Author + information
- Received December 4, 2001
- Revision received May 30, 2002
- Accepted July 2, 2002
- Published online October 16, 2002.
- Michael N Zairis, MD*,* (, )
- John A Ambrose, MD, FACC†,
- Stavros J Manousakis, MD*,
- Alexander S Stefanidis, MD*,
- Olga A Papadaki, MD*,
- Helen I Bilianou, MD*,
- Mary C DeVoe, RN†,
- Constantine N Fakiolas, MD*,
- Evangelos G Pissimissis, MD*,
- Christopher D Olympios, MD*,
- Stefanos G Foussas, MD, FESC, FACC*,
- GENERATION Study Group
- ↵*Reprint requests and correspondence:
Dr. Michael N. Zairis, 273-277 Alkiviadou Street, Piraeus, 18536, Greece.
Objectives The objective of this study was to evaluate the association of high plasma levels of either C-reactive protein (CRP), lipoprotein (a) (Lp[a]) or total homocysteine (tHCY) with the long-term prognosis after successful coronary stenting (CS)
Background High plasma levels of either CRP, Lp(a) or tHCY may have an impact in coronary artery disease. However, long-term prospective data after coronary stenting (CS) are limited.
Methods Four-hundred and eighty-three consecutive patients with either stable or unstable coronary syndromes were followed for up to three years after successful CS. The composite of cardiac death, myocardial infarction or rehospitalization for rest unstable angina, whichever occurred first, was the prespecified primary end point. Moreover, the one-year incidence of clinical recurrence of symptoms, in-stent restenosis (ISR) and progression of atherosclerosis to a significant lesion (PTSL) were additionally evaluated. PTSL was defined as an increase by at least 25% in the luminal diameter stenosis of a known nonsignificant lesion (≤50% luminal diameter stenosis) that was located in a nonintervened vessel at restudy, resulting in an angiographically significant lesion (≥70% luminal diameter stenosis).
Results By the end of the follow-up, high plasma levels of either CRP or Lp(a) but not tHCY were independently associated with the primary end point. In particular, CRP ≥0.68 mg/dl (p < 0.001) or Lp(a) ≥25 mg/dl (p = 0.003) conferred a significantly increased risk. By 1 year, a CRP ≥0.68 mg/dl conferred a significantly increased risk for clinical recurrence of symptoms (p < 0.001) or PTSL (p < 0.001). None of the studied biochemical markers was related to ISR.
Conclusions High plasma levels of either CRP or Lp(a) but not tHCY may be associated with a higher incidence of late adverse events after successful CS. PTSL in vessels not previously intervened upon may play a significant role in the underlying pathophysiology as opposed to ISR.
Coronary atherosclerosis is a treatable but incurable disease, and percutaneous coronary intervention (PCI) constitutes an established therapeutic approach. Although the introduction of stents coupled with new antiplatelet agents has substantially reduced early and late complications, a significant proportion of successfully treated patients experience late nonfatal or fatal ischemic events. The pathophysiologic mechanisms for these late events include in-stent restenosis (ISR) or coronary disease progression in other lesions that had not undergone intervention. The identification of these high-risk patients is essential in everyday clinical practice but remains a challenge. To more accurately risk-stratify these patients with late events, several clinical and angiographic characteristics as well as biochemical markers have been suggested as possible predictors. In particular, plasma C-reactive protein (CRP), lipoprotein (a) (Lp[a]) and total homocysteine (tHCY) have all been proposed as markers for secondary risk stratification. However, long-term prospective studies concerning these biochemical markers are limited or their results are inconsistent (1–8).
The Global Evaluation of New Events and Restenosis After Stent Implantation (GENERATION) study was a prospectively designed observational study to investigate the impact of high plasma levels of either CRP, Lp(a) or tHCY on the long-term incidence of clinical events after successful coronary stenting (CS) and to evaluate globally the contribution of either ISR or progression of coronary atherosclerosis to these events. For this purpose, 903 consecutive patients who were admitted from January 1998 through April 2000 with the diagnosis of either stable angina, non–ST-segment elevation acute coronary syndrome (NSTACS) or ST-segment elevation acute myocardial infarction (STEMI) were considered eligible for the study. Only patients with STEMI who presented within the first 6 h and NSTACS patients who presented within 24 h of index pain were included. Patients with left bundle-branch block, infection or inflammatory disease, malignancy, hepatic dysfunction, plasma creatinine >1.5 mg/dl (upper normal value), a history of myocardial infarction, coronary revascularization in the last six months or a left ventricular ejection fraction <35% before current PCI were excluded.
Of the 903 patients, 501 eligible patients underwent PCI with stent implantation. Eighteen patients who needed urgent coronary surgery or experienced in-hospital death or myocardial infarction were excluded. Thus, 483 patients with successful CS in 539 vessels (554 lesions) comprised the study cohort. The Hospital Ethics Committee approved the study and informed consent was obtained from all participants.
Upon admission, venous blood samples were obtained before the intravenous administration of drugs. Coded plasma samples were stored at −24°C for CRP and Lp(a) analysis, which were performed 24 to 72 h later. Coded plasma for tHCY measurement was immediately stored at −80°C and analyzed at the end of the study.
CRP was measured using a quantitative turbidimetric method (Turbicant, Dade Behring, Germany). Lp(a) was measured by immunoassay method (N Latex Lp(a), Nephelometer BN 100, Dade Behring, Germany). A range of 0.5 to 60 mg/dl and 8 to 160 mg/dl were covered for CRP and Lp(a), respectively. The upper normal value of CRP is 0.5 mg/dl in the laboratory. Plasma tHCY was measured by an immunoassay method (FPIA assay, IMx system, Abbott) with a range of 4 to 500 μmol/l. For values below the limit of detection, the lower limit value was used for statistical analysis. The results of biochemical markers were not decoded until the end of the study.
Two independent and experienced angiographers who had no knowledge of the study performed all quantitative measurements, using off-line quantitative computerized analysis. Values were calculated as the mean score given by the two observers. Intra-observer and inter-observer variability were <5% and 10%, respectively.
Follow-up and study end points
Before discharge, all patients were advised for a six-month angiographic restudy. After discharge, patients were followed up clinically at one, three and six months and subsequently every six months for up to three years.
The composite of cardiac death, nonfatal myocardial infarction or rehospitalization for rest unstable angina (whichever occurred first) up to three years after discharge was the prespecified primary end point. The incidence of clinical recurrence of symptoms (CLR), angiographic ISR or progression to a significant coronary lesion (PTSL) in vessels other than the treated ones during the one-year of follow-up were additionally evaluated. ISR was defined as >50% in-stent stenosis on re-evaluation. In case of ISR in two stented lesions in the same patient, the more significant was taken into account. PTSL was defined as an increase by at least 25% in luminal diameter stenosis of a known nonsignificant lesion (≤50% luminal diameter stenosis) located in a nonintervened vessel at restudy, resulting in an angiographically significant lesion (≥70% luminal diameter stenosis).
Coronary angiography and PCI
Patients with NSTACS or STEMI were scheduled for coronary angiography and PCI beyond the third day of admission and after their initial clinical stabilization or earlier if it was clinically indicated. Unfractioned heparin (10,000 IU) was given intravenously at the start of the PCI. Balloon catheters and stents were selected by the interventionalist. Predilation was performed, and all stents were deployed at high pressures. The procedure was considered successful if: 1) there were no adverse events, including death, myocardial infarction or the need for urgent coronary surgery; 2) the visually estimated residual stenosis was <30% and 3) there was TIMI 3 flow distally immediately after stent implantation. Aspirin was given orally (160 to 325 mg) upon admission to all patients and was continued indefinitely. Glycoprotein IIb/IIIa inhibitors were administered periprocedurally in 21.1% of patients. All patients received ticlopidine (250 mg b.i.d.) for four weeks after PCI.
Values were expressed as mean ± SD for normally distributed and as median (range) for non-normally distributed variables. Normal distribution was evaluated with Kolmogorov–Smirnov test. Values were compared using t test or Mann-Whitney U test as appropriate. Associations between categorical variables were tested by Fisher’s test. Receiver operating characteristics (ROC) curves were constructed for biochemical markers to evaluate their accuracy in the prediction of the primary end point (measured by the area under the ROC curve, range 0.5 to 1). To avoid arbitrary cutoff points of biochemical markers for the prediction of the primary end point, the optimal cutoff points with the highest predictive accuracy, which separated the cohort into two populations, were estimated by ROC analysis. If a biochemical marker didn’t accurately predict the primary end point, its median value was selected as the cutoff point. Event-free survival was analyzed with Kaplan–Maier method and log-rank used for comparisons between curves. Univariate and multivariate Cox proportional hazards regression analyses were constructed for the determination of univariate and multivariate predictors of the composite primary end point or of or its components. Univariate and multivariate binary logistic regression analysis was constructed to evaluate the univariate or multivariate predictors of ISR or PTSL. All variables, presented in Tables 1 and 2, ⇓were evaluated as possible predictors of the study end points ⇓and those with a p < 0.1 were included into the multivariate models. All tests were two-tailed, and a p < 0.05 was considered as significant. Statistical analysis was performed with SPSS statistical software (release 10.0, SPSS, Chicago, Illinois).
Baseline clinical characteristics are shown into Tables 1 and 2. The majority of the included STEMI patients had received intravenous thrombolysis (147/177; 83%) after admission and five underwent either primary or rescue PCI. Twenty-five of the included STEMI patients (with a pre-PCI left ventricular ejection fraction >35%) had not received acute reperfusion therapy because of contraindications.
Significant univariate associations between biochemical markers and other baseline variables
Stable angina patients had significantly lower baseline CRP values than those with either NSTACS (median, range 0.52, 0.50 to 1.20 mg/dl vs. 0.59, 0.50 to 4.10 mg/dl; p < 0.001), or STEMI (median, range 0.59, 0.50 to 5.10 mg/dl; p < 0.001). There was no significant difference in plasma levels of either Lp(a) or tHCY among the groups defined by the qualified coronary syndrome (data not presented). The median baseline tHCY level was 8 μmol/l higher in patients with a history of coronary bypass grafting than in patients without (21 μmol/l vs. 13 μmol/l; p = 0.01); 5 μmol/l higher in patients with a history of cerebrovascular or peripheral artery disease than the patients without such a history (18 μmol/l vs. 13 μmol/l, p < 0.001); 4 μmol/l higher in patients ≥75 years of age (17 μmol/l vs. 13 μmol/l in patients <75 years; p = 0.01). Moreover, median baseline tHCY was 2 μmol/l higher in patients with multivessel coronary artery disease than those with single artery disease (14 μmol/l vs. 12 μmol/l; p = 0.001) and 1 μmol/l higher in diabetics than nondiabetics (14 μmol/l vs. 13 μmol/l; p = 0.08). No other significant associations were found between the studied biochemical markers and the other baseline variables. Furthermore, biochemical markers were not significantly interrelated (data not presented).
Clinical follow-up was obtained in 465 (465/483; 96.3%) patients at 22.1 ± 8.3 months (1 to 37 months). By the end of follow-up, 20 patients (20/465; 4.3%) died, 21 (21/465; 4.5%) suffered from a nonfatal myocardial infarction and 35 (35/465; 7.5%) were rehospitalized because of rest unstable angina for a total of 76 primary events. By one year, 121 patients (121/465 26%) developed CLR, including 10 (10/465; 2.1%) who died, 15 (15/465 3.2%) with a nonfatal myocardial infarction, 25 (25/465; 5.4%) who rehospitalized because of rest unstable angina, 68 (68/465; 14.6%) who developed new exertional angina and 3 (3/465; 0.6%) with exacerbated heart failure.
During this one-year time period, 309 (309/465; 66.5%) patients underwent angiographic restudy, including 114 (114/121; 94.2%) with CLR and 195 (195/344; 57%) asymptomatic. There were no significant differences concerning the plasma levels of the study biochemical markers between asymptomatic patients with or without angiographic restudy (data not presented). ISR was observed in 108 (108/309; 34.9%) patients, including 69 (69/74 93.2%) with new exertional angina or heart failure, 7 (7/15; 46.6%) with nonfatal myocardial infarction, 14 (14/25; 56%) who rehospitalized because of rest unstable angina, and 18 (18/195; 9.2%) without symptoms. PTSL was observed in 47 (47/309; 15.2%) patients, including 13 (13/74; 17.6%) with new exertional angina or exacerbated heart failure, 9 (9/15; 60%) with new nonfatal myocardial infarction, 14 (14/25; 56%) with rest unstable angina and 11 (11/195; 5.6%) asymptomatic.
Biochemical markers and primary end point
Receiver operating characteristics analysis indicated that either CRP or Lp(a) but not tHCY had reasonable accuracy for the composite primary end point (Fig. 1 A). The optimal cutoff points of CRP and Lp(a) were 0.68 mg/dl and 25 mg/dl, respectively (Figs. 1B and 1C). Patients with CRP values ≥0.68 mg/dl were at higher risk for the composite end point or its components than patients with CRP values <0.68 mg/dl for up to three years (Table 3, Figs. 2A and 3A). In particular, the former were at higher risk for either rehospitalization for rest unstable angina (hazards ratio [HR] 7.10; 95% CI 3.30 to 15.26; p < 0.001), nonfatal myocardial infarction (HR 5.67; 95% CI 2.19 to 14.66; p < 0.001) or cardiac death (HR 3.16; 95% CI 1.25 to 7.98; p = 0.01) than the latter. Moreover, patients with Lp(a) values ≥25 mg/dl were at higher risk for either the composite primary end point, rehospitalization for rest unstable angina (HR 2.09; 95% CI 1.16 to 4.12; p = 0.03) or nonfatal myocardial infarction (HR 3.31; 95% CI 1.33 to 8.22; p = 0.01) than those with Lp(a)<25 mg/dl but not for cardiac death (HR 1.27; 95% CI 0.48 to 3.34; p = 0.87) (Table 3, Figs. 2B and 3B). There was no significant difference in the incidence of the end points between the groups of tHCY (Table 3, Figs. 2C and 3C).
Biochemical markers and CLR, ISR or PTSL during the first year
The distribution of CLR among the groups is presented in Figure 4 (Figs. 4A to 4C). By univariate Cox proportional hazards analysis, several variables, including a plasma CRP ≥0.68 mg/dl (p < 0.001) or Lp(a) ≥25 mg/dl (p = 0.05), were related to the one-year CLR. However, by multivariate Cox proportional hazards analysis, only diabetes mellitus (HR 1.88; 95% CI 1.27 to 2.78; p = 0.002), post-PCI in-stent residual stenosis (HR 1.08; 95% CI 1.05 to 1.10; p < 0.001), B2- or C-treated lesion (HR 1.72; 95% CI 1.15 to 2.57; p = 0.008), a plasma CRP ≥ 0.68 mg/dl (HR 2.70; 95% CI 1.64 to 3.81; p < 0.001) and left ventricular ejection fraction (HR 0.97; 95% CI 0.95 to 0.99; p = 0.04) were independent predictors of the one-year CLR. None of the study biochemical marker was related to ISR (Figs. 4A to 4C). The distribution of PTSL between the groups is depicted in Figures 4A to 4C. By univariate logistic regression analysis, several variables, including a plasma CRP ≥ 0.68 mg/dl (p < 0.001) or Lp(a) ≥25 mg/dl (p = 0.05), were positively related to PTSL. However, by multivariate logistic analysis, only a plasma CRP ≥0.68 mg/dl (relative risk 19.34; 95% CI 8.80 to 62.97; p < 0.001) conferred a significant increased risk.
C-reactive protein is a sensitive and nonspecific inflammatory marker (9). In this study increased CRP levels conferred a significant increased risk for new fatal or nonfatal cardiac events up to three years after successful CS. This association was significant by univariate and multivariate analysis and after adjustment for the qualifying event. These results are in agreement with previous prospective studies, which have consistently showed that elevated plasma CRP levels are associated with the incidence of short- or long-term ischemic complications after PCI with or without stent implantation (1,10–12). Moreover, in this study, a strong and independent significant association was also found between elevated CRP and progression of atherosclerosis (in not previously treated vessels) during the first year of follow-up. This is consistent with the findings of a previous retrospective study from our group in a different cohort of patients. Thus, a low-grade inflammatory process associated with an elevated plasma CRP may predispose to atherosclerotic plaque progression and the induction of future fatal or nonfatal ischemic complications.
However, elevated plasma CRP levels were not related to the rate of ISR. This finding is in disagreement with the results reported by Walter et al. (11). These investigators studied a cohort of 256 patients in whom 86% had a 6-month angiographic follow-up and observed that tertiles of CRP upon admission were significantly associated with ISR. In our study, 114 symptomatic (114/121; 94%) and 195 (195/344; 56.7%) asymptomatic patients underwent restudy during the first year, corresponding to a 67% angiographic follow-up. The lower rate of recatheterization in the present study, as well as differences in either the structure of the studied populations or the statistical approach used, may account for these disparate findings. However, in GENERATION, 34% (26/76) of primary acute coronary events occurred after the first year, when ISR was less likely to occur. This also suggests that progression of atherosclerosis in untreated coronary plaques rather than ISR may be a responsible mechanism for the high CRP-related excess in clinical events. More studies are needed to elucidate this issue.
Lipoprotein (a) has been considered as a link between atherosclerosis and thrombosis because of its similarity to both low density lipoprotein and plasminogen (13). Two larger previous angiographic studies could not show a significant relationship between high plasma Lp(a) levels and the 6-month ISR rate (5,6) or one-year clinical outcome (5). In concordance with these reports, we found no relationship between Lp(a) and one-year CLR or ISR. However, Lp(a) conferred a significantly increased risk of the primary end point at the end of the follow-up (Fig. 2B). An elevated plasma Lp(a) level conferred only a marginally significant increased risk of PTSL, and this association was not as strong as between CRP and PTSL. It is possible that a more protracted latent period may be needed in order to manifest clinically the unfavorable influence of an elevated Lp(a) on atherosclerotic plaque instability.
Although elevated plasma tHCY levels have been considered as an independent risk factor of coronary artery disease, the role of hyperhomocystinemia in the prediction of restenosis and clinical long-term prognosis after PCI is unclear (6–8). No significant association between plasma tHCY and any end point was found in this study.
Small previous studies have found a positive relationship between elevated plasma tHCY levels and angiographic restenosis after PCI with (8) or without (7) stent implantation. However, because of the small sample size in these studies, confounding variables (14,15) may not have been controlled. In GENERATION, a strong association between tHCY levels and cardiovascular risk profile was found, which is consistent with previous epidemiological studies (15).
The determined cutoff points of either CRP or Lp(a) must not be considered as generally applied. These cutoff points optimally dichotomize the GENERATION population for the prediction of the study end points. Moreover, the recently recommended high-sensitivity CRP assay was not available at the beginning of the study and there is not a standardized analytic method for Lp(a) levels determination.
In addition, our findings concerning ISR or PTSL may have been modified if more asymptomatic patients underwent an angiographic restudy.
The results of the GENERATION study indicate that high plasma levels of either CRP or Lp(a) but not tHCY confer an increased risk for the long-term cardiac mortality and morbidity after successful CS. Although the role of ISR in this increased risk could not be definitely excluded, a significant progression of atherosclerosis in not previously intervened atherosclerotic plaques appears to be involved.
- clinical recurrence of symptoms
- C-reactive protein
- coronary stenting
- Global Evaluation of New Events and Restenosis After Stent Implantation
- in-stent restenosis
- lipoprotein (a)
- non–ST-segment elevation acute coronary syndrome
- percutaneous coronary intervention
- progression to a significant lesion
- receiver operating characteristics
- ST-segment elevation myocardial infarction
- total homocysteine
- Received December 4, 2001.
- Revision received May 30, 2002.
- Accepted July 2, 2002.
- American College of Cardiology Foundation
- Heeschen C.,
- Hamm C.W.,
- Bruemmer J.,
- Simoons M.L.
- Desmarais R.L.,
- Sarembock I.J.,
- Ayers C.R.,
- Vernon S.M.,
- Powers E.R.,
- Gimple L.W.
- Ribichini F.,
- Steffenino G.,
- Dellavalle A.,
- et al.
- Wehinger A.,
- Kastrati A.,
- Elezi S.,
- et al.
- Buffon A.,
- Liuzzo G.,
- Biasucci L.M.,
- et al.
- Walter D.H.,
- Fichtlscherer S.,
- Sellwig M.,
- Auch-Schwelk W.,
- Schachinger V.,
- Zeiher A.M.
- Chew D.P.,
- Bhatt D.L.,
- Robbins M.A.,
- et al.