J Clin Pathol 1996;49:403-406 403 Necropsy study of the association between

sudden cardiac , cardiac isoenzymes and J Clin Pathol: first published as 10.1136/jcp.49.5.403 on 1 May 1996. Downloaded from contraction band

D J Hopster, C M Milroy, J Bums, N B Roberts

Abstract carried out for the .' In such cases, Aims-To assess whether a quantitative however, occlusive coronary artery atheroma is analysis of myocardial contraction bands often the only finding. This is because the could aid the postmortem identification of time interval involved is too short to allow the early myocardial , especially if development of identifiable infarction. Al- used in conjunction with cardiac iso- though many and various methods have been enzyme activities. described to identify early infarction, at present Methods-Sixty four coroner's necropsies there is no one single technique in regular use were grouped by gross and histological that can recognise these early signs ofcell injury findings into 26 cases of definite non-car- and death. Recently, however, three of us have diac death, 12 cases ofdefinite myocardial been able to demonstrate that the measurement infarction and 26 cases in which there was ofcardiac isoenzyme activities can aid the iden- occlusive coronary artery atheroma, but tification of early myocardial infarction.2 The no demonstrable evidence of infarction. present study was undertaken to assess whether Using multiple sections of left ventricular a quantitative analysis of myocardial con- myocardium stained with Heidenhain's traction bands could improve the accuracy of iron haematoxylin, the number of myo- such a diagnosis, when used in conjunction cardial cells containing contraction bands with cardiac isoenzyme activities. per unit area was quantified. The results Contraction band necrosis, also known as were analysed statistically using logistic myofibrillar degeneration3 and coagulative regression, and were then compared and ,4 is a type of that is combined with results from the statistical identifiable as early as two minutes after ir- analysis of postmortem cardiac iso- reversible cell injury has occurred.5 The lesion

enzymes that had recently been under- is produced by a disruption of the linear ar- http://jcp.bmj.com/ taken on the same cases. rangement of myofibrils via hypercontraction, Results-The number of cells containing leaving dense aggregates of both thick and thin contraction bands per unit area was higher filaments that are separated by clearer areas in cases of definite myocardial infarction now devoid of myofibrils.6 These aggregates compared with those of non-cardiac are seen as faint transverse eosinophilic bands . In addition, cases ofocclusive cor- on routine staining with haematoxylin and

onary artery atheroma only could be iden- eosin, but are much easier to identify (fig 1) on September 29, 2021 by guest. Protected copyright. tified, indicating the presence of early using special histochemical stains, such as myocardial infarction. The accuracy of Masson's or Gomori's trichrome,7 Luxol fast this identification could be improved by blue,8 Mallory's phosphotungstic acid haema- combining these results with the results toxylin,9 and Heidenhain's iron haematoxylin. from the statistical analysis ofpostmortem Other features of this process include an inter- Department of cardiac stitial , isoenzymes. mononuclear cell proliferation, relative University of Conclusion-The quantitative assessment sparing of stroma, vessels and nerves, a lack Liverpool of myocardial contraction band necrosis of a neutrophilic infiltrate, and the frequent D J Hopster can provide useful additional information appearance of a granularity within the muscle Sub-Department of in cases of sudden death, where myo- cells. This latter phenomenon corresponds to cardial infarction is suspected but not the mineralisation oftranslocated mitochondria J Bums identified on routine histological staining. at an ultrastructural level. Electron microscopy Department of The value of the information obtained is also demonstrates myofilaments out ofregister, increased when used in conjunction with aggregating to produce contraction bands at a N B Roberts the postmortem measurement of cardiac microscopic level. With time, there is loss of Department of isoenzyme activities. nuclear structure and degenerative cytoplasmic Forensic Pathology, (J Clin Pathol 1996;49:403-406) changes, and healing eventually occurs via University of Sheffield phagocytosis of debris and subsequent fib- C M Milroy Keywords: contraction band necrosis, cardiac iso- rosis. 0 enzymes, sudden death, myocardial infarction, post- Correspondence to: Dr J Bums, mortem. Sub-Department of Forensic Pathology, Royal Liverpool University Methods Hospital, The identification of early myocardial in- Necropsies carried out for Her Majesty's cor- PO Box 147, Liverpool L69 3BX. farction is a common problem in necropsy oner were grouped into non-cardiac deaths, Accepted for publication pathology, as sudden cardiac death is the most deaths from definite myocardial infarction and 16 January 1996 common finding in postmortem examinations deaths with occlusive coronary atheroma but 404 Hopster, Milroy, Burns, Roberts

containing contraction bands and remeasuring the area. In total 64 cases were collected for which the full biochemical data were included and the number of cells/mm2 containing J Clin Pathol: first published as 10.1136/jcp.49.5.403 on 1 May 1996. Downloaded from contraction bands could be quantified. Additionally, in all cases the postmortem ex- amination interval (PMI) was known, which was the time in hours between death and nec- ropsy.

Results The 64 cases had already been classified into three groups from the results of the necropsy, in combination with the histological evidence. There were 26 cases of non-cardiac death, 12 cases of definite myocardial infarction, and 26 cases of occlusive coronary artery atheroma without evidence ofinfarction-that is, possible myocardial infarction. The section with the ,*.. -. highest number of cells/mm2 containing con- I * t ...er..r. traction bands was selected from the multiple Figure 1 Section of myocardium showing numerous transverse contraction bands, sections examined in each of the 64 cases, and pronounced in the upper half of the figure (Heidenhain s iron haematoxylin). was used for statistical analysis. The number of cells/mm2 containing contraction bands varied from 0-02 to 12-25 with a mean of 1 87, and Table 1 Non-cardiac deaths (n =26) with no definite macroscopic evidence of in- decreased with increases in the PMI, which farction. Multiple blocks of paraffin wax em- itself varied from three to 70 Bands/mm2 PMI hours, with a bedded left ventricular myocardium were then mean of 25 hours. 0-02 26 stained with haematoxylin and eosin and ex- The results were analysed 0-02 30 statistically using 0-03 22 amined to identify or exclude histological evi- logistic regression and discriminant analysis,'2 0-14 6 dence of infarction. At necropsy, blood from both of which gave similar results. 0-03 26 Logistic 0-06 15 the right atrium (into which the coronary sinus regression, however, is the preferred technique, 0-07 16 drains), blood from a peripheral vein and peri- as it does not assume multivariate normality of 0-10 20 0-04 70 cardial fluid were collected. The following car- the predictor variables and leads to a direct 0-15 21 diac isoenzymes were measured in each case: estimate of the probability of death from a 0-05 62 1-25 3 aspartate aminotransferase; hydroxybutarate cardiac cause without resorting to Bayes' the- http://jcp.bmj.com/ 0-42 10 dehydrogenase; total creatine phosphokinase orem'3-'5 which permits the combination of 0-25 21 0-27 23 (CPK); and the creatine phosphokinase iso- many individual test results to give a final odds 2-01 4 enzymes CPK-MM, CPK-MB and CPK-BB. ratio. It predicts a future event based upon 2-22 4 0-36 26 The creatine phosphokinase isoenzymes were associated events of the past and although the 2-68 4 measured by quantitative electrophoresis and principle is simple, the implementation of the 0-39 25 0-20 50 the enzyme activity by a standard procedure theorem requires much tedious calculation. 0-18 56 using a centrifugal analyser. The total CPK Using logistic regression, L= logit(p) - ln(p/ on September 29, 2021 by guest. Protected copyright. 0-16 66 0-31 56 for each sample was measured, then a second [l-p]), where p=exp(l)+exp(l); p is the es- 0-54 51 sample with activity up to 1000 IU/l was used timated probability of the death being due to 0-77 37 for electrophoresis. Samples with activity myocardial infarction, and a "predicted pos- greater than 1000 IU/l were diluted accordingly itive" result is indicated by values of p>0 5. Table 2 Possible infarct using isotonic saline. The results of this bio- Tables 1, 2 and 3 contain the relevant data (n = 26) chemical analysis have been published pre- collected in this investigation and also show viously in a study from our departments.2 the resulting p values from the biochemical BandsImm2 PMI For the purposes of this investigation, a sec- analysis of the same cases. When the number 0-02 8 tion of each block was stained with Hei- of cells/mm2 containing contraction bands and 0-08 3 0-03 15 denhain's iron haematoxylin and the number the PMI were combined, the results in table 4 0-04 14 of cells/mm2 containing contraction bands were obtained. In this model, 92% (24/26) of 0-03 24 0-15 13 (bands/mm2) was quantified, using a method the non-cardiac deaths and 67% (eight of 12) 0-12 20 that has been described previously," where the ofthe deaths due to myocardial infarction were 012 21 0 12 24 total number of cells containing contraction correctly classified. In addition, 27% (seven of 0-13 36 bands was counted in each section, using a 26) of the cases of possible myocardial in- 0-17 40 0 52 14 Leitz dialux microscope with a 40 x objective farction could be reclassified as cases of myo- 0-12 61 lens. The area of each of these sections was cardial infarction. 092 10 4-46 3 then measured using a Kontron mop-30 image The biochemical data had also been analysed 4-84 3 analyser system connected to a digitising tablet. using logistic regression, and the best model 1-20 14 0-85 24 All of these results were obtained without prior was found to be a combination of total CPK 5-97 5 knowledge of the group to which each case had in the right atrial blood, percentage of iso- 1-19 26 4.73 11 been assigned. The reproducibility of these enzyme MB in the peripheral venous blood, 0-85 69 results was assessed by measuring the intra- percentage of isoenzyme MB in the pericardial 3-49 24 5-50 16 observer error, which was found to be less than fluid, percentage ofisoenzyme MM in the peri- 7-26 15 5%, and was established by taking 10 sections cardial fluid, and the PMI. The results are 2-82 39 at random, and recounting the number of cells shown in table 5. Using this model, 92% (24/ Association between sudden cardiac death, isoenzymes and contraction band necrosis 405

Table 3 Definite Table 4 Model using PMI and number of cellslmm2 and in subsequent experiments by others, con- infarction (n = 12) containing contraction bands traction band necrosis was produced by re- Bandslmm' PMI Classification Predicted Predicted perfusion following temporary ischaemia.22-25 non-infarct infarct It also occurred at the boundary of an area J Clin Pathol: first published as 10.1136/jcp.49.5.403 on 1 May 1996. Downloaded from 0 03 45 0 07 27 Non-cardiac 24 2 of coagulative necrosis following permanent 0 51 16 Infarct 4 8 ischaemia. Other studies have shown that in- 0-46 26 Possible infarct 19 7 12 25 10 creased catecholamine concentrations, either 637 19 Logit(p)=- 3503 + 24251og(PMI) + 22831og(bands/mm'). exogenous or endogenous, can produce con- 3*43 36 10-84 16 traction band necrosis,102128 which has con- 10-74 18 sequently been reported in the following: 942 21 Table 5 Model using PMI, total CPK and percentage 10 33 20 isoenzymes administration ofpolymxyin, neomycin, plasm- 541 38 ocid, streptokinase, trypsin, ficin, papain, co- Classification Predicted Predicted non-infarct infarct balt, and corticoids; potassium and magnesium deficiency; hypothermia; oxygen toxicity; elec- Non-cardiac 24 2 Infarct 8 4 tric shock; haemorrhagic shock; intra-cranial Possible infarct 16 10 haemorrhage; near drowning; cocaine abuse; and severe emotional Logit(p) = 5-529-0-6531og(PMI) pheochromocytomas; + 1-0031og (total right atrial CPK) stress.1029 31 These conditions are not relevant -0 187 (%isoenzyme MB in peripheral venous blood) -0-162 (%isoenzyme MB in pericardial fluid) to the current study though, as the majority -0 078 (%isoenzyme MM in pericardial fluid) can be excluded from the clinical details alone, and the remainder eliminated by routine post- mortem examination, biochemistry and tox- Table 6 Model combining results from tables 4 and 5 icology. Of more importance, however, is the Classification Predicted Predicted fact that contraction band necrosis can be pro- non-infarct infarct duced by resuscitation, via both catecholamine Non-cardiac 22 4 administration and defibrillation,32 thereby Infarct 2 10 making their quantification unhelpful in such Possible infarct 11 15 cases. In this study all cases were of sudden death without resuscitation. However, even in those cases where resuscitation has been at- 26) of the non-cardiac deaths and 33% (four tempted, there is likely to be other evidence of of 12) ofthe deaths due to myocardial infarction myocardial infarction, such as an electro- were correctly classified. In addition, 38% (10/ cardiogram. 26) of the cases of possible myocardial in- In 1954, Schlesinger and Reimer identified farction could be reclassified as cases of myo- the presence of contraction band necrosis in cardial infarction. 51 of 80 hearts showing recent coagulative It is possible to combine the results of the necrosis, compared with 45 of278 hearts show- http://jcp.bmj.com/ contraction band quantification and the bio- ing previous and five of 213 normal chemical analysis, as the same cases have been hearts.33 In addition, the present study showed used in each, and the same statistical analyses that the extent ofthe contraction band necrosis applied. In doing so, p>05 in either study was greatest in those hearts where coagulation (or both studies) has been used to indicate a necrosis was also identified. Bouchardy and "predicted positive" result. The results of this Majno, in 1973, seem to have been the first to combination are shown in table 6. In this suggest that contraction band necrosis rep- on September 29, 2021 by guest. Protected copyright. model, 85% (22/26) of the non-cardiac deaths resented a discrete but important sign of and 83% (10/12) of the deaths due to myo- cellular in early infarcts and that con- cardial infarction were correctly identified. In sequently it could be a useful adjunct in the addition, 58% (15/26) of the cases of possible diagnosis of early myocardial ischaemia.34 In myocardial infarction could be reclassified as 1975, Baroldi reviewed the functional sig- cases of myocardial infarction. nificance of different types of myocardial nec- rosis.35 In 100 cases of definite myocardial infarction, contraction band necrosis was docu- Discussion mented in a total of 96% of cases, and in 200 Contraction band necrosis has been known cases ofsudden death with obstructive coronary about for several decades, although various artery atheroma, but no demonstrable in- synonymous terms have been in use. In 1904, farction, contraction band necrosis was the only Smith noted that focal myocytolysis occurred lesion in 67% of cases. In 16% of these cases, at the border of myocardial infarcts16 and in coagulative necrosis was also identified, 76% 1906 Pearce observed the occurrence ofhyaline of which also contained contraction band nec- necrosis in rabbit myocardium after multiple rosis. Later, in 1975, Reichenbach and Moss injections of adrenaline. 7 The actual term published a study of myocardial cell necrosis "contraction bands" was first introduced by and sudden death.36 Of 50 cases of sudden Caulfield and Klionsky in 1959.18 However, death, contraction band necrosis was found in they were not referring to contraction band 88%. Furthermore, 16 of these cases showed necrosis proper, but to the thickening of Z coagulative necrosis. However, severe coronary bands that they observed in rabbit myocardium artery atheroma was rare. In 1979, Baroldi after 20 minutes of coronary artery ligation. It et al published another postmortem study of was Jennings et al who first recognised that sudden coronary death.4 Of208 cases ofsudden the presence of contraction bands indicated a death, contraction band necrosis was the specific type of necrosis.'9 21 In their studies, unique lesion in 72% of cases and was seen 406 Hopster, Milroy, Bums, Roberts

8 Arnold G, Fischer R. Myocardial "contraction bands" (let- in a total of 86% of cases, when those with ter). Hum Pathol 1987;18:99-101. coagulative necrosis were included. They con- 9 Stahl E. Screening of myocardial contraction bands: A comparison between two histological staining methods. cluded tat contraction band necrosis may be Forensic Sci Int 1990;45:151-7. either the first sign of myocardial infarction or 10 Reichenbach DD, Benditt EP. Catecholamines and car- J Clin Pathol: first published as 10.1136/jcp.49.5.403 on 1 May 1996. Downloaded from a diomyopathy: The and potential importance primary sympathetic disorder linked with of myofibrillar degeneration. Hum Pathol 1970;1:125-50. ventricular fibrillation. In 1970, Wu et al pub- 11 Adomian GE, Laks MM, Billingham ME. The incidence lished a and significance of contraction bands in endomyocardial quantitative analysis of contraction biopsies from normal human hearts. Am Heartjt 1978;95: band necrosis and infarction in cases with near 348-51. 12 Altman DG. Relation between several variables. In: Practical normal arteries.37 They found that in cases of statisticsformedical research. London: Chapman Hall, 1991: infarction, an average area of86% was occupied 351-60. 13 Bayes T. (1764). An essay towards solving a problem in the by contraction band necrosis. doctrine of chances. Philosophical Transactions of the The results of our study show that eight Royal Society ofLondon for 1763, 53, 370-418. Reprinted with Barnard (1958), Pearson and Kendal (1970), 131- of 12 cases of definite infarction contained 153. contraction band necrosis. Although in four 14 Pearson ES, Kendall MG (eds). Studies in the history of cases was statistics and probability. London: Charles Griffin, 1970. contraction band necrosis not iden- 15 Barnard, G.A. (1958) Thomas Bayes - a biographical note tified, this is not a diagnostic problem, as rou- (together with a reprinting of Bayes, 1764). Biometrika 45, 293-315. Reprinted in Pearson and Kendall (1970), tine histological examination had revealed the 131-153 presence of coagulative necrosis. Twenty four 16 Smith AJ. On the histological behaviour of the cardiac of cases muscle in two examples of organization of myocardial 26 of non-cardiac death were also infarct. Univ Pennsylvania Med Bull 1904;17:227-34. identified, confirming that contraction band 17 Pearce RM. Experimental myocarditis: A study ofthe histo- not an logical changes following intravenous injections of ad- necrosis is artefactual occurrence. In the renalin. J Exp Med 1906;8:400. two non-cardiac cases that showed significant 18 Caulfield J, Klionsky B. Myocardial ischaemia and early infarction: An electron microscopic study. Am J' Pathol contraction band necrosis, identifiable factors 1959;35:489-523. were present which could account for its pres- 19 Sommers HM, Jennings RB. Experimental acute myocardial ence. Of the infarction. Histologic and histochemical studies of early possible infarct group, seven of myocardial infarcts induced by temporary or permanent 26 cases contained significant contraction band occlusion of a coronary artery. Lab Invest 1964;13:1491- 503. necrosis, which approaches the diagnostic ac- 20 Herdson PB, Sommers HM, Jennings RB. A comparative curacy of isoenzyme analysis. The best results, study of the fine structure of normal and ischaemic dog come from a myocardium with special reference to early changes fol- however, using combination of lowing temporary occlusion of a coronary artery. Am J contraction band necrosis quantification and Pathol 1965;46:367-86. 21 Kloner RA, Ganote CE, Whalen DA, Jennings RB. Effect biochemical analysis, when 15 of 26 test cases of a transient period of ischaemia on myocardial cells. II. can be identified as cases of infarction, whilst Fine structure during the first few minutes of reflow. Am still JTPathol 1974;74:399-413. maintaining good identification of both 22 Ganote CE, Worstell J, Iannotti JP, Kaltenbach JP. Cellular definite infarction and non-cardiac deaths. Al- swelling and irreversible myocardial injury. Effects ofpoly- not ethylene glycol and mannitol in perfused rat hearts. Am though all cases of early myocardial in- J Pathol 1977;88:95-109. farction can be predicted using contraction 23 Hutchins GM, Silverman KJ. Pathology of the stone heart band necrosis syndrome. Massive myocardial contraction band necrosis http://jcp.bmj.com/ quantification and isoenzyme and widely patent coronary arteries. Am J Pathol 1979; analysis, any additional information is always 95:745-52. useful in cases. 24 Miyazaki S, Fujiwara H, Onodera T, Kihara Y, Matsuda controversial It is also note- M, Wu DJ, et al. Quantitative analysis ofcontraction band worthy that most failures were false negatives and coagulation necrosis after ischaemia and reperfusion in the porcine heart. Circulation 1987;75:1074-82. rather than false positives, indicating that a 25 Matsuda M, Fujiwara H, Onodera T, Tanaka M, Wu DJ, positive result can be viewed with some con- Fujiwara T, et al Quantitative analysis of infarct size, contraction band necrosis, and coagulation necrosis in fidence. human autopsied hearts with acute myocardial infarction In conclusion, we would recommend that after treatment with selective intracoronary thrombolysis. on September 29, 2021 by guest. Protected copyright. cases Circulation 1987;76:981-9. in of sudden unexplained death, when 26 Szakacs JE, Cannon A. 1-Norepinephrine myocarditis. Am myocardial infarction is suspected but not iden- J Clin Pathol 1958;30:425-34. 27 Szakacs JE, Dimmette RM, Cowart EC. Pathologic im- tified, valuable information can be obtained plication of the catecholamines, epinephrine and nor- when cardiac isoenzyme analysis is combined epinephrine. US Armed Forces MedJ. 1959;10:908-25. 28 Csapo Z, Dusek J, Rona G. Early alterations of the cardiac with quantification of contraction band nec- muscle cells in isoproterenol-induced necrosis. Arch Pathol rosis. 1972;93:356-65. 29 Lunt DWR, Rose AG. Pathology of the human heart in drowning. Arch Pathol Lab Med 1987;111:939-42. We would like to thank Joan Holtom for her technical assistance 30 Tazelaar HD, Karch SB, Stephens BG, Billingham ME. and West for his statistical advice. Cocaine and the heart. Hum Pathol 1987;18:195-9. Christopher 31 Cebelin MS, Hirsch CS. Human stress cardiomyopathy. Myocardial lesions in victims ofhomicidal assaults without internal injuries. Hum Pathol 1980;11:123-32. 1 Knight B. The pathology of sudden death. In: Forensic 32 Karch SB. Resuscitation-induced myocardial niecrosis. Ca- pathology. London: Edward Arnold, 1991:444-54. techolamines and defibrillation. Am JfForensic Med Pathol 2 Burns J, Milroy CM, Hulewicz B, West CR, Walkley SM, 1987;8:3-8. Roberts NB. Necropsy study of association between sud- 33 Schlesinger MJ, Reiner L. Focal myocytolysis of the heart. den death and cardiac enzymes. J Clin Pathol 1992;45: AmJPathol 1955;31:443-53. 217-20. 34 Bouchardy B, Majno G. of early myocardial 3 Reichenbach DD, Benditt EP. Myofibrillar degeneration: A infarcts. A new approach. Am J Pathol 1974;74:301-30. response ofthe myocardial cell to injury. Arch Pathol 1968; 35 Baroldi G. Different types ofmyocardial necrosis in coronary 85:189-99. heart disease: A pathophysiologic review oftheir functional 4 Baroldi G, Falzi G, Mariani F. Sudden coronary death. A significance. Am HeartJf1975;89:742-52. post-mortem study in 208 selected cases compared to 97 36 Reichenbach DD, Moss NS. Myocardial cell necrosis and "control" subjects. Am Heart J 1979;98:20-31. sudden death in humans. Circulation 1975;51&52(Suppl 5 Oehmichen M, Pedal I, HohmannP. Diagnostic significance HI):III60s-2s. of myofibrillar degeneration of cardiocytes in forensic 37 Wu DJ, Fujiwara H, Matsuda M, Ishida M, Kawamura A, pathology. Forensic Sci Int 1990;48:163-73. Takemura G, et al. Clinicopathological study ofmyocardial 6 Karch SB, Billingham ME. Myocardial contraction bands infarction with normal or nearly normal extracardiac cor- revisited. Hum Pathol 1986;17:9-13. onary arteries. Quantitative analysis of contraction band 7 Morales AR, Fine G, Taber RE. Cardiac surgery and myo- necrosis, coagulation necrosis, haemorrhage, and infarct cardial necrosis. Arch Pathol 1967;83:71-9. size. Heart Vessels 1990;6:55-62.