In 1985, the Thrombolysis in Myocardial Infarction (TIMI) Study Group introduced a grading system, the TIMI Flow Classification, that categorically describes qualitative changes in coronary flow (Table 1).1 This classification, which is easy and quick to apply, is used widely not only in heart catheterization laboratories during clinical routine, but also as an endpoint in large multicenter studies (e.g., to express the 90 minute patency rates in patients treated with thrombolytics during acute myocardial infarction). In contrast to this categorical, subjective method, Gibson described a new quantitative method for assessing coronary artery flow, the TIMI Frame Count. It is a simple continuous angiographic index, which counts the number of cineframes required for contrast material to reach standard distal coronary landmarks.2 To analyze deviations between TIMI Flow Classification and TIMI Frame Count in assessing coronary flow we retrospectively examined a group of 102 patients before and after stent implantation and at the 6-month control angiography. Methods A retrospective analysis of the angiographic data was performed in 102 patients with coronary artery disease and an indication to percutaneous coronary intervention (77% men, 23% women) before stent implantation, after stent implantation and at 6-month control angiography (100% follow-up). Fifty-five percent of the patients had stable angina pectoris and 45% had an acute coronary syndrome. One stent was implanted in 64% of patients, while the other patients had 2 or more stents implanted; the maximum was one patient with 4 stents. The mean age of the patients was 59 years (minimum, 30 years; maximum, 80 years). The study population is described in detail in Table 2. We assessed the TIMI flow, TIMI frame count (TFC) and the corrected TIMI frame count (CTFC) for the coronary vessel with the culprit lesion before, immediately after stent implantation and 6 months post-intervention. Coronary angiography was performed by hand injection of contrast medium via 6 French catheters and filmed at 12.5 frames/s. The original TFC is described for a cinefilm speed of 30 frames/s. To make the results of TFC analysis comparable between different studies, an adaptation to the cinefilm speed of 30 frames/s was performed as previously described.3 Therefore, the frame counts were multiplied by 2.4 (correction factor, which was calculated by the division of the cinefilm speed of 30 frames/s through the actual cinefilm speed of 12.5 frames/s). The number of cineframes required for contrast medium to reach the first standardized distal coronary landmark in the artery with the culprit lesion was measured with the frame counter on the ARRIPRO 35 cineviewer. The analysis of the TIMI flow grade was performed following the criteria of the TIMI Study Group introduced in 1985 (Table 1). The TFCs were determined as published by Gibson 1996.2 The TFC is defined as the number of cineframes required for the contrast material to reach a standard distal coronary landmark. The first frame used for counting was the frame in which the contrast material completely entered the artery. Therefore, three conditions have to be fulfilled: 1) a column of nearly full or fully concentrated contrast material must extend across the entire width of the origin of the artery; 2) the dye must touch both borders of the origin of the artery; and 3) there must be antegrade motion of the contrast medium. If the left anterior descending coronary artery (LAD) is subselectively engaged and the left circumflex artery (LCX) is the culprit vessel, the count of TFC begins when dye first touches both borders of the LCX. The same rule holds when the LAD is the culprit vessel. That frame is defined as the last frame, where dye enters the sidebranch for the first time, which is defined as the distal anatomic landmark, instead of where the branch is completely filled with contrast medium. For the LAD, the distal bifurcation is defined as the landmark branch (“mustache”, “pitchfork”, “whale’s tale”). If the LAD goes around the apex of the heart, the branch nearest to the apex is taken. For the LCX, the frames are counted until dye enters the distal bifurcation with the longest total distance that includes the culprit lesion. For the RCA, the first branch of the posterolateral artery is defined as the distal anatomic landmark. Because of the larger vessel length of the LAD, the TFC for coronary arteries with normal flow is different, as proven by an analysis of patients with normal coronary artery flow and their mean TFCs: LAD, 36.2 frames; RCA, 20.4 frames; and LCX, 22.2 frames.2 To adapt those differences in vessel length and to make the results of TFC comparable for the different vessels, the corrected TFC (CTFC) was introduced. The CTFC is calculated by division of the actual measured frame counts of the LAD through a correction factor of 1.7, which was calculated by the division of the mean value of LAD frame count through the mean values for RCA and LCX in a normal collective.2 The values of the frame counts for the RCA and LCX remain uncorrected. When no CTFC could be determined because the contrast medium did not reach the distal landmark, 100 frames were imputed as the value for the CTFC, as previously described.4,5 Quantitative coronary analysis was performed after balloon angioplasty, after stent implantation and at 6-month control. Angiographic pictures were projected on an ARRIPRO 35-Projector (Arnold and Richter, Munich, Germany), digitalized with a Sparc II PC with Parallax Board (Parallax, Santa Clara, California) and transferred to the AWOS-System (Siemens, Erlangen, Germany). Minimum lumen diameter of stenosis was measured. The statistical analysis was performed either with SAS or SPSS. Graphical presentation of the data is based on box plots (SPSS) or chart diagrams done with EXCEL. Intraindividual changes in TFCs were assessed via sign rank tests (paired Wilcoxon tests). Due to the rather descriptive nature of our analyses, the resulting p-values were not formally adjusted for multiplicity. Therefore, the p-values should be regarded as descriptive measures of intraindividual trends and a p-value = 0.80, sufficient agreement for at least kappa >= 0.60. Results The median of minimum lumen diameter of the culprit lesion before stent implantation was 0.74 mm, which increased after stent implantation to 2.63 mm and decreased after 6 months to 1.92 mm. Focusing on the differences in minimum lumen diameter before and after stent implantation in a paired analysis, an increase in lumen diameter of 1.85 mm (1.54/2.20 = 25%/75% interquartile range; p = 0.0001 in paired Wilcoxon Test) could be found. The difference in minimum lumen diameter between preinterventional and 6-month values was 1.17 mm (0.66/1.59 = 25%/75% interquartile range; p = 0.0001 in paired Wilcoxon Test), an effective increase in lumen diameter which is still highly statistically significant. The analysis of TIMI flow grades (Figure 1) demonstrated that 20% of the patients had a totally occluded vessel preintervention. A TIMI flow grade 1 was seen in 5%, grade 2 in 13% and grade 3 in 63%. After stent implantation, no patient had TIMI grade 0 flow and 1 patient had TIMI grade 1 flow. Postintervention, 99% of the patients had TIMI grade 2 or 3 flow. At the 6-month control angiography, 4% of the patients had TIMI grade 0 or 1 flow, while 96% of the patients had TIMI grade 2 or 3 flow. The median of the CTFC values was 43.2 frames (26.4/100.0 = 25%/75% interquartile range) before stent implantation, 16.9 frames (14.4/31.1 = 25%/75% interquartile range) after stent implantation and 21.6 frames (16.8/32.5 = 25%/75% interquartile range) after 6 months. Increasing coronary flow after stent implantation becomes evident by the value of -19.2 frames for CTFC differences before and after stent implantation (-52.8/-8.47 = 25%/75% interquartile range; p = 0.0001 in paired Wilcoxon Test). After 6 months, a difference in CTFC compared to the preinterventional point of time of -16.8 frames (-48.0/-4.8 = 25%/75% interquartile range; p = 0.0001 in paired Wilcoxon Test) was determined. The preinterventional CTFC values showed a bimodal, not a normal distribution with a peak of values between 20 and 29 frames and a second peak over 100 frames (Figure 2). After stent implantation, an unimodal left-skewed distribution of CTFC values was observed. The maximum of CTFC values was reached earlier between 10 and 19 frames. After stent implantation, the peak of CTFC values ranged again between 10 and 19 frames, but the intracoronary flow was again reduced as expressed by higher frame counts. A comparison of CTFC values for patients with TIMI flow grade 3 and TIMI flow grade 2 showed an overlap between 30 and 69 frames (Figure 3). No patients were found with TIMI flow grade 2 under 30 frames or with TIMI flow grade 3 over 70 frames. We also investigated whether or not CTFC changes (> 5 frames) and TIMI flow changes were both able to detect an alteration of intracoronary flow (Tables 2 and 3). CTFC is able to find more flow changes than TIMI Flow Classification. Comparing preinterventional versus postinterventional or postinterventional versus 6 months results, CTFC detected intracoronary flow changes in 46% and 39% of the patients, which were not detected by TIMI Flow Classification. Corresponding results were obtained for rather sensitive cutpoints of > 10 and > 15 frames (data not shown). Discussion The TIMI Flow Classification was introduced in 1985 as a system for the description of coronary flow changes. This classification, which is easy and quick to apply, is used widely not only in heart catheterization laboratories during clinical routine, but also as an endpoint in large multicenter studies (e.g., to express the 90-minute patency rates in patients treated with thrombolytics during acute myocardial infarction). There are several limitations of this qualitative, categorical classification system. The TIMI flow grading system classifies perfusion as either grade 2 (partial perfusion) or grade 3 (complete perfusion).6 Vessels with TIMI grade 0 or 1 flow are considered to be completely or functionally occluded.7 Focusing on one vessel region, the complete occlusion of a coronary artery (TIMI 0) or the functional occlusion characterized by penetration of contrast medium through a stenosis without flow (TIMI 1) are relatively easy to classify: contrast medium fails completely or partially “to opacify the entire coronary bed distal to the obstruction for the duration of the cineangiographic filming sequence”.1 But the classification of TIMI grade 2 or 3 is more complex because it is based on the subjective comparison of perfusion (rate of contrast medium entry and clearance) in two vessel regions: the coronary artery bed distal to an obstruction and an uninvolved coronary bed proximal to the obstruction or the opposite coronary artery. The correct detection of incomplete reperfusion expressed as TIMI flow grade 2 is of great importance. Patients with acute myocardial infarction and TIMI flow grade 2 determined 90 minutes after thrombolysis, compared to those with TIMI flow grade 3, have a worse left ventricular function with an increased left ventricular volume.8 In a meta-analysis published in 1997, the relationship between TIMI flow grades 90 minutes after thrombolysis and mortality in 4,687 pooled patients was quantified. An increased mortality for patients with TIMI flow grade 2 compared with TIMI flow grade 3 was found (6.6% vs. 3.7%, respectively; p = 0.0003).9 The TIMI flow grading system classifies coronary flow as either normal or reduced. Two or more distinct subpopulations with either normal or reduced flow do not exist, so that such a categorical classification system is arbitrary.10 The classification into TIMI grade 2 or 3 is based on the subjective comparison of perfusion in the coronary artery bed distal to an obstruction with an uninvolved coronary bed proximal to the obstruction or the opposite coronary artery. This comparison is based on the assumption that the flow in the uninvolved coronary arteries is always normal. Especially 90 minutes after thrombolysis, when TIMI flow grade is used to classify the results of different thrombolytics, the intracoronary flow in the uninvolved coronary bed proximal to the obstruction is also reduced (21%).2 Comparing the analysis of TIMI flow grades between an angiographic core laboratory and clinical centers, the rate of agreement is the best in the assessment of TIMI flow grade 0/1 (kappa value = 0.84), moderate for TIMI flow grade 3 (kappa value = 0.55) and poor for TIMI flow 2 (kappa value = 0.38).2 In 1996, Gibson described a new quantitative method for assessing coronary artery flow, the TIMI frame count (TFC).2 It is a simple continuous angiographic index that counts the number of cineframes required for contrast material to reach standard distal coronary landmarks. In the meantime, the TFC is adapted for use in saphenous vein grafts by counting the cineframes required for dye to travel from the ostium of the graft to the graft anastomotic site and to the standardized distal coronary landmark.11 We retrospectively analyzed deviations between TIMI flow classification and TIMI frame count in assessing coronary flow in a group of 102 patients before, after stent implantation and at 6-month control angiography. Particularly in the area of intracoronary stenting, in 99% of the cases the postinterventional TIMI flow grade 3 or 2 is reached. However, the TIMI flow classification is not able to separate clearly between complete perfusion (TIMI 3) or partial perfusion (TIMI 2). As demonstrated in Figure 4, an overlap between the CTFC values for TIMI flow grade 3 and 2 is found between 30 and 69 frames with a smooth, steady transition from TIMI 3 to TIMI 2. Although the objective index of coronary flow, the CTFC, may be identical, in some cases the subjective perception of contrast medium entry and clearance rate may be assessed in one patient as TIMI grade 3 flow and in the other patient as TIMI grade 2 flow. When CTFC was lower than 30 frames, all patients were rated grade 3 in the TIMI Flow Classification. CTFC values under 30 frames represent in our examination a normal coronary flow. The same result was published by Gibson 1996, who described a CTFC value of 27 frames to be representative for the upper boundary of normal flow in the absence of acute myocardial infarction.2 We compared the frequency of agreement in the detection of coronary flow changes between TIMI flow classification and TIMI frame count. Before and after stent implantation, the rate of agreement is very poor (kappa value = 0.15 ± 0.06), as it is for the comparison of flow changes after stent implantation and after 6 months (kappa value = 0.22 ± 0.06). The TIMI frame count grading system is able to detect changes that were not detected by TIMI flow classification in an additional 46% and 39% of patients. It seems to be presumptive that in multicenter studies angiographic differences in the efficacy between treatment strategies (e.g., various thrombolytics) could be detected by tTIMI frame count where the TIMI flow classification fails to show relevant differences. Despite the superiority of TIMI frame count compared with TIMI flow classification, recent studies use TIMI flow classification to measure coronary perfusion12,13 and TIMI flow classification is discussed to be the “gold standard” for myocardial reperfusion assessment14 or to be the primary endpoint in thrombolytic trials.15 The CTFC is not only an index of coronary blood flow. It is also related to the risk of adverse outcomes in acute coronary syndromes and the risk of restenosis. In the TIMI 4 trial, a relationship between CTFC and adverse outcomes (death, recurrent myocardial infarction, left ventricular ejection fraction 20 to
1. The TIMI Study Group. The thrombolysis in myocardial infarction (TIMI) trial. N Engl J Med 1985;312:932‚Äì936. 2. Gibson CM, Cannon CP, Daley WL, et al. TIMI frame count. A quantitative method of assessing coronary artery flow. Circulation 1996;93:879‚Äì888. 3. Gibson CM, Murphy SA, Rizzo MJ, et. al. Relationship between TIMI frame count and clinical outcomes after thrombolytic administration. Thrombolysis in Myocardial Infarction (TIMI) Study Group. Circulation 1999;99:1945‚Äì1950. 4. Gibson CM, Goel M, Dotani I, et al. The post-PTCA TIMI frame count and mortality in RESTORE. Circulation 1996;94(Suppl I):I-86. 5. Gibson M, Cannon CP, Marble SJ, et al. The corrected TIMI Frame Count (CTFC) predicts clinical outcomes in acute MI. Circulation 1996;94(Suppl I):I-441. 6. Chesebro JH, Knatterud G, Roberts R, et al. Thrombolysis in Myocardial Infarction (TIMI) Trial, Phase I: A comparison between intravenous tissue plasminogen activator and intravenous streptokinase: Clinical findings through hospital discharge. Circulation 1987;76:142‚Äì154. 7. White HD, Braunwald E. Applying the open artery theory: Use of predictive survival markers. Eur Heart J 1998;19:1132‚Äì1139. 8. The GUSTO Angiographic Investigators. The effects of tissue plasminogen activator, streptokinase, or both on coronary-artery patency, ventricular function, and survival after acute myocardial infarction. N Engl J Med 1993;329:1615‚Äì1622. 9. Fath-Ordoubadi F, Huehns TY, Al-Mohammad A, et al. Significance of the thrombolysis in myocardial infarction scoring system in assessing infarct-related artery reperfusion and mortality rates after acute myocardial infarction. Am Heart J 1997;134:62‚Äì69. 10. Dotani I, Dodge JT, Goel M, et al. Techniques in the angiographic analysis of coronary flow: Past, present, and future. J Intervent Cardiol 1996;9:429‚Äì444. 11. Al-Mousa EN, Dodge JT, Rizzo M, et al. Thrombolysis in myocardial infarction frame count in saphenous vein grafts. Am Heart J 1998;135:323‚Äì328. 12. Andrews J, Straznicky IT, French JK, et al. ST-segment recovery adds to the assessment of TIMI 2 and 3 flow in predicting infarct wall motion after thrombolytic therapy. Circulation 2000;101:2138‚Äì2143. 13. Akasaka T, Yoshida K, Kawamoto T, et al. Relation of phasic coronary flow velocity characteristics with TIMI perfusion grade and myocardial recovery after primary percutaneous transluminal coronary angioplasty and rescue stenting. Circulation 2000;101:2361‚Äì2367. 14. Shah A, Wagner GS, Granger CB, et al. Prognostic implications of TIMI flow grade in the infarct related artery compared with continuous 12-lead ST-segment resolution analysis. Reexamining the ‚Äúgold standard‚Äù for myocardial reperfusion assessment. J Am Coll Cardiol 2000;35:666‚Äì672. 15. Gibson CM, Murphy SA, Marble SJ, et al. Can we replace the 90-minute thrombolysis in myocardial infarction (TIMI) flow grades with those at 60 minutes as a primary end point in thrombolytic trials? Am J Cardiol 2001;87:450‚Äì453. 16. Gibson CM, Cannon CP, Baim DS, et al. TIMI Frame Count: A new standardization of infarct-related flow grade and its relationship to clinical adverse outcomes in the TIMI 4 trial (Abstr). Circulation 1994;90(Suppl I):I-220. 17. Gibson M, Rizzo M, McLean C, et al. The TIMI frame count and restenosis: Faster is better. J Am Coll Cardiol 1997;29:201A.