Abstract
Background
The PROMISE trial found that initial use of ≥64 detector-row CT angiography versus standard functional testing strategies (exercise ECG, stress nuclear, or stress echo) did not improve clinical outcomes in 10,003 stable symptomatic patients with suspected coronary artery disease (CAD) requiring noninvasive testing. Symptom burden and quality of life (QOL) were major secondary outcomes.
Methods and Results
We prospectively collected a battery of QOL instruments in 5985 patients at baseline and 6, 12, and 24 months post-randomization. The pre-specified primary QOL measures were the Duke Activity Status Index (DASI) and the Seattle Angina Questionnaire (SAQ) frequency and QOL scales. All comparisons were made as randomized. Baseline variables were well balanced in the 2982 patients randomized to CTA testing and 3003 to functional testing. The DASI improved substantially in both groups over the first 6 months following testing but we found no evidence for a strategy-related difference [mean difference (anatomic − functional) at 24 months of follow-up 0.1 (95% CI −0.9 to 1.1)]. Similar results were seen for the SAQ frequency scale (mean difference at 24 months −0.2, 95% CI −0.8 to 0.4) and quality of life scale (mean difference at 24 months −0.2, CI −1.3 to 0.9). None of the secondary QOL measures showed a consistent strategy-related difference.
Conclusions
In symptomatic patients with suspected CAD who required noninvasive testing, symptoms and QOL improved significantly. However, a strategy of initial anatomic testing, as compared with functional testing, did not provide an incremental benefit for QOL over two years of follow-up.
Clinical Trial Registration Information: clinicaltrials.gov. Identifier: NCT01174550.
Keywords: chest pain diagnosis, chest pain, patient reported outcome, quality of life, coronary artery disease
INTRODUCTION
Each year in the United States, over 10 million people present to emergency rooms and doctors’ offices with chest pain symptoms, over 40 million non-invasive cardiac tests are performed,1 and approximately 4 million people receive a new diagnosis of angina. Conventional clinical practice (“usual care”) routinely involves diagnostic imaging, often with stress testing, as part of the initial evaluation of many of these patients. Several factors, including the lack of high-quality clinical trial data on clinical or patient-reported outcomes to guide test selection combined with varying beliefs about the value of specific tests and different patterns of test availability result in highly variable, often redundant, patterns of testing employed in an inefficient search for “diagnostic certainty.”2 The advent of high quality coronary computed tomographic angiography (CTA) has, for the first time, provided clinicians with the ability to visualize the coronary arteries without the risks of invasive angiography. However, the benefits of this technology for patient care have remained controversial, and adoption of CTA into practice has been quite variable. In addition to the uncertainty about the effects of CTA relative to functional testing on clinical outcomes, very little information is available about the differential effects of these testing strategies on patient reported outcomes.
The PROspective Multicenter Imaging Study for Evaluation of Chest Pain (PROMISE) trial was the first large scale randomized trial to compare clinical outcomes from the initial use of coronary computed tomographic angiography (CTA) versus the usual care strategy of functional testing in patients with chest pain and low-intermediate coronary artery disease (CAD) risk. Comparison of quality of life outcomes from these two strategies was a major secondary objective of the PROMISE research program.
METHODS
Patient Population and Primary Clinical Results
Between July 27, 2010 and September 19, 2013, PROMISE enrolled 10,003 symptomatic outpatients without known CAD at 193 clinical sites in the US and Canada. Patients were eligible if their physician believed that non-urgent noninvasive stress testing was indicated for evaluation of their symptoms.3 Inclusion criteria were selected in part to achieve an intermediate-risk cohort and included men over 54 years old or 45 to 54 with at least one cardiac risk factor (diabetes, peripheral arterial disease, cerebrovascular disease, current or past tobacco use, hypertension, or dyslipidemia). Eligible women had to be more than 64 years old if no cardiac risk factors were present or 50 to 64 with one or more cardiac risk factors. Patients were excluded if they were unstable or if they had a history of CAD, CAD evaluation within the previous 12-month period, or clinically significant cardiac comorbidity. All patients provided informed consent, and study protocol approval was obtained from each site’s institutional review board or ethics committee.
Patients were randomized to either initial CTA (anatomic strategy) or initial functional testing with stress ECG, stress echo, or stress nuclear methods (67.3% stress nuclear, 22.5% stress echo and 10.2% stress electrocardiography [ECG]) and stratified by site and prespecified intended functional test if randomized to that arm.4 The choice of the functional test was left to the patient’s clinician, as was the interpretation of the test and all subsequent management following the randomized test. Median follow-up time was 25 months (interquartile range, 18 to 34 months). One-year follow-up was complete for 93.5% of patients. The primary study outcome, a composite of all-cause mortality, myocardial infarction, hospitalization for unstable angina, and major 72-hour complications of diagnostic tests or cardiovascular procedures (stroke, major bleeding, renal failure, anaphylaxis), occurred in 3.3% of the CTA arm and 3.0% of the functional testing arm (adjusted HR 1.04, 95% CI 0.83 to 1.29, p=0.75).3 None of the individual primary endpoint components or combinations of components was significantly different between groups. Prespecified subgroup analyses as randomized were consistent with the overall comparison.
Patient Selection and Health-Related Quality of Life Data Collection
Our original study design called for all randomized patients to be included in the quality of life study. Because of budget constraints, the QOL sample was revised to include all patients enrolled in the trial before July 1, 2012. Original projections determined that the PROMISE trial would have more than 90% power to detect a difference of ¼ standard deviation in the primary QOL endpoints with 500 patients per group. We did not attempt to predict the effect on study power of variations in the proportion of PROMISE patients without CAD, who would likely not show any treatment effect. Health-related QOL data were collected using interviews at baseline and at 6, 12, and 24 months post-randomization.4 Site coordinators were specially trained to administer the QOL baseline interviews. Follow-up interviews were conducted via telephone by trained interviewers from the Duke Clinical Research Institute’s Outcomes Research Group.
Quality of Life Measures
We used a battery of validated measures to provide a comprehensive but efficient assessment of QOL in PROMISE.5 No prior QOL data related to the testing strategies being compared was available to guide our design of this portion of the study. Two principal QOL measures were prespecified as believed most likely to show an effect: the Duke Activity Status Index (DASI) and the Seattle Angina Questionnaire (SAQ). The DASI is a 12-item instrument used to assess cardiac-related functional status.6 Item selection and weighting in DASI was performed using a cardiopulmonary exercise testing reference standard. Scores range from 0 (worst) to 58.2 (best) and reflect the ability of patients to do physical activities without difficulty or assistance in 12 domains. For an individual patient, a clinically significant change is considered to be 4 points or more.7
The Seattle Angina Questionnaire (SAQ) was collected to assess the impact of angina symptoms on QOL outcomes.8 All scales are scored 0 (worst) to 100 (best) and no summary score for the full SAQ instrument has been established. In clinical trial applications of the SAQ, the two scales most often used to look for a treatment effect have been the angina frequency scale and the quality of life scale. The SAQ angina frequency scale assesses the frequency of angina symptoms in the previous 4 weeks. Higher scores reflect fewer angina symptoms. The SAQ quality of life scale measures the effect of angina symptoms on patients’ perceptions of their quality of life, with higher scores being more favorable. Additional SAQ scales include the physical limitation scale, the angina stability scale, and the treatment satisfaction scale. For SAQ scales, 5 or more points represent a benchmark for a clinically significant change for an individual patient.8
Depressive symptoms were assessed using the 9-item Patient Health Questionnaire (PHQ-9), which has been used to make criteria-based diagnoses of depressive disorders and depression severity.9 The PHQ-9 is scored from 0 to 27, with 0 indicated no depressive symptoms and 27 indicating severe depressive symptoms. Scores from 15 to 19 indicate major depression with moderately severe symptoms, and scores greater than or equal to 20 indicate major depression with severe symptoms.
To provide an overview of functioning and well-being from a generic perspective, we also collected four individual scales from the Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36): mental health inventory (MHI-5), general health, social functioning, and vitality (energy/fatigue).10 These instruments asked patients to recall health status “during the past 4 weeks.” Scores range from 0 to 100 with higher scores representing better health status. A clinically significant change for a patient using this scoring system has not been established but can be approximated by a ¼ standard deviation, or 5 points or more.
We also collected the EQ-5D-3L, which is a generic instrument consisting of two parts: a 5-dimension single summary health status index and a self-rated visual analog scale that ranges from 0 (best imaginable health state) to 100 (worst imaginable health state).11
Work and Productivity Measures
Baseline assessments included the patient’s employment status, and follow-up assessments included work status and interval changes, including time lost from work. We also used the 6-item Stanford Presenteeism Scale to assess whether or not patients felt they experienced a decrease in their work productivity or below-normal work quality because of their chest pain or symptoms of coronary artery disease.12 The Stanford Presenteeism Scale is a generic instrument that uses a 5-point Likert scale ranging from “Strongly Agree” to “Strongly Disagree” to rate 6 job performance measures over the past month. The responses are then summed to produce a Total Score that ranges from 6 to 30, with higher scores indicating increased work productivity.
Statistical Analyses
Primary Analyses
All primary comparisons were performed “as randomized” to initial diagnostic testing strategy. Descriptive statistics included percentages for discrete variables, and medians with 25th to 75th percentiles plus means with standard deviations for continuous variables. The chi-square test was used for discrete variable comparisons. Strategy comparisons of continuous variables were performed using a mixed effects regression model to account for repeated measures within a patient, with strategy group, time point, and the respective QOL measure from the baseline questionnaire as predictor variables. Using PROC MIXED in SAS version 9.4, the 6-, 12-, and 24-month measurements within a patient were fitted using an unstructured covariance matrix. At each time point, estimated strategy differences, 95% confidence intervals, and p-values were obtained using estimate statements. These models assumed missing data were missing at random. Time and strategy groups were coded using a cell means model with eight binary variables representing the four QOL assessment points and two strategy groups.
All reported P values were two sided and the conventional Neyman-Pearson decision rule for rejecting the null hypothesis was used (p <0.05). No adjustment was made in significance levels for multiple comparisons.
As an aid to interpretation, we considered evidence supporting a clinically important QOL effect to be present if we saw either a consistent initial diagnostic testing strategy difference over multiple consecutive follow-up contacts or a concordant strategy difference in multiple related QOL measures.5
Subgroups
We compared QOL outcomes for the seven subgroups that were prespecified in the overall PROMISE trial analysis: age ≥65, sex, minorities, CAD risk equivalent (history of diabetes, peripheral arterial disease, or cerebrovascular disease), preselected functional test type, pretest estimated risk, and Diamond and Forrester and Coronary Artery Surgery Study risk score.
RESULTS
Baseline Characteristics and QOL Data Collection Rates
The first 5985 (59.8%) of the 10,003 patients enrolled in the PROMISE trial were also included in the QOL substudy. In this QOL cohort, 2982 patients were randomized to CTA and 3003 to functional testing. The baseline characteristics of the QOL subset were similar to the non-QOL subset (Supplemental Table 1) and were well balanced between the two treatment arms (Table 1). Overall, the mean age was 60.0 years, 52.0% were female, and 78.3% were non-Hispanic whites.
Table 1.
Baseline Characteristics of the Quality of Life Subset in the PROMISE Trial.
| Randomized Treatment Arm | |||
|---|---|---|---|
| Characteristic | Overall (N= 5985) |
Anatomic (N= 2982) |
Functional (N= 3003) |
| Mean age – years | 60.0 | 59.8 | 60.1 |
| Female sex (%) | 52.0 | 51.5 | 52.5 |
| Racial or ethnic minority (%) | 21.7 | 22.4 | 21.0 |
| Cardiac risk factor | |||
| Mean body mass index (kg/m2) | 29.7 | 29.6 | 29.9 |
| Hypertension (%) | 65.9 | 65.8 | 66.0 |
| Diabetes (%) | 21.2 | 21.2 | 21.2 |
| Dyslipidemia (%) | 67.9 | 67.3 | 68.4 |
| Family history of premature CAD (%) | 31.0 | 32.0 | 30.1 |
| Peripheral arterial or cerebrovascular disease (%) | 5.9 | 5.5 | 6.3 |
| CAD risk equivalent (%) | 25.4 | 24.9 | 25.8 |
| Metabolic syndrome (%) | 37.9 | 37.4 | 38.4 |
| Current or past tobacco use (%) | 51.5 | 51.6 | 51.4 |
| Sedentary lifestyle (%) | 51.4 | 51.6 | 51.3 |
| History of depression (%) | 21.1 | 20.2 | 22.0 |
| Relevant medication | |||
| Beta blocker (%) | 23.5 | 23.8 | 23.1 |
| ACE inhibitor or ARB (%) | 42.2 | 41.7 | 42.7 |
| Statin (%) | 44.7 | 44.5 | 44.8 |
| Aspirin (%) | 42.7 | 43.1 | 42.3 |
| Primary presenting symptom | |||
| Chest pain (%) | 73.3 | 74.2 | 72.5 |
| Dyspnea on exertion (%) | 15.4 | 14.8 | 15.9 |
| Other (%) | 11.3 | 11.0 | 11.6 |
| Type of angina | |||
| Typical (%) | 11.4 | 11.4 | 11.4 |
| Atypical (%) | 78.0 | 77.7 | 78.2 |
| Non-cardiac (%) | 10.6 | 10.9 | 10.4 |
CAD=coronary artery disease; ACE=angiotensin-converting enzyme; ARB=angiotensin II receptor blocker Race was self-reported on the clinical case report form.
From a total of 23,055 expected patient QOL contacts from baseline out to 24 months, 21,448 (93.0%) QOL questionnaires were collected. Expected contacts exclude deaths and withdrawals. The rate of missing QOL assessments did not differ by treatment group at baseline or any follow-up interval (Table 2).
Table 2.
Reasons for Missing Data in the PROMISE Trial Quality of Life Substudy.
| 5985 Randomized | ||
|---|---|---|
| Anatomic (N=2982) |
Functional (N=3003) |
|
| Baseline | 2960 (99.3%) questionnaire data
0 (0%) dead 22 (0.7%) missing
|
2963 (98.7%) questionnaire data
0 (0%) dead 40 (1.3%) missing
|
| Interim losses* | 3 died, 36 withdrawn | 6 died, 60 withdrawn |
| 6 Month | 2772/2943 (94.2%) questionnaire data 1 (0.03%) dead 170 (5.7%) missing
|
2700/2937 (91.9%) questionnaire data
6 (0.2%) dead 231 (7.9%) missing
|
| Interim losses | 4 died, 50 withdrawn | 9 died, 96 withdrawn |
| 12 Month | 2715/2889 (93.9%) questionnaire data 4 (0.1%) dead 170 (5.9%) missing
|
2575/2832 (90.9%) questionnaire data
1 (0.04%) dead 256 (9.0%) missing
|
| Interim losses | 16 died, 73 withdrew | 10 died, 153 withdrew |
| 24 Month | 2479/2800 (88.5%) questionnaire data 7 (0.25%) dead 314 (11.2%) missing
|
2284/2669 (85.6%) questionnaire data
4 (0.15%) dead 381 (14.3%) missing
|
Interim losses include deaths and number of withdrawn from previous assessment window.
Quality of Life Outcomes
Pre-specified Principal Patient-Reported Quality of Life Endpoints
Both groups showed an improvement relative to baseline in patient-reported cardiac-related physical functioning, as assessed using the DASI (Table 3). At baseline, the CTA group had a mean DASI score of 25.5, which improved to 30.4 at 6 months, 29.9 at 12 months, and 31.2 at 24 months. The functional testing group had a mean DASI score of 25.5 at baseline, 30.5 at 6 months, 30.7 at 12 months, and 31.1 at 24 months. After adjustment, between-group differences averaged 0.1 DASI points at 24 months of QOL follow-up (95% CI −0.9 to 1.1) and there were no statistically significant differences at any time point.
Table 3.
Pre-specified Primary Quality of Life Measures as Randomized in the PROMISE Trial.
| Anatomic (N=2982) |
Functional (N=3003) |
Difference Between Anatomic and Functional (95% CI) |
|
|---|---|---|---|
| Duke Activity Status Index (DASI) | |||
| 0-58.2 scale; 0=worst, 58.2=best | |||
| Baseline | |||
| Number of patients | 2923 | 2918 | |
| Median (Q1, Q3) | 21.5 (10.7, 38.2) | 22.9 (10.0, 39.5) | |
| Mean (standard deviation) | 25.5 (17.7) | 25.5 (17.8) | |
| 6 month | |||
| Number of patients | 2677 | 2563 | |
| Median (Q1, Q3) | 30.5 (14.8, 44.7) | 30.2 (15.2, 44.7) | |
| Mean (standard deviation) | 30.4 (17.8) | 30.5 (17.7) | −0.1 (−1.0 to 0.8) |
| 12 month | |||
| Number of patients | 2593 | 2396 | |
| Median (Q1, Q3) | 29.2 (13.5, 44.7) | 31.2 (15.2, 44.7) | |
| Mean (standard deviation) | 29.9 (17.8) | 30.7 (17.7) | −0.6 (−1.6 to 0.3) |
| 24 month | |||
| Number of patients | 2247 | 2002 | |
| Median (Q1, Q3) | 31.5 (16.0, 45.5) | 31.2 (16.0, 45.5) | |
| Mean (standard deviation) | 31.2 (17.4) | 31.1 (17.4) | 0.1 (−0.9 to 1.1) |
| Seattle Angina Questionnaire | |||
| Angina Frequency Scale | |||
| Higher scores = less frequent angina | |||
| Baseline | |||
| Number of patients | 2929 | 2933 | |
| Median (Q1, Q3) | 80 (70, 90) | 80 (70, 90) | |
| Mean (standard deviation) | 78.5 (16.9) | 78.2 (17.2) | |
| 6 month | |||
| Number of patients | 2726 | 2648 | |
| Median (Q1, Q3) | 100 (90, 100) | 100 (90, 100) | |
| Mean (standard deviation) | 93.1 (12.1) | 92.9 (13.1) | 0.2 (−0.4 to 0.9) |
| 12 month | |||
| Number of patients | 2672 | 2525 | |
| Median (Q1, Q3) | 100 (90, 100) | 100 (90, 100) | |
| Mean (standard deviation) | 94.0 (11.8) | 94.1 (11.6) | −0.1 (−0.7 to 0.5) |
| 24 month | |||
| Number of patients | 2413 | 2218 | |
| Median (Q1, Q3) | 100 (100, 100) | 100 (100, 100) | |
| Mean (standard deviation) | 95.0 (11.2) | 95.1 (10.8) | −0.2 (−0.8 to 0.4) |
| Quality of Life Scale | |||
| 0-100 scale with higher scores representing
better quality of life |
|||
| Baseline Number of patients Median (Q1, Q3) Mean (standard deviation) |
2866 58.3 (41.7, 75.0) 56.3 (22.7) |
2884 58.3 (41.7, 75.0) 56.0 (22.8) |
|
| 6 month Number of patients Median (Q1, Q3) Mean (standard deviation) |
2685 83.3 (75.0, 91.7) 80.4 (19.7) |
2581 83.3 (75.0, 100) 80.6 (20.1) |
−0.2 (−1.2 to 0.9) |
| 12 month | |||
| Number of patients | 2563 | 2387 | |
| Median (Q1, Q3) | 83.3 (75.0, 100) | 91.7 (75.0, 100) | |
| Mean (standard deviation) | 81.7 (19.1) | 82.4 (18.8) | −0.5 (−1.5 to 0.6) |
| 24 month | |||
| Number of patients | 2201 | 1997 | |
| Median (Q1, Q3) | 91.7 (75.0, 100) | 91.7 (75.0, 100) | |
| Mean (standard deviation) | 82.8 (18.8) | 83.0 (18.6) | −0.2 (−1.3 to 0.9) |
| Physical Limitation Score | |||
| 0-100 scale with higher scores representing
better physical functioning |
|||
| Baseline | |||
| Number of patients | 2814 | 2801 | |
| Median (Q1, Q3) | 86.1 (66.7, 97.2) | 86.1 (66.7, 97.2) | |
| Mean (standard deviation) | 78.6 (21.5) | 79.1 (22.0) | |
| 6 month | |||
| Number of patients | 2403 | 2297 | |
| Median (Q1, Q3) | 100.0 (94.4, 100) | 100.0 (94.4, 100) | |
| Mean (standard deviation) | 93.6 (14.6) | 94.0 (13.9) | −0.4 (−1.2 to 0.4) |
| 12 month | |||
| Number of patients | 2282 | 2141 | |
| Median (Q1, Q3) | 100 (97.2, 100) | 100 (97.2, 100) | |
| Mean (standard deviation) | 94.3 (14.1) | 95.0 (12.4) | −0.8 (−1.6 to -0.0) |
| 24 month | |||
| Number of patients | 1810 | 1986 | |
| Median (Q1, Q3) | 100 (97.2, 100) | 100 (100, 100) | |
| Mean (standard deviation) | 95.3 (12.1) | 94.8 (13.2) | −0.7 (−1.5 to 0.1) |
| Angina Stability | |||
| 0-100 scale, lower scores indicate more
frequent angina, higher scores less frequent angina, a score of 50 indicates no change in angina frequency at most strenuous level of activity |
|||
| Baseline | |||
| Number of patients | 2903 | 2909 | |
| Median (Q1, Q3) | 50.0 (25.0, 50.0) | 50.0 (25.0, 50.0) | |
| Mean (standard deviation) | 44.2 (23.9) | 43.8 (25.2) | |
| 6 month | |||
| Number of patients | 2716 | 2640 | |
| Median (Q1, Q3) | 50.0 (50.0, 50.0) | 50.0 (50.0, 50.0) | |
| Mean (standard deviation) | 54.4 (16.5) | 53.7 (15.5) | 0.7 (−0.1 to 1.6) |
| 12 month | |||
| Number of patients | 2666 | 2518 | |
| Median (Q1, Q3) | 50.0 (50.0, 50.0) | 50.0 (50.0, 50.0) | |
| Mean (standard deviation) | 52.2 (14.5) | 52.8 (14.8) | −0.7 (−1.5 to 0.1) |
| 24 month | |||
| Number of patients | 2406 | 2217 | |
| Median (Q1, Q3) | 50.0 (50.0, 50.0) | 50.0 (50.0, 50.0) | |
| Mean (standard deviation) | 95.0 (12.9) | 51.3 (12.0) | −0.1 (−0.8 to 0.6) |
| Treatment Satisfaction | |||
| 0-100 scale with higher scores indicating more
satisfaction with treatment |
|||
| Baseline | |||
| Number of patients | 2810 | 2819 | |
| Median (Q1, Q3) | 91.7 (75.0, 100) | 91.7 (75.0, 100) | |
| Mean (standard deviation) | 83.7 (20.6) | 83.4 (20.3) | |
| 6 month | |||
| Number of patients | 2667 | 2556 | |
| Median (Q1, Q3) | 100 (83.3, 100) | 100 (83.3, 100) | |
| Mean (standard deviation) | 89.4 (18.6) | 88.2 (20.9) | 1.4 (0.3 to 2.4) |
| 12 month | |||
| Number of patients | 2547 | 2367 | |
| Median (Q1, Q3) | 100 (83.3, 100) | 100 (83.3, 100) | |
| Mean (standard deviation) | 89.4 (18.4) | 88.7 (20.0) | 1.0 (−0.1 to 2.1) |
| 24 month | |||
| Number of patients | 2163 | 1939 | |
| Median (Q1, Q3) | 100 (83.3, 100) | 100 (83.3, 100) | |
| Mean (standard deviation) | 89.8 (17.6) | 89.4 (18.6) | 1.0 (−0.1 to 2.1) |
Continuous variables shown as medians (Q1, Q3) and means with standard deviations. 95% confidence intervals on treatment differences calculated as CTA minus functional testing.
Scores on the Seattle Angina Questionnaire angina frequency scale were similar in both groups at all time points throughout the trial (Table 3). At baseline, the CTA group had a mean score of 78.5, which improved to 93.1 at 6 months, 94.0 at 12 months, and 95.0 at 24 months. The functional testing group had a mean score of 78.2 at baseline, 92.9 at 6 months, 94.1 at 12 months, and 95.1 at 24 months. After adjustment, between-group differences averaged −0.2 points at 24 months of QOL follow-up (95% CI −0.8 to 0.4) and there were no statistically significant differences at any time point. The SAQ quality of life scale, reflecting patient satisfaction with aspects of their life affected by angina, showed improvement in both groups from baseline to initial follow-up with little change thereafter and with no differential pattern according to testing strategy group (Table 3). No treatment effect was evident in any of the other SAQ scales (Table 3).
Additional Patient-Reported Quality of Life Measures
Moderate to severe depressive symptoms were present in 18% to 19% of the study cohort at baseline and diminished to 11% to 12% by 24 months, but no differential effect of testing strategy was evident (Table 4). Other secondary QOL measures, including the SF-36 general health scale and the MHI-5, did not show any clinically significant treatment differences between groups at any follow-up interval (Supplemental Table 2).
Table 4.
PHQ-9 Results in the PROMISE Trial.
| CTA (N=2982) |
Functional (N=3003) |
P-value* | |
|---|---|---|---|
|
PHQ-9 (% with moderate-severe
depression,
score ≥15) |
|||
| Baseline | |||
| Number of patients | 2899 | 2895 | |
| Patients with moderate to severe depression (%) | 19.2% | 18.2% | 0.34 |
| 6 month | |||
| Number of patients | 2669 | 2550 | |
| Patients with moderate to severe depression (%) | 14.8% | 13.1% | 0.07 |
| 12 month | |||
| Number of patients | 2585 | 2392 | |
| Patients with moderate to severe depression (%) | 13.0% | 11.5% | 0.12 |
| 24 month | |||
| Number of patients | 2238 | 2001 | |
| Patients with moderate to severe depression (%) | 11.5% | 10.7% | 0.45 |
| PHQ-9 Composite Score |
Difference Between
Anatomic
and Functional (95% CI) |
||
| 0 (absence of depressive symptoms) to 27
(severe depressive symptoms) scale |
|||
| Baseline | |||
| Number of patients | 2899 | 2895 | |
| Median (Q1, Q3) | 4 (2, 8) | 4 (1, 8) | |
| Mean (standard deviation) | 5.6 (5.2) | 5.4 (5.4) | |
| 6 month | |||
| Number of patients | 2669 | 2550 | |
| Median (Q1, Q3) | 3 (1, 6) | 3 (1, 6) | |
| Mean (standard deviation) | 4.3 (5.0) | 4.5 (4.9) | 0.1 (−0.2 to 0.3) |
| 12 month | |||
| Number of patients | 2585 | 2392 | |
| Median (Q1, Q3) | 3 (1, 6) | 2 (1, 5) | |
| Mean (standard deviation) | 4.2 (4.9) | 3.9 (4.6) | 0.2 (−0.0 to 0.5) |
| 24 month | |||
| Number of patients | 2238 | 2001 | |
| Median (Q1, Q3) | 2 (0, 5) | 2 (0, 5) | |
| Mean (standard deviation) | 3.8 (4.7) | 3.7 (4.5) | 0.1 (−0.2 to 0.4) |
CTA=computed tomography angiography
P-values were calculated from the chi-square test.
Diagnostic testing strategy groups are as randomized.
95% confidence intervals on treatment differences calculated as CTA minus functional testing.
Mean EQ-5D scores were 72.0 at baseline in the CTA group, 75.1 at 6 months, 74.7 at 12 months, and 73.2 at 24 months (Supplemental Table 3). In the functional testing group, mean EQ-5D scores were 72.4 at baseline, 75.0 at 6 months, 74.9 at 12 months, and 72.7 at 24 months (Supplemental Table 3).
Treatment Changes Following Testing
Following a positive PROMISE diagnostic test, modest differential changes were seen in some aspects of medical therapy. In particular, a positive CTA result (i.e., CTA showing obstructive CAD [defined as ≥50% epicardial vessel stenosis]; found in 13% of patients with interpretable tests) was associated with more initiation of aspirin, statins, and beta blockers than a positive functional study (ST segment changes consistent with ischemia in a stress ECG, inducible ischemia in at least one coronary territory during stress test imaging, or early termination due to ischemic ST segment changes; found in 12% of patients with interpretable tests): aspirin initiation 28.3% CTA, 19.0% for functional testing; statin initiation 28.3% for CTA, 19.9% for functional testing; beta blocker initiation 28.9% for CTA, 17.3% for functional testing (Supplemental Table 4). No effect of testing strategy on ACE-inhibitor/ARB use was seen. Negative tests were associated with low rates of medication discontinuation without a difference between testing strategies.
Following a positive test, use of invasive catheterization with revascularization was increased for the CTA arm relative to the functional testing arm: 41.6% for CTA, 23.8% for functional testing (Supplemental Table 5). Catheterization without revascularization was modestly lower in the CTA arm: 24.2% versus 28.0% for functional testing. Invasive catheterization rates were low and not different following a negative PROMISE diagnostic test.
Pre-specified Subgroup Analysis
There were no consistent treatment effects on QOL in the pre-specified subgroups. CTA showed a trend toward improved QOL outcomes in patients with high pre-test estimated risk, but no such trend was seen in patients with high risk according to the Diamond-Forrester/CASS risk score (Figures 1, 2, and 3).
Figure 1.
Subgroup analysis of the Duke Activity Status Index. CAD equivalent was defined as diabetes, peripheral vascular disease, or cerebrovascular disease.
Figure 2.
Subgroup analysis of the Seattle Angina Questionnaire Anginal Frequency scale. CAD equivalent was defined as diabetes, peripheral vascular disease, or cerebrovascular disease.
Figure 3.
Subgroup analysis of the Seattle Angina Questionnaire Quality of Life scale. CAD equivalent was defined as diabetes, peripheral vascular disease, or cerebrovascular disease.
Work Status and Presenters
At baseline, 47.5% of patients reported working full time in the CTA group versus 45.9% in the functional testing group, and 9.3% of the CTA group was working part time versus 9.7% in the functional testing group (Supplemental Table 6). Approximately 91% of working patients planned to return to work. At 6 months, 2.5% of the CTA group and 1.5% of the functional testing group had stopped working because of their health (Supplemental Table 7). At 24 months, 42.1% of patients in the CTA group were working full time versus 38.9% in the functional testing group, and 8.9% of the CTA group was working part time versus 10.1% in the functional group. Self-assessed effectiveness in the workplace (presenteeism) improved from baseline in both groups with little difference between the groups (Supplemental Table 8).
DISCUSSION
A key secondary objective of the PROMISE research program was to compare the effects of an “anatomic” strategy using CTA and a “functional testing” strategy using a clinician-directed mix of stress nuclear, stress echo, and stress ECG modalities on patient-reported outcomes, including presenting symptoms, quality of life, and return to work. When PROMISE was designed, we had no prior empirical research to help guide our expectations about the type and magnitude of quality of life changes that might be created by the differences between these two approaches to testing in the target stable chest pain population. In a subset of almost 6000 PROMISE patients, we found clinically important improvements in quality of life over the first 6 months following enrollment for both strategies, with little subsequent change out to 2 years. However, we did not detect any evidence that the CTA testing strategy had a measurable differential effect on these outcomes at the cohort level. For generic measures where normative population data exists (the three SF-36 scales, Supplemental Table 2), values had reached population norms by 6 months. Conceptually, the primary mechanism for a diagnostic test to drive QOL changes is through increased deployment of symptomatically active therapies in appropriate patients. One notable finding is that doctors made major therapeutic changes infrequently, even following a positive test. Despite similar overall rates of test positivity (13% for CTA, 12% for functional testing), a positive CTA test was followed by modestly more frequent initiation of aspirin, statins, and beta-blockers and more frequent invasive catheterization with revascularization. The magnitude of these changes in absolute patient numbers was small. Of the medication shifts, only the beta-blocker differences might be expected to have altered QOL through an anti-anginal effect, but the absolute difference in numbers of patients shifted (1.7 patients per 100) was likely too small to have a detectable effect. The absolute increment in patients undergoing revascularization after a positive test was similarly small (1.7 patients out of 100).
Our finding that the results of noninvasive testing in chest pain patients have an unexpectedly modest impact on patient management are very consistent with a more detailed examination of this issue reported by the Study of Myocardial Perfusion and Coronary Artery Imaging Roles in Coronary Artery Disease (SPARC).13 Both the modest overall effects of positive test results and the incremental differences between CTA and functional testing for medical therapy and invasive catheterization seen in our study were also seen in that multicenter registry of subjects with intermediate to high likelihood of CAD.
Since PROMISE was designed, a few studies have addressed the patient reported outcomes associated with the use of CTA versus functional testing. A multicenter pilot study of 180 stable chest pain patients compared CTA versus stress imaging and found no difference in the SAQ scales at a mean of 55 days.14 An earlier observational claims data analysis suggested a 2 per 100 lower rate of angina for CTA patients.15 Four recent studies are also relevant to PROMISE. The Scottish Computed Tomography of the Heart (SCOT-HEART) randomized trial showed that CTA added to standard care in the evaluation of chest pain (including stress ECG in most and stress radionuclide imaging in about 10%) did not alter the SAQ Angina Frequency scale at 6 weeks relative to usual care alone, despite having a significant effect on the diagnostic and therapeutic strategies employed.16 No difference was noted in rates of hospital admission for chest pain. The Angina Frequency scores at 6 weeks in this study (mean values about 80 where 0 is worst and 100 is no angina) were lower than the PROMISE mean 6-month Angina Frequency scores (about 93). Whether the less complete resolution of the chest pain symptoms in SCOT-HEART is a function of a different population, the different testing strategies used, the different time point of assessment, or different downstream management cannot be determined at present. An observational comparison of CTA with fractional flow reserve (FFR) calculation (FFRCT) versus usual care in 204 intermediate likelihood chest pain patients in Europe (Prospective Longitudinal Trial of FFRCT: Outcomes and Resource Impacts [PLATFORM]) found improved quality of life at 90 days with the FFRCT strategy: 19.5 points on the SAQ 7-item short form versus 11.4 points for the usual care arm.17 The authors indicated that they could not explain this magnitude of differential improvement with FFRCT. Given that the difference in the use of PCI was about 6 per 100 more with FFRCT while the use of CABG was 1 per 100 more with usual care, differential use of revascularization alone is not adequate as a mechanism to explain this relatively large difference in the SAQ. The Cost Effectiveness of Noninvasive Cardiac Testing (CeCAT) study randomized 898 patients with stable chest pain referred for invasive angiography at one U.K. hospital to one of four initial diagnostic strategies: stress SPECT, stress CMR, stress echocardiography, or invasive coronary angiography.18 The EQ-5D was used to assess QOL at baseline and at 6 and 18 months after randomization. Absolute values were modestly higher than were seen in the present study and were not different by diagnostic strategy. The Cardiac CT for the Assessment of Pain and Plaque (CAPP) trial randomized 500 patients with suspected nonacute CAD to exercise ECG treadmill testing (Bruce protocol) or cardiac CTA at two sites in Northern Ireland.19 The exercise ECG test arm had substantially more inconclusive test results that were followed by a second diagnostic test, most often myocardial perfusion imaging, and the CTA arm had more patients who went on to receive invasive coronary angiography and PCI. These patterns are consistent with what we observed in PROMISE. CAPP also measured QOL and found more improvement in several components of the SAQ in the CTA arm than in the exercise ECG arm. However, interpretation of these results are made difficult by baseline imbalances (for angina stability and quality of life subscales) and by a lack of durability of the differences over the 12-month follow-up.
To change patient outcomes, a diagnostic test must pass two sequential hurdles: 1) it must provide novel actionable information about the patient not available from the comparator care strategy, and 2) doctors must actually use this testing information to change care in ways that are known to affect outcomes. When researchers have considered this problem in the context of decision models comparing CTA with functional testing, the models were generally constructed assuming that quality of life differences would be created through differential relief of angina with medical therapy, PCI and CABG and through prevention of non-fatal cardiac events.20, 21 PROMISE is the first study to provide empirical data on these questions in a large scale randomized trial context. Conceptually, three physician actions can be postulated by which CTA related patient outcomes in this study might have differed from functional testing outcomes: 1) more effective targeting of medical therapy and revascularization to patients whose symptoms would be most likely to respond, 2) more persuasive reassurance provided to patients without obstructive CAD, 3) earlier initiation of secondary prevention in patients seen by CTA to have atherosclerotic arterial disease leading to prevention of quality of life-impairing events. Of these, the first would generally be expected to account for most of any clinically important outcome differences observed. However, as was also seen in SCOT-HEART and PLATFORM, the differences in revascularization due to CTA versus functional testing/usual care in PROMISE were not large enough to cause significant shifts in population level quality of life scores.3
The present analysis extends our understanding by showing that most stable low-risk patients with chest pain become asymptomatic or minimally symptomatic regardless of testing strategy. These results are also consistent with a significant body of research showing that medical interventions often have a much more modest impact on quality of life outcomes than physicians expect.
Three caveats should be considered in the interpretation of our study. First, QOL data collection was truncated earlier than initially planned due to budgetary constraints. We did not see any differences in the clinical characteristics of the final 4000 PROMISE patients that would raise concerns that our findings would have been different if we had collected data on all 10,000 PROMISE patients. Second, PROMISE was designed to enroll a low risk stable outpatient population with chest pain and many of these patients turned out not to have significant obstructive CAD. The trial results might have been different had a higher risk population been studied. Finally, PROMISE did not include a strategy of “no testing” so we cannot calculate how much the diagnostic testing per se contributed to the significant early quality of life improvements we observed.
In conclusion, in symptomatic stable patients with suspected CAD who were felt by their doctors to require noninvasive testing, chest pain symptoms and quality of life improved significantly. However, a strategy of initial anatomic testing, as compared with functional testing, did not provide an incremental benefit for quality of life over two years of follow-up.
Supplementary Material
Clinical Perspectives.
Over 10 million people present to emergency departments and doctors’ offices with chest pain symptoms each year. Coronary computed tomographic angiography (CTA) offers clinicians the ability to visualize the coronary arteries of their patients without the risks of invasive angiography. The PROspective Multicenter Imaging Study for Evaluation of Chest Pain (PROMISE) trial compared clinical outcomes from the initial use of coronary CTA versus functional (stress) testing in 10,003 patients with chest pain and low-intermediate coronary artery disease risk and found no difference in clinical outcomes between the testing strategies. Patient-reported outcomes related to quality of life were also assessed as part of the trial. We found no significant differences in quality of life outcomes between groups through 2 years of follow-up. Both diagnostic-testing strategies resulted in similar improvements in quality of life and there were few differences in patient care seen after a positive test result. Both groups reported decreased quality of life at baseline, and both groups showed high levels of chest pain resolution by 6 months. This quality of life substudy demonstrates for the first time that the selection of diagnostic tests for patients with stable chest pain may be made based upon availability and clinician and patient preference without significantly impacting patient reported outcomes.
Acknowledgments
We are particularly indebted to the coordinators at the PROMISE sites that collected data and to the patients who agreed to provide their quality of life data for the trial.
Funding Sources: This work was supported by grants R01HL098237, R01HL098236, R01HL98305, and R01HL098235 from the National Heart, Lung, and Blood Institute. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Heart, Lung, and Blood Institute or the National Institutes of Health.
Footnotes
Journal Subject Terms: Computerized Tomography (CT); Diagnostic Testing; Exercise Testing; Echocardiography; Nuclear Cardiology and PET
Disclosures: None.
References
- 1.Mark DB, Shaw LJ, Lauer MS, O'Malley PG, Heidenreich P. 34th Bethesda Conference: Task force #5--Is atherosclerosis imaging cost effective? J Am Coll Cardiol. 2003;41:1906–17. doi: 10.1016/s0735-1097(03)00362-0. [DOI] [PubMed] [Google Scholar]
- 2.Kassirer JP. Our stubborn quest for diagnostic certainty. A cause of excessive testing. N Engl J Med. 1989;320:1489–91. doi: 10.1056/NEJM198906013202211. [DOI] [PubMed] [Google Scholar]
- 3.Douglas PS, Hoffmann U, Patel MR, Mark DB, Al-Khalidi HR, Cavanaugh B, Cole J, Dolor RJ, Fordyce CB, Huang M, Khan MA, Kosinski AS, Krucoff MW, Malhotra V, Picard MH, Udelson JE, Velazquez EJ, Yow E, Cooper LS, Lee KL, PROMISE Investigators Outcomes of anatomical versus functional testing for coronary artery disease. N Engl J Med. 2015;372:1291–300. doi: 10.1056/NEJMoa1415516. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Douglas PS, Hoffmann U, Lee KL, Mark DB, Al-Khalidi HR, Anstrom K, Dolor RJ, Kosinski A, Krucoff MW, Mudrick DW, Patel MR, Picard MH, Udelson JE, Velazquez EJ, Cooper L, the PROMISE Investigators PROspective Multicenter Imaging Study for Evaluation of chest pain: rationale and design of the PROMISE trial. Am Heart J. 2014;167:796–803 e1. doi: 10.1016/j.ahj.2014.03.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Mark DB. Assessing quality-of-life outcomes in cardiovascular clinical research. Nat Rev Cardiol. 2016 Feb 18; doi: 10.1038/nrcardio.2016.10. doi: 10.1038/nrcardio.2016.10. [Epub ahead of print] [DOI] [PubMed] [Google Scholar]
- 6.Hlatky MA, Boineau RE, Higginbotham MB, Lee KL, Mark DB, Califf RM, Cobb FR, Pryor DB. A brief self-administered questionnaire to determine functional capacity (the Duke Activity Status Index) Am J Cardiol. 1989;64:651–4. doi: 10.1016/0002-9149(89)90496-7. [DOI] [PubMed] [Google Scholar]
- 7.Hlatky MA, Rogers WJ, Johnstone I, Boothroyd D, Brooks MM, Pitt B, Reeder G, Ryan T, Smith H, Whitlow P, Wiens R, Mark DB. Medical care costs and quality of life after randomization to coronary angioplasty or coronary bypass surgery. Bypass Angioplasty Revascularization Investigation (BARI) Investigators. N Engl J Med. 1997;336:92–9. doi: 10.1056/NEJM199701093360203. [DOI] [PubMed] [Google Scholar]
- 8.Spertus JA, Winder JA, Dewhurst TA, Deyo RA, Prodzinski J, McDonell M, Fihn SD. Development and evaluation of the Seattle Angina Questionnaire: a new functional status measure for coronary artery disease. J Am Coll Cardiol. 1995;25:333–41. doi: 10.1016/0735-1097(94)00397-9. [DOI] [PubMed] [Google Scholar]
- 9.Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001;16:606–13. doi: 10.1046/j.1525-1497.2001.016009606.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ware JE, Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care. 1992;30:473–83. [PubMed] [Google Scholar]
- 11.The EuroQoL Group EuroQol--a new facility for the measurement of health-related quality of life. Health Policy. 1990;16:199–208. doi: 10.1016/0168-8510(90)90421-9. [DOI] [PubMed] [Google Scholar]
- 12.Koopman C, Pelletier KR, Murray JF, Sharda CE, Berger ML, Turpin RS, Hackleman P, Gibson P, Holmes DM, Bendel T. Stanford presenteeism scale: health status and employee productivity. J Occup Environ Med. 2002;44:14–20. doi: 10.1097/00043764-200201000-00004. [DOI] [PubMed] [Google Scholar]
- 13.Hachamovitch R, Nutter B, Hlatky MA, Shaw LJ, Ridner ML, Dorbala S, Beanlands RS, Chow BJ, Branscomb E, Chareonthaitawee P, Weigold WG, Voros S, Abbara S, Yasuda T, Jacobs JE, Lesser J, Berman DS, Thomson LE, Raman S, Heller GV, Schussheim A, Brunken R, Williams KA, Farkas S, Delbeke D, Schoepf UJ, Reichek N, Rabinowitz S, Sigman SR, Patterson R, Corn CR, White R, Kazerooni E, Corbett J, Bokhari S, Machac J, Guarneri E, Borges-Neto S, Millstine JW, Caldwell J, Arrighi J, Hoffmann U, Budoff M, Lima J, Johnson JR, Johnson B, Gaber M, Williams JA, Foster C, Hainer J, Di Carli MF, Investigators S Patient management after noninvasive cardiac imaging results from SPARC (Study of myocardial perfusion and coronary anatomy imaging roles in coronary artery disease) J Am Coll Cardiol. 2012;59:462–74. doi: 10.1016/j.jacc.2011.09.066. [DOI] [PubMed] [Google Scholar]
- 14.Min JK, Koduru S, Dunning AM, Cole JH, Hines JL, Greenwell D, Biga C, Fanning G, LaBounty TM, Gomez M, Horowitz JM, Hadimitzsky M, Hausleiter J, Callister TQ, Rosanski AR, Shaw LJ, Berman DS, Lin FY. Coronary CT angiography versus myocardial perfusion imaging for near-term quality of life, cost and radiation exposure: a prospective multicenter randomized pilot trial. J Cardiovasc Comput Tomogr. 2012;6:274–83. doi: 10.1016/j.jcct.2012.06.002. [DOI] [PubMed] [Google Scholar]
- 15.Min JK, Kang N, Shaw LJ, Devereux RB, Robinson M, Lin F, Legorreta AP, Gilmore A. Costs and clinical outcomes after coronary multidetector CT angiography in patients without known coronary artery disease: comparison to myocardial perfusion SPECT. Radiology. 2008;249:62–70. doi: 10.1148/radiol.2483071453. [DOI] [PubMed] [Google Scholar]
- 16.SCOT-HEART Investigators CT coronary angiography in patients with suspected angina due to coronary heart disease (SCOT-HEART): an open-label, parallel-group, multicentre trial. Lancet. 2015;385:2383–91. doi: 10.1016/S0140-6736(15)60291-4. [DOI] [PubMed] [Google Scholar]
- 17.Hlatky MA, De Bruyne B, Pontone G, Patel MR, Norgaard BL, Byrne RA, Curzen N, Purcell I, Gutberlet M, Rioufol G, Hink U, Schuchlenz HW, Feuchtner G, Gilard M, Andreini D, Jensen JM, Hadamitzky M, Wilk A, Wang F, Rogers C, Douglas PS, the PLATFORM Investigators Quality-of-life and economic outcomes of assessing fractional flow reserve with computed tomography angiography: PLATFORM. J Am Coll Cardiol. 2015;66:2315–23. doi: 10.1016/j.jacc.2015.09.051. [DOI] [PubMed] [Google Scholar]
- 18.Thom H, West NE, Hughes V, Dyer M, Buxton M, Sharples LD, Jackson CH, Crean AM, CECaT study group Cost-effectiveness of initial stress cardiovascular MR, stress SPECT or stress echocardiography as a gate-keeper test, compared with upfront invasive coronary angiography in the investigation and management of patients with stable chest pain: mid-term outcomes from the CECaT randomised controlled trial. BMJ Open. 2014;4:e003419. doi: 10.1136/bmjopen-2013-003419. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.McKavanagh P, Lusk L, Ball PA, Verghis RM, Agus AM, Trinick TR, Duly E, Walls GM, Stevenson M, James B, Hamilton A, Harbinson MT, Donnelly PM. A comparison of cardiac computerized tomography and exercise stress electrocardiogram test for the investigation of stable chest pain: the clinical results of the CAPP randomized prospective trial. Eur Heart J Cardiovasc Imaging. 2015;16:441–8. doi: 10.1093/ehjci/jeu284. [DOI] [PubMed] [Google Scholar]
- 20.Ladapo JA, Jaffer FA, Hoffmann U, Thomson CC, Bamberg F, Dec W, Cutler DM, Weinstein MC, Gazelle GS. Clinical outcomes and cost-effectiveness of coronary computed tomography angiography in the evaluation of patients with chest pain. J Am Coll Cardiol. 2009;54:2409–22. doi: 10.1016/j.jacc.2009.10.012. [DOI] [PubMed] [Google Scholar]
- 21.Priest VL, Scuffham PA, Hachamovitch R, Marwick TH. Cost-effectiveness of coronary computed tomography and cardiac stress imaging in the emergency department: a decision analytic model comparing diagnostic strategies for chest pain in patients at low risk of acute coronary syndromes. JACC Cardiovasc Imaging. 2011;4:549–56. doi: 10.1016/j.jcmg.2011.03.008. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.