Author + information
- Received March 7, 2001
- Revision received July 13, 2001
- Accepted August 9, 2001
- Published online November 15, 2001.
- Stéphane Rinfret, MD, MSc, FRCP(C)*,†,
- Cindy L Grines, MD, FACC§,
- Roberta S Cosgrove, MSc, MPH*,
- Kalon K.L Ho, MD, MSc, FACC*,†,
- David A Cox, MD, FACC∥,
- Bruce R Brodie, MD, FACC¶,
- Marie-Claude Morice, MD, FACC#,
- Gregg W Stone, MD, FACC**,
- David J Cohen, MD, MSc*,*,†,‡ (, )
- the Stent-PAMI Investigators
- ↵*Reprint requests and correspondence:
Dr. David J. Cohen, Cardiovascular Division, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, Massachusetts 02215 USA
The goal of this study was to compare the impact of primary stenting or percutaneous transluminal coronary angioplasty (PTCA) on health-related quality of life (HRQOL) in patients undergoing direct angioplasty for acute myocardial infarction (AMI).
Previous studies have demonstrated that coronary stenting reduces clinical and angiographic restenosis compared with PTCA. However, the impact of stenting on HRQOL from the patient’s perspective remains unknown.
We administered the Seattle Angina Questionnaire and the Medical Outcomes Study Short-form Survey at 1, 6 and 12 months after initial treatment to all North American patients in the Stent-Primary Angioplasty for Myocardial Infarction trial (Stent-PAMI) (n = 509)—a randomized trial comparing primary stenting to conventional PTCA for patients with AMI.
At one month, most HRQOL measures were similar for the two groups, but stent patients reported less bodily pain than PTCA patients (p = 0.03). At six-month follow-up, stenting resulted in significant improvements in several dimensions of HRQOL including reduced anginal frequency and bodily pain as well as improved disease perception (all p ≤ 0.03) and a trend towards better anginal stability (p = 0.056). By 12-month follow-up, however, none of these differences remained statistically significant. These differences in HRQOL were largely explained by the greater need for ischemia-driven target-vessel repeat revascularization procedures in PTCA patients during the first six months (16.0% vs. 6.2%, p < 0.001).
In patients undergoing revascularization for AMI, initial stent placement is associated with improvements in several dimensions of health status during the first six months of follow-up. In the absence of differences in mortality, these findings add to the overall argument in favor of initial stenting in patients treated with mechanical reperfusion for myocardial infarction.
Over the past decade, coronary stenting has been shown to reduce the rates of angiographic and clinical restenosis compared with conventional balloon angioplasty. Although these benefits were initially established in the setting of elective percutaneous coronary intervention (PCI) (1–3), recent studies have shown that these benefits extend to patients undergoing direct PCI for acute myocardial infarction (AMI) as well (4–6). As a result, stenting has emerged as the dominant form of percutaneous coronary revascularization. Currently, more than 70% of all PCIs performed in the U.S. involve placement of one or more coronary stents (7).
Despite their prominent position in the current practice of interventional cardiology, certain unresolved questions remain about the true value of coronary stenting (8). Although stents have been clearly demonstrated to reduce follow-up medical care costs compared with balloon angioplasty alone, overall medical care costs remain higher in most studies (9–11). Moreover, no randomized studies to date have demonstrated that coronary stenting improves short or long-term survival compared with conventional percutaneous transluminal coronary angioplasty (PTCA) (10–12). Thus, the ultimate value of coronary stenting is highly dependent on its impact on overall health-related quality of life (HRQOL).
Although several studies have examined HRQOL after stenting or balloon angioplasty, these studies are limited by their retrospective nature (13)or extremely high rates of treatment crossover (14). In addition, no studies have specifically examined the impact of stenting on quality of life (QOL) after AMI. To address these unresolved issues, we performed a prospective HRQOL substudy in the Stent-Primary Angioplasty for Myocardial Infarction trial (Stent-PAMI).
Between December 1996 and November 1997, 900 patients with AMI were randomized to routine coronary stent placement or primary PTCA as part of the Stent-PAMI trial. Details of the study protocol have been described previously (5). Briefly, patients presenting with AMI within 12 h of the onset of symptoms were referred for emergent coronary angiography. If angiography confirmed the presence of a high-grade native coronary stenosis suitable for either balloon angioplasty or stent implantation, the patient was eligible for randomization. The study protocol was approved by the institutional review board at each site, and all patients provided informed consent before enrollment.
Randomization and treatment protocol
Eligible patients were randomized to treatment with either conventional PTCA or the heparin-coated, Palmaz-Schatz stent (Cordis, Johnson and Johnson, Warren, New Jersey). Balloon angioplasty and stenting were performed according to standard techniques, including high-pressure postdilation after stent placement. Crossover from PTCA to stent placement was discouraged unless required to treat a severe dissection or an unacceptable PTCA result (>50% residual stenosis). After stent placement, patients received ticlopidine (250 mg twice a day) for one month. The remainder of each patient’s care was left to the discretion of the primary treating physician.
Assessment of in-hospital outcomes and clinical follow-up
Detailed case report forms including baseline demographics, clinical data and events during the initial hospitalization and follow-up were completed at each site and forwarded to the data-coordinating center (Cardialysis, Rotterdam, the Netherlands). All patients underwent clinical follow-up at 1, 6 and 12 months after initial treatment. Ischemia-driven target-vessel revascularization (TVR) was defined as any repeat revascularization of the original treated vessel that was required because of recurrent angina (Canadian Cardiovascular Society [CCS] class 2 to 4) or ischemia on functional testing. All end points including ischemia-driven TVR were adjudicated by an independent clinical events committee who were blinded to treatment assignment.
Health-related QOL was assessed at 1, 6 and 12 months after randomization using the Medical Outcomes Study (MOS) Short-form Survey (SF-36) and the Seattle Angina Questionnaire (SAQ) (15–18). Baseline health status was not assessed during the index hospitalization because of concerns about reliability of these instruments in the immediate post-myocardial infarction (MI) setting. The SF-36 is a generic health status instrument that evaluates eight dimensions of health including physical function, social function, mental health, general health perception, pain, energy and fatigue and two types of role function (15,16). The SAQ is a disease-specific instrument that measures five dimensions of QOL that are particularly important to patients with coronary artery disease (CAD): physical limitations due to CAD, anginal stability, anginal frequency, treatment satisfaction and disease perception (17,18). All SF-36 and SAQ domains are scaled from 0 to 100 points, with higher scores indicating better QOL (e.g., less pain, less frequent angina, better disease perception, etc.). Both instruments have undergone extensive validity and reliability testing and are sensitive to clinically meaningful differences in QOL (19). In our study population, the SF-36 demonstrated excellent internal consistency according to Cronbach’s α, with values ranging between 0.80 and 0.94 for all domains at 1, 6 and 12 months. Internal consistency for the SAQ was somewhat lower but still acceptable, with Cronbach’s α >0.62 for each subscale at each time point (20). Because the SAQ was validated only in its English version at the time of study initiation, only patients enrolled at U.S. and Canadian hospitals were eligible for the HRQOL substudy.
At each follow-up time point, patients were provided with a self-administered survey booklet and stamped return envelope. Any patient who failed to return the survey by the time of phone contact was given the survey by telephone (or in person at the time of the clinic visit).
Discrete data are reported as frequencies and continuous data as mean ± SD. Baseline characteristics and outcome variables were compared between treatment groups with the use of the two-sample ttest for continuous variables and Fisher exact test for categorical variables. Nonnormally distributed continuous data were compared using the Wilcoxon rank-sum test.
We used two-way analysis of variance to assess the extent to which HRQOL scores improved over time within the two treatment groups. Two-sample ttests were performed to compare HRQOL scores between treatment arms at each time point. Since each HRQOL domain was evaluated at three time points, we used the method of Benjamini and Hochberg (21)to adjust for multiple comparisons. Although our primary analyses were based on all of the available HRQOL data, extensive sensitivity analyses were performed using various imputation strategies to assess the impact of missing data on our findings. Since there were minor imbalances in several covariates at baseline, we also used multiple linear regression analyses to evaluate the treatment effect at six months while adjusting for baseline differences. All analyses were based on the intention-to-treat principle. All p values are two-tailed, and a p value ≤0.05 was considered statistically significant. All statistical analyses were performed with SAS statistical software, versions 6.12 and 8 (SAS Institute, Cary, North Carolina).
Of the 900 patients randomized in the Stent-PAMI trial, 554 were enrolled in North American centers and were eligible for the HRQOL substudy. Of these, 509 (92%) completed at least one QOL questionnaire and formed the population for the substudy. In general, responders and nonresponders were similar. However, questionnaire nonresponders were more likely to have a history of transient ischemic attack (11.1% vs. 4.0%, p = 0.01) or stroke (10.8 vs. 3.4%, p = 0.05) and had lower baseline ejection fractions (44% vs. 47%, p = 0.05). There were no differences in treatment allocation between responders and nonresponders.
Among the 509 responders, baseline characteristics were generally comparable between the two treatment groups (Table 1). Patients assigned to coronary stenting tended to be somewhat older than patients in the PTCA group (60.7 vs. 58.8 years, p = 0.07), had a slightly higher prevalence of hypercholesterolemia (46.4% vs. 37.8%, p = 0.08) and stroke (4.7% vs. 2.0%, p = 0.09) and a lower prevalence of chronic obstructive pulmonary disease (COPD) (2.0% vs. 5.5%, p = 0.06).
Mortality rates in the HRQOL substudy were lower than those in the overall clinical trial since patients who died within the first month of follow-up were not eligible for QOL evaluation. Otherwise, clinical outcomes in the HRQOL substudy closely paralleled the outcomes seen in the overall randomized trial (Table 2). Over the one-year follow-up period, there were no significant differences between the treatment groups in the incidence of death, nonfatal reinfarction or stroke. In contrast, patients randomized to initial stenting were less likely than those assigned to conventional PTCA to require repeat TVR because of recurrent angina or ischemia over the follow-up period (12.1% vs. 24.2%, p < 0.001). Of these repeat revascularization procedures, 64% were performed during the first six months of follow-up while 36% were performed during months 7 to 12.
Results of the QOL surveys at each follow-up time point are displayed graphically in Figures 1 and 2. ⇓⇓Health-related QOL scores improved over the follow-up period for three of five domains of the SAQ (anginal frequency, disease perception and physical limitations due to heart disease; all p < 0.001) and for seven of eight domains of the SF-36 (all except general health perception; all p ≤ 0.001). The only exception to this pattern was the results for the SAQ anginal stability score, which measures the extent of change in a patient’s disease status over the preceding month (50 = no change, 51 to 100 = improvement, 0 to 49 = decline). Thus, the high scores for this scale at one month reflect dramatic improvements in cardiovascular health relative to the time of enrollment in the study. Although scores tended to decline at months 6 and 12, they remained higher than 50 for both treatment groups, which indicates continued improvement in anginal stability, but to a lesser degree, over the succeeding months.
At the one-month time point, there were few differences in HRQOL between the treatment groups. The only exceptions were for the SF-36 bodily pain (mean difference = 7.0, p = 0.03) and mental health scales (mean difference = 4.4, p = 0.056), both of which favored the stent group. At the six-month time point, patients assigned to initial stenting had higher scores for multiple dimensions of HRQOL including reduced anginal frequency (mean difference = 5.7, p = 0.02), better disease perception (mean difference = 6.8, p = 0.03) and improved anginal stability (mean difference = 7.3, p = 0.056) according to the SAQ, as well as reduced bodily pain (mean difference = 7.0, p = 0.03) according to the SF-36. By 12-month follow-up, however, none of these differences remained statistically significant.
In order to adjust for slight imbalances between the stent and PTCA groups, we used multiple linear regression to examine the relationship between HRQOL at six months and treatment assignment. Of those factors that differed between the stent and PTCA groups at baseline (p < 0.10), only the presence of COPD was associated with QOL during follow-up. In multivariable analyses that controlled for the presence of COPD, randomization to coronary stenting remained independently associated with improved six-month QOL across all of the domains identified in the primary analyses.
Health-related QOL data were missing for 14% to 36% of our study population (mean = 22%) across the various domains and time points. There were no differences in the rates of missing data between the two treatment groups. To examine the robustness of our results, we performed a variety of sensitivity analyses in which we imputed missing data using the worst health status score, the best health status score, the mean value for the treatment group and a last-value-carried-forward approach. None of our principal findings were altered in any of these sensitivity analyses.
Since the major benefit of stenting in our patient population was a reduction in the need for clinically-driven TVR, we performed stratified analyses to explore the relationship between treatment assignment, TVR and HRQOL at the six-month time point (Table 3). The occurrence of ischemia-driven TVR during follow-up was associated with substantially worse HRQOL scores across virtually all domains, with differences ranging from 4.2 points for the SAQ treatment satisfaction subscale to 14.5 for the physical role function subscale of the SF-36. Nonetheless, among patients who did not require TVR during follow-up, stenting was still associated with improved six month anginal frequency, disease perception, anginal stability and bodily pain scores. We observed similar findings in favor of stenting in the subgroup of patients who did require TVR during follow-up as well. At one-year follow-up, freedom from ischemia-driven revascularization was associated with persistent HRQOL benefits across nearly all the SAQ and SF-36 domains (data not shown).
Although coronary stenting reduces restenosis compared with conventional PTCA, most studies have found that aggregate long-term costs remain somewhat higher with stenting regardless of the clinical setting (9–11,22). Thus, in the current era of constrained resources, justification for routine coronary stenting requires evidence of meaningful long-term benefits as assessed from the patient’s perspective. Since most studies of stenting have not demonstrated reduced long-term mortality compared with PTCA alone (10–12,23,24), the value of routine coronary stenting depends critically on its ability to improve HRQOL compared with alternative revascularization procedures.
In this study, we prospectively evaluated HRQOL in a large population of North American patients who were randomized to stenting or conventional PTCA as treatment for AMI. Using two well-validated instruments (the SF-36 health status instrument and the SAQ), we found that stenting produced substantial improvements in HRQOL compared with PTCA alone. These benefits were principally manifest at the six-month follow-up time point and were no longer apparent 12 months after initial treatment. Nonetheless, patients who required one or more repeat revascularization procedures during follow-up demonstrated persistent reductions in QOL across both general and cardiac-specific domains.
Comparison with previous studies
Two previous studies have attempted to assess the HRQOL benefits of stenting compared with PTCA for patients undergoing nonemergent PCI (13,14). Krumholz et al. (13)used the SF-36, Duke Activity Status Index and CCS anginal classification to examine HRQOL in 160 patients who were randomly assigned to stenting or PTCA as part of the Stent Restenosis study (13). At a median follow-up interval of 15 months, they found no major differences in HRQOL with the exception of the SF-36 pain scale, which favored stenting. However, this study was limited by its retrospective nature, modest sample size, lack of a disease-specific instrument and the prolonged interval between initial treatment and follow-up health assessment. More recently, the Optimum PTCA compared with Routine Stent Strategy (OPUS) trial used the SAQ to evaluate HRQOL six months after randomization to routine stenting or a provisional stenting strategy (14). In this prospective study of 431 patients, there were no differences in any of the SAQ subscales despite a significant reduction in TVR with stenting. Whether these differences between OPUS and our study are due to inherent differences in the patient populations (planned PCI vs. AMI intervention), the high rate of crossover from angioplasty to stent implantation in the provisional stent arm of OPUS (37%) or our larger sample size and greater statistical power is uncertain. Of note is that, similar to our study, the OPUS trial demonstrated substantial differences in six-month HRQOL between patients who did and did not require repeat revascularization during follow-up.
At the six-month time point, the differences in HRQOL between stenting and PTCA ranged from 5.7 points for the SAQ anginal frequency scale to 6.8 points and 7.0 points for the SAQ disease perception and the SF-36 bodily pain scales, respectively. On a population basis, such differences in health status are fairly large. For example, a cross-sectional study of population-based means by age group revealed that decrements of 4 to 7 points in SF-36 scores were associated with five years of aging (25). Moreover, a previous longitudinal study of patients with chronic stable angina found that differences in SAQ scores of approximately 8 points were clinically meaningful for all but the anginal stability subscale (18).
Examination of the individual items that comprise each subscale can also improve the interpretability of the observed differences in HRQOL. For example, in our study the 7 point difference on the SF-36 bodily pain subscale indicates that at six-month follow-up 25% of patients assigned to initial PTCA reported “moderate, severe or very severe pain during the last four weeks” as compared with only 16% of patients assigned to initial stenting (p = 0.03). Similarly, 23% of patients who had received PTCA had reported “moderate to extreme interference with normal work due to pain” compared with only 13% of patients receiving stents (p < 0.01). For the SAQ anginal frequency subscale, the observed 6 point difference at six-month follow-up indicates that 43% of patients who had received PTCA experienced one or more chest pain episodes during the month preceding the questionnaire compared with 26% of patients who had received stents (p < 0.01), and 18% of patients who had received PTCA reported that they have used some nitroglycerin over the same period (vs. 10% of patients receiving stents, p = 0.03). On the basis of these comparisons, it appears that the HRQOL benefits that were associated with stenting in Stent-PAMI are clinically relevant.
Mechanisms of benefit
The major clinical benefit of stenting compared with PTCA in our study was a reduction in the need for ischemia-driven TVR, which was apparent by six-month follow-up and sustained through one year. The major benefit of stenting on HRQOL was observed at the six-month time point as well. The timing of these benefits suggests that much of the HRQOL benefits of stenting relate to its ability to reduce restenosis and its associated complications. In this light, it is not surprising that the QOL benefits associated with stenting were no longer apparent at one-year follow-up. By this time point, it is likely that most episodes of symptomatic restenosis had already occurred and been successfully treated with repeat revascularization.
Nonetheless, our stratified analyses suggest that restenosis (or at least its clinical manifestation as repeat TVR) does not fully account for the observed differences in HRQOL between treatment groups. In fact, stenting was associated with slightly better QOL scores regardless of whether patients underwent ischemia-driven TVR or not (Table 3). There are several possible explanations for these findings. First, recurrence of anginal symptoms does not necessarily lead to repeat revascularization after initially successful PCI. Thus, it is possible that some of the HRQOL benefits associated with stenting in our study reflect cases of symptomatic restenosis that were felt (by either the patient or the treating physician) to be sufficiently mild as not to warrant repeat revascularization. Similarly, it is possible that among patients who required repeat revascularization, the extent of recurrent symptoms was less severe in the stent group. Finally, it is possible that some of the HRQOL benefits associated with stenting were due to a placebo effect in this nonblinded trial as suggested by better mental health and less bodily pain reported by stent patients at one month. However, a placebo effect seems unlikely to account for much of the six-month benefits we observed for the anginal frequency and disease perception domains, given the limited benefits of stenting on these subscales at the one-month time point.
This study has several important limitations. Baseline HRQOL data were not collected because of concerns that their assessment would not be valid in the immediate post-MI setting. Thus, we cannot exclude the possibility that despite randomization there were differences in baseline HRQOL that accounted for some of our findings. The fact that baseline characteristics were quite well-matched in our population and that our findings were unchanged in multivariable analyses that adjusted for minor imbalances in baseline characteristics tends to mitigate against such selection bias, however.
From an analytic standpoint, the use of multiple statistical tests increases the risk of false positive findings. To mitigate this effect, we reported p values adjusted for multiple HRQOL comparisons in our primary analyses. Moreover, the consistency of our findings across multiple HRQOL domains argues against type I error as an important explanation for our results. Missing data pose an additional challenge for studies of HRQOL, where they may occur nonrandomly. Not surprisingly, in our study HRQOL scores tended to be higher for those patients who responded at each of the three time points compared with patients who missed one or more evaluations. Nonetheless, missing data were relatively infrequent in our study, and our findings were unchanged when we restricted our analysis to complete responders (n = 342), as well as with various data imputation scenarios.
Finally, it is important to recognize that our findings apply only to patients undergoing primary angioplasty for AMI. It is possible that some of the benefits associated with stenting in this study reflect specific differences in post-MI recovery that may not be attained in the setting of elective PCI. On the other hand, the HRQOL benefits associated with reduced restenosis are similar to those previously reported in the elective setting (14)and may, in fact, have been blunted in certain patients since the target lesion supplied infarcted myocardium.
In patients undergoing primary angioplasty for AMI, initial stent placement is associated with significantly better HRQOL at six-month follow-up but no differences at one year. These findings were explained, in part, by the lower incidence of clinically significant restenosis in the stent group but were also independent of the restenosis benefits of stenting. In the absence of proven mortality benefits, these findings provide the major rationale for stenting in the AMI setting. Further studies are needed to determine whether the benefits of stenting on HRQOL extend to patients undergoing nonemergent PCI as well.
The authors thank John Orav, PhD, and Fran Cook, ScD, for their guidance with the statistical analyses.
☆ Dr. Rinfret was supported, in part, by a Detweiler Fellowship Grant from the Royal College of Physicians and Surgeons of Canada and by the Samuel McLaughlin Foundation. Dr. Cohen was supported, in part, by a Clinician-Scientist Award from the American Heart Association. Additional support was provided by an unrestricted grant from Cordis, Inc., Warren, New Jersey.
- acute myocardial infarction
- coronary artery disease
- Canadian Cardiovascular Society
- chronic obstructive pulmonary disease
- health-related quality of life
- myocardial infarction
- Optimum PTCA compared with Routine Stent Strategy trial
- percutaneous coronary intervention
- percutaneous transluminal coronary angioplasty
- quality of life
- Seattle Angina Questionnaire
- Medical Outcomes Study (MOS) Short-form Survey
- Stent-Primary Angioplasty for Myocardial Infarction trial
- target-vessel revascularization
- Received March 7, 2001.
- Revision received July 13, 2001.
- Accepted August 9, 2001.
- American College of Cardiology
- Suryapranata H,
- van’t Hof A.W,
- Hoorntje J.C,
- de Boer M.J,
- Zijlstra F
- Maillard L,
- Hamon M,
- Khalife K,
- et al.
- Williams D.O,
- Holubkov R,
- Yeh W,
- et al.
- Cohen D.J,
- Krumholz H.M,
- Sukin C.A,
- et al.
- Macaya C,
- Serruys P.W,
- Ruygrok P,
- et al.
- Ware J.E
- Spertus J.A,
- Winder J.A,
- Dewhurst T.A,
- et al.
- Streiner D.L,
- Norman G.R
- Cohen D.J,
- Taira D.A,
- Berezin R,
- et al.
- Taira D.A,
- Seto T.B,
- Ho K.K,
- et al.