Author + information
- Received June 22, 2001
- Revision received October 31, 2001
- Accepted November 13, 2001
- Published online February 20, 2002.
- Gregg W Stone, MD, FACC*,* (, )
- Michael A Peterson, MD†,
- Alexandra J Lansky, MD, FACC*,
- George Dangas, MD, FACC*,
- Roxana Mehran, MD, FACC* and
- Martin B Leon, MD, FACC*
- ↵*Reprint requests and correspondence:
Dr. Gregg W. Stone, The Cardiovascular Research Foundation, 55 E. 59th St., 6th Floor, New York, NY 10022, USA.
Objectives We sought to evaluate and validate the ability of the angiographic myocardial blush grade to risk stratify patients after successful angioplasty in acute myocardial infarction (AMI).
Background Although epicardial Thrombolysis In Myocardial Infarction (TIMI)-3 flow is restored in >90% of patients undergoing primary percutaneous coronary intervention (PCI), normal myocardial perfusion may be present less frequently and may detrimentally impact survival.
Methods A cohort of 173 consecutive patients undergoing intervention within 24 h of AMI onset were studied. High-risk features of this population included failed thrombolysis in 39%, cardiogenic shock in 17% and saphenous vein graft culprit in 11% of patients.
Results Despite the restoration of TIMI-3 flow in 163 (94.2%) patients, myocardial perfusion, as evidenced by normal contrast opacification of the myocardial bed subtended by the infarct artery (myocardial blush), was normal in only 29.4% of patients with TIMI-3 flow following PCI, and in no patient with TIMI 0 to 2 flow. In patients in whom TIMI-3 flow was restored, survival was strongly dependent on the myocardial perfusion grade; one-year cumulative mortality was 6.8% with normal myocardial blush, 13.2% with reduced myocardial blush and 18.3% in patients with absent myocardial blush (p = 0.004).
Conclusions Abnormal myocardial perfusion is present in most patients following primary or rescue PCI in AMI, despite restoration of brisk epicardial coronary flow. In high risk patients achieving TIMI-3 flow after intervention, the myocardial blush score may be used to stratify prognosis into excellent, intermediate and poor survival. Further study is warranted to examine whether adjunctive mechanical or pharmacologic strategies can further improve myocardial perfusion and survival of patients with acute myocardial infarction undergoing intervention.
Brisk epicardial coronary (Thrombolysis In Myocardial Infarction [TIMI]-3) flow is restored in >90% of patients with acute myocardial infarction (AMI) undergoing primary angioplasty at skilled centers (1,2). Despite this high rate of patency, myocardial recovery is often suboptimal, and mortality—especially in high-risk patients, such as those with anterior myocardial infarction and cardiogenic shock—is still considerable (3–5). Recent studies with myocardial contrast echocardiography (6), positron emission tomography (7)and sophisticated nuclear imaging techniques (8)have clearly demonstrated that myocardial perfusion and metabolism is often abnormal, despite restoration of TIMI-3 flow. This has prompted the search for a simple, reproducible means to assess myocardial microcirculation.
In 1998, van’t Hof et al. (9), from Zwolle, the Netherlands, described the concept of myocardial blush, the contrast opacification of the myocardial bed subtended by the infarct artery, and found that the degree of myocardial perfusion so assessed correlated with survival after primary percutaneous coronary intervention (PCI) (9). An advantage of this method is the potential for simple “on-line” determination of myocardial perfusion immediately after angioplasty in the catheterization laboratory, while still in the therapeutic window during which measures may be taken to improve a suboptimal result. However, the Zwolle method has not been validated at other centers, and the predictors of normal myocardial blush and the impact of normalized myocardial perfusion in a nonselected high-risk referral population with AMI after successful PCI remains incompletely characterized. We therefore examined the acute and late impact of myocardial blush grade in a high risk AMI cohort undergoing mechanical reperfusion therapy at a tertiary referral center.
Consecutive patients at the Washington Cardiology Center undergoing percutaneous intervention within 24 h of symptom onset of AMI with ST segment elevation or left bundle branch block from January 1995 through April 1999 were identified. All patients were considered, regardless of age, hemodynamic status, infarct vessel and whether thrombolytic therapy had been given prior to the intervention. Of 191 such patients, angiographic films were available and technically adequate for quantitative coronary analysis in 173 (91%). All patients provided informed written consent for longitudinal data collection and reporting.
Data collection and methodology
In-hospital clinical data were collected prospectively on site by dedicated research nurses. Follow-up was conducted using physician office visits and study nurse interview by a separate group of research associates. All adverse events were adjudicated by an independent adjudication committee at the Cardiovascular Research Foundation after review of original source documentation. Qualitative and quantitative angiographic core laboratory analyses were performed as previously described by technicians blinded to clinical outcomes (10). Antegrade blood flow in the infarct vessel was evaluated using the TIMI scale (11). All films were over-read for accuracy by a physician (A. J. L.).
Myocardial blush was graded densitometrically based on visual assessment of relative contrast opacification of the myocardial territory subtended by the infarct vessel in relation to epicardial density as described by van’t Hof et al. (9). Blush score 0 = absence of contrast opacification in the myocardial infarct zone or persistent stain without washout; blush 1 = minimal contrast opacification; blush 2 = reduced but clearly evident blush in the infarct zone compared to the ipsilateral or contralateral noninvolved epicardial vessel(s); and blush 3 = myocardial contrast filling equal to or greater than that seen in the noninvolved epicardial vessel(s). From multiple orthogonal projections, the single view was chosen that best isolated the myocardial infarct zone in question, most commonly the right anterior oblique (RAO) projection with cranial angulation for the left anterior descending artery (LAD) distribution, the RAO caudal or left lateral projection for the left circumflex artery (LCX), and either the left anterior oblique or RAO projections for the right coronary artery (RCA) distribution. For this analysis, blush grades were scored by two physicians not involved in the TIMI flow analysis (M. A. P. and G. W. S.). Inter- and intraobserver variability were determined from a random sample of 40 films scored by both reviewers (Table 1).
Categorical variables were compared by chi-square analysis or Fisher’s exact test. Continuous variables are presented as mean ± standard deviation and were compared by Student ttest, Mann-Whitney Utest, or ANOVA, as appropriate. All p values are two-tailed. Follow-up clinical events were analyzed with actuarial methods and Kaplan-Meier curves were constructed. The influence of baseline demographic and angiographic variables on mortality during the follow-up period was evaluated with the log rank test. Cox proportional hazard regression was used to determine the independent predictors of late adverse events. For all analyses, significance was determined at the p = 0.05 level.
Patient population and baseline characteristics
The study cohort at this tertiary referral center consisted of a high-risk group of patients (Table 2), with 29 (16.8%) presenting in cardiogenic shock, 68 patients (39.3%) undergoing rescue angioplasty for failed thrombolysis and 19 patients (11.0%) having an occluded saphenous vein graft as the infarct-related artery. An intra-aortic balloon pump was required in 76 (43.9%) patients, including placement in 11.0% of patients prior to arrival in the catheterization laboratory for sustained hypotension.
Procedural results and late outcomes
Balloon angioplasty was performed in all patients; stents were implanted in 103 patients (59.5%), and thrombectomy or thromboablative devices (atherectomy, laser) were used in 20 patients (11.6%). Glycoprotein IIb/IIIa receptor antagonists were administered in 32 (18.5%) of patients, mostly for suboptimal results. A mean of 1.5 ± 0.8 lesions were treated per patient. The final minimal luminal diameter was 2.55 ± 0.69 mm, and the residual stenosis was 12 ± 32%. By core laboratory angiographic analysis, TIMI-3 flow was restored in 163 patients (94.2%).
In-hospital adverse events included death in 14 patients (8.1%), reinfarction in 1 patient (0.6%) and target vessel revascularization in 12 patients (6.9%). The mean follow-up duration was 1.5 ± 1.4 years. By actuarial analysis, 13.7% of patients had died at one year, and 4.0% had developed reinfarction. By Cox proportional hazards regression, independent determinants of mortality were advanced age, cardiogenic shock on admission, LAD intervention and inability to restore TIMI-3 flow (Table 3).
Myocardial perfusion status, infarct size and effect on outcomes
As seen in Figure 1, normal myocardial perfusion prior to intervention as evidenced by a blush score of 3 was significantly less likely to be present than was normal epicardial TIMI-3 flow (9.3% vs. 35.2%, p < 0.0001); normal blush was present in 0% of patients with TIMI 0-2 flow at baseline, versus 20.5% of patients with initial TIMI-3 flow (p < 0.0001). Similarly, normal blush after angioplasty was present in only 48 patients (27.9%), despite the fact that TIMI-3 flow was restored in 163 patients (94.2%) (p < 0.0001, Fig. 1). Normal blush was not present in any patient with TIMI 0-2 flow after intervention, compared with 29.4% of patients with TIMI-3 flow after angioplasty (p = 0.04).
In addition to TIMI flow, the only other significant baseline correlate of restoration of normal myocardial blush was the coronary vessel undergoing intervention (Table 4). A blush score of 3 was achieved in 15.0% of patients undergoing intervention in the LAD myocardial territory, in 25.8% of patients undergoing intervention in the LCX distribution and in 45.2% of patients undergoing intervention in the RCA distribution (p = 0.0003). In this series, procedural stent implantation was also positively associated with achievement of normal blush, whereas there was no relationship with procedural glycoprotein IIb/IIIa inhibitor use (Table 4).
The degree of angiographic myocardial perfusion after intervention correlated with infarct size as estimated by peak creatine phosphokinase-MB isozyme (CPK-MB) in all patients, as well as in the subset of 163 patients in whom TIMI-3 flow was restored (Fig. 2). Abnormal postprocedure myocardial blush also correlated with in-hospital and cumulative mortality, though not as strongly in this regard as did reduced postintervention epicardial TIMI flow (Fig. 3). However, in patients in whom TIMI-3 flow was restored, the degree of restored myocardial perfusion strongly correlated with early and late survival (Fig. 4); this relationship held in both patients with anterior and nonanterior infarction (Fig. 5, left), and was especially strong in patients with cardiogenic shock (Fig. 5, right). There were no differences in cumulative rates of reinfarction as a function of myocardial perfusion in patients with TIMI-3 flow; reinfarction occurred in 4.2% of patients with final blush score 3, 2.8% of patients with blush score 2 and 2.3% of patients with blush score 0/1 (p = 0.87).
The present study demonstrates that despite restoration of brisk antegrade epicardial blood flow by mechanical reperfusion therapy in a broad cross section of patients with evolving AMI, angiographic evidence of normal myocardial perfusion is achieved in only a minority of patients. Furthermore, the simple assessment of the degree of myocardial contrast opacification immediately after percutaneous intervention allows prognostic stratification of patients with TIMI-3 flow into three well circumscribed strata of survival, with an approximate threefold difference in mortality.
Correlates and impact of normal myocardial blush
The strongest predictors of not achieving normal myocardial blush were the inability to achieve TIMI-3 flow and angioplasty in the LAD territory. Although it is intuitive that myocardial perfusion is likely to be suboptimal if epicardial flow is depressed, the reason that LAD intervention is less likely to result in normal myocardial perfusion is less obvious and may relate to increased thrombus burden and atheroembolism, more active vasohumoral responses or greater myocardial edema and necrosis in LAD infarction (12–14). Regardless of the cause, it is interesting to speculate whether the greater mortality of patients with anterior myocardial infarction compared with nonanterior myocardial infarction is due not only to the increased muscle mass at risk (15), but also to a lower likelihood of achieving normal myocardial perfusion despite restored epicardial flow.
In addition to validating the reproducibility and utility of the Zwolle method, these data complement and extend the prior findings of van’t Hof et al. (9)in several important ways. The current patient population, in whom both primary and rescue PCI was performed for failed thrombolysis and who were characterized by a high prevalence of cardiogenic shock, was at significantly higher risk of mortality (14% 1-year mortality) and less selected than the pure primary PCI cohort in the prior study (9). Furthermore, as seen in Figure 3, absent epicardial flow (TIMI 0/1) was found to be the strongest angiographic predictor of mortality after PCI in AMI, whereas the achievement of normal myocardial perfusion as assessed by blush was a more powerful predictor of survival than attaining TIMI-3 flow. Moreover, in patients achieving normal epicardial perfusion (TIMI-3 flow), the degree of myocardial perfusion restored was strongly associated with infarct size and subsequent early and late survival. The observation that normal myocardial blush was achieved in only 28% of patients in this series (despite 94% TIMI-3 flow) and in only 19% of patients in the Zwolle population (9), is strong evidence that most patients are not achieving an optimal result after mechanical reperfusion therapy. Further investigation is required to determine whether adjunctive therapies can be identified, which, if applied before or after angioplasty in evolving AMI, are able to facilitate restoration of normal myocardial perfusion in a greater percentage of patients, and whether such interventions translate into improved myocardial salvage and enhanced survival. In this regard, both glycoprotein IIb/IIIa inhibitors (16)and devices to protect the distal microcirculation from atheroembolic and thrombotic debris (14)have shown early promise.
Of note, the beneficial effect on survival of restoring normal myocardial perfusion was most evident in patients with the most myocardium at risk, that is, those with cardiogenic shock and LAD infarction. This finding suggests that efforts to restore normal myocardial perfusion may be most critical in patients with extensive infarction, though advantageous trends were also seen for patients with smaller amounts of myocardium at risk, those without shock and with non-LAD occlusion.
Myocardial blush methodologies
The myocardial blush scale used in the present study was adapted from the original description of van’t Hof et al. (9)and relies on simple visual assessment of contrast opacification in the affected myocardial bed, relative to nonaffected myocardial territories. This method has now been validated in both labs to have high intra- and interobserver reproducibility and has also been demonstrated to be of prognostic import in a broad variety of patients undergoing angioplasty in AMI, including primary and rescue PCI and cardiogenic shock. In contrast, Gibson et al. (17)have proposed an alternative, automated, computer-enhanced videodensitometric technique of measuring myocardial blush, which relies heavily on washout rates, as well as opacification. The prognostic utility of this instrument has been examined in an AMI population receiving thrombolytic therapy without intervention in the TIMI 10B study (17), but not after primary angioplasty, and the intra- and interobserver variability of this method has not yet been described. Whether the two methods are comparable and equally predictive is presently unknown.
Conceivably, all variables affecting myocardial blush may not have been collected. For example, ascertainment of baseline left ventricular ejection fraction was incomplete, and thus the contribution of left ventricular function to achievement of myocardial perfusion and prognosis is inconclusive. Time delays to intervention are also an important determinant of myocardial recovery and survival (18)and theoretically may affect attainment of normal myocardial perfusion. Nonsignificant trends were present relating delayed time to presentation and treatment with decreased blush, which may have reached significance in a larger population. Despite these limitations, confidence in the analysis presented regarding the prognostic importance of blush may be derived from the fact that all data were prospectively collected by independent research nurses, and adverse events were adjudicated by separate committee. Even the degree of myocardial blush was determined by separate physicians than those who graded TIMI flow, with neither group having knowledge of clinical outcomes. Second, the greater rates of normal myocardial blush achieved in patients receiving stents in this study may seem counterintuitive, given the observation in recent studies that stent implantation in AMI may degrade TIMI-3 flow (19,20). However, the use of stents (and IIb/IIIa inhibitors) were not prespecified nor controlled in this series, and data are not available to determine whether they were used in a primary fashion or for bail-out after a suboptimal result. Thus, prospective randomized studies are required to determine the impact of these and other modalities on myocardial perfusion. Third, higher grades of myocardial blush have been found to correlate with ST-segment resolution (21,22), which was not evaluated in the present study; whether this indice or other surrogates of ischemia are complementary or additive to the assessment of myocardial blush is unknown. Finally, although myocardial blush offers the potential for simple “real-time” assessment and therapeutic response, no studies have been done evaluating the prospective accuracy and predictive value of blush determined in the catheterization laboratory, as compared with in a core lab. Further investigation is also required to determine whether therapeutic measures during intervention are able to improve suboptimal myocardial blush and subsequently enhance prognosis.
☆ Supported in part by an unrestricted research grant by PercuSurge, Inc., Sunnyvale, California.
- acute myocardial infarction
- coronary artery prior bypass grafting
- creatine phosphokinase-MB isozyme
- left anterior descending artery
- left circumflex artery
- percutaneous coronary intervention
- right anterior oblique
- right coronary artery
- Thrombolysis In Myocardial Infarction
- Received June 22, 2001.
- Revision received October 31, 2001.
- Accepted November 13, 2001.
- American College of Cardiology
- Stone G.W.,
- Brodie B.R.,
- Griffin J.J.,
- et al.
- Brodie B.R.,
- Stuckey T.D.,
- Wall T.C.,
- et al.
- Ito H.,
- Tomooka T.,
- Sakai N.,
- et al.
- Maes A.,
- Van de Werf F.,
- Nuyts J.,
- et al.
- Kondo M.,
- Nakano A.,
- Saito D.,
- et al.
- van’t Hof A.W.J.,
- Liem A.,
- Suryapranata H.,
- et al.
- Lansky A.,
- Popma J.
- The TIMI Study Group
- Ito H.,
- Maruyama A.,
- Iwakura K.,
- et al.
- Stone P.F.,
- Raabe D.S.,
- Jaffe A.S.
- Neumann F.-J.,
- Blasini R.,
- Schmitt C.,
- et al.
- Gibson C.M.,
- Cannon C.P.,
- Murphy S.A.,
- et al.
- Schröder R.,
- Wegscheider K.,
- Schröder K.,
- et al.