- Email: [email protected]
Earlobe crease and coronary artery disease: Association or coincidence?
CORRESPONDENCE
tells, nursing homes, and emergency medical services would also enable the DNR orders to be portable across state lines, providing an advantage that local efforts cannot. We agree that DNR orders are appropriate in the ED under certain circumstances. But we also agree that the ED is not the best place to have this discussion, and we support efforts to encourage discussion before the need arises. KAREN S. RITCHIE, M.D., M.A. EDWARD B. RUBENSTEIN, M.D. ALAN D. VALENTINE, M.D.
University of Texas M.D. Anderson Cancer Center Houston, Texas 1. Wrenn K, Brody SL. Do-not-resuscitate orders in the emergency department. Am J Med 1992; 92: 129-33. 2. Fox E, Siegter M. Redefining the emergency physician's role in do-not-resuscitate decision-making. Am J Med 1992; 92: 125-8. 3. Youngner SJ. Who defines futility? JAMA 1988; 260: 2904-5. Submitted April 27, 1992, and accepted May 13, 1992
TUMOR NECROSISFACTOR IN PATHOGENESISOF FAMILIAL MEDITERRANEAN FEVER To the Editor: We read with great interest the article of Schattner et al [1] describing the possible participation of t u m o r necrosis factor (TNF) in the pathogenesis of fam i l i a l M e d i t e r r a n e a n fever (FMF). They found that induced TNF production was markedly decreased in patients with acute FMF, and increased in asymptomatic FMF patients to levels over t h o s e of c o n t r o l s u b j e c t s . Schattner et al also suggested that the exhaustion of already highly activated cells might explain the decreased production of TNF during an acute attack. However, in 1989, the same authors had reported that there was no evidence of the involvement of TNF in the pathogenesis of FMF, and that the levels of TNF pro-
duced in vitro by peripheral blood mononuclear cells spontaneously, or in response to any of the inducers used, had not differed significantly in (acute and asymptomatic) FMF patients and controls [2]. These findings support our results of a still ongoing study in asymptomatic FMF patients. We could not demonstrata any difference in the levels of spontaneous and induced TNF production in 10 patients and healthy subjects. There was no change in the study design of the two articles of the authors, and only the number of patients enrolled in the study increased from 42 to 60. Since the authors did not mention the resuits of their previous study and did not discuss the possible explanations in their latter report, we think these conflicting results should be clarified and further investigations made for the demonstration of the participation of TNF or other cytokines in the pathogenesis of FMF.
sponse in acute FMF) led us to increase the number of patients included in the study and to employ a more accurate statistical analysis that resulted in the significant differences in induced TNF production reported in The American Journal of Medicine. However, again, no serum TNF or spontaneous TNF production could be detected. A letter to the editor was submitted to the Lancet in December 1989, reporting the modification in our results. However, it was not accepted for publication. The editor of the Lancet considered that the results should be reported in full, which we did.
MURIEL VERONIQUE SANGUEDOLCE
I read with interestthe articleby Elliottand Karrison [1] on the association of a diagonal earlobe crease and coronary events. Although the firstreport suggesting a connection between the earlobe crease (ELC) and coronary artery disease (CAD) was made in 1973 [2], such an association existed even in ancient Rome, where Emperor Hadrian had C A D and bilateral E L C [3].Since 1973, numerous studies have been reported throughout the world to investigate this association [4].These studies generally indicate the significance of E L C as a marker for CAD, whereas some other studies have claimed that E L C is correlated only with advancing age rather than with the presence of C A D [5]. That age is a major confounding variable is evident in the
AMI SCHATTNER, M.D.
Kaplan Hospital Rehovot, Israel 1. Schattner A, Lachmi M, Hahn T, et al. Tumour necrosis factor in familial Mediterranean fever [letter]. Lancet 1989; 2: 1050.
EARLOBECREASEAND CORONARY ARTERY DISEASE: N~HAT D~L~EN, M.D. AItMET GUL, M.D. ASSOCIATION OR Istanbul University COINCIDENCE? Istanbul, Turkey JEAN LOUIS MEGE, M.D., Ph.D. To the Editor:
Hopital de Saints-Marguerite MarseiUe, France 1. Schattner A, Lachmi M, Livneh A, Pras M, Hahn T. Tumor necrosis factor in familial Mediterranean fever. Am J Med 1991; 90: 434-8. 2. Schattner A, Lachmi M, Hahn T, et al. Tumour necrosis factor in familial Mediterranean fever [letter]. Lancet 1989; 2: 1050. Submitted September 18, 1991, and accepted July 21, 1992
The Reply: Our continued studies on the role of TNF in FMF led us to modify our initial results as published in the Lancet [1]. After this publication, we started evaluating patients first studied during acute disease by reexamining them for the second time several weeks later, when the attack of FMF was long over. The marked differences we found in induced TNF production using all three inducers (namely, a depressed re-
November 1992 The American Journal of Medicine Volume 93
587
CORRESPONDENCE
Figure 1. Acupuncture points on the ear corresponding to different parts of the body including the heart (arrow), shown both diagrammatically (left) and descriptively (right).
study by Elliott and Karrison, which included mostly older patients. T h e y also stated t h a t "most series in the literature have reported an association between the ELC and CAD, although the association has not been found in several smaller or more recent series." A large, recently published clinical study of a Chinese population involving 3,155 persons concluded t h a t ELC was a phenomenon of age and had no predictive significance for CAD in the aged population [6]. The fact that the connection between ELC and CAD has not been convincingly proven from an epidemiologic standpoint does not necessarily reduce the importance of CAD risk factors in atherogenesis. The association of ELC and CAD or CAD risk factors should be further investi588
gated. Meanwhile, ELC should not be used as a clinical diagnostic sign of CAD [5]. Incidentally, Elliott and Karrison cited the "acupuncture theory" of Dang [7] that maps the heart to the pinna of the ear to explain the connection between the earlobe and the coronary arteries. But if one looks at the points on the ear used for ear acupuncture, the area on the earlobe where creases are found is nowhere near the "heart" point (Figure 1). A more provocative theory of connecting ELC with CAD--that of macrophage receptors--was recently proposed by Sapira [8]. Earlobe collagen consists of peptide chains with a repeating triplet of glycine-X-Y whose flexibility permits monomers to wrap around each other, forming a triple helix. Similar stretches of
November 1992 The American Journal of Medicine Volume 93
such collagen-like repeating triplets are also found in acetylcholinesterase, in the complement component Clq, and, most curiously, on the surface of the scavenger macrophage at the receptor used for ingesting atheromatous cholesterol [9]. Such macrophages may be thought of as keeping arterial atheromas within manageable size by such ingestions. The macrophages later secrete the cholesterol with apolipoprotein E, which permits combination with circulating high-density lipoprotein (HDL) as a special HDL-C. This targets the cholesterol for the hepatocyte, which can then capture it and secrete it into the bile, after which it is excreted from the body and can no longer form atheromas [10]. According to the theory of macrophage receptors, if the aging body can no longer maintain
CORRESPONDENCE
r = 0 . 7 6 ; P < 0.01 10000 0 0 0 r v
n
Finland J
800-
U.S
600-
n•
Israel
400-
o
> c) F i g u r e 2. Relationship between the 1976 cardiac-specific death rates and the prevalence of diagonal earlobe creases in nine nations.
earlobe collagen, then it is reasonable to assume that it can no longer maintain the chemically similar macrophage receptor. This would decrease macrophage activity, which serves to control the size of atheromas, and could account for the correlation between ELC and CAD [8]. This theory seems worthy of further exploration. As was stated in a caption for a poster of Confucius deep in thought, you may have your eyes closed while meditating but should always keep your ears open.
200-
J
0(5)
0
0
20
30
40
50
60
70
Percent of Population with ELCs 10. Brown MS, Goldstein JL. Lipoprotein metabolism in the macrophage: implications for cholesterol deposition in atherosclerosis. Annu Rev Biochem
1983; 52: 223-61. Submitted October 14,1991, and accepted July 21,
The Reply:
1992
We thank Dr. Cheng for his interest in our work on the possible relationship between ELC and CAD. Although Petrakis [1] has pointed out the existence of typical diagonal ELC in statues of the Roman Emperor Hadrian, and postulated that he may have had CAD and congestive heart failure, the first teaching of a possible relationship between these TSUNG O. CHENG, M.D. George Washington University p h e n o m e n a was probably by Medical Center Dickerson in the late 1950s, as Washington, D.C. cited (reference 3) in our recent paper [2]. We also gave proper at1. Elliott WJ, Karrison T. Increased all-cause and tribution to the work of Lichstein cardiac morbidity and mortality associated with the et al [3], who had begun a clinical diagonal earlobe crease: a prospective cohort study. Am J Med 1991; 91: 247-54. study of the possible relationship 2. Frank ST. Aural sign of coronary artery disease. N of ELC and CAD in 1970, 3 years Engi J Med 1973; 289: 327-8. before the letter of Frank [4]. The 3. Petrakis NL. Diagonal earlobe creases, type A beabstract of their work [5] was also havior, and the death of Emperor Hadrian. West J Med 1980; 132: 87-91. published earlier. Thus, the issue 4. Cheng TO. The international textbook of cardioloof the original attribution of the gy. New York: Pergamon Press, 1987: 26. association between ELC and 5. Cheng TO. Ear lobe crease and coronary artery heart disease is somewhat more disease. J Am Geriatr Soc 1991; 39: 315-6. 6. Hu B, Zhang J, Hong X. Ear-lobe crease: prevadifficult than has been noted prelence and clinical significance. Chin J Cardio11988; viously in the medical literature. 16: 21. Dr. Cheng's major concern 7. Dan8 CV. The earlobe crease: chromosomes, about our study, it seems, is the acupuncture, and atherosclerosis.Lancet 1984; I: 1083. issue of age as a confounding vari8. Sapira JD. Earlobe creases and macrophage reable in clinical studies of ELC ceptors. South Med J 1991; 84: 537-8. and heart disease. This issue was 9. Brown MS, Goldstein JL. Scavenging for receptors. Nature 1990; 343: 508-9. addressed in Figure 2 of the first
publication of Lichstein et al [3] and has been of major concern ever since. Special mention has been made of this point in nearly every well-done study of ELC since then, including some of our previous work [6]. We disagree with Dr. Cheng's interpretation of our recent study, in which there is concern that our results may have been influenced by age as a confounder. There are two reasons why this is unlikely: first, since we (and others) have recognized that age is an important factor in many studies of ELC, we wished to eliminate this factor as much as possible when designing the study. This is exactly why the individual patients in each cohort were matched for age, sex, and race (all of which were found to be statistically significant covariables in ×2 analyses of the original population studied [6]). This matching process therefore "balances" the cohorts regarding age, sex, and race, and makes it unlikely that any of these factors were confounding covariables. It is true, of course, that by designing the study in this way, we were unable to study much younger and much older patients, due to the low and high prevalence of ELC in these age groups, respectively; this is discussed on page 253, right-hand column, lines 2 to 7, of our article. The second rea-
November 1992 The American Journal of Medicine Volume 93
589
tells, nursing homes, and emergency medical services would also enable the DNR orders to be portable across state lines, providing an advantage that local efforts cannot. We agree that DNR orders are appropriate in the ED under certain circumstances. But we also agree that the ED is not the best place to have this discussion, and we support efforts to encourage discussion before the need arises. KAREN S. RITCHIE, M.D., M.A. EDWARD B. RUBENSTEIN, M.D. ALAN D. VALENTINE, M.D.
University of Texas M.D. Anderson Cancer Center Houston, Texas 1. Wrenn K, Brody SL. Do-not-resuscitate orders in the emergency department. Am J Med 1992; 92: 129-33. 2. Fox E, Siegter M. Redefining the emergency physician's role in do-not-resuscitate decision-making. Am J Med 1992; 92: 125-8. 3. Youngner SJ. Who defines futility? JAMA 1988; 260: 2904-5. Submitted April 27, 1992, and accepted May 13, 1992
TUMOR NECROSISFACTOR IN PATHOGENESISOF FAMILIAL MEDITERRANEAN FEVER To the Editor: We read with great interest the article of Schattner et al [1] describing the possible participation of t u m o r necrosis factor (TNF) in the pathogenesis of fam i l i a l M e d i t e r r a n e a n fever (FMF). They found that induced TNF production was markedly decreased in patients with acute FMF, and increased in asymptomatic FMF patients to levels over t h o s e of c o n t r o l s u b j e c t s . Schattner et al also suggested that the exhaustion of already highly activated cells might explain the decreased production of TNF during an acute attack. However, in 1989, the same authors had reported that there was no evidence of the involvement of TNF in the pathogenesis of FMF, and that the levels of TNF pro-
duced in vitro by peripheral blood mononuclear cells spontaneously, or in response to any of the inducers used, had not differed significantly in (acute and asymptomatic) FMF patients and controls [2]. These findings support our results of a still ongoing study in asymptomatic FMF patients. We could not demonstrata any difference in the levels of spontaneous and induced TNF production in 10 patients and healthy subjects. There was no change in the study design of the two articles of the authors, and only the number of patients enrolled in the study increased from 42 to 60. Since the authors did not mention the resuits of their previous study and did not discuss the possible explanations in their latter report, we think these conflicting results should be clarified and further investigations made for the demonstration of the participation of TNF or other cytokines in the pathogenesis of FMF.
sponse in acute FMF) led us to increase the number of patients included in the study and to employ a more accurate statistical analysis that resulted in the significant differences in induced TNF production reported in The American Journal of Medicine. However, again, no serum TNF or spontaneous TNF production could be detected. A letter to the editor was submitted to the Lancet in December 1989, reporting the modification in our results. However, it was not accepted for publication. The editor of the Lancet considered that the results should be reported in full, which we did.
MURIEL VERONIQUE SANGUEDOLCE
I read with interestthe articleby Elliottand Karrison [1] on the association of a diagonal earlobe crease and coronary events. Although the firstreport suggesting a connection between the earlobe crease (ELC) and coronary artery disease (CAD) was made in 1973 [2], such an association existed even in ancient Rome, where Emperor Hadrian had C A D and bilateral E L C [3].Since 1973, numerous studies have been reported throughout the world to investigate this association [4].These studies generally indicate the significance of E L C as a marker for CAD, whereas some other studies have claimed that E L C is correlated only with advancing age rather than with the presence of C A D [5]. That age is a major confounding variable is evident in the
AMI SCHATTNER, M.D.
Kaplan Hospital Rehovot, Israel 1. Schattner A, Lachmi M, Hahn T, et al. Tumour necrosis factor in familial Mediterranean fever [letter]. Lancet 1989; 2: 1050.
EARLOBECREASEAND CORONARY ARTERY DISEASE: N~HAT D~L~EN, M.D. AItMET GUL, M.D. ASSOCIATION OR Istanbul University COINCIDENCE? Istanbul, Turkey JEAN LOUIS MEGE, M.D., Ph.D. To the Editor:
Hopital de Saints-Marguerite MarseiUe, France 1. Schattner A, Lachmi M, Livneh A, Pras M, Hahn T. Tumor necrosis factor in familial Mediterranean fever. Am J Med 1991; 90: 434-8. 2. Schattner A, Lachmi M, Hahn T, et al. Tumour necrosis factor in familial Mediterranean fever [letter]. Lancet 1989; 2: 1050. Submitted September 18, 1991, and accepted July 21, 1992
The Reply: Our continued studies on the role of TNF in FMF led us to modify our initial results as published in the Lancet [1]. After this publication, we started evaluating patients first studied during acute disease by reexamining them for the second time several weeks later, when the attack of FMF was long over. The marked differences we found in induced TNF production using all three inducers (namely, a depressed re-
November 1992 The American Journal of Medicine Volume 93
587
CORRESPONDENCE
Figure 1. Acupuncture points on the ear corresponding to different parts of the body including the heart (arrow), shown both diagrammatically (left) and descriptively (right).
study by Elliott and Karrison, which included mostly older patients. T h e y also stated t h a t "most series in the literature have reported an association between the ELC and CAD, although the association has not been found in several smaller or more recent series." A large, recently published clinical study of a Chinese population involving 3,155 persons concluded t h a t ELC was a phenomenon of age and had no predictive significance for CAD in the aged population [6]. The fact that the connection between ELC and CAD has not been convincingly proven from an epidemiologic standpoint does not necessarily reduce the importance of CAD risk factors in atherogenesis. The association of ELC and CAD or CAD risk factors should be further investi588
gated. Meanwhile, ELC should not be used as a clinical diagnostic sign of CAD [5]. Incidentally, Elliott and Karrison cited the "acupuncture theory" of Dang [7] that maps the heart to the pinna of the ear to explain the connection between the earlobe and the coronary arteries. But if one looks at the points on the ear used for ear acupuncture, the area on the earlobe where creases are found is nowhere near the "heart" point (Figure 1). A more provocative theory of connecting ELC with CAD--that of macrophage receptors--was recently proposed by Sapira [8]. Earlobe collagen consists of peptide chains with a repeating triplet of glycine-X-Y whose flexibility permits monomers to wrap around each other, forming a triple helix. Similar stretches of
November 1992 The American Journal of Medicine Volume 93
such collagen-like repeating triplets are also found in acetylcholinesterase, in the complement component Clq, and, most curiously, on the surface of the scavenger macrophage at the receptor used for ingesting atheromatous cholesterol [9]. Such macrophages may be thought of as keeping arterial atheromas within manageable size by such ingestions. The macrophages later secrete the cholesterol with apolipoprotein E, which permits combination with circulating high-density lipoprotein (HDL) as a special HDL-C. This targets the cholesterol for the hepatocyte, which can then capture it and secrete it into the bile, after which it is excreted from the body and can no longer form atheromas [10]. According to the theory of macrophage receptors, if the aging body can no longer maintain
CORRESPONDENCE
r = 0 . 7 6 ; P < 0.01 10000 0 0 0 r v
n
Finland J
800-
U.S
600-
n•
Israel
400-
o
> c) F i g u r e 2. Relationship between the 1976 cardiac-specific death rates and the prevalence of diagonal earlobe creases in nine nations.
earlobe collagen, then it is reasonable to assume that it can no longer maintain the chemically similar macrophage receptor. This would decrease macrophage activity, which serves to control the size of atheromas, and could account for the correlation between ELC and CAD [8]. This theory seems worthy of further exploration. As was stated in a caption for a poster of Confucius deep in thought, you may have your eyes closed while meditating but should always keep your ears open.
200-
J
0(5)
0
0
20
30
40
50
60
70
Percent of Population with ELCs 10. Brown MS, Goldstein JL. Lipoprotein metabolism in the macrophage: implications for cholesterol deposition in atherosclerosis. Annu Rev Biochem
1983; 52: 223-61. Submitted October 14,1991, and accepted July 21,
The Reply:
1992
We thank Dr. Cheng for his interest in our work on the possible relationship between ELC and CAD. Although Petrakis [1] has pointed out the existence of typical diagonal ELC in statues of the Roman Emperor Hadrian, and postulated that he may have had CAD and congestive heart failure, the first teaching of a possible relationship between these TSUNG O. CHENG, M.D. George Washington University p h e n o m e n a was probably by Medical Center Dickerson in the late 1950s, as Washington, D.C. cited (reference 3) in our recent paper [2]. We also gave proper at1. Elliott WJ, Karrison T. Increased all-cause and tribution to the work of Lichstein cardiac morbidity and mortality associated with the et al [3], who had begun a clinical diagonal earlobe crease: a prospective cohort study. Am J Med 1991; 91: 247-54. study of the possible relationship 2. Frank ST. Aural sign of coronary artery disease. N of ELC and CAD in 1970, 3 years Engi J Med 1973; 289: 327-8. before the letter of Frank [4]. The 3. Petrakis NL. Diagonal earlobe creases, type A beabstract of their work [5] was also havior, and the death of Emperor Hadrian. West J Med 1980; 132: 87-91. published earlier. Thus, the issue 4. Cheng TO. The international textbook of cardioloof the original attribution of the gy. New York: Pergamon Press, 1987: 26. association between ELC and 5. Cheng TO. Ear lobe crease and coronary artery heart disease is somewhat more disease. J Am Geriatr Soc 1991; 39: 315-6. 6. Hu B, Zhang J, Hong X. Ear-lobe crease: prevadifficult than has been noted prelence and clinical significance. Chin J Cardio11988; viously in the medical literature. 16: 21. Dr. Cheng's major concern 7. Dan8 CV. The earlobe crease: chromosomes, about our study, it seems, is the acupuncture, and atherosclerosis.Lancet 1984; I: 1083. issue of age as a confounding vari8. Sapira JD. Earlobe creases and macrophage reable in clinical studies of ELC ceptors. South Med J 1991; 84: 537-8. and heart disease. This issue was 9. Brown MS, Goldstein JL. Scavenging for receptors. Nature 1990; 343: 508-9. addressed in Figure 2 of the first
publication of Lichstein et al [3] and has been of major concern ever since. Special mention has been made of this point in nearly every well-done study of ELC since then, including some of our previous work [6]. We disagree with Dr. Cheng's interpretation of our recent study, in which there is concern that our results may have been influenced by age as a confounder. There are two reasons why this is unlikely: first, since we (and others) have recognized that age is an important factor in many studies of ELC, we wished to eliminate this factor as much as possible when designing the study. This is exactly why the individual patients in each cohort were matched for age, sex, and race (all of which were found to be statistically significant covariables in ×2 analyses of the original population studied [6]). This matching process therefore "balances" the cohorts regarding age, sex, and race, and makes it unlikely that any of these factors were confounding covariables. It is true, of course, that by designing the study in this way, we were unable to study much younger and much older patients, due to the low and high prevalence of ELC in these age groups, respectively; this is discussed on page 253, right-hand column, lines 2 to 7, of our article. The second rea-
November 1992 The American Journal of Medicine Volume 93
589