Assessment of Flow
All angiographic end points were prospectively assessed at 90
minutes. The TIMI flow grade, as previously defined,1 was assessed
at the TIMI angiographic core laboratory by a single observer
(C.M.G.) who was blinded to treatment assignment and clinical
outcome. The corrected TIMI frame count (CTFC) is the number of
cine frames required for contrast to first reach standardized distal
coronary landmarks in the culprit artery and is measured by use of a
frame counter on a cine viewer.7,8 A frame count of 100, a value that
is the 99th percentile of patent vessels, was imputed to an occluded
vessel.7,8 The CTFC is a measure of time, and the data were
converted when necessary to be based on the most common filming
speed in the United States of 30 frames per second.7,8 The collateral grade was assessed at 90 minutes10 and was based on the presence of collaterals to the culprit artery. TMP grades are defined in Table 1. Blush was assessed distal to the culprit lesion, and views were chosen to minimize superimposition of noninfarcted territories in the assessment of the TMP grade the culprit artery. The duration of cine filming was required to exceed 3 cardiac cycles in the washout phase to assess washout of the myocardial blush. Care was taken not to mistake filling of the venous system, such as the great cardiac vein, as blush. Blush was assessed during the same phase of the cardiac cycle, because it may be less intense during diastole. Mortality was confirmed by a clinical events committee.

Statistical Analysis

Analyses were performed with Stata statistical software version 6.0.11 Variables were compared with the Fisher’s exact test or x2 test for categorical data. The Student’s t test or ANOVA was used for analysis of normally distributed continuous variables. The nonparametric Wilcoxon rank sum test (for 2-way comparisons) or the Kruskal-Wallis test (for 3-way comparisons) was used to compare continuous variables when the data were not normally distributed or
when data were imputed to an occluded vessel. Data are summarized as mean6SD.

Results
Baseline Demographic and Angiographic
Characteristics of the TMP Grades
There was no difference among TMP grades with respect to
many demographic and angiographic variables: age, sex,
blood pressure, pulse rate, ejection fraction, history of myocardial
infarction (MI), and presence of angiographically
visible collaterals (Tables 2 and 3). Compared with patients
who exhibited some myocardial perfusion (TMP grades 1, 2,
or 3), patients without detectable perfusion (TMP grade 0)
had significantly slower epicardial flow (higher CTFCs and a
lower incidence of TIMI grade 3 flow), a significantly greater
thrombus burden, and tighter epicardial stenoses (Table 3). In
these cases, LAD infarcts tended to be involved more
frequently (Table 3).
Relationship of TMP Grade to Mortality
Patients with TMP grade 0 had a higher 30-day mortality rate
(6.2%, 27 of 434 patients) than patients with TMP grade 1
(5.1%, 4 of 79 patients), TMP grade 2 (4.4%, 2 of 46
patients), or TMP grade 3 (2.0%, 4 of 203 patients; TMP
grades 0 and 1 combined to achieve adequate power,
P50.055 by Fisher’s exact test, P50.046 by logistic regression)
(Figure 1). Likewise, when TMP grades 2 and 3 flow
were combined, the mortality rate was lower than that in
patients with TMP grade 0 or 1 (2.4% [6 of 249 patients]
versus 6.0% [31 of 513 patients]; P50.03).
Risk Stratification Within TIMI Grade 3 Flow by
Use of TMP Grades
Among patients with TIMI grade 3 flow in the epicardial
artery, use of TMP grades allowed further risk stratification
such that reduced myocardial perfusion was related to a
higher risk of 30-day mortality: the mortality rate was 0.7%
among those with TMP grade 3 (1/137) versus 4.7% among
all others (15/318; P50.05) (Figure 2). When the patients
were further divided into 3 TMP grades, the same relationship
held true: the mortality rate was 0.73% for TMP grade 3,
2.9% for TMP grade 2, and 5.0% for TMP grades 0 and 1
(P50.03 for TMP grade 3 versus grades 0, 1, and 2; 3-way
P50.066) (Figure 2). For an open (TMP grade 2 or 3) versus
a closed (TMP grade 0 or 1) microvasculature, the P value
was 0.04. Among those patients with less than TIMI grade 3
flow in the epicardial artery, those with TMP grade 3 flow
also tended to have better outcomes (Figure 2). Likewise,
among patients with a CTFC of ,40 (a value that quantitatively
characterizes TIMI grade 3 flow),7 TMP grade 3 was
associated with reduced mortality (0.8%, 1 of 131 patients)
compared with TMP grades 0 through 2 (4.6%, 14 of 306
patients; P50.05) (Figure 3). A similar gradient was seen in
patients with CTFC $40, with a 4.5% (3/67) mortality rate in
TMP grade 3 compared with 7.8% (18/232) in TMP grades 0
through 2 (4-way P50.02) (Figure 3).
To evaluate the independent contribution of myocardial
perfusion to mortality, a multivariate model was developed
that included angiographic and demographic variables previously
shown to be related to mortality.8 The presence of TMP
grade 3 flow was an independent correlate of 30-day mortality
(OR 0.35, 95% CI 0.12 to 1.02, P50.054) in a multivariate
model that adjusted for variables that have been previously
identified in the TIMI studies as correlates of
mortality8: TIMI grade 3 flow (P5NS), CTFC (OR 1.02 per
1-frame rise, P50.06), presence of an anterior MI (OR 2.3,
P50.03), pulse rate on admission (P5NS), female sex
(P5NS), and age (OR 1.1 per 1-year rise, P,0.001).

Combination of TIMI Epicardial Flow and TMP
Grades and Their Relationship to Mortality
Those patients with both epicardial TIMI grade 3 flow and
myocardial perfusion grade 3 flow (successful epicardial and
tissue-level perfusion) had a low mortality rate of 0.73%
(1/137), whereas those with grades of 0 or 1 for both TIMI
epicardial flow and myocardial perfusion had a mortality rate
of 10.9% (14 of 129 patients) (Figure 4). Patients with either
incomplete epicardial or myocardial flow (ie, patients with
neither the combination of TIMI flow grade 3 and TMP grade
3 or the combination of TIMI flow grade 0/1 and TMP grade
0/1) had an intermediate mortality rate of 4.4% (21/483;
3-way P,0.001) (Figure 4). The presence of both TIMI
epicardial flow and myocardial perfusion grade 3 (successful
epicardial and tissue-level reperfusion) was a multivariate
predictor of low mortality (OR 0.056, P50.006), even after
adjustment for anterior MI location and age (overall model
n5742, P,0.0001). Thus, in the multivariate model, the odds
of death by 30 days for patients with an occluded epicardial
artery and no tissue-level reperfusion (TIMI flow grade 0/1
and TMP grade 0/1) were 18 times as great as in those with
both successful epicardial and successful tissue-level reperfusion
(TIMI flow grade 3 and TMP grade 3).
Discussion
Improved epicardial blood flow assessed by use of either
TIMI flow grades1–6 or the TIMI frame count7,8 has been
related to reduced mortality after thrombolytic administration.
The data presented here indicate that improved myocardial
perfusion at 90 minutes after thrombolytic administration
is related to reduced mortality independent of epicardial
blood flow. Its reliance on ordinary visual inspection of the
angiogram without the use of sophisticated equipment allows
the method to be conveniently and broadly applied. Although
simple, the myocardial perfusion grade scheme is semiquantitative
and is adjusted for the heart rate of the patient.
These findings extend those of previous investigators1–6
who have reported that patients with TIMI grade 3 flow have
a reduced incidence of mortality. Use of the TMP grades
allows additional risk stratification into low- and high-risk
subgroups such that slower myocardial perfusion among
patients with TIMI grade 3 flow is related to higher mortality
(0.7% for TMP grade 3 versus 4.7% for TMP grades 0 to 2;
Figure 2). Interestingly, TMP grade 3 appeared to be a better
marker of reduced mortality (2.0%) than the presence of
TIMI flow grade 3 (3.5%; 1-sided P50.2), which has been
the “gold standard” for assessment of complete reperfusion
over the past 15 years. Likewise, the TMP grade was an
independent predictor of mortality when adjustments were
made for the epicardial TIMI flow grades, infarct artery
location, and age. Indeed, those patients with TIMI grade 3
flow with absent or near-absent myocardial perfusion (TMP
grade 0 or 1) had a mortality rate (5.0%) as high as that in
patients with unsuccessful restoration of epicardial artery
patency (TIMI 0 to 2; 4.7%) but preservation of myocardial
perfusion (TMP grade 3), presumably through collaterals.

Finally, the combined use of the TMP grade and the TIMI
flow grade appears to identify 2 subgroups of patients with
extremely low and high risks of mortality, respectively.
Patients with both normal epicardial flow and myocardial
perfusion (both grade 3) had a mortality rate of 0.73%. As
we8 have reported in the past, patients with hyperemic flow
(CTFCs faster than the 95th percentile, ,14 frames, TIMI
grade 4 flow) were found to have a mortality rate of 0%
(0/41), and these patients had nearly twice the incidence of
excellent myocardial perfusion (TMP grade 3) as other
groups (44.8% versus 26.2%; P50.03). Improved myocardial
perfusion may explain in part the favorable mortality rate that
we have reported for this subgroup of patients. Thus, the
TMP grade adds additional prognostic information to the
conventional epicardial TIMI flow grades and TIMI frame
counts.
Relationship to Previous Work in the Field
Van’t Hof et al12 showed that the presence of no, minimal,
moderate, or normal blush (relative to the contrast density in
uninvolved territories) is related to mortality after primary
angioplasty. The method used in the present study differs
from that study in that we characterize the duration of the
blush rather than the brightness or density of the blush. The
patients in the present study were treated with thrombolysis,
whereas those in the study by Van’t Hof et al were treated
with primary PTCA. Thus, it appears that both the contrast
density and the duration of blush may be related to mortality,
but both measures have not been implemented simultaneously
in the same study to determine whether they are
independent of one another.
Myocardial contrast echocardiography (MCE) has also
been used to characterize the no-reflow phenomenon.13–16
The incidence of no reflow varies across studies. Whereas we
observed that nearly half of the patients had minimal or no
blush on the coronary arteriogram, prior reports have ranged
from 23%12 to 56%14 of patients having no-reflow after
restoration of patency (via either thrombolytic administration
or primary PTCA) when MCE was used. The lower percentages
in some MCE studies likely reflect the lower number of
patients with no reflow after patency is restored, whereas our
series includes patients with occluded epicardial arteries. In
the study by Ito et al,13 patients were excluded if they had a
tight residual stenosis, and 29 of 39 patients were treated with
primary PTCA.
Myocardial tissue perfusion has also been assessed by
Maes et al17 using PET. Among patients with TIMI grade 3
epicardial flow, both regional and global ejection fraction at
5 days and 3 months after infarction were lower in patients
with severely impaired myocardial flow than in patients with
moderately decreased flow or adequate tissue reperfusion.
This reduced contractile function may explain in part the
mortality risk observed in patients with TMP grades 0 or 1.
Study Limitations
TMP grades were available in 88% of patients with 30-day
mortality data in the TIMI 10B trial (762 of 865 patients).
With prospective emphasis on a longer duration of cine
filming, adequate panning, and the use of a 9-in image
intensifier in coronary angiography, it is likely that the rate of
ascertainment will be greater. The mortality rate among
patients in whom TMP grades were assessed (4.9%, 37 of 762
patients) was no different from that in the study group overall
(5.3%, 46 of 865 patients). The reproducibility of the TMP
grades remains to be determined. It must be borne in mind
that although 90-minute myocardial perfusion and epicardial
coronary blood flow are both related to mortality, there are
other causes of death that may be unrelated to 90-minute
perfusion, such as intracranial hemorrhage, reinfarction, ventricular
arrhythmias, and mechanical complications. Both
rescue and adjunctive angioplasty may have obscured differences
in outcomes that would have been attributable to
90-minute TIMI flow grades and TMP grades. However,
even when the analysis was stratified by those patients who
did not subsequently undergo rescue or adjunctive PTCA or
stenting and those who did, the same relationships were
observed (3-way P50.003 and P50.088, respectively).
Conclusions
After administration of thrombolytic drugs in patients with acute
MI, impaired perfusion of the myocardium on coronary arteriography
as assessed by TMP grade is related to a higher risk of
mortality that is independent of flow in the epicardial artery. The
use of the TMP grade permits risk stratification, even among
patients with TIMI grade 3 flow. Patients with both normal
epicardial flow (TIMI grade 3 flow) and normal tissue-level
perfusion (TMP grade 3 flow) had an extremely low risk of
mortality (0.73%) and in a multivariate model were 18 times less
likely to die by 30 days than patients with occluded epicardial
flow (TIMI grade 0 or 1 flow) and no tissue perfusion (TMP
grade 0 or 1). The TMP grade represents a simple, readily
available method to assess myocardial perfusion in patients
undergoing reperfusion therapy.
Acknowledgment
This study was supported in part by a grant from Genentech, Inc,
South San Francisco Calif.