Establishing the validity of appropriate use criteria (AUC) for percutaneous coronary intervention (PCI) in the setting of stable ischemic heart disease can support their adoption for quality improvement. We conducted a post hoc analysis of 2,287 Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation trial patients with stable ischemic heart disease randomized to PCI with optimal medical therapy (OMT) or OMT alone. Within appropriateness categories, we compared rates of death, myocardial infarction, revascularization subsequent to initial therapy, and angina-specific health status as determined by the Seattle Angina Questionnaire in patients randomized to PCI + OMT to those randomized to OMT alone. A total of 1,987 patients (87.9%) were mapped to the 2012 publication of the AUC, with 1,334 (67.1%) classified as appropriate, 551 (27.7%) uncertain, and 102 (5.1%) as inappropriate. There were no significant differences between PCI and OMT alone in the rate of mortality and myocardial infarction by appropriateness classification. Rates of revascularization were significantly lower in patients initially receiving PCI + OMT who were classified as appropriate (hazard ratio 0.65; 95% confidence interval 0.53 to 0.80; p <0.001) or uncertain (hazard ratio 0.49; 95% confidence interval 0.32 to 0.76; p = 0.001). Furthermore, among patients classified as appropriate by the AUC, Seattle Angina Questionnaire scores at 1 month were better in the PCI-treated group compared with the medical therapy group (80 ± 23 vs 75 ± 24 for angina frequency, 73 ± 24 vs 68 ± 24 for physical limitations, and 68 ± 23 vs 60 ± 24 for quality of life; all p <0.01), with differences generally persisting through 12 months. In contrast, health status scores were similar throughout the first year of follow-up in PCI + OMT patients compared with OMT alone in patients classified as uncertain or inappropriate. In conclusion, these findings support the validity of the AUC in efforts to improve health care quality through optimal use of PCI.
The appropriate use criteria (AUC) were developed through the collaborative efforts of multiple cardiovascular professional organizations to quantify the anticipated benefits of percutaneous coronary intervention (PCI), in terms of survival or health status outcomes, relative to the procedural risks for a given clinical scenario. Some have criticized the AUC for lacking evidence to support their validity, particularly as applied to patients with stable ischemic heart disease (SIHD). Providing empirical evidence to support the AUC ratings could enhance their use in supporting safer and potentially more cost-effective care. We leveraged the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) study, which randomized patients with SIHD to PCI and optimal medical therapy (OMT) or OMT alone, to assess the incremental benefits of PCI across different categories of appropriateness. Finding heterogeneity of health status treatment benefit with PCI that was greater in the more appropriate patients would validate the AUC as a tool for improving evidence-based and patient-centered care for patients with SIHD.
Methods
Details of the COURAGE trial have been described previously. Briefly, from 1999 to 2004, 2,287 of 3,071 eligible patients with stable coronary artery disease (CAD) were randomized at 50 US and Canadian centers to PCI with OMT or OMT alone. The study enrolled patients with at least a 70% diameter stenosis in 1 or more major epicardial coronary arteries and evidence of myocardial ischemia or a stenosis of at least 80% in 1 or more coronary arteries and classic angina pectoris without provocative testing. Patients were excluded if they had persistent Canadian Cardiovascular Society (CCS) class IV angina symptoms, a markedly positive stress test (substantial ST-segment depression or hypotensive response during stage 1 of the Bruce protocol), refractory heart failure or cardiogenic shock, the ejection fraction of <30%, revascularization in the previous 6 months, and coronary anatomy that was not amenable to PCI. Randomized patients were followed for a median 4.6 years (range 2.5 to 7.0 years).
Details regarding the methodology of AUC development have been described previously. In the AUC, PCI was considered “appropriate” for a clinical scenario when the expected benefits, in terms of survival or quality of life, exceeded the expected negative consequences of the procedure and “inappropriate” when the perceived risks outweighed potential benefits. Classification as “uncertain” in the AUC implies inadequate data to classify the balance of anticipated risk and benefit definitively. In patients with SIHD being considered for PCI, the principal determinants of procedural appropriateness were (1) extent of CAD, (2) ischemic risk, as determined by noninvasive testing, (3) intensity of antianginal therapy, and (4) symptom burden, as determined by CCS class. COURAGE was designed before to the first publication of AUC for PCI. We used baseline clinical and angiographic data to map COURAGE trial patients to an AUC scenario, which was then categorized as “appropriate,” “uncertain,” or “inappropriate” based on the 2012 publication of the AUC.
Ascertainment of clinical outcomes in the COURAGE trial has been previously described. The primary outcome measure was a composite of death from any cause and nonfatal myocardial infarction (MI). Clinical outcomes were adjudicated by an independent committee whose members were blinded to patients’ treatment assignments. COURAGE also collected information on patient-reported health status, which was assessed using the Seattle Angina Questionnaire (SAQ) at baseline and at the 1-, 3-, 6-, 12-, 24-, and 36-month follow-up visits. The SAQ is a 19-item questionnaire that quantifies the frequency of angina, any recent change in the severity of angina, physical limitations because of angina, satisfaction with treatment, and quality of life. Scores range from 0 to 100, with higher scores indicating better health status. For this study, we focused our analyses on the most relevant SAQ domains: angina frequency, physical limitation, and quality of life. A clinically significant change is defined as a difference of ≥20 on the angina frequency scale, ≥8 points on the physical limitation scale, and ≥16 on the quality of life scale.
To ensure balance between treatment strategies within each appropriateness category, we stratified COURAGE participants by their appropriateness class and compared baseline characteristics between PCI and OMT-only patients, including demographics, coronary anatomy, ischemic risk by noninvasive testing, medical therapy, symptom burden by CCS, and patient-reported health status of patients, within each AUC classification. Categorical variables were compared using frequencies and the chi-square tests or Fisher’s exact test, and means ± SD and t tests were used for continuous variables.
Within each appropriateness category (appropriate, uncertain, and inappropriate for PCI), we compared rates of death, nonfatal MI, and revascularization procedures subsequent to initial therapy in patients randomized to PCI + OMT to those randomized to OMT alone with Kaplan-Meier survival curves and log-rank tests. Similarly, within each appropriateness category, we evaluated whether there were differences in SAQ scores over time between patients randomized to PCI versus OMT alone. We compared SAQ scores for the angina frequency, physical limitations, and quality of life domains at baseline, 1, 3, 6, 12, 24, and 36 months using t tests. We then determined if SAQ scores were different for PCI versus OMT alone for different levels of appropriateness by test of interaction (appropriateness category by treatment) at these time points.
We also conducted repeated-measures analyses of the scores and the change from baseline of scores over time using maximum likelihood methods to estimate the average benefit of PCI over time. For these analyses, intermittent missing data (e.g., a missing observation at 6 but not at 12 months) were imputed using multivariate imputation by chained equations. We included in the models the squared and cubed effect of time and the interactions with treatment and appropriateness because of the rapid improvement in patients’ health status after randomization. A level of significance of p <0.01 was used for all analyses, consistent with all COURAGE post hoc analyses.
Results
Of the 2,287 patients randomized in the COURAGE trial, 1,987 (88%) could be mapped to the AUC at baseline, with most unmapped patients because of a lack of stress testing. Among patients mapped to the AUC, 986 (50%) were randomized to PCI + OMT as initial treatment and 1,001 (50%) to OMT alone. Of the patients who received PCI, 654 (66%) were classified as appropriate, 279 (28%) as uncertain, and 53 (5%) as inappropriate. Of the patients who received OMT alone, 680 (68%) were classified as appropriate, 272 (27%) as uncertain, and 49 (5%) as inappropriate.
As would be expected by randomization, patient characteristics were similar by treatment strategy within each appropriateness category ( Table 1 ). Although patients were similar across treatment strategies, within categories of appropriateness, patient characteristics differed across categories of appropriateness ( Supplementary Table 1 ). Compared with patients classified as inappropriate, those patients classified as appropriate were more likely to have multivessel disease, use ≥2 antianginal medications, and have at least CCS I angina. Furthermore, baseline SAQ scores for angina frequency, physical limitation, and quality of life were much lower, reflecting worse baseline health status in patients classified as appropriate.
| Characteristic | Appropriate | Uncertain | Inappropriate | ||||||
|---|---|---|---|---|---|---|---|---|---|
| PCI (n=654) | OMT (n=680) | P value | PCI (n=279) | OMT (n=272) | P value | PCI (n=53) | OMT (n=49) | P value | |
| Age, mean (SD) (years) | 62±10 | 63±10 | 0.33 | 61±10 | 60±10 | 0.71 | 61±11 | 60±10 | 0.32 |
| Men | 86% | 85% | 0.58 | 85% | 86% | 0.63 | 77% | 78% | 0.98 |
| White | 85% | 87% | 0.24 | 89% | 86% | 0.28 | 91% | 83% | 0.28 |
| Body mass index, mean (SD) (kg/m 2 ) | 30±5 | 30±5 | 0.65 | 29±5 | 30±5 | 0.46 | 29±4 | 29±5 | 0.48 |
| Diabetes mellitus | 35% | 36% | 0.68 | 28% | 34% | 0.14 | 22% | 41% | 0.04 |
| Heart failure | 5% | 5% | 0.65 | 6% | 2% | 0.02 | 6% | 4% | 0.71 |
| Chronic lung disease | 10% | 13% | 0.12 | 7% | 12% | 0.04 | 6% | 10% | 0.48 |
| Current smoker (<30 days) | 27% | 26% | 0.51 | 31% | 32% | 0.84 | 34% | 43% | 0.42 |
| Cancer (other than skin) | 5% | 4% | 0.58 | 4% | 6% | 0.33 | 15% | 0% | <0.01 |
| Family history of CAD | 55% | 51% | 0.16 | 54% | 52% | 0.65 | 58% | 55% | 0.83 |
| Hypertension | 70% | 69% | 0.66 | 63% | 65% | 0.52 | 45% | 55% | 0.32 |
| Hypercholesterolemia | 47% | 52% | 0.08 | 53% | 52% | 0.74 | 49% | 59% | 0.31 |
| Prior myocardial infarction | 36% | 36% | 0.77 | 37% | 42% | 0.23 | 37% | 44% | 0.46 |
| Prior percutaneous coronary intervention | 16% | 15% | 0.64 | 13% | 14% | 0.71 | 8% | 12% | 0.51 |
| Prior coronary artery bypass grafting | 10% | 11% | 0.71 | 9% | 7% | 0.48 | 2% | 4% | 0.61 |
| Prior Stroke or TIA | 9% | 9% | 0.71 | 7% | 6% | 0.60 | 2% | 4% | 0.61 |
| Peripheral Vascular Disease | 8% | 9% | 0.49 | 4% | 7% | 0.16 | 11% | 10% | 0.86 |
| Systolic blood pressure, mean (SD) (mmHg) | 134±21 | 133±20 | 0.28 | 131±19 | 131±18 | 0.79 | 132±18 | 130±17 | 0.57 |
| Diastolic blood pressure, mean (SD) (mmHg) | 74±12 | 74±11 | 0.69 | 74±11 | 75±10 | 0.27 | 74±11 | 74±10 | 0.92 |
| Estimated GFR, mean (SD) (mL/min/1.73 m 2 ) | 80±20 | 77±20 | <0.01 | 82±19 | 81±19 | 0.68 | 84±23 | 80±20 | 0.34 |
| Ejection fraction, mean (SD) (%) | 62±11 | 61±10 | 0.98 | 62±11 | 62±10 | 0.94 | 63±10 | 60±10 | 0.08 |
| No. of Coronary Arteries Narrowed | |||||||||
| 1 | 22% | 24% | 57% | 51% | 45% | 37% | |||
| 2 | 41% | 37% | 0.21 | 38% | 44% | 0.37 | 53% | 59% | 0.59 |
| 3 | 36% | 40% | 5% | 6% | 2% | 4% | |||
| Proximal Left Anterior Descending | 58% | 66% | 0.02 | 20% | 26% | 0.14 | 0% | 2% | 0.48 |
| Baseline medications | |||||||||
| β-blockers | 77% | 75% | 0.54 | 60% | 61% | 0.81 | 67% | 74% | 0.52 |
| Nitrates | 65% | 67% | 0.61 | 41% | 47% | 0.23 | 29% | 29% | 0.97 |
| ACE-I/ARBs | 52% | 52% | 0.84 | 53% | 51% | 0.61 | 43% | 66% | 0.05 |
| Statins | 68% | 68% | 0.92 | 63% | 66% | 0.44 | 53% | 58% | 0.67 |
| Calcium-channel blockers | 38% | 34% | 0.22 | 23% | 21.9% | 0.83 | 6.1% | 7.9% | 1.00 |
| ≥ 2 antianginal medications | 69% | 67% | 0.53 | 34% | 39.3% | 0.23 | 26.5% | 34.2% | 0.48 |
| Risk from stress test results | |||||||||
| Low | 2% | 1% | 4% | 4.0% | 36.6% | 51.3% | |||
| Intermediate | 54% | 50% | 0.35 | 86% | 88.6% | 0.64 | 63.4% | 48.7% | ND |
| High | 45% | 49% | 10% | 7.5% | 0% | 0% | |||
| Baseline health status | |||||||||
| CCS class | |||||||||
| None | 7% | 8% | 12% | 17% | 77% | 69% | |||
| I | 29% | 30% | 0.12 | 36% | 37% | 0.15 | 4% | 16% | ND |
| II | 35% | 39% | 41% | 39% | 19% | 14% | |||
| III | 29% | 24% | 11% | 7% | 0% | 0% | |||
| SAQ angina frequency score, mean ± SD (IQR) | 66±27 (40,90) | 66±26 (50,90) | 0.58 | 73±24 (60,100) | 76±24 (60,100) | 0.26 | 83±24 (75,100) | 85±23 (75,100) | 0.72 |
| SAQ physical limitation score, mean ± SD (IQR) | 64±25 (44,83) | 64±25 (44,83) | 0.88 | 70±24 (50,92) | 73±24 (53,94) | 0.29 | 78±20 (64,96) | 79±25 (69,100) | 0.88 |
| SAQ quality of life score, mean ± SD (IQR) | 50±25 (33,67) | 49±24 (33,67) | 0.36 | 52±26 (33,75) | 57±26 (42,83) | 0.05 | 64±25 (46,83) | 63±27 (42,83) | 0.86 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
