Author + information
- Received March 18, 1998
- Revision received June 19, 1998
- Accepted July 6, 1998
- Published online November 1, 1998.
- Bruce R Brodie, MD, FACCa,c,*,
- Thomas D Stuckey, MD, FACCa,c,
- Thomas C Wall, MD, FACCa,c,
- Grace Kissling, PhD∗,c,
- Charles J Hansen, MAa,c,
- Denise B Muncy, BSNa,c,
- Richard A Weintraub, MD, FACCa,c and
- Thomas A Kelly, MD, FACCa,c
- ↵*Address for correspondence: Dr. Bruce R. Brodie, 520 North Elam Avenue, Greensboro, North Carolina 27403
Objectives. The purpose of this study was to evaluate the importance of time to reperfusion for outcomes after primary angioplasty for acute myocardial infarction.
Background. Survival benefit of thrombolytic therapy for acute myocardial infarction is strongly dependent on time to treatment. Recent observations suggest that time to treatment may be less important for survival with primary angioplasty.
Methods. Consecutive patients (n = 1,352) with acute myocardial infarction treated with primary angioplasty were followed for up to 13 years. Paired acute and follow-up ejection fraction data were obtained at cardiac catheterization in 606 patients.
Results. Reperfusion was achieved within 2 h in 164 patients (12%). Thirty-day mortality was lowest with early reperfusion (4.3% at <2 h vs. 9.2% at ≥2 h, p = 0.04) and was relatively independent of time to reperfusion after 2 h (9.0% at 2 to 4 h, 9.3% at 4 to 6 h, 9.5% at >6 h). Thirty-day–plus late cardiac mortality was also lowest with early reperfusion (9.1% at <2 h vs. 16.3% at ≥2 h, p = 0.02) and relatively independent at time to reperfusion after 2 h (16.4% at 2 to 4 h, 16.9% at 4 to 6 h, 15.6% at >6 h). Improvement in left ventricular ejection fraction was greatest in the early reperfusion group and relatively modest after 2 h (6.9% at <2 h vs. 3.1% at ≥2 h, p = 0.007).
Conclusions. Time to reperfusion, up to 2 h, is important for survival and recovery of left ventricular function. After 2 h, recovery of left ventricular function is modest and survival is relatively independent of time to reperfusion. These data suggest that factors other than myocardial salvage may be responsible for survival benefit in patients treated with primary angioplasty after 2 h.
Over the past decade the use of reperfusion therapy for acute myocardial infarction has dramatically reduced mortality (1–3). Thrombolytic therapy is thought to be beneficial when coronary reperfusion can be established early enough to salvage myocardium with consequent improvement in left ventricular function and better survival. Data from a number of randomized trials have shown that the mortality benefit of thrombolytic therapy is strongly dependent on the time from symptom onset until treatment (1,4–6). However, recent observations suggest that time to treatment may be less important for survival after primary angioplasty (percutaneous transluminal coronary angioplasty [PTCA]) than after thrombolytic therapy (7). The purpose of this study is to evaluate the importance of time to reperfusion for 30-day and late survival and recovery of left ventricular function after primary PTCA for acute myocardial infarction.
The study population consisted of 1,352 consecutive patients with acute myocardial infarction treated with primary PTCA without prior thrombolytic therapy by one cardiology group at our institution from 1984 through 1996. Patient selection criteria have been previously described (8). Patients were selected for intervention if they presented with chest pain of <12 h duration (<6 h before 1992). Patients were selected for intervention after 12 h (after 6 h before 1992) only if they had persistent ischemic chest pain or hemodynamic compromise. Patients were included in the study only if they had a diagnostic electrocardiogram with ST-segment elevation of ≥1 mm in ≥2 contiguous leads (or reciprocal ST-segment depression of ≥1 mm in leads V1and V2) or left bundle branch block. Primary PTCA has been the preferred reperfusion strategy at our institution since 1984. Of all patients with acute myocardial infarction seen at our institution with diagnostic electrocardiograms during the study period, approximately 70% received primary PTCA, 10% thrombolytic therapy (usually given at a referring hospital prior to transfer) and 20% received no reperfusion therapy usually because of late presentation, resolution of symptoms or comorbid disease (9).
The protocol for primary PTCA at our institution has been previously described (8). Patients were given 5,000 to 10,000 U of heparin intravenously and 325 mg chewable aspirin in the emergency department and transferred promptly to the catheterization laboratory. Reperfusion was established mechanically with primary PTCA without antecedent use of thrombolytic therapy. Adjunctive therapy with temporary transvenous pacing or intraaortic balloon pumping were performed at the discretion of the operator. Following the procedure, heparin was continued for 2 to 3 days, adjusted to prolong the activated partial thromboplastin time to 2 to 3 times the control value. Beta-adrenergic blocking agents and nitrates were used at the discretion of the operator and became standard treatment in the last 3 years of the study. Coronary stents were used in 103 patients during the last 2 years of the study. Treatment with the monoclonal antibody directed against the platelet IIb/IIIa glycoprotein receptor (abciximab) was used in 38 patients.
Clinical and angiographic follow-up
Clinical follow-up was obtained by hospital and office chart review and telephone contact. Follow-up catheterization and angiography were performed routinely during the first 3 years of the study and during participation in several clinical trials (the Primary Angioplasty Registry in 1990 to 1991, the Second Primary Angioplasty in Myocardial Infarction [PAMI-2] Trial in 1993 to 1994, the PAMI Stent Pilot Trial in 1995 to 1996 and the Stent PAMI Randomized Trial in 1997) (10–12). Otherwise, follow-up catheterization was performed for recurrent ischemic symptoms or after abnormal functional testing. Left ventricular ejection fractions were calculated from tracing contours of right anterior oblique cineangiograms using the area–length method with correction for the right anterior oblique projection (13).
Time to reperfusion was measured as the time from the onset of symptoms until coronary reperfusion was established with balloon inflation. If Thrombolysis In Myocardial Infarction (TIMI)-3 flow was present on the initial angiogram, the time of the initial angiogram was used as the time of reperfusion. For the purpose of this analysis, time to reperfusion was divided into four categories: <2 h, 2 to <4 h, 4 to <6 h and ≥6 h. Coronary flow in the infarct artery after primary PTCA was assessed visually by the operator and classified according to the TIMI grading system on a scale of 0 to 3 (14). Cardiogenic shock was defined as hypotension (systolic blood pressure <85 mm Hg) not responsive to volume expansion and associated with severe left ventricular dysfunction or right ventricular infarction.
Variables examined as predictors of 30-day survival and 30-day–plus late cardiac survival included age, gender, diabetes, prior myocardial infarction, prior coronary bypass surgery, cardiogenic shock before intervention, infarction location, 3-vessel coronary artery disease, acute left ventricular ejection fraction, TIMI flow post intervention and reperfusion time. Statistical comparisons of baseline, procedural and outcome variables between subgroups were performed using the chi-square statistic for categorical variables and Student’s unpaired ttest and analysis of variance for continuous variables. Multiple logistic regression was used to assess the relation between predictor variables and 30-day mortality. Differences in 30-day–plus late cardiac survival across categories of discrete predictor variables were examined with Kaplan–Meier survival curves and their associated log-rank test statistics. Multivariable analyses of predictors of 30-day–plus late cardiac survival were performed using Cox proportional hazards regression models. All analyses were performed with SAS (SAS Institute Inc., Cary, North Carolina) and SPSS (SPSS, Inc., Chicago, Illinois) statistical software.
Of 1,352 patients treated with primary PTCA for acute myocardial infarction, 164 patients (12%) were reperfused within 2 h, 581 patients (43%) within 2 to 4 h, 332 patients (25%) within 4 to 6 h and 275 patients (20%) after 6 h. The TIMI-3 flow was achieved in 1,248 patients (92.3%), TIMI-2 flow in 65 patients (4.8%) and TIMI-0-1 flow in 39 patients (2.9%). Thirty-day mortality was 8.6% (116 patients). Clinical follow-up was obtained in 1,222 of 1,238 30-day survivors (98.7%) at a mean follow-up time of 5.0 ± 3.4 years. There were 93 (6.9%) late cardiac deaths.
Baseline characteristics by time to reperfusion
Baseline variables for the four categories of time to reperfusion are shown in Table 1. The proportion of women and the incidence of diabetes were highest in the late reperfusion group (>6 h), whereas the incidence of prior myocardial infarction and anterior wall myocardial infarction was highest in the early reperfusion group (<2 h). Acute ejection fraction was higher in the early reperfusion group (<2 h) than in the three later reperfusion groups combined (≥2 h) (54.7 ± 13% vs. 52.0 ± 13%, p = 0.02).
Relationship between time to reperfusion and TIMI flow postintervention
The frequency of achieving TIMI-3 flow in the infarct artery after primary PTCA was high (90 to 93%) in each of the four categories of time to reperfusion and was similar regardless of the time to reperfusion (Table 2).
Importance of time to reperfusion for 30-day mortality and 30-day–plus late cardiac mortality
Thirty-day mortality was lowest in the early reperfusion group (4.3% at <2 h vs. 9.2% at ≥2 h, p = 0.04) and was relatively independent of time to reperfusion after 2 h (Table 2, Fig. 1). ⇓Late cardiac mortality (>30 days) was lowest in the early reperfusion group (<2 h), but this was not statistically different from the later reperfusion groups (Table 2). Thirty-day–plus late cardiac mortality was also lowest in the early reperfusion group (9.1% at <2 h vs. 15.3% at ≥2 h, p = 0.02) and was relatively independent of time to reperfusion after 2 h (Table 2, Fig. 1).
When the effect of late reperfusion (≥2 h) on 30-day mortality was adjusted for differences in baseline variables with multiple logistic regression, late reperfusion was a relatively weak correlate of 30-day mortality (odds ratio [OR] 1.54, 95% confidence interval [CI] 0.89 to 2.66, p = 0.12) (Fig. 2). When acute ejection fraction was excluded from the model (since some patients had missing ejection fraction data), late reperfusion (≥2 h) became a slightly stronger predictor of 30-day mortality (OR 1.59, 95% CI 1.00 to 2.51, p = 0.05). Cardiogenic shock and the inability to restore TIMI-3 flow were the strongest predictors of 30-day mortality (Fig. 2).
Kaplan–Meier survival curves of 30-day–plus late cardiac survival for each of the four categories of time to reperfusion are shown in Figure 3. Survival in the early reperfusion group (<2 h) was significantly better than the three later reperfusion groups combined (p = 0.03 by log rank test). Survival was very similar in the three later reperfusion groups with no significant differences between groups. Using a Cox proportional hazards regression model to adjust for differences in baseline variables between early and late reperfusion groups, late reperfusion (≥2 h) was a modest predictor of 30-day–plus late cardiac mortality (relative risk [RR] 1.72, 95% CI 0.96 to 3.11, p = 0.07) (Table 3). Cardiogenic shock, the inability to restore TIMI-3 flow, age >70 years and anterior myocardial infarction were the strongest predictors of 30-day–plus late cardiac mortality.
Importance of time to reperfusion for recovery of left ventricular function
Follow-up catheterization was performed in 751 surviving patients. Of these, 606 patients had acute and follow-up (paired) catheterization data that were adequate for measurement of left ventricular ejection fractions at a mean follow-up time of 8.0 ± 10.3 months. A comparison of baseline characteristics in patients with and without paired ejection fraction data is shown in Table 4. Older patients, patients with cardiogenic shock and patients with 3-vessel coronary artery disease were less likely to have follow-up angiography.
In patients with paired ejection fraction data, acute ejection fraction was similar in all categories of time to reperfusion (Table 2). Late ejection fraction was highest in the early reperfusion group (<2 h) and significantly higher than in the three later reperfusion groups combined (59.5 ± 14% vs. 55.3 ± 13.5%, p = 0.02). Improvement in ejection fraction was also greatest in the early reperfusion group and relatively modest after 2 h (6.9 ± 11% at <2 h vs. 3.1 ± 12% at ≥2 h, p = 0.007).
Previous studies of the importance of time to treatment with reperfusion therapy
The beneficial effect of reperfusion therapy for acute myocardial infarction is thought to be due to myocardial salvage, which results in improved left ventricular function and better survival (15,16). In concordance with this, thrombolytic therapy is thought to be most beneficial in the treatment of acute myocardial infarction when reperfusion is established early. Data from a number of randomized trials support this concept (1,4–6). The Fibrinolytic Therapy Trialists Collaborative Group summarized 9 randomized trials with over 58,000 patients and found a strong relationship between mortality benefit and time to treatment up to 12 h (4). Likewise, the Global Utilization of Streptokinase and tPA for Occluded Coronary Arteries (GUSTO) Trial showed that hospital mortality progressively increased with increasing time to treatment (5.3% at ≤2 h, 5.9% at 2 to 4 h, 8.5% at 4 to 6 h, and 8.9% at >6 h) (5). However, recent data from the PAMI-2 Trial suggest that the relationship between time to treatment and mortality may be different with primary PTCA (7). In 1,100 patients treated with primary PTCA, the PAMI group found that mortality was lowest when patients were treated very early (<2 h), but that mortality was relatively independent of time to treatment after 2 h (2.3% at <2 h, 3.7% at 2 to 4 h, 3.7% at 4 to 6 h and 3.2% at >6 h). (Seventy-three of the patients in the PAMI-2 Trial were enrolled at our institution and are included in both studies.)
Major findings of the present study
Our data show that very early reperfusion (<2 h) with primary PTCA was associated with a lower 30-day and late mortality compared with later reperfusion (≥2 h). The most interesting and possibly the most important finding of this study is that 30-day mortality and late mortality were relatively independent of time to reperfusion in patients with reperfusion times ≥2 h. This is in contrast to data from thrombolytic trials that show mortality continues to increase with increasing time to treatment up to 6 to 12 h (1,4,5).
Our data also show that early reperfusion (<2 h) was associated with substantial recovery of left ventricular function whereas later reperfusion (≥2 h) was associated with only modest recovery of left ventricular function. A partial exception to this is in patients with reperfusion times >6 h in which recovery of left ventricular function is somewhat better than expected. We believe this is related in part to our selection criteria. Up until 1992, patients were selected for intervention after 6 h only if they had persistent chest pain or hemodynamic compromise. We and others have shown previously that patients who present after 6 h with persistent chest pain have an increased incidence of collateral flow to the infarct region, and these patients can have substantial recovery of left ventricular function after reperfusion (17–20).
Reasons for differences in the results of the present study and prior thrombolytic trials
There are several possible reasons that might explain why mortality continues to increase with increasing time to treatment with thrombolytic therapy, yet seems to remain relatively constant after 2 h with primary PTCA. First, data from several trials suggest that successful reperfusion after thrombolytic therapy is achieved in a smaller proportion of patients with increasing time to treatment (21–23, and data from the GUSTO Database, which is on file with Genentech, South San Francisco, California). The TIMI-1 Trial found that infarct artery patency (TIMI-2 or -3 flow) was achieved less often with increasing time to treatment after the administration of intravenous streptokinase (45% at 2 to 4 h, 27% at 4 to 6 h and 17% at >6 h) (21).
Angiographic data from the GUSTO Trial indicate that TIMI-3 flow is achieved less often with tissue plasminogen activator (tPA) with increasing time to treatment, although these differences did not reach statistical significance (63% at <2 h, 54% at 2 to 4 h and 50% at 4 to 6 h; data from the GUSTO database). Steg et al. (22)found that patency and TIMI-3 flow rates decreased with increasing time to treatment with streptokinase but not with tPA. The RAPID-2 investigators found that TIMI-3 flow rates decreased with increasing time to treatment with both reteplase and accelerated tPA (62% at ≤6 h vs. 40% at >6 h for reteplace and 47% at ≤6 h vs. 35% at >6 h for tPA) (23). In contrast, in this study with primary PTCA, TIMI-3 flow was achieved in >90% of patients regardless of time to reperfusion. Because TIMI flow is a major determinant of hospital mortality with both thrombolytic therapy and primary PTCA, this may partially explain why mortality continues to rise with increasing time to treatment with thrombolytic therapy but remains relatively constant after 2 h with primary PTCA.
Second, mortality due to cardiac rupture appears to be higher in patients treated with thrombolytic therapy than in patients not treated, and mortality due to rupture increases with increasing time to treatment (24–26). If thrombolytic therapy is given early, the risk of death from myocardial rupture is reduced, but if thrombolytic therapy is given late, the risk of death from myocardial rupture is increased (24). In the Gruppo Italiano per lo Studio della Streptochinasi nell’Infarcto Miocardico Trial, the mortality rate due to cardiac rupture with thrombolytic therapy increased progressively with increasing time to treatment (0.7% at 0 to 3 h, 1.2% at 3 to 6 h, 1.3% at 6 to 9 h and 2.0% at 9 to 12 h) (26). In contrast, death due to myocardial rupture is very low with primary PTCA (27,28).
There also is an increasing incidence of hemorrhagic stroke (often fatal) with increasing time to treatment with thrombolytic therapy. The GUSTO investigators found by univariate analysis that the incidence of hemorrhagic stroke increased progressively with increasing time to treatment (0.5% at ≤2 h, 0.7% at 2 to 4 h, 0.8% at 4 to 6 h and 1.0% at >6 h) (5). The occurrence of hemorrhagic stroke with primary PTCA is rare (29).
The combined results of the PAMI, Zwolle and Mayo Clinic randomized trials have shown improved survival with PTCA compared to thrombolytic therapy in all time to treatment intervals, but patients presenting late showed the most survival benefit (30). This is consistent with the observation that mortality increases with increasing time to treatment with thrombolytic therapy (for the reasons given above), but is relatively constant after 2 h with primary angioplasty.
Our study may have profound implications regarding the mechanism of benefit of reperfusion therapy for acute myocardial infarction. The traditional paradigm regarding the benefit of reperfusion therapy for acute myocardial infarction is that early reperfusion results in myocardial salvage, which results in improved left ventricular function and better survival. This paradigm has been expanded to include benefit of late reperfusion when myocardial salvage is no longer expected. An open infarct artery, even if opened too late for myocardial salvage, may result in survival benefit by preventing ventricular dilatation, promoting electrical stability and providing a source of collateral flow should occlusion occur in another coronary artery (15,16). The survival benefit due to myocardial salvage is felt to be strongly time dependent and, based on the results of thrombolytic trials, has been thought to extend to 6 to 12 h. The survival benefit due to late reperfusion not related to myocardial salvage is felt to be relatively independent of time to reperfusion. Our data are consistent with this expanded paradigm, but our data suggest that the time period for myocardial salvage is relatively short. Recovery of left ventricular function is very modest after 2 h, and the predominant mechanism of benefit of reperfusion with primary PTCA after 2 h may be related to factors other than myocardial salvage.
Other studies support the concept that the time period for myocardial salvage is relatively short. Reimer et al. (31)first described the “wavefront” of myocardial necrosis in the canine model after coronary occlusion and found that after 2 h of occlusion little myocardium was viable (30). The Myocardial Infarction Triage and Intervention investigators (32)found that patients treated with thrombolytic therapy within 70 min of symptom onset had a smaller thallium infarct size (4.9% vs. 11.2%, p = 0.001), greater mean ejection fraction at 30 days (53% vs. 49%, p = 0.03) and reduced 30-day mortality (1.2% vs. 8.7%, p = 0.04) compared to patients treated at >70 min. The Mayo Clinic group performed paired sestamibi perfusion scans prior to reperfusion (with either thrombolytic therapy or primary PTCA) and at hospital discharge, and found substantial myocardial salvage when reperfusion was established within 2 h and only modest salvage with reperfusion after 2 h in the absence of residual blood flow (33).
A second clinical implication of this study relates to whether withholding thrombolytic therapy to transfer patients from a community hospital to an interventional facility to preferentially perform primary PTCA would result in an unacceptable amount of further myocardial necrosis (34). Our data suggest that unless infarct artery patency can be restored within 2 h, the time delay in transferring patients with acute myocardial infarction to tertiary centers for primary PTCA may not be prohibitive. If this is true, the proportion of patients eligible for transfer would be relatively high because only 12% of patients presented early enough to allow reperfusion within 2 h in our study. Two randomized trials have been designed to answer the questions when and which patients with acute myocardial infarction should be transferred to an interventional facility for primary PTCA. The PAMI group has an ongoing randomized trial (AIR-PAMI) to test the hypothesis that outcomes in high-risk acute myocardial infarction patients transferred for primary PTCA, when there is up to 2 h of transport delay, will be superior to thrombolytic treatment given at the local hospital (30). Likewise, Ziljstra and colleagues have a similar trial planned in The Netherlands. It is hoped the results of these trials will help to answer these important questions.
Finally, our data reinforce the importance of very earlytime to treatment (<2 h) with primary PTCA for acute myocardial infarction. Mortality in patients in whom patency of the infarct artery was restored within 2 h was less than one-half the mortality of patients treated after 2 h. Efforts to reduce patient delays in contacting the health care system, to provide rapid transport to hospitals and to establish protocols for rapid treatment in hospitals should continue with emphasis on increasing the proportion of patients treated at <2 h.
Our study has the following limitations.
1. The number of patients in the early reperfusion group is relatively small, which limits statistical power to detect differences in outcomes between early and later reperfusion groups.
2. Our study has follow-up angiography and paired ejection fraction data in a limited number of patients. There are some differences in baseline variables between patients with and without paired follow-up ejection fraction data, so it is possible that our ejection fraction data may not be representative of our entire patient population.
3. Assessment of the time of onset of symptoms of acute myocardial infarction is often difficult, and this could lead to errors in determining time to reperfusion. In patients with TIMI-3 flow on the initial angiogram, the time of the initial angiogram was used as a time of reperfusion. In these patients, the actual time of reperfusion may have occurred much earlier.
4. The TIMI flow in the infarct artery postintervention was assessed by the operator, and left ventricular ejection fraction measurements were performed on site rather than by core lab analysis.
5. Reinfarction or reocclusion after successful reperfusion could negate the effects of time to treatment on both mortality and recovery of left ventricular function and could affect our results. The fact that reinfarction and reocclusion were distributed similarly across all categories of time to reperfusion makes it unlikely that this affected our results.
6. Finally, our data represent an observational experience from one institution. Further multicenter studies will be needed to confirm our findings.
We thank Gregg W. Stone, MD, at the Cardiovascular Institute, Mountain View, California, for his thoughtful review and suggestions for this manuscript.
☆ This study was supported by a grant from the LeBauer Cardiovascular Research Foundation.
- global utilization of streptokinase and tPA for occluded coronary arteries
- primary angioplasty in myocardial infarction (trial)
- percutaneous transluminal coronary angioplasty
- Thrombolysis in Myocardial Infarction (trial)
- tissue plasminogen activator
- Received March 18, 1998.
- Revision received June 19, 1998.
- Accepted July 6, 1998.
- American College of Cardiology
- Fibrinolytic Therapy Trialists’ (FTT) Collaborative Group
- Newby L.K,
- Rutsh W.R,
- Califf R.M,
- et al.
- Rawls J.M
- Stone G.W,
- Brodie B.R,
- Griffin J,
- et al.
- Brodie B.R
- Brodie B.R,
- Grines C.L,
- Ivanhoe R,
- et al.
- Stone G.W,
- Marsalese D,
- Brodie B.R,
- et al.
- Stone G.W,
- Brodie B.R,
- Griffin J.J,
- et al.
- Braunwald E
- Califf R.M,
- Topol E.J,
- Gersh B.J
- Rogers W.J,
- Hood W.P,
- Mantle J.A,
- et al.
- Cheseboro J.H,
- Knatterud G,
- Roberts R,
- et al.
- Steg P.G,
- Laperche T,
- Goldmard J.L,
- et al.
- Bode C,
- Smalling R.W,
- Berg G,
- et al.
- Becker R.C,
- Gore J.M,
- Lambrew C,
- et al.
- Honan M.B,
- Harrell F.E,
- Reimer K.A,
- et al.
- Grines C.L
- Reimer K.A,
- Lowe J.E,
- Rasmussen M.M,
- Jennings R.B
- Milavetz J.J,
- Giebel D.W,
- Christian T.S,
- Schwartz R.S,
- Holmes D.R,
- Gibbons R.J
- Brodie B.R