Author + information
- Received March 16, 2005
- Revision received June 23, 2005
- Accepted June 27, 2005
- Published online October 18, 2005.
- Wissam A. Jaber, MD,
- Ryan J. Lennon, MS,
- Verghese Mathew, MD,
- David R. Holmes Jr, MD,
- Amir Lerman, MD and
- Charanjit S. Rihal, MD, MBA⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Charanjit S. Rihal, Director, Cardiac Catheterization Laboratory, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905
Objectives We sought to determine whether the prescription of evidence-based medications at discharge after successful percutaneous coronary intervention (PCI) can predict long-term clinical outcome.
Background The association of standard-of-care drug utilization and long-term mortality and morbidity after PCI is not well studied.
Methods We performed a retrospective cohort study of successful PCI procedures performed on 7,745 patients between March 1, 1998, and December 31, 2004. Discharge medications were analyzed, and a medication score (MEDS) was developed. A MEDS of 1 was assigned for each of the following medication classes: 1) antiplatelet, 2) lipid-lowering, 3) beta-blocker, and 4) angiotensin-converting enzyme (ACE) inhibitor. The outcomes measured were long-term death, myocardial infarction, and revascularization.
Results Patients with MEDS of 3 to 4 had higher-risk profiles based upon standard clinical and angiographic criteria. Despite this, at a median follow-up of 36 months, patients with a MEDS of 3 or 4 were at lower risk of death than those with a MEDS of 0 or 1 (8.9%, 7.5%, and 13% for MEDS of 4, 3, and 0 to 1, respectively; p = 0.014). After adjustment for covariates, a MEDS of 3 to 4 was associated with significantly lower mortality or myocardial infarction in follow-up than a MEDS of 0 to 1 (hazard ratios of 0.72 and 0.67 for MEDS of 3 and 4, respectively; p < 0.01). There was no association between MEDS and target vessel revascularization.
Conclusions After successful PCI, the use of multiple evidence-based classes of cardiovascular medications—antiplatelet, lipid-lowering, beta-blockers, and ACE inhibitors—is associated with improved outcome free of death or MI.
Despite the effect percutaneous coronary intervention (PCI) has on cardiovascular symptoms, it does not mitigate the underlying atherosclerotic process and, more importantly, does not prevent death or myocardial infarction (MI) (1). Antiplatelet therapy, angiotensin-converting enzyme (ACE) inhibitors, beta-blockers, and lipid-lowering therapy have each been shown to improve survival and decrease cardiovascular events in patients with coronary artery disease, partly by modifying the disease process (2–7). These classes form the basis for evidence-based medicine guidelines. Despite the beneficial effects these medical therapies have on clinical outcome, their use after PCI has not been universal (8,9). Whether the extent to which they are prescribed after PCI can be used as a predictor for clinical outcome has not been established.
In this study, we sought to define what effect the application of evidence-based medical therapy in terms of medical treatment after successful PCI has on clinical outcome. We hypothesized that a simple medication score (MEDS) based on the number of cardioprotective medications a patient is discharged on will correlate with improved outcome after PCI.
All patients undergoing PCI at the Mayo Clinic are prospectively enrolled in a registry that includes baseline demographic, clinical, and angiographic characteristics. These patients are contacted 6 and 12 months after the procedure and yearly thereafter by trained research technicians. The supervisor of the data entry technicians performs routine audits of the data every one to two months: data variables that are out of range are identified, and the corresponding medical records are reviewed; approximately 10% of all patients in the registry also are selected routinely for chart review so that the accuracy of the data that were put into the database is correct.
After approval by the Mayo Clinic Institutional Review Board, the PCI registry was searched for all successful PCI procedures performed between March 1, 1998, and January 31, 2004, at Saint Mary’s Hospital in Rochester, Minnesota. These dates were chosen because our practice remained relatively stable with regard to adjunctive medical therapies (in particular clopidogrel) and stent types (bare-metal). Successful PCI was defined as the presence of <50% residual stenosis in the treated coronary lesion after dilation and the absence of in-hospital death, Q-wave MI, or coronary artery bypass graft surgery (CABG) immediately after the PCI. For patients undergoing two procedures during the study period, only the first procedure was included. Baseline data on the patients were available from the database, including clinical characteristics, details of the PCI, and medication use. Patients who refused authorization for medical record research were excluded, as per the state of Minnesota statute.
A MEDS was assigned to each patient based on the medications at discharge from the hospital. A MEDS of 1 was added for each of the following medication classes: antiplatelets (aspirin), beta-blockers, ACE inhibitors, and lipid-lowering agents. Therefore, a MEDS of 0 means the patients were discharged on none of these medication classes and a MEDS of 4 means they were discharged on all four of them. The patients were then grouped into four categories: 1) MEDS 0 or 1; 2) MEDS 2; 3) MEDS 3; or 4) MEDS 4.
Follow-up was achieved through telephone calls at six months and then yearly after PCI, in addition to the review of hospital records. Measured outcomes were death, MI (defined by increase in cardiac biomarkers), CABG, and target vessel revascularization (TVR). The study was approved by the institutional committee on human research.
Continuous variables are summarized as mean ± SD. Discrete variables are presented as group percentages. Kaplan-Meier estimates were used to describe survival rates. For time-to-event analyses, follow-up began at the date of discharge. One-way analysis of variance, Pearson’s chi-square test, and the log-rank test were used to test the significance of group differences.
Cox proportional hazards survival models were used to estimate the hazard ratios for the MEDS. The MEDS group 0 to 1 was designated as the reference group. Covariate models for the three end points (death, death/MI, and death/MI/TVR) were built using backward selection on risk factors significantly associated with the end points in simple Cox regression models. The three parameters for MEDS effects were then added to the covariate models to assess the “independence” of these effects.
There were 10,060 PCI procedures performed on 8,465 patients. A total of 205 patients denied authorization for their records to be used for research. The earliest PCI for the remaining 8,260 patients were selected for analysis. Of these, 7,769 were successful PCI procedures. A total of 24 patients did not have enough medication information to calculate a MEDS, resulting in a final sample size of 7,745. The median follow up was 36 months.
Only 26 patients had a MEDS of 0; thus, the 0 and 1 MEDS groups were combined to form a sample of 507 patients. A total of 1,739 patients had a MEDS of 2, 3,321 had a MEDS of 3, and 2,178 had a MEDS of 4. The baseline characteristics of the patients in each of these groups are shown in Table 1.The angiographic and procedural characteristics are detailed in Table 2.The median amount of time between the last contact with the patients and the time of data analysis was similar between the four groups, suggesting similar compliance and follow-up. As compared with patients with lower MEDS, patients with a MEDS of 4 were more likely to have had a recent MI (50% for MEDS 4 vs. 13% for MEDS 0 to 1; p < 0.01), to present with congestive heart failure (CHF) (14% vs. 8%, p < 0.01), to have an ejection fraction ≤40% (14% vs. 7%; p < 0.01), and to have comorbid illnesses and coronary artery disease risk factors. Coronary lesions also were more likely to be complex (type C lesion, presence of thrombus, and/or no flow), and PCI was more likely to be an emergency intervention (29% vs. 10%; p < 0.01). Patients with a MEDS of 0 generally were older, with more renal disease and malignancies. The groups were not different in the number of coronary segments treated and stents inserted, and were only slightly different in the extent of their coronary artery disease on angiography (29% of patients with a MEDS of 4 had one-vessel disease vs. 31% of patients with a MEDS of 0 to 1; 38% vs. 42% for two-vessel disease; 32% vs. 27% for three-vessel disease; p = 0.03). Although the use of clopidogrel was less common in the lowest score group, it was similar among the three other groups.
At the end of the follow-up period, there were a total of 667 deaths; 1,045 patients with death or MI; and 1,824 with death, MI, CABG, or TVR. Despite the fact that the patients with MEDS of 3 and 4 had a higher risk profile based on the parameters described in Tables 1 and 2, their observed mortality was lower than that of patients with lower MEDS (8.9%, 7.5%, and 13% for MEDS of 4, 3, and 0 to 1, respectively, p = 0.014; Table 3).After including MI and MI/CABG/TVR with death in a composite end point, the difference between the four groups became nonsignificant, although there was a trend toward better outcome with higher scores compared with scores of 0 or 1 (Table 3).
Relevant baseline clinical and angiographic characteristics were included in a multivariable model. These characteristics included age, diabetes, CHF on presentation, history of CHF, peripheral vascular disease, history of cardiovascular accident, smoking, renal disease, malignancy, previous CABG, presence of angina, recent MI, hyperlipidemia, multivessel disease, glycoprotein IIb/IIIa use, use of intra-aortic balloon pump, angiographically complex lesion, and Thrombolysis In Myocardial Infarction (TIMI) flow grade after procedure.
In this multivariable analysis, many of the baseline clinical and procedural characteristics were associated with a worse clinical outcome (Table 4).Although age (hazard ratio [HR] 1.03), diabetes (HR 1.42), previous cardiovascular history (HR 1.26 for previous MI and 2.27 for CHF), renal disease (HR 2.05), hypertension (HR 1.19), recent MI (HR 1.25), and smoking (HR 1.38) were related to increased risk of death or MI, elective PCI (HR 0.84) and TIMI flow grade 3 after PCI (HR 0.69) were associated with a lower risk. Adjusting for all these factors, a higher MEDS appears to be related to a decrease in the rate of mortality, although the results did not reach statistical significance (HR 0.77 and 0.74 for MEDS 3 and 4, respectively, compared with MEDS 0 to 1, p = 0.058). There was, however, a significant decrease in the risk of death or MI associated with higher MEDS (HR 0.72 and 0.67 for MEDS of 3 and 4, respectively, p < 0.01; Table 5).The MEDS 3 and 4 group MI-free survival curves showed a significant divergence from that of the MEDS 0 to 1 group (Fig. 1).
When TVR was included as an outcome variable, the adjusted effect of a higher MEDS became nonsignificant (p > 0.05; Table 5). The MEDS was not associated with a higher risk of the composite end point of death, MI, or TVR.
Compared with patients without a history of a previous MI, patients with such a history had a considerably higher observed mortality rate (HR 3.4, 95% confidence interval [95% CI] 2.2 to 5.5), higher risk of death or MI (HR 2.9, 95% CI 1.9 to 4.4), and a higher incidence of death, MI, or TVR (HR 1.8, 95% CI 1.3 to 2.6). In a multivariable analysis, the effect of having a history of a previous MI on the outcomes became less pronounced, though still slightly significant (HR 1.74, CI 1.07 to 2.83 for death; HR 1.79, CI 1.17 to 2.72 for death or MI; and HR 1.46, 95% CI 1.04 to 2.03 for death, MI, or TVR).
In patients with a history of a previous MI, the effect of MEDS on outcomes in a multivariate analysis was in general more pronounced than in patients without a history of MI. The adjusted HR of death or MI for MEDS 4 versus 0 to 1 was 0.56 (95% CI 0.40 to 0.77) for patients with a history of previous MI and 0.79 (95% CI 0.53 to 1.18) for patients with no history of previous MI. Although the differential effect of MEDS is not statistically significant, there appears to be a trend towards more benefit derived from medical therapy with the former group.
The main findings of this study are: 1) patients with MEDS of 3 or 4, that is, discharged on three or four of the cardioprotective drug classes after successful PCI, have a lower observed mortality at 36 months than patients with a MEDS of 0 to 1; 2) in a multivariate analysis, higher MEDS were associated with a lower risk of death or MI; and 3) there was no association between medication use at discharge and the need for TVR.
A possible implication from our findings is that the use of a MEDS based on the number of cardioprotective medications used at hospital discharge may be used as a quality marker after PCI. The more medications from the four examined drug classes the patient was prescribed, the lower the risk of mortality and death or MI during follow up.
Correlation with other studies
Antiplatelet therapy, ACE inhibitors, beta-blockers, and lipid-lowering agents have all been shown to improve outcome in patients with coronary artery disease (2–5,7,10). After PCI, most also have been shown to reduce adverse cardiovascular outcome, although no beneficial effect on restenosis has been established.
In the Clopidogrel in Unstable Angina to Prevent Recurrent Events (CURE) trial, the long-term administration of clopidogrel after PCI was associated with a lower rate of cardiovascular death, MI, or any revascularization (relative risk, 0.7) (6). In an analysis of the Second Primary Angioplasty in Myocardial Infarction (PAMI-2) trial, use of beta-blockers in patients undergoing primary PCI for acute MI was independently associated with lower six-month mortality (2). Beta-blockers also were shown to be related to a marked long-term survival benefit after elective PCI (HR 0.63 at 1 year) (2). Similarly ACE inhibitors reduced the incidence of cardiac death, acute MI, and clinical heart failure in patients with stable angina and mildly depressed ejection fraction undergoing invasive revascularization (11). Other studies have similarly associated the use of statins to improved clinical outcomes after PCI (12,13).
In our study, patients with MEDS of 4 had a lower unadjusted mortality than patients with a lower score despite the fact that they were on average sicker and had more comorbid illnesses. After adjusting for multiple variables associated with a worse outcome (Table 4), a higher score predicted a lower risk of death or MI at a median of three years of follow-up. The difference between the higher score groups and the group with MEDS of 0 to 1 starts appearing early, and persists through follow up (Fig. 1). Although the difference between MEDS 4 and MEDS 3 or 2 was not statistically significant, there is some trend observed in the MI-free survival curves.
Effect on TVR
The statistical significance seen in the composite outcome of death or MI was lost when TVR was included in the outcome. The most likely explanation for this is that these cardioprotective medications do not reduce clinically significant restenosis rate and thus do not affect TVR rate. In fact, a few studies have examined the effect of ACE inhibitors, beta-blockers, and statins on restenosis after PCI. In particular, quinapril, carvedilol, pravastatin, and fluvastatin all failed in separate studies to prevent or delay the progression of angiographic restenosis at the target site after angioplasty (12–15). Their beneficial effect on mortality and MI is more likely related to their effect on the general atherosclerotic disease, which is not limited to the specific region targeted by the percutaneous intervention.
MEDS in patients with a history of previous MI
The association between medication use and the clinical outcomes is more pronounced in patients with a history of previous MI as opposed to those without such a history. However, the interaction is not statistically significant, so it cannot be solidly concluded that patients with a history of MI constitute the only group in which the MEDS is associated with outcome. It is possible that sicker patients in general derive the most benefit from medical therapy.
This study has limitations inherent to registry data. The use of medications after intervention was not randomized and, thus, resulted in study groups with different prognostic characteristics. Although the number of patients was large enough to allow us to correct for many clinical variables, there may have been some other factors unaccounted for in the patients’ medical history, other medication use, and angiographic data. However, one strength of our findings is that the unadjusted risk of death for the group with MEDS of 4 was lower than that for the other groups despite the fact that it generally was a sicker population that otherwise would be expected to have a higher baseline mortality. It is possible that some of the patients in the MEDS 0 to 1 group received too few medications because they were too sick to tolerate them. However, on average, this was not the case, as is clearly seen in Tables 1 and 2.
It was also unknown to us how compliant the patients were with the prescribed medications once outside the hospital. However, this should not affect the implication of our data that medication prescription profile at discharge is linked to the long-term clinical outcome.
In summary, this large observational study demonstrates that positive clinical outcomes after successful PCI are associated with the prescription of evidence-based medical therapies at discharge.
- Abbreviations and Acronyms
- angiotensin-converting enzyme
- coronary artery bypass graft surgery
- congestive heart failure
- confidence interval
- hazard ratio
- medication score
- myocardial infarction
- percutaneous coronary intervention
- target vessel revascularization
- Received March 16, 2005.
- Revision received June 23, 2005.
- Accepted June 27, 2005.
- American College of Cardiology Foundation
- Chan A.W.,
- Quinn M.J.,
- Bhatt D.L.,
- et al.
- Kernis S.J.,
- Harjai K.J.,
- Stone G.W.,
- et al.
- Fox K.A.,
- Mehta S.R.,
- Peters R.,
- et al.
- Mehta S.R.,
- Yusuf S.,
- Peters R.J.,
- et al.,
- Clopidogrel in Unstable angina to prevent Recurrent Events trial (CURE) Investigators
- Fonarow G.C.,
- French W.J.,
- Parsons L.S.,
- Sun H.,
- Malmgren J.A.
- Rogers W.J.,
- Canto J.G.,
- Lambrew C.T.,
- et al.
- Mukherjee D.,
- Fang J.,
- Chetcuti S.,
- Moscucci M.,
- Kline-Rogers E.,
- Eagle K.A.
- Kjoller-Hansen L.,
- Steffensen R.,
- Grande P.
- Herrmann J.,
- Lerman A.,
- Baumgart D.,
- et al.
- Bertrand M.E.,
- McFadden E.P.,
- Fruchart J.C.,
- et al.