232

Lumen Diameter of Normal

Human Coronary Arteries
Influence of Age, Sex, Anatomic Variation, and Left
Ventricular Hypertrophy or Dilation
J. Theodore Dodge Jr., MD; B. Greg Brown, MD, PhD;
Edward L. Bolson, MS; and Harold T. Dodge, MD

Background. Precise knowledge of the expected "normal" lumen diameter at a given coronary anatomic
location is a first step toward developing a quantitative estimate of coronary disease severity that could be
more useful than the traditional "'percent stenosis."
Methods and Resuls. Eighty-three arteriograms were carefuly selected from among 9,160 consecutive
studies for their smooth lumen borders indicating freedom from atherosclerotic disease. Of these, 60 men
and 10 women had no abnormalities of cardiac function, seven men had idiopathic dilated cardiomyop-
athy, and six men had left ventricular hypertrophy associated with significant aortic stenosis. Lumen
diameter was measured at 96 points in 32 defined coronary segments or major branches. Measurements
were scaled to the catheter, corrected for imaging distortion, and had a mean repeat measurement error
of 0.12 mm. When sex, anatomic dominance, and branch length were accounted for, normal lumen
diameter at each of the standard anatomic points could usually be specified with a population variance
of ±O.6 mm or less (SD) and coefficient of variation of <0.25 (SD/mean). For example, the left main artery
measured 4.5±0.5 mm, the proximal left anterior descending coronary artery (LAD) 3.7+0.4 mm, and the
distal LAD 1.9+0.4 mm. For the LAD, lumen diameter was not affected by anatomic dominance (right
versus left), but for the right coronary artery, proximal diameter varied between 3.9±0.6 and 2.8±0.5 mm
Downloaded from http://ahajournals.org by on September 16, 2019

(p<O.Ol) and for the left circumflex, between 3.4±0.5 and 4.2±0.6 mm (p<0.01). Women had smaller
epicardial arterial diameter than men (-9%o; p<0.001), even after normalization for body surface area
(p<O.Ol). Branch artery caliber was unaffected by the anatomic dominance but increased with branch
length, expressed as a fraction of the origin-to-apex distance (p<0.001). Lumen diameter was not affected
by age or by vessel tortuosity but was significantly increased among men with left ventricular hypertrophy
(+17%; p<O.001) or dilated cardiomyopathy (+12%; p<O.OOl).
Conclusions. This is a reference normal data set against which to compare lumen dimensions in various
pathological states. It should be of particular value in the investigation of diffuse atherosclerotic disease.
(Circulation 1992;86:232-246)
KEY WoRDs * coronary anatomy * vessels, coronary, size * computer graphics * arteriography,
quantitative

In humans, certain pathological and physiological
processes influence coronary artery lumen caliber.
Diffuse arteriosclerotic intimal thickening is said
to reduce lumen diameter at pathological examina-
tion,1-3 although a compensatory enlargement occurs.4
Reduced flow demand contracts vessel size by an endo-
thelium-dependent process.5 Aging,6'7 ectasia,8-'0 Ka-
wasaki syndrome," increased myocardial flow demands,
From the Cardiovascular Research and Training Center, Car-
diology Division, Department of Medicine, University of Wash-
ington, Seattle.
Supported in part by US Public Health Service grants HL-
13517, HL-19451, HL-18805, and HL-30086, in part by an Estab-
lished Investigator Award (B.G.B.) from the American Heart
Association, and in part by a grant from the John L. Locke Jr.
Charitable Trust.
Address for correspondence: B. Greg Brown, MD, PhD, Cardi-
ology Division, RG-22, University of Washington School of Med-
icine, Seattle, WA 98195.
Received December 4, 1990; revision accepted April 8, 1992.
and high-flow coronary fistulae12 are processes that may
increase lumen caliber. These processes may occur in
the absence of focal narrowing and/or luminal irregu-
larity, so we cannot be certain that normal-appearing
arterial segments are truly normal. This presents a
problem for the traditional radiographic estimate of
arterial disease severity, the "percent stenosis," which is
based on the ratio of a focal minimum to a nearby
"normal" diameter. Percent stenosis is, experimentally,
a uniquely effective correlate of the physiological impact
of focal coronary narrowing,13-'6 but unfortunately, to
the extent that we cannot accurately assess normal
lumen diameter in humans, the clinical utility of this
estimate is diminished.2
One solution to this problem is to find methods
whereby we may predict normal lumen diameter at a
given point in the coronary anatomy and to use this
diameter as a normal reference for the percent stenosis.
At present, data on the "true" normal diameter of
human coronary arteries are not available in a system-
 Dodge et al Normal Lumen Diameter 233

atic format. In this report, we extend the early measure-

ments of MacAlpin et al'7 to demonstrate that, when RCA DOMINANT
anatomic variations are accounted for, it is usually
possible to specify normal coronary segment diameter
in men and women to within +25% (coefficient of
variation). We also show that left ventricular (LV)
hypertrophy (LVH) and, to a lesser extent, dilated
cardiomyopathy (DCM) result in enlargement of nor-
mal vessel caliber.
A)
Methods A;
Patients
Eighty-three patients were selected for the absence of
arterial disease from among 9,160 consecutive clinical
catheterization studies. Blood pressure was routinely
obtained during the catheterization admission in all
cases, and a fasting lipid profile (total cholesterol [TC],
low density lipoprotein [LDL], high density lipoprotein
[HDL], and apolipoprotein B [apo B]) was obtained in
the great majority. Patients with anemia or diabetes
mellitus, those who had received nitroglycerin in the
catheterization laboratory, and those who had taken
long-acting nitrates on the day of catheterization were
excluded. Films were very carefully inspected by an
See p 331
experienced arteriographer; those with any luminal
irregularity even faintly suggestive of an atherosclerotic
plaque were excluded, leaving only films with smooth, FIGURE 1. Diagram of coronary artery segment and branch
regular arterial borders. In 70 normal patients (10 nomenclature. Top panel: Right coronary artery (RCA)-
women), heart size on radiography, global ejection dominant distribution; bottom panel: balanced and left
Downloaded from http://ahajournals.org by on September 16, 2019

fraction, and end-diastolic LV chamber volume were
within normal limits (20.52; 50-120 ml/m2). Five pa-
tients had mild mitral insufficiency resulting from pro-
lapse without evidence of a dilated ventricle. None of
these normal patients had structural cardiac abnormal-
ities. Indications for catheterization were the evaluation
of chest pain in 53 patients, history of ventricular
arrhythmia (premature contraction, tachycardia, fibril-
lation) in eight, Wolff-Parkinson-White disease in four,
abnormal exercise tolerance test in three, and being
potential participants in a drug trial in two. Six other
patients had concentric LVH caused by aortic stenosis,
with or without aortic insufficiency, with aortic valve
area of <1.2 cm2 and/or aortic valve gradient >60
mm Hg, evidence of LVH on ECG, and/or wall thick-
ening > 13 mm on echocardiogram. Another seven
patients had DCM with global ejection fraction <0.25
(six patients)- and increased end-diastolic LV chamber
volume (six patients; .160 ml/m2).
Arteriography
Coronary arteries were cannulated by the Judkins
technique with catheters of known dimension (5F to
8F). Selective coronary injections of Renografin 76 were
filmed in standard projections with a General Electric
Fluoricon 300 radiographic system (Milwaukee, Wis.)
with a resolution of 2.4 line pairs per millimeter.
Anatomic Representation
The anatomic segments shown in Figure 1 were
located in each coronary arteriogram. Although coro-
nary anatomy tends to be somewhat more variable than
Figure 1 suggests, it was possible in every case to specify
coronary artery (LCA)-dominant distribution. Not shown is
the "small-RCA" distribution, in which the RCA branches
into the posterior descending branch (RD) and the inferior
wall branch (RI), and the posterior wall branch (CP) arises
from the left circumflex artery. Artery segment abbreviations
are fully detailed in Table 5.
segments that corresponded approximately to these
standard ones. In most cases, these segments were easy
to locate, and in no case was a major branch ignored in
this process. We have described this coronary artery
segment and branch nomenclature18 and have applied it
in this report with minor alterations including a "small"
right coronary artery (RCA) anatomic distribution and
anterior (OA) and posterior (OP) branches of the
obtuse marginal (OM) branch. A complete description
of coronary artery segment and branch nomenclature is
included in the "Appendix."
Measurements: Data Entry
The coronary angiographic films were projected with
magnification to 4.4 times life-size. Frames with clear
images were selected. All segments and branches of the
coronary artery tree were identified, and the anatomy
was reduced to a set of up to 96 defined subsegments in
the following manner. With the exception of Al, A2,
and A3, each segment shown in Figure 1 was trisected,
and the midpoint of each third of each segment was
identified as shown in Figure 2. For each artery, we
digitized the two-point diameter of the catheter into a
Digital Equipment Corporation VAX-11/750 (May-
nard, Mass.). We then digitized the lumen diameters of
 234 Circulation Vol 86, No 1 July 1992
Y-AxIS (CEPHALAD)
X-AXIS
(LATERAL)
FIGURE 2. Diagram showing example sites selected for
diameter measurement. Each artery segment was trisected
along its length, and the midpoint of each resulting subseg-
ment was found. Along the segments of the major arteries, the
lumen borders of the proximal midpoints, middle midpoints,
and distal midpoints (0) were digitized, whereas in arteries
Downloaded from http://ahajournals.org by on September 16, 2019
branching from the main arteries, the lumen diameters at the
point oforigin (c), proximal midpoints, and middle midpoints
(.) were digitized, as illustrated here for the middle left
anterior descending coronary artery (L2), the first septal (Si),
and the second diagonal (D2). r is the radial distance from the
coronary ostium to the artery segment. 0 is the azimuth angle
between the projection of r onto the transverse plane (X-Z
plane) and the anterior-posterior (AP) axis. (D is the angle
between r and the transverse plane (X-Z plane). '8
the midpoint of each subsegment in the left main (LM),
left anterior descending (LAD) (L1, L2, L3, IA), left
circumflex (LCx) (Ci, C2, C3, C4), and right coronary
arteries (RCA) (R1, R2, R3, R4) that was present. For
the three septal (S1, S2, S3), diagonal (D1, D2, D3),
marginal (M1, M2, M3), and inferior wall branches
(from RD, RI, RP, CP, CI, CD) and for the median
ramus, we digitized the lumen diameter at the branch
origin and the proximal and middle midpoints. For an
OM, the length of artery between its origin and its
bifurcation was divided into three equal parts, and the
lumen diameters were measured at the midpoint of each
of these subsegments. Diameters of the two longest
branches continuing beyond the bifurcation, OA and
OP, were measured as described above for the
branches.
Terminal branches (diagonal, marginal, septal, etc.)
were classified by size into one of five groups: long,
medium, short, absent, or unseen. This branch size
rating refers to the vessel's length of distribution, not its
width per se; where these vessels branched, their longest
extension was used. This classification was made by
FIGURE 3. Diagram showing example of vessel length clas-
sification. Branches were classified by size into one of five
groups: short, medium, long, absent, or unseen. The classifi-
cation was performed from views roughlyperpendicular to the
long axis of the heart by comparing the visible length of the
vessel with the distance between its origin and the apex of the
heart as illustrated here for the first marginal branch (M])
arising from the left circumflex artery. If the visible path of the
vessel was less than 25% of this distance, then the vessel was
neglected ("absent"). When the visible path of the vessel was
between 25% and 50%,o or 50% and 75%, or .75% of this
distance, then it was classified as short, medium, or long,
respectively. Vessels that could not be clearly visualized were
classified as unseen. This left coronary artery model was
generated by an updated computer program for anatomic
display.'8
comparing the visible length of the branch to the
distance between its origin and the apex of the heart in
a view perpendicular to the long axis of the heart, as
illustrated in Figure 3. If the visible length of the branch
was less than 25% of this distance, then it was neglected
("absent"). When the visible length of the vessel was
between 25% and 50%, or 50% and 75%, or >75% of
this distance, then it was classified as short, medium, or
long, respectively. Branches that could not be clearly
visualized were classified as unseen. Occasionally, a
vessel branched extensively and thus supplied a larger
myocardial mass than would otherwise be predicted by
its length; it was accordingly given a longer classifica-
tion. The three largest septal branches were classified; a
septal was long if it had at least three branches and
extended >65% of the distance from its origin to the
base of the septum, medium if it had two branches and
extended approximately 50% of this distance, and oth-
erwise was short. If a single branch of the LCx was
much larger than its neighbors, it was considered to be
an OM, although a first marginal (M1) or third marginal
(M3) was also permitted if present. OM, OA, and OP
length were classified according to the same reference
length: the OM origin-to-apex distance. The posterior
descending artery, RD, and CD were required to reach
the apex to be long.
The L4 segment was categorized to investigate
whether the length of apical LAD wrap-around affects
LAD lumen diameter. L4 was short if the LAD barely
passed the apex, medium if the LAD rounded the apex
and made a small ramification on the inferior apical
region, and long if the L4 provided inferior septal
branches. When the LAD did not reach the cardiac
apex, L4 was absent.

FIGURE 4. Representative photographs of the left anterior descending artery with four different degrees of tortuosity:
nontortuous, slightly tortuous, moderately tortuous, and very tortuous (from left to right). This figure was present for reference
when tortuosity was classified.
The continuous spectrum of LAD tortuosity was
classified into four groups: nontortuous, slightly tortu-
ous, moderately tortuous, and very tortuous. Examples
of each degree of tortuosity for the LAD are displayed
in Figure 4. Figure 4 was present for reference when
tortuosity was graded.
Our primary goal was to provide a reference data set
Downloaded from http://ahajournals.org by on September 16, 2019
of normal men with right dominant anatomy; a second-
ary goal was to highlight differences between this group
and other groups with normal coronary arteries. We
sought to completely measure 20 normal RCA-domi-
nant men and 10 patients from each of the other six
groups (small-RCA-dominant men, balanced men,
LCA-dominant men, RCA-dominant women, men with
LVH, and men with DCM) and .5 measurements from
each branch-length group from among the normal men.
Though we drew from a pool of 9,160 patients, it was
difficult to meet these goals. The occurrence of men
with LVH with RCA-dominant anatomy and normal
coronary arteries was approximately 1 in 1,500. As the
study progressed, it became clear that the low preva-
lence of some rarely occurring branch arteries (e.g., long
CI) would result in counts too low for a reliable
population average. In 73 patients (50 normal men, 10
normal women, six men with LVH, and seven men with
DCM), all artery subsegments present were measured.
The counts for some branches were then augmented by
making measurements at selected artery sites in 10
additional normal male patients, accounting for some of
the apparent discrepancies in patient numbers in the
tables. For groups in which patients could be found
more frequently, data were collected only until the
target sample was obtained.
Measurements: Theoretical
To reduce an angiographic image to true scale, opti-
cal and uniform x-ray magnifications were compensated
directly by use of the catheter, which was of known
diameter, as a scale factor. Out-of-plane magnification
Dodge et al Normal Lumen Diameter 235
caused by divergence of the x-ray beam was also cor-
rected in the scaling process.'8'19 Pincushion distortion
is a small source of selective image magnification with
our modern equipment and small (5-6-in.) image fields
and was neglected in this analysis for all arterial sub-
segments. However, the diameter of the scaling cathe-
ter, usually found in the periphery of the image, was
adjusted for pincushion effects.'9
Validation of Vessel, Dominance, and
Tortuosity Classification Methods
We determined the repeatability of our classification
of anatomic dominance, tortuosity, branch artery iden-
tity, and length after an interval of more than 3 months.
All 73 completely measured arteriograms were ran-
domly ordered, and the anatomic dominance of each
was consecutively reclassified. Tortuosity was consecu-
tively reclassified in a random subset of 36 normal male
patients. Each branch artery in 15 consecutive, ran-
domly selected patients was reidentified with our no-
menclature system (Figure 1 and "Appendix"), and the
length of each was reclassified with the above origin-to-
apex system. These reclassifications were then com-
pared with the original set.
Validation of Measurement Methods
We determined the accuracy and variability of this
method by measuring known diameters from films of a
roughly heart-sized styrofoam block in multiple viewing
angles with the General Electric Fluoricon 300 radio-
graphic equipment. This block contained a household
pin, the head representing the location of a coronary
ostium, and two brass tubes positioned to represent the
catheter and a subsegment with 3.25-mm lumen diam-
eter. The spatial relations between these locations were
determined.'8 At the selected site, opposite sides of the
tube were digitized in a fashion identical to that de-
scribed for the arteries. The widths at these locations
 236 Circulation Vol 86, No 1 July 1992
TABLE 1. Demographic Summary
Coronary
Sex Disease group dominance n Age (years)
Men Normal Right 24
Small right 14
Balanced 11
Left 11
LVH Right 6 63±5
DCM Right 7 44±13
Women Normal Right 10
BSA, body surface area; LVH, left ventricular hypertrophy; DCM,
were computed and compared with the known
diameters.
The measurement variability was determined by re-
tracing and analyzing arterial subsegments in the coro-
nary arteriograms of six normal RCA-dominant men.
The first and second sets of measurements were then
compared.
Statistics
Population variation of arterial lumen diameter has
been expressed in terms of the SD and the coefficient of
variation (SD/mean) of a group of measurements at a
specified point for a given sex, anatomic variation, and
pathological state. Comparisons of mean lumen caliber
among those groups used pooled Student's t tests.
Comparisons of body surface area (BSA)-normalized
lumen area for 12 main artery segments from normal
men, women, and men with LVH or DCM were made
Downloaded from http://ahajournals.org by on September 16, 2019
by Sign test,20 as was a comparison between long,
medium-length, and short terminal branch arteries at 19
segments, each with three subsegments. The influence
of age on arterial lumen diameter was tested by
ANOVA. A difference was considered statistically sig-
nificant if the two-sided probability of the observed
result, under the null hypothesis, was p<0.05.
Results
Baseline Patient Characteristics
The 83 patients (Table 1) were of mean age 45+± 13
years and had normal blood pressure (systolic, 118+15
mm Hg; diastolic, 71+9 mm Hg) and normal lipids (TC,
187±36 mg/dl; LDL, 120±32 mg/dl; HDL, 44±12 mg/
dl; apo B, 101±25 mg/dl). Of 50 normal men in whom
all artery segments were measured, diastolic blood
pressure exceeded 85 mm Hg (maximum 95 mm Hg) in
six. Those with normal blood pressure had a mean
proximal subsegment (Llp, Rlp, and Clp) lumen diam-
eter of 3.56 mm, and those with 285 mm Hg measured
3.66 mm (p=NS). Among 50 normal men, the TC,
LDL/HDL ratio, and apo B were measured in 48 and
were normal in 46, 46, and 44, respectively (<90th
population percentile for age21.22). Abnormal lipid levels
had no effect on lumen diameter in this group with
normal arteries. For example, 44 patients with normal
apo B had a mean proximal subsegment lumen diameter
of 3.57 mm; four with an elevated level measured 3.52
mm (p=NS). We were able to obtain medication infor-
mation for 73 of these 83 patients. By the exclusion
criterion, none of these patients continued chronic
nitrates at the time of catheterization; for three who had
Left ventricular end
Left ventricular diastolic volume/BSA
ejection fraction (mi/m2)
41+12 0.65±0.07 (n= 15) 81+19 (n=15)
46±12 0.67±0.08 (n=12) 76+16 (n=12)
43±16 0.63±0.06 (n=9) 67±15 (n=9)
41±6 0.65±0.10 (n=8) 73+8 (n=8)
0.56±0.18 (n=4) 125+37 (n=4)
0.20±0.05 (n=6) 208+43 (n=6)
47±17 0.67±0.06 (n=8) 68+16 (n=8)
dilated cardiomyopathy.
been taking chronic nitrates, lumen diameters were not
different from patients not on nitrates. Similarly, lumen
caliber was unaffected in three patients receiving cal-
cium channel blockers; we know that these do not
acutely affect epicardial lumen caliber.23
Error Calculations on Brass Tube Phantom Model
A comparison between computed and actual mea-
surements of the brass tube showed a high degree of
accuracy. In five views, from 480 RAO to 480 LAO with
200 caudal to 300 cranial angulation, errors between
-0.13 and 0.10 mm were observed. The mean error
(+SD) was <0.03 mm (±0.12), and the mean absolute
error was 0.10 mm (+0.03), suggesting that the complex
scaling logic was correctly implemented for these
measurements.
Repeatability of Lumen Diameter Measurement
A high degree of repeatability was observed when
arteriograms were reanalyzed. In six angiograms, the
lumen diameters of a total of 213 subsegments were
remeasured, 132 by the original tracer and 81 by a
second tracer. For all vessels, the mean difference
between first and second measurements (error) was
0.12+0.34 mm, independent of vessel size over the
diameter range of 0.5-5.0 mm. The mean intraobserver
and interobserver errors were similar: 0.13±0.30 and
0.10+0.40 mm, respectively.
Repeatability in Classification of Dominance, Vessel
Identity, Branch Length, and Tortuosity
Comparisons between the original and second classi-
fication sets reveal a high degree of precision. Anatomic
dominance was exactly reidentified in 92% of 73 angio-
grams; 8% had one-step differences involving large-
RCA versus small-RCA dominance. Of the 161 branch
arteries present on first reading, 95% were correctly
reidentified. Among these, branch length (long, me-
dium, short) was correctly reclassified in 77%; one
septal branch had more than a one-step difference.
Tortuosity classification was exactly repeated in 83% of
36 angiograms; the remainder had one-step differences.
Diameter of the Main Arteries in
Normal Right-Dominant Men
The main coronary arteries are the conduit vessel
skeleton from which the branches ramify. They include
the LM, the interventricular course of the LAD, and the
atrioventricular groove courses of the RCA and LCx.
Each patient has 12 such segments; because of varying
anatomic dominance, there are 13 main artery segments
 Dodge et al Normal Lumen Diameter 237

TABLE 2. Diameter Measurements of the Main Coronary Arteries in Normal Men

RCA dominant Small RCA dominant Balanced LCA dominant
Location n Diameter (mm) n Diameter (mm) n Diameter (mm) n Diameter (mm)
Ri middle 20 3.9±0.6 9(a) 3.8±0.5 10 3.0±0.5t 10 2.8±0.5t
R3 middle 20 3.1±0.5 10 2.6±0.6* 10 2.0±0.6t 10 1.1±0.4t
LM middle 18(b) 4.5±0.5 10 4.6±0.7 9(b) 4.4±0.4 10 4.6±0.4
Li middle 20 3.6±0.5 10 3.8±0.4 10 3.6±0.4 10 3.7±0.2
L3 middle 20 1.7±0.5 10 1.9±0.5 10 1.8±0.4 10 2.0±0.3
Cl middle 20 3.4±0.5 10 3.5±0.8 10 3.4±0.5 10 4.2±0.6t
C3 middle 19(c) 1.6±0.6 10 2.2±0.8* 10 2.5±0.5t 10 3.2±0.5t
RCA, right coronary artery; LCA, left coronary artery; R, right coronary artery; LM, left main; L, left anterior
descending; C, circumflex (refer to Figure 1 and Table 5). Values are mean±SD.
Reported segment numbers vary because of: (a) catheter-induced Rl spasm; (b) absent LM; (c) absent C3.
*p<0.05, tp<0.01, small-RCA, balanced, or LCA-dominant groups compared with RCA-dominant group.
For a complete tabulation see Table 6.

in this description (a patient may have either R4 or C4).
As shown in Table 2 and Figure 5, lumen caliber at the
midpoint of each segment varied from 4.5 -mm diameter
in the LM artery to 1.6 mm in the terminal LCx (C3).
The coefficient of variation (SD/mean) of lumen diam-
eter at these selected points is <0.25 in 75% of locations
and <0.20 in 58%. (A complete tabulation of the 12
main artery segments in RCA-dominant men, each with
three subsegments, is presented in Table 6.)
Differences Among Main Arteries Caused by
Variations in Anatomic Distribution
As shown in Table 2, the diameters of the LM artery
and segments of the LAD were unaffected by anatomic
distribution. The LCx, however, was usually significantly
larger when measured in a left-dominant distribution
than in a right. Conversely, the RCA was significantly
larger when measured in a right-dominant than in a
left-dominant or a balanced distribution. The normal

Rlp a
a
Rlm a 10 1
Downloaded from http://ahajournals.org by on September 16, 2019

Rid m71arlwr~lm,m,Zgp~
R2p 1 --------

RCA m a
p 25
R3M

=4p
R4m 0 9
a a,
a
PAd X5XXXt 0 .1

1 0 0 A
LMp a
LM I-
Mm_
JLMM 02 ^4

Lbm 1 1
9 9 a
a! !
Llm
FIGuRE5. Bargraph showingmean lumen
Lld diameters of main artery segments in normal
12p %%%%%M% ' * right-dominant men. RCA, right coronary
artery; LM, left main artery; LAD, left
LAD anterior descending coronary artery; LCx,
L3m,p left circumflex artery. Artery segment abbre-
13d_ viations are presented in Table 5.
L4p'
L4m y^XX,,,
L4d_

Clmd
LCx C2m
C2d_
C3p EM-Ar-"a
1 9
C3m
C3d I **
a c a1
1
0
5
a
1,
a
1
5

0 11 2 3 4 5
Mean Lumen Diameter ± SD (mm)
 238 Circulation Vol 86, No 1 July 1992
TABLE 3. Diameter Measurements of Selected Coronary Artery Branches in Normal Men
Short
Location n Diameter (mm)
RD origin 12 1.5±0.3t
CD origin 2 1.6±0.2
RP origin 6 1.4±0.5
Middle 1.1±0.3
CPorigin 9 1.4±0.3*
Middle 1.1 ±0.2t
Dl origin 10 1.4±0.3t
Ml origin 12 1.5±0.2t
S1 origin 6 1.0±0.3t
RD, right descending artery; CD, circumflex descending artery; RP, right posterior artery; CP, circumflex posterior
artery; Dl, first diagonal; Ml, first marginal; S1, first septal (refer to Figure 1 and Table 5). Values are mean±SD.
*p<0.05, tp<0.0l, short and long vessel groups compared with medium length group.
For a complete tabulation see Table 7.
width of most arterial subsegments in men can be
specified to within ±0.6 mm (±SD) when anatomic
dominance is accounted for. (A complete tabulation of
all 13 main artery segments possible in normal men,
each with three subsegments, is presented in Table 6.)
For these normal arteries, a round cross section was
assumed,24'25 and the lumen cross-sectional area was
estimated to be (i7/4)(Diameter)2. The summed cross-
sectional area of Rlm, Llm, and Clm was calculated
and, per MacAlpin et al,17 was called the total coronary
area (TCA). In men with the large-RCA-dominant
distribution, the TCA was 32.1±7.3 mm2 with the RCA
Downloaded from http://ahajournals.org by on September 16, 2019
contributing 38% of the total area, the LCx 29%, and
the LAD 33%. In normal men with small-right-coro-
nary, balanced, and dominant-left-coronary distribu-
tions, the TCAs were 33.5 ±9.3, 26.8 +5.2, and 30.7 +5.5
mm2, respectively (p=NS).
Influence of Artery Length
Table 3 shows considerable variation in branch diam-
eter. For these branches of the three main arteries,
diameter was unaffected by the dominance of the ana-
tomic distribution. For example, there was no difference
in diameter of descending (RD, CD) or posterior (RP,
CP) branches arising from the RCA or the LCx to
supply the inferior wall. The principal determinant of
branch diameter was the extent of its epicardial distri-
bution, as characterized by its length relative to the
distance from its origin to the LV apex (long, medium,
short), as described above and in Figure 3. The diame-
ter of the principal diagonal, marginal, and inferior wall
branches averaged 2.4-2.6 mm at their origin if classi-
fied as long, 1.8-2.2 mm if of medium extent, and
1.4-1.6 mm if short. For a given branch and size
classification, diameters were relatively uniform; the
coefficient of population variation of the lumen diame-
ter of a given subsegment was <0.25 for 76% of the
measured points and <0.20 for 48%. The coefficient of
variation of lumen diameter decreased as vessel length
increased. The expected width of most normal terminal
branches can be specified to within ±0.5 mm (SD) when
grouped by vessel length. (A complete tabulation of the
19 possible branch segments, each with three subseg-
ments, in normal men is presented in Table 7. The
statistical comparison from the more complete Table 7
Medium Long
n Diameter (mm) n Diameter (mm)
17 2.2±0.3 11 2.4±0.3
8 2.1±0.3 1 3.2
13 1.8±0.4 1 2.6
1.4±0.2 2.3
17 2.0±0.6 7 2.6±0.5*
1.6±0.5 1.7±0.4
28 2.1±0.5 10 2.4±0.3*
10 2.1±0.5 6 2.6±0.4*
34 1.4±0.2 12 1.8±0.5t
shows that long branches were usually significantly
wider than medium-length branches [p<0.001], and
short vessels were usually significantly narrower than
medium length vessels [p<O.OOi].)
The influence of terminal LAD (L4) length upon
LAD lumen diameter was investigated. We found no
effect of L4 length (long, medium, or short) on LAD
lumen diameter when L4 was present. When L4 was
present, the mean lumen diameters at Llm, L2m, and
L3m were 3.7+0.4, 2.8±0.5, and 1.9±0.4 mm, compared
with 3.5±0.4 mm (p=NS), 2.3±0.6 mm (p<0.01), and
1.2±0.3 mm (p<O.01), respectively, when L4 was ab-
sent. LA was absent in 14% of the normal men studied.
Influence of Sex and Pathological State
Selected and representative lumen diameters are
given in Table 4 for the large-right-dominant distribu-
tion of women, normal men, and men with LVH caused
by aortic valve stenosis or with idiopathic DCM. In
nearly all comparisons, lumen caliber was smaller in
normal women than in normal men (-9±8% [SD];
p<O.OOi for group), although this difference at individ-
ual points was statistically significant for only the prox-
imal artery subsegments (R1, LM, Llp, Clp, and Clm;
each p<0.05). In men, those with LVH or with DCM
had wider vessels (+17±12% and +12±11%, respec-
tively; p<0.001 each group), but these differences for
individual subsegments were significant only in the RCA
(Ri through R4; each with p<0.05) with LVH and
portions of the RCA (Rlm through R3; each with
p<0.05) with DCM. (A complete listing of lumen diam-
eters of the main coronary arteries in women, normal
men, and men with LVH or DCM is found in Table 8.)
Figure 6 compares BSA-normalized segment lumen
areas in women and in men with LVH or DCM with
those in normal men. It contains coronary lumen area
estimates derived from 1,452 measurements of lumen
diameter. Mean segment areas (averaged from proxi-
mal, median, and distal measures) of each of the 12
main artery segments were calculated and divided by
BSA to normalize for differences in body size between
groups. Within a group, a 1.0 ratio occurs when the
BSA-normalized area at a specific segment is the same
as that found in normal men. Women tend to have
 *v
Dodge et al Normal Lumen Diameter 239

TABLE 4. Effect of Sex, Left Ventricular Hypertrophy, and Dilated Cardiomyopathy on Selected Coronary Artery
Diameter Measurements in Right Coronary Artery-Dominant Patients
Normal men Normal women DCM men LVH men
Diameter Diameter Diameter Diameter
Location n (mm) n (mm) n (mm) n (mm)
Rl middle 20 3.9±0.6 10 3.3±0.6* 7 4.5±0.5* 6 4.6±0.7*
(1.0±0.2) (0.8±0.2) (1.2±0.1) (1.2±0.2)
LM middle 18* 4.5±0.5 10 3.9±0.4t 7 4.8±0.3 6 4.9±0.4
(1.0±0.1) (0.9±0.1) (1.1±0.1) (1.1±0.1)
Li middle 20 3.6±0.5 10 3.2±0.5 7 3.8±0.5 6 3.9±0.5
(1.0±0.1) (0.9±0.1) (1.1±0.1) (1.1±0.1)
Cl middle 20 3.4±0.5 10 2.9±0.6* 7 3.3±0.7 6 3.6±0.6
(1.0±0.1) (0.9±0.2) (1.0±0.2) (1.1±0.2)
DCM, dilated cardiomyopathy; LVH, left ventricular hypertrophy; R, right coronary artery; LM, left main; L, left
anterior descending artery; C, circumflex artery (refer to Figure 1 and Table 5). For each patient group/artery location,
the mean±SD in millimeters is above, and the ratio (mean±SD) in normal men is below in parentheses.
tReported segment numbers vary because of absent LM.
*p<0.05, tp<0.01.
A complete tabulation may be found in Table 8.

15.3% smaller BSA-normalized segment area for main
artery branches (p<O.O1), and men with LVH or with
DCM have 36.8% (p<O.O1) or 30.5% (p<0.05) larger
segment area, respectively, than normal men.
Influence ofAge
Figure 7 shows that there is no age-related trend in
Downloaded from http://ahajournals.org by on September 16, 2019
constant (15.2±3.6 mm2/m2) for men of all ages. Nor-
malized cross-sectional areas of individual main arteries
in all normal men likewise showed no trend with aging.
Influence of Tortuosity
We found a positive correlation between age and
LAD tortuosity but none between LAD tortuosity and

TCA, as defined above, in normal men when normalized lumen diameter. The four tortuosity categories were
for BSA (p=NS). Thus, BSA-normalized TCA is a numbered in increasing order of tortuousness as illus-

Left Ventricular Dilated

2.01
Normal Females Hypertrophy Cardiomyopathy
_ ~ ~
~ ~ ~ ~ ~ ~ ~ --11-, . ...
&.I)

1.8
a 1.8
1.6 1.6

MEO
L
1.4 * l
1.4
a * a
.0i 1.2 a
m U 1.2
a o

1.0 1.0
a a
E .5 0.8' 0.8
Eo 4
0.6 0.6
0.4 0.4
0.2' p < 0.01 p < 0.01 p < 0.05 0.2
ll
0.0 . -3. --r- n
1.0 3.0 5.0 1.0 3.0 5.0 1.0 3.0 5.0

Lumen Diameter in Normal Men (mm)

FIGURE 6. Plots showing body surface area (BSA)-normalized lumen area at 12 main artery segment locations in women and
in men with left ventricular hypertrophy and dilated cardiomyopathy as a fraction of the area at that same location in normal men.
 240 Circulation Vol 86, No 1 July 1992
30
A tomic distribution but is strongly influenced by branch
m length. The coefficient of population variation is less
20- CA
E dominant normal men.
li.
V
U.---O ° -.
E artery disease have been reported infrequently.'726-28
04 10*
* *o A A
U, U the coronary artery tree using the techniques available
0.
0
10
.
1
20 30
5 9
40
5 5
s0
1
60
A~
FIGURE 7. Scatterplot showing body surface area (BSA)-
normalized total coronary area (TCA) in normal men versus
age. The best-fit line is y=15. 7-O.lAge. Thus, TCA in men
with apparently normal coronary arteries is virtually constant
at 15.2+3.6 mm21m2. *, Large-right dominant (n=20); A,
small-right dominant (n=9); *, balanced (n=10); o, left
dominant (n=10).
trated in Figure 4 (nontortuous, 1; slightly tortuous, 2;
moderately tortuous, 3; and severely tortuous, 4). In
patients younger than 38 years (n=20), those aged
Downloaded from http://ahajournals.org by on September 16, 2019
38-47 years (n=19), and those older than 47 years
(n=21), the average tortuosity (±SD) was 2.3+0.8,
3.0+0.8, and 3.2±0.9, respectively (p<0.01 for the
comparison of the younger group with the middle and
with the older group, p=NS for comparison of the
middle and older groups). L3m is in the region of the
LAD that would appear most likely to be affected by
tortuosity when L4 is present. In such patients, when
the LAD was nontortuous or slightly tortuous (n=16),
moderately tortuous (n = 15), and severely tortuous
(n=9), average L3m lumen diameters relative to means
for the entire group were +4%, -3%, and -2%
different, respectively (p=NS).
Computer-Generated Model of Coronary Arteries
The lumen diameters of normal men were integrated
into a computer program, previously described,'8 to
generate a two-dimensional projection of the coronary
artery tree with expected normal lumen diameters in
any of the four anatomic dominances from any desired
viewing angle. This program generated the LCA tree
seen in Figure 3.
Discussion
The lumen diameters at 96 specified points in the
coronary anatomy have been measured in a group of 83
patients who have no evidence of atherosclerotic dis-
ease. The measurement method used has a mean error
of <0.03 mm for a measurement of 3.25 mm (<4%) in
phantom models and a mean error of 0.12 mm in
repeated coronary arteriogram analysis. We found that
the lumen diameter of most arterial subsegments can be
specified to within ±0.6 mm (SD) when the sex, ana-
tomic variation, branch length, and certain determi-
nants of myocardial mass are accounted for. Branch
artery caliber is unaffected by dominance of the ana-
than 0.25 in 75% of artery subsegments in large-RCA-
Arteriographic measurements of coronary artery lu-
men diameter during life in patients without coronary
MacAlpin et al'7 made measurements at 12 locations in
in the early 1970s. Despite method differences, our
4,290 measurements of lumen diameters agree in most
cases with those obtained by others arteriographi-
cally'7,26 and at postmortem examination.2728 For exam-
70 ple, we report LM at 4.5±0.5 mm versus 4.3+0.6 mm17
and 4.0 mm.27 At comparable proximal and middle LAD
artery subsegments, MacAlpin17 reported 3.5 +0.5 and
2.0±0.4 mm versus our finding of 3.6±0.5 mm at the
mid-Li and 2.3±0.4 mm at the distal L2. A mid-Li
measure of 3.3±0.5 mm has been reported elsewhere.26
RD has been reported as 2.11 mm27; pooling vessels of
different lengths, we found 2.0±0.5 mm. Previous re-
ports of normal TCA have been 25.7±4.8 mm2 (Refer-
ence 17) and 33.3±10.4 mm2 (Reference 28); we report
a range of 26.8±5.2 to 33.5±9.3 mm2 (mean, 31.0+7.2
mm2).
It has been suggested that both lumen diameter and
vessel length are necessary for a quantitative descrip-
tion of the coronary vascular network.29 To categorize
length, we compared the visible length of a branch
artery with the distance between its origin and the apex
of the heart (Figure 3). With this method, it is possible
to classify branch arteries of different absolute lengths
into groups that have significantly different lumen di-
ameters with reasonably small coefficients of population
variation. In the LAD, lumen diameter was significantly
affected by length only when the terminal LAD failed to
reach the cardiac apex (approximately 15% of
occurrences).
TCA is known to correlate directly with myocardial
mass in both normal and hypertrophied hearts.28 Lumen
cross-sectional area in an individual coronary artery
after maximum dilation is linearly related to the myo-
cardial volume perfused in dogs30 and nearly linearly
related to the myocardial weight perfused in normal
human hearts.31 Thus, we chose to compare lumen
cross-sectional area and not diameter between RCA-
dominant normal men, women, and men with LVH or
DCM, with normalization using BSA to compensate for
differences between groups in body size.
In contrast to MacAlpin's'7 statement that differences
in lumen caliber between men and women could be
acceptably resolved when TCA was normalized to BSA,
we found this not to be the case. In our set of 1,011
lumen measurements, women tended to have narrower
coronary arteries than men, a difference that persisted
after adjustment for body size. BSA-normalized coro-
nary area in women is approximately 90% that in men,
a finding in agreement with the observation that BSA-
normalized LV mass in women is approximately 90% of
that in men.32 The cause of this association with sex is
unknown. This difference in vessel size may contribute
to the greater operative mortality in women after coro-

nary artery bypass graft surgery33-35; however, our
measurements at two of the five most commonly grafted
sites (L2 and R3) were not significantly different be-
tween men and women. Too few of the remaining three
common graft sites (M1, Dl, and RD) were measured in
women to allow comparisons. Other explanations for
this operative mortality difference have been offered.36
With 441 lumen measurements in the main arteries of
male patients with LVH or DCM, our results compare
favorably with measurements made by others during
life7"26 and postmortem.28'37 Resting blood flow per unit
myocardium is normal in patients with severe aortic
valve disease38 and decreased in those with DCM,39
whereas myocardial mass is greatly increased in
both.40,4l The total blood flow is thus increased in
patients with aortic stenosis4243 and may be increased in
patients with DCM.39 We found coronary area to be
significantly increased in hypertrophied and dilated
hearts. As total flow equals cross-sectional area multi-
plied by velocity, this is probably an adaptive phenom-
enon, with coronary size increasing to satisfy myocardial
demand for an augmented blood supply37 while preserv-
ing velocity by means of partially characterized autoreg-
ulatory pathways.5"12
Arteries are said to dilate with age, as they diffusely
narrow with progressive intimal disease.46,7 Depending
on the predominant effect, either greater or lesser
lumen diameter would result. During aging, the histo-
logical increase in diameter of the internal elastica
offsets increased intimal thickness, possibly as an adap-
tation to the advance of intimal thickening with age
rather than a direct effect of aging on the arterial wall.4
Downloaded from http://ahajournals.org by on September 16, 2019
We found no evidence that lumen caliber is influenced
by age in these adults without angiographically apparent
atherosclerosis, a finding in agreement with that of
others.44
The population variation in tortuosity of the coronary
artery tree, increasing with age, has long been noted.45
As did others,44 we found a significant positive relation
between age and tortuosity but no significant relation
between tortuosity and lumen diameter. Tortuosity may
be a normal aging phenomenon and not secondary to a
disease process.
The percent diameter stenosis is the most commonly
used arteriographic index of coronary disease severity,
but objections have been raised to its use in patients
with atherosclerosis.1-3 Chief among these is that nar-
rowed segments are frequently compared with adjacent
less narrowed but diffusely diseased segments, and thus
the actual severity of a lesion may be underestimated.
As a result, in humans, measures of percent stenosis do
not correlate with the physiological significance of cor-
onary atherosclerotic lesions ranging from 20% to
60%.3
The physiological impact of a stenosis is related to the
blood pressure loss (AP) that occurs as blood flows
through it. This is reasonably well approximated by the
dynamic loss, AP=3.8 (Q/A )2, where Q is blood flow
(cm3/sec), A, is minimum lumen cross-sectional area
(mm2) in the stenosis, and AP is in millimeters of
mercury.23 Because the arterial flow volume differs
greatly at different points in the coronary anatomy, no
single value of stenosis area, A,, can be used to estimate
a lesion's severity. However, the ratio (normal flow/
normal area) (QN/AN) is experimentally a fairly constant
Dodge et al Normal Lumen Diameter 241
velocity in arteries of various sizes'6 and can be esti-
mated at QN/AN= 11.1 cm/sec for a proximal LAD of
area 10.2 mm2 (3.6 mm diameter; see Table 2) supplying
its basal myocardial needs of 68 ml/min.46 As normal
total flow is maintained in the face of substantial
narrowing, the physiological impact of the stenosis can
be approximated as AP=0.05 (AN/A,)2 using the two
expressions above. Thus, hypothetically, a better arte-
riographic index of the hemodynamic impact of a ste-
notic coronary artery lesion at any site in the coronary
anatomy might result from the above comparison of the
measured minimum area with its expected normal area
at that site. The value of this approach remains to be
tested. This report represents an important first step
toward investigating this hypothesis by determining
whether and with what degree of confidence it is
possible to specify the expected lumen caliber of a given
normal coronary artery segment.
These normal widths have been included in an edu-
cational Macintosh computer program'8 (see Figure 3)
permitting a two-dimensional projection of the three-
dimensional normal coronary artery tree from any
user-specified viewing angle. This may be of benefit to
first-year cardiology fellows; medicine, surgery, and
radiology residents; medical students; and others trying
to develop an appreciation for the normal coronary
anatomy.
Certain limitations of our study may have increased
measurement variance. These include infrequent in-
stances of retrospectively unavailable information on
LV size or wall thickness, concomitant drug use (10 of
83 patients), and lack of direct measures of scaling
catheter dimensions, which may introduce 1-2% addi-
tional variance.47 Automated quantitative angiographic
measurements of nonnylon catheters may systematically
underestimate true size (by micrometer) by 3.5%47;
however, this does not affect dimensional accuracy with
the approach we use.19 Finally, a relatively small num-
ber of normal women and men with DCM or LVH were
studied.
With this article, an anatomic nomenclature scheme
has now been fully defined. It is concise, easy to
remember because of the correspondence of the loca-
tion code to the widely accepted coronary arterial
nomenclature, and repeatable. We have shown in nor-
mal men that the combination of subsegment location,
anatomic distribution pattern (RCA dominant, small-
RCA dominant, balanced, and LCA dominant), and the
branch artery length categories (long, medium, short,
and absent) permits an estimate of normal lumen
diameter with a relatively small population variance. We
believe that this anatomic scheme18 corresponds to the
clinically relevant coronary anatomy and should be
considered for use as a standardized description of
these vessels in future investigations.
Appendix
Quantitative Description of the Human Coronary
Anatomy: Nomenclature and Measurements of
Subsegment Spatial Location and
Normal Lumen Diameter
In a coronary arteriogram, the anatomic segments shown in
Figure 1 can be located (not all segments are labeled).
Although coronary anatomy tends to be somewhat more
variable than Figure 1 suggests, it is usually easy to specify
 242 Circulation Vol 86, No 1 July 1992

TABLE 5. Abbreviations of Coronary Artery Nomenclature

Abbreviation Name Description
Main arteries
LAD Left anterior descending
LCA Left coronary artery
LCx Left circumflex
LM Left main
RCA Right coronary artery
Main artery segments
Cl LCx first segment LCx from its origin at the LM to Mi (or OM, if Ml is absent)
C2 LCx second segment LCx from Ml to M2 (or OM) (not present if Ml is absent)
C3 LCx third segment LCx from M2 (or OM) to CP (if present, otherwise to end of LCx)
C4 LCx fourth segment LCx from CP along atrioventricular groove to end of LCx (absent in RCA
and small-RCA-dominant distributions)
Ll LAD first segment LAD from its origin at the LM to Si
L2 LAD second segment LAD from Si to S3
L3 LAD third segment LAD from S3 to the cardiac apex
L4 LAD fourth segment LAD from the cardiac apex to its terminal point on the inferior wall
LM Left main LCA from ostium to bifurcation of LCA into LAD and LCx
RI RCA first segment RCA from its origin to Al
R2 RCA second segment RCA from Al to A3
R3 RCA third segment RCA from A3 to RD (if present, otherwise to end of RCA)
R4 RCA fourth segment RCA from the RD along atrioventricular groove to end of RCA (absent in
balanced and LCA-dominant distributions)
Branch artery segments
A1-A3 Acute marginals Three largest branches arising from the RCA to supply the right
ventricular wall, numbered from most proximal to most distal
CD Posterior descending Distal most inferior wall branch arises from C4, present only in
left-dominant anatomy
CI Inferior Inferior wall branch arises from C4, present only in balanced and
left-dominant anatomy
CP Posterior Proximal most inferior wall branch arises from junction of C3 and C4,
present in small-right, balanced, and left-dominant anatomy
DI-D3 Diagonals Three largest branches arising from the LAD to supply the left ventricular
anterior free wall, numbered from most proximal to most distal
M1-M3 Marginals Three largest branches arising from the LCx to supply the left ventricular
Downloaded from http://ahajournals.org by on September 16, 2019

lateral free wall, numbered from most proximal to most distal

MR Median ramus An anatomic variant arising at a trifurcation of the LM
OA Anterior branch OM Anterior distal branch of OM
OM Obtuse marginal Anatomic variant present when one branch artery off the LCx is much
larger than its neighbors supplying the left ventricular lateral free wall
OP Posterior branch OM Posterior distal branch of OM
RD Posterior descending Proximal most inferior wall branch arises from junction of R3 and R4,
present in right, small-right, and balanced-dominant anatomy
RI Inferior Inferior wall branch arises from R4, present only in right and
small-right-dominant anatomy
RP Posterior Distal most inferior wall branch arises from R4, present only in
right-dominant anatomy
S1-S3 Septals Three largest branches arising from the LAD to supply the interventricular
septal wall, numbered from most proximal to most distal
Suffixes (subsegments)
d Distal Midpoint of the distal third of an artery segment
m Middle Midpoint of the middle third of an artery segment
0 Origin Origin of an artery segment (proximal border)
p Proximal Midpoint of the proximal third of an artery segment

segments that correspond approximately to these standard
ones as described in Table 5. The LM artery has a single
segment. The LAD is separated into four segments, as illus-
trated in Figure 1, defined by 1) its origin from the LM, 2) the
first septal perforator (S1), 3) the third septal perforator (S3,
not labeled in Figure 1), 4) the cardiac apex, and 5) its terminal
point on the inferior wall. S3 commonly arises near the bend of
the LAD in RAO views and often is near the origin of the
second diagonal. These landmarks may serve as alternative
markers. If none of these are identified, the L2-L3 transition
is defined as halfway from Si to the cardiac apex. Branches
from the main arteries are numbered from proximal (1) to
distal (3). The possible branches of the LAD included the
three largest septal branches (S1-S3) and the three largest
diagonal branches (D1-D3). A median ramus branch is an
anatomic variant arising at a trifurcation of the LM and is
called a branch. Because of the variable nature of the branches
that originate at or very near the bifurcation of the LM, the
angiographer must decide whether a given case is best classi-
fied as a first diagonal (D1), first marginal (M1), or if inter-
mediate between the two, as a median ramus.
The LCx, with a right- and small-right-dominant anatomy,
is divided into three parts (Cl, C2, and C3) by the first and
second marginal branches (Ml and M2). In this case, the
atrioventricular groove continuation of the LCx (C3) is usually
small. In the case of small-right-dominant anatomy, the
posterior wall branch (CP) originates at the distal end of C3.
With a balanced anatomic distribution, the LCx has a fourth
segment (C4) distal to CP and gives rise to an inferior wall
branch (CI). With a left dominant anatomy, the LCx also gives
 Dodge et al Normal Lumen Diameter 243

TABLE 6. Lumen Diameter and Spatial Location of 13 Main Artery Segments, Each With Three Subsegments, in Normal Men
Lumen diameter
RCA dominant Small RCA dominant Balanced dominant LCA dominant Spatial location'8
Subsegment Diameter (mm) n Diameter (mm) n Diameter (mm) n Diameter (mm) n r 6 4' n
Left main
LMp 4.5±0.6 18(a) 4.6+0.5 10 4.3±0.4 9(a) 4.6+0.4 10 0.3 830 70 19
m 4.5±0.5 18(a) 4.6±0.5 10 4.4±0.4 9(a) 4.6±0.4 10 0.7 800 40 19
d 4.5±0.4 18(a) 4.6+0.5 10 4.4±0.3 9(a) 4.5 ±0.4 10 1.1 790 00 19
Left anterior
descending
Llp 3.7±0.5 20 3.9+0.4 10 3.8±0.2 10 3.8+0.3 10 1.6 700 -30 19
m 3.6±0.5 20 3.8±0.4 10 3.6±0.4 10 3.7+0.2 10 2.3 630 -50 19
d 3.5±0.6 20 3.8±0.5 10 3.4±0.5 10 3.5±0.3 10 3.0 590 -80 19
L2p 2.9±0.5 20 3.5±0.6t 10 2.8±0.7 10 3.1±0.4 10 4.0 510 -90 19
m 2.5±0.5 20 2.9±0.7 10 2.5±0.4 10 2.8+0.4 10 5.3 44° -120 19
d 2.3±0.4 20 2.5±0.6 10 2.3±0.4 10 2.6±0.4* 10 6.5 390 -170 19
L3p 2.0±0.5 20 2.1±0.6 10 2.0±0.4 10 2.4±0.4* 10 8.0 330 -230 19
m 1.7±0.5 20 1.9±0.5 10 1.8±0.4 10 2.0+0.3 10 10.1 300 -320 19
d 1.4±0.5 20 1.7±0.4 10 1.6±0.3 10 1.9±0.5t 10 11.8 310 -400 19
L4p 1.4±0.5 15(b) 1.3±0.3 8(c) 1.3±0.4 8(d) 1.7±0.4 7(e) 12.2 320 -460 16
m 1.1±0.4 15(b) 1.1±0.4 8(c) 1.2±0.4 8(d) 1.5±0.2* 7(e) 11.8 330 -490 16
d 0.9±0.3 15(b) 1.0±0.3 8(c) 0.9±0.3 8(d) 1.1+0.2 7(e) 11.4 340 -520 16
Left circumflex
Clp 3.4±0.5 20 3.6±0.7 10 3.6±0.6 10 4.3±0.6t 10 1.4 860 -12° 19
m 3.4±0.5 20 3.5±0.8 10 3.4±0.5 10 4.2+0.6t 10 1.8 940 -240 19
d 3.3±0.5 20 3.5±0.8 10 3.5±0.5 10 4.1±0.6t 10 2.2 1020 -310 19
C2p 2.8±0.5 11(f) 3.0±0.8 9(g) 3.1±0.5 7(g) 3.4±0.6* 7(g) 2.6 1070 -330 18
m 2.8±0.6 11(f) 3.0±0.8 9(g) 3.2±0.4 7(g) 3.4+0.6* 7(g) 3.1 1160 -370 18
d 2.7±0.5 11(f) 2.9±0.8 9(g) 3.1±0.6 7(g) 3.3+0.7 7(g) 3.6 1200 -410 18
C3p 1.7±0.6 19(h) 2.2±0.8* 10 2.6±0.6t 10 3.3+0.6t 10 4.1 1260 -450 17
m 1.6±0.6 19(h) 2.2±0.8* 10 2.5±0.5t 10 3.2+0.5t 10 4.7 1350 -510 17
d 1.3±0.5 19(h) 2.1±0.8t 10 2.3±0.5t 10 2.9+0.5t 10 5.4 1410 -560 17
C4p 0 0 1.7±0.3 9 (i) 2.4+0.5 10 5.7 1560 -67° 6
m 0 0 1.5±0.4 9 (i) 2.2+0.5 10 5.9 1660 -710 6
Downloaded from http://ahajournals.org by on September 16, 2019

d 0 0 1.5±0.4 9 (i) 2.1±0.5 10 6.2 1180 -730 6

Right coronary
Rlp 4.0±0.6 20 3.9±0.5 9(j) 3.1±0.4t 10 2.8+0.5t 10 0.5 -710 -220 24
m 3.9±0.6 20 3.8±0.5 9(j) 3.0±0.5t 10 2.8+0.St 10 1.7 -680 -24° 24
d 3.8±0.5
3.5±0.6
20 3.7±0.4 9(j) 3.0±0.5t 10 2.7+0.St 10 2.6 -66` -300 24
3.4 -690 -390 24
R2p 20 3.3±0.4 10 2.7±0.4t 10 2.3+0.3t 10
m 3.4±0.5 20 3.2±0.5 10 2.7±0.4t 10 2.2+0.3t 10 4.6 -720 -480 24
d 3.2±0.5 20 2.8±0.6 10 2.5±0.3t 10 1.5+0.4t 10 5.6 -770 -550 24
R3p 3.2±0.6 20 2.6±0.5* 10 2.1±0.5t 10 1.2+0.3t 10 6.2 -840 -620 24
m 3.1±0.5 20 2.6±0.6* 10 2.0±0.6t 10 1.1+0.4t 10 6.4 -910 -690 24
d 3.1±0.6 20 2.6±0.6 10 2.0±0.6t 10 0.8+0.3t 10 6.7 150 -730 24
R4p 2.4±0.5 20 2.0±0.6 10 0 0 6.6 1400 -710 24
m 2.2±0.5 20 1.6±0.3t 10 0 0 6.2 1810 -660 24
d 1.9±0.6 20 1.5±0.4* 10 0 0 6.2 1710 -570 24
RCA, right coronary artery; LCA, left coronary artery; r, radius; 6, azimuth angle between projection of r onto transverse plane and
anterior-posterior axis; '1, angle between r and transverse plane. Subsegment artery abbreviations may be found in Table 5. Values are
mean±SD.
Reported subsegment numbers vary because of (a) LM absent; (b) L4 absent; (c) L4 absent in one and unseen in one; (d) L4 unseen;
(e) L4 absent in one and unseen in two; (f) C2 absent in eight (when an OM was present) and inadequately visualized in one; (g) C2 absent;
(h) no atrioventricular groove continuation of LCx after OM; (i) C4 absent because of shared origin of CP and CI; (j) catheter-induced Rl
spasm.
*p<0.05, tp<0.01, small-RCA, balanced, or LCA-dominant groups compared with the RCA dominant group.
rise to the posterior descending artery (CD). A large obtuse posterior wall branch (RP). An inferior wall branch (RI) is
marginal branch (OM) with anterior (OA) and posterior (OP) also specified. With a small-right-dominant anatomy, the RD
branches may substitute for Ml and M2 if a single branch of and RI are present, and the posterior wall vessel arises from
the LCx is much larger than its neighbors, although an Ml or the LCx. In the case of a balanced anatomy, only the RD arises
M3 is permitted if they are also present. With an OM present from the RCA to supply the inferior septum of the LV. With
and Ml absent, C2 is absent by definition. A third marginal a left-dominant anatomy, the inferior wall vessels arise from
(M3) is usually small and is often absent. The possible the LCx, as described above. The possible terminal branches
branches of the LCx include M1-M3, OM, OA, OP, CP, CI, of the RCA include the three largest acute marginal branches
and CD. (A1-A3), which supply the wall of the right ventricle, in
The RCA, with a right-dominant anatomy, is divided into addition to the previously described RD, RI, and RP.
four parts between 1) its origin, 2) the first acute marginal Subsegment nomenclature is as follows. All artery segments
branch (A1), 3) the third acute marginal branch (A3), 4) the can be trisected to create three subsegments. The first segrnent
posterior descending branch (RD), and 5) the origin of the third is the proximal (p) subsegrnent. The second segment
 244 Circulation Vol 86, No 1 July 1992

TABLE 7. Lumen Diameter and Spatial Location of Branch Artery Subsegments by Length in Normal Men
Lumen diameter
Short Medium Long Spatial location18
Subsegment Diameter (mm) n Diameter (mm) n Diameter (mm) n r 0 ( n
Septals
Slo 1.0L0.3t 6 1.4±0.2 34 1.8±0.5t 12 3.3 570 _90 19
p 1.1+0.2* 6 1.3±0.2 34 1.8±0.4t 12 3.7 500 -170 19
m 0.9+0.2* 6 1.0±0.2 34 1.1±0.3 12 4.3 350 -31° 19
S2o 0.8±0.0* 7 1.1±0.3 36 1.9±0.3t 9 5.1 440 -12° 19
p 0.8±0.1* 7 1.1±0.3 36 1.7±0.2t 9 5.3 400 -18° 19
m 0.7±0.1 7 0.9±0.2 36 1.4±0.3t 9 5.8 310 -290 19
S3o 0.9±0.2* 12 1.1±0.3 36 1.4±0.2 2 6.9 370 -190 19
p 0.8±0.2t 12 1.0+0.2 36 1.3±0.1 2 7.0 340 -230 19
m 0.7±0.2* 12 0.9±0.2 36 1.2±0.0 2 7.3 300 -310 19
Diagonals
Dlo 1.4±0.3t 10 2.1±0.5 28 2.4±0.3* 10 3.2 610 -80 19
p 1.3±0.3t 10 1.9±0.4 28 2.3±0.2* 10 4.2 610 -13° 19
m 1.1±0.2t 10 1.5±0.3 28 1.8±0.2* 10 6.2 610 -230 19
D2o 1.2±0.3t 11 1.9±0.4 22 2.6±0.3t 11 4.9 470 -120 19
p 1.2±0.2t 11 1.7±0.4 22 2.4±0.4t 11 5.9 470 -160 19
m 1.0±0.2t 11 1.4±0.2 22 1.9±0.4t 11 7.7 490 -24` 19
D3o 1.1 ±0.2t 19 1.7±0.3 13 2.4 1 6.6 410 -180 18
p 1.0±0.2t 19 1.6±0.3 13 2.3 1 7.4 410 -210 18
m 0.9±0.2t 19 1.3±0.2 13 1.7 1 9.0 430 -260 18
Marginals
MRo 1.4±0.3t 8 1.8±0.2 6 2.7±0.3t 10 1.2 850 -100 7

p 1.3±0.3* 8 1.7±0.2 6 2.4±0.3t 10 2.7 870 -20` 6

m 1.0±0.2* 8 1.3±0.2 6 2.0±0.3t 10 5.6 800 340 6
OMp 0 3.3±0.2 3 3.3±0.5 20 2.8 1050 -300 5t
m
. . .

0 2.9±0.2 3 3.0±0.5 20 3.5 1070 -320 5

d . . .

0 3.1±0.5 3 2.7±0.5 20 4.1 1070 -360 5t

OAo 1.4±0.3t 6 2.2±0.3 11 2.5±0.6 5 4.4 1070 370 5
p 1.3±0.3t 6 2.0±0.3 11 2.4±0.5 5 5.3 1020 430 S
m 1.2±0.3t 6 1.7±0.3 11 2.0±0.6 5 7.1 840 490 5
1.6±0.6* 10 2.2±0.5 2.5±0.4
Downloaded from http://ahajournals.org by on September 16, 2019

OPo 9 3 4.4 1070 370 5

p 1.6±0.5* 10 2.1±0.5 9 2.3±0.4 3 4.9 1140 430
m 1.3±0.5* 10 1.9±0.5 9 1.9±0.1 3 6.3 1160 530 5
Mlo 1.5±0.2t 12 2.1±0.5 10 2.6±0.4* 6 2.2 1020 -31° 13

p 1.4±0.2t 12 2.0±0.5 10 2.5±0.2* 6 3.5 1010 -280 13

m 1.1±0.2t 12 1.5±0.4 10 2.0±0.1t 6 5.6 910 -38° 13
M2o 1.5±0.3* 7 2.0±0.4 10 2.5±0.6* 8 3.4 1180 -41` 13

p 1.4±0.3* 7 1.9±0.3 10 2.3+0.6 8 4.4 1180 -430 13

m 1.0±0.1* 7 1.4±0.3 10 1.9±0.4t 8 6.0 109° -510 13
M3o 1.2±0.3t 10 1.7±0.2 12 2.8±0.3t 4 4.7 1300 490 15
p 1.1±0.3t 10 1.7±0.3 12 2.6±0.2t 4 5.3 1290 -520 15
m 0.9±0.3t 10 1.4±0.3 12 2.3±0.3t 4 6.5 1180 -58° 15
Inferior wall arteries
CPo 1.4+0.3* 9 2.0±0.6 17 2.6±0.5* 7 6.0 1510 -640 7
p 1.2±0.2t 9 1.8±0.6 17 2.1±0.6 7 6.5 1510 -67` 7
m 1.1±0.2t 9 1.6±0.5 17 1.7±0.4 7 7.6 1450 -71` 7
CIo 1.1±0.3t 13 1.7±0.4 6 0 6.5 1170 -670 S
p 1.1±0.2t 13 1.7±0.5 6 0 7.0 1170 -690 5
m 0.9±0.2t 13 1.5±0.4 6 0 8.0 1190 730 5
CDo 1.6±0.2 2 2.1±0.3 8 3.2 1 6.7 -1190 740 2
p 1.6±0.2 2 2.0±0.3 8 3.1 1 7.4 720 -740 2
m 1.4±0.2 2 1.8±0.3 8 2.5 1 8.7 350 -810 2
RDo 1.5±0.3t 12 2.2±0.3 17 2.4±0.3 11 6.7 1050 730 24
p 1.3±0.2t 12 2.1±0.3 17 2.3±0.4 11 7.0 1030 770 24
m 1.1±0.3t 12 2.0±0.3 17 2.0±0.3 11 7.6 800 -760 24
RIo 1.2±0.3* 11 1.5±0.4 17 2.1 1 6.3 1760 -630 24
p 1.1±0.2t 11 1.5±0.4 17 1.9 1 7.0 1700 -650 24
m 1.0±0.3t 11 1.4±0.4 17 1.5 1 7.6 1430 -680 23
RPo 1.4+0.5 6 1.8±0.4 13 2.6 1 6.5 1690 490 17
p 1.3±0.4* 6 1.7±0.4 13 2.4 1 7.3 1660 -52° 17
m 1.1±0.3* 6 1.4±0.2 13 2.3 1 8.0 1450 -560 17
r, Radius; 0f azimuth angle between projection of r onto transverse plane and anterior-posterior axis; (D, angle between r and transverse
plane. Subsegment abbreviations may be found in Table 5. Values are mean±+SD.
*p<0.05, tp<0.01, compared with medium length group.
*Location derived from reported data.
 Dodge et al Normal Lumen Diameter 245

TABLE 8. Lumen Diameter of 12 Main Artery Segments, Each With Three Subsegments, in Normal Men, Normal Women, Men With
Left Ventricular Hypertrophy, and Men With Dilated Cardiomyopathy, All light Coronary Artery Dominant
Normal men Normal women Men with DCM Men with LVH
Subsegment Diameter (mm) n Diameter (mm) n Diameter (mm) n Diameter (mm) n
Left main
LMp 4.5±0.6 18(a) 4.0±0.5* 10 4.8±0.2 7 4.8+0.3 6
m 4.5±0.5 18(a) 3.9+0.4t 10 4.8+0.3 7 4.90.4 6
d 4.5+0.4 18(a) 3.8±0.3t 10 4.7±0.3 7 4.7+0.4 6
Left anterior descending
Llp 3.7±0.5 20 3.3±-0.4* 10 3.9±0.5 7 4.2±0.4* 6
m 3.6±0.5 20 3.2+0.5 10 3.8±0.5 7 3.90.5 6
d 3.5±0.6 20 3.1±0.5 10 3.6±0.3 7 3.6±0.7 6
L2p 2.9±0.5 20 2.8±0.5 10 3.1±0.5 7 3.2±0.4 6
m 2.5±0.5 20 2.4±0.4 10 2.8±0.5 7 3.0±0.4 6
d 2.3±0.4 20 2.2±0.5 10 2.4±0.4 7 2.8±0.4t 6
L3p 2.0±0.5 20 1.9±0.4 10 2.2±0.4 7 2.5±0.5 6
m 1.7±0.5 20 1.6±0.4 10 2.0±0.4 7 2.1±0.5 6
d 1.4±0.5 20 1.4±0.3 10 1.7±0.3 7 1.8±0.6 6
L4p 1.4+0.5 15(b) 1.1±0.5 8(c) 1.3±0.3 6(d) 1.1±0.2 3(e)
m 1.1±0.4 15(b) 0.9±0.3 8(c) 1.1±0.3 6(d) 1.1±0.4 3(e)
d 0.9±0.3 15(b) 0.7±0.2 8(c) 0.8±0.2 6(d) 1.0±0.2 3(e)
Left circumflex
Clp 3.4±0.5 20 2.9±0.5* 10 3.5±0.6 7 3.6±0.8 6
m 3.4±0.5 20 2.9±0.6* 10 3.3±0.7 7 3.6±0.6 6
d 3.3±0.5 20 2.9±0.6 10 3.1±0.7 7 3.6±0.6 6
C2p 2.8±0.5 11(f) 3.1±0.4 6(g) 3.1±0.5 4(g) 3.3±0.3* 4(g)
m 2.8±0.6 11(f) 3.1±0.4 6(g) 3.1±0.5 4(g) 3.3±0.2 4(g)
d 2.7±0.5 11(f) 3.0±0.4 6(g) 3.1±0.6 4(g) 3.1±0.3 4(g)
C3p 1.7±0.6 19 1.5±0.5 10 2.2±1.0 7 2.0±0.6 6
m 1.6±0.6 19 1.4±0.6 10 2.1±1.0 7 1.7±0.6 6
d 1.3±0.5 19 1.2±0.6 10 1.9±1.0 7 1.5±0.6 6
Right coronary
Rlp 4.0±0.6 20 3.4±0.7* 10 4.6±0.6 7 4.9±0.7t 6
m 3.9±0.6 20 3.3±0.6* 10 4.5±0.5* 7 4.6±0.7* 6
Downloaded from http://ahajournals.org by on September 16, 2019

d 3.8±0.5 20 3.2±0.6* 10 4.4±0.4* 7 4.5±0.7* 6

R2p 3.5±0.6 20 3.3±0.6 10 4.2±0.6* 7 4.4±0.7t 6
m 3.4±0.5 20 3.0±0.5 10 3.9±0.5* 7 4.3±0.5t 6
d 3.2±0.5 20 2.9±0.7 10 3.9±0.5t 7 4.1±0.5t 6
R3p 3.2±0.6 20 2.8±0.5 10 3.9±0.4t 7 4.0-0.5t 6
m 3.1±0.5 20 2.8±0.5 10 3.8±0.4t 7 4.0±0.6t 6
d 3.1±0.6 20 3.0±0.5 10 3.7±0.3t 7 4.1±0.5t 6
R4p 2.4±0.5 20 2.1±0.4 10 2.7±0.4 7 3.2±0.6t 6
m 2.2±0.5 20 1.8±0.5* 10 2.5±0.3 7 3.0±0.7* 6
d 1.9±0.6 20 1.7±0.5 10 2.1±0.4 7 2.7±0.7* 6
DCM, dilated cardiomyopathy; LVH, left ventricular hypertrophy. Subsegment artery abbreviations may be found in Table 5.
Reported segment numbers vary because of (a) LM absent; (b) IA absent; (c) L4 absent in one and unseen in one; (d) L4 unseen; (e)
L4 absent in two and unseen in one; (f) C2 absent in eight (when an OM was present) and inadequately visualized in one; (g) C2 absent.
*p<0.05, tp<0.01, each group compared with the group of normal men.
third is the middle (m) subsegment. The third segment is the normal men, normal women, men with LVH, and men with
distal (d) subsegment. The point at the beginning of a segment DCM is presented in Table 8.
(the proximal subsegment's proximal border) is designated as
the origin (o). References
A complete listing of lumen diameters in the 13 possible 1. Arnett EN, Isner JM, Redwood DR, Kent KM, Baker WP, Ack-
main artery segments, each with three subsegments, is pre- erstein H, Roberts WC: Coronary artery narrowing in coronary
sented in Table 6 for RCA-, small-RCA-, balanced, and heart disease: Comparison of cineangiographic and necropsy find-
LCA-dominant anatomy. Spatial location of these artery sub- ings. Ann Intern Med 1979;91:350-356
segments, from a previous report,"8 is included for complete- 2. White CW, Wright CB, Doty DB, Hiratza LF, Eastham CL, Har-
ness. Although the number of measurements at each segment rison DG, Marcus ML: Does visual interpretation of the coronary
appears to vary in this table, no segments were excluded arteriogram predict the physiologic importance of a coronary ste-
except as below. The LM was absent in three of 50 patients. [A nosis? N Engl J Med 1984;310:819-824
was absent in seven of 50 patients and unseen in five. C2 was 3. Harrison DG, White CW, Hiratzka LF, Doty DB, Barnes DH,
absent in 15 of 50 patients (when an obtuse marginal branch Eastham CL, Marcus ML: The value of lesion cross-sectional area
was present) and inadequately visualized in one. Similar
determined by quantitative coronary angiography in assessing the
reasons explain other numerical discrepancies.
physiologic significance of proximal left anterior descending coro-
nary arterial stenoses. Circulation 1984;69:1111-1119
Table 7 contains a complete listing of lumen diameters at 19 4. Glagov S, Weisenberg E, Zarins CK, Stankunavicius R, Kolettis
branch artery segments, each with three subsegments, for GJ: Compensatory enlargement of human atherosclerotic coro-
short, medium, and long vessels. The spatial locations of these nary arteries. N Engl J Med 1987;316:1371-1375
artery subsegments in normal men has been included.18 5. Langille BI, O'Donnell F: Reductions in arterial diameter pro-
A complete list of lumen diameters at three subsegments in duced by chronic decreases in blood flow are endothelium depen-
each of the 12 main artery segmnents present in RCA-dominant dent. Science 1986;231:405-407
 246 Circulation Vol 86, No 1 July 1992
6. Learoyd BM, Taylor MG: Alterations with age in the viscoelastic
properties of human arterial walls. Circ Res 1966;18:278-292
7. Bader H: Dependence of wall stress in the human thoracic aorta
on age and pressure. Circ Res 1967;20:354-361
8. Markis JE, Joffe CD, Cohn PF, Feen DJ, Herman MV, Gorlin R:
Clinical significance of coronary artery ectasia. Am J Cardiol 1976;
37:217-222
9. Swaye PS, Fisher LD, Litwin P, Vignola PA, Judkins MP, Kemp
HG, Mudd JG, Gosselin AJ: Aneurysmal coronary artery disease.
Circulation 1983;67:134-138
10. Hartnell GG, Parnell BM, Pridie RB: Coronary artery ectasia: Its
prevalence and clinical significance in 4,993 patients. Br Heart J
1985;54:392-395
11. Takahashi M, Mason W, Lewis AB: Regression of coronary aneu-
rysms in patients with Kawasaki syndrome. Circulation 1987;75:
387-394
12. Jaffe RB, Glancy DL, Epstein SE, Brown BG, Morrow AG: Cor-
onary arterial-right heart fistulae: Long-term observations in
seven patients. Circulation 1973;48:133-143
13. Young DF, Cholvin NR, Roth AC: Pressure drop across artificially
induced stenoses in the femoral arteries of dogs. Circ Res 1975;36:
735-743
14. Gould KL, Lipscomb K: Effects of coronary stenoses on coronary
flow reserve and resistance. Am J Cardiol 1974;34:48-55
15. Gould KL, Lipscomb K, Hamilton GW: Physiologic basis for
assessing critical coronary stenosis: Instantaneous flow response
and regional distribution during coronary hyperemia as measures
of coronary flow reserve. Am J Cardiol 1974;33:87-94
16. Gould KL, Kelley KO: Physiological significance of coronary flow
velocity and changing stenosis geometry during coronary vasodi-
lation in awake dogs. Circ Res 1982;50:695-704
17. MacAlpin RN, Abbasi AS, Grollman JH, Eber L: Human coronary
artery size during life. Radiology 1973;108:567-576
18. Dodge JT, Brown BG, Bolson EL, Dodge HT: Intrathoracic spatial
location of specified coronary segments on the normal human
heart. Circulation 1988;78:1167-1180
19. Brown BG, Bolson E, Frimer M, Dodge HT: Quantitative coro-
nary arteriography: Estimation of dimensions, hemodynamic resis-
Downloaded from http://ahajournals.org by on September 16, 2019
tance, and atheroma mass of coronary artery lesions using the
arteriogram and digital computation. Circulation 1977;55:329-337
20. Dixon WJ, Massey FJ Jr: Introduction to Statistical Analysis. New
York, McGraw-Hill Book Co, Inc, 1969, pp 335-340
21. Heiss G, Tamir I, Davis CE, Tyroler HA, Rifkind BM, Schonfeld
G, Jacobs D, Frantz ID Jr: Lipoprotein-cholesterol distributions in
selected North American populations: The Lipid Research Clinics
Program prevalence study. Circulation 1980;61:302-315
22. Wahl PW, Warnick GR, Albers JJ, Hoover JJ, Walden CE, Berge-
lin RO, Olgilvie JT, Hazzard WR, Knopp RH: Distribution of
lipoprotein triglyceride and lipoprotein cholesterol in an adult
population by age, sex, and hormone use. Atherosclerosis 1981;39:
111-124
23. Brown BG, Bolson EL, Dodge HT: Dynamic mechanisms in
human coronary stenosis. Circulation 1984;70:917-922
24. Liu LB, Richardson T, Holt JH, Taylor CB: Luminal eccentricity
and deformity in atherosclerotic coronary arteries. ParoiArterielle
1976;3:167-169
25. Brown BG, Petersen RB: Computer-assisted measurements of cor-
onary artery stenosis. (reply) Circulation 1979;60:1196
26. Kimball BP, LiPreti VL, Bui S, Wigle ED: Comparison of proximal
left anterior descending and circumflex coronary artery dimen-
sions in aortic valve stenosis and hypertrophic cardiomyopathy.Am
J Cardiol 1990;65:767-771
27. Baroldi G, Scomazzoni G: Coronary Circulation in the Normal and
the Pathologic Heart. Washington, DC, Armed Forces Institutes of
Pathology, 1967, pp 5-23
28. Harrison CV, Wood P: Hypertensive and ischemic heart disease: A
comparative clinical and pathological study. Br Heart J 1949;11:
205-229
29. Zamir M, Chee H: Segment analysis of human coronary arteries.
Blood Vessels 1987;24:76-84
30. Koiwa Y, Bahn RC, Ritman EL: Regional myocardial volume
perfused by the coronary artery branch: Estimation in vivo. Circu-
lation 1986;74:157-163
31. Woods JD: Relative ischaemia in the hypertrophied heart. Lancet
1961;1:696-698
32. Kennedy JW, Baxley WA, Figley MM, Dodge HT, Blackmon JR:
Quantitative angiocardiography: I. The normal left ventricle in
man. Circulation 1966;34:272-278
33. Kennedy JW, Kaiser GC, Fisher LD, Fritz JK, Meyers W, Mudd
JG, Ryan TJ: Clinical and angiographic predictors of operative
mortality from the Collaborative Study in Coronary Artery Sur-
gery (CASS). Circulation 1981;63:793-802
34. Spray TL, Roberts WC: Status of the grafts and the native coro-
nary arteries proximal and distal to coronary anastomotic sites of
aortocoronary bypass grafts. Circulation 1977;55:741-749
35. Fisher LD, Kennedy JW, Davis KB, Maynard C, Fritz JK, Kaiser
G, Myers WO: Association of sex, physical size, and operative
mortality after coronary artery bypass in the Coronary Artery
Surgery Study (CASS). J Thorac Cardiovasc Surg 1982;84:334-341
36. Kahn SS, Nessim S, Gray R, Czer LS, Chaux A, Matloff J:
Increased mortality of women in coronary artery bypass surgery:
Evidence for referral bias. Ann Intern Med 1990;112:561-567
37. Rodriguez FL, Robbins SL: Capacity of human coronary arteries.
Circulation 1959;19:570-578
38. Rowe GG, Afonso S, Lugo JE, Castillo CA, Boake WC, Crumpton
CW: Coronary blood flow and myocardial oxidative metabolism at
rest and during exercise in subjects with severe aortic valve disease.
Circulation 1965;32:251-257
39. Henry PD, Eckberg D, Gault JH, Ross J Jr: Depressed inotropic
state and reduced myocardial oxygen consumption in the human
heart. Am J Cardiol 1973;31:300-306
40. Kennedy JW, Twiss RD, Blackmon JR, Dodge HT: Quantitative
angiocardiography: III. Relationships of left ventricular pressure,
volume, and mass in aortic valve disease. Circulation 1968;38:
838-845
41. Benjamin JI, Schuster EH, Bulkley BH: Cardiac hypertrophy in
idiopathic dilated congestive cardiomyopathy: A clinicopathologic
study. Circulation 1981;64:442-447
42. Fallen EL, Elliott WC, Gorlin R: Mechanisms of angina in aortic
stenosis. Circulation 1967;36:480-488
43. Bertrand ME, LaBlanche JM, Tilmant PY, Thieuleux FP, Delforge
MR, Carre AG: Coronary sinus blood flow at rest and during
isometric exercise in patients with aortic valve disease. Am J Car-
diol 1981;47:199-205
44. Hutchins GM, BuLldey BH, Miner MM, Boitnott JK: Correlation
of age and heart weight with tortuosity and caliber of normal
human coronary arteries. Am Heart J 1977;94:196-202
45. Gross L: The Blood Supply of the Heart. New York, Hoeber, 1921,
p 151
46. Ganz W, Tamura K, Marcus HS, Donoso R, Yoshida S, Swan
HJC: Measurement of coronary sinus blood flow by continuous
thermodilution in man. Circulation 1971;44:181-195
47. Reiber JHC, Kooijman CJ, den Boer A, Serruys PW: Assessment
of dimensions and image quality of coronary contrast catheters
from cineangiograms. Cathet Cardiovasc Diagn 1985;11:521-531