Quick viewing(Text Mode)

Association of Cardiomegaly with Coronary Artery Histopathology and Its Relationship to Atheroma

Association of Cardiomegaly with Coronary Artery Histopathology and Its Relationship to Atheroma

32 Journal of and Thrombosis Vol.18, No.1 Coronary Histopathology in 33 Original Article

Association of Cardiomegaly with Coronary Histopathology and its Relationship to

Richard Everett Tracy

Department of Pathology, Louisiana State University Health Sciences Center, New Orleans, USA

Aims: Hypertrophied at autopsy often display excessive coronary artery atherosclerosis, but the histopathology of coronary in hearts with and without cardiomegaly has rarely been com- pared. Methods: In this study, forensic autopsies provided hearts with unexplained enlargement plus com- parison specimens. Right coronary artery was opened longitudinally and flattened for formalin fixa- tion and H&E-stained paraffin sections were cut perpendicular to the endothelial surface. The mi- croscopically observed presence or absence of a necrotic atheroma in the specimen was recorded. At multiple sites far removed from any form of atherosclerosis, measurements were taken of intimal thickness, numbers of smooth muscle cells (SMC) and their ratio, the thickness per SMC, averaged over the entire nonatheromatous arterial length. When the mean thickness per SMC exceeded a cer- tain cutoff point, the artery was declared likely to contain a necrotic atheroma. Results: The prevalence of specimens with necrotic increased stepwise with increasing weight, equally with fatal or with incidental cardiomegaly, and equally with - or -related hypertrophy, rejecting further inclusion of appreciable age, race, or gender effects. The prevalence of specimens with thickness per SMC exceeding the cutoff point was almost always nearly identical to the prevalence of observed necrotic atheroma, showing the two variables to be tightly linked to each other with quantitative consistency across group comparisons of every form. Conclusions: In summary, cardiomegaly, irrespective of the specific cause, seems to accelerate the risk of atheromas, and to do so by first altering the arterial architecture, especially by increasing intimal thickness per SMC.

J Atheroscler Thromb, 2011; 18:32-41.

Key words; Aging, Atherosclerosis, Human, Hypertension, Obesity

Abbreviations; SMC: smooth muscle , FS: fibroplastic thickness per SMC in intima, CHD: coronary heart , YesA: 1 if specimen has an atheroma, NoA: specimen has no atheroma, FI: fibrolastic thickness of intima, SI: SMC numbers in intima, MAP: mean arterial pressure, BMI:

the condition is usually classed among the cardiomy- Introduction opathies. The most common of these are ischemic, hy- When cardiomegaly is encountered at autopsy in pertensive, obesity-related, idiopathic dilated, and hy- the absence of evidence of myocardial lesions, valvular pertrophic , as variably defined 1-5). deformities, or other anatomically evident causes, then When the cause of hypertrophy is determined to be hypertension6), obesity7-10), or a combination of the Address for correspondence: Richard E Tracy, Department of two, then the are expected to display Pathology, Louisiana State University Health Sciences Center, 1901 Perdido St 1P5, New Orleans, LA 70112, USA excessive sclerosis. , clinically E-mail: [email protected] recognized, is said to occur upon finding significant Received: January 28, 2010 coronary by or noninvasive Accepted for publication: August 19, 2010 methods, whether or not this is confirmed by abnor- 32 Journal of Atherosclerosis and Thrombosis Vol.18, No.1 Coronary Histopathology in Cardiomegaly 33 Original Article

mal chamber wall motility2-3). Idiopathic dilated car- said to manifest a “predicted atheroma”, a convention diomyopathy, clinically recognized, manifests dilated adopted to contribute brevity to the exposition and to hypertrophy without anatomic cause, hypertension, or reduce the sizes of some tables. obesity 11-13), and coronary arteries are generally thought to have little atherosclerosis3-13), although Aim contrary evidence has been observed 12-13). Recent stud- ies applying newly introduced technologies for quanti- The presence or absence of observed or predicted fying coronary artery calcifications have reported an atheromas is examined here in relation to heart weight increased extent of such plaques in relation to cardiac in a series of forensic autopsies. The question of cen- hypertrophy2, 6, 14-16); however, these studies in living tral interest is whether cardiomegaly itself might pro- patients were handicapped by the inability to clearly mote observed or predicted atheromas, irrespective of distinguish from other forms of cardiomyop- the cause of cardiomegaly, a question that has been athy. Whether hypertrophy can itself promote coro- raised by accumulating epidemiological evidence21). nary plaques (rather than result from them) therefore remains a persistent question. Methods Postmortem examination offers some advantages Source of material and some shortcomings for exploring the relationships Specimens of coronary artery were prepared at of cardiomegaly to coronary artery sclerosis. Ischemia autopsy from 462 men and women of black and white from coronary artery stenosis is widely thought to pro- ethnic groups aged 35-98 years in the Orleans Parish duce only when myocardial Coroner’s Office from 1993-2005, accepting all causes scarring interferes with ventricular function, but not of except for some selected forms of anatomical- otherwise17); hence, inspection of the myocardium for ly evident heart disease, cerebrovascular disease, and visible lesions should allow for elimination of nearly aortic aneurysms. A basal category using subjects with all instances of ischemic cardiomyopathy. Further- causes of death from violence, or natural causes unre- more, coronary arteries can be examined in gross and lated to hypertension or atherosclerosis, offers an ap- histological detail18-20); however, important clinical in- proximation of a representative sample of the popula- formation is usually unavailable in forensic autopsies. tion22), but this report does not require the use of such In the autopsy series to be reported here only a subset a basal group. This autopsy study, omitting patient of subjects had data on height and weight to use for identifiers, was declared exempt from Internal Review assessing obesity. No information on blood pressure Board review. was available, and evidence of a hypertensive status was extracted from renal histology. Cardiomegaly The aging of coronary arteries is accompanied by The arbitrary cutoff points for defining cardio- progressive alterations of their basic architecture megaly were chosen as heart weight ＞399 grams in throughout sites that are far removed from any form women and ＞449 grams in men. These values corre- of atherosclerosis. The architectural component of late well with the values sometimes applied to echo- particular interest here, revealed in H&E-stained par- cardiographic data as thresholds for diagnosing left affin sections, is thickening of the intima, which oc- ventricular hypertrophy23). These values were abstract- curs with little or no change in the numbers of smooth ed from the autopsy protocols without special han- muscle cells (SMCs)19, 20), so that an increase in the dling, and are expected to include measurement errors amount of interstitial matrix material occurs per of uncertain magnitude. SMC, a quantity abbreviated here as FS, fibroplastic thickness per SMC (excluding atherosclerotic thicken- Cardiac diagnosis ing). Much evidence has accumulated to suggest that Coronary heart disease (CHD) was recognized as the likelihood of finding a necrotic atheroma some- , or myocardial lesions (scars or where in the specimen strongly increases with FS in a infarcts) in company with arterial stenosis or occlu- simple mathematical relationship 19, 20). When the sion. All instances of CHD were omitted. Also omit- mean magnitude of FS exceeds a certain defined ted were the few instances of cardiac disorders attrib- amount, the probability of finding an atheroma ex- uted to valvular deformities, right ventricular abnor- ceeds 50%, and the specimen can be declared likely to malities, and chronic renal disease. In 40 cases the contain an atheroma. In this report, any specimen cause of death was determined at autopsy to be left whose mean FS is found to exceed this threshold is ventricular hypertrophy, usually in the form of dilated 34 Tracy Coronary Histopathology in Cardiomegaly 35

cardiomyopathy. These 40 subjects constitute the group of cases referred to as “fatal cardiomegaly” (mean B heart weight, 697 grams). Incidental cardiomegaly in the presence of a clearly apparent cause unrelated to occurred in 73 cases, 46 instanc- es of death by violence or poisoning and 27 by assort- ed non-cardiovascular natural (mean heart A weight, 501 grams). This group of 73 subjects was designated “incidental cardiomegaly”. Hypertrophic cardiomyopathy in its original sense of asymmetric septal hypertrophy marked by myocyte disarray4) was not encountered in this series.

Processing of specimens The first 9 cm of each right coronary artery was opened longitudinally, dissected from the heart, com- pressed by a sponge to flatten for fixation in 4 % ace- tate-buffered formaldehyde, and cut into 3 to 5 longi- tudinal segments to be embedded in paraffin on edge to allow sectioning perpendicular to the luminal sur- face. Sections of 6 μm thickness were stained with he- Fig.1. Hearts with (B) or without (A) cardiomegaly show av- matoxylin and eosin (H&E). erage SMC numbers (above, N) and layer thicknesses (below, μm) similar to these illustrations. Evaluation of slides Blood flow left to right. Vertical lines 100 μm apart; arrows mark Atheromatous necrotic cores, recognized when intimal surface, intima-media junction and adventitia; H&E. necessary by crystals easily discerned under a 40X objective lens, were marked in black ink on the cover slip. No specific size limit was placed on these in were excluded, but nearly always a nearby location was the protocol, but in practice the sizes always exceeded found to substitute for the otherwise excluded data the 100 μm width marker incorporated in the ob- point. Omission of sites with atherosclerotic features served microscopic field. The variable named YesA was assures that nearly pure populations of SMCs are ob- assigned the value of 1 if an atheromatous necrotic tained, without dilution by appreciable numbers of core was found anywhere in the specimen; otherwise immigrant leukocytes from the blood25). Omitting it was assigned the value of zero. At nine equally sites with atherosclerosis has the hypothetical possibil- spaced positions along the coronary sample (approxi- ity of generating a biased sample of the artery that re- mately 1 cm separations), excluding black marked lo- mains to be observed, because the atheroma may not cations and sites with microscopically visible fatty select and censor sites at random. This biasing effect streak elements24), measurements were taken of inti- was examined elsewhere in detail26), and the only sug- mal and medial thicknesses and counts were made of gested effect was a minor increase in the observed SI in nuclei, presumed to be nearly all those of smooth YesA specimens, possibly explained by selective ather- muscle cells (SMCs), within a band of 100 μm width omatous censoring of sites with the greatest fibroplasia drawn with an eyepiece reticle through the full intimal and the fewest cells; the effect was judged to be negli- and medial thicknesses, as in Fig.1 (in that figure, gibly small for the purposes of this report. SMC numbers (above) and layer thicknesses (below) for these examples are given for intima and media). Hypertension Means of the fibroplastic thicknesses for intima and Measuring the intimal thickness of interlobular media (omitting atheromatous thickening) are called arteries in PAS-stained paraffin sections of renal the FI and FM, respectively, and their associated SMC cortex provided information for calculating mean ar- numbers are SI and SM. Their ratios, defined as FS =FI/ terial pressure (MAP). In a previous series of 236 au- 27, 28) SI and MS =FM/SM, denote the amount of matrix ma- topsies , data on mean blood pressure were avail- terial associated with an average SMC, and the FS ra- able from clinical records. Measurements of interlobu- tio in particular is of major interest. Sites with an ath- lar arteries in those cases were used to generate a re- eromatous necrotic core or fatty streak elements12) gression equation to use for calculating the predicted 34 Tracy Coronary Histopathology in Cardiomegaly 35

clinical MAP, and that equation is now available for across groupings by cardiac diagnosis, using chi square application to the current study. By tests for the significance of differences. The influence analysis, the occurrence of clinically measured hyper- of heart weight upon these two variables, altered by tension (observed MAP ＞106.7 mmHg) coincided the inclusion of the covariates, age, race, sex, MAP, with predicted hypertension (calculated MAP ＞106.7 and BMI, were tested by logistic regression. mmHg) in 87 cases, observed normotension coincid- ed with predicted normotension in 91 cases, and the Results false positive rate was 24.3%, while the false negative rate was 24.8%, odds ratio=9.5. Since blood pressure Relations of selected variables to atheroma status of 140/90 mm Hg corresponds to MAP =106.7 Specimens containing at least one necrotic ather- mmHg, MAP ＞106.7 mmHg was used in this data oma (YesA cases) tended to occur at older age, to be set to define hypertension. Renal samples were used in more often from white males and the hearts were this study to calculate predicted MAP, and it was ex- more often heavier than NoA cases (Table 1, YesA and pected that about 1/4 of the cases in the hypertension NoA heart weights, respectively, 421 vs 377 grams, category thus defined would be false positives, while adjusted for age, race, and gender). Atheroma-bearing 1/4 of the normotensives would be false negatives. arteries, evaluated at nonatherosclerotic sites distant from any form of atherosclerosis, were also in marked Obesity histological contrast to atheroma-free specimens, with Body mass index (BMI) was defined as weight special interest in the enhancement of intimal thick- over height squared, and obesity was defined as BMI ness and intimal thickness per SMC (FI and FS, re- ＞29 Kg/m2. Only 304 subjects had data on both spectively in Table 1, P＜0.001 for both variables). BMI and calculated MAP, and this subset of cases was SMCs showed a small but significantly higher number used for analyzing these variables. in the same comparison (respectively 25.3 vs 22.9 SMCs per unit cut through the artery wall, P=0.007). Predicted atheroma Measurements of intimal thickness and SMCs Predictor variables for atheroma status counts at sites distant from any form of atherosclerosis Univariate logistic regression (models 1 and 2 in can be used to generate the mean ratio FS as defined Table 2) showed both variables, excessive intimal above. Elevations of this ratio have consistently re- thickness and SMC numbers, to be predictors of ath- vealed, under a wide range of conditions, a strong as- eroma presence, thickness more strongly than SMCs. sociation with the presence of necrotic atheroma Taking the two variables together (bivariate model 3), somewhere in the specimen18, 20). When mean FS ＞ greatly strengthens the thickness effect (beta rising to 8.873 μm/SMC, a quantity derived later, the proba- 3.53 from 2.04 SD units) while reversing the sign of bility for finding an atheroma somewhere in the speci- beta for SMC numbers and enhancing its magnitude men exceeds 50%, and the specimen is said to contain (the product term, FI×SI, is rejected as adding no fur- a “predicted atheroma” (exemplified by Fig.1B). ther significance P=0.21). The ratio, FS=FI/SI, serves to indicate how much collagenous matrix is associated Statistical analyses with an average SMC in the intima, and this quantity The measure of coronary atherosclerosis of prin- statistically summarizes the prediction of atheroma us- cipal interest here is the presence (YesA) or absence ing both SMC numbers and intimal thickness taken (NoA) of necrotic atheroma observed somewhere in together (Table 2, model 4). Specimens with the few- the specimen. The SAS package of computer programs est SMCs require the least intimal thickening to exert (SAS Institute, Cary, NC, USA) was used for analyses. an effect upon the prediction of atheroma presence The means of selected variables in the YesA specimens while those with the most SMCs require the greatest were compared with those in NoA specimens, after thickness. Models 3 and 4, assessing the combined ef- adjustment for selected covariates, by PROC GLM. fects of FI and SI, generate 2×2 contingency tables; The relationships of YesA (yes/no variable) to predic- that for model 3 yields an odds ratio of OR=14.8, tor variables were examined at the outset by logistic while the ratio term, FS =FI/SI, yields OR=12.8, indi- regression, leading to an equation that was used to cal- cating that the entire action of FI plus SI in this data culate a second principal quantity called “predicted set is not fully captured by the ratio FS. This difference atheroma”. These two variables, observed and predict- seems small enough to treat as negligible, and the in- ed atheroma, are dichotomous variables which were tuitively meaningful simpler model 4 was chosen for compared across groupings according to heart size and further exploration. By recalculating model 4 using 36 Tracy Coronary Histopathology in Cardiomegaly 37

Table 1. Descriptions of variables (N=462) and means for YesA (N=229) and NoA (N=233) subjects aged 35-98 years＊ Descriptions Means Variable Symbol Units Mean SD NoA YesA P-value† Age A Years 51.5 13.0 47.1 55.9 ＜0.001 Gender (M=0, F=1) G 0,1 0.325 0.469 0.408 0.240 ＜0.001‡ Race (W=0, B=1) R 0,1 0.528 0.500 0.586 0.469 0.020§ Heart Weight HW Grams 398.9 126.6 377.0 421.2 ＜0.001‖ ‖ Intimal SMCs SI Number 24.1 9.4 22.9 25.3 0.007 ‖ Intimal thickness FI μm 224.2 126.6 166.1 283.3 ＜0.001 Thickness per SMC ‖ Intimal FS μm/SMC 9.35 3.93 7.51 11.22 ＜0.001 ‖ Medial MS μm/SMC 2.49 0.58 2.42 2.56 ＜0.001

＊ YesA specimens have at least one instance of atheromatous necrotic core elsewhere in the specimen at a distance from measurements sites are taken and NoA specimens have no such instance. The data set excludes instances of chronic renal, valvular, right ventricular, and coronary heart disease. †Significance levels are determined by F-tests within PROC GLM analysis. ‡NoA and YesA means are adjusted for age within PROC GLM analysis. §Means are adjusted for age and gender. ‖ Means are adjusted for age, race, and gender; FS =FI/SI and MS =FS/MS are calculated using means of observations within cases.

Table 2. Logistic regression models using predictor variables to calculate expected outcomes for presence or absence of at least one observed necrotic atheroma elsewhere in the specimen from sites of measure- ment (N=462)＊. Beta Coefficient

Model Constant Estimate ±2 SD Odds ratio

1. Intimal fibroplastic thickness; FI －3.35 2.04 1.63 to 2.45 9.8 2. Intimal SMCs; SI －1.03 0.40 0.20 to 0.60 2.0 3. Intimal fibroplastic thickness; FI －2.15 3.53 2.82 to 4.24 14.8 Intimal SMCs; SI －1.39 －1.82 to －0.96 4. Thickness per SMC; FS =FI/SI －5.06 2.24 1.81 to 2.67 12.8 5. Thickness per SMC; FS －4.38 2.19 1.71 to 2.67 14.4 Race －0.44 －0.69 to －0.19 Gender －0.31 －0.57 to －0.05 Heart weight 0.26 －0.01 to 0.53 Age 0.21 －0.09 to 0.50

＊ The coefficients here are standardized so that each predictor variable is measured in units of its own standard deviation; raw coefficients can be regenerated using standard deviations in Table 1, where variables are described. Odds ratio is calculated from the two by two table of classification outcomes.

variables in their raw units of measurement, FS =8.873 centages of specimens revealing at least one observed μm/SMC is found as the cutoff point for separating necrotic atheroma are also shown. The two variables, YesA from NoA cases, so that a predicted atheroma is percentage of cases with predicted and observed ather- declared to be present if this cutoff is exceeded (i.e. a omas, both repeatedly showed almost perfect stepwise necrotic atheroma is predicted to be found somewhere increases as the heart became progressively heavier in the specimen). This model rejects further entry of (Spearman rank order correlation between the 6 pairs age (P=0.108) and heart weight (P=0.053). of observed and predicted atheromas is r=0.989); the regression between them rejects the additional qua- Relation of atheroma to heart weight dratic term of YesA (P=0.32) implying linearity, al- Percentages of specimens displaying predicted though comparing pairs of means by the chi square atheroma, as just defined, are recorded in Table 3 test indicates that the three largest heart sizes do not within 6 ranges of increasing heart weight. The per- differ significantly in either observed or predicted ath- 36 Tracy Coronary Histopathology in Cardiomegaly 37

Table 3. Percentage＊ of cases having necrotic atheromas in the right coronary artery, observed or predicted, ac- cording to the total heart weight in 462 men and women aged 35-98 years. Necrotic atheroma (% cases) Range of Mean heart Number of Size Category weight (g) Observed Predicted Weight (g) cases 1 150-300 33.3A 40.7AB 261 81 2 310-380 47.2B 36.2A 342 163 3 390-460 48.8BC 48.8B 421 129 4 470-540 65.9CD 58.5BC 499 41 5 550-620 69.6CD 65.2C 580 23 6 630-990 76.0D 68.0C 768 25 Chi square† 23.9 18.5 (P) (＜0.001) (0.002)

＊ Predicted atheroma is defined as FS ＞8.873 μm/SMC. Heart weights were recorded to with 10 grams. † Tabulated percentages were used to calculate numbers of YesA and NoA cases in each of the six heart size categories, there- by generating a 2×6 continency matrix for each column, and chi square tests rejected randomness of atheroma dispersion in both of these matrices. Pairs within a column that fail to share a symbol A, B, C, or D differ significantly at P＜0.05 by chi square test.

Table 4. Percentage＊ of cases of necrotic atheromas, observed or predicted, and hypertension or obesity, ac- cording to the fatal or incidental category of cardiomegaly in men and women aged 35-98 years. Necrotic Atheroma % cases‡ Percent of cases§ Mean heart Cardiomegaly category Observed Predicted Hypertensive Obese Weight (g) No cardiomegaly 44.3A 41.1A 28.6A 23.4A 344 N cases (348) (348) (231) (231) Cardiomegaly Fatal 62.5B 62.5B 55.6B 63.0B 697 N cases (40) (40) (27) (27) Incidental 67.6B 58.1B 50.0B 60.9B 501 N cases (74) (74) (46) (46) Chi square† 16.2 12.1 13.8 37.0 (P) (0.001) (0.001) (＜0.001) (＜0.001)

＊See footnote Table 3 †See footnote Table 3. With three categories of cardiac condition, the contingency matrices for chi square testing are 2×3 in each column. ‡ Number of cases=462 includes those with missing data on hypertension and obesity. §Number of cases=304 includes only those with complete data.

eroma. It remains ambiguous, therefore, whether ath- mas, and also preserving the near identity of observed eroma prevalence reached a limit as heart sizes ap- with predicted atheromas in all comparisons (Table 4). proach the extreme of their upper range. Relations of hypertension and obesity to cardiomegaly Relations of atheroma to cardiac diagnosis In the subset of 304 subjects with data on body Of the 114 specimens with cardiomegaly, 40 mass index (BMI=kg/m2) and renal-derived informa- were classed as cause of death and 74 as incidental ac- tion on mean arterial pressure (MAP, mm Hg), the companiment of some other conspicuous cause of percentages classed as obese were 23.4, 63.0, and death, usually violence (46 cases). When compared 60.9% (P＜0.001) and the percentages classed as hy- with subjects lacking cardiomegaly, both categories of pertensive were 28.6, 55.6, and 50.0 % (P＜0.001), cardiomegaly were conspicuously similar in displaying respectively, in the three cardiac diagnosis categories in nearly identical increases in the prevalence of athero- Table 4. The two categories of cardiomegaly were 38 Tracy Coronary Histopathology in Cardiomegaly 39

Table 5. Logistic regression models relating observed or predicted atheromas to selected variables＊.

Beta Coefficient

Model Constant Estimate ± 2 SD Odds ratio

RESPONSE VARIABLE=YESA OBSERVED

1. Heart weight (HW) －1.58 0.50 0.28 to 0.72 1.7 Number of cases=462 2. HW －4.53 0.50 0.18 to 0.86 2.4 MAP 0.30 0.05 to 0.55 BMI (p=0.50)† －0.15 －0.42 to 0.12 Number of cases=304 3. HW －2.85 0.40 0.20 to 0.63 4.1 Age (A) 0.82 0.57 to 1.07 Gender (G) －0.31 －0.53 to －0.09 Race (R) －0.23 －0.44 to －0.02 Number of cases=462

RESPONSE VARIABLE=YESA PREDICTED

1. HW －1.39 0.38 0.18 to 0.58 2.6 Number of cases=462 2. HW －7.17 0.40 0.09 to 0.71 2.9 MAP 0.53 0.27 to 0.84 BMI (p=0.27)† －0.09 －0.35 to 0.18 Number of cases=304 3. HW －5.27 0.32 0.10 to 0.54 4.8 A 1.07 0.81 to 1.33 G (p=0.15)† －0.16 －0.38 to 0.06 R (p=0.28)† 0.11 －0.10 to 0.31

＊See footnote Table 2. MAP is mean arterial pressure and BMI is body mass index. † Beta does not differ significantly from zero with the specified p level. nearly identical in the relevant comparisons. The 20 (0.50, model 2). Similarly, the beta for predicted ath- cardiomegaly cases of both hypertension and obesity eroma was almost unchanged (0.38 vs 0.40). These had equal prevalence to the cases of either hyperten- are the expected results if the heart weight effect is sion or obesity alone, and these cases are therefore not similar whether correlated with hypertension or obesi- detailed in Table 4. These findings point to about ty or some other unidentified agent. equal elevated incidences of hypertension and obesity among cardiomegaly cases. Relationship of atheroma to heart weight adjusted for age, race, and gender Relations of atheroma prevalence to MAP and BMI These adjustments are introduced in model 3 in By logistic regression analysis, mean arterial pres- Table 5. In this multivariate setting, race and gender sure (MAP) was significantly associated with necrotic had negligibly small effects on observed and predicted atheroma prevalence, both observed and predicted, atheroma prevalences, while age was strongly related adjusted for heart weight (model 2 in Table 5, to atheromas of both types. The beta coefficient relat- P=0.01 and P＜0.001 respectively). Body mass index ing heart weight to observed atheroma (0.50, model 1 (BMI) had no significant association with atheroma, in Table 5) was little changed by adjusting for age, either observed or predicted, after adjustment for race, and gender (0.40, model 3). The same outcome heart weight (P=0.50 and P=0.27, respectively, mod- was obtained for predicted atheroma (beta declining el 2 in Table 5). The beta coefficient relating heart from 0.38 to 0.32). Although a small component of weight to observed atheroma (0.50, model 1 in Table heart weight association with atheroma seems to arise 5) was unaffected by adjusting for MAP and BMI from their common correlation with age, the effect is 38 Tracy Coronary Histopathology in Cardiomegaly 39

small, as expected if heart weight has almost identical after adjusting for numerous covariates, including associations with atheromas at all ages. blood pressure and obesity6, 9). The entity here assigned the name “predicted SMC numbers atheroma” refers to a property of the age-altered basic The mean number of intimal SMCs (SI in Ta- architecture of the artery, recognized by an excess of bles 1 and 2) in the three cardiac diagnosis categories mean intimal fibroplasia per SMC, FS, at sites lacking of Table 4 were 23.6, 25.5, and 26.3 SMCs per unit any form of atherosclerosis (Fig.1B). The data tabu- band through the artery wall, respectively (P=0.04). lated here reiterate prior reports18, 20) showing that the After adjustment for the YesA condition as a covariate, mean magnitude of FS displays a strong linear associa- these means became 23.7, 25.1, and 25.7 SMCs per tion with the probability of finding a necrotic athero- unit band, and lost their significance of difference (P= ma somewhere in the specimen (Table 2). Further- 0.18) For medial SMCs the respective means were more, the probability of finding observed atheromas 77.4, 87.3, and 80.4 SMCs per unit band through the varies across group comparisons in nearly perfect par- artery wall; 87.3 differs significantly from 77.4 (P＜ allel with comparable variations of predicted athero- 0.001) while 80.4 does not (P=0.18). Later adjust- mas (Tables 3 & 4). The enhanced probability of ment for the YesA condition had only negligible ef- finding an observed or predicted atheroma in the cor- fects on these tabulated values for the media and the onary artery that accompanies cardiomegaly appears differences retained their significance or its absence, to occur in all enlarged hearts, irrespective of the cause P=0.003 and P=0.22, respectively. of hypertrophy in these selected classes. Obesity in particular shows evidence of enhancing atheroma only in those obese subjects that manifest cardiomegaly. Discussion It has been proposed that sufficient increases of In the present data set, cardiomegaly is seen to FS in aging coronary arteries can act to promote the display a strong association with the likelihood of sequestration of collagen-linked interstitial lipid30) and finding necrotic atheroma, either observed or predict- evolution to atheroma20, 26, 30). Therefore, if enlarged ed, in the right coronary artery (Table 3). It is as- arteries that grow bigger along with the hypertrophy- sumed here that nearly all cases of ischemic cardiomy- ing heart should accelerate the progression of FS, then opathy should be eliminated by finding postmortem the enhancement of atheromas could be explained. evidence of coronary thrombosis or visible ischemic It is generally accepted that little if any increase lesions in the myocardium2, 3, 17). To the extent that in cardiac myocyte numbers occurs during ventricular this assumption may be false, the following conclu- hypertrophy. Perhaps something similar may be in op- sions must be viewed with caution. eration in the intimal SMC population of coronary Hypertension and obesity were far more com- arteries31). These cells can proliferate after injury in mon in the groups with cardiomegaly than in the the process of wound healing, but this is unlikely to group lacking a cardiac diagnosis, and these agents are be what is happening in natural aging26, 31). The num- likely candidates causing excessive heart weights29). bers of intimal SMCs in the coronary arteries of this However, the data reviewed here seem to implicate study showed little if any increase with cardiomegaly; only hypertension, and not obesity, as a serious addi- rather, the data showed an expansion of the collage- tional factor in the coexistence of atheromas in hyper- nous matrix, thereby thickening the arterial intima in trophied hearts, because the introduction of MAP to company with a nearly stable population of SMCs. the logistic regression model significantly improved SMC numbers therefore appear to increase little or the forecasting of atheroma, while BMI had no such not at all with the coronary enlargement of cardio- effect (Table 5, model 2). Furthermore, the introduc- megaly, irrespective of its cause, thereby leaving the in- tion of MAP and BMI had no influence on the persis- tima with an enhanced mass of matrix insufficiently tent relationship of heart weight to atheroma presence supported by overly few SMCs. (Table 5, models 1 and 2 both report beta=0.50 for This study has a number of serious limitations heart weight). This evidence taken together suggests that are inherent in studies of forensic material. These that heart weight itself may have independent activity include the absence of clinical information about dys- for directly promoting atheroma, whether the cardio- lipidemia, , alcoholism, and metabol- megaly results from hypertension, obesity, or other ic diseases. Also, the measure of hypertension is con- unidentified causes. Clinical studies of coronary calci- fined to an undesirable range of false positives and fications have found similar results, with heart size re- negatives. Of particular concern is the uncertainty that taining a persistent correlation with coronary plaques instances of ischemic cardiomyopathy have been suffi- 40 Tracy Coronary Histopathology in Cardiomegaly 41

ciently eliminated; however, the histological data un- in hypertrophic cardiomyopathy with regard to age, asym- der inquiry are accessible only in autopsy materials, metric septal hypertrophy, and concentric hypertrophy and forensic material has the unique advantage of of- beyond the pediatric age group. Arch Pathol Lab Med, 1998; 122: 434-441 fering instances of death by violence or poisoning al- 5) Ikeda H, Maki S, Yoshida N, Murohara T, Adachi H, lowing, observation of the histology unaltered by oth- Koga Y & Imaizumi T: Predictors of death from conges- er complicating diseases. These limitations need to be tive i hypertrophic cardiomyopathy. Am J tolerated in order to perform a study. The findings Cardiol, 1999; 83: 1280-1283 should be held as tentative working hypotheses. 6) Altunkan S, Erdogan N, Altin L, Budoff MJ: Relation of coronary artery calcium to left ventricualr mass and ge- ometry in patients with essential hypertension. Blood Conclusions Press Monit, 2003; 8: 9-15 7) Kortelainen ML & Särkioja T: Extent and composition of When hearts are examined at autopsy, it is possi- coronary lesions and degree of cardiac hypertrophy in re- ble that the evidence used to recognize ischemic car- lation to abdominal fatness in men under 40 years of age. diomyopathy is sufficiently reliable so that nearly all Arterioscler Thromb Vasc Biol, 1997; 17: 574-579 instances of this condition are eliminated. This would 8) Tanaka K, Kodama H, Sasazuki S, Yoshimasu K, Liu Y, offer the opportunity to avoid the possible mischief Washio M, Tokunaga S, Kono S, Arai H, Koyanagi S, Hi- from ascribing to a primary coronary disease the find- yamuta K, Doi Y, Kawano T, Nakagaki O, Takada K, Nii T, ing of an increased likelihood of coronary plaques Shirai K, Ideishi M, Arakawa K, Mohri M, Takeshita A: Obesity, body distribution and coronary atherosclero- with cardiomegaly. The data cautiously introduced sis among Japanese men and women. Int J Obesity, 2001; here point to a correlation of atheromas with heart 25: 191-197 weight that is independent of the specific cause of car- 9) Schunkert H: Obesity and target organ damage: the heart. diac enlargement. It could be, therefore, that enlarg- Int J Obes, 2002; 26, Suppl 4, S15-S20 ing coronary arteries in the hypertrophied heart are 10) McGill HC Jr., McMahan CA, Herderick EE, Zieske AW, not exhibiting physiological growth, but rather a form Malcom GT, Tracy RE, Strong JP, Pathobiological Deter- of pathological ectasia, marked especially by an in- minants of Atherosclerosis in Youth (PDAY) Research creased amount of fibroplastic matrix per SMC, and Group: Obesity accelerates the progression of coronary atherosclerosis in young men: Circulation, 2002; 105: that this might bear upon the correlation of heart size 2712-2718 with coronary atheromas. 11) Waynne J & Braunwald E: The and myocarditides. In: Heart Disease 6th Ed, ed by Braun- wald E, Zipes DP & Libby P, pp1751-1797, Philadelphia Acknowledgements W.B. Saunders Co, Philadelphia, 2001 This work received no grant funding. The skilled 12) Sharp SM, Sawada SG, Segar DS, Ryan T, Kovacs R, technical input of Nicole Jackson was indispensable to Fineberg NS & Feigenbaum H: echo- cardiography: Detection of in pa- this project. tients with . J Am Coll Cardiol, 1994; 24: 934-939 13) Repetto A, Dal Bello B, Pasotti M, Agozzino M, Viganò M, Klersy C, Tavazzi L & Arbustini E: Coronary athero- References sclerosis in end-stage idiopathic dilated cardiomyopathy: 1) Richardson CP, McKenna W, Bristow M, Maisch B, an innocent bystander? Eur Heart J, 2005; 26: 1519-1527 Mautner B, O’Connell J, Olsen E, Thiene G, Goodwin J, 14) Tong W, Lima JA, Lai H, Celentano DD, Dai S, Lai S: Gyarfas I, Martin I, Nordet P: Report of the 1995 World Relation of coronary artery calcium to left ventricular Health Organization/International Society and Federation mass in African-Americans. Am J Cardiol, 2004; 93: of Task Force on the definition and classifica- 490-492 tion of cardiomyopathies. Circulation, 1996; 93: 841-842 15) Gardin JM, Iribarren C, Detrano RC, Liu K, Schreiner 2) Sawada SG, Ryan T, Segar D, Atherton L, Fineberg N, PJ, Loria CM & Wong ND: Relation of echocardiograph- Davis C & Feigenbaum H: Distinguishing ischemic car- ic left ventricular mass, geometry and wall stress, and left diomyopathy from nonischemic dilated cardiomyopathy atrial dimension to coronary calcium in young adults (The with coronary . J Am Coll Cardiol, CARDIA Study). Am J cardiol, 2005; 95: 626-629 1992; 19: 1223-1228 16) Tang W, Arnett DK, Province MA, Lewis CE, North K, 3) Atak R, Ileri M, Yetkin O, Yetkin E, Turhan H, Senen K, Carr JJ, Pankow JS, Hopkins PN, Devereux RB, Wilk JB Sahin O, Ozbakir C & Demikan D: The role of valvular & Wagenknecht L: Racial differences in the association of and thoracic aortic calcifications in distinction between coronary calcified plaque with left ventricular hypertro- ischemic and nonischemic cardiomyopathy. Angiol, 2004; phy: The National Heart, Lung, and Blood Institute Fam- 55: 661-667 ily Heart Study and Hypertension Genetic Epidemiology 4) Litovsky SH & Rose AG: Clinicopathologic heterogeneity Network. Am J Cardiol, 2006; 97: 1441-1448 40 Tracy Coronary Histopathology in Cardiomegaly 41

17) Correa P, Restrepo C, Montenegro MR & Solberg LA: 24) Committee on Vascular Lesions of the Council on Arte- Coronary atherosclerosis and heart weight. Lab Invest, riosclerosis, American Heart Association: A definition of 1968; 18: 580-585 initial, fatty streak, and intermediate lesions of atheroscle- 18) Tracy RE: Declining density of intimal smooth muscle rosis. Arterioscl Thromb, 1994; 14: 840-56 cells as a precondition for atheronecrosis in the coronary 25) Emeson EE & Robertson ALJr: T lymphocytes in aortic artery. Virchows Arch, 1997; 430: 155-162 and coronary intimas. Amer J Pathol, 1988; 130: 369-376 19) Tracy RE: The course of fibroplasia per SMC over time in 26) Tracy RE: Evidence concerning resistance to atheroma by aging arteries. In: The Role of Aging in Atherosclerosis: media-like islands in the intima of coronary arteries. Ath- The Sequestration Hypothesis, ed by Tracy RE, pp erosclerosis, 2005; 178: 49-65 159-163, Kluwer Academic Publishers, Dordrecht Neth- 27) Tracy, RE, Velez-Duran, M, Heigle, T, Oalmann, MC: erlands, 2003 Two variants of nephrosclerosis separately related to age 20) Tracy RE & Eigenbrodt ML: Coronary artery circumfer- and blood pressure. Amer J Pathol, 1988; 131: 270-282 ential stress: Departure from Laplace expectations with 28) Tracy, RE, Lanjewar, DN, Ghorpade, KG, Valand, AG, aging. TheScientificWorldJournal, 2009; 9: 946-960. DOI and Raghuwanshi, SR 1997: Renovasculopathies of hy- 10.1100/tsw.2009.109 pertension in Bombay, India. Geriat Nephrol Urol, 1997; 21) Kannel WB, Dannenberg AL & Levy D: Population im- 7: 101-109 plications of electrocardiographic left ventricular hyper- 29) Messerli FH, Sundgaard-Riise K, Reisin E, Dreslinski G, trophy. Am J Cardiol, 1987; 60: 85I-93I Dunn FG & Frohlich E: Disparate cardiovascular effects 22) McFarlane MJ, Feinstein AR, Wells CK, Chan CK: The of obesity and arterial hypertension. Am J Med, 1983; 74: ‘epidemiologic necropsy’. JAMA, 1987; 258: 331-338 808-812 23) Devereux RB, Wachtell K, Gerdts E, Boman K, Niemin- 30) Tracy RE: Low smooth muscle cell densities characterize en MS, Papademetriou V, Rokkedal J, Harris K, Aurup P, sites with isolated interstitial lipid in coronary artery inti- Dahlöf B: Prognostic significance of left ventricular mass ma. Arch Pathol Lab Med, 2007; 131: 755-760 change during treatment of hypertension. JAMA, 2004; 31) Zak R: Development and proliferative capacity of cardiac 292: 2350-2356 muscle cells. Circ Res, 1974; 35(2):suppl Ⅱ: 17-26

Top View