Author + information
- Received August 8, 2000
- Revision received November 14, 2000
- Accepted December 20, 2000
- Published online April 1, 2001.
- Christodoulos Stefanadis, MD, FACCa,* (, )
- Konstantinos Toutouzas, MDa,
- Eleftherios Tsiamis, MDa,
- Costas Stratos, MDa,
- Manolis Vavuranakis, MD, FACCa,
- Ioannis Kallikazaros, MDa,
- Dimosthenis Panagiotakos, MSa and
- Pavlos Toutouzas, MD, FACCa
- ↵*Reprint requests and correspondence: Dr. Christodoulos Stefanadis, 9 Tepeleniou Street, 15452 Paleo Psychico, Athens, Greece
We investigated the midterm clinical significance of human coronary atherosclerotic plaques temperature after a successful percutaneous coronary intervention.
Previous studies have shown an increased temperature in human atherosclerotic plaques. However, the prognostic significance of atherosclerotic plaque temperature in patients undergoing a successful percutaneous intervention is unknown.
We prospectively investigated the relation between the temperature difference (ΔT) between the atherosclerotic plaque and the healthy vessel wall and event-free survival among 86 patients undergoing a successful percutaneous intervention. Temperature was measured by a thermography catheter, as previously validated. The study group consisted of patients with effort angina (EA) (34.5%), unstable angina (UA) (34.5%) and acute myocardial infarction (AMI) (30%).
The ΔT increased progressively from EA to AMI (0.132 ± 0.18°C in EA, 0.637 ± 0.26°C in UA and 0.942 ± 0.58°C in AMI). The median clinical follow-up period was 17.88 ± 7.16 months. The ΔT was greater in patients with adverse cardiac events than in patients without events (ΔT: 0.939 ± 0.49°C vs. 0.428 ± 0.42°C; p < 0.0001). The ΔT was a strong predictor of adverse cardiac events during the follow-up period (odds ratio 2.14, p = 0.043). The threshold of the ΔT value, above which the risk for an adverse cardiac event was significantly increased, was 0.5°C. The incidence of adverse cardiac events in patients with ΔT ≥0.5°C was 41%, as compared with 7% in patients with ΔT <0.5°C (p < 0.001).
Increased local temperature in atherosclerotic plaques is a strong predictor of an unfavorable clinical outcome in patients with coronary artery disease undergoing percutaneous interventions.
Previous studies have revealed in vivo thermal heterogeneity within human atherosclerotic plaques (1–3). Thermal heterogeneity within the atherosclerotic plaques was shown in 20%, 40% and 67% of the patients with stable angina (SA), unstable angina (UA) and acute myocardial infarction (AMI), respectively, whereas no heterogeneity was shown in the control subjects. Ex vivo studies have shown that heat is released by activated inflammatory cells in the atherosclerotic plaques of the carotid arteries (4). Moreover, temperature correlated positively with cell density and mainly with the density of macrophages (4).
Furthermore, recent studies have shown that inflammatory cells such as macrophages and T lymphocytes are present in the majority of restenotic lesions of patients presenting with UA after a percutaneous intervention (5). In contrast, the smooth muscle cell area is similar in the restenotic lesions of patients with SA or UA (5). Thus, inflammation may lead to destabilization of atherosclerotic plaques after percutaneous interventions. However, increased plaque temperature associated with acute-phase reactants expresses an aggressive inflammation (3). Accordingly, the identification of lesions with increased temperature may indicate the lesions that are predisposed to future alterations of plaque morphology, producing adverse cardiac events.
The impact of increased local temperature of atherosclerotic plaques on clinical events after percutaneous coronary balloon angioplasty is unknown, however. The present study evaluated the mid-term clinical significance of the temperature of human coronary atherosclerotic plaques after a successful percutaneous intervention.
The study group consisted of 86 patients. Inclusion criteria were a single lesion <20 mm in length in a major native coronary artery and a proximal reference vessel diameter ≥2.5 mm. Thirty patients had effort angina (EA). Thirty patients with UA were unresponsive to maximal medical treatment after a two-day hospital period. In addition, 26 patients had AMI in which primary angioplasty was performed within 6 h after the onset of pain.
Patients medicated with corticosteroids or nonsteroidal anti-inflammatory drugs, except for aspirin, were excluded from the study. Moreover, patients with an intercurrent inflammatory or neoplastic condition likely to be associated with an acute-phase response were not enrolled in the study. Exclusion criteria also included multivessel disease and myocardial infarction less than one month before the intervention.
The study was performed between November 1998 and January 1999 at the Department of Cardiology, Athens Medical School, Athens, Greece. Baseline demographic and procedural variables were recorded and entered prospectively in a prespecified data base. The study protocol was approved by the institutional Ethics Committee, and each patient provided written, informed consent.
The design and construction characteristics of the coronary thermography catheter (3F diameter), which was developed in our institution, have been previously described in detail (1–3). In brief, the technical characteristics of the polyamide thermistor include 1) temperature accuracy of 0.05°C; 2) time constant of 300 ms; 3) spatial resolution of 0.5 mm; and 4) linear correlation of resistance versus temperature over the range of 33 to 43°C (1–3).
The culprit lesion of interest was outlined in two or more well-opacified views with biplane angiography. These projections were obtained before and after the procedure. A computer-assisted analysis system was used for quantitative coronary angiography (DCI-S, Automated Coronary Analysis, Philips, The Netherlands). Automated edge-detection of the minimal lumen diameter and reference diameter were measured by use of the guiding catheter filled with contrast agents as a scaling factor. Moreover, an intravascular ultrasound (IVUS) catheter (20-MHz; Visions Five-64 F/XTM Endosonics Corp., Rancho Cordoba, California) was introduced to the target vessel for detailed mapping of the area of interest. Thereafter, patients underwent percutaneous balloon angioplasty using standard technique (6). The balloon was dilated with a mixture of contrast medium and normal saline at 37°C. Procedural success was defined as final diameter stenosis <50% and Thombolysis In Myocardial Infarction (TIMI) flow grade 3. Subsequently, IVUS examination was repeated to exclude residual thrombus in patients with thrombus-containing lesions and to guide selection of healthy, without even small atheromatous plaques, and diseased segments where temperature measurements were planned to be performed. All patients received preprocedural aspirin and periprocedural heparin.
Thus, 5 min after contrast injection, the thermography catheter was advanced through the guiding catheter, and blood temperature was measured when the thermistor had just emerged from the tip of the guiding catheter without being in contact with the vessel wall. Thereafter, temperature measurements were made at five locations over a length of ∼1 cm of the normal vessel wall proximal to the lesion. The most frequent temperature of these locations was designated as the background temperature. In addition, measurements were performed at five different lesion sites at the region of interest (ROI). The coronary segments in which the temperature was measured were guided by IVUS findings. Accordingly, the temperature difference (ΔT) between the atherosclerotic plaque and the healthy vessel wall was calculated by subtracting the background temperature from the maximal temperature at the ROI.
After temperature measurements, deployment of stents was scheduled in all lesions to achieve optimal long-term results (7). All operators had no knowledge of the temperature measurements. Optimal stent expansion was defined as residual diameter stenosis <10% by quantitative coronary angiography. All patients were discharged from the hospital with ticlopidine, aspirin, statin and standard medical therapy.
All patients were prospectively followed for the occurrence of the following adverse events: 1) in-hospital repeat myocardial infarction, recurrent angina and death. Reinfarction was diagnosed on the basis of the recurrence of persistent ischemic chest pain followed by at least two-fold re-elevation of creatine kinase from the last measured value. Any recurrent chest pain that was accompanied by ST-T segment changes on the electrocardiogram was considered as recurrent angina if reinfarction was ruled out. 2) Long-term follow-up: the patients without in-hospital complications were followed periodically in our outpatient clinic, or they were serially interviewed by telephone at a mean of 17.88 ± 7.16 months. All patients underwent a treadmill stress test at three and six months and annually after the implantation. Patients with recurrence of symptoms or a positive treadmill stress test were scheduled for repeat angiography. The long-term occurrence of recurrent angina, myocardial infarction and death was recorded. For recurrent angina and myocardial infarction, we used the same definitions for in-hospital events.
All data are presented as the mean value ± SD. Univariate analysis was based on regression models, exact chi-square tests and Wilcoxon and Mann-Whitney criteria. Multivariate and exact logistic regression analyses were applied to assess the relation of ΔT with the adverse cardiac event, adjusted for several cofactors. A Cox proportional hazards model was applied to estimate the corresponding relative risk (i.e., odds ratio [OR]) and the relationship between survival and explanatory variables. All risk factors were tested for pairwise interactions. The main risk factor (ΔT) was adjusted for age, treated vessels, total cholesterol, diabetes mellitus, current smoking, hypertension, left ventricular dysfunction, previous myocardial infarction, acute coronary syndromes (UA and AMI), reference diameter and minimal lumen diameter. Left ventricular dysfunction was defined as ejection fraction <40%.
Sensitivity and specificity values and receiver-operating characteristics (ROC) curves were calculated to estimate the crucial threshold (cut-off point) of ΔT. To evaluate the diagnostic ability of the method, we calculated the area under the curve, and we examined its significance with the Wilcoxon criterion. A p value <0.05 was considered significant on univariate analysis, and p < 0.10 was considered significant on multivariate survival analysis. The analyses were performed with SPSS, release 8 (SPSS Inc., Chicago, Illinois), LogXact (Cytel Corp.) and STATA release 6 (STATA Corp., Lakeway Drive, Texas) statistical software programs.
The demographic characteristics of the study group (n = 86) are demonstrated in Table 1. There were no differences between the subgroups of patients with respect to baseline clinical characteristics. In all patients, the balloon angioplasty procedure was successful by achieving final diameter stenosis <50% and mechanical lysis of the thrombus in patients with thrombus-containing lesions. In 82 patients, 82 metallic stents were implanted after completing the temperature measurements. In all stented segments, optimal expansion was performed. In four patients, the stents could not be delivered to the target lesion due to tortuosity of the coronary vessel.
Procedural temperature measurements
The five measurements obtained for determination of background temperature were constant in each patient of the total study group, varying by only 0.05°C, with a standard deviation for each of the patients ranging from 0 to 0.0312. The temperature of the blood and healthy vessel wall did not differ (p = 0.45). The ΔT was similar between patients receiving aspirin or statin treatment before the percutaneous intervention and those patients who did not receive such medications (p = NS, respectively). The ΔT was significantly different between the subgroups of the study. In the patients with EA, ΔT was 0.132 ± 0.18°C; in patients with UA, it was 0.637 ± 0.26°C; and patients with AMI, it was 0.942 ± 0.58°C (p < 0.01) (Fig. 1).
Predictive value of δT on event-free survival
In-hospital complications were observed in four patients. Two of them had UA and two had AMI. Three of these patients died, and one patient with UA had a subacute thrombosis. No statistical significance was observed between the three subgroups. The remainder of the patients were discharged from the hospital with no adverse events.
After a median follow-up period of 17.88 ± 7.16 months, 17 more patients had an event. Thus, in total, 21 patients of the whole study group had an adverse cardiac event. Of them, 4 (13%) had EA, 8 (27%) had UA and 9 (35%) had AMI (p < 0.01). Three patients died after discharge from the hospital. Of them, one patient had a fatal AMI and two had a sudden death. Two patients had AMI, five had UA, four had SA and three had a positive treadmill stress test. In these patients, the repeat angiogram revealed restenosis of the treated vessel. The ΔT was greater in patients with adverse cardiac events than in patients without events (ΔT: 0.939 ± 0.49 [first quartile: 0.57, median quartile: 0.87, third quartile: 1.15] vs. 0.428 ± 0.42°C [first quartile: 0.07, median quartile: 0.26, third quartile: 0.70]; p < 0.0001) (Fig. 2). Moreover, ΔT was greater in the patients with EA and UA with adverse cardiac events as compared with those without events (Table 2)(Fig. 3). In patients with AMI, ΔT was greater, although this difference did not reach significant statistical difference. Cox regression analysis after adjustment for age, treated vessels, total cholesterol, diabetes mellitus, current smoking, hypertension, left ventricular dysfunction, previous myocardial infarction, acute coronary syndrome, reference diameter and minimal lumen diameter revealed that ΔT was a strong predictor of adverse cardiac events during the follow-up period (OR 2.14, 95% confidence interval 1.31 to 6.85, p = 0.043). In addition, left ventricular dysfunction, age, current smoking, minimal lumen diameter after the procedure and acute coronary syndrome were also predictors of an adverse cardiac event. (Table 3). However, the interactions between the covariates and the main factor of the analysis (ΔT) were not statistically significant.
Sensitivity and specificity analyses showed that the threshold of the ΔT value (cut-off point) above which the risk for an adverse outcome after the intervention was significantly increased, was 0.5°C (ROC area = 77%) (Fig. 4)(8). The sensitivity for this cut-off point was 86% (18 of 21 patients), and the specificity was 60%. We categorized the study group into those with ΔT ≥0.5°C and those with ΔT <0.5°C. The incidence of adverse cardiac events in patients with ΔT ≥0.5°C was 41%, as compared with 7% in patients with ΔT <0.5°C (p < 0.001). A Cox survival plot adjusted for ΔT and stratified for the cut-off point showed a clear relationship between ΔT and event-free survival (Fig. 5).
The present study demonstrated that after a successful percutaneous intervention, ΔT was a strong predictor of adverse cardiac events. Moreover, mean ΔT was greater in patients with adverse cardiac events within each subgroup, although in patients with AMI, a trend was found without achieving statistical significance. The threshold of the ΔT value above which the rate of adverse cardiac events was significantly increased was 0.5°C. The observed differences were significant and maintained at follow-up. According to these results, heat release from human coronary atherosclerotic plaques after a successful percutaneous intervention was associated with an adverse clinical outcome during a mid-term follow-up period. The majority of the adverse cardiac events were related to restenosis of the treated lesion. Furthermore, we confirmed previous observations indicating that ΔT is increased in unstable syndromes as compared with EA.
A considerable overlap of plaque temperature between patients with EA and patients with acute coronary syndromes was observed in previous studies (1–3). Thus, we enrolled patients with all clinical syndromes to include all temperature values and to correlate the individual ΔT with the clinical outcome, regardless of the clinical syndrome. Therefore, patients with adverse events had an increased ΔT as compared with patients without events during the follow-up period. In contrast, the treatment of lesions without an increased ΔT was associated with a favorable outcome, regardless of the initial clinical syndrome. However, in patients with AMI, although ΔT tended to be greater in patients with adverse events than in those without an event, statistical significance was not achieved. This finding may be due to the number of patients participating in this subgroup or the duration of follow-up period.
A critical question is whether the relationship between heat release and clinical outcome is due to confounding by an association of ΔT with other strong predictors of event-free survival. Whether ΔT and these strong prognostic factors share a common causal pathway still remains unknown. However, adjustment for all of these factors determines the true relation between ΔT and event-free survival, because any interaction among these covariates was not detected.
Furthermore, the relationship between intake of aspirin and statins after the procedure and event-free survival in patients with increased ΔT could not be determined, because the whole study group was discharged from the hospital with these medications. Thus, the effect of statins on plaque stabilization (9)and on local temperature of the culprit atherosclerotic plaques is unknown. In addition, the beneficial effects of aspirin in reducing the risks of a first myocardial infarction are directly related to high plasma concentrations of C-reactive protein (10,11). However, in previous studies, a positive correlation was observed between ΔT and concentrations of C-reactive protein and serum amyloid A (2,3). Thus, the administration of aspirin and statins may cause the true relationship between ΔT and event-free survival to be underestimated.
Potential explanations of clinical outcome associated with temperature measurements
The exact mechanism by which heat is generated from the atherosclerotic plaques, resulting in more adverse events, in unknown. Several factors may contribute to heat release from the atherosclerotic plaques, which theoretically result in poor prognosis after percutaneous interventions. It seems that the process leading to increased temperature precedes plaque rupture and formation of thrombus. This hypothesis is based on the considerable overlap of plaque temperature between patients with EA and patients with acute coronary syndromes, indicating that a percentage of patients with clinically stable angina had plaques with increased risk of ischemic events.
The first study showing thermal heterogeneity in vitro within atherosclerotic plaques, by Casscells et al. (4), also demonstrated a relationship between the number of infiltrated macrophages and thermal heterogeneity. It has been reported that a fifth of culprit lesions in patients with chronic SA show features of recent injury and repair, and inflammation is present in these lesions (12). In another study, elevated concentrations of acute-phase proteins are found to be predictors of coronary events in patients with EA and UA (13). Our results support previous reports demonstrating that an inflammatory process is involved in the onset of acute coronary syndromes (14)and is also present in culprit lesions of patients with EA (13). Moreover, previous studies from our group showed that increased plaque temperature associated with acute-phase reactants expresses an aggressive inflammation (3). Thus, inflammation plays a significant role in heat release from atherosclerotic plaques, leading to destabilization of restenotic lesions during the follow-up period (5,15). Furthermore, inflammatory cells (macrophages, T lymphocytes) are present in the majority of restenotic lesions of patients presenting with UA. In contrast, the smooth muscle cell area is similar in the restenotic lesions of patients with SA or UA (5). This inflammatory reaction is related to the underlying morphology. Thus, the detection of lesions with infiltration of inflammatory cells during a percutaneous intervention may indicate the lesions that are predisposed to future alterations of plaque morphology.
In addition, inflammation is correlated with increased vascular injury (16). The oversizing of stented segments leads to injury of the arterial wall, which is associated with thrombus formation and increased neointimal growth. Accordingly, stent oversizing may appear to be an undesirable goal in lesions with chronic inflammation that leads to heat production, because the rate of restenosis in this subgroup of patients may be considerably increased.
This study was prospective but observational. However, the study group was rather small. Nevertheless, the results are significant and were maintained during the follow-up period. Follow-up angiography was not performed, except for the majority of patients with clinical events. Moreover, follow-up angiography could lead to an increased rate of interventions based only on angiographically demonstrated stenosis at follow-up (17). Despite these results, repeat angiography in all patients would provide useful information on the correlation of ΔT with the angiographic characteristics of the lesion. This was not the aim of the present study, however. Also, the results of this study seem unlikely to be affected by progression of the disease in other stenoses, because all patients had a single significant stenosis.
In addition, despite our effort, it was impossible to eliminate all potential confounding factors. Finally, we cannot reach the conclusion that the association between increased temperature and an adverse clinical outcome is causal. The increase in temperature of atheromatous plaques may be explained by inflammatory activity. However, the results of a recent study demonstrated that there is a positive association between the clinical syndrome and the extent of inflammation in atherectomy specimens of patients with a restenotic lesion (5).
The results of this study showed that the difference in temperature between the atherosclerotic plaque and the healthy vessel wall is a strong predictor of event-free survival after a successful percutaneous intervention in patients with coronary artery disease. This study does not prove a causal relationship between increased temperature of the atherosclerotic plaque and adverse cardiac events. Our results, however, serve as an incentive to consider strategies to stabilize the atherosclerotic plaques after a percutaneous coronary intervention.
☆ The study was supported by a grant from the Hellenic Cardiac Foundation, Athens, Greece.
- acute myocardial infarction
- temperature difference
- area under the curve
- effort angina
- intravascular ultrasound
- odds ratio
- receiver-operating characteristics
- region of interest
- stable angina
- unstable angina
- Received August 8, 2000.
- Revision received November 14, 2000.
- Accepted December 20, 2000.
- American College of Cardiology
- Stefanadis C.,
- Diamantopoulos L.,
- Vlachopoulos C.,
- et al.
- Stefanadis C.,
- Diamantopoulos L.,
- Dernellis J.,
- et al.
- Piek J.,
- van der Wal A.C.,
- Meuwissen M.,
- et al.
- Ellis S.G.
- Stone G.W.,
- Brodie B.R.,
- Griffin J.J.,
- et al.
- Bustos C.,
- Hernandez-Presa M.A.,
- Ordego M.,
- et al.
- van der Wal A.C.,
- Becker A.E.,
- Koch K.T.,
- et al.
- Moreno P.R.,
- Falk E.,
- Palacios I.F.,
- et al.
- Grewe P.,
- Deneke T.,
- Machraoui A.,
- et al.
- Roders C.,
- Edelman E.R.