analysed the impact of the number of diseased vascular territories on top of CAD on long-term outcome in a large cohort of patients undergoing PCI. In addition, we addressed whether subclinical polyvascular disease leads to a comparable outcome as clinical polyvascular disease.

Methods

Study design and patient population

The Second Manifestations of ARTerial disease (SMART) study is an ongoing prospective follow-up study at the University Medical Centre Utrecht in the Netherlands. Since 1996, newly referred patients, aged 18–80 years, with traditional cardiovascular risk factors or with clinical arterial disease were included. Patients with end-stage malignancy, those dependent in daily activities or not sufficiently fluent in the Dutch language were excluded. A detailed description of the study was previously published [6]. The SMART study protocol was approved by The Ethics Committee of our hospital. In short, patients who gave their written informed consent were asked to fill in a standardized health questionnaire and to undergo a standardized vascular screening that includes physical examination, laboratory tests, electrocardiogram (ECG), ankle-brachial index and ultrasonographic examination of the abdominal aorta and carotid arteries. All patients received a personalized cardiovascular secondary prevention therapy advice based on the findings of the screening. In the present analysis, only patients who participated in the SMART study after undergoing a PCI were included (n = 2299). These patients were included between April 1996 and March 2012 and were followed until March 2012 or death.

Definitions, follow-up procedure and endpoint evaluation

Definitions

The number of diseased vascular territories on top of their known CAD was determined in all patients. The vascular territories that were taken into account were CVD, PAD, aneurysm of the abdominal aorta (AAA) and vascular renal disease. The definition of atherosclerosis in a vascular territory consisted of either clinical arterial disease (medical history) or subclinical atherosclerosis that was determined by the SMART vascular screening. A medical history of CVD consisted of either a stroke or carotid endarterectomy. Patients were considered to have PAD when they underwent an amputation, bypass surgery or percutaneous transluminal angioplasty (PTA) of the peripheral arteries in the past. A medical history of AAA comprised the diagnosis of an AAA treated conservatively or with open/endovascular surgery. Macrovascular renal disease was considered present when patients had documented renal artery disease at angiography. The standardized SMART cardiovascular screening for the evaluation of subclinical disease comprised a duplex ultrasonography of the carotid artery (cut-off > 50% stenosis), an ankle-brachial index to determine PAD (cut-off < 0·9) and ultrasonography of the abdominal aorta (cut-off ≥ 3·5 cm). Microvascular renal disease was considered present when there was either macroproteinuria (>30 mg albumin/mmol creatinine in 24-h urine sample) or microproteinuria (>3 < 30 mg albumin/mmol creatinine in 24-h urine sample) in combination with an estimated glomerular filtration rate (eGFR) in a blood sample below 60 mL/min/1.73 m2 or an eGFR below 30 mL/min/1.73 m2. Patients were divided in subgroups based on the number of diseased atherosclerotic arterial territories. Patients with solely CAD were considered to have monovascular disease. Patients with one concomitant diseased vascular territory on top of CAD were considered to have two diseased vascular territories. Patients with two or more diseased vascular territories in addition to their CAD were categorized as more than two diseased vascular territories.

Hypertension was defined by a systolic blood pressure ≥ 140 mmHg and/or the use of antihypertensive drugs. In patients with diabetes, a blood pressure above 130/85 mmHg was classified as hypertension [7, 8]. Hyperlipidaemia at screening was defined by a LDL cholesterol ≥ 2·5 mM or triglycerides > 2·0 mM or HDL cholesterol ≤ 1·0 mM in men and HDL cholesterol ≤ 1·3 mM in women. Among subjects without a history of diabetes, those with a fasting plasma glucose level >11·1 mM at baseline or with fasting plasma glucose ≥7·0 mM at baseline and receiving treatment with glucose-lowering agents within 1 year after baseline were considered as having diabetes at baseline. High-sensitive CRP (hs-CRP) was measured in all patients to evaluate the inflammatory state.

Follow-up procedure

All patients were followed every 6 months with use of a standardized questionnaire or by telephone to find out whether a cardiovascular event or arterial intervention had occurred. When a possible event was reported, hospital discharge letters were retrieved to verify the diagnosis. Three members of the SMART Endpoint Committee independently adjudicated all events. This Committee consists of physicians from different cardiovascular specialties. In case of disagreement, the event was evaluated in detail by members of the SMART study group.

Data-analysis

Patients with polyvascular disease (two diseased vascular territories or more than two diseased vascular territories) were compared to patients with monovascular disease (CAD only). The measures of outcome were all-cause mortality, cardiovascular mortality and a combined cardiovascular endpoint (composite of cardiovascular mortality, stroke and myocardial infarction). Cardiovascular mortality was defined by sudden death (unexpected cardiac death occurring within 1 h after onset of symptoms or within 24 h given convincing circumstantial evidence) or death from stroke, myocardial infarction, congestive heart failure or ruptured AAA. Myocardial infarction was defined by a combination of at least two of the following: (i) chest pain for at least 20 min, (ii) ST elevation >1 mm in at least two consecutive leads or a new left bundle branch block on the ECG, (iii) CK elevation of at least two times the normal value of CK and a MB-fraction >5% or a troponin rise exceeding the upper limit of normal.

Descriptives are expressed as mean (standard deviation) for continuous variables that have a normal distribution and as mean (range) for continuous variables that are not normally distributed. Categorical variables are presented as numbers (percentages). Difference in all-cause mortality, cardiovascular death and cardiovascular outcome between patients with 1 (CAD), 2 (CAD + 1) and at least 3 (CAD + 2 or more) diseased vascular territories was calculated with Cox proportional hazard model analysis. Patients with monovascular disease (CAD only) served as the reference category in the analyses. Any first occurrence of an event during the follow-up period was used in the model. Results are expressed as hazard ratios (HR) with 95% confidence intervals (95% CI). All analyses were conducted with three different models. One model adjusting for age and gender (model I) and a second model adjusting for potential confounding factors besides age and gender: diabetes mellitus, packyears (20 cigarettes a day/year, in quartiles), hypertension, hyperlipidaemia, body mass index (BMI) and previous myocardial infarction. The latter was added in an attempt to approximate the left ventricular function. Model III was performed in the subgroup of patients with known extent of CAD. On top of all the potential confounding factors of model II, we adjusted in model III also for the extent of CAD.

A subgroup analysis was performed to compare the impact of subclinical and clinical polyvascular disease on the prognosis of PCI patients. All statistical analyses were performed with IBM SPSS 20.0 for Windows (IBM Corporation, Armonk, NY, USA).

Results

Baseline characteristics

A total of 2299 patients who underwent a PCI between 1996 and 2012 at the University Medical Center Utrecht, the Netherlands, were included in the SMART study. A total of 462 (21%) patients had polyvascular disease and were categorized by the number of diseased vascular territories. The baseline characteristics are summarized in Table 1. Patients with polyvascular disease were older and had an unfavourable risk profile as illustrated by more extensive CAD, heavy smoking and a higher prevalence of hypertension and diabetes mellitus. The median hs-CRP levels increased gradually with the number of diseased vascular territories. Patients with polyvascular disease use more ACE inhibitors, diuretics, glucose-lowering medication and oral anticoagulants, but less antiplatelet drugs. The use of statins and beta-blockers was not different.

Table 1. Baseline characteristics according to the number of diseased vascular territories
 One territory n (%) 1837 (80)Two territories n (%) n = 375 (16)More than two territories n (%) n = 87 (4)
  1. N, Number; SD, standard deviation; IQR, interquartile range.

Age in years, mean (SD)58·6 (9·4)63·0 (9·1)64·7 (7·0)
Male1518 (83)289 (77)70 (81)
Packyears, median (IQR)21·6 (9·8–33·6)21·6 (9·9–35·1)27·3 (10·8–44·1)
Previous myocardial infarction850 (46)172 (46)42 (48)
Extent of coronary artery disease
One vessel785 (43)141 (38)27 (31)
Two vessel538 (29)120 (32)25 (29)
Three vessel234 (13)59 (16)26 (30)
Unknown280 (15)55 (15)9 (10)
Hypertension805 (44)231 (62)55 (63)
Hyperlipidaemia613 (33)119 (32)37 (43)
Diabetes251 (14)77 (21)26 (30)
Body mass index (kg/m²), median (IQR)27·1 (24·9–29·4)26·9 (25·0–29·7)27·2 (25·5–30·7)
High-sensitive CRP, median (IQR)1·5 (0·8–3·0)2·1 (1·1–4·1)3·3 (1·8–5·8)
Medication use
Beta-blockers1481 (81)289 (77)68 (78)
ACE inhibitors520 (28)123 (33)36 (41)
Diuretics267 (15)105 (28)33 (38)
Statins1162 (63)262 (70)58 (67)
Platelet aggregation inhibitors1664 (91)336 (90)72 (83)
Oral anticoagulants111 (6)47 (13)21 (24)
Platelet aggregation inhibitors or oral anticoagulants1704 (93)349 (93)78 (90)
Oral glucose-lowering medication154 (8)54 (14)14 (16)
Insulin64 (4)28 (8)6 (7)

Type of diseased vascular territories

The type of diseased vascular territories in the patients with polyvascular disease is shown in Table 2. Remarkably, 48% of patient with more than two diseased territories had a carotid stenosis of more than 50%. Moreover, 59% of these patients had a diminished ankle-brachial index (<0·9) and 30% had microvascular renal disease.

Table 2. Type of diseased vascular territory according to clinical (medical history) or subclinical (finding at screening) polyvascular disease
 Two territories number (%) 375More than two territories number (%) 87
  1. CAD, Coronary artery disease.

  2. *Including CAD.

  3. a

    Subclinical: atherosclerotic disease diagnosed with SMART standardized cardiovascular screening; PTA, percutaneous transluminal angioplasty.

  4. b

    Estimated glomerular filtration rate (eGFR) ≤ 30 or eGFR ≤60 and microalbuminuria or macroalbuminuria.

Cerebrovascular disease
Clinical50 (13)22 (25)
Stroke46 (12)15 (17)
Carotid endarterectomy6 (2)9 (10)
Subclinicala
Carotid stenosis > 50%66 (18)42 (48)
Peripheral arterial disease
Clinical48 (13)24 (28)
Amputation8 (2)2 (2)
Bypass surgery/PTA43 (11)23 (26)
Subclinicala
Ankle-brachial index < 0·999 (26)51 (59)
Aneurysm of the abdominal aorta
Clinical26 (7)21 (24)
Open/endovascular repair4 (1)7 (8)
Conservative management2214
Subclinicala
Aorta diameter of ≥ 3·5 cm5 (1)7 (8)
Renal disease
Clinical49 (13)13 (15)
PTA1 (0)0 (0)
Subclinicala
Microvascular renal diseaseb57 (15)26 (30)

Incidence of cardiovascular events

During a mean follow-up time of 7·3 (±4·0) years, 62 patients (2·7%) were lost to follow-up. A total of 211 patients (9%) died of whom 92 (44%) from a cardiovascular death (Table 3). A total of 243 patients (11%) developed the combined cardiovascular endpoint.

Table 3. The risk of the number of diseased vascular territories on long-term mortality, cardiovascular mortality and a combined cardiovascular endpoint
 One territory n = 1837 Reference categoryTwo territories n = 375 HR (95% CI)More than two territories n = 87 HR (95% CI)
  1. HR, Hazard ratio; CI, confidence interval. Model I, hazard ratios (HRs) adjusted for age and gender; Model II, HRs adjusted for age, gender, diabetes mellitus, hypertension, hyperlipidaemia, body mass index, packyears (in quartiles) and previous myocardial infarction; Model III*, subgroup analysis in patients with known extent of coronary artery disease (CAD) and adjusted for extent of CAD on top of age, gender, diabetes mellitus, hypertension, hyperlipidaemia, body mass index, packyears (in quartiles) and previous myocardial infarction. Number of patients in subgroup analysis: one territory: 1555 patients, two territories: 319 patients, more than two territories: 78 patients.

All-cause mortality [no. of events (%)]137 (7)48 (13)26 (30)
Model I1·001·55 (1·11–2·17)3·92 (2·57–5·99)
Model II1·001·60 (1·14–2·25)3·81 (2·45–5·92)
Model III*1·001·53 (1·05–2·22)3·41 (2·11–5·50)
Cardiovascular mortality [no. of events (%)]57 (3)22 (6)13 (15)
Model I1·001·78 (1·08–2·94)4·85 (2·63–8·92)
Model II1·002·13 (1·29–3·50)4·40 (2·32–8·35)
Model III*1·001·80 (1·02–3·18)3·38 (1·65–6·92)
Combined cardiovascular endpoint [no. of events (%)]173 (9)50 (13)20 (23)
Model I1·001·58 (1·15–2·18)2·92 (1·82–4·67)
Model II1·001·66 (1·20–2·31)2·75 (1·69–4·47)

Model III*

(myocardial infarction, stroke, cardiovascular death)

1·001·55 (1·08–2·23)2·25 (1·33–3·82)

Survival analysis

Compared to patients with monovascular disease (CAD only), patients with polyvascular disease were associated with a higher risk of all-cause mortality, cardiovascular mortality and the combined cardiovascular endpoint (Table 3). This increased risk was different among the patients with polyvascular disease as patients with more than two diseased vascular territories had an even higher risk than patients with two atherosclerotic vascular territories. The HRs (with 95% confidence intervals) for patients with two diseased vascular territories compared to patients with monovascular disease were for all-cause mortality 1·60 (1·14–2·25), cardiovascular mortality 2·13 (1·29–3·50) and the combined cardiovascular endpoint 1·66 (1·20–2·31). Moreover, the HRs (95% confidence intervals) for patients with more than two diseased vascular territories compared to patients with monovascular disease were for all-cause mortality 3·81 (2·45–5·92), cardiovascular mortality 4·40 (2·32–8·35) and the combined cardiovascular endpoint 2·75 (1·69–4·47). No large differences were found between the two different statistical models. Event-free survival curves, derived from the second Cox proportional hazard model, are shown for all-cause mortality and the combined cardiovascular outcome (Figs 1–2).

Figure 1.

All-cause mortality in percutaneous coronary intervention (PCI) patients, subdivided in subgroups of number of diseased atherosclerotic territories (derived from an age-, gender-, diabetes mellitus-, hypertension-, packyears (in quartiles), previous myocardial infarction-, body mass index-, hyperlipidaemia-adjusted Cox proportional hazards model. The X-axis shows follow-up duration in days. The Y-axis shows the event rate.

Figure 2.

The combined cardiovascular endpoint (myocardial infarction, stroke, cardiovascular mortality) in percutaneous coronary intervention (PCI) patients, subdivided in subgroups by the number of diseased atherosclerotic territories (derived from an age-, gender-, diabetes mellitus-, hypertension-, packyears (in quartiles), previous myocardial infarction-, body mass index-, hyperlipidemia-adjusted Cox proportional hazards model. The X-axis shows follow-up duration in days. The Y-axis shows the event rate.

After adjusting for the extent of CAD in the subgroup of patients in whom the extent of CAD was known (model III), the relation between the number of territories and the outcomes was slightly attenuated but evidently still present.

Subclinical polyvascular disease

We repeated the survival analysis after subdividing the polyvascular patients based on whether the concomitant atherosclerotic disease was previously diagnosed (clinical) or detected at the SMART screening (subclinical) (Table 4). A total of 226 patients (49%) of the 462 patients with polyvascular disease had clinical atherosclerotic disease. The remaining 236 patients (51%) were identified with subclinical atherosclerotic disease. Patients with clinical or subclinical polyvascular disease had an impaired long-term prognosis as compared with monovascular disease. The hazard ratios for different outcomes in patients with clinical and subclinical polyvascular disease compared to patients with monovascular disease were as follows: all-cause mortality: 2·05 (1·40–3·01) and 1·94 (1·34–2·81); cardiovascular mortality 2·28 (1·29–4·06) and 2·85 (1·68–4·83) and for the combined cardiovascular endpoint: 1·51 (1·00–2·28) and 2·22 (1·55–3·17).

Table 4. The risk of subclinical and clinical polyvascular disease on long-term mortality, cardiovascular mortality and a combined cardiovascular endpoint
 One territory n = 1837 Reference categoryClinical disease More than one territory n = 226 HR (95%CI)Subclinical disease More than one territory n = 236 HR (95%CI)
  1. HR, Hazard ratio; CI, confidence interval. Model I, hazard ratios (HRs) adjusted for age and gender; Model III, HRs adjusted for age, gender, diabetes mellitus, hypertension, hyperlipidaemia, body mass index, packyears (In quartiles) and previous myocardial infarction. Given the relatively small numbers of clinical and subclinical polyvascular disease, there was no room for subanalysis of two or more diseased vascular territories.

All-cause mortality [no. of events (%)]137 (7)35 (15)39 (17)
Model I1·002·04 (1·40–2·97)1·91 (1·33–2·75)
Model II1·002·05 (1·40–3·01)1·94 (1·34–2·81)
Cardiovascular mortality [no. of events (%)]57 (3)15 (7)20 (8)
Model I1·002·19 (1·23–3·90)2·44 (1·45–4·11)
Model II1·002·28 (1·29–4·06)2·85 (1·68–4·83)
Combined cardiovascular endpoint [no. of events (%)]173 (9)28 (12)42 (18)
Model I1·001·52 (1·01–2·27)2·09 (1·48–2·95)

Model II

(myocardial infarction, stroke, cardiovascular death)

1·001·51 (1·00–2·28)2·22 (1·55–3·17)

Discussion

In this study, we demonstrate that the presence of polyvascular disease in patients undergoing PCI is associated with an unfavourable long-term outcome. Impaired prognosis is present for both clinical and subclinical atherosclerosis. Moreover, the outcome is highly influenced by the number of diseased vascular territories.

Several explanations are possible for the impaired prognosis of patients with polyvascular disease. First, it is certain that the high atherosclerotic burden plays an important role. It is established that a high atherosclerotic burden in just one vascular bed (CAD) is associated with a worse 1-year mortality [9-11]. The same holds for patients with polyvascular disease as found by previous studies [12-14]. Secondly, there is evidence that several plasma biomarkers such as fibrinogen and CRP are higher in CAD patients with concomitant PAD compared to patients with exclusive CAD [15]. It has been hypothesized that these procoagulant and proinflammatory states are related to a worse outcome in terms of cardiac death and nonfatal cardiac events [16, 17]. We also found a statistically significant higher CRP in our cohort of patients with polyvascular disease as compared to those with monovascular disease. Moreover, among the polyvascular patients, CRP was significantly higher in patients with more than two diseased vascular territories compared to patients with two diseased territories. The question remains whether there is a causal relation between CRP levels and outcome or that CRP levels rose secondary to the severity of the atherosclerotic process involved [18]. Nevertheless, intensive treatment of inflammation and cholesterol lowering may be beneficial to stabilize the atherosclerotic process [19, 20]. Finally, several studies showed that patients with polyvascular disease are treated suboptimal [2, 15, 21, 22]. For example, beta-blockers and statins were prescribed less often in patients with polyvascular disease than in patients with monovascular disease [15, 23]. Potentially, the outcome of polyvascular patients might be improved if treated according to current guidelines. Although not proven in polyvascular patients, a more aggressive statin therapy might be more beneficial because of its atherosclerotic disease stabilization/regression properties [19, 24]. In contrast, in the GRACE registry and in our study population, there were no differences in the use of evidence-based medication between patients with or without polyvascular disease [13]. In our study, patients with more than two diseased vascular territories did use less platelet aggregation inhibitors compared to the other two groups. However, after combining platelet aggregation inhibitors and oral coagulants, no difference was found between the three groups.

We are the first to assess the number of diseased vascular territories, clinical and subclinical, in a comprehensive way. We demonstrate that not only the presence of clinical but also subclinical polyvascular disease is associated with a worse prognosis. Our findings are in line with three small studies in which the presence of subclinical concomitant disease is associated with a worse prognosis compared to patients with monovascular disease (CAD) [5, 25, 26]. However, two studies only describe subclinical PAD and the other one PAD and carotid atherosclerotic lesions. We confirm these findings in a large cohort of patients in which subclinical disease was determined systematically in all vascular territories.

The presence of subclinical disease was determined by a standardized cardiovascular screening comprising noninvasive ultrasound imaging and routine laboratory tests. Half of the patients with polyvascular disease were identified with this screening protocol. Of note, our standardized vascular screening protocol carries no safety issues compared to novel screening techniques such as radiation-based calcium scoring of vascular disease. The prognosis of patients with polyvascular disease is associated with the number of diseased vascular territories. According to our protocol, a thorough enquiry of the medical history in combination with a simple standardized screening is helpful to reclassify the risk of patients undergoing PCI [26-32]. Reclassification may be important for optimization of personalized secondary prevention strategies.

According to the SMART study protocol, all participating patients received patient-tailored secondary prevention and therapy recommendations from the multidisciplinary team to improve patients' prognosis. These recommendations include guideline-based lifestyle management, improvement of medical treatment (e.g. antihypertensive drugs) and – if indicated – invasive treatment. However, as all patients, with and without polyvascular disease, receive the same personalized advice based on general risk factors, we did not expect any influence on the results.

The strengths of this study include the prospective cohort design, large sample size, long follow-up duration and the thorough assessment of the clinical endpoints. In addition, all patients received extensive screening of subclinical atherosclerotic disease leading to a more accurate identification of patients with polyvascular disease.

A limitation of our study is that compliance of patients to the multidisciplinary therapy advice and medication use is unknown. Furthermore, data on left ventricle function were not available in most patients and therefore could not be included in the current analysis. Moreover, we acknowledge that there have been innovations in coronary stents during the relatively long time frame of inclusion of patients. The emerging use of drug-eluting stents, for example, may have influenced the outcome of the PCI procedures in general. However, as patients with and without polyvascular disease were equally presented during all timeframes, we expect no influence on the findings.

Conclusions

In patients undergoing PCI, the presence of subclinical and clinical polyvascular disease is associated with an increased long-term mortality and morbidity. Moreover, the outcome is highly influenced by the number of diseased vascular territories.

Acknowledgements

We gratefully acknowledge the contribution of the SMART research nurses; Rutger van Petersen (data manager); Harry Pijl (vascular manager); and the Secondary Manifestations of ARTerial disease (SMART) Study Investigators: Ale Algra, MD, PhD, Julius Center for Health Sciences and Primary Care and Rudolf Magnus Institute for Neurosciences, Department of Neurology; Pieter A. Doevendans, MD, PhD, Department of Cardiology; Yolanda van der Graaf, MD, PhD, Diederick E.Grobbee, MD, PhD, and Guy E. H. M. Rutten, MD, PhD, Julius Center for Health Sciences and Primary Care; L.Jaap Kappelle, MD, PhD, Department of Neurology; Willem P. Th. M. Mali, MD, PhD, Department of Radiology; Frans L. Moll, MD, PhD, Department of Vascular Surgery; and Frank L. J. Visseren, MD, PhD, Department of Vascular Medicine, University Medical Center Utrecht.

Contributions

MGM design/analyses/writing paper; MJC design/contribution writing; YG design/contribution analyses/contribution writing; YA design/contribution writing; PAD design/contribution writing; HMN design/contribution analyses/contribution writing.

Address

Department of Cardiology, University Medical Center, Room E 03.511, POB 85500, 3508 GA, Utrecht, the Netherlands (M. G. van der Meer, M. J. Cramer, P. A. Doevendans, H. M. Nathoe); Department of Clinical Epidemiology, University Medical Center, IHP Str. 6.131, P.O. Box 85500, 3508 GA, Utrecht, the Netherlands (Y. van der Graaf); Department of Invasive Cardiology, De Boelelaan 1117, 1081 HV, Amsterdam, the Netherlands (Y. Appelman).

Ancillary