The diagonal ear lobe crease (Frank\'s sign) as a marker of cardiovascular disease. A systematic review

Letters to the Editor

calculate other important markers of cardiovascular risk like baroreflex sensitivity, at no extra cost and with considerable time saving. Supplementary data to this article can be found online at http:// dx.doi.org/10.1016/j.ijcard.2014.03.179. PH Groop has gained lecture fees from Eli Lilly, Boehringer Ingelheim, Novartis, Genzyme, MSD, and Novo Nordisk, and he is an advisory board member of Boehringer Ingelheim (global), Novartis (global), Abbott (local), and Cebix (global). He has received grants from Eli Lilly, and Roche. The other authors declare that there is no duality of interest associated with this manuscript. The skilled assistance of Anna Sandelin is gratefully acknowledged. The study was supported by the Folkhälsan Research Foundation, Helsinki University Central Hospital Research Funds (EVO), the Wilhelm and Else Stockmann Foundation, the Waldemar von Frenckell Foundation, the Liv och Hälsa Foundation, the Signe and Ane Gyllenberg Foundation, the Finnish Medical Society (Finska Läkaresällskapet), the Novo Nordisk Foundation and the Academy of Finland. LB. researched data, performed statistical analyses, wrote part of the analysis software and wrote the manuscript. D.G. researched data, performed statistical analyses, contributed to discussion, and reviewed and edited the manuscript. V-P.M. performed statistical analyses, wrote part of the analysis software, contributed to discussion, and reviewed and edited the manuscript. RM and AdT researched data, contributed to discussion, and reviewed and edited the manuscript. M.R-B. and P.H.G. contributed to discussion and reviewed and edited the manuscript. P.-H.G. is the guarantor of the study.

171

References [1] Vlachopoulos C, Aznaouridis K, O'Rourke MF, Safar ME, Baou K, Stefanadis C. Prediction of cardiovascular events and all-cause mortality with central haemodynamics: a systematic review and meta-analysis. Eur Heart J 2010;31:1865–71. [2] Chirinos JA, Zambrano JP, Chakko S, et al. Aortic pressure augmentation predicts adverse cardiovascular events in patients with established coronary artery disease. Hypertension 2005;45:980–5. [3] Gordin D, Wadén J, Forsblom C, et al. Arterial stiffness and vascular complications in patients with type 1 diabetes: the Finnish Diabetic Nephropathy (FinnDiane) Study. Ann Med 2012;44:196–204. [4] Prince CT, Secrest AM, Mackey RH, Arena VC, Kingsley LA, Orchard TJ. Cardiovascular autonomic neuropathy, HDL cholesterol, and smoking correlate with arterial stiffness markers determined 18 years later in type 1 diabetes. Diabetes Care 2010;33:652–7. [5] Millasseau SC, Kelly RP, Ritter JM, Chowienczyk PJ. Determination of age-related increases in large artery stiffness by digital pulse contour analysis. Clin Sci 2002;103:371–7. [6] Karamanoglu M, Feneley MP. On-line synthesis of the human ascending aortic pressure pulse from the finger pulse. Hypertension 1997;30:1416–24. [7] Chen CH, Nevo E, Fetics B, et al. Estimation of central aortic pressure waveform by mathematical transformation of radial tonometry pressure. Validation of generalized transfer function. Circulation 1997;95:1827–36. [8] Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10. [9] Wilkinson IB, Fuchs SA, Jansen IM, et al. Reproducibility of pulse wave velocity and augmentation index measured by pulse wave analysis. J Hypertens 1998;16:2079–84. [10] Frimodt-Møller M, Nielsen AH, Kamper AL, Strandgaard S. Reproducibility of pulse-wave analysis and pulse-wave velocity determination in chronic kidney disease. Nephrol Dial Transplant 2008;23:594–600.

http://dx.doi.org/10.1016/j.ijcard.2014.03.179 0167-5273/© 2014 Published by Elsevier Ireland Ltd.

Ear lobe crease as a marker of coronary artery disease: A meta-analysis Ersilia Lucenteforte a, Marco Romoli b, Giovanni Zagli c, Gian Franco Gensini d, Alessandro Mugelli a, Alfredo Vannacci a,⁎ a

NEUROFARBA, Department of Neurosciences, Psychology, Drug Research and Child Health, Section of Pharmacology and Toxicology, Center for Integrative Medicine, Florence University, Careggi General Hospital, Florence, Italy b Florence University, Careggi General Hospital, Florence, Italy c Anaesthesiology Unit, Florence University, Careggi General Hospital, Florence, Italy d Department of Critical Care Medicine and Surgery, Florence University, Careggi General Hospital, Florence, Italy

a r t i c l e

i n f o

Article history: Received 25 March 2014 Accepted 2 April 2014 Available online 13 April 2014 Keywords: Coronary artery disease Ear lobe crease Meta-analysis

The possible relation between ear lobe crease (ELC) and coronary artery disease (CAD) was described for the first time by Frank in 1973 in a letter to the editor of the New England Journal of Medicine [1]. The author observed that 19 out of 20 patients with ELC had 1 or more risk factors for CAD, and he suggested that the presumptive relation should have been tested. Since Frank's first publication several trials and necropsy studies have been debating the association between ELC and CAD. Since then ⁎ Corresponding author at: NEUROFARBA, Department of Neurosciences, Psychology, Drug Research and Child Health, University of Florence, viale G. Pieraccini 6, 50139 Florence, Italy. Tel.: + 39 055 4271270; fax: +39 055 4271280. E-mail address: alfredo.vannacci@unifi.it (A. Vannacci).

several articles were published confirming or denying this relationship. A recent review addressing the issue of cutaneous markers associated with atherosclerosis, concluded that it appears prudent to examine ELC in a suspected case of CAD as additional indirect evidence of atherosclerosis [2]. Two reviews also addressed the relationship between diagonal ear-lobe crease and atherosclerosis focusing on dental [3] and maxillofacial [4] implications, and concluded that ELC in combination with patient's medical history, vital signs and panoramic radiograph should be considered in atherosclerosis risk assessment. Despite the notable number of publications available on this subject, a comprehensive review and a meta-analysis of published studies is still lacking. The aim of our study was to provide a thorough evaluation and meta-analysis of existing evidence regarding the possible role of ELC as a marker of CAD, providing also statistical indexes potentially useful for clinicians such as sensitivity, specificity and diagnostic odds ratios. The PUBMED database was searched for published reports, using the following terms: (“ear lobe”[tiab] OR “earlobe”[tiab]) AND “crease*”[tiab]. Three investigators (EL, MR, and AV) independently reviewed titles and abstracts for relevance. Disagreements were resolved by discussion and consensus. Most reports were retrieved for full-text review. Reference list from these reports were hand searched and relevant articles were also retrieved for full-text review. Original articles published in Chinese, English, French, German, Italian, and Spanish and including number of

172

Letters to the Editor

subjects with and without CAD and with and without ELC were considered. The search strategy yielded 120 papers, 36 papers were included in the present work [5–40]. One paper [18] showed results of two different studies, thus final sample included 37 different studies. For each study, patients with CAD were considered as cases and patients without CAD as controls. Depending on the definition of CAD used in the different studies, cases were categorized as angiographically defined (Ang cases), autoptically defined (Aut cases), and clinically defined (Clin cases). Controls were categorized as cardiovascular controls (i.e. patients undergoing cardiovascular examinations for some medical reason but not satisfying diagnostic criteria for CAD–CV controls) and non-cardiovascular controls (i.e. healthy subjects with no symptom or history of cardiovascular disease–NonCV controls). Cases with ELC were considered as true-positives, controls with ELC as false-positives, cases without ELC as false-negatives, and controls without ELC as true-negatives. Sensitivities and specificities with corresponding 95% confidence intervals (CIs) for each study were calculated from the contingency tables of true-positives, false-positives, false-negatives, and true negatives. The random-effect bivariate models [41], which take into account the negative correlation between sensitivity and specificity, were used to calculate summary estimates of sensitivity, specificity, and diagnostic odds ratio (DOR) with corresponding 95% confidence interval (CI). Heterogeneity within studies was assessed by Cochran's Q test. Potential sources of heterogeneity were investigated by carrying out several predefined subgroup analyses: location, year of publication, sample size, definition of patient with CAD and without CAD, and selected patients' characteristics.

Because bivariate models do not converge with less than four studies, few subgroup analyses were performed by the use of univariate models [42]. Heterogeneity across strata was tested using the χ2 statistic. All analyses were performed using Midas [43], Metandi [44], and Metan [45] commands in STATA 11. Eleven studies (5239 subjects, Inline Supplementary Table) were conducted in USA, 14 (19,931 subjects) in Europe, and 12 (6018 subjects) in other countries (Argentina, Australia, Brazil, Canada, Israel, Japan, and Mexico). Twenty-three studies (9599 subjects) were published before 1989, and 14 after (21,598 subjects). Nineteen studies (3414 subjects) included less than 340 subjects, and 18 more (27,774 subjects) (Table 1). With regard to definition of cases and controls, 16 studies (6135 subjects) and two sub-studies (312 patients) used angiographically or autoptically defined cases, while 20 (24,053 subjects) used clinically defined cases. Twenty-five studies (27,399 subjects) used non-cardiovascular controls, while 12 (2789 subjects) and 1 sub-study (200 patients) used cardiovascular controls not satisfying case criteria. One study (1000 patients) included mixed cases and mixed controls, and also performed a subgroup analysis (112 subjects) on angiographycally defined cases and mixed controls. The number of patients refers to all participants included in the present analysis, and in some cases marginally differs from those in the published reports, generally due to missing data for relevant variables. The sensitivities ranged between .21 (95% CI: .07–.42) and .88 (95% CI: .76–.95), while specificities ranged between .13 (95% CI: .03–.34) and .95 (95% CI: .93–.97) (Fig. 1). There was heterogeneity between studies.

Table 1 Pooled sensitivity, specificity, and diagnostic odds ratio (DOR) with corresponding 95% confidence interval (95% CI) overall and in strata of selected studies and patients' characteristics.

All studies Stratum according to study characteristics Location USA Europe Other countries Year of paper publication ≤1989 N1989 Sample size ≤340 subjects N340 subjects Definition of casesa Angiography/autopsy cases Clinical cases Type of controlsa Non-cardiovascular controls (healthy subjects) Cardiovascular controls not satisfying case criteria Type of study a,b Studies comparing Ang/Aut cases versus NonCV controls Studies comparing Ang/Aut cases versus CV controls Studies comparing Clin cases versus NonCV controls Stratum according to patients characteristics Gender Male Female Age ≤60 years N60 years History of diabetesc No Yes History of hypertensionc No Yes a b c

No. of studies (no. of subjects)

Sensitivity (95% CI)

Specificity (95% CI)

DOR (95% CI)

37 (31,188)

0.62 (0.56–0.67)

0.67 (0.61–0.73)

3.27 (2.47–4.32)

11 (5239) 14 (19,931) 12 (6018)

0.64 (0.53–0.73) 0.62 (0.54–0.69) 0.60 (0.49–0.70)

0.64 (0.53–0.74) 0.64 (0.52–0.74) 0.72 (0.62–0.81)

3.18 (1.77–5.71) 2.80 (1.86–4.22) 4.03 (2.48–6.54)

0.589

23 (9599) 14 (21,589)

0.63 (0.57–0.70) 0.59 (0.51–0.67)

0.67 (0.58–0.74) 0.67 (0.58–0.75)

3.48 (2.43–4.99) 2.94 (1.89–4.56)

0.567

19 (3414) 18 (27,774)

0.66 (0.60–0.71) 0.58 (0.50–0.65)

0.61 (0.52–0.68) 0.73 (0.65–0.79)

2.97 (2.00–4.41) 3.67 (2.53–5.31)

0.456

18 (6447) 20 (24,053)

0.64 (0.57–0.69) 0.59 (0.49–0.68)

0.66 (0.55–0.76) 0.68 (0.59–0.75)

3.44 (2.34–5.06) 3.02 (2.03–4.48)

0.653

25 (27,399) 12 (2989)

0.62 (0.55–0.69) 0.57 (0.48–0.66)

0.66 (0.58–0.73) 0.71 (0.57–0.81)

3.20 (2.35–4.37) 3.21 (1.89–5.44)

0.992

6 (3558) 11 (2777) 19 (24,053)

0.68 (0.60–0.74) 0.59 (0.51–0.67) 0.61 (0.51–0.69)

0.62 (0.45–0.76) 0.71 (0.56–0.82) 0.67 (0.58–0.75)

3.38 (2.20–5.20) 3.56 (2.07–6.13) 3.20 (2.16–4.73)

0.978

8 (4770) 4 (425)

0.60 (0.53–0.68) 0.57 (0.48–0.66)

0.60 (0.49–0.71) 0.78 (0.55–0.91)

2.31 (1.40–3.82) 4.66 (1.22–17.91)

0.337

8 (3013) 8 (2221)

0.58 (0.47–0.68) 0.68 (0.54–0.78)

0.81 (0.71–0.89) 0.60 (0.48–0.70)

5.99 (2.48–14.48) 3.14 (1.54–6.40)

0.264

2 (1170) 3 (1197)

0.52 (0.48–0.56) 0.61 (0.52–0.69)

0.74 (0.70–0.77) 0.60 (0.44–0.74)

3.79 (1.17–12.25) 2.35 (0.99–5.57)

0.520

2 (563) 2 (1027)

0.54 (0.47–0.60) 0.53 (0.49–0.58)

0.71 (0.60–0.80) 0.73 (0.69–0.77)

3.01 (0.47–19.33) 3.92 (1.31–11.75)

0.810

The sum does not add up to the total because two studies included mixed patients or performed subgroup analyses. Only one study compared Clin cases versus CV controls. Because bivariate models do not converge with less than four studies, subgroup studies were combined by using univariate models.

p-Value for interaction

Letters to the Editor

173

Fig. 1. Forest plot showing study-specific (box) and overall (diamond) sensitivities and specificities of the studies included in the meta-analysis, with corresponding 95% CIs (in brackets).

The 37 studies showed an overall pooled sensitivity of .62 (95% CI: .56–.67), specificity of .67 (95% CI: .61–.73), and DOR of 3.27 (95% CI: 2.47–4.32) (Table 1). DOR reached 4.03 (95% CI: 2.48–6.54) in studies conducted in countries other than USA and Europe, 3.48 (95% CI: 2.43–4.99), in those published before 1989, and 3.67 (95% CI: 2.53–5.31) in studies including more than 340 subjects (Table 1, Fig. 2a). With regard to definition of cases and controls, DOR was higher in studies including angiography or autopsy cases (3.44; 95% CI: 2.34–5.06), in those including cardiovascular controls not satisfying case criteria (3.21; 95% CI: 1.89–5.44), and in those comparing angiography or autopsy cases versus cardiovascular controls (3.56; 95% CI: 2.07–6.13). Moreover, DOR seemed to be stronger among females (4.66; 95% CI: 1.22– 17.91), in patients with less than 60 years (5.99; 95% CI: 2.48–14.48), without diabetes (3.79; 96% CI: 1.17–12.25), and with hypertension (3.92; 95% CI: 1.31–11.75) (Table 1, Fig. 2b). However, the overlap of 95% CIs and the high p-values for interaction demonstrated that all these differences were not statistically significant. Based on our large meta-analysis, covering more than 31,100 subjects, it would be expected that 62% of patients with CAD have ELC, while 67% of those without CAD would not have any ELC. Moreover, the risk of CAD is 3.3-fold higher in patients with ELC compared to those without ELC. There was a trend toward a higher risk in females, in subjects under 60, in subjects without diabetes and in hypertensive patients, although these differences were not statistically significant. No heterogeneity was found in different strata of selected study characteristics, showing a substantial consistency among them. A minority of the studies included in the present meta-analysis looked at ELC as a marker of CAD calculating sensitivity and specificity

[9,11,13,17,22,23,28,34,37,40]. In particular, the lowest sensitivity (51%) was reported in a study on 415 consecutive patients admitted to a cardiology department for coronary angiography [13]; the same study also reported the highest specificity (85%). The highest sensitivity (85%) was instead found in an investigation on 530 patients undergoing elective surgery [22]; it is worth nothing that the study reporting the lowest specificity (43%), also registered a high sensitivity (78%) [40]. Overall, sensitivity and specificity both ranged from 45 to 50% to 85% in the different studies, with ELC being presented as a sensitive but not specific marker for CAD by some authors, as well as just the opposite from other authors; most investigations reported intermediate values for both indexes. These data are quite consistent with sensitivity and specificity values calculated by us for all 37 studies included in our meta-analysis. In fact, as shown in Fig. 1, sensitivity ranged from 0.21 to 0.88, with a combined value of 0.62, and specificity ranged from 0.13 to 0.95, with a combined value of 0.67. With regard to underlying mechanisms, although the link between ELC and CAD was suggested by many studies and confirmed by the present meta-analysis, a pathophysiologic explanation for this association is still lacking and only hypotheses can be forwarded. Isunado et al. have shown that both earlobe and myocardium are supplied by “end arteries” without the possibility for collateral circulation [46] and Shoenfeld et al. found pre-arteriolar wall thickening and elastic-fiber tears on biopsy sections from the ear lobes at the site of the crease [30]. We might therefore hypothesize that in chronic pathological conditions characterized by atherosclerosis, ear lobe tissues might undergo a diminished blood supply similar to that of myocardium, determining the premature destruction of elastic fibers which becomes manifest as the early creasing and folding seen by the naked eye.

174

Letters to the Editor

Fig. 2. a. Forest plot showing pooled DOR with corresponding 95% CIs overall and according to selected study characteristics. b. Forest plot showing pooled DOR with corresponding 95% CIs overall and according to selected patients' characteristics.

In conclusion, data from our meta-analysis strongly support the hypothesis that ELC could be a marker for CAD, with a sensitivity of 0.62, a specificity of 0.67 and a diagnostic odd ratio of 3.3. Nevertheless, further studies are needed to better define histological, cellular and molecular mechanisms leading to the development of ELC during the course of CAD, as well as prospective population-based studies aimed at defining time of onset and evolution of ELC presence and characteristics with regard to cardiovascular risk factors and CAD development. Appendix A. Supplementary data Supplementary data to this article can be found online at http:// dx.doi.org/10.1016/j.ijcard.2014.04.025. References [1] Frank ST. Aural sign of coronary-artery disease. N Engl J Med 1973;289:327–8. [2] Dwivedi S, Jhamb R. Cutaneous markers of coronary artery disease. World J Cardiol 2010;2:262–9. [3] Friedlander AH, Lopez-Lopez J, VO. E. Diagonal ear lobe crease and atherosclerosis: a review of the medical literature and oral and dental implications. Med Oral Patol Oral Cir Bucal Jan 2012;17(1):e153–9.

[4] Friedlander AH, Scully C. Diagonal ear lobe crease and atherosclerosis: a review of the medical literature and oral and maxillofacial implications. J Oral Maxillofac Surg 2010;68:3043–50. [5] Bahcelioglu M, Isik AF, Demirel B, Senol E, Aycan S. The diagonal ear-lobe crease. As sign of some diseases. Saudi Med J 2005;26:947–51. [6] Brady PM, Zive MA, Goldberg RJ, Gore JM, Dalen JE. A new wrinkle to the earlobe crease. Arch Intern Med 1987;147:65–6. [7] Christiansen JS, Mathiesen B, Andersen AR, Calberg H. Letter: diagonal ear-lobe crease in coronary heart disease. N Engl J Med 1975;293:308–9. [8] Cumberland GD, Riddick L, Vinson R. Earlobe creases and coronary atherosclerosis. The view from forensic pathology. Am J Forensic Med Pathol 1987;8:9–11. [9] Davis TM, Balme M, Jackson D, Stuccio G, Bruce DG. The diagonal ear lobe crease (Frank's sign) is not associated with coronary artery disease or retinopathy in type 2 diabetes: the Fremantle Diabetes Study. Aust N Z J Med 2000;30:573–7. [10] Doering C, Ruhsenberger C, Phillips DS. Ear lobe creases and heart disease. J Am Geriatr Soc 1977;25:183–5. [11] Edston E. The earlobe crease, coronary artery disease, and sudden cardiac death: an autopsy study of 520 individuals. Am J Forensic Med Pathol 2006;27:129–33. [12] Elliott WJ. Ear lobe crease and coronary artery disease. 1,000 patients and review of the literature. Am J Med 1983;75:1024–32. [13] Evrengul H, Dursunoglu D, Kaftan A, et al. Bilateral diagonal earlobe crease and coronary artery disease: a significant association. Dermatology 2004;209:271–5. [14] Farrell RP, Gilchrist AM. Diagonal ear-lobe crease: an independent risk factor in coronary heart disease? Ulster Med J 1980;49:171–2. [15] Gral T, Thornburg M. Earlobe creases in a cohort of elderly veterans. J Am Geriatr Soc 1983;31:134–6.

Letters to the Editor [16] Gutiu I, el Rifai C, Mallozi M. Relation between diagonal ear lobe crease and ischemic chronic heart disease and the factors of coronary risk. Med Interne 1986;24:111–6. [17] Higuchi Y, Maeda T, Guan JZ, Oyama J, Sugano M, Makino N. Diagonal earlobe crease are associated with shorter telomere in male Japanese patients with metabolic syndrome. Circ J 2009;73:274–9. [18] Kaukola S. The diagonal ear-lobe crease, a physical sign associated with coronary heart disease. Acta Med Scand Suppl 1978;619:1–49. [19] Kenny DJ, Gilligan D. Ear lobe crease and coronary artery disease in patients undergoing coronary arteriography. Cardiology 1989;76:293–8. [20] Kirkham N, Murrells T, Melcher DH, Morrison EA. Diagonal earlobe creases and fatal cardiovascular disease: a necropsy study. Br Heart J 1989;61:361–4. [21] Kuon E, Pfahlbusch K, Lang E. The diagonal ear lobe crease for evaluating coronary risk. Z Kardiol 1995;84:512–9. [22] Kuri M, Hayashi Y, Kagawa K, Takada K, Kamibayashi T, Mashimo T. Evaluation of diagonal earlobe crease as a marker of coronary artery disease: the use of this sign in pre-operative assessment. Anaesthesia 2001;56:1160–2. [23] Lesbre JP, Castier B, Tribouilloy C, Labeille B, Isorni C. Frank's sign and coronary disease. Ann Cardiol Angeiol (Paris) 1987;36:37–41. [24] Lichstein E, Chadda KD, Naik D, Gupta PK. Diagonal ear-lobe crease: prevalence and implications as a coronary risk factor. N Engl J Med 1974;290:615–6. [25] Mehta J, Hamby RI. Letter: diagonal ear-lobe crease as a coronary risk factor. N Engl J Med 1974;291:260. [26] Miric D, Fabijanic D, Giunio L, et al. Dermatological indicators of coronary risk: a case–control study. Int J Cardiol 1998;67:251–5. [27] Moncada B, Ruiz JM, Rodriguez E, Leiva JL. Ear-lobe crease. Lancet 1979;1:220–1. [28] Pasternac A, Sami M. Predictive value of the ear-crease sign in coronary artery disease. Can Med Assoc J 1982;126:645–9. [29] Rhoads GG, Klein K, Yano K, Preston H. The earlobe crease—sign of obesity in middle-aged Japanese men. Hawaii Med J 1977;36:74–7. [30] Shoenfeld Y, Mor R, Weinberger A, Avidor I, Pinkhas J. Diagonal ear lobe crease and coronary risk factors. J Am Geriatr Soc 1980;28:184–7. [31] Shrestha I, Ohtsuki T, Takahashi T, Nomura E, Kohriyama T, Matsumoto M. Diagonal ear-lobe crease is correlated with atherosclerotic changes in carotid arteries. Circ J 2009;73:1945–9.

175

[32] Sprague DH. Diagonal ear-lobe crease as an indicator of operative risk. Anesthesiology 1976;45:362–4. [33] Toyosaki N, Tsuchiya M, Hashimoto T, et al. Earlobe crease and coronary heart disease in Japanese. Heart Vessels 1986;2:161–5. [34] Tranchesi Junior B, Barbosa V, de Albuquerque CP, et al. Diagonal earlobe crease as a marker of the presence and extent of coronary atherosclerosis. Am J Cardiol 1992;70:1417–20. [35] Verma SK, Khamesra R, Mehta LK, Bordia A. Ear-lobe crease and ear-canal hair as predictors of coronary artery disease in Indian population. Indian Heart J 1989;41:86–91. [36] Wagner Jr RF, Reinfeld HB, Wagner KD, et al. Ear-canal hair and the ear-lobe crease as predictors for coronary-artery disease. N Engl J Med 1984;311:1317–8. [37] Lamot SB, Lonegro GG, Hernandez M, Lamot JM, Lapresa S, Sobrino E. Diagonal earlobe crease, a sign of coronary artery disease. Medicina (B Aires) 2007;67:321–5. [38] Christoffersen M, Frikke-Schmidt R, Schnohr P, Jensen GB, Nordestgaard BG, Tybjaerg-Hansen A. Visible age-related signs and risk of ischemic heart disease in the general population: a prospective cohort study. Circulation 2014;129:990–8. [39] Kwai-Ping Suen L, Lau YK, Ma HC, Lai KW, Holroyd E. Predictive value of auricular diagnosis on coronary heart disease. Evid Based Complement Altern Med 2012;2012:706249. [40] Shmilovich H, Cheng VY, Rajani R, et al. Relation of diagonal ear lobe crease to the presence, extent, and severity of coronary artery disease determined by coronary computed tomography angiography. Am J Cardiol 2012;109:1283–7. [41] Reitsma JB, Glas AS, Rutjes AW, Scholten RJ, Bossuyt PM, Zwinderman AH. Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews. J Clin Epidemiol 2005;58:982–90. [42] Simel DL, Bossuyt PM. Differences between univariate and bivariate models for summarizing diagnostic accuracy may not be large. J Clin Epidemiol 2009;62:1292–300. [43] MIDAS. Stata module for meta-analytical integration of diagnostic test accuracy studies. Available at http://ideas.repec.org/c/boc/bocode/s456880.html. [44] METANDI. Stata module to perform meta-analysis of diagnostic accuracy. Avalilabe at http://ideas.repec.org/c/boc/bocode/s456932.html. [45] METAN. METAN: Stata module for fixed and random effects meta-analysis. Avalilabe at http://ideas.repec.org/c/boc/bocode/s456798.html. [46] Isunado T, Ito I, Katabira Y, Takahashi G. Histological study on the ear-lobe crease. Hihu 1982;24:352–60.

http://dx.doi.org/10.1016/j.ijcard.2014.04.025 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.

Incidence of sudden cardiac death in congestive heart failure: Chagas disease versus systemic arterial hypertension Henrique Horta Veloso ⁎ Department of Cardiology (Cardioteam), Hospital do Rio, Rio de Janeiro, Brazil

a r t i c l e

i n f o

Article history: Received 11 March 2014 Accepted 2 April 2014 Available online 5 May 2014 Keywords: Chagas disease Heart failure Systemic arterial hypertension Sudden cardiac death

In a recent article, Bestetti et al. [1] compared the outcome of patients with congestive heart failure (CHF) secondary to Chagas cardiomyopathy (n = 244) with those with CHF secondary to systemic arterial hypertension (n = 130). In this interesting research, patients were followed for 33 months and the probability of survival for Chagas disease at 12, 24, 36, 48, and 60 months was 76%, 56%, 45%, 37%, and 29%, respectively; nonetheless, the probability of survival for hypertensive cardiomyopathy at 12, 24, 36, 48, and 60 months was 96%, 92%, 82%, 77%, and 73%, respectively (p b 0.05). Thus, the authors concluded that patients with Chagas heart disease have a poorer outcome in comparison to those with CHF secondary to hypertensive ⁎ Rua Conde de Bonfim 255/505, Tijuca, Rio de Janeiro, RJ, ZIP 20520-051, Brazil.

cardiomyopathy. Unfortunately, in this article, the authors did not mention the mechanism of death in the studied population. The concept of a poorer prognosis of patients with CHF due to Chagas disease in comparison to those with hypertensive cardiomyopathy was already demonstrated in a Brazilian cohort of outpatients followed for 26 months [2]. In this study, the mortality rate was of 45% in the chagasic group (110 of 242 patients) versus 26% in the hypertensive group (45 of 170 patients) (p b 0.0001; relative risk 2.73, 95% confidence interval 1.82 to 4.07). Bestetti et al. [1] discussed that the worse prognosis of patients with CHF secondary to Chagas disease in comparison to those secondary to hypertensive cardiomyopathy could be the consequence of a less aggressive ventricular remodeling process observed in the latter, and also to the lower proportion of patients on, and the lower dose of, beta-blocker therapy in the Chagas group. Both of these explanations can be related to a higher propensity to fatal arrhythmias and sudden cardiac death (SCD) in Chagas cardiomyopathy, but this mechanism of death was not explored. SCD is the main mechanism of death in patients with Chagas disease (62%), followed by progressive heart failure (15%) [3,4]. In chagasic patients with non-sustained ventricular tachycardia, this difference is still more evident (73% of SCD and 9% of heart failure) [5]. In opposition, it is well established that SCD is not a common event in patients with CHF secondary to hypertension. Among patients with SCD and nonischemic dilated cardiomyopathy of several etiologies,