The Scientific World Journal

The Scientific World Journal / 2012 / Article

Research Article | Open Access

Volume 2012 |Article ID 284851 |

Hassan Al Thani, Ayman El-Menyar, Khalid F. AlHabib, Ahmed Al-Motarreb, Ahmad Hersi, Hussam AlFaleh, Nidal Asaad, Shukri Al Saif, Wael Almahmeed, Kadhim Sulaiman, Haitham Amin, Alawi A. Alsheikh-Ali, Khalid AlNemer, Jassim Al Suwaidi, "Polyvascular Disease in Patients Presenting with Acute Coronary Syndrome: Its Predictors and Outcomes", The Scientific World Journal, vol. 2012, Article ID 284851, 7 pages, 2012.

Polyvascular Disease in Patients Presenting with Acute Coronary Syndrome: Its Predictors and Outcomes

Academic Editor: A. J. Kirtane
Received09 Oct 2011
Accepted02 Nov 2011
Published04 Jan 2012


We evaluated prevalence and clinical outcome of polyvascular disease (PolyVD) in patients presenting with acute coronary syndrome (ACS). Data for 7689 consecutive ACS patients were collected from the 2nd Gulf Registry of Acute Coronary Events between October 2008 and June 2009. Patients were divided into 2 groups (ACS with versus without PolyVD). All-cause mortality was assessed at 1 and 12 months. Patients with PolyVD were older and more likely to have cardiovascular risk factors. On presentation, those patients were more likely to have atypical angina, high resting heart rate, high Killip class, and GRACE risk scoring. They were less likely to receive evidence-based therapies. Diabetes mellitus, renal failure, and hypertension were independent predictors for presence of PolyVD. PolyVD was associated with worse in-hospital outcomes (except for major bleedings) and all-cause mortality even after adjusting for baseline covariates. Great efforts should be directed toward primary and secondary preventive measures.

1. Introduction

Atherothrombosis is a systemic disease that often occurs at more than one vascular site and should be considered in practice as an integral disease [13]. Moreover, patients with affected arterial disease are more likely to develop higher event rates than patients with multiple risk factors only [3, 4]. Polyvascular disease (PolyVD) defined as presence of more than one affected vascular bed, that is, any combination of the following: coronary artery disease (CAD), peripheral arterial disease (PAD), and cerebrovascular disease (CVD) [59]. The frequency and impact of PolyVD in patients with acute coronary syndrome (ACS) in the Gulf region of the Middle East have not been studied yet. These countries have higher prevalence of the traditional risk factors in a unique fashion [1]. The aim of the current study is to evaluate the prevalence of PolyVD and its impact on the in-hospital major adverse events and 1-year mortality across ACS in a Middle Eastern population.

2. Methods

2.1. Study Population

Data were collected from a prospective, multicenter study of the 2nd Gulf Registry of Acute Coronary Events (Gulf RACE-2) between October 2008 and June 2009. We recruited 7,930 consecutive patients with ACS from 6 adjacent Middle Eastern Gulf countries (Bahrain, Kingdom of Saudi Arabia, Qatar, Oman, United Arab Emirates, and Yemen). Patients diagnosed with ACS, including unstable angina (UA) and non-ST- and ST-elevation myocardial infarction (NSTEMI and STEMI, resp.), were enrolled from 65 hospitals. On-site cardiac catheterization laboratory was available in 43% of the participating hospitals. All prospective patients with ACS were eligible for enrollment. The study received ethics approval from the institutional review boards in all participating hospitals. Full details of the methods have been previously published [10].

2.2. Definitions

ACS: diagnosis of the different types of ACS and definitions of data variables were based on the American College of Cardiology clinical data standards [10, 11]. Peripheral arterial disease (PAD): defined as presence of any of the following: intermittent claudication, critical limb ischemia (ulcer or gangrene), peripheral bypass surgery (surgical bypass for PAD indication), or peripheral percutaneous transluminal angioplasty. Polyvascular disease (PolyVD) was defined as presence of more than one affected vascular bed, that is, CAD, cerebrovascular disease (CVD), and asymptomatic or symptomatic peripheral arteries (PAD) [6]. A case report form (CRF) for each patient with suspected ACS was filled out upon hospital admission by assigned physicians and/or research assistants using standard definitions and was completed throughout the patient’s hospital stay. All CRFs were verified by a cardiologist then sent online to the principal coordinating center, where the forms were further checked for mistakes before submission for final analysis.

2.3. Statistical Analysis

Data are presented as proportions or mean ± standard deviation (SD) as appropriate. Baseline demographic characteristics, past medical history, clinical presentation, and clinical outcomes were compared between 2 groups (ACS with versus without PolyVD). Subanalysis was performed comparing the clinical outcomes among various combinations of vascular bed affection (ACS alone, ACS plus PAD, ACS plus CVD, and ACS plus PAD and CVD). Statistical analyses were conducted using the Student’s t-test for continuous variables and Pearson chi-square ( 𝜒 2 ) test for categorical variables. Primary endpoints included recurrent ischemia, heart failure (HF), stroke, major bleedings, and mortality. All-cause mortality was assessed at 1-month and 12-month follow-up period. In order to assess the independent association of PolyVD with clinical outcomes, logistic regression analysis models were used and crude and adjusted odd ratios were calculated. We included significant baseline variables in the analysis such as age, sex, diabetes mellitus, hypertension, dyslipidemia, smoking, renal failure, prior coronary artery disease and prior coronary revascularization in addition to the type of ACS at presentation. Univariate and multivariate analysis for clinical outcomes were tested. Global Registry of ACS events (GRACE) risk scores for hospital mortality were used to stratify the risk status of patients at presentation as low, intermediate, or high [12]. All P values were the results of 2-tailed tests and values <0.05 were considered significant. Data analysis was carried out using the Statistical Package for Social Sciences version 18 (SPSS Inc., USA).

3. Results

3.1. Clinical and Biochemical Profiles

Out of the 7689 patients who were admitted with ACS, PolyVD was documented in 428 patients (5.6%) in terms of ACS plus PAD (110 patients), ACS plus CVD (284 patients), and ACS plus PAD and CVD (34 patients). Table 1 shows the baseline characteristics and risk factors of patients with PolyVD including ACS in comparison to ACS-alone patients. Patients with PolyVD were 6 years older ( 6 2 ± 1 1 versus 5 6 ± 1 2 , 𝑃 < 0 . 0 0 0 1 ) and were more likely to be female (31% versus 20%, 𝑃 = 0 . 0 0 1 ). PolyVD patients were more likely to have risk factors such as diabetes mellitus (67% versus 38%, 𝑃 = 0 . 0 0 1 ), hypertension (78% versus 45%, 𝑃 = 0 . 0 0 1 ), dyslipidemia (55% versus 36%, 𝑃 = 0 . 0 0 1 ), and renal failure (14% versus 3%, 𝑃 = 0 . 0 0 1 ). They were less likely to be smokers (45% versus 54%, 𝑃 = 0 . 0 0 1 ). On presentation with ACS, PolyVD patients had higher heart rate, and GRACE risk score and were more likely to present with a higher Killip class ( 𝑃 = 0 . 0 0 1 for all). NSTEACS was the most frequent diagnosis in PolyVD patients whereas STEMI was the predominant diagnosis in ACS-alone group.

ACS alone ( 𝑛 = 7 2 6 1 )Polyvascular disease ( 𝑛 = 4 2 8 )P value

Age 5 6 ± 1 2 6 2 ± 1 2 0.001
Diabetes mellitus38670.001
Renal failure3140.001
Prior coronary artery disease39640.001
Prior revascularization11250.001
Khat chewing19200.92

Clinical presentation
Atypical presentation ACS15280.001
Heart rate 8 4 ± 2 0 9 0 ± 2 2 0.001
Systolic blood pressure 1 3 5 ± 2 9 1 4 0 ± 3 5 0.01
Diastolic blood pressure 8 1 ± 1 7 7 9 ± 2 0 0.02
Unstable angina2424
Killip class > 121450.001
Low GRACE risk score4119 0.001
High GRACE risk score2043

Laboratory findings
Fasting blood sugar (mmol/L) 7 . 3 ± 3 8 . 1 ± 3 0.001
First hemoglobin (g/dL) 1 3 . 7 ± 2 1 2 . 5 ± 2 0.001
Serum creatinine (umol/L) 1 0 0 ± 7 2 1 3 0 ± 1 1 0 0.001
Hemoglobin A1c (%) 7 . 6 ± 2 . 4 8 ± 2 . 4 0.15
Total cholesterol (mmol/L) 4 . 9 ± 1 . 6 4 . 4 ± 1 . 6 0.001
Serum triglyceride (mmol/L) 1 . 8 ± 1 . 1 1 . 6 ± 0 . 8 0.001
Low-density lipoprotein 3 . 2 ± 1 . 3 2 . 8 ± 1 . 2 0.001
High-density lipoprotein 1 . 0 5 ± 0 . 5 0 . 9 9 ± 0 . 4 0.01
Peak troponin T 1 . 4 ± 0 . 7 1 . 3 ± 0 . 6 0.02
LV ejection fraction <50%73850.001
1-vessel CAD29480.001
2-vessel CAD25180.12
3-vessel CAD30170.004

All categorical and continuous variables are given in percent and mean ± SD, respectively.
3.2. In-Hospital Treatment Pattern

Table 2 demonstrates the treatment pattern for patients with and without PolyVD. In regard to on-admission therapy, there were no differences between the two groups in the use of heparin and glycoprotein inhibitors. Apart from thrombolysis therapy use that was in favor of ACS-alone group, patients with PolyVD were more likely to be treated with oral antiplatelet, angiotensin-converting enzyme inhibitors, statin, and β-blockers. Coronary intervention was more frequently used in ACS-alone group in comparison to PolyVD patients. Aspirin, statin, and β-blockers were more frequently used during hospitalization and at discharge in ACS-alone patients.

ACS alone ( 𝑛 = 7 2 6 1 )Polyvascular disease ( 𝑛 = 4 2 8 )P value

On admission therapy %
𝛽 -blockers2843.50.001
ACE inhibitors24460.001
Unfractionated heparin42400.46
LMW heparin37410.12
Glycoprotein inhibitors7.780.82

During hospitalization %
β blockers75680.002
ACE inhibitors71690.31
Coronary angiography39310.002

Discharge medications %
β blockers80740.01
ACE inhibitors7263.50.001

Outcomes %
Heart failure11.629.70.001
Hospital death4.29.10.001
1-month death7.416.30.001
12-month death11.224.60.001

*ST-elevation MI, PCI: percutaneous coronary intervention.
3.3. Predictors for Polyvascular Disease

Multivariate logistic regression analysis showed that the important independent predictors for the presence of PolyVD in ACS patients were diabetes mellitus (adjusted OR 2.28; 95% CI 1.81–2.89), renal failure (adjusted OR 2.32; 95% CI 1.54–3.03) and hypertension (adjusted OR 2.66; 2.03–3.49), 𝑃 = 0 . 0 0 1 for all (Table 3).

Odds ratio (95% confidence interval)P value

Age1.04 (1.036–1.056)0.001
Male gender1.13 (0.865–1.466)0.34
Dyslipidemia1.10 (0.882–1.392)0.37
Diabetes mellitus2.28 (1.809–2.889)0.001
Hypertension2.66 (2.028–3.489)0.001
Smoking1.24 (0.970–1.591)0.07
Obesity1.003 (0.986–1.020)0.62
Renal failure2.32 (1.543–3.029)0.001

3.4. Clinical Outcomes

Figure 1 demonstrates the clinical outcomes in different vascular disease combinations. Table 2 shows hospital outcomes in PolyVD versus ACS-alone group. All primary endpoints were significantly worse in PolyVD except for major bleedings. One- and 12-month all-cause mortality was 2-times greater in PolyVD group. Univariate analysis showed that PolyVD was predictor for reischemia, HF, stroke, and hospital, 1, and 12-month mortality ( 𝑃 = 0 . 0 0 1 for all). Similarly, multivariate logistic regression analysis denoted that PolyVD was an independent predictor for all outcomes except for major bleeding. PolyVD was strong predictor for in-hospital stroke (adjusted OR 5.40; 95% CI 2.17–13.29, 𝑃 = 0 . 0 0 1 Table 4 and Figure 2).

Unadjusted odds ratio (95% CI)Adjusted odds ratio* (95% CI)

Re-ischemia1.7 (1.37–2.15), 𝑃 = 0 . 0 0 1 1.50 (1.12–2.01), 𝑃 = 0 . 0 0 7
Heart failure3.3 (2.67–4.09), 𝑃 = 0 . 0 0 1 2.00 (1.47–2.61), 𝑃 = 0 . 0 0 1
In-hospital Stroke4.1 (2.03–8.22), 𝑃 = 0 . 0 0 1 5.40 (2.17–13.29), 𝑃 = 0 . 0 0 1
Major bleedings1.61 (0.57–4.52), 𝑃 = 0 . 3 6 1.68 (0.54–5.17), 𝑃 = 0 . 3 7
Hospital death2.3 (1.66–3.26), 𝑃 = 0 . 0 0 1 1.85 (1.25–2.74), 𝑃 = 0 . 0 1
1-month death2.5 (1.89–3.29), 𝑃 = 0 . 0 0 1 2.03 (1.46–2.83), 𝑃 = 0 . 0 0 1
12-month death2.6 (2.02–3.32), 𝑃 = 0 . 0 0 1 1.83 (1.83–2.45), 𝑃 = 0 . 0 0 3

CI: confidence interval, *variables adjusted for age, sex, diabetes mellitus, hypertension, dyslipidemia, smoking, renal failure, prior coronary artery disease, and prior coronary revascularization in addition to the type of ACS at presentation.
3.5. Gender and Polyvascular Disease

In comparison to their counterparts with ACS alone, men and women with PolyVD were older and had higher percent of commodities except for smoking. When compared to men, women with ACS alone had significant worse outcomes in terms of re-ischemia (18% versus 14%), in-hospital mortality (6.3% versus 3.7%), 1-month mortality (9.7% versus 6.8%), 𝑃 = 0 . 0 0 1 for all. In all forms of PolyVD, there were no significant morbidity and mortality differences between men and women (Table 5).

ACS alonePolyvascular disease
MenWomenP valueMenWomenP value

Age (yrs, mean) 5 5 ± 1 2 6 1 ± 1 2 0.001 6 4 ± 1 2 6 5 ± 1 3 0.19
Diabetes mellitus %35510.00163740.03
Hypertension %40.5640.00176810.24
Dyslipidemia %34440.00153630.55
Smoking %65100.00157160.001
Renal failure %350.00116130.44
Admission therapy (%)
 Beta blockers76710.00170630.11
 ACE inhibitors71720.7270650.31
 Glycoprotein inhibitors8.35.20.001960.23
Outcomes (%)
 Heart failure10.715.50.00129340.28
 Hospital mortality3.76.30.0018.610.80.46
 1-month mortality6.89.70.00115200.26
 12-month mortality10.3150.00123290.22

Table 6 shows risk factors and in-hospital mortality in patients with versus without polyvascular disease presenting with acute coronary syndrome in 3 major clinical studies.

GRACE (n = 32735) [7]MASCARA (n = 6745) [9]GULFRACE-2 ( 𝑛 = 7 6 8 9 ) ALLIANCE (n = 8904) [6]
APolyVD (15.6%)APolyVD (16.6%)APolyVD (5.6%)APolyVD (13%)

Patients %847628396294140.687841
Age (mean yrs)6471737367707372566365666572
Diabetes mellitus2238344228494352387761821934
Renal failureN/AN/AN/AN/AN/AN/AN/AN/A3201030312


A: acute coronary syndrome (ACS) alone; B: ACS plus peripheral arterial disease (PAD); C: ACS plus cerebrovascular disease (CVD); D: ACS plus PAD plus CVD; PolyVD: polyvascular disease, all categorical variables represents in percentage (%); DM: diabetes mellitus; N/A: data not available; *prior/current smoking. Data were collected from [6, 7, 9].

4. Discussion

The present study reports on the frequency, predictors and implication of PolyVD in patients presenting with ACS in the Gulf Region of the Middle East. Up to our knowledge, PolyVD in ACS patients was not reported in the Middle East before. There are several key findings in this study. First, PolyVD is a common disorder in that region in the setting of acute coronary events and represents a marker of high-risk population. The prevalence of PolyVD in the current study (5.6%) is over 2-fold lower versus GRACE study (15.6%), Alliance project (13%), and MASCARA study (16.6% Table 6). Among the 4 studies, the high percent of young age, DM, and renal failure were observed in the current study. However, the difference in presentations and outcomes in the 4 studies may in part relate to diversity of biological and environmental factors between the different ethnicities. This observation needs further confirmatory studies. Second, PolyVD is associated with higher mortality rate even after adjusting for baseline variables. Third, at presentation, those patients were more likely to have atypical presentation, high resting heart rate, high Killip class, and high GRACE risk score. Moreover, NSTEACS was the most frequent diagnosis in PolyVD in comparison to ACS-alone group. Our findings are consistent with data from Western reports [59]. These studies showed that the management of PolyVD group was less aggressive in terms of less in-hospital coronary intervention and less evidence-based therapies during hospitalization and at discharge [68]. Several factors could explain the undertreatment trend in PolyVD group such as high percent of comorbidities with subsequent contraindication of some medications, physician and patient discretion, and possible socioeconomic factors. Understanding such reasons may reduce this therapy imbalance. Fourth, the current study shows that increasing the number of the affected vascular bed is associated with the worst clinical scenario (Figure 1). Apart from major bleedings, PolyVD is independent predictor for all in-hospital adverse outcomes. PolyVD increased risk of stroke 5-times even after adjustment for other covariates. PAD increased rate of HF and hospital mortality by 3-times whereas presence of CVD increased rate of in-hospital stroke 4-times. Data from Table 5 indicates that the highest mortality rate was observed in patients with PAD in our study, in patients with CVD in ALLIANCE project, and in patients with triple vascular bed affection in MASCARA and our study. Data from the Japanese REACH registry [13] denoted that the rates of 1-year stroke and myocardial infarction were higher for patients with CVD and PAD than for patients with CVD and CAD; however, that study was not carried out during the acute setting of coronary artery disease and did not report on the in-hospital rate of stroke. Fifth, the present study expands the previous reports and shows that diabetes mellitus, hypertension, and renal failures are independent predictors for the presence of PolyVD. This finding highlights the importance of primary and secondary prevention in this high-risk population. Sixth, male gender has better outcomes in ACS alone in comparison to female. These unfavorable outcomes associated with female gender were shown in our previous work as well [14]. Interestingly, presence of PolyVD is not associated with significant differences in the outcomes between men and women, this finding warrants further exploration.

5. Limitation

The current study is an observational study; however, well-designed observational studies may provide valid results. Another limitation of the current retrospective analysis is that the diagnosis of PAD relied on the clinical history and not on the measurement of ankle-brachial index (ABI). This limitation could be explained in part by facts that clinical variables used are incompletely sensitive to identify PAD, approximately half of subjects with ABI <0.90 are asymptomatic, and of those that are symptomatic, only a minority have classic intermittent claudication. The presence of PolyVD is underdiagnosed in our daily practice that may underestimate its true prevalence and impact on the outcome.

6. Conclusion

Although PolyVD’s patients are high-risk population in the setting of ACS, they received less aggressive therapy. Apart from major bleedings, PolyVD is an independent predictor for adverse hospital outcomes and short- and long-term mortality. Great efforts should be directed to primary and secondary prevention. Looking for the other affected vascular bed in ACS patients will add an important step in risk stratification and management.

Conflict of Interests

The authors have no conflict of interests.


Gulf RACE is a Gulf Heart Association (GHA) project and was financially supported by the GHA, Sanofi Aventis, and the College of Medicine Research Center at King Khalid University Hospital, King Saud University, Riyadh, Saudi Arabia. The sponsors had no role in study design, data collection, data analysis, writing of the report, or submission of the paper. Ethical approvals were obtained prior to the study.


The authors thank the staff in all the participating centers for their invaluable cooperation. Special thanks to Zenaida Ramoso and Kazi Nur Asfina for data coordination and secretarial assistance, and also to the physicians who participated in this registry.


  1. A. El-Menyar, H. Amin, I. Rashdan et al., “Ankle-brachial index and extent of atherosclerosis in patients from the middle east (the AGATHA-ME study): a cross-sectional multicenter study,” Angiology, vol. 60, no. 3, pp. 329–334, 2009. View at: Publisher Site | Google Scholar
  2. F. G. R. Fowkes, L. P. Low, S. Tuta, and J. Kozak, “Ankle-brachial index and extent of atherothrombosis in 8891 patients with or at risk of vascular disease: results of the international AGATHA study,” European Heart Journal, vol. 27, no. 15, pp. 1861–1867, 2006. View at: Publisher Site | Google Scholar
  3. G. Steg, D. L. Bhatt, P. W.F. Wilson et al., “One-year cardiovascular event rates in outpatients with atherothrombosis,” Journal of the American Medical Association, vol. 297, no. 11, pp. 1197–1206, 2007. View at: Publisher Site | Google Scholar
  4. P. P. Cacoub, U. Zeymer, T. Limbourg et al., “Effects of adherence to guidelines for the control of major cardiovascular risk factors on outcomes in the REduction of Atherothrombosis for Continued Health (REACH) Registry Europe,” Heart, vol. 97, no. 8, pp. 660–667, 2011. View at: Publisher Site | Google Scholar
  5. D. L. Bhatt, E. D. Peterson, R. A. Harrington et al., “Prior polyvascular disease: risk factor for adverse ischaemic outcomes in acute coronary syndromes,” European Heart Journal, vol. 30, no. 10, pp. 1195–1202, 2009. View at: Publisher Site | Google Scholar
  6. A. Meizels, D. M. Zeitoun, V. Bataille et al., “Impact of polyvascular disease on baseline characteristics, management and mortality in acute myocardial infarction. The Alliance project,” Archives of Cardiovascular Diseases, vol. 103, no. 4, pp. 207–214, 2010. View at: Publisher Site | Google Scholar
  7. D. Mukherjee, K. A. Eagle, E. Kline-Rogers et al., “Impact of prior peripheral arterial disease and stroke on outcomes of acute coronary syndromes and effect of evidence-based therapies (from the Global Registry of Acute Coronary Events),” American Journal of Cardiology, vol. 100, no. 1, pp. 1–6, 2007. View at: Publisher Site | Google Scholar
  8. G. Cotter, C. P. Cannon, C. H. McCabe et al., “Prior peripheral arterial disease and cerebrovascular disease are independent predictors of adverse outcome in patients with acute coronary syndromes: are we doing enough? Results from the Orbofiban in Patients with Unstable Coronary Syndromes-Thrombolysis in Myocardial Infarction (OPUS-TIMI) 16 study,” American Heart Journal, vol. 145, no. 4, pp. 622–627, 2003. View at: Publisher Site | Google Scholar
  9. I. Ferreira-González, G. P. Miralda, M. Heras et al., “Investigadores del Estudio MASCARA. Prognosis and management of patients with acute coronary syndrome and polyvascular disease,” Revista Espanola de Cardiologia, vol. 62, no. 9, pp. 1012–1021, 2009. View at: Publisher Site | Google Scholar
  10. K. F. AlHabib, A. Hersi, A. A. Alsheikh-Ali et al., “Prevalence, predictors, and outcomes of conservative medical management in non-ST-segment elevation acute coronary syndromes in Gulf RACE-2,” Angiology. In press. View at: Google Scholar
  11. A. El-Menyar, M. Zubaid, K. Sulaiman et al., “In-hospital major clinical outcomes in patients with chronic renal insufficiency presenting with acute coronary syndrome: data from a registry of 8176 patients,” Mayo Clinic Proceedings, vol. 85, no. 4, pp. 332–340, 2010. View at: Publisher Site | Google Scholar
  12. C. B. Granger, R. J. Goldberg, O. Dabbous et al., “Predictors of hospital mortality in the Global Registry of Acute Coronary Events,” Archives of Internal Medicine, vol. 163, no. 19, pp. 2345–2353, 2003. View at: Publisher Site | Google Scholar
  13. S. Uchiyama, S. Goto, M. Matsumoto et al., “Cardiovascular event rates in patients with cerebrovascular disease and atherothrombosis at other vascular locations: results from 1-year outcomes in the Japanese REACH Registry,” Journal of the Neurological Sciences, vol. 287, no. 1-2, pp. 45–51, 2009. View at: Publisher Site | Google Scholar
  14. A. El-Menyar, M. Zubaid, W. Rashed et al., “Comparison of men and women with acute coronary syndrome in six Middle Eastern countries,” American Journal of Cardiology, vol. 104, no. 8, pp. 1018–1022, 2009. View at: Publisher Site | Google Scholar

Copyright © 2012 Hassan Al Thani et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

More related articles

 PDF Download Citation Citation
 Download other formatsMore
 Order printed copiesOrder

Related articles

Article of the Year Award: Outstanding research contributions of 2020, as selected by our Chief Editors. Read the winning articles.