INTRODUCTION
Atrial fibrillation (AF) is the most common arrhythmia in the elderly with evidence of an increasing prevalence and incidence worldwide.1–3 AF is associated with an increased risk of ischemic stroke, incident myocardial infarction, all-cause mortality, and heart failure.4
Epicardial adipose tissue (EAT) represents a real and unique visceral fat deposit of the heart as indicated by the size of its adipocytes, metabolic activity, and biochemical composition. EAT has significantly higher rates of lipolysis and lipogenesis than visceral fat depots of other sites. Because no definite barrier exists between the EAT and the contiguous myocardium, EAT influences the myocardium and adjacent coronary arteries via a paracrine or vasocrine mechanism. Increased EAT thickness is strongly associated with diabetes mellitus, cardiovascular (CV) disease, visceral obesity, subclinical atherosclerosis at multiple locations, and the metabolic syndrome.5–9 EAT is also increased in subjects with AF10 and is associated with AF severity and the recurrence of AF after catheter ablation.11,12 In non-AF patients, increased EAT thickness is shown to be positively associated with the severity of coronary artery disease and left ventricular diastolic dysfunction5,13 and is a useful parameter in predicting adverse CV events.14,15 However, there was no study to analyze the ability of EAT thickness in prediction of adverse CV events in AF patients. Hence, this study was conducted to evaluate whether EAT thickness was a useful parameter in prediction of adverse CV events in patients with AF. Besides, we also evaluated the major correlates of EAT thickness in these patients.
METHODS
Study Patients
From April 2010 to June 2012, this prospective observational study consecutively enrolled patients diagnosed with persistent AF and referred for transthoracic echocardiography. The definition of persistent AF was sustained arrhythmia beyond 7 days, which was confirmed by 12-lead eletrocardiography (ECG), ECG recording during transthoracic echocardiography, or 24-hour Holter ECG. Our strategy in treating patients with persistent AF was rate control. A total of 225 patients were initially enrolled but those with inadequate echocardiographic visualization (n = 11) were excluded first. Then patients with moderate-to-severe valvular heart disease were also excluded (n = 14). Ten patients were lost to follow-up. Finally, 190 AF patients were enrolled for further evaluation. The protocol of this study was reviewed and approved by the Institutional Review Board and all enrolled patients signed informed consent forms.
Echocardiographic Evaluation
The transthoracic echocardiography was performed using VIVID 7 (General Electric Medical Systems, Horten, Norway) by one independent cardiologist that blind to original patient characteristics according to the standardized protocol, which was also described in our previous studies.16–19 The cardiologist was blinded to the clinical data. Patients stayed in the left decubitus position to acquire the anatomic M-mode and 2-dimensional echocardiographic images. The gain was minimized and the wall filter settings were well adapted to eliminate the high-frequency signal. From the apical 4-chamber view, the Doppler sample volume was positioned at the tips of the mitral leaflets to acquire the left ventricular inflow waveforms and at the septal and lateral corners of the mitral annulus to acquire the pulsed tissue Doppler imaging. Early diastolic mitral annulus velocity (E′) was averaged from septal and lateral ones. The modified Simpson method, the Devereux-modified method, and the biplane area-length method were used to calculate left ventricular ejection fraction (LVEF), left ventricular mass, and left atrial volume respectively.20,21 Left ventricular mass index (LVMI) was defined as left ventricular mass divided by body surface area. Left atrial volume index (LAVI) was defined as left atrial volume divided by body surface area.
Measurement of EAT Thickness
EAT thickness was measured on the free wall of right ventricle from the parasternal long-axis views by one independent cardiologist. EAT was identified as an echo-free space located between the visceral pericardium and the outer wall of myocardium on the 2-dimensional echocardiography, and EAT thickness was calculated perpendicularly on the free wall of the right ventricle at end-diastole for 3 cardiac cycles.8,22 The mean value from 3 cardiac cycles was calculated for the later analysis. The aortic annulus was used as anatomical reference of standardized measurement, which was performed at one location on the free wall of the right ventricle along the midline of the ultrasound beam, perpendicular to the aortic annulus. The raw data of echocardiography were stored for off-line analysis using the EchoPAC software (EchoPAC version 08; GE-Vingmed Ultrasound AS GE Medical Systems, Horten, Norway).
Left ventricular dimensions, LVMI, LVEF, and LAVI were measured using the index beat method.23–25 As their measurements were rapid and easy, the early mitral inflow velocity (E), E-wave deceleration time, and E′ were averaged from 3 beats. However, a heartbeat was avoided if the duration of a cardiac cycle was too short to finish the diastolic process. Therefore, the measurement of E, E-wave deceleration time, and E′ was not consecutive every time.
Collection of Clinical Data
Clinical data of AF patients comprising sex, age and history of diabetes mellitus, hypertension, coronary artery disease (CAD), chronic heart failure (CHF), stroke, and smoking were acquired from interviews with patients or medical records. The following definitions were also described in our previous studies.16,18 The readings of systolic and diastolic blood pressures were obtained using a mercury sphygmomanometer. The definition of hypertension was based on a systolic blood pressure >140 mmHg or a diastolic blood pressure >90 mmHg or antihypertensive drugs prescribed for blood pressure control. The definition of diabetes mellitus was based on antidiabetic agents prescribed to control blood glucose, a fasting blood glucose level >126 mg/dL, or a postprandial blood glucose level >200 mg/dL obtained 2 hours after meals. The definition of body mass index (BMI) was a person's weight in kilograms divided by the square of height in meters. The definition of stroke was based on a history of cerebrovascular accident including infarction or cerebral bleeding. The definition of CAD was based on any documented coronary angiography, any positive stress test in patients with a history of typical angina, a history of old myocardial infarction, or a history of percutaneous coronary intervention or bypass surgery. Heart failure was defined according to the Framingham criteria.26 Myocardial infarction was defined according to the universal definition.27 Estimated glomerular filtration rate (eGFR) was obtained by using the 4-item Modification of Diet in Renal Disease Study equation.28
The prescribed medications comprising antiplatelets, anticoagulants, angiotensin II receptor blockers, angiotensin-converting enzyme inhibitors, β-blockers, calcium channel blockers, and diuretics were surveyed form the formal medical records in the study period.
Definition of CV Events
In AF patients, CV events were defined as CV mortality, myocardial infarction, stroke, and hospitalization for heart failure, which resulted from dyspnea with concurrent intravenous diuretics and with pulmonary congestion confirmed radiographically. The diagnosis of stroke was confirmed by clinical assessment of a neurologist combined with computed tomographic or magnetic resonance imaging findings.
Two cardiologists were selected to determine and judge CV events. If any disagreement was encountered, a third cardiologist was assigned for final judgement after reviewing all clinical data and medical records. If patients had several CV events, only the first event was coded. If patients died after episodes of heart failure, myocardial infarction, or stroke during the same admission, they were coded as CV deaths. In patients reaching the study end points, they were followed until the first episode of adverse events, whereas the other patients received follow-up until March 2014.
Reproducibility
To investigate the interobserver variability of measurement of EAT thickness, 30 patients were randomly chosen by 2 independent observers. Then identical measurement of echocardiographic parameters was repeated 1 week later to investigate the intraobsever variability. The mean percent error was obtained by dividing the absolute difference by an average of above 2 observations.
Statistical Analysis of Clinical Data
We used the SPSS 18.0 software (SPSS Inc, Chicago, IL) for further statistical analysis. Clinical data were listed as mean ± standard deviation, percentage, or median values (25th–75th percentile) for the follow-up period. We compared categorical and continuous variables between 2 groups by using the χ2 test and independent samples t test, respectively, and investigated any relationship between 2 continuous variables by using the bivariate correlation method. Significant variables in the univariable analysis were chosen for further multivariable analysis. Cox proportional hazards model with forward selection was used to identify the predictors of adverse CV events. To define the best cut-off of EAT thickness, we calculated the Youden index (sensitivity + specificity − 1) by using the receiver-operating characteristic (ROC) curve. We also evaluated incremental model performance by calculating sequential change of the χ2 value and performed the Kaplan–Meier survival analysis of CV events. Besides, we performed stepwise multiple linear regression analysis to identify the correlates of EAT thickness. Finally, incremental model performance was also assessed by Integrated Discrimination Improvement (IDI) using SAS 9.3 (SAS Institute, Cary, NC). Statistical significance was recognized if P < 0.05 in 2-sided tests.
RESULTS
Our final study population included 190 persistent AF patients (128 male and 62 female) with a mean age of 70 ± 10 years. In ROC curve analysis, we found that 6.0 mm was the best cutoff value of EAT thickness in the prediction of CV events. Table 1 demonstrates the comparison of traditional clinical and echocardiographic characteristics according to EAT thickness (6.0 mm). We found that age, the prevalence of CAD, the presence of CHF, BMI, CHA2DS2-VASc score, eGFR, E-wave deceleration time, E′, the ratio of transmitral E-wave velocity to early diastolic mitral annulus velocity (E/E′), and EAT thickness were significantly different between patients with EAT thickness ≦6.0 mm and >6.0 mm.
;)
TABLE 1:
Comparison of Traditional Clinical and Echocardiographic Characteristics According to the Value of EAT Thickness (6.0 mm)
Table 2 demonstrates univariable and multivariable correlates of EAT thickness in AF patients. We found that older age, the existence of CAD, higher CHA2DS2-VASc score, lower eGFR, the prescription of calcium channel blockers, higher LVMI, longer E-wave deceleration time, lower E′, and higher E/E′ were correlated with higher EAT thickness in the univariable analysis. After performing the multivariable analysis, older age, the prescription of calcium channel blockers, and higher E/E′ were still correlated with higher EAT thickness.
TABLE 2:
Univariable and Multivariable Correlates of EAT Thickness in Study Patients
The follow-up duration of AF patients experiencing CV events was 29 (25th–75th percentile: 17–36) months. Sixty-nine CV events, including CV death (n = 19), hospitalization for heart failure (n = 32), myocardial infarction (n = 3), and stroke (n = 15) were recognized during the follow-up period. The total amount of person-years was 459 and incidence rate of CV events was 15.0% per year during follow-up. Table 3 shows predictors of CV events using the Cox proportional hazards regression analysis. We found that older age, the existence of CAD and CHF, non-smoking status, lower BMI, higher CHA2DS2-VASc score, lower eGFR, diuretic use, larger left ventricular end-systolic dimension, lower LVEF, higher LVMI, lower E′, higher E/E′, and higher EAT thickness were associated with adverse CV events in the univariable analysis. After performing the multivariable analysis, the existence of CHF, lower BMI and eGFR, and higher LVMI and EAT thickness (hazard ratio, 1.211; 95% confidence interval [CI], 1.084–1.351; P < 0.001) were still independent predictors of adverse CV events in AF patients.
TABLE 3:
Predictors of Cardiovascular Events Using Cox Proportional Hazards Model
Figure 1 demonstrates the Kaplan–Meier curves for CV event-free survival in study patients subdivided according to EAT thickness ≦6.0 mm and >6.0 mm (log-rank P < 0.001).
FIGURE 1:
Kaplan–Meier analysis of cardiovascular (CV) event-free survival according to epicardial adipose tissue (EAT) thickness ≦6.0 mm and >6.0 mm (log-rank P < 0.001).
Table 4 shows incremental values of EAT thickness in the prediction of adverse CV events. Adding EAT thickness to a Cox model containing CHA2DS2-VASc score, LAVI, LVEF, and E/E′ demonstrated significant improvement in the prediction of future CV events (difference inχ2 value: 14.65, P < 0.001 and IDI 0.10, 95% CI 0.04–0.16, P < 0.001). Subgroup analysis, which focused only on patients with a CHA2DS2-VASc score of 0 for males (n = 19) and 1 for females (n = 6), was performed and EAT thickness could further identify CV events in males with a CHA2DS2-VASc score of 0 (IDI 0.13, 95% CI 0.05–0.21, P = 0.002) and showed a trend in further identifying CV events in females with a CHA2DS2-VASc score of 1 (IDI 0.10, 95% CI -0.01–0.20, P = 0.056), when compared with a Cox model containing CHA2DS2-VASc score, LAVI, LVEF, and E/E′.
TABLE 4:
Incremental Values of EAT Thickness in Relation to CV Events
For measurement of EAT thickness, our result demonstrated that the intraobserver mean percent error was 7.1% ± 10.1%, whereas the interobserver mean percent error was 8.3% ± 12.1%.
DISCUSSION
Our study estimated the association of EAT thickness with adverse CV events in patients with documented AF. We found that higher EAT thickness was associated independently with further risks of adverse CV events in AF patients. Additionally, the EAT thickness could add significant incremental values beyond the conventional clinical and echocardiographic parameters in prediction of adverse CV events.
EAT represents a type of visceral adipose tissue adjacent the heart and has a unique character in terms of the biochemical composition, the size of its adipocytes, and metabolic activity. EAT has been proposed to influence the development of coronary atherosclerosis owing to its endocrine and paracrine activity by secreting anti-inflammatory and proinflammatory cytokines and chemokines.29–31 Jeong et al found that in patients with documented CAD, EAT thickness was correlated significantly with its severity of atherosclerosis.5 Several studies showed age and BMI were associated significantly with EAT thickness.5,7,14,32 In patients diagnosed with metabolic syndrome, Park et al13 also found that E/E′ was associated significantly with EAT thickness. In the present study, we similarly demonstrated that compared to AF patients with EAT thickness ≦6.0 mm, those with EAT thickness >6.0 mm had an older age, more prevalent CAD, and higher BMI and E/E′.
In the general population, Mahabadi et al14 found an association between EAT and incident myocardial infarction and CV risk factors. In the cross-section study by Akil et al,33 the EAT thickness in patients diagnosed with ischemic stroke was found to be higher than those in healthy controls. Several studies have showed EAT thickness is associated significantly with left ventricular diastolic dysfunction.13,34 Left ventricular diastolic dysfunction was correlated to poor CV prognosis.35–37 In our study, the multivariable analysis consistently revealed that increased E/E′ and old age were independently associated with increased EAT thickness. Old age is a well-recognized risk factor of adverse CV events.38 E/E’ was also a strong predictor of adverse CV events in various population including patients with type 2 diabetes mellitus, chronic kidney disease, and dialysis.39–41 Our present results found EAT thickness as a useful predictor for adverse CV events even after adjustment for many important clinical and echocardiographic parameters might be partially explained by the older age and higher E/E′ in our patients with higher EAT thickness. Besides, adding EAT thickness to a Cox model containing CHA2DS2-VASc score, LAVI, LVEF, and E/E′ demonstrated significant improvement in the prediction of future CV events. Hence, EAT thickness may be a useful indicator of poor CV outcomes in AF patients.
STUDY LIMITATIONS
Several limitations existed in our study. First, EAT thickness was measured by echocardiography and by one independent cardiologist, but not by computed tomography scan. Hence, we could not analyze the relationship between periatrial fat and adverse CV events in AF patients. Second, we did not withdraw antihypertensive, antiplatelet, or anticoagulant medications for ethical reasons. Hence, pharmacological influence on the present findings could not be excluded. However, we have taken the use of medicine into consideration in the multivariable analysis. Third, the using rate of anticoagulation medication was low in our study, but it was similar to our national survey for AF.30 Fouth, because AF patients were selected from those who received transthoracic echocardiographic examination, it was predisposed to selection bias and resulted in findings potentially less generalized. Finally, the limited power and the possibility of chance findings should be considered owing to the large amount of variables in the analysis but with only 69 outcomes.
CONCLUSIONS
In AF, EAT thickness was useful in predicting future CV events. Additionally, EAT thickness could provide incremental value for CV outcome prediction over traditional clinical and echocardiographic parameters in AF.
Acknowledgments
We are grateful to the assistance from the Statistical Analysis Laboratory, Department of Medical Research, Kaohsiung Medical University Hospital, Kaohsiung Medical University.
REFERENCES
1. Camm AJ, Lip GY, De Caterina R, et al. 2012 focused update of the ESC Guidelines for the management of atrial fibrillation: an update of the 2010 ESC Guidelines for the management of atrial fibrillation. Developed with the special contribution of the European Heart Rhythm Association.
Eur Heart J 2012; 33:2719–2747.
2. Chugh SS, Havmoeller R, Narayanan K, et al. Worldwide epidemiology of atrial fibrillation: a Global Burden of Disease 2010 Study.
Circulation 2014; 129:837–847.
3. Mozaffarian D, Benjamin EJ, Go AS, et al. Heart disease and stroke statistics – 2015 update: a report from the American Heart Association.
Circulation 2015; 131:e29–322.
4. January CT, Wann LS, Alpert JS, et al. 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the Heart Rhythm Society.
J Am Coll Cardiol 2014; 64:e1–76.
5. Jeong JW, Jeong MH, Yun KH, et al. Echocardiographic epicardial fat thickness and coronary artery disease.
Circ J 2007; 71:536–539.
6. Karastergiou K, Fried SK. Multiple adipose depots increase cardiovascular risk via local and systemic effects.
Curr Atheroscler Rep 2013; 15:361.
7. Natale F, Tedesco MA, Mocerino R, et al. Visceral adiposity and arterial stiffness: echocardiographic epicardial fat thickness reflects, better than waist circumference, carotid arterial stiffness in a large population of hypertensives.
Eur J Echocardiogr 2009; 10:549–555.
8. Iacobellis G, Ribaudo MC, Assael F, et al. Echocardiographic epicardial adipose tissue is related to anthropometric and clinical parameters of metabolic syndrome: a new indicator of cardiovascular risk.
J Clin Endocrinol Metab 2003; 88:5163–5168.
9. Bos D, Shahzad R, van Walsum T, et al. Epicardial fat volume is related to atherosclerotic calcification in multiple vessel beds.
Eur Heart J Cardiovasc Imag 2015; 16:1264–1269.
10. Al Chekakie MO, Welles CC, Metoyer R, et al. Pericardial fat is independently associated with human atrial fibrillation.
J Am Coll Cardiol 2010; 56:784–788.
11. Tsao HM, Hu WC, Wu MH, et al. Quantitative analysis of quantity and distribution of epicardial adipose tissue surrounding the left atrium in patients with atrial fibrillation and effect of recurrence after ablation.
Am J Cardiol 2011; 107:1498–1503.
12. Chao TF, Hung CL, Tsao HM, et al. Epicardial adipose tissue thickness and ablation outcome of atrial fibrillation.
PLoS One 2013; 8:e74926.
13. Park HE, Choi SY, Kim M. Association of epicardial fat with left ventricular diastolic function in subjects with metabolic syndrome: assessment using 2-dimensional echocardiography.
BMC Cardiovasc Disord 2014; 14:3.
14. Mahabadi AA, Berg MH, Lehmann N, et al. Association of epicardial fat with cardiovascular risk factors and incident myocardial infarction in the general population: the Heinz Nixdorf Recall Study.
J Am Coll Cardiol 2013; 61:1388–1395.
15. Ulucan S, Kaya Z, Efe D, et al. Epicardial fat tissue predicts increased long-term major adverse cardiac event in patients with moderate cardiovascular risk.
Angiology 2015; 66:619–624.
16. Chen SC, Chang JM, Liu WC, et al. Echocardiographic parameters are independently associated with increased cardiovascular events in patients with chronic kidney disease.
Nephrol Dial Transplant 2012; 27:1064–1070.
17. Hsu PC, Tsai WC, Lin TH, et al. Association of arterial stiffness and electrocardiography-determined left ventricular hypertrophy with left ventricular diastolic dysfunction.
PLoS One 2012; 7:e49100.
18. Lee WH, Hsu PC, Chu CY, et al. Anemia as an independent predictor of adverse cardiac outcomes in patients with atrial fibrillation.
Int J Med Sci 2015; 12:618–624.
19. Chen SC, Lee WH, Hsu PC, et al. Association of brachial-ankle pulse wave velocity with cardiovascular events in atrial fibrillation.
Am J Hypertens 2016; 29:348–356.
20. Devereux RB, Alonso DR, Lutas EM, et al. Echocardiographic assessment of left ventricular hypertrophy: comparison to necropsy findings.
Am J Cardiol 1986; 57:450–458.
21. Lang RM, Bierig M, Devereux RB, et al. Recommendations for chamber quantification: a report from the American Society of Echocardiography's Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology.
J Am Soc Echocardiogr 2005; 18:1440–1463.
22. Iacobellis G, Assael F, Ribaudo MC, et al. Epicardial fat from echocardiography: a new method for visceral adipose tissue prediction.
Obes Res 2003; 11:304–310.
23. Govindan M, Kiotsekoglou A, Saha SK, et al. Validation of echocardiographic left atrial parameters in atrial fibrillation using the index beat of preceding cardiac cycles of equal duration.
J Am Soc Echocardiogr 2011; 24:1141–1147.
24. Lee CS, Lin TH, Hsu PC, et al. Measuring left ventricular peak longitudinal systolic strain from a single beat in atrial fibrillation: validation of the index beat method.
J Am Soc Echocardiogr 2012; 25:945–952.
25. Kusunose K, Yamada H, Nishio S, et al. Index-beat assessment of left ventricular systolic and diastolic function during atrial fibrillation using myocardial strain and strain rate.
J Am Soc Echocardiogr 2012; 25:953–959.
26. Ho KK, Anderson KM, Kannel WB, et al. Survival after the onset of congestive heart failure in Framingham Heart Study subjects.
Circulation 1993; 88:107–115.
27. Thygesen K, Alpert JS, White HD, et al. Joint ESCAAHAWHFTFftRoMI. Universal definition of myocardial infarction.
Circulation 2007; 116:2634–2653.
28. Levey AS, Coresh J, Greene T, et al. Using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate.
Ann Intern Med 2006; 145:247–254.
29. Mazurek T, Zhang L, Zalewski A, et al. Human epicardial adipose tissue is a source of inflammatory mediators.
Circulation 2003; 108:2460–2466.
30. Laine P, Kaartinen M, Penttila A, et al. Association between myocardial infarction and the mast cells in the adventitia of the infarct-related coronary artery.
Circulation 1999; 99:361–369.
31. Hirata Y, Tabata M, Kurobe H, et al. Coronary atherosclerosis is associated with macrophage polarization in epicardial adipose tissue.
J Am Coll Cardiol 2011; 58:248–255.
32. Park JS, Ahn SG, Hwang JW, et al. Impact of body mass index on the relationship of epicardial adipose tissue to metabolic syndrome and coronary artery disease in an Asian population.
Cardiovasc Diabetol 2010; 9:29.
33. Akil E, Akil MA, Varol S, et al. Echocardiographic epicardial fat thickness and neutrophil to lymphocyte ratio are novel inflammatory predictors of cerebral ischemic stroke.
J Stroke Cerebrovasc Dis 2014; 23:2328–2334.
34. Lin HH, Lee JK, Yang CY, et al. Accumulation of epicardial fat rather than visceral fat is an independent risk factor for left ventricular diastolic dysfunction in patients undergoing peritoneal dialysis.
Cardiovasc Diabetol 2013; 12:127.
35. Park SJ, Lee SC, Jang SY, et al. E/e’ ratio is a strong prognostic predictor of mortality in patients with non-valvular atrial fibrillation with preserved left ventricular systolic function.
Circ J 2011; 75:2350–2356.
36. Hillis GS, Moller JE, Pellikka PA, et al. Noninvasive estimation of left ventricular filling pressure by E/e’ is a powerful predictor of survival after acute myocardial infarction.
J Am Coll Cardiol 2004; 43:360–367.
37. Gharacholou SM, Scott CG, Takahashi PY, et al. Left ventricular diastolic function and long-term outcomes in patients with normal exercise echocardiographic findings.
Am J Cardiol 2013; 112:200–207.
38. Go AS, Mozaffarian D, Roger VL, et al. Heart disease and stroke statistics – 2013 update: a report from the American Heart Association.
Circulation 2013; 127:e6–e245.
39. Blomstrand P, Engvall M, Festin K, et al. Left ventricular diastolic function, assessed by echocardiography and tissue Doppler imaging, is a strong predictor of cardiovascular events, superior to global left ventricular longitudinal strain, in patients with type 2 diabetes.
Eur Heart J Cardiovasc Imag 2015; 16:1000–1007.
40. Han JH, Han JS, Kim EJ, et al. Diastolic dysfunction is an independent predictor of cardiovascular events in incident dialysis patients with preserved systolic function.
PLoS One 2015; 10:e0118694.
41. Chen SC, Chang JM, Tsai YC, et al. Ratio of transmitral E-wave velocity to early diastole mitral annulus velocity with cardiovascular and renal outcomes in chronic kidney disease.
Nephron Clin Pract 2013; 123:52–60.
