313
International Journal of Cardiologv, 32 (1991) 313-322 0 1991 Elsevier Science Publishers B.V. 0167-5273/91 ADONIS 016752739100189A
CARD10
01288
Segmental analysis of coronary arterial stenoses in patients presenting with angina or first myocardial infarction D.P. Department
de Bono
and AK.
of Cardiology,
Bhattacharrya
*
University of Leicester, Ikeester,
U.K.
(Received 26 November 1990; revision accepted 14 February 1991)
De Bono DP, Bhattacharrya AK. Segmental analysis of coronary arterial with angina or first myocardial infarction. Int J Cardiol 1991;32:313-322.
stenoses
in patients
presenting
The segmental distribution of stenoses within the coronary arteries was anafysed in a population of 258 patients with a first myocardial infarction undergoing coronary angiography to evaluate the effect of thrombolytic therapy, and in a population of 466 patients undergoing elective coronary angiography for stable angina. Mean ages were 53.7 and 56.7 years respectively (P = NS). As judged angiographically, coronary arterial disease was more extensive in the group suffering angina, with a greater proportion of patients with two- or three-vessel disease (odds ratio 2.56, 95% confidence interval 1.87 to 3.52) and more patients having stenoses in two or more coronary arterial segments (odds ratio 1.52,95% confidence interval 1.12 to 2.08). For each coronary vessel, the probability of finding a stenosis > 50% in an individual segment was greater in the group presenting with angina. There was a relative deficiency of stenoses within the main stem of the left coronary artery or its proximal left anterior descending branch among the patients suffering myocardii infarction. Within those having angina, subgroups were identified with “isolated” and “diffuse” coronary arterial disease: the latter patients tended to have a lower concentration of total cholesterol in the serum, but an increased prevalence of diabetes mellitus. Patients presenting clinically with a first myocardial infarction, and patients with severe angina, constitute distinct populations selected by different mechanisms from the overall pool of patients with atheromatous coronary arterial disease. Key words:
Coronary
atheroma;
Angiographic
analysis;
Introduction The distribution of atheromatous lesions in the coronary arterial tree is important for understand-
Correspondence to: Prof. D.P. de Bono, Dept. of CardioIogy, Clinical Sciences Wing, Glenfield General Hospital, Leicester LE3 9QP, U.K. * Present address: A.K. Bhattacharrya, Gawhati State University, Assam, India.
Myocardial
infarction;
Stable angina
ing the pathogenesis and natural history of coronary arterial disease. This coronary “ topography” has been extensively studied in autopsy material [l]. Coronary arteriography enables similar information to be obtained during life, and to be correlated with clinical findings [2] and the natural progression of disease [3]. The advent of computerised systems for tabulating cineangiographic data [4] has made it easier to compare patterns of disease in different populations of patients.
314
The European Co-operative Study Group trials of coronary thrombolysis in acute myocardial infarction provided an opportunity for centralised reading and analysis of coronary arteriograms from a relatively large population of patients studied shortly after the onset of symptoms of acute myocardial infarction [5,6]. Besides identifying the presumed stenosis responsible for infarction, and assessing its effect on arterial patency, the distribution of stenoses in the entire arterial tree was recorded in a form suitable for subsequent analysis. For comparison, we applied a similar analysis to a series of 466 patients undergoing angiography for ischaemic heart disease at one of the centres within the Co-operative study group. In this paper, we describe and compare the distribution of arterial lesions in the two populations.
Patients and Methods The “myocardial infarction” group consisted of patients taking part in two simultaneous multicentric trials of thrombolytic therapy in acute myocardial infarction. In one trial, recombinant human tissue type plasminogen activator was being compared with placebo, and in the other trial with intravenous streptokinase. Patients admitted to the trial were under the age of 71, had had chest pain typical of myocardial infarction for 30 minutes or longer, and fulfilled stringent electrocardiographic criteria for myocardial ischaemia (> 2 mV ST segment elevation in at least two precordial leads or > 1 mV ST segment elevation in at least two frontal plane leads). To assist in the unequivocal identification of the vessel responsible for infarction, only patients with a first myocardial infarction were studied. Detailed protocols for both studies have been published elsewhere [5,6]. Angiograms in those patients suffering myocardial infarction were taken as close as possible to 90 minutes after the start of administration of thrombolytic therapy which, in turn, was initiated within six hours of the onset of major symptoms. All angiograms were recorded on 35 mm tine film and were read centrally by a small group of assessors. The assessment group was convened by D. de B., who was also involved in over 80% of the assessments. The scoring system used has been
RCA
Fig. 1. Diagram of coronary arterial segments, after American Heart Association task force [7]. RCA = right coronary artery; LAD = left anterior descending; CFX = circumflex.
previously described [5]. Briefly, the coronary arterial tree was divided into 15 notional segments according to the scheme of the American Heart Association [7] (Fig. 1). Each segment was graded visually on a scale of 1 to 5 for stenoses (Table 1). The grading system allowed for segments to be noted as congenitally absent; as “unknown” (for example, if there were a proximal occlusion); or as filled by collaterals. Scoring was validated by comparing independent assessments. Patients suffering angina were consecutive cases undergoing coronary arteriography for ischaemic heart disease at the Royal Infirmary of Edinburgh. The policy for selecting patients for angiography during this period was for angiography to be performed with a view to subsequent bypass grafting in patients with symptomatic angina, despite adequate medical therapy, who had a positive treadmill exercise test. Patients were not considered for angiography if there was severe functional
TABLE 1 Scale for semiquantitative segment analysis. 0 = No stenosis 1 = Diameter stenosis < 50% 2 = Diameter stenosis 50-90s 3 = Diameter stenosis > 908, complete filling of distal vessel within 3 cycles 4 = Subtotal occlusion, no complete filling within 3 cycles 5 = Occluded vessel 6 = Congenitally absent 7 = Unknown, for instance because of proximal occlusion 8 = Retrograde collateral filling
315
impairment of the left ventricle (ejection, fraction < 20%). Patients with previous myocardial infarction were not excluded. There was no formal age limit to investigation. All the angiograms in the “angina group” were reviewed and re-analysed by A.K.B. and D. de B. using the same criteria as for those suffering a myocardial infarction. All data were stored in coded form for computer analysis. The data on grades of stenosis in individual segments for each patient were used to compute a score for the extent of coronary arterial disease. This was defined as the number of segments in which there was a stenosis greater than 50% luminal diameter. Segments in which there was complete occlusion of the coronary vessels (stenosis grade = 5) were included but segments which were congenitally absent, unassessable because of proximal occlusion or filled by collaterals were excluded (codes 6, 7 and 8). An attempt was made to fit a Poisson distribution curve to the data for numbers of stenosed segments in the myocardial infarct group. The equation used was:
where P, is the probability of finding n = X,2.. . stenosed segments and m represents the hypothetical prevalence of stenosed segments in the parent population. e is the base of natural logarithms, n! is the factorial of n; m was derived using an iterative approach to give the best fit to the experimental data, predicted and observed frequencies were compared using the Chi-square test. The equation assumes that the probability of having a coronary thrombosis is proportional to the number of stenoses, but is otherwise constant from patient to patient. The coded data for the “angina” group were used to identify two subgroups. One subgroup (“isolated stenosis”) consisted of patients with a stenosis > 70% in a single coronary segment and no other stenosis identified. The second subgroup (“diffuse disease”) consisted of patients with a stenosis of > 50% in at least eight different segments of the coronary arteries. Records of patients in these two subgroups were reviewed and infor-
50
40 30 20 10 0 1
2
3
4
5
6
8
7
9 101112131415
SEGMENTS Fig. 2. Segmental stenosis grades in 466 patients presenting with angina. Vertical axis indicates percentage of patients with occlusion, stenosis or collateral filling of the segment indicated on the horizontal axis. “Occluded” means grade 4 and grade 5 (Table l), “stenosed” means grade 2 and grade 3.
mation collected on weight, height. age, smoking habits, hypertension or diabetes mellitus (defined as currently receiving medication for either condition), alcohol ingestion, haemoglobiu, plasma total cholesterol, fasting plasma triglycerides, and fasting plasma glucose concentration. Statistical tests used in comparing the groups were Wilcoxon’s test for unpaired samples, Pearson’s correlation coefficient r, and the Chi square test. Odds ratios and confidence intervals were calculated according to Morris and Gardner [S]. 50
40
30
20
IO
0 1
2
3
4
5
6
7
8
9 101112131415
SEGMENTS Fig. 3. Segmental stenosis grades in 193 patients with acute myocardial infarction treated with thrombolytic agents. (key as for Fig. 2).
316
Results
PATIENTS 30R
There were 466 patients suffering angina and 258 who had undergone myocardial infarction. Of the latter, 193 had received thrombolytic therapy and 65 had received a placebo. The segment by segment distributions of grades of stenosis in those with angina; the patients suffering myocardial in-
20%
10%
TABLE 2
O-69
Distribution of occluded segments (grades 4 and 5) in patients with myccardial infarction treated with and without a thrombolytic agent. Numbers of occluded segments No thrombolysis (n = 65)
Bight coronary
Left main Left anterior descending Circumflex Total
Thrombolytic used (n =193)
Segment *: 10 1 5 2 5 3+4 0 5 8 6 15 I 2 8+9+10 9 11 to15 54 1 to15
12 19 11 0 13 15 11 12 93
Odds ratio
2.74 (1.12 to 6.69) 0.85 (0.3 to 2.4) 1.37 (0.46 to 4.17) 1.9 3.9 0.52 2.42 4.74
(0.76 to 4.9) (1.77 to 8.67) (0.11 to 2.43) (0.97 to 6.05) (2.36-9.44)
* Segments have been pooled where numbers of occlusions small.
are
Fig. 5. Age distribution
>70
YEARS
of patients having myocardial tion and angina.
infarc-
farction treated with thrombolytic therapy and the patients with infarction receiving placebo are given in Figs. 2-4. The principal difference between the patients having thrombolytic therapy as opposed to those receiving placebo is the increased prevalence of totally or subtotally occluded segments in the latter (odds ratio 4.74, 95% confidence intervals 2.38-9.44). The majority of the occluded segments in the patients receiving placebo occur proximally (Table 2). 83% of patients with myocardial infarction, and 79% of patients with stable angina, were male (P = NS). The age distribution of the groups with angina and pooled myocardial infarction is shown in Fig. 5. The mean age of those having myocardial infarction was 53.7 years, and of those with 45 40 35
n
ANGINA
q
INFARCTION
30 25 20 15 10 5
1
2
3
4
5
6
7
8
9 101112131415
SEGMENTS Fig. 4. Segmental stenosis grades in 65 patients with acute myocardial infarction not treated with thrombolytic agents. (key as for Fig. 2).
0 0
12
3
4
5
(I
7
>7
Fig. 6. Histogram showing the numbers of patients with 0, 1, 2.. . > 7 segments having a diameter stenosis z- 50% in those with angina and myocardial infarction.
317 PATIENTS 100
1 2
3
I
I
4
5
I
I 6
7
I
a
1 9
Number of segments wth stenosis> 50%
Fig. 7. Distribution of patients with 0, 1, 2.. segments having a diameter stenosis > 50% in patients with myocardial infarction (from Fig. 4) compared with a theoretical distribution based on a random distribution of stenoses (vertical bars). There is reasonably good correlation for segments with 1, 2, 3 stenoses (Sigma Chi squared = 2.7). The value m in n in the equation
corresponds
to the average number of stenosed segments in the “parent population”. providing the best fit to the data.
angina 56.7years (P = NS). Linear regression analysis indicated a weak positive relationship between age and number of stenosed segments: r = 0.142 (P c 0.01) for those with angina and r = 0.145 (P < 0.05) for those with myocardial infarction. After correction of the scores for the effect of age, the mean score was 2.61 for those with angina and 2.07 for those with myocardial infarction (P < 0.001). The distribution of the number of
ANGINA GROUP 466 PTS
Fig. 8. Distribution
The value m = 0.6 was derived empirically
as
stenosed segments in the two groups is shown in Fig. 6. The odds ratio for a score greater than 1 (i.e. stenosis affecting more than one coronary segment) in those with angina compared with those receiving placebo was 1.52 (95% confidence intervals 1.12-2.08). The theoretical model described above, which compared the distribution of the number of stenosed segments with a random Poisson distribution, gave the best fit with a value for
FIRST MI GROUP 258 PTS
of l-, 2- and 3-vessel disease in patients (pts) with angina and myocardial
infarction (MI)
318 TABLE
3
TABLE
Odds ratios and confidence intervals for stenosis > 50% in proximal, mid and distal portions of coronary arteries in the patients with angina compared with those haying myocardial infarction.
Right coronary Proximal (1) Mid (2) Distal (3 + 4) Left main (5) Left anterior descending Proximal (6) Mid (7) Distal (8 + 9 + 10) Circumflex Proximal (11) Mid (12 + 13) Distal (14+ 15)
Odds ratio
95% confidence interval
2.25 1.17 1.2 12.1
1.57 to 3.22 1.2 to 2.56 0.82 to 1.95 1.6 to 90
3.59 0.61 1.16
2.58 to 4.99 0.44 to 0.86 0.87 to 1.56
2.45 1.27 3.69
1.56 to 3.87 0.97 to 1.49 1.39 to 9.49
DinUse
Left main Left anterior descending
Isolated stenosis
.
Rank order of segments with stenosis angina and myocardial infarction.
Right coronary
m (equivalent to the mean number of stenosed segments per patient in the hypothetical parent population) of 0.6 (Fig. 7). There was fairly good correlation for n = 1, 2, 3 (Chi square 2.71, 0.3 > P > 0.2) but less good correlation for n > 3. The segmental data can also be used to divide patients into groups of patients with disease in single, double or three vessels. The distribution of these in patients suffering myocardial infarction and angina is shown in Fig. 8. The odds ratio for
disease
4 > 50% in patients
Segment
Angina
Myocardial infarction
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
2= 2= 4 10 14 1 5 10 = 6 10 = 7 8= 8= 13 15
5 3 8 12 14 2 1 11 3= 12 9 6= 6= 13 15
with
two or three vessel disease (defined as a diameter stenosis > 50% in segments in two or three different vessels) in the patients with angina compared with those suffering a first myocardial infarction was 2.56 (95% confidence intervals 1.87 to 3.52). The groups with angina and pooled infarction were further compared in Table 3, which lists the odds ratio and 95% confidence intervals for stenosis > 50% in proximal, middle and distal portions of each coronary artery in the patients with angina
Isolated stenosis
Diffuse lisease
Diffuse disease nlYOl 12
0 :
;_
TOTAL CHOLESTEROL Fig. 9. Distribution
of concentrations catheterisation
of plasma in patients
(I
:ASTING TRIGLYCERIDE
’
FASTING SUGAR
total cholesterol, fasting triglyceride and fasting blood with isolated stenosis or diffuse coronary arterial disease.
glucose
at the time of
319 TABLE
5
Comparison of frequencies of risk factors in patients with angina and “isolated stenosis” or “diffuse” patterns of coronary arterial disease. Diffuse disease
Isolated
Height Weight
M=49, F=9 (85% M) 1.69 metre 74.78 kg
Smokers Gigs/day Alcohol Units/wk
73% 16.8 378 5.3
High BP Diabetes Hb (g/dI) Cholesterol Triglycerides Fasting sugar
32% 6/58 14.24 6.04 mM 2.11 mM 6.24 mM
M=37,F=8 (82% M) 1.69 metre 74.66 kg 77% 16.7 40% 3.8 28% o/45 14.75 6.72 2.21 5.60
Gender
stenosis
P
NS NS NS NS NS NS NS i 0.05 NS c 0.05 NS NS
as compared with those having myocardial infarction and Table 4, which gives the rank order of segments with stenosis of greater than 50% in the two groups. Within the patients suffering angina, the distributions of concentrations of total cholesterol and fasting triglycerides in the plasma, fasting glucose in the blood in the subgroups with “isolated stenosis” and “diffuse disease” are shown in Fig. 9. Other “risk factors” are listed in Table 5. Discussion Differences between patients suffering myocardial infarction treated with and without a thrombolytic agent The differences in perfusion of vessels feeding an infarcted area in patients treated with and without a thrombolytic agent have been well documented in previous studies [5,6]. At segmental level (Table 2) retrograde propagation of thrombus from the site of its initiation often leads to occlusion of a more proximal segment of the vessel. This is more likely to occur in the right coronary artery, which has few major side branches, and probably explains the excess of proximal (segment 1) occlusions in the right coronary artery and the relative lack of occlusions
in the mid-right coronary artery (segment 2). It has been claimed [9] that occlusions of the right coronary artery are more resistant to thrombolysis, but this was not the case in the studies from which the present data are derived. In the left anterior descending coronary artery, thrombus seldom extends retrogradely beyond the first septal perforating artery, which helps to explain the excess of occlusions in segment 7 as opposed to segment 6. A further reason for the paucity of proximal stenoses in the left anterior descending artery and the main stem in the group suffering myocardial infarction is that sudden occlusion at these sites is likely to be associated with a high immediate mortality. Such patients may not survive to undergo angiography. Differences between the groups with myocardial infarction and angina The groups are broadly similar in the distribution of coronary stenosis (Table 4) but differ markedly in the extent of stenotic disease (Figs. 6 and 9, Table 3). Patients with established stable angina tended to have a higher proportion of multi vessel, multi segment disease, more occluded segments and more segments filling from collaterals. Since the groups were selected on the basis of clinical symptoms and electrocardiographic changes (evidence of infarction or a positive exercise test), the predominance of proximal lesions in both is to be expected. In both groups, the modal number of stenotic segments was one, that is to say, patients tended to present with significant stenosis confined to a single coronary arterial segment. It should be emphasised that the detection of a stenosis is not the same as the detection of coronary atheroma, although the two are loosely correlated [lo]. Coronary angiography is only capable of detecting atheromatous lesions which obtrude into the coronary arterial lumen: in comparison with post-mortem studies it will consistently underestimate the number of atheromatous lesions. Plaques which do not obstruct the lumen may, nevertheless, rupture and cause occlusive thrombus - this presumably accounts for the
320
4.3% of patients suffering myocardial infarction with “no stenosis > 50%“. In the 8.8% of patients having stable angina with no significant stenosis, the clinical diagnosis of angina may have been erroneous, or other mechanisms such as “small vessel disease” may have been involved. We included this group in our analysis because to do so would reduce rather than increase the differences between those suffering myocardial infarction and angina. Weak correlation stenoses
between
age and number
of
The difference between the groups in terms of the extent of stenotic disease cannot be explained by differing age structures of the two populations. The relationship between age and number of affected segments, though significant because of the large numbers involved, was very weak (1 - r 2 between 0.97 and 0.98). The weakness of the correlation between age and number of stenoses in those with angina was somewhat surprising. Age was explicitly not a contraindication to investigation in our unit, but it is conceivable that older patients with impaired ventricular function might not have been offered angiography. On the other hand, older patients might be expected to have had a higher level of symptoms before referral. A more likely explanation is that clinical angina is usually due to a limited number of severe stenoses rather than to the extent of diffuse mural coronary disease. This would accord with the relatively poor correlation described by others between the extent of coronary arterial disease and the results of exercise testing [ll]. Myocardial infarction as a random event in a population with few stenoses per patient Myocardial infarction is usually due to thrombosis on a ruptured atheromatous plaque [12]. The risk of infarction is partly dependent on the number of plaques and partly on other factors which affect plaque stability and susceptibility to thrombosis. The present data would be compatible with a model in which most first infarcts occur in a population with a low prevalence of arterial
stenosis (Fig. 7), and in which the majority of patients will, therefore, have only a single lesion. The distribution of stenoses in our population with infarction correlates closely with that described by Koren and colleagues in an Israeli population [13], and approximates to that described by Hamsten and colleagues [lo] and by Hughes and colleagues [14] in white male patients suffering their first infarction. Others who have studied the distribution of artery lesions in patients subsequent to infarction [15,16] have shown a more nearly equal distribution of single-, double- and triple-vessel disease. Their population, however, consisted of patients who had had one, two or sometimes three infarcts compared with the population studied here suffering their initial infarction. It has been suggested that patients of Asian or Mediterranean origin suffering infarction have a tendency to a more diffuse distribution of coronary arterial stenoses. We cannot comment on this, as our study populations were almost exclusively North European in origin. Why do patients with angina have more stenoses? Conversely, the population presenting with angina in the present study largely represents patients who have become symptomatic because of the relatively gradual progression of coronary arterial stenoses. A proportion will also have suffered myocardial infarction, but this was explicitly not the reason for angiography in our study, and patients with severely impaired left ventricular function were excluded. Patients with angina also constitute a selected group but, in this instance, the bias will be towards excluding those with major selective risk factors for thrombosis. This may help to explain the greater extent of coronary stenotic disease in the patients with angina. A similar argument may help to explain the excess of stenosis in the main stem of the left coronary artery and its proximal anterior descending branch in those with angina: gradual development of such lesions will permit the formation of an adequate collateral circulation. Longitudinal studies involving repeated angiography [17,18] have, in general, shown that stenoses tend to progress with time. Lesions which develop
321
rapidly from angiographically “normal” or mildly stenosed segments are more likely to progress to complete occlusion and myocardial infarction. Stenoses greater than 90’7%are also likely to become occluded, but this occlusion may be clinically silent if it occurs sufficiently gradually for a collateral circulation to develop. Our work supports the concept that patients presenting with acute infarction and with chronic stable angina constitute different subsets of the overall group of patients with atheromatous coronary arterial disease.
with ischaemic heart disease from the subgroups who present with infarction or stable angina particularly when generalising from one subgroup to the other. Second, they indicate that the impact on the overall incidence of myocardial infarction of identifying and intervening in patients with stable angina is likely to be small. Finally, the heterogeneity within the population with stable angina provides scope for further studies on the interaction between specific risk factors and the distribution of arterial stenoses.
References Subsets within patients having angina Within the overall patterns of distribution of stenoses in the patients with both infarction and angina, there is clearly considerable heterogeneity. We attempted to see whether subpopulations of those with angina selected for having isolated stenoses or diffuse disease differed in respect of a limited population of risk factors. Obviously, there is a continuous range of patterns of stenosis between these extremes, but we chose these groups for maximum potential contrast. Diabetes was more common in the group with diffuse disease and total cholesterol in those with isolated stenoses. The latter is most closely analogous to the pattern of disease in the patients having a first myocardial infarction and it is noteworthy that Hamsten and colleagues have identified elevated levels of apolipoprotein B in the plasma as the major risk marker correlating with the extent of atheroma in a group of young male patients after an infarction [19]. Hangartner et al. [19] have pointed out that, although many individuals have a mixed population of plaques, some fibrous and some fatty, certain individuals have a predilection for specific type of plaque. We do not yet have any way of inferring the composition of plaque from angiographic appearances. Practical implications There are three principal practical implications of these data. First, they emphasise the need for caution in generalising to the overall population
1 Montenegro MR. Eggen DA. Topography of atherosclerosis in the coronary arteries. In: McGill HC, ed. The geographic pathology of atheroma. Baltimore: Williams & Wilkins 1968;126-133. 2 Proudfit WL, Shirey EK, Sones FM Jr. Selective tine coronary arteriography: correlation with clinical findings in 1000 patients. Circulation 1966;33:90-910. 3 Proudfit WL, Bruschle AVG, Sones FM. Natural history of obstructive coronary disease: ten year study of 601 non surgical cases. Progr Cardiovasc Dis 1978;21:53-64. R. Computerized tabulation of tine coronary 4 Favaloro angiograms: its implication for results of randomized trials. Circulation 1990;81:1992-2003. 5 Verstraete M, Bleifeld W, Brower RW et al. Double blind randomized trial of intravenous recombinant tissue type plasminogen activator versus placebo in acute myocardial infarction. Lancet 1985;ii:965-969. 6 Verstraete M, Bernard R, Bory M et al. Randomized trial of intravenous recombinant tissue type plasminogen activator versus intravenous streptokinase in acute myocardial infarction. Lancet 1985;i:842-847. 7 Austen WG, Edwards JE, Faye RL et al. AHA committee report: a reporting system on patients evaluate for coronary artery disease. Report of the Ad Hoc Committee for grading of coronary artery disease; Council of Cardiovascular Surgery, American Heart Association. Circulation 1975; 51:7-21. 8 Morris JA, Gardner MJ. Calculating confidence intervals for relative risks (odds ratios) and standardised ratios and rates. Br Med J 1988;296:1313-1315. 9 Bates ER, Califf RM, Stack RS et al. Thrombolysis and angioplasty in myocardial infarction(TAMI-1) trial: influence of infarct location on arterial patency, left ventricular function and mortality. J Am Co11 Cardiol 1989;13:1218. 10 Hamsten A, Walldius G, Szamosi A, Dahlen G, de Faire U. Relationship of angiographically defined coronary artery disease to serum lipoprotein and apolipoprotein concentrations in young survivors of myocardial infarction. Circulation 1986;73:1097-1110.
322 11 Balcon
12
13
14
15
R, Brooks N, Layton C. Correlation of heart rate/ST slope and coronary angiographic findings. Br Heart J 1984;52:304-307. Davies MJ, Thomas AC. Plaque fissuring: the cause of acute myocardial infarction, sudden ischaemic death and crescendo angina. Br Heart J 1985;53:363-373. Koren G, Weiss AT, Hasin Y et al. Prevention of myocardial damage in acute myocardial ischaemia by early treatment with intravenous streptokinase. N Engl J Med 1985;313:1384-1389. Hughes LO, Raval U, Raftery EB. First myocardial infarctions in British Asian and white men. Br Med J 1989;298:1340-1345. Sam G, Castover A, Betrui A. Determinants of prognosis in survivors of myocardial infarction: a prospective clinical angiographic study. N Engl J Med 1982;306:1065-1069.
16 Bertrand ME, Lefebre JM, Laisne CL, Rousseau MF, Carre AG, Le Kieffre JP. Coronary arteriography in acute transluminal myocardial infarction. Am Heart J 1979;97:61-69. 17 Brushke AVG. Wijers TS, Kolsters W, Londman J. The anatomic evolution of coronary artery disease demonstrated by coronary arteriography in 256 non-operated patients. Circulation 1981;63:527-536. 18 Singh RN. Progressions of coronary atherosclerosis: clues to pathogenesis from serial coronary arteriography. Br Heart J 1984;32:451-461. 19 Hangartner JRW, Thomas AC. Morphological characteristics of clinical significant coronary artery stenoses in stable angina. Br Heart J 1986;56:501-508.
International Journal of Cardiologv, 32 (1991) 313-322 0 1991 Elsevier Science Publishers B.V. 0167-5273/91 ADONIS 016752739100189A
CARD10
01288
Segmental analysis of coronary arterial stenoses in patients presenting with angina or first myocardial infarction D.P. Department
de Bono
and AK.
of Cardiology,
Bhattacharrya
*
University of Leicester, Ikeester,
U.K.
(Received 26 November 1990; revision accepted 14 February 1991)
De Bono DP, Bhattacharrya AK. Segmental analysis of coronary arterial with angina or first myocardial infarction. Int J Cardiol 1991;32:313-322.
stenoses
in patients
presenting
The segmental distribution of stenoses within the coronary arteries was anafysed in a population of 258 patients with a first myocardial infarction undergoing coronary angiography to evaluate the effect of thrombolytic therapy, and in a population of 466 patients undergoing elective coronary angiography for stable angina. Mean ages were 53.7 and 56.7 years respectively (P = NS). As judged angiographically, coronary arterial disease was more extensive in the group suffering angina, with a greater proportion of patients with two- or three-vessel disease (odds ratio 2.56, 95% confidence interval 1.87 to 3.52) and more patients having stenoses in two or more coronary arterial segments (odds ratio 1.52,95% confidence interval 1.12 to 2.08). For each coronary vessel, the probability of finding a stenosis > 50% in an individual segment was greater in the group presenting with angina. There was a relative deficiency of stenoses within the main stem of the left coronary artery or its proximal left anterior descending branch among the patients suffering myocardii infarction. Within those having angina, subgroups were identified with “isolated” and “diffuse” coronary arterial disease: the latter patients tended to have a lower concentration of total cholesterol in the serum, but an increased prevalence of diabetes mellitus. Patients presenting clinically with a first myocardial infarction, and patients with severe angina, constitute distinct populations selected by different mechanisms from the overall pool of patients with atheromatous coronary arterial disease. Key words:
Coronary
atheroma;
Angiographic
analysis;
Introduction The distribution of atheromatous lesions in the coronary arterial tree is important for understand-
Correspondence to: Prof. D.P. de Bono, Dept. of CardioIogy, Clinical Sciences Wing, Glenfield General Hospital, Leicester LE3 9QP, U.K. * Present address: A.K. Bhattacharrya, Gawhati State University, Assam, India.
Myocardial
infarction;
Stable angina
ing the pathogenesis and natural history of coronary arterial disease. This coronary “ topography” has been extensively studied in autopsy material [l]. Coronary arteriography enables similar information to be obtained during life, and to be correlated with clinical findings [2] and the natural progression of disease [3]. The advent of computerised systems for tabulating cineangiographic data [4] has made it easier to compare patterns of disease in different populations of patients.
314
The European Co-operative Study Group trials of coronary thrombolysis in acute myocardial infarction provided an opportunity for centralised reading and analysis of coronary arteriograms from a relatively large population of patients studied shortly after the onset of symptoms of acute myocardial infarction [5,6]. Besides identifying the presumed stenosis responsible for infarction, and assessing its effect on arterial patency, the distribution of stenoses in the entire arterial tree was recorded in a form suitable for subsequent analysis. For comparison, we applied a similar analysis to a series of 466 patients undergoing angiography for ischaemic heart disease at one of the centres within the Co-operative study group. In this paper, we describe and compare the distribution of arterial lesions in the two populations.
Patients and Methods The “myocardial infarction” group consisted of patients taking part in two simultaneous multicentric trials of thrombolytic therapy in acute myocardial infarction. In one trial, recombinant human tissue type plasminogen activator was being compared with placebo, and in the other trial with intravenous streptokinase. Patients admitted to the trial were under the age of 71, had had chest pain typical of myocardial infarction for 30 minutes or longer, and fulfilled stringent electrocardiographic criteria for myocardial ischaemia (> 2 mV ST segment elevation in at least two precordial leads or > 1 mV ST segment elevation in at least two frontal plane leads). To assist in the unequivocal identification of the vessel responsible for infarction, only patients with a first myocardial infarction were studied. Detailed protocols for both studies have been published elsewhere [5,6]. Angiograms in those patients suffering myocardial infarction were taken as close as possible to 90 minutes after the start of administration of thrombolytic therapy which, in turn, was initiated within six hours of the onset of major symptoms. All angiograms were recorded on 35 mm tine film and were read centrally by a small group of assessors. The assessment group was convened by D. de B., who was also involved in over 80% of the assessments. The scoring system used has been
RCA
Fig. 1. Diagram of coronary arterial segments, after American Heart Association task force [7]. RCA = right coronary artery; LAD = left anterior descending; CFX = circumflex.
previously described [5]. Briefly, the coronary arterial tree was divided into 15 notional segments according to the scheme of the American Heart Association [7] (Fig. 1). Each segment was graded visually on a scale of 1 to 5 for stenoses (Table 1). The grading system allowed for segments to be noted as congenitally absent; as “unknown” (for example, if there were a proximal occlusion); or as filled by collaterals. Scoring was validated by comparing independent assessments. Patients suffering angina were consecutive cases undergoing coronary arteriography for ischaemic heart disease at the Royal Infirmary of Edinburgh. The policy for selecting patients for angiography during this period was for angiography to be performed with a view to subsequent bypass grafting in patients with symptomatic angina, despite adequate medical therapy, who had a positive treadmill exercise test. Patients were not considered for angiography if there was severe functional
TABLE 1 Scale for semiquantitative segment analysis. 0 = No stenosis 1 = Diameter stenosis < 50% 2 = Diameter stenosis 50-90s 3 = Diameter stenosis > 908, complete filling of distal vessel within 3 cycles 4 = Subtotal occlusion, no complete filling within 3 cycles 5 = Occluded vessel 6 = Congenitally absent 7 = Unknown, for instance because of proximal occlusion 8 = Retrograde collateral filling
315
impairment of the left ventricle (ejection, fraction < 20%). Patients with previous myocardial infarction were not excluded. There was no formal age limit to investigation. All the angiograms in the “angina group” were reviewed and re-analysed by A.K.B. and D. de B. using the same criteria as for those suffering a myocardial infarction. All data were stored in coded form for computer analysis. The data on grades of stenosis in individual segments for each patient were used to compute a score for the extent of coronary arterial disease. This was defined as the number of segments in which there was a stenosis greater than 50% luminal diameter. Segments in which there was complete occlusion of the coronary vessels (stenosis grade = 5) were included but segments which were congenitally absent, unassessable because of proximal occlusion or filled by collaterals were excluded (codes 6, 7 and 8). An attempt was made to fit a Poisson distribution curve to the data for numbers of stenosed segments in the myocardial infarct group. The equation used was:
where P, is the probability of finding n = X,2.. . stenosed segments and m represents the hypothetical prevalence of stenosed segments in the parent population. e is the base of natural logarithms, n! is the factorial of n; m was derived using an iterative approach to give the best fit to the experimental data, predicted and observed frequencies were compared using the Chi-square test. The equation assumes that the probability of having a coronary thrombosis is proportional to the number of stenoses, but is otherwise constant from patient to patient. The coded data for the “angina” group were used to identify two subgroups. One subgroup (“isolated stenosis”) consisted of patients with a stenosis > 70% in a single coronary segment and no other stenosis identified. The second subgroup (“diffuse disease”) consisted of patients with a stenosis of > 50% in at least eight different segments of the coronary arteries. Records of patients in these two subgroups were reviewed and infor-
50
40 30 20 10 0 1
2
3
4
5
6
8
7
9 101112131415
SEGMENTS Fig. 2. Segmental stenosis grades in 466 patients presenting with angina. Vertical axis indicates percentage of patients with occlusion, stenosis or collateral filling of the segment indicated on the horizontal axis. “Occluded” means grade 4 and grade 5 (Table l), “stenosed” means grade 2 and grade 3.
mation collected on weight, height. age, smoking habits, hypertension or diabetes mellitus (defined as currently receiving medication for either condition), alcohol ingestion, haemoglobiu, plasma total cholesterol, fasting plasma triglycerides, and fasting plasma glucose concentration. Statistical tests used in comparing the groups were Wilcoxon’s test for unpaired samples, Pearson’s correlation coefficient r, and the Chi square test. Odds ratios and confidence intervals were calculated according to Morris and Gardner [S]. 50
40
30
20
IO
0 1
2
3
4
5
6
7
8
9 101112131415
SEGMENTS Fig. 3. Segmental stenosis grades in 193 patients with acute myocardial infarction treated with thrombolytic agents. (key as for Fig. 2).
316
Results
PATIENTS 30R
There were 466 patients suffering angina and 258 who had undergone myocardial infarction. Of the latter, 193 had received thrombolytic therapy and 65 had received a placebo. The segment by segment distributions of grades of stenosis in those with angina; the patients suffering myocardial in-
20%
10%
TABLE 2
O-69
Distribution of occluded segments (grades 4 and 5) in patients with myccardial infarction treated with and without a thrombolytic agent. Numbers of occluded segments No thrombolysis (n = 65)
Bight coronary
Left main Left anterior descending Circumflex Total
Thrombolytic used (n =193)
Segment *: 10 1 5 2 5 3+4 0 5 8 6 15 I 2 8+9+10 9 11 to15 54 1 to15
12 19 11 0 13 15 11 12 93
Odds ratio
2.74 (1.12 to 6.69) 0.85 (0.3 to 2.4) 1.37 (0.46 to 4.17) 1.9 3.9 0.52 2.42 4.74
(0.76 to 4.9) (1.77 to 8.67) (0.11 to 2.43) (0.97 to 6.05) (2.36-9.44)
* Segments have been pooled where numbers of occlusions small.
are
Fig. 5. Age distribution
>70
YEARS
of patients having myocardial tion and angina.
infarc-
farction treated with thrombolytic therapy and the patients with infarction receiving placebo are given in Figs. 2-4. The principal difference between the patients having thrombolytic therapy as opposed to those receiving placebo is the increased prevalence of totally or subtotally occluded segments in the latter (odds ratio 4.74, 95% confidence intervals 2.38-9.44). The majority of the occluded segments in the patients receiving placebo occur proximally (Table 2). 83% of patients with myocardial infarction, and 79% of patients with stable angina, were male (P = NS). The age distribution of the groups with angina and pooled myocardial infarction is shown in Fig. 5. The mean age of those having myocardial infarction was 53.7 years, and of those with 45 40 35
n
ANGINA
q
INFARCTION
30 25 20 15 10 5
1
2
3
4
5
6
7
8
9 101112131415
SEGMENTS Fig. 4. Segmental stenosis grades in 65 patients with acute myocardial infarction not treated with thrombolytic agents. (key as for Fig. 2).
0 0
12
3
4
5
(I
7
>7
Fig. 6. Histogram showing the numbers of patients with 0, 1, 2.. . > 7 segments having a diameter stenosis z- 50% in those with angina and myocardial infarction.
317 PATIENTS 100
1 2
3
I
I
4
5
I
I 6
7
I
a
1 9
Number of segments wth stenosis> 50%
Fig. 7. Distribution of patients with 0, 1, 2.. segments having a diameter stenosis > 50% in patients with myocardial infarction (from Fig. 4) compared with a theoretical distribution based on a random distribution of stenoses (vertical bars). There is reasonably good correlation for segments with 1, 2, 3 stenoses (Sigma Chi squared = 2.7). The value m in n in the equation
corresponds
to the average number of stenosed segments in the “parent population”. providing the best fit to the data.
angina 56.7years (P = NS). Linear regression analysis indicated a weak positive relationship between age and number of stenosed segments: r = 0.142 (P c 0.01) for those with angina and r = 0.145 (P < 0.05) for those with myocardial infarction. After correction of the scores for the effect of age, the mean score was 2.61 for those with angina and 2.07 for those with myocardial infarction (P < 0.001). The distribution of the number of
ANGINA GROUP 466 PTS
Fig. 8. Distribution
The value m = 0.6 was derived empirically
as
stenosed segments in the two groups is shown in Fig. 6. The odds ratio for a score greater than 1 (i.e. stenosis affecting more than one coronary segment) in those with angina compared with those receiving placebo was 1.52 (95% confidence intervals 1.12-2.08). The theoretical model described above, which compared the distribution of the number of stenosed segments with a random Poisson distribution, gave the best fit with a value for
FIRST MI GROUP 258 PTS
of l-, 2- and 3-vessel disease in patients (pts) with angina and myocardial
infarction (MI)
318 TABLE
3
TABLE
Odds ratios and confidence intervals for stenosis > 50% in proximal, mid and distal portions of coronary arteries in the patients with angina compared with those haying myocardial infarction.
Right coronary Proximal (1) Mid (2) Distal (3 + 4) Left main (5) Left anterior descending Proximal (6) Mid (7) Distal (8 + 9 + 10) Circumflex Proximal (11) Mid (12 + 13) Distal (14+ 15)
Odds ratio
95% confidence interval
2.25 1.17 1.2 12.1
1.57 to 3.22 1.2 to 2.56 0.82 to 1.95 1.6 to 90
3.59 0.61 1.16
2.58 to 4.99 0.44 to 0.86 0.87 to 1.56
2.45 1.27 3.69
1.56 to 3.87 0.97 to 1.49 1.39 to 9.49
DinUse
Left main Left anterior descending
Isolated stenosis
.
Rank order of segments with stenosis angina and myocardial infarction.
Right coronary
m (equivalent to the mean number of stenosed segments per patient in the hypothetical parent population) of 0.6 (Fig. 7). There was fairly good correlation for n = 1, 2, 3 (Chi square 2.71, 0.3 > P > 0.2) but less good correlation for n > 3. The segmental data can also be used to divide patients into groups of patients with disease in single, double or three vessels. The distribution of these in patients suffering myocardial infarction and angina is shown in Fig. 8. The odds ratio for
disease
4 > 50% in patients
Segment
Angina
Myocardial infarction
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
2= 2= 4 10 14 1 5 10 = 6 10 = 7 8= 8= 13 15
5 3 8 12 14 2 1 11 3= 12 9 6= 6= 13 15
with
two or three vessel disease (defined as a diameter stenosis > 50% in segments in two or three different vessels) in the patients with angina compared with those suffering a first myocardial infarction was 2.56 (95% confidence intervals 1.87 to 3.52). The groups with angina and pooled infarction were further compared in Table 3, which lists the odds ratio and 95% confidence intervals for stenosis > 50% in proximal, middle and distal portions of each coronary artery in the patients with angina
Isolated stenosis
Diffuse lisease
Diffuse disease nlYOl 12
0 :
;_
TOTAL CHOLESTEROL Fig. 9. Distribution
of concentrations catheterisation
of plasma in patients
(I
:ASTING TRIGLYCERIDE
’
FASTING SUGAR
total cholesterol, fasting triglyceride and fasting blood with isolated stenosis or diffuse coronary arterial disease.
glucose
at the time of
319 TABLE
5
Comparison of frequencies of risk factors in patients with angina and “isolated stenosis” or “diffuse” patterns of coronary arterial disease. Diffuse disease
Isolated
Height Weight
M=49, F=9 (85% M) 1.69 metre 74.78 kg
Smokers Gigs/day Alcohol Units/wk
73% 16.8 378 5.3
High BP Diabetes Hb (g/dI) Cholesterol Triglycerides Fasting sugar
32% 6/58 14.24 6.04 mM 2.11 mM 6.24 mM
M=37,F=8 (82% M) 1.69 metre 74.66 kg 77% 16.7 40% 3.8 28% o/45 14.75 6.72 2.21 5.60
Gender
stenosis
P
NS NS NS NS NS NS NS i 0.05 NS c 0.05 NS NS
as compared with those having myocardial infarction and Table 4, which gives the rank order of segments with stenosis of greater than 50% in the two groups. Within the patients suffering angina, the distributions of concentrations of total cholesterol and fasting triglycerides in the plasma, fasting glucose in the blood in the subgroups with “isolated stenosis” and “diffuse disease” are shown in Fig. 9. Other “risk factors” are listed in Table 5. Discussion Differences between patients suffering myocardial infarction treated with and without a thrombolytic agent The differences in perfusion of vessels feeding an infarcted area in patients treated with and without a thrombolytic agent have been well documented in previous studies [5,6]. At segmental level (Table 2) retrograde propagation of thrombus from the site of its initiation often leads to occlusion of a more proximal segment of the vessel. This is more likely to occur in the right coronary artery, which has few major side branches, and probably explains the excess of proximal (segment 1) occlusions in the right coronary artery and the relative lack of occlusions
in the mid-right coronary artery (segment 2). It has been claimed [9] that occlusions of the right coronary artery are more resistant to thrombolysis, but this was not the case in the studies from which the present data are derived. In the left anterior descending coronary artery, thrombus seldom extends retrogradely beyond the first septal perforating artery, which helps to explain the excess of occlusions in segment 7 as opposed to segment 6. A further reason for the paucity of proximal stenoses in the left anterior descending artery and the main stem in the group suffering myocardial infarction is that sudden occlusion at these sites is likely to be associated with a high immediate mortality. Such patients may not survive to undergo angiography. Differences between the groups with myocardial infarction and angina The groups are broadly similar in the distribution of coronary stenosis (Table 4) but differ markedly in the extent of stenotic disease (Figs. 6 and 9, Table 3). Patients with established stable angina tended to have a higher proportion of multi vessel, multi segment disease, more occluded segments and more segments filling from collaterals. Since the groups were selected on the basis of clinical symptoms and electrocardiographic changes (evidence of infarction or a positive exercise test), the predominance of proximal lesions in both is to be expected. In both groups, the modal number of stenotic segments was one, that is to say, patients tended to present with significant stenosis confined to a single coronary arterial segment. It should be emphasised that the detection of a stenosis is not the same as the detection of coronary atheroma, although the two are loosely correlated [lo]. Coronary angiography is only capable of detecting atheromatous lesions which obtrude into the coronary arterial lumen: in comparison with post-mortem studies it will consistently underestimate the number of atheromatous lesions. Plaques which do not obstruct the lumen may, nevertheless, rupture and cause occlusive thrombus - this presumably accounts for the
320
4.3% of patients suffering myocardial infarction with “no stenosis > 50%“. In the 8.8% of patients having stable angina with no significant stenosis, the clinical diagnosis of angina may have been erroneous, or other mechanisms such as “small vessel disease” may have been involved. We included this group in our analysis because to do so would reduce rather than increase the differences between those suffering myocardial infarction and angina. Weak correlation stenoses
between
age and number
of
The difference between the groups in terms of the extent of stenotic disease cannot be explained by differing age structures of the two populations. The relationship between age and number of affected segments, though significant because of the large numbers involved, was very weak (1 - r 2 between 0.97 and 0.98). The weakness of the correlation between age and number of stenoses in those with angina was somewhat surprising. Age was explicitly not a contraindication to investigation in our unit, but it is conceivable that older patients with impaired ventricular function might not have been offered angiography. On the other hand, older patients might be expected to have had a higher level of symptoms before referral. A more likely explanation is that clinical angina is usually due to a limited number of severe stenoses rather than to the extent of diffuse mural coronary disease. This would accord with the relatively poor correlation described by others between the extent of coronary arterial disease and the results of exercise testing [ll]. Myocardial infarction as a random event in a population with few stenoses per patient Myocardial infarction is usually due to thrombosis on a ruptured atheromatous plaque [12]. The risk of infarction is partly dependent on the number of plaques and partly on other factors which affect plaque stability and susceptibility to thrombosis. The present data would be compatible with a model in which most first infarcts occur in a population with a low prevalence of arterial
stenosis (Fig. 7), and in which the majority of patients will, therefore, have only a single lesion. The distribution of stenoses in our population with infarction correlates closely with that described by Koren and colleagues in an Israeli population [13], and approximates to that described by Hamsten and colleagues [lo] and by Hughes and colleagues [14] in white male patients suffering their first infarction. Others who have studied the distribution of artery lesions in patients subsequent to infarction [15,16] have shown a more nearly equal distribution of single-, double- and triple-vessel disease. Their population, however, consisted of patients who had had one, two or sometimes three infarcts compared with the population studied here suffering their initial infarction. It has been suggested that patients of Asian or Mediterranean origin suffering infarction have a tendency to a more diffuse distribution of coronary arterial stenoses. We cannot comment on this, as our study populations were almost exclusively North European in origin. Why do patients with angina have more stenoses? Conversely, the population presenting with angina in the present study largely represents patients who have become symptomatic because of the relatively gradual progression of coronary arterial stenoses. A proportion will also have suffered myocardial infarction, but this was explicitly not the reason for angiography in our study, and patients with severely impaired left ventricular function were excluded. Patients with angina also constitute a selected group but, in this instance, the bias will be towards excluding those with major selective risk factors for thrombosis. This may help to explain the greater extent of coronary stenotic disease in the patients with angina. A similar argument may help to explain the excess of stenosis in the main stem of the left coronary artery and its proximal anterior descending branch in those with angina: gradual development of such lesions will permit the formation of an adequate collateral circulation. Longitudinal studies involving repeated angiography [17,18] have, in general, shown that stenoses tend to progress with time. Lesions which develop
321
rapidly from angiographically “normal” or mildly stenosed segments are more likely to progress to complete occlusion and myocardial infarction. Stenoses greater than 90’7%are also likely to become occluded, but this occlusion may be clinically silent if it occurs sufficiently gradually for a collateral circulation to develop. Our work supports the concept that patients presenting with acute infarction and with chronic stable angina constitute different subsets of the overall group of patients with atheromatous coronary arterial disease.
with ischaemic heart disease from the subgroups who present with infarction or stable angina particularly when generalising from one subgroup to the other. Second, they indicate that the impact on the overall incidence of myocardial infarction of identifying and intervening in patients with stable angina is likely to be small. Finally, the heterogeneity within the population with stable angina provides scope for further studies on the interaction between specific risk factors and the distribution of arterial stenoses.
References Subsets within patients having angina Within the overall patterns of distribution of stenoses in the patients with both infarction and angina, there is clearly considerable heterogeneity. We attempted to see whether subpopulations of those with angina selected for having isolated stenoses or diffuse disease differed in respect of a limited population of risk factors. Obviously, there is a continuous range of patterns of stenosis between these extremes, but we chose these groups for maximum potential contrast. Diabetes was more common in the group with diffuse disease and total cholesterol in those with isolated stenoses. The latter is most closely analogous to the pattern of disease in the patients having a first myocardial infarction and it is noteworthy that Hamsten and colleagues have identified elevated levels of apolipoprotein B in the plasma as the major risk marker correlating with the extent of atheroma in a group of young male patients after an infarction [19]. Hangartner et al. [19] have pointed out that, although many individuals have a mixed population of plaques, some fibrous and some fatty, certain individuals have a predilection for specific type of plaque. We do not yet have any way of inferring the composition of plaque from angiographic appearances. Practical implications There are three principal practical implications of these data. First, they emphasise the need for caution in generalising to the overall population
1 Montenegro MR. Eggen DA. Topography of atherosclerosis in the coronary arteries. In: McGill HC, ed. The geographic pathology of atheroma. Baltimore: Williams & Wilkins 1968;126-133. 2 Proudfit WL, Shirey EK, Sones FM Jr. Selective tine coronary arteriography: correlation with clinical findings in 1000 patients. Circulation 1966;33:90-910. 3 Proudfit WL, Bruschle AVG, Sones FM. Natural history of obstructive coronary disease: ten year study of 601 non surgical cases. Progr Cardiovasc Dis 1978;21:53-64. R. Computerized tabulation of tine coronary 4 Favaloro angiograms: its implication for results of randomized trials. Circulation 1990;81:1992-2003. 5 Verstraete M, Bleifeld W, Brower RW et al. Double blind randomized trial of intravenous recombinant tissue type plasminogen activator versus placebo in acute myocardial infarction. Lancet 1985;ii:965-969. 6 Verstraete M, Bernard R, Bory M et al. Randomized trial of intravenous recombinant tissue type plasminogen activator versus intravenous streptokinase in acute myocardial infarction. Lancet 1985;i:842-847. 7 Austen WG, Edwards JE, Faye RL et al. AHA committee report: a reporting system on patients evaluate for coronary artery disease. Report of the Ad Hoc Committee for grading of coronary artery disease; Council of Cardiovascular Surgery, American Heart Association. Circulation 1975; 51:7-21. 8 Morris JA, Gardner MJ. Calculating confidence intervals for relative risks (odds ratios) and standardised ratios and rates. Br Med J 1988;296:1313-1315. 9 Bates ER, Califf RM, Stack RS et al. Thrombolysis and angioplasty in myocardial infarction(TAMI-1) trial: influence of infarct location on arterial patency, left ventricular function and mortality. J Am Co11 Cardiol 1989;13:1218. 10 Hamsten A, Walldius G, Szamosi A, Dahlen G, de Faire U. Relationship of angiographically defined coronary artery disease to serum lipoprotein and apolipoprotein concentrations in young survivors of myocardial infarction. Circulation 1986;73:1097-1110.
322 11 Balcon
12
13
14
15
R, Brooks N, Layton C. Correlation of heart rate/ST slope and coronary angiographic findings. Br Heart J 1984;52:304-307. Davies MJ, Thomas AC. Plaque fissuring: the cause of acute myocardial infarction, sudden ischaemic death and crescendo angina. Br Heart J 1985;53:363-373. Koren G, Weiss AT, Hasin Y et al. Prevention of myocardial damage in acute myocardial ischaemia by early treatment with intravenous streptokinase. N Engl J Med 1985;313:1384-1389. Hughes LO, Raval U, Raftery EB. First myocardial infarctions in British Asian and white men. Br Med J 1989;298:1340-1345. Sam G, Castover A, Betrui A. Determinants of prognosis in survivors of myocardial infarction: a prospective clinical angiographic study. N Engl J Med 1982;306:1065-1069.
16 Bertrand ME, Lefebre JM, Laisne CL, Rousseau MF, Carre AG, Le Kieffre JP. Coronary arteriography in acute transluminal myocardial infarction. Am Heart J 1979;97:61-69. 17 Brushke AVG. Wijers TS, Kolsters W, Londman J. The anatomic evolution of coronary artery disease demonstrated by coronary arteriography in 256 non-operated patients. Circulation 1981;63:527-536. 18 Singh RN. Progressions of coronary atherosclerosis: clues to pathogenesis from serial coronary arteriography. Br Heart J 1984;32:451-461. 19 Hangartner JRW, Thomas AC. Morphological characteristics of clinical significant coronary artery stenoses in stable angina. Br Heart J 1986;56:501-508.
