- Email: [email protected]
Development and validation of a risk calculator predicting exercise-induced ventricular arrhythmia in patients with cardiovascular disease
International Journal of Cardiology 220 (2016) 625–628
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Development and validation of a risk calculator predicting exercise-induced ventricular arrhythmia in patients with cardiovascular disease☆ Ilarraza-Lomelí Hermes a,⁎, García-Saldivia Marianna a, Rojano-Castillo Jessica a, Barrera-Ramírez Carlos b, Chávez-Domínguez Rafael a, Rius-Suárez María Dolores a, Iturralde Pedro c a b c
Cardiac Rehabilitation Department, National Institute of Cardiology Ignacio Chavez, Mexico City, Mexico Hospital La Concepción, Saltillo, Coahuila, Mexico Electrocardiology Department, National Institute of Cardiology Ignacio Chavez, Mexico City, Mexico
a r t i c l e
i n f o
Article history: Received 27 April 2016 Accepted 24 June 2016 Available online 29 June 2016 Keywords: Heart failure Sudden death Exercise test Arrhythmia Myocardial infarction
a b s t r a c t Background: Mortality due to cardiovascular disease is often associated with ventricular arrhythmias. Nowadays, patients with cardiovascular disease are more encouraged to take part in physical training programs. Nevertheless, high-intensity exercise is associated to a higher risk for sudden death, even in apparently healthy people. During an exercise testing (ET), health care professionals provide patients, in a controlled scenario, an intense physiological stimulus that could precipitate cardiac arrhythmia in high risk individuals. There is still no clinical or statistical tool to predict this incidence. The aim of this study was to develop a statistical model to predict the incidence of exercise-induced potentially life-threatening ventricular arrhythmia (PLVA) during high intensity exercise. Methods and results: 6415 patients underwent a symptom-limited ET with a Balke ramp protocol. A multivariate logistic regression model where the primary outcome was PLVA was performed. Incidence of PLVA was 548 cases (8.5%). After a bivariate model, thirty one clinical or ergometric variables were statistically associated with PLVA and were included in the regression model. In the multivariate model, 13 of these variables were found to be statistically significant. A regression model (G) with a X2 of 283.987 and a p b 0.001, was constructed. Significant variables included: heart failure, antiarrhythmic drugs, myocardial lower-VD, age and use of digoxin, nitrates, among others. Conclusion: This study allows clinicians to identify patients at risk of ventricular tachycardia or couplets during exercise, and to take preventive measures or appropriate supervision. © 2016 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Worldwide, cardiovascular diseases are the cause of non-traumatic sudden death, due to ventricular arrhythmia in more than 50% of cases [1,2]. Intense physical exercise is associated with the appearance several kinds of arrhythmia, with a rate of premature ventricular complexes during an exercise testing (ET) around 70% of patients with heart disease, and 34% in apparently healthy people [3]. Furthermore, the presence of frequent ventricular ectopy in a stress testing is also associated with a higher risk of all causes mortality, sudden death or acute coronary syndrome [4–8]. Consequently, the attending physician needs to identify which patients are at high risk of exercise-related arrhythmia but, at present there is no stratification tool to predict it [9–14]. The aim
☆ This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. ⁎ Corresponding author at: Juan Badiano 01, Colonia Seccion XVI, Tlalpan, Mexico City 14080, Mexico. E-mail address: [email protected] (I.-L. Hermes).
http://dx.doi.org/10.1016/j.ijcard.2016.06.196 0167-5273/© 2016 Elsevier Ireland Ltd. All rights reserved.
of this study was to find the statistical association among baseline characteristics of patients with cardiovascular disease with potentially life-threatening ventricular arrhythmia (PLVA) during intense exercise. 2. Methods We studied a database of over 6000 patients with cardiovascular disease, who underwent exercise testing in a cardiology referral center, between 2006 and 2015. Clinical files without complete information were ruled out. Patients with more than one exercise testing were included. Medication was not suspended before exercise and all patients signed a consent form. Every ET was performed using a Balke ramp protocol, and tests were stopped according to symptoms or an absolute indication. A Schiller CS-200 © device with a Trackmaster treadmill was used to perform all exercise tests. An electrocardiographic signal (ECG) was recorded throughout the test. Blood pressure (BP) was measured every minute during exercise, and at the 1st, 3rd, 5th and 8th min of recovery using a calibrated aneroid sphygmomanometer. The primary composite outcome (dependent variable) was the occurrence of potentially life-threatening ventricular arrhythmia during the ET, defined as: ventricular fibrillation, ventricular tachycardia or couplets. Independent variables included those obtained from the clinical record. The nominal, categorical or numerical variables were presented as frequency (percentage), mean (median) and standard deviation (range) as appropriate.
626
I.-L. Hermes et al. / International Journal of Cardiology 220 (2016) 625–628
Table 1 Demographic characteristics of the patients, according to the presence of potentially life-threatening ventricular arrhythmia Variable
PLVA n = 548 (8%)
No PLVA n = 5867 (92%)
Total n = 6415 (100%)
p value
Age (years) Height (cm) Weight BMI (kg/m2) Female:Male
55 ± 15 165 ± 1 74 ± 14 26.7 ± 5 97 (7%): 451 (9%)*
48 ± 20 161 ± 1 68 ± 18 25.4 ± 5 1317 (93%): 4550 (91%)
49 ± 20 162 ± 1.4 69 ± 18 25.5 ± 5 1414 (22%): 5001 (78%)
b0.001 b0.001 b0.001 b0.001 b0.05*
Diagnosis Ischemic heart disease Congenital heart disease Idiopathic cardiomyopathy Valvular heart disease Mixed diagnosis Healthy individuals Other diagnosis CABG PTCA History of ventricular arrhythmia ICD Heart failure
362 (67%) 10 (2%) 102 (18%) 19 (3%) 32 (6%) 6 (1%) 17 (3%) 95 (17%) 161 (29%) 31 (6%) 26 (5%) 266 (48 %)
3855 (66%) 488 (8%) 425 (7%) 170 (3%) 158 (3%) 301 (5%) 470 (8%) 772 (13%) 1736 (29%) 124 (2%) 143 (2%) 1521 (26%)
4217 (66%) 498 (7%) 527 (8%) 189 (3%) 190 (3%) 307 (5%) 487 (8%) 867 (14%) 1897 (29%) 155 (2%) 169 (3%) 1787 (28%)
b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001
Drugs Betablocker ACE/ARB Nitrates Digoxin Calcium channel blocker Anti-arrhythmic Statins Diuretics Anti-platelet
416 (76%) 446 (81%) 92 (17%) 170 (31%) 86 (16%) 110 (20%) 328 (60%) 248 (45%) 3968 (72%)
3920 (67 %) 3806 (65%) 1263 (22%) 881 (15%) 896 (15%) 509 (9%) 3375 (58%) 1429 (24%) 3960 (67%)
4336 (68%) 4252 (66%) 1357 (21%) 1051 (16%) 982 (15%) 619 (10%) 3703 (58%) 1677 (26%) 4356 (68%)
b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001
* Differences between women and men. Abbreviations: potentially life-threatening ventricular arrhythmia (PLVA), body mass index (BMI), coronary artery bypass grafting (CABG), percutaneous transluminal coronary angioplasty (PTCA), implantable cardioverter-defibrillator (ICD), angiotensin-converting enzyme/angiotensin receptor blocker (ACE/ARB).
Table 2 Relative risk of exercise-induced potentially life-threatening ventricular arrhythmia according to bivariate analysis. Variable
RR
CI 95%
p
Age (N54 years) BMI (≥25.7 kg/m2) Male gender Ischemic heart disease Anterior Wall myocardial infarction Inferior Wall, right ventricle myocardial infarction Re-infarction CABG PTCA Congenital heart disease Dilated cardiomyopathy Valvular heart disease Mixed diagnosis Healthy individuals History of ventricular arrhythmia ICD Heart failure
1.53 1.45 1.35 1.03 0.69 1.47 1.33 1.38 0.99 3.73 3.19 1.81 2.24 0.20 2.8 1.99 2.70
1.28 a 1.82 1.22 a 1.73 1.07 a 1.70 0.86 a 1.25 0.50 a 0.96 1.22 a 1.79 0.97 a 1.82 1.10 a 1.75 0.82 a 1.20 2.22 a 6.27 2.56 a 3.97 1.32 a 2.51 1.52 a 3.31 0.09 a 0.46 1.86 a 4.16 1.3 a 3.06 2.26 a 3.22
b0.001 b0.001 b0.05 ns b0.05 b0.05 ns b0.01 ns b0.01 b0.001 b0.001 b0.001 b0.001 b0.001 b0.01 b0.001
Drugs Betablocker ACE/ARB Nitrates Digoxin Calcium channel blocker Anti-arrhythmic Statins Diuretics Anti-platelet Resting heart rate (b60 bpm) Resting systolic blood pressure (≥140 mm Hg) Resting dyastolic blood pressure (≥90 mm Hg) Resting double product (≥7750 bpm ∗ mm Hg)
1.57 2.37 0.73 2.55 1.03 2.64 1.10 2.57 1.25 0.81 1.64 1.47 1.29
1.28 a 1.92 1.90 a 2.96 0.58 a 0.93 2.10 a 3.10 0.81 a 1.31 2.10 a 3.32 0.92 a 1.32 2.15 a 3.07 1.03 a 1.5 0.66 a 1.0 1.27 a 2.12 1.18 a 1.82 1.08 a 1.53
b0.001 b0.001 b0.01 b0.001 ns b0.001 ns b0.001 b0.05 0.06 b0.001 b0.001 b0.01
Abbreviations: body mass index (BMI), coronary artery bypass grafting (CABG), percutaneous transluminal coronary angioplasty (PTCA), implantable cardioverter-defibrillator (ICD), angiotensin-converting enzyme/angiotensin receptor blocker (ACE/ARB), beats per minute (bpm).
A bivariate analysis was performed including those clinical or paraclinical variables that had been previously recognized as independent risk factors. The variables with a statistically significant association with the outcome (p b 0.05), were recruited into a logisticregression multivariable analysis. Variables were included in the regression model using a step by step forward-Wald modality. The entry value was 0.05 and variables with a p value above 0.1 were removed. The statistical analysis was performed using a SPSS-19 program. Finally, the linear general model (G) was incorporated to the logistic equation to find the probability for PLVA incidence. PLVAðriskÞ ¼
1 : ‐G 1þe
Table 3 Variables included in the multiple logistic regression model.
Constant Heart failure (0/1) Anti-arrhythmic drugs (0/1) Inferior wall, right ventricle MI (0/1) Age (years) Nitrates (0/1) Diuretics (0/1) Anterior Wall MI (0/1) Valvular heart disease (0/1) Resting double product
B
ES
Wald
exp β
p
−4.712 0.610 0.464 0.412 0.014 −0.401 0.362 −0.485 0.483 0.270
0.316 0.114 0.126 0.107 0.003 0.125 0.112 0.171 0.177 0.097
222.924 28.82 13.458 14.791 19.843 10.222 10.424 8.087 7.462 7.697
0.009 1.840 1.590 1.510 1.014 0.670 1.436 0.615 1.621 1.310
b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 0.004 0.006 0.006
0.325
0.120
7.359
1.384
0.007
0.026 0.292 0.272
0.011 0.126 0.126
5.738 5.397 4.673
1.027 1.339 1.313
0.017 0.020 0.031
(≥7750 bpm ∗ mm Hg)
Resting dyastolic blood pressure (≥90 mm Hg)
BMI (kg/m2) ACE/ARB (0/1) Digoxin (0/1)
Abbreviations: myocardial infarction (MI), beats per minute (bpm), body mass index (BMI), angiotensin-converting enzyme/angiotensin receptor blocker (ACE/ARB). Dichotomous variables should be codified as follows: absence (0), presence (1).
I.-L. Hermes et al. / International Journal of Cardiology 220 (2016) 625–628
627
Finally, the probability of PLVA appearance may be calculated as 1 follows: PLVAðriskÞ ¼ ½1þe ‐G ; where G = (− 4.759) + (heart failure × 0.694) + (antiarrhythmic drug (0/1) × 0.491) + (Inferior (0/1) wall or right ventricle myocardial infarction (0/1) × 0.389) + (age (years) ∗ 0.015) + (digoxin (0/1) × 0.402) + (nitrates (0/1) × [−0.405]) + (di(diuretics (0/1) × 0.362) + (valvular heart disease (0/1) × 0.534) + (anterior wall myocardial infarction (0/1) × [− 0.503]) + (resting doubleproduct ≥ 7750 (0/1) × 0.276) + (resting diastolic blood pressure ≥ 90 mm Hg (0/1) × 0.309) + (body mass index (kg/m2) × 0.028) + (ACE or ARB (0/1) × 0.315). Fig. 1 shows the receiver operating characteristic curve (ROC) of the logistic regression model. Thereby, if a 65 years old patient has heart failure, takes digoxin, diuretic, ACE, and has a body mass index of 26 kg/m2, the risk of VPLA during intense exercise is 21.7% (see Fig. 2)
4. Discussion Fig. 1. shows the receiver operating characteristic curve (ROC) of the logistic regression model Area under curve (AUC).
3. Results A total of 6415 ET were analyzed, and 548 (8.5%) outcome events were detected, including 118 (21%) episodes of ventricular tachycardia. Table 1 shows the demographic characteristics according to the presence of PLVA. Bivariate analysis (Table 2) identified several predictive variables for arrhythmia: the presence of congenital heart disease; dilated cardiomyopathy; a history of ventricular arrhythmia; heart failure, and some drugs like antiarrhythmic or diuretics among others. Thirty one independent variables were stochastically associated with arrhythmia by bivariate analysis and were recruited into a logistic multivariable regression model (G). Table 3 shows the 13 variables that set up the statistically significant G-model (X2 = 283,987, p b 0.001). The Hosmer-Lemeshow goodness of fit model at 13rd step was X2 = 8.43, p = 0.39.
Several authors have shown that exercise testing is a helpful tool for patient evaluation, particularly assessing variables like exercise tolerance and the presence of ischemia or electrical stability. We observed that the exercise-induced ventricular potentially life-threatening arrhythmia among patients with heart disease is not low (8.5%). The frequency of exercise-induced arrhythmia is generally explained with at least one of the following mechanisms: sympathetic system activation, myocardial late potentials or re-entry phenomena. In this statistical model, the reason that some of these 13 variables were associated with arrhythmia is clearly base on the presence of myocardial damage (infarction or heart failure) or on electrical conduction heterogeneity (myocardial scar, ischemia or QT dispersion due to some drugs) [10, 11,15–18]. According to reports of many researchers, anterior and inferior wall myocardial infarctions are associated with a higher incidence of ventricular arrhythmia. [19,20] A reasonable explanation for this topographic distribution is founded on the anatomical model presented by professor Torrent-Guasp, who stated that the myocardium of both ventricles is, in fact, the same helical band, disposed as a complex interaction of muscle
Fig. 2. This clinimetric tool, allows physician to predicting ventricular arrhythmia during intense physical exercise according to several clinical variables. In this example, we calculate the risk of VPLA in a 65 years old patient with heart failure, taking digoxin, diuretic, ACE, and with a body mass index of 26 kg/m2. Abbreviations: right ventricle (RV), myocardial infarction (MI), heart rate (HR), blood pressure (BP), body mass index (BMI), angiotensin-converting enzyme (ACE), angiotensin receptor blocker (ARB), potentially life-threatening ventricular arrhythmia (PLVA).
I.-L. Hermes et al. / International Journal of Cardiology 220 (2016) 625–628
628
fibers. The inferior wall is composed of myocardial fibers located in several directions and any damage can produce different sites of re-entry [21,22]. On the other hand, in our study, the anterior wall infarction was related to a lower probability of ventricular arrhythmia, independent of the presence of heart failure. This observation has not yet been described in medical literature, and we do not have any convincing explanation. Habitually, patients with anterior myocardial infarction are treated more intensively than those with inferior infarction, and this could explain that finding. Although Anastassiades et al. stated that patients that underwent cardiac surgery show higher rates of exercise-induced ventricular arrhythmia, in our study this association was stochastically significant only in the bivariate analysis [23]. The nitrates intake was clearly a preventive factor of ventricular arrhythmia, and this could be attributed to a lower incidence of coronary spasm [24]. Obesity has been linked to certain kinds of arrhythmia, particularly atrial fibrillation, mainly due to an increment of pericardial fat. Our results showed that obesity is also associated with ventricular arrhythmia [25–27]. In this study we also found that diuretic and digoxin intake, were strongly correlated with exercise-induced arrhythmia. These two kind of drugs increase sudden death rate by conditioning hydroelectrolytic disorders like hypokalemia or high levels of intracellular concentrations of calcium [28–31]. The ACE/ARB medication is statistically associated with PLVA, and we do not have any plausible explanation at the moment [32]. Study limitations are centered on the retrolective nature of the study, and the lack of a healthy control group. This results show that clinical variables can predict the PLVA in patients with cardiovascular disease, and further research is needed to study other groups of patients. 5. Conclusion The frequency of exercise-induced ventricular arrhythmias in patients with cardiovascular disease is elevated, and strongly associated with several clinical variables. This arrhythmia-predictive calculator may be a useful risk stratification tool for the physician Financial support No financial support has been granted. Conflict of interest The authors declare no conflict of interest. Acknowledgements We would like to thank all people that work in the Cardiac Rehabilitation Service in the National Institute of Cardiology Ignacio Chavez, for their daily labor. References [1] O. Yousuf, J. Chrispin, G. Tomaselli, G. Berger, Clinical management and prevention of sudden cardiac death, Circ. Res. 116 (2015) 2020–2040. [2] R. Chávez Domínguez, J.A. Ramírez Hernández, J.M. Casanova Garcés, Coronary heart disease in Mexico and the clinical epidemiological and preventive relevance, Arch. Cardiol. Mex. 73 (2) (Apr-Jun 2003) 105–114. [3] R.A. Candinas, P.J. Podrid, Evaluation of cardiac arrhythmias by exercise testing, Herz 15 (1) (1990 Feb) 21–27. [4] P.A. Maggioni, G. Zuanetti, et al., Prevalence and prognostic significance of ventricular arrhythmias after acute myocardial infarction in the fibrinolytic era, Circulation 87 (1993) 312–322. [5] J.A. Udall, M.H. Ellestad, Predictive implications of ventricular premature contractions associated with treadmill stress testing, Circulation 56 (1977) 985–989.
[6] R.A. Schweikert, F.J. Pashkow, C.E. Snader, T.H. Marwick, M.S. Lauer, Clinical relevance of exercise-induced ventricular ectopic activity with thallium myocardial perfusion and angiographic coronary artery disease in stable, low-risk populations, Am. J. Cardiol. 83 (1999) 530–534. [7] C.K. Nair, W. Thomson, W.S. Aronow, T. Pagano, K. Ryschon, M.H. Sketch, Prognostic significance of exercise-induced complex ventricular arrhythmias in coronary artery disease with normal and abnormal left ventricular ejection fraction, Am. J. Cardiol. 54 (1984) 1136–1138. [8] R.E. Eckart, M.E. Field, T.W. Hruczkowski, et al., Association of electrocardiographic morphology of exercise-induced ventricular arrhythmia with mortality, Ann. Intern. Med. 149 (2008) 451–460. [9] X. Jouven, A. Zureik, I. Desnos, D. Courbo, P. Ducimetière, Long-term outcome in asymptomatic men with exercise-induced premature ventricular depolarizations, N. Engl. J. 343 (2000) 826–833. [10] G.F. Fletcher, G.J. Balady, Exercise standards for testing and training, Circulation 104 (2001) 1694–1740. [11] A.N. DeMaria, Z. Vera, E.A. Amsterdam, D.T. Mason, R.A. Massumi, Disturbances of cardiac rhythm and conduction induced by exercise: diagnostic, prognostic and therapeutic implications, Am. J. Cardiol. 33 (1974) 732–736. [12] N. Franco, G. Lepera, A. Rolli, Arrhythmic risk stratification of post-myocardial infarction patients, Curr. Opin. Cardiol. 15 (2000) 1–6. [13] R.A. Rourke, Risk stratification after myocardial infarction, Circulation 84 (Suppl. I) (1991) I-177–I-181. [14] J. Frolkis, C. Pothier, E. Blackstone, L. Lauer, Frequent ventricular ectopy after exercise as a predictor of death, N. Engl. J. 348 (2003) 781–790. [15] T.M. Kelly, Exercise testing and training of patients with malignant ventricular arrhythmias, Med. Sci. Sports Exerc. 28 (1996) 53–61. [16] D. Katritsis, J. Camm, Nonsustained ventricular tachycardia: where do we stand? Eur. Heart J. 25 (2004) 1093–1099. [17] J.O. O'Neill, J.B. Young, C.E. Pothier, M.S. Lauer, Severe frequent ventricular ectopy after exercise as a predictor of death in patients with heart failure, J. Am. Coll. Cardiol. 44 (2004) 820–826. [18] G.F. Fletcher, G.J. Balady, et al., Exercise standards for testing and training, Circulation 104 (2001) 1694–1740. [19] M.J. Shah, N.R. Bhatt, A. Dabhi, P.B. Thorat, K. Chudasama, J. Patel, A study of 100 cases of arrhythmias in first week of acute myocardial infarction (AMI) in Gujarat: a high risk and previously undocumented population, J. Clin. Diagn. Res. 8 (1) (Jan 2014) 58–61. [20] M. Olimulder, K. Kraaier, M. Galjee, M. Scholten, J. van Es, L. Wagenaar, J. van der Palen, C. von Birgelen, Infarct tissue characterization in implantable cardioverter defibrillator recipients for primary versus secondary prevention following myocardial infarction: a study with contrast enhancement cardiovascular magnetic resonance imaging, Int. J. Card. Imaging 29 (2013) 169–176. [21] M.J. Kocica, A.F. Corno, V. Lackovic, V.I. Kanjuh, The helical ventricular myocardial band of torrent-Guasp, Semin. Thorac. Cardiovasc. Surg. Pediatr. Card. Surg. Annu. (2007) 52–60. [22] F. Torrent-Guasp, M.J. Kocica, A.F. Corno, M. Komeda, F. Carreras-Costa, A. Flotats, J. Cosin-Aguillar, H. Wen, Towards new understanding of the heart structure and function, Eur. J. Cardiothorac. Surg. 27 (2) (Feb 2005) 191–201. [23] L.C. Anastassiades, A.G. Antonopoulos, A.A. Petsas, The effect of coronary revascularization on exercise-induced ventricular ectopic activity, Eur. Heart J. 8 (Suppl. D) (Aug 1987) 75–78. [24] K.L. Looi, A. Grace, S. Agarwal, Coronary artery spasm and ventricular arrhythmias, Postgrad. Med. J. 88 (1042) (Aug 2012) 465–471. [25] C.A. Goudis, P. Korantzopoulos, I.V. Ntalas, E.M. Kallergis, D.G. Ketikoglou, Obesity and atrial fibrillation: a comprehensive review of the pathophysiological mechanisms and links, J. Cardiol. (May 7 2015) (pii: S0914-5087(15)00118-5). [26] R.K. Pathak, A. Elliott, M.E. Middeldorp, M. Meredith, A.B. Mehta, R. Mahajan, et al., Impact of cardiorespiratory fitness on arrhythmia recurrence in obese individuals with atrial fibrillation: the CARDIO-FIT study, J. Am. Coll. Cardiol. (Jun 20 2015) (pii: S0735-1097(15)03339-2). [27] A. Sabbag, Y. Sidi, S. Kivity, R. Beinart, M. Glikson, S. Segev, I. Goldenberg, E. Maor, Obesity and exercise-induced ectopic ventricular arrhythmias in apparently healthy middle aged adults, Eur. J. Prev. Cardiol. (Jun 11 2015) (pii: 2047487315591442. [Epub ahead of print]). [28] D.E. Grobbee, A.W. Hoes, Non-potassium-sparing diuretics and risk of sudden cardiac death, J. Hypertens. 13 (1995) 1539–1545. [29] C.E. Leonard, H. Razzaghi, C.P. Freeman, J.A. Roy, C.W. Newcomb, S. Hennessy, Empiric potassium supplementation and increased survival in users of loop diuretics, PLoS One 16 (2014 Jul) 9(7). [30] M. Shenasa, H. Shenasa, N. El-Sherif, Left ventricular hypertrophy and Arrhythmogenesis, Card Electrophysiol. Clin. 7 (2) (2015 Jun) 207–220. [31] L.A. Gonano, M. Sepúlveda, Y. Rico, M. Kaetzel, C.A. Valverde, J. Dedman, A. Mattiazzi, P.M. Vila, Calcium-calmodulin kinase II mediates digitalis-induced arrhythmias, Circ. Arrhythm. Electrophysiol. 4 (6) (2011 Dec) 947–957. [32] W.A. AlJaroudi, M.M. Refaat, R.H. Habib, L. Al-Shaar, M. Singh, R. Gutmann, H.L. Bloom, S.C. Dudley, P.T. Ellinor, S.F. Saba, A.A. Shalaby, R. Weiss, D.M. McNamara, I. Halder, B. London, Genetic risk assessment of defibrillator events investigators. Effect of angiotensin-converting enzyme inhibitors and receptor blockers on appropriate implantable cardiac defibrillator shock in patients with severe systolic heart failure (from the GRADE multicenter study), Am. J. Cardiol. 115 (7) (Apr 1 2015) 924–931, http://dx.doi.org/10.1016/j.amjcard.2015.01.020 Epub 2015 Jan 15.
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Development and validation of a risk calculator predicting exercise-induced ventricular arrhythmia in patients with cardiovascular disease☆ Ilarraza-Lomelí Hermes a,⁎, García-Saldivia Marianna a, Rojano-Castillo Jessica a, Barrera-Ramírez Carlos b, Chávez-Domínguez Rafael a, Rius-Suárez María Dolores a, Iturralde Pedro c a b c
Cardiac Rehabilitation Department, National Institute of Cardiology Ignacio Chavez, Mexico City, Mexico Hospital La Concepción, Saltillo, Coahuila, Mexico Electrocardiology Department, National Institute of Cardiology Ignacio Chavez, Mexico City, Mexico
a r t i c l e
i n f o
Article history: Received 27 April 2016 Accepted 24 June 2016 Available online 29 June 2016 Keywords: Heart failure Sudden death Exercise test Arrhythmia Myocardial infarction
a b s t r a c t Background: Mortality due to cardiovascular disease is often associated with ventricular arrhythmias. Nowadays, patients with cardiovascular disease are more encouraged to take part in physical training programs. Nevertheless, high-intensity exercise is associated to a higher risk for sudden death, even in apparently healthy people. During an exercise testing (ET), health care professionals provide patients, in a controlled scenario, an intense physiological stimulus that could precipitate cardiac arrhythmia in high risk individuals. There is still no clinical or statistical tool to predict this incidence. The aim of this study was to develop a statistical model to predict the incidence of exercise-induced potentially life-threatening ventricular arrhythmia (PLVA) during high intensity exercise. Methods and results: 6415 patients underwent a symptom-limited ET with a Balke ramp protocol. A multivariate logistic regression model where the primary outcome was PLVA was performed. Incidence of PLVA was 548 cases (8.5%). After a bivariate model, thirty one clinical or ergometric variables were statistically associated with PLVA and were included in the regression model. In the multivariate model, 13 of these variables were found to be statistically significant. A regression model (G) with a X2 of 283.987 and a p b 0.001, was constructed. Significant variables included: heart failure, antiarrhythmic drugs, myocardial lower-VD, age and use of digoxin, nitrates, among others. Conclusion: This study allows clinicians to identify patients at risk of ventricular tachycardia or couplets during exercise, and to take preventive measures or appropriate supervision. © 2016 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Worldwide, cardiovascular diseases are the cause of non-traumatic sudden death, due to ventricular arrhythmia in more than 50% of cases [1,2]. Intense physical exercise is associated with the appearance several kinds of arrhythmia, with a rate of premature ventricular complexes during an exercise testing (ET) around 70% of patients with heart disease, and 34% in apparently healthy people [3]. Furthermore, the presence of frequent ventricular ectopy in a stress testing is also associated with a higher risk of all causes mortality, sudden death or acute coronary syndrome [4–8]. Consequently, the attending physician needs to identify which patients are at high risk of exercise-related arrhythmia but, at present there is no stratification tool to predict it [9–14]. The aim
☆ This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. ⁎ Corresponding author at: Juan Badiano 01, Colonia Seccion XVI, Tlalpan, Mexico City 14080, Mexico. E-mail address: [email protected] (I.-L. Hermes).
http://dx.doi.org/10.1016/j.ijcard.2016.06.196 0167-5273/© 2016 Elsevier Ireland Ltd. All rights reserved.
of this study was to find the statistical association among baseline characteristics of patients with cardiovascular disease with potentially life-threatening ventricular arrhythmia (PLVA) during intense exercise. 2. Methods We studied a database of over 6000 patients with cardiovascular disease, who underwent exercise testing in a cardiology referral center, between 2006 and 2015. Clinical files without complete information were ruled out. Patients with more than one exercise testing were included. Medication was not suspended before exercise and all patients signed a consent form. Every ET was performed using a Balke ramp protocol, and tests were stopped according to symptoms or an absolute indication. A Schiller CS-200 © device with a Trackmaster treadmill was used to perform all exercise tests. An electrocardiographic signal (ECG) was recorded throughout the test. Blood pressure (BP) was measured every minute during exercise, and at the 1st, 3rd, 5th and 8th min of recovery using a calibrated aneroid sphygmomanometer. The primary composite outcome (dependent variable) was the occurrence of potentially life-threatening ventricular arrhythmia during the ET, defined as: ventricular fibrillation, ventricular tachycardia or couplets. Independent variables included those obtained from the clinical record. The nominal, categorical or numerical variables were presented as frequency (percentage), mean (median) and standard deviation (range) as appropriate.
626
I.-L. Hermes et al. / International Journal of Cardiology 220 (2016) 625–628
Table 1 Demographic characteristics of the patients, according to the presence of potentially life-threatening ventricular arrhythmia Variable
PLVA n = 548 (8%)
No PLVA n = 5867 (92%)
Total n = 6415 (100%)
p value
Age (years) Height (cm) Weight BMI (kg/m2) Female:Male
55 ± 15 165 ± 1 74 ± 14 26.7 ± 5 97 (7%): 451 (9%)*
48 ± 20 161 ± 1 68 ± 18 25.4 ± 5 1317 (93%): 4550 (91%)
49 ± 20 162 ± 1.4 69 ± 18 25.5 ± 5 1414 (22%): 5001 (78%)
b0.001 b0.001 b0.001 b0.001 b0.05*
Diagnosis Ischemic heart disease Congenital heart disease Idiopathic cardiomyopathy Valvular heart disease Mixed diagnosis Healthy individuals Other diagnosis CABG PTCA History of ventricular arrhythmia ICD Heart failure
362 (67%) 10 (2%) 102 (18%) 19 (3%) 32 (6%) 6 (1%) 17 (3%) 95 (17%) 161 (29%) 31 (6%) 26 (5%) 266 (48 %)
3855 (66%) 488 (8%) 425 (7%) 170 (3%) 158 (3%) 301 (5%) 470 (8%) 772 (13%) 1736 (29%) 124 (2%) 143 (2%) 1521 (26%)
4217 (66%) 498 (7%) 527 (8%) 189 (3%) 190 (3%) 307 (5%) 487 (8%) 867 (14%) 1897 (29%) 155 (2%) 169 (3%) 1787 (28%)
b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001
Drugs Betablocker ACE/ARB Nitrates Digoxin Calcium channel blocker Anti-arrhythmic Statins Diuretics Anti-platelet
416 (76%) 446 (81%) 92 (17%) 170 (31%) 86 (16%) 110 (20%) 328 (60%) 248 (45%) 3968 (72%)
3920 (67 %) 3806 (65%) 1263 (22%) 881 (15%) 896 (15%) 509 (9%) 3375 (58%) 1429 (24%) 3960 (67%)
4336 (68%) 4252 (66%) 1357 (21%) 1051 (16%) 982 (15%) 619 (10%) 3703 (58%) 1677 (26%) 4356 (68%)
b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001
* Differences between women and men. Abbreviations: potentially life-threatening ventricular arrhythmia (PLVA), body mass index (BMI), coronary artery bypass grafting (CABG), percutaneous transluminal coronary angioplasty (PTCA), implantable cardioverter-defibrillator (ICD), angiotensin-converting enzyme/angiotensin receptor blocker (ACE/ARB).
Table 2 Relative risk of exercise-induced potentially life-threatening ventricular arrhythmia according to bivariate analysis. Variable
RR
CI 95%
p
Age (N54 years) BMI (≥25.7 kg/m2) Male gender Ischemic heart disease Anterior Wall myocardial infarction Inferior Wall, right ventricle myocardial infarction Re-infarction CABG PTCA Congenital heart disease Dilated cardiomyopathy Valvular heart disease Mixed diagnosis Healthy individuals History of ventricular arrhythmia ICD Heart failure
1.53 1.45 1.35 1.03 0.69 1.47 1.33 1.38 0.99 3.73 3.19 1.81 2.24 0.20 2.8 1.99 2.70
1.28 a 1.82 1.22 a 1.73 1.07 a 1.70 0.86 a 1.25 0.50 a 0.96 1.22 a 1.79 0.97 a 1.82 1.10 a 1.75 0.82 a 1.20 2.22 a 6.27 2.56 a 3.97 1.32 a 2.51 1.52 a 3.31 0.09 a 0.46 1.86 a 4.16 1.3 a 3.06 2.26 a 3.22
b0.001 b0.001 b0.05 ns b0.05 b0.05 ns b0.01 ns b0.01 b0.001 b0.001 b0.001 b0.001 b0.001 b0.01 b0.001
Drugs Betablocker ACE/ARB Nitrates Digoxin Calcium channel blocker Anti-arrhythmic Statins Diuretics Anti-platelet Resting heart rate (b60 bpm) Resting systolic blood pressure (≥140 mm Hg) Resting dyastolic blood pressure (≥90 mm Hg) Resting double product (≥7750 bpm ∗ mm Hg)
1.57 2.37 0.73 2.55 1.03 2.64 1.10 2.57 1.25 0.81 1.64 1.47 1.29
1.28 a 1.92 1.90 a 2.96 0.58 a 0.93 2.10 a 3.10 0.81 a 1.31 2.10 a 3.32 0.92 a 1.32 2.15 a 3.07 1.03 a 1.5 0.66 a 1.0 1.27 a 2.12 1.18 a 1.82 1.08 a 1.53
b0.001 b0.001 b0.01 b0.001 ns b0.001 ns b0.001 b0.05 0.06 b0.001 b0.001 b0.01
Abbreviations: body mass index (BMI), coronary artery bypass grafting (CABG), percutaneous transluminal coronary angioplasty (PTCA), implantable cardioverter-defibrillator (ICD), angiotensin-converting enzyme/angiotensin receptor blocker (ACE/ARB), beats per minute (bpm).
A bivariate analysis was performed including those clinical or paraclinical variables that had been previously recognized as independent risk factors. The variables with a statistically significant association with the outcome (p b 0.05), were recruited into a logisticregression multivariable analysis. Variables were included in the regression model using a step by step forward-Wald modality. The entry value was 0.05 and variables with a p value above 0.1 were removed. The statistical analysis was performed using a SPSS-19 program. Finally, the linear general model (G) was incorporated to the logistic equation to find the probability for PLVA incidence. PLVAðriskÞ ¼
1 : ‐G 1þe
Table 3 Variables included in the multiple logistic regression model.
Constant Heart failure (0/1) Anti-arrhythmic drugs (0/1) Inferior wall, right ventricle MI (0/1) Age (years) Nitrates (0/1) Diuretics (0/1) Anterior Wall MI (0/1) Valvular heart disease (0/1) Resting double product
B
ES
Wald
exp β
p
−4.712 0.610 0.464 0.412 0.014 −0.401 0.362 −0.485 0.483 0.270
0.316 0.114 0.126 0.107 0.003 0.125 0.112 0.171 0.177 0.097
222.924 28.82 13.458 14.791 19.843 10.222 10.424 8.087 7.462 7.697
0.009 1.840 1.590 1.510 1.014 0.670 1.436 0.615 1.621 1.310
b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 0.004 0.006 0.006
0.325
0.120
7.359
1.384
0.007
0.026 0.292 0.272
0.011 0.126 0.126
5.738 5.397 4.673
1.027 1.339 1.313
0.017 0.020 0.031
(≥7750 bpm ∗ mm Hg)
Resting dyastolic blood pressure (≥90 mm Hg)
BMI (kg/m2) ACE/ARB (0/1) Digoxin (0/1)
Abbreviations: myocardial infarction (MI), beats per minute (bpm), body mass index (BMI), angiotensin-converting enzyme/angiotensin receptor blocker (ACE/ARB). Dichotomous variables should be codified as follows: absence (0), presence (1).
I.-L. Hermes et al. / International Journal of Cardiology 220 (2016) 625–628
627
Finally, the probability of PLVA appearance may be calculated as 1 follows: PLVAðriskÞ ¼ ½1þe ‐G ; where G = (− 4.759) + (heart failure × 0.694) + (antiarrhythmic drug (0/1) × 0.491) + (Inferior (0/1) wall or right ventricle myocardial infarction (0/1) × 0.389) + (age (years) ∗ 0.015) + (digoxin (0/1) × 0.402) + (nitrates (0/1) × [−0.405]) + (di(diuretics (0/1) × 0.362) + (valvular heart disease (0/1) × 0.534) + (anterior wall myocardial infarction (0/1) × [− 0.503]) + (resting doubleproduct ≥ 7750 (0/1) × 0.276) + (resting diastolic blood pressure ≥ 90 mm Hg (0/1) × 0.309) + (body mass index (kg/m2) × 0.028) + (ACE or ARB (0/1) × 0.315). Fig. 1 shows the receiver operating characteristic curve (ROC) of the logistic regression model. Thereby, if a 65 years old patient has heart failure, takes digoxin, diuretic, ACE, and has a body mass index of 26 kg/m2, the risk of VPLA during intense exercise is 21.7% (see Fig. 2)
4. Discussion Fig. 1. shows the receiver operating characteristic curve (ROC) of the logistic regression model Area under curve (AUC).
3. Results A total of 6415 ET were analyzed, and 548 (8.5%) outcome events were detected, including 118 (21%) episodes of ventricular tachycardia. Table 1 shows the demographic characteristics according to the presence of PLVA. Bivariate analysis (Table 2) identified several predictive variables for arrhythmia: the presence of congenital heart disease; dilated cardiomyopathy; a history of ventricular arrhythmia; heart failure, and some drugs like antiarrhythmic or diuretics among others. Thirty one independent variables were stochastically associated with arrhythmia by bivariate analysis and were recruited into a logistic multivariable regression model (G). Table 3 shows the 13 variables that set up the statistically significant G-model (X2 = 283,987, p b 0.001). The Hosmer-Lemeshow goodness of fit model at 13rd step was X2 = 8.43, p = 0.39.
Several authors have shown that exercise testing is a helpful tool for patient evaluation, particularly assessing variables like exercise tolerance and the presence of ischemia or electrical stability. We observed that the exercise-induced ventricular potentially life-threatening arrhythmia among patients with heart disease is not low (8.5%). The frequency of exercise-induced arrhythmia is generally explained with at least one of the following mechanisms: sympathetic system activation, myocardial late potentials or re-entry phenomena. In this statistical model, the reason that some of these 13 variables were associated with arrhythmia is clearly base on the presence of myocardial damage (infarction or heart failure) or on electrical conduction heterogeneity (myocardial scar, ischemia or QT dispersion due to some drugs) [10, 11,15–18]. According to reports of many researchers, anterior and inferior wall myocardial infarctions are associated with a higher incidence of ventricular arrhythmia. [19,20] A reasonable explanation for this topographic distribution is founded on the anatomical model presented by professor Torrent-Guasp, who stated that the myocardium of both ventricles is, in fact, the same helical band, disposed as a complex interaction of muscle
Fig. 2. This clinimetric tool, allows physician to predicting ventricular arrhythmia during intense physical exercise according to several clinical variables. In this example, we calculate the risk of VPLA in a 65 years old patient with heart failure, taking digoxin, diuretic, ACE, and with a body mass index of 26 kg/m2. Abbreviations: right ventricle (RV), myocardial infarction (MI), heart rate (HR), blood pressure (BP), body mass index (BMI), angiotensin-converting enzyme (ACE), angiotensin receptor blocker (ARB), potentially life-threatening ventricular arrhythmia (PLVA).
I.-L. Hermes et al. / International Journal of Cardiology 220 (2016) 625–628
628
fibers. The inferior wall is composed of myocardial fibers located in several directions and any damage can produce different sites of re-entry [21,22]. On the other hand, in our study, the anterior wall infarction was related to a lower probability of ventricular arrhythmia, independent of the presence of heart failure. This observation has not yet been described in medical literature, and we do not have any convincing explanation. Habitually, patients with anterior myocardial infarction are treated more intensively than those with inferior infarction, and this could explain that finding. Although Anastassiades et al. stated that patients that underwent cardiac surgery show higher rates of exercise-induced ventricular arrhythmia, in our study this association was stochastically significant only in the bivariate analysis [23]. The nitrates intake was clearly a preventive factor of ventricular arrhythmia, and this could be attributed to a lower incidence of coronary spasm [24]. Obesity has been linked to certain kinds of arrhythmia, particularly atrial fibrillation, mainly due to an increment of pericardial fat. Our results showed that obesity is also associated with ventricular arrhythmia [25–27]. In this study we also found that diuretic and digoxin intake, were strongly correlated with exercise-induced arrhythmia. These two kind of drugs increase sudden death rate by conditioning hydroelectrolytic disorders like hypokalemia or high levels of intracellular concentrations of calcium [28–31]. The ACE/ARB medication is statistically associated with PLVA, and we do not have any plausible explanation at the moment [32]. Study limitations are centered on the retrolective nature of the study, and the lack of a healthy control group. This results show that clinical variables can predict the PLVA in patients with cardiovascular disease, and further research is needed to study other groups of patients. 5. Conclusion The frequency of exercise-induced ventricular arrhythmias in patients with cardiovascular disease is elevated, and strongly associated with several clinical variables. This arrhythmia-predictive calculator may be a useful risk stratification tool for the physician Financial support No financial support has been granted. Conflict of interest The authors declare no conflict of interest. Acknowledgements We would like to thank all people that work in the Cardiac Rehabilitation Service in the National Institute of Cardiology Ignacio Chavez, for their daily labor. References [1] O. Yousuf, J. Chrispin, G. Tomaselli, G. Berger, Clinical management and prevention of sudden cardiac death, Circ. Res. 116 (2015) 2020–2040. [2] R. Chávez Domínguez, J.A. Ramírez Hernández, J.M. Casanova Garcés, Coronary heart disease in Mexico and the clinical epidemiological and preventive relevance, Arch. Cardiol. Mex. 73 (2) (Apr-Jun 2003) 105–114. [3] R.A. Candinas, P.J. Podrid, Evaluation of cardiac arrhythmias by exercise testing, Herz 15 (1) (1990 Feb) 21–27. [4] P.A. Maggioni, G. Zuanetti, et al., Prevalence and prognostic significance of ventricular arrhythmias after acute myocardial infarction in the fibrinolytic era, Circulation 87 (1993) 312–322. [5] J.A. Udall, M.H. Ellestad, Predictive implications of ventricular premature contractions associated with treadmill stress testing, Circulation 56 (1977) 985–989.
[6] R.A. Schweikert, F.J. Pashkow, C.E. Snader, T.H. Marwick, M.S. Lauer, Clinical relevance of exercise-induced ventricular ectopic activity with thallium myocardial perfusion and angiographic coronary artery disease in stable, low-risk populations, Am. J. Cardiol. 83 (1999) 530–534. [7] C.K. Nair, W. Thomson, W.S. Aronow, T. Pagano, K. Ryschon, M.H. Sketch, Prognostic significance of exercise-induced complex ventricular arrhythmias in coronary artery disease with normal and abnormal left ventricular ejection fraction, Am. J. Cardiol. 54 (1984) 1136–1138. [8] R.E. Eckart, M.E. Field, T.W. Hruczkowski, et al., Association of electrocardiographic morphology of exercise-induced ventricular arrhythmia with mortality, Ann. Intern. Med. 149 (2008) 451–460. [9] X. Jouven, A. Zureik, I. Desnos, D. Courbo, P. Ducimetière, Long-term outcome in asymptomatic men with exercise-induced premature ventricular depolarizations, N. Engl. J. 343 (2000) 826–833. [10] G.F. Fletcher, G.J. Balady, Exercise standards for testing and training, Circulation 104 (2001) 1694–1740. [11] A.N. DeMaria, Z. Vera, E.A. Amsterdam, D.T. Mason, R.A. Massumi, Disturbances of cardiac rhythm and conduction induced by exercise: diagnostic, prognostic and therapeutic implications, Am. J. Cardiol. 33 (1974) 732–736. [12] N. Franco, G. Lepera, A. Rolli, Arrhythmic risk stratification of post-myocardial infarction patients, Curr. Opin. Cardiol. 15 (2000) 1–6. [13] R.A. Rourke, Risk stratification after myocardial infarction, Circulation 84 (Suppl. I) (1991) I-177–I-181. [14] J. Frolkis, C. Pothier, E. Blackstone, L. Lauer, Frequent ventricular ectopy after exercise as a predictor of death, N. Engl. J. 348 (2003) 781–790. [15] T.M. Kelly, Exercise testing and training of patients with malignant ventricular arrhythmias, Med. Sci. Sports Exerc. 28 (1996) 53–61. [16] D. Katritsis, J. Camm, Nonsustained ventricular tachycardia: where do we stand? Eur. Heart J. 25 (2004) 1093–1099. [17] J.O. O'Neill, J.B. Young, C.E. Pothier, M.S. Lauer, Severe frequent ventricular ectopy after exercise as a predictor of death in patients with heart failure, J. Am. Coll. Cardiol. 44 (2004) 820–826. [18] G.F. Fletcher, G.J. Balady, et al., Exercise standards for testing and training, Circulation 104 (2001) 1694–1740. [19] M.J. Shah, N.R. Bhatt, A. Dabhi, P.B. Thorat, K. Chudasama, J. Patel, A study of 100 cases of arrhythmias in first week of acute myocardial infarction (AMI) in Gujarat: a high risk and previously undocumented population, J. Clin. Diagn. Res. 8 (1) (Jan 2014) 58–61. [20] M. Olimulder, K. Kraaier, M. Galjee, M. Scholten, J. van Es, L. Wagenaar, J. van der Palen, C. von Birgelen, Infarct tissue characterization in implantable cardioverter defibrillator recipients for primary versus secondary prevention following myocardial infarction: a study with contrast enhancement cardiovascular magnetic resonance imaging, Int. J. Card. Imaging 29 (2013) 169–176. [21] M.J. Kocica, A.F. Corno, V. Lackovic, V.I. Kanjuh, The helical ventricular myocardial band of torrent-Guasp, Semin. Thorac. Cardiovasc. Surg. Pediatr. Card. Surg. Annu. (2007) 52–60. [22] F. Torrent-Guasp, M.J. Kocica, A.F. Corno, M. Komeda, F. Carreras-Costa, A. Flotats, J. Cosin-Aguillar, H. Wen, Towards new understanding of the heart structure and function, Eur. J. Cardiothorac. Surg. 27 (2) (Feb 2005) 191–201. [23] L.C. Anastassiades, A.G. Antonopoulos, A.A. Petsas, The effect of coronary revascularization on exercise-induced ventricular ectopic activity, Eur. Heart J. 8 (Suppl. D) (Aug 1987) 75–78. [24] K.L. Looi, A. Grace, S. Agarwal, Coronary artery spasm and ventricular arrhythmias, Postgrad. Med. J. 88 (1042) (Aug 2012) 465–471. [25] C.A. Goudis, P. Korantzopoulos, I.V. Ntalas, E.M. Kallergis, D.G. Ketikoglou, Obesity and atrial fibrillation: a comprehensive review of the pathophysiological mechanisms and links, J. Cardiol. (May 7 2015) (pii: S0914-5087(15)00118-5). [26] R.K. Pathak, A. Elliott, M.E. Middeldorp, M. Meredith, A.B. Mehta, R. Mahajan, et al., Impact of cardiorespiratory fitness on arrhythmia recurrence in obese individuals with atrial fibrillation: the CARDIO-FIT study, J. Am. Coll. Cardiol. (Jun 20 2015) (pii: S0735-1097(15)03339-2). [27] A. Sabbag, Y. Sidi, S. Kivity, R. Beinart, M. Glikson, S. Segev, I. Goldenberg, E. Maor, Obesity and exercise-induced ectopic ventricular arrhythmias in apparently healthy middle aged adults, Eur. J. Prev. Cardiol. (Jun 11 2015) (pii: 2047487315591442. [Epub ahead of print]). [28] D.E. Grobbee, A.W. Hoes, Non-potassium-sparing diuretics and risk of sudden cardiac death, J. Hypertens. 13 (1995) 1539–1545. [29] C.E. Leonard, H. Razzaghi, C.P. Freeman, J.A. Roy, C.W. Newcomb, S. Hennessy, Empiric potassium supplementation and increased survival in users of loop diuretics, PLoS One 16 (2014 Jul) 9(7). [30] M. Shenasa, H. Shenasa, N. El-Sherif, Left ventricular hypertrophy and Arrhythmogenesis, Card Electrophysiol. Clin. 7 (2) (2015 Jun) 207–220. [31] L.A. Gonano, M. Sepúlveda, Y. Rico, M. Kaetzel, C.A. Valverde, J. Dedman, A. Mattiazzi, P.M. Vila, Calcium-calmodulin kinase II mediates digitalis-induced arrhythmias, Circ. Arrhythm. Electrophysiol. 4 (6) (2011 Dec) 947–957. [32] W.A. AlJaroudi, M.M. Refaat, R.H. Habib, L. Al-Shaar, M. Singh, R. Gutmann, H.L. Bloom, S.C. Dudley, P.T. Ellinor, S.F. Saba, A.A. Shalaby, R. Weiss, D.M. McNamara, I. Halder, B. London, Genetic risk assessment of defibrillator events investigators. Effect of angiotensin-converting enzyme inhibitors and receptor blockers on appropriate implantable cardiac defibrillator shock in patients with severe systolic heart failure (from the GRADE multicenter study), Am. J. Cardiol. 115 (7) (Apr 1 2015) 924–931, http://dx.doi.org/10.1016/j.amjcard.2015.01.020 Epub 2015 Jan 15.