Coronary restenosis is the most important clinical limitation after percutaneous coronary intervention (PCI), and coronary flow reserve (CFR) is reduced in the presence of significant coronary stenosis. This study evaluated whether detection of early reduction of Doppler echocardiographically derived CFR in the left anterior descending coronary artery can identify patients at high risk for developing restenosis after successful PCI.
Doppler echocardiographically derived CFR was studied in 124 consecutive patients at 1-month and 6-month follow-up after PCI in the left anterior descending coronary artery, together with coronary angiography.
Restenosis was detected in 39 angiographic examinations (group A) and no coronary restenosis in the remaining 85 (group B) at 6 months. At 1 month, CFR was reduced in group A compared with group B ( P < .0001), and a significant reduction of CFR in group A ( P < .0001) but not in group B ( P = .89) was detected at 6 months. CFR ≤ 2.5 at 1 month was 67% sensitive and 87% specific for predicting significant restenosis, with positive and negative predictive values of 67% and 87%, respectively.
CFR ≤ 2.5 detected 1 month after PCI in the left anterior descending coronary artery has the potential to identify patients at higher risk for developing coronary restenosis and indicates the need for close clinical follow-up.
Although the use of stents has reduced the angiographic incidence of restenosis after percutaneous coronary intervention (PCI), it remains one of the major clinical problems, occurring randomly and often unexpectedly a few months after PCI. In addition, although the recent introduction of drug-eluting stents has reduced the post-PCI restenosis rate, the need for longer monitoring for late restenosis or thrombosis has been suggested, especially in patients with “off-label” indications. Coronary flow reserve (CFR) and its derivative by means of Doppler flow wire have been used in the catheterization laboratory to evaluate post-PCI results, showing only fairly good prediction of the restenosis rate at follow-up, probably because CFR was attained immediately after the procedure, at a moment when it may be temporarily impaired because of microvascular coronary dysfunction. The ability to identify patients who may need closer follow-up for restenosis may be particularly useful in an era of regional referral centers that perform coronary interventions. Moreover, because the issues of early recoil and microvascular dysfunction appear to be linked to increased risk for restenosis, 1-month evaluation may be a reasonable approach if done using a technique that is economically feasible. A method capable of monitoring CFR noninvasively and repeatedly after a sufficient time lapse from PCI would clearly be advisable to clarify the role of CFR in predicting the restenosis process in the long term. Doppler transthoracic echocardiography (TTE) allows the serial assessment of CFR, has been shown to follow intracoronary flow–derived data closely, and has been shown to be useful in detecting coronary restenosis after elective PCI in coronary artery disease as well as in monitoring coronary microcirculatory dysfunction.
We therefore sought to evaluate whether “early” (1-month) determination of noninvasive CFR by TTE is capable of predicting which patients are at higher risk for developing significant coronary restenosis after elective PCI during clinical follow-up.
We studied 124 consecutive patients (92 men; mean age, 63 ± 9 years; age range, 38–80 years) with coronary artery disease, submitted to elective PCI on the left anterior descending coronary artery (LAD). They were enrolled in a multicenter study, from June 2001 to February 2005, conducted by the Clinical Cardiology Unit at the University of Cagliari ( n = 64) and the Cardiology Unit at the University of Genoa ( n = 60). Patients enrolled in the study agreed to undergo follow-up angiography at approximately 6 months irrespective of symptoms.
Exclusion criteria included previous myocardial infarction in the LAD territory, total or functional coronary occlusion, grade II or III atrioventricular block, severe chronic obstructive pulmonary disease, and bronchospasm. Patients were not selected on the basis of good transthoracic Doppler signals; no patient was excluded for inadequate echocardiographic image quality. An institutional review committee approved the study, and all patients gave informed consent.
All study patients underwent follow-up angiographic studies. Of the 124 follow-up studies, 20 were performed earlier than the initially scheduled time, because of new symptoms of typical angina ( n = 14) and/or evidence of ischemia on routine noninvasive cardiac imaging tests (exercise stress test or dobutamine stress echocardiography; n = 6). The remaining 104 angiographic studies were performed at the scheduled time (6 ± 1 months after percutaneous transluminal coronary angioplasty). All patients had a median follow-up period of 6 months (range, 2–12 months). Over this time span, they underwent clinical evaluation and CFR assessment twice: early, approximately 1 month after the procedure (median, 1 month; range, 0.5–4 months), and late, about 6 months after the procedure (median, 6 months; range, 3–10 months). Coronary angiography was performed at the end of the follow-up period after elective PCI, 1 day after the late CFR assessment. The late assessment of CFR and coronary angiography were performed before the scheduled time, when new symptoms of typical angina and/or evidence of ischemia was found on routine noninvasive cardiac imaging tests (exercise stress test or dobutamine stress echocardiography). During follow-up, a new PCI procedure, at the previous PCI site, was performed if clinically indicated. All patients received bare-metal stents in the LAD, and a total of 148 stents (in 124 PCI procedures) were used. In particular, 77 patients had one stent delivered, 34 had two stents, and one had three stents, and balloon angioplasty alone was performed in 12 patients.
Echocardiography was performed for CFR evaluation with TTE, as previously described. Briefly, CFR was measured in the distal LAD, resulting in a modified foreshortened two-chamber view, with Doppler recordings of the basal flow, and during adenosine infusion at a rate of 0.14 mg/kg/min for 3 to 5 min. Cardiac medications were not interrupted before adenosine, although all medications or methylxanthine-containing substances were withheld 48 hours before the study. Beverages containing methylxanthine substances (cola, tea, coffee, etc.) were restricted for 24 hours before the study. CFR in the LAD was calculated as the ratio of hyperemic to basal diastolic flow velocity (for each parameter, the highest of three cycles was averaged) by one experienced echocardiographer performing the test blinded to the angiographic and clinical data. Adenosine stress echocardiography was performed only to assess CFR; stress-induced left ventricular wall motion abnormalities were evaluated in different sessions with standard dobutamine stress echocardiography in each patient at follow-up of 1 and 6 months.
Coronary angiography was performed according to the standard Judkins method with the femoral or radial artery approach, as indicated. Coronary stenosis was assessed from orthogonal angiographic projections by one investigator, unaware of the CFR results. Coronary restenosis in the LAD was defined as >50% luminal diameter narrowing at the previous PCI site on follow-up angiography assessed by quantitative coronary angiography. In addition, in patients with significant restenosis, we also used the classification of Mehran et al. to characterize in-stent restenosis: pattern I includes focal lesions (≤10 mm in length; 10 patients), pattern II includes in-stent restenosis (>10 mm; 12 patients), pattern III involves proliferative disease outside the stent restenosis (15 patients), and pattern IV is total occlusion restenosis (two patients). These data were matched with noninvasive CFR detected at 1-month and 6-month follow-up.
Continuous data distribution was assessed using the Kolmogorov-Smirnov test. Continuous variables with no skew are presented as mean ± SD and skewed variables as medians and interquartile ranges. Student’s t tests, analysis of variance, or Mann-Whitney tests (two sided) were used to compare continuous variables between groups. Differences between frequencies were assessed using χ 2 tests. Sensitivity, specificity, and positive and negative predictive values were determined according to standard definitions. Both the relationship between CFR and quantitative coronary angiography and between CFR and the Mehran classification were evaluated using Spearman’s nonparametric test and expressed as ρ coefficients. Angiographic evidence of restenosis was considered the positive reference standard. Receiver operating characteristic curve analysis was performed to test the predictive discrimination of patients with or without restenosis. A manual Cox regression model with backward elimination was performed on blocks of variables until a regression model with only significant or marginally significant ( P < .10) variables was obtained; the independent predictive values of selected covariates were then evaluated. P values < .05 were considered significant. Data were analyzed using SPSS version 18.0 (SPSS, Inc., Chicago, IL).
All 124 patients who were initially enrolled in the study underwent angiographic studies at follow-up. Of the 124 follow-up studies, 20 were performed earlier, before the initially scheduled time, because of new symptoms of typical angina ( n = 14) and/or evidence of ischemia on routine noninvasive cardiac imaging tests ( n = 6); 12 of these 20 patients showed coronary restenosis caused by severe in-stent endothelial proliferation. The remaining 104 angiographic studies were performed at the established end of follow-up (6 ± 1 months after PCI). At follow-up, coronary angiography revealed significant coronary restenosis (luminal narrowing > 50%) in 39 angiographic examinations (31.5%; group A) and none in the remaining 85 (68.5%; group B).
Among the clinical and demographic findings, angina become significantly more frequent in the restenosis group, with early CFR ≤ 2.5, in late follow-up (54% vs 18%, P < .0001). The other variables, such as age, diabetes, hypertension, hypercholesterolemia, smoking habit, and the use of statins, did not differ in the two groups ( Table 1 ).
|Variable||Group A (restenosis ∗ ) ( n = 39)||Group B (no restenosis † ) ( n = 85)||P|
|Age (y)||64.3 ± 8.9||62.7 ± 9.6||.40|
|Men||35 (90%)||58 (68%)||.014|
|Systolic blood pressure (mm Hg)||139.8 ± 16.2||134.4 ± 18.5||.12|
|Diastolic blood pressure (mm Hg)||85 ± 8.3||82,2 ± 11,9||.20|
|Heart rate (beats/min)||65.1 ± 12.3||69,1 (14,4)||.14|
|Family history of CAD||12 (31%)||33 (39%)||.54|
|Total cholesterol >6.5 mmol/L||21 (54%)||57 (67%)||.22|
|Diabetes mellitus||3 (8%)||13 (15%)||.39|
|Statin use||31 (79%)||66 (78%)||.81|
|Hypertension||23 (59%)||59 (69%)||.31|
|Cigarette smoking||21 (54%)||24 (48%)||.33|
|Number of vessels > 1||19 (49%)||41 (48%)||.60|
|Angina at 1-month follow-up||3 (8%)||11 (13%)||.54|
|Angina at 6-month follow-up||21 (54%)||13 (18%)||<.001|
There were no significant differences in heart rate, systolic and diastolic aortic pressure, or angiographic variables, in terms of number of vessels involved, in either patient group ( Table 1 ).
CFR at 1-Month Follow-Up
Pulsed-wave Doppler tracings, suitable for CFR evaluation, were obtained in the distal portion of the LAD in all patients. CFR studies were well tolerated by all subjects. Wall motion during stress echocardiography was normal in all patients. In the entire study population, there was an inverse relationship between CFR detected at 1 and 6 months and the grade of coronary restenosis expressed by quantitative coronary angiographic values at follow-up ( P < .0001, ρ = −0.47, and P < .0001, ρ = −0.69, respectively; Figure 1 ).
At baseline, peak diastolic velocity in the LAD was similar in the two groups (median, 31 cm/sec [interquartile range, 23–43 cm/sec] vs 27 cm/sec [interquartile range, 23–32 cm/sec]; P = .12). However, during hyperemia, flow velocity increased more in group B than in group A (86 ± 27 cm/sec vs 73 ± 31 cm/sec, P = .03). Figures 2 and 3 show a representative example. Consequently, CFR was significantly reduced in group A at 1-month follow-up (2.3 ± 0.65 vs 3.1 ± 0.69, P < .0001; Figure 4 ). A cutoff value of ≤2.5 at 1 month was identified as optimal by receiver operating characteristic curve analysis (area under the curve, 0.79 ± 0.05; 95% confidence interval [CI], 0.70–0.88; P < .0001) for the best prediction of the occurrence of restenosis at the previous PCI site at follow-up ( Figure 5 ). On the basis of this cutoff, the event rate and calculated relative risk for coronary restenosis were both much higher in the group with lower (≤2.5) than with higher (>2.5) CFR (67% vs 15%, P < .0001). The probability of coronary restenosis was >10 times higher in patients with lower CFR than in those with higher CFR (odds ratio, 11; 95% CI, 4.5–26.9; P < .0001). Detection of low CFR at 1 month (CFR at 1 month ≤ 2.5) was the only independent risk factor of restenosis (hazard ratio, 4.2; 95% CI, 2.1–8.2; P < .0001). Male gender was only a marginally significant risk factor (hazard ratio, 2.4; 95% CI, 1.16–6.8; P = .09) (multivariate model χ 2 = 24.496, P < .001).
CFR ≤ 2.5 at 1 month was 67% sensitive and 87% specific for detecting significant restenosis, with positive and negative predictive values of 67% and 87% respectively, and accuracy of 79%. Moreover, in the restenosis group, CFR at 1 month decreased significantly, with an increase in in-stent restenosis, according to the Mehran classification (pattern I, 2.64 ± 0.5; pattern II, 2.3 ± 0.6; pattern III, 2.2 ± 0.6; pattern IV, 1.15 ± 0.2; P = .023), and there was an inverse relationship between CFR detected at 1 month and the Mehran classification ( P < .05, ρ = −0.32; Table 2 ).
|Variable||Mehran classification of in-stent restenosis ( n = 39)||P|
|Pattern I (focal) |
( n = 10)
|Pattern II (in-stent) |
( n = 12)
|Pattern III (proliferative) |
( n = 15)
|Pattern IV (Total occlusion) |
( n = 2)
|Age (y)||64 ± 7||63.5 ± 8.2||66 ± 10||58 ± 18.4||.60|
|Implanted stents/lesion||1.2 ± 0.4||1.4 ± 0.5||1.2 ± 0.4||1 ± 0||.60|
|Stent diameter (mm)||3 ± 0.33||3.1 ± 0.3||3 ± 0.4||2.5 ± 0||.16|
|Stent length (mm)||20 ± 78||18.1 ± 6.4||15 ± 5.4||11.5 ± 5||.18|
|Quantitative coronary angiography (%)||55.8 ± 3.9||69 ± 14.6||68.7 ± 10.4||90 ± 8||.001|
|CFR at 1 month||2.64 ± 0.5||2.3 ± 0.6||2.3 ± 0.6||1.15 ± 0.21||.023|
|CFR at 1 month ≤ 2.5||50%||66%||73%||100%||.46|
|CFR at 6-month follow-up||2.0 ± 0.6||1.7 ± 0.3||1.4 ± 0.4||1.31 ± 0.4||.03|