Relationship Between TIMI Frame Count and Clinical Outcomes After Thrombolytic Administration
Background—The corrected TIMI frame count (CTFC) is the number of cine frames required for dye to first reach standardized distal coronary landmarks, and it is an objective and quantitative index of coronary blood flow.
Methods and Results—The CTFC was measured in 1248 patients in the TIMI 4, 10A, and 10B trials, and its relationship to clinical outcomes was examined. Patients who died in the hospital had a higher CTFC (ie, slower flow) than survivors (69.6±35.4 [n=53] versus 49.5±32.3 [n=1195]; P=0.0003). Likewise, patients who died by 30 to 42 days had higher CTFCs than survivors (66.2±36.4 [n=57] versus 49.9±32.1 [n=1059]; P=0.006). In a multivariate model that excluded TIMI flow grades, the 90-minute CTFC was an independent predictor of in-hospital mortality (OR=1.21 per 10-frame rise [95% CI, 1.1 to 1.3], an ≈0.7% increase in absolute mortality for every 10-frame rise; P<0.001) even when other significant correlates of mortality (age, heart rate, anterior myocardial infarction, and female sex) were adjusted for in the model. The CTFC identified a subgroup of patients with TIMI grade 3 flow who were at a particularly low risk of adverse outcomes. The risk of in-hospital mortality increased in a stepwise fashion from 0.0% (n=41) in patients with a 90-minute CTFC that was faster than the 95% CI for normal flow (0 to 13 frames, hyperemia, TIMI grade 4 flow), to 2.7% (n=18 of 658 patients) in patients with a CTFC of 14 to 40 (a CTFC of 40 has previously been identified as the cutpoint for distinguishing TIMI grade 3 flow), to 6.4% (35/549) in patients with a CTFC >40 (P=0.003). Although the risk of death, recurrent myocardial infarction, shock, congestive heart failure, or left ventricular ejection fraction ≤40% was 13.0% among patients with TIMI grade 3 flow (CTFC ≤40), the CTFC tended to segregate patients into lower-risk (CTFC ≤20, risk of adverse outcome of 7.9%) and higher-risk subgroups (CTFC >20 to ≤40, risk of adverse outcome of 15.5%; P=0.17).
Conclusions—Faster (lower) 90-minute CTFCs are related to improved in-hospital and 1-month clinical outcomes after thrombolytic administration in both univariate and multivariate models. Even among those patients classified as having normal flow (TIMI grade 3 flow, CTFC ≤40), there may be lower- and higher-risk subgroups.
The conventional Thrombolysis In Myocardial Infarction (TIMI) flow grade classification scheme1 has been a valuable tool to compare angiographic and clinical outcomes after thrombolysis.2 3 4 5 6 7 However, this classification scheme is limited by interobserver variability, its categorical nature, its limited statistical power, and the fact that nonculprit flow (used to gauge TIMI grade 3 flow) is abnormal.8 Recently, we described a new index of coronary blood flow called the corrected TIMI frame count (CTFC), in which the number of frames required for dye to reach standardized distal landmarks is counted.8 We hypothesized that faster flow (lower CTFCs) would be related to improved clinical outcomes.
A total of 1248 patients from the TIMI 4, 10A, and 10B trials had CTFC and outcomes data available for analysis. The TIMI 4 trial was a randomized, double-blind comparison of 3 thrombolytic regimens in 416 patients9 : anistreplase (Eminase or APSAC, 30 U IV), front-loaded recombinant tissue plasminogen activator (rtPA) (Activase or alteplase), or combination therapy. The TIMI 10A trial was a nonrandomized, open-label study of 8 ascending doses of TNK (5, 7.5, 10, 15, 20, 30, 40, and 50 mg IV over 5 seconds) in 113 patients.10 TIMI 10B was a randomized comparison of TNK (30, 40, and 50 mg) and 90-minute infusion of rtPA (Activase or alteplase) in 854 patients.11 Angiography was performed at 60, 75, and 90 minutes after thrombolysis.8 9 10 11 Nitroglycerin was administered every 15 minutes if the systolic blood pressure exceeded 110 mm Hg.8 9 10 11
Assessment of Flow
TIMI flow grade was assessed at an angiographic core laboratory as previously defined.1 The CTFC is the number of cine frames required for contrast to first reach standardized distal coronary landmarks in the culprit artery and is measured with a frame counter on a cine viewer.8 A frame count of 100, a value that is the 99th percentile of patent vessels, was imputed to an occluded vessel.8 CTFC is a measure of time, and data were converted when necessary to be based on the most common filming speed used in the United States (30 frames/s). The CTFC was divided by 30 to calculate the transit time for dye to traverse the length of the artery to the landmark in seconds and multiplied by 1000 to calculate the time in milliseconds. This was used along with the heart rate to calculate the fraction of a cardiac cycle required for dye to traverse the artery: fraction of cardiac cycle=(CTFC/30 seconds)/(60 seconds/heart rate). Calculation of the fraction of a cardiac cycle required for dye to traverse the culprit artery normalizes the CTFC for heart rate.
In-hospital mortality and recurrent myocardial infarction (MI) were ascertained in all 3 trials (TIMI 4, 10A, and 10B). A composite end point was ascertained in TIMI 4 and TIMI 10A and consisted of any of the following in-hospital adverse events: death; recurrent MI; development of new severe congestive heart failure (CHF) or shock; or ejection fraction <40%. In TIMI 4, a predischarge ejection fraction was obtained, and in TIMI 10A, a 90-minute ejection fraction was obtained. Mortality and morbidity committees confirmed all adverse events. Analyses were performed with STATA statistical software.12 Variables were compared by Fisher’s exact test for categorical data and ANOVA for continuous variables. The nonparametric Wilcoxon rank sum test was used to compare continuous variables that were not normally distributed. Data are summarized as mean±SD. Due to the relatively small number of deaths (n=53) in the sample of 1248 patients, the 5 most significant clinical and angiographic univariate predictors of in-hospital mortality (one for each 10 events) with a value of P<0.01 were entered into multivariate models. A linear spline analysis was performed to reject the null hypothesis that the slope of the relationship between the CTFC and outcomes remained constant across intervals of the CTFC: 0 to 20, 21 to 40, 41 to 60, 61 to 80, and ≥81. Fractional polynomial regression was also used to determine if the use of a higher-order polynomial (quadratic, cubic, or greater) was superior to linear regression in fitting the relationship between outcomes and the CTFC as a continuous variable.
Univariate Predictors of In-Hospital Mortality
The 5 univariate correlates (Table 1⇓) that successfully competed for entry into the multivariate model of in-hospital mortality (Tables 2⇓ and 3⇓) included age (χ2=63.9, P<0.001), CTFC (χ2=17.3, P<0.001), anterior MI location (χ2=12.5, P<0.001), heart rate (χ2=9.3, P<0.001), and female sex (χ2=8.3, P=0.003). TIMI flow grades and CTFC were colinear, and TIMI flow grades were dropped from the model.
Relationship of TIMI Frame Count to Mortality
The CTFC method was introduced during the TIMI 4 trial, and in this exploratory data set, the 90-minute CTFC was a univariate predictor of in-hospital mortality (OR=1.24 expressed per 10-frame rise throughout; 95% CI, 1.1 to 1.4; P=0.007; n=318). This observation was then prospectively validated in the TIMI 10 trials (OR=1.16; 95% CI, 1.1 to 1.3; P=0.002; n=930). In all 3 trials combined, the CTFC was a univariate predictor of in-hospital mortality (OR=1.18; 95% CI, 1.1 to 1.3; P<0.001). When the percent change in the CTFC was used instead, the OR was 1.07 for every 20% increase in the CTFC (P<0.001).
Patients who died in the hospital had higher CTFCs (69.6±35.4 [n=53] versus 49.5±32.3 [n=1195]; P=0.0003) (Figure 1⇓). When the analysis was performed with either the CTFC at 90 minutes after thrombolytic administration or, if the patient went on to have an intervention, immediately after 90-minute angiography (rescue or adjunctive percutaneous intervention [PCI]), the CTFC at the completion of any and all therapy (the final CTFC) was also significantly higher in patients who died in the hospital (53.0±36.7 [n=45] versus 38.4±25.9 [n=1110]; P=0.03).
The CTFC identified a subgroup of patients at particularly low mortality risk among those classified as having normal flow. The risk of in-hospital mortality increased in a stepwise fashion from 0.0% (n=41) in patients with a 90-minute CTFC that was faster than the 95% CI for normal flow (0 to 13 frames, hyperemia, TIMI grade 4 flow), to 2.7% (18/658) in patients with a CTFC of 14 to 40 (a CTFC of 40 has previously been identified as the cutpoint for distinguishing TIMI grade 3 versus 2 flow),8 to 6.4% (35/549) in patients with a CTFC >40 (3-way P=0.003) (Figure 2⇓).
The 90-minute CTFC (OR=1.21; z=4.29; P<0.001), the final CTFC (OR=1.15; z=2.81; P=0.005), the fraction of the cardiac cycle (OR=1.46; P<0.001), and the dye transit time in seconds (OR=1.64; P<0.001) were each independent predictors of in-hospital mortality even when the 4 other most significant variables (age, heart rate, anterior MI, and female sex) were adjusted for in multivariate models. When the analysis was confined to the 881 patients who did not undergo PCI, the 90-minute CTFC (OR=1.15; z=2.43; P=0.02) was a multivariate predictor of in-hospital mortality. The 90-minute CTFC was also related to mortality later, at 30 to 42 days (OR=1.15; P<0.001), and likewise, the 90-minute CTFC was higher for patients who died at 30 to 42 days than for those who lived (66.2±36.4 [n=57] versus 49.9±32.1 [n=1059]; P=0.006).
When the conventional TIMI flow grades were used instead of CTFC data, TIMI grade 0/1 flow differed significantly from TIMI grade 3 flow (OR=4.4; P<0.001), but TIMI grade 2 flow did not differ from TIMI grade 3 flow (OR=0.6; P=NS) in the multiple logistic regression model.
Comparison of 60-, 75-, and 90-Minute TIMI Frame Count Data in Predicting Mortality
To ensure a consistent sample size and statistical power, clinical outcomes were compared in the 563 patients in whom the CTFC was assessed at all 3 time points. Although the 90-minute CTFC was associated with mortality (OR=1.12; P=0.049) neither the 75-minute CTFC (OR=1.07; P=0.24) nor the 60-minute CTFC (OR=1.08; P=0.21) achieved significance. The mortality rate for patients with an occluded artery at 60 minutes (6.1%, 12/197) tended to be lower than that for patients with an occluded artery at 90 minutes (10.2%, 29/284; P=0.14). Patients with an occluded artery at 60 minutes that opened by 90 minutes had a mortality rate of 0.0% (0/49), which was lower than the 8.2% (12/146) mortality rate for those whose arteries remained occluded at both 60 and 90 minutes (P=0.05).
Relationship Between CTFC and Death or Recurrent MI
The CTFC for patients who died or sustained a recurrent MI was significantly higher (60.0±34.7 [n=100] versus 49.5±32.4 [n=1144]; P=0.01) (Figure 3⇓), as was the final CTFC (47.5±33.3 [n=89] versus 38.2±25.8 [n=1066]; P=0.05). The 90-minute CTFC (OR=1.10; z=3.17; P=0.002), the final CTFC (OR=1.11; z=2.78; P=0.005), the fraction of the cardiac cycle (OR=1.18; P=0.01), and the dye transit time (OR=1.35; P=0.002) were all multivariate predictors of death/recurrent MI. When the analysis was confined to the 881 patients who did not undergo PCI, the 90-minute CTFC (OR=1.10; P=0.04) was a multivariate predictor of death/recurrent MI.
When the conventional TIMI flow grades were used instead of the CTFC, TIMI grade 0/1 differed from TIMI grade 3 flow (OR=4.4; P<0.001), but TIMI grade 2 flow did not differ from TIMI grade 3 flow (OR=0.6; P=NS) in the multiple logistic regression model.
Relationship Between CTFC and Composite Adverse Events
The 90-minute CTFC for patients with any adverse outcome (death, recurrent MI, shock/severe CHF, or left ventricular ejection fraction <40%) was significantly higher than that for patients without an event (66.3±36.9 [n=83] versus 52.4±31.9 [n=341]; P=0.002) (Figure 4⇓), as was the final CTFC (55.8±33.7 [n=74] versus 41.4±26.1 [n=316]; P=0.0006). The 90-minute CTFC (OR=1.13; z=3.13; P=0.003), the final CTFC (OR=1.13; z=2.56; P=0.01), the fraction of the cardiac cycle (OR=1.30; P=0.002), and the dye transit time (OR=1.45; P=0.003) were multivariate predictors of adverse outcomes. When the analysis was confined to those patients who did not undergo PCI, the 90-minute CTFC (OR=1.13; P=0.02) was a multivariate predictor of adverse outcomes.
When the conventional TIMI flow grades were used instead of frame count data, TIMI grade 0/1 differed from TIMI grade 3 flow (OR=2.3; P=0.006), and TIMI grade 2 flow differed from TIMI grade 3 flow (OR=1.9; P=0.02) in the multiple logistic regression model. Within the categories of TIMI grade 2 and 3 flows, the CTFC was a multivariate predictor of adverse outcomes (OR=1.17; P=0.015). Figure 5⇓ shows the risk of adverse outcomes within the categories of TIMI 2 and 3 flow. Although the overall rate of adverse outcomes for TIMI grade 3 flow was 13.0% (33 of 253 patients), those patients with a CTFC ≤20 had an incidence of 7.9% (3 of 38), and those with a CTFC >20 and ≤40 (the previously defined upper bound of TIMI grade 3 flow) tended to have a higher incidence at 15.5% (26 of 168 patients; P=0.17).
Linearity of Relationship Between CTFC and Clinical Outcomes
A linear spline analysis failed to reject the null hypothesis that there was a linear relationship between the CTFC and outcomes (ie, the null hypothesis that the ORs were the same across intervals of the CTFC was not rejected). A logarithmic transformation of the CTFC data provided no better explanatory power in any of the outcome variables. Fractional polynomial regression demonstrated that the use of a higher-order polynomial (quadratic, cubic, or greater) was not superior to linear regression in fitting the relationship between outcomes and the CTFC as a continuous variable.
Higher CTFCs 90 minutes after thrombolytic administration are associated with an increased risk of mortality and composite end points in both univariate and multivariate models. This was true of events in the hospital and by 30 to 42 days after treatment. These findings extend those of the Global Utilization of Streptokinase and TPA for Occluded Arteries (GUSTO) angiographic investigators, who have reported that patients with TIMI grade 3 flow have improved outcomes compared with patients with TIMI grade 2 flow at 90 minutes.3 The use of the CTFC identifies a subgroup of patients with TIMI grade 3 flow who are at extremely low risk of mortality: none of those patients who achieved a CTFC <14 after thrombolysis died (TIMI grade 4 flow, or flow that is faster than the 95th percentile for normal flow). Likewise, in the setting of PTCA for acute coronary syndromes (1057 trial patients in the Randomized Efficacy Study of Tirofiban for Outcomes and Restenosis [RESTORE] trial), we have recently demonstrated that survivors had a lower CTFC (faster flow) than patients who died after PTCA, and none of the 376 patients with a CTFC <14 after PTCA (hyperemic flow or TIMI grade 4 flow) died.13 14 With respect to the composite end point of death, recurrent MI, shock, or left ventricular ejection fraction <40%, the CTFC was a multivariate predictor of adverse outcomes (P=0.015) among patients with TIMI grade 2 and 3 flows in a model that controlled for age, anterior MI, heart rate, and sex. Even within TIMI grade 3, there tended to be lower- and higher-risk subgroups: a CTFC ≤20 was associated with a risk of adverse outcome of 7.9%, whereas a CTFC >20 to ≤40 was associated with a risk of adverse outcome of 15.5% (Figure 4⇑). Thus, although the TIMI flow grades and the TIMI frame count are colinear and therefore do not appear to provide independent predictive power when combined in the same multivariate model, this analysis suggests that the CTFC may add additional prognostic information within TIMI grade 3 flow. The observation that the CTFC is an independent multivariate correlate of outcomes also extends observations from GUSTO I by Lee et al15 in which nonangiographic clinical variables such as age, heart rate, and anterior location of the MI were demonstrated to be multivariate determinants of mortality. When we normalized the CTFC for heart rate by estimating the fraction of a cardiac cycle required for the dye to reach the landmark, this index of flow was also related to clinical outcomes in multivariate models.
In GUSTO I, a 20 percentage point or a 64% relative improvement in the rate of TIMI grade 3 flow of rtPA over streptokinase (54% versus 33%) was associated with a 1% improvement in mortality.3 The corresponding analysis using the CTFC in these 3 TIMI trials indicates that a 20% relative improvement in the CTFC would alter mortality by a factor of 1.07 or by ≈0.3% (overall mortality was 4.25% in these trials), and a 64% improvement would therefore reduce mortality by another 1%. Thus, the mortality benefits estimated by the relative improvements in the TIMI flow grades and the TIMI frame count closely parallel each other.
This analysis assumes that there is a linear relationship between 90-minute coronary blood flow and mortality, and likewise, the results in this study are presented as the OR for every 10-frame rise in the CTFC. A linear spline analysis failed to reject the null hypothesis that there was a linear relationship between the CTFC and outcomes (ie, the null hypothesis that the ORs were the same across intervals of the CTFC was not rejected). Furthermore, neither a logarithmic transformation of the CTFC data nor the use of higher-order polynomials to fit the relationship to complex curves provided any better explanatory power in any of the outcome variables. The number of events in these trials, however, was rather small, and it is possible that a larger number of events might demonstrate that the relationship between flow and mortality improvements is instead nonlinear.
Many patients in thrombolytic trials now undergo angiography by 60 minutes, and the relative merits of waiting to obtain 90-minute angiographic data over 60-minute data are unclear. This study demonstrates that in a data set with both 60- and 90-minute data available, the conventional 90-minute end point was related to mortality, whereas the 60-minute data were not. Although this could be interpreted as demonstrating the superiority of 90-minute angiography as a surrogate end point, this interpretation must be weighed carefully in light of the observation that patients whose closed arteries failed to open spontaneously between 60 and 90 minutes had a significantly higher mortality than those patients whose arteries did open spontaneously (without intervention) between 60 and 90 minutes (0% versus 8.2%; P=0.05). Although these observations suggest that reperfusion between 60 and 90 minutes is related to improved outcomes, randomized trial data are lacking regarding the clinical benefit of percutaneous intervention for an occluded artery at 60 rather than 90 minutes after administration of a thrombolytic. If newer pharmacological reperfusion strategies can demonstrate superiority at 60 minutes rather than waiting until 90 minutes, this may obviate the need for 90-minute angiography. Indeed, data from the TIMI 14 trial suggest that there is a greater advantage in CTFCs at 60 minutes for the combination of abciximab plus reduced-dose tPA versus front-loaded tPA (14 frames, P=0.001) compared with 90 minutes (7 frames, P=0.005).16
Another issue that has arisen is whether use of the CTFC at 90 minutes is superior to use of the final posttreatment CTFC (ie, either the CTFC at 90 minutes after thrombolytic administration or, if the patient went on to have an intervention [rescue or adjunctive PTCA or stent], the CTFC at the completion of the intervention). For all outcomes, the 90-minute CTFC had higher z values and more significant probability values in relationship to clinical outcomes than the final posttreatment CTFC.
When prospectively applied in the TIMI 10 trials, the CTFC was evaluable in 97.1% of open arteries (741 of 763). The mortality rate in patients with CTFC data available (4.3%, 53 of 1248) was no different from that in the study group overall (4.3%, 60 of 1383). These mortality rates may be lower than those observed outside of randomized trials because of the selective nature of the inclusion/exclusion criteria. Rescue and adjunctive angioplasty may have obscured differences in outcomes that would have been attributable to 90-minute flow, but the same relationships were observed when the analysis was confined to those patients who did not undergo interventions. Although 90-minute coronary blood flow is related to mortality, other causes of death such as intracranial hemorrhage, reinfarction, ventricular arrhythmias, and mechanical complications may be unrelated to 90-minute blood flow.
Higher CTFCs at 90 minutes after thrombolytic administration are related to an increased risk of early and late adverse outcomes in both univariate and multivariate models. The CTFC may add additional prognostic information by segregating patients with TIMI grade 3 flow into lower- and higher-risk subgroups.
This study was supported in part by grants from Smith Kline Beecham, Philadelphia, Pa, Genentech, Inc, South San Francisco, Calif, and Boehringer Ingelheim, Rhein, Germany.
Reprint requests to C. Michael Gibson, MS, MD, Director of Invasive Cardiology, Allegheny General Hospital, 320 E North Ave, Pittsburgh, PA 15212.
- Received August 12, 1998.
- Revision received November 6, 1998.
- Accepted January 11, 1999.
- Copyright © 1999 by American Heart Association
The TIMI Study Group. The Thrombolysis In Myocardial Infarction (TIMI) trial. N Engl J Med. 1985;31:932–936.
Vogt A, Von Essen R, Tebbe U, Feuerer W, Appel KF, Neuhaus KL. Impact of early perfusion status of the infarct-related artery on short-term mortality after thrombolysis for acute myocardial infarction: retrospective analysis of four German multicenter studies. J Am Coll Cardiol. 1993;21:1391–1395.
Karagounis L, Sorensen SG, Menlove RI, Moreno F, Anderson JL. Does thrombolysis in myocardial infarction TIMI perfusion grade 2 represent a mostly patent artery or a mostly occluded artery? Enzymatic and electrocardiographic evidence from the TEAM-2 study. J Am Coll Cardiol. 1992;17:1–10.
Anderson JL, Karagounis LA, Becker LC, Sorensen SG, Menlove RL, for the TEAM-3 Investigators. TIMI perfusion grade 3 but not grade 2 results in improved outcome after thrombolysis for myocardial infarction: ventriculographic, enzymatic, and electrocardiographic evidence from the TEAM-3 study. Circulation. 1993;87:1829–1839.
Gibson CM, Cannon CP, Daley WL, Dodge JT, Alexander B, Marble SJ, McCabe CH, Raymond L, Fortin T, Poole WK, Braunwald E. The TIMI frame count: a quantitative method of assessing coronary artery flow. Circulation. 1996;93:879–888.
Cannon CP, McCabe CH, Diver DJ, Herson S, Greene RM, Shah PK, Sequeira RF, Leya F, Kirshenbaum JM, Magorien RD, Palmeri S, Davis V, Gibson CM, Poole WK, Braunwald E, for the TIMI 4 Investigators. Comparison of front-loaded recombinant tissue-type plasminogen activator, anistreplase and combination thrombolytic therapy for acute myocardial infarction: results of the Thrombolysis in Myocardial Infarction (TIMI) 4 trial. J Am Coll Cardiol. 1994;24:1602–1610.
Cannon CP, McCabe CH, Gibson CM, Ghali M, Sequeira RF, McKendall GR, Breed J, Modi NB, Fox NL, Tracy RP, Love TW, Braunwald E, and the TIMI 10A Investigators. TNK-tissue plasminogen activator in acute myocardial infarction: results of the Thrombolysis In Myocardial Infarction (TIMI) 10A dose-ranging trial. Circulation. 1997;95:351–356.
Cannon CP, Gibson CM, McCabe CH, Adgey AAJ, Schweiger MJ, Sequeira RF, Grollier G, Giugliano RP, Frey M, Mueller HS, Steingart RM, Fox NL, Weaver WD, Van de Werf F, Braunwald E, for the TIMI 10B Investigators. TNK-tissue plasminogen activator compared with front-loaded alteplase in acute myocardial infarction: results of the TIMI 10B Trial. Circulation.. 1998;98:2805–2814.
Stata Corp. Stata Statistical Software: Release 5.0. College Station, Tex: Stata Corp; 1997.
Gibson CM, Goel M, Dotani I, Rizzo MJ, McLean C, Martin NE, Al-Mousa EN, Marble SJ, Daley WL, Dodge T. The post-PTCA TIMI frame count and mortality in RESTORE. Circulation. 1996;94(suppl I):I-85. Abstract.
Lee KL, Woodlief LH, Topol EJ, Weaver DW, Betriu A, Col J, Simoons M, Aylward P, Van de Werf F, Maliff RM. Predictors of 30-day mortality in the era of reperfusion for acute myocardial infarction: results from an international trial of 41 021 patients. Circulation. 1995;91:1659–1668.
Gibson M, Giugliano RP, Anderson K, Scherer JC, McCabe CH, Antman EM. Abciximab enhances thrombolysis. a comparison of abciximab alone versus abciximab plus low dose thrombolytics using the corrected TIMI frame count. Circulation. 1998;98(suppl):I-559. Abstract.