Background
Abnormal cardiac stress imaging findings are not always associated with angiographically significant coronary artery disease. The outcomes of patients with such false-positive findings have not been extensively examined. The aim of this retrospective study was to describe the characteristics and outcomes of patients with abnormal stress echocardiographic findings who had false-positive results compared with those who had true-positive results.
Methods
Of 1,477 consecutive patients (mean age, 66 ± 12 years; 61% men) with abnormal stress echocardiographic findings who underwent coronary arteriography within 30 days, death from any cause was ascertained.
Results
At coronary arteriography, 997 patients (67.5%) had true-positive results, defined by the presence of angiographically significant coronary artery disease (≥50% stenoses), and 480 (32.5%) had false-positive results, defined by <50% stenoses or normal coronary arteries. Of the subgroup of patients with markedly abnormal stress echocardiographic findings (n = 605), 28% had <50% stenoses or normal coronary arteries. During an average follow-up period of 2.4 ± 1.0 years, there were 140 deaths. The adjusted likelihood of subsequent death for patients with <50% stenoses compared to patients with ≥50% stenoses after abnormal stress echocardiography was 1.05 (95% confidence interval, 0.86-1.31; P = .62).
Conclusions
A sizable proportion of patients with abnormal stress echocardiographic results who are referred for coronary angiography have false-positive findings. The outcomes of patients with false-positive results were similar to those of patients with true-positive results. This finding suggests that patients with false-positive results on stress echocardiography should still receive intensive risk factor management and careful clinical follow-up.
Over the past 20 years, stress echocardiography has become an integral part of the evaluation of patients with known or suspected coronary artery disease (CAD). Stress echocardiography provides both diagnostic and prognostic information. Some patients who have abnormal results on stress echocardiography, including those who have markedly abnormal results, such as an abnormal response of the left ventricular end-systolic size, a decrease in the left ventricular ejection fraction, or a high wall motion score index at peak stress, have been found to have normal or only mildly (<50%) stenotic coronary arteries on subsequent coronary arteriography. In this situation, the abnormal stress echocardiographic result is often described as a false positive test. Possible explanations for false-positive results on stress echocardiography include the presence of left bundle branch block, hypertensive heart disease, cardiomyopathy, epicardial coronary artery spasm, small-vessel CAD, and overinterpretation of data by reviewing echocardiographers. The prognoses of patients with false-positive stress echocardiograms have not been extensively evaluated. Although it is assumed that patients with false-positive results have better survival than patients who have significant coronary disease, the underlying mechanisms potentially responsible for false-positive results suggest that the outcomes of such patients are not necessarily favorable.
The aims of this retrospective study in patients referred for coronary arteriography after abnormal results on stress echocardiography were (1) to identify any clinical or echocardiographic characteristics that are associated with the angiographic findings of mild (<50%) CAD or normal coronary arteries and (2) to ascertain the subsequent all-cause mortality of patients with false-positive results compared to that of patients with true-positive results. We hypothesized that patients with false-positive results would have better survival than patients with true-positive results.
Methods
Identification of Patients
The Mayo Clinic Institutional Review Board approved the study. Data from patients undergoing stress echocardiography at the Mayo Clinic (Rochester, MN) are routinely entered into a prospectively maintained database. All patients who underwent coronary arteriography <30 days after stress echocardiography from November 1, 2003, to December 31, 2006, were identified each month by a research assistant. Patients who did not authorize the use of their records for research purposes were not included.
Clinical Characteristics
Patient demographics, including cardiac risk factors and medications, were abstracted at the time of stress echocardiography and entered prospectively into the stress echocardiography research database. A history of CAD was defined by a prior myocardial infarction and/or prior coronary revascularization.
Stress Echocardiography
Patients underwent clinically indicated symptom-limited treadmill exercise or dobutamine stress echocardiography. Their data were entered into the database. Medications such as β-blockers were stopped prior to stress testing only if the referring physician did so. A hypertensive response to stress was defined as peak systolic blood pressure ≥ 210 mm Hg in men and ≥ 190 mm Hg in women, as previously described. The results of stress electrocardiography (ECG) were considered positive for ischemia if there was horizontal or down-sloping ST-segment depression ≥ 1 mm at 80 ms after the J point.
Two-dimensional echocardiographic images at rest and peak stress were obtained, digitized, recorded, and analyzed according to previously published protocols. Both digitized and videotaped images were used for interpretation. The left ventricular ejection fraction was determined by visual assessment or by a modification of the Quinones method. Regional wall motion was assessed and graded on a scale of 1 to 5 at rest and peak stress in each of 16 segments. The left ventricular wall motion score index was calculated by adding the scores and dividing the sum by 16. By definition, patients with abnormal results on stress echocardiography had rest regional wall motion abnormalities and/or stress-induced regional wall motion abnormalities. The overall change in left ventricular end-systolic size was assessed visually by comparing side-by-side rest and stress digitized images with each other. The left ventricular end-systolic size response was considered abnormal if it did not change appreciably or if it increased in response to stress. The study population was stratified according to the severity of the abnormal stress echocardiographic results. Markedly abnormal results were defined as those with an abnormal response of left ventricular end-systolic size and ≥5 abnormal segments (hypokinesis or worse) at peak stress. Patients with these abnormalities were compared with those who had less extensive echocardiographic abnormalities.
Coronary Arteriography
Coronary angiograms were interpreted by experienced invasive cardiologists and scored according to the Coronary Artery Surgery Study (CASS) criteria. The definition of normal coronary arteries was the absence of luminal irregularities or any other CAD in any of the epicardial vessels. Angiographically mild CAD was defined by the presence of <50% stenoses and angiographically significant CAD by the presence of ≥50% stenoses in ≥1 of the coronary arteries or major branches. Patients with ≥70% stenoses were also identified.
Ascertainment of Mortality Data
All-cause mortality data were retrospectively abstracted from the Social Security Death Index in January 2008. Causes of death were not available. Patients who did not have Social Security numbers could not be included in this follow-up.
Statistical Analysis
Categorical variables are summarized as percentages and continuous variables as mean ± SD. Comparisons between groups were based on 2-sided t tests for continuous variables and Pearson’s χ 2 test for categorical variables. Logistic regression analysis using mild CAD (<50% stenoses) and/or normal coronary arteries on subsequent angiography (yes or no) as the outcome variable was used to identify characteristics that were significantly associated with normal coronary arteries on subsequent angiography. Cox proportional-hazards regression modeling using time from echocardiography to death for those who died or time to linkage with the Social Security Death Index for those who were not identified as having died was used to identify whether false-positive stress results (defined as abnormal results in patients with <50% stenoses on subsequent angiography) were associated with death during follow-up. A backward selection approach was used to identify the independent predictors for both the logistic and Cox regression analyses. Analyses were performed using JMP version 6.0.0 (SAS Institute Inc, Cary, NC).
Results
Study Population
A total of 23,107 patients underwent clinically indicated stress echocardiography at the Mayo Clinic during the above-described 38-month period. Stress echocardiographic results were normal in 15,755 patients (68%) and abnormal in 7,352 patients (32%). A total of 1,645 patients subsequently underwent coronary angiography within 30 days. Of these patients, 168 had normal results on stress echocardiography prior to coronary arteriography and were not included in this analysis. The other 1,477 patients who underwent coronary arteriography had abnormal stress echocardiographic findings and constituted the study population.
Clinical Characteristics
The mean age of the study population was 66 ± 12 years, 61% of the patients were men, 35% had histories of CAD, and 72% had diagnoses of hypertension. Most patients underwent stress echocardiography for the evaluation of symptoms of chest pain (48%) or dyspnea (39%). The patients’ clinical characteristics are shown in Table 1 according to their findings on coronary arteriography.
| ≥70% stenosis | 50%-69% stenosis | <50% stenosis | Normal coronary arteries | |
|---|---|---|---|---|
| Variable | (n = 851) | (n = 146) | (n = 339) | (n = 141) |
| Age (y) | 69 ± 11 | 67 ± 12 | 64 ± 11 | 60 ± 11 |
| Women | 231 (27%) | 61 (42%) | 190 (56%) | 100 (71%) |
| Body mass index (kg/m 2 ) | 28.8 ± 10.6 | 31.1 ± 25.0 | 30.6 ± 16.1 | 28.6 ± 6.1 |
| History of hypertension | 661 (78%) | 109 (75%) | 218 (64%) | 78 (56%) |
| Diabetes | 240 (28%) | 50 (34%) | 58 (17%) | 19 (13%) |
| Hyperlipidemia | 674 (79%) | 111 (76%) | 217 (64%) | 67 (47%) |
| History of smoking | 471 (56%) | 87 (60%) | 156 (46%) | 52 (36%) |
| History of CAD | 427 (50%) | 37 (25%) | 53 (16%) | 2 (1%) |
| Medications | ||||
| ACE inhibitors/ARBs | 366 (43%) | 60 (41%) | 113 (33%) | 30 (21%) |
| β-blockers | 490 (58%) | 76 (52%) | 143 (42%) | 48 (34%) |
| Statins | 494 (60%) | 73 (51%) | 117 (35%) | 25 (18%) |
| Presenting symptom | ||||
| CP | 390 (46%) | 72 (49%) | 168 (50%) | 84 (59%) |
| Typical angina | 161 (19%) | 14 (10%) | 30 (9%) | 15 (11%) |
| Atypical chest pain | 228 (27%) | 58 (40%) | 138 (41%) | 69 (49%) |
| Dyspnea | 313 (37%) | 63 (43%) | 143 (42%) | 61 (43%) |
| CP and dyspnea | 168 (20%) | 32 (22%) | 82 (24%) | 39 (27%) |
Echocardiographic and Angiographic Characteristics
The modality of stress testing was exercise echocardiography in 750 patients (51%), dobutamine stress echocardiography in 725 patients (49%), and transesophageal atrial pacing in 1 patient. Echocardiographic characteristics at rest and peak stress are shown in Table 2 .
| ≥70% stenosis | 50%-69% stenosis | <50% stenosis | Normal coronary arteries | |
|---|---|---|---|---|
| Variable | (n = 851) | (n = 146) | (n = 339) | (n = 141) |
| Study type | ||||
| Exercise | 385 (45%) | 71 (49%) | 196 (58%) | 96 (69%) |
| Dobutamine | 465 (55%) | 75 (51%) | 142 (42%) | 45 (31%) |
| Transesophageal pacing | 1 (0.1%) | 0 (0%) | 0 (0%) | 0 (0%) |
| Rest variables | ||||
| EF | 54 ± 11 | 55 ± 11 | 57 ± 9 | 57 ± 8 |
| EF < 50% | 202 (24%) | 32 (22%) | 41 (12%) | 15 (11%) |
| WMSI | 1.33 ± 0.41 | 1.28 ± 0.42 | 1.16 ± 0.31 | 1.15 ± 0.29 |
| Number of abnormal segments at rest | ||||
| 0 | 347 (41%) | 77 (53%) | 222 (65%) | 94 (67%) |
| 1-2 | 140 (16%) | 18 (12%) | 42 (12%) | 22 (16%) |
| 3-4 | 127 (15%) | 14 (10%) | 25 (7%) | 6 (4%) |
| ≥5 | 237 (28%) | 37 (26%) | 50 (15%) | 19 (14%) |
| Stress variables | ||||
| Hypertensive response | 60 (7%) | 25 (17%) | 42 (12%) | 24 (17%) |
| Positive results on ECG | 189 (22%) | 43 (29%) | 55 (16%) | 29 (21%) |
| EF | 54 ± 14 | 56 ± 13 | 60 ± 12 | 59 ± 11 |
| WMSI | 1.68 ± 0.43 | 1.63 ± 0.43 | 1.49 ± 0.35 | 1.51 ± 0.38 |
| Abnormal LVESS response | 417 (49%) | 63 (43%) | 145 (43%) | 57 (41%) |
| Number of abnormal segments at peak stress | ||||
| 1-2 | 70 (8%) | 14 (9%) | 44 (13%) | 20 (14%) |
| 3-4 | 128 (15%) | 31 (21%) | 93 (27%) | 39 (28%) |
| ≥5 | 653 (77%) | 101 (70%) | 202 (60%) | 82 (58%) |
The median time to angiography was 4 days (interquartile range, 1-8 days); 88% underwent angiography within 14 days, and 95% within 21 days after stress echocardiography. The angiographic findings were as follows: 851 patients (57.6%) had ≥70% stenoses, 146 patients (9.9%) had 50% to 69% stenoses, 339 (23%) had <50% stenoses, and 141 patients (9.5%) had normal coronary arteries. Compared with patients who had angiographic evidence of CAD, those who had normal coronary arteries were younger and were more often female.
Only a small minority of patients who underwent angiography had hypertensive responses to stress (10%). This was true for patients with and those without angiographic evidence of CAD ( Table 2 ). Positive results on stress ECG were present in a minority of patients. The majority of patients in all angiographic subgroups developed extensive wall motion abnormalities at peak stress.
Multivariate analysis revealed that younger age, female gender, the absence of diabetes, the absence of a history of CAD, the absence of a history of hypertension, and negative results on stress ECG were independently associated with angiographically mild (<50%) CAD and/or normal coronary arteries ( Table 3 ).
| Normal | Normal or <50% stenosis | |||
|---|---|---|---|---|
| Variable | OR (95% CI) | P | OR (95% CI) | P |
| Age ∗ | 0.67 (0.57-0.78) | <.001 | 0.78 (0.71-0.87) | <.001 |
| Female gender | 2.99 (2.01-4.53) | <.001 | 2.73 (2.14-3.50) | <.001 |
| Absence of diabetes mellitus | 1.98 (1.16-3.53) | .01 | 1.94 (1.42-2.67) | <.001 |
| Absence of history of CAD | 30.4 (9.48-186.9) | <.001 | 4.91 (3.58-6.84) | <.001 |
| Absence of history of hypertension | 1.33 (0.89-1.97) | NS | 1.40 (1.07-1.82) | .01 |
| Body mass index | 0.99 (0.97-1.01) | NS | 1.01 (1.00-1.01) | NS |
| History of chest pain or dyspnea | 1.22 (0.80-1.91) | NS | 0.90 (0.69-1.17) | NS |
| Exercise (vs dobutamine stress) | 1.00 (0.80-1.01) | NS | 1.00 (0.86-1.00) | NS |
| Hypertensive response | 1.59 (0.92-2.65) | NS | 1.40 (0.95-2.06) | NS |
| Negative results on stress ECG | 1.44 (0.96-2.18) | NS | 1.77 (1.38-2.28) | <0.001 |
| Resting ejection fraction <50% | 1.33 (0.89-1.97) | NS | 1.02 (0.71-1.46) | NS |
| Markedly abnormal stress echocardiographic results † | 1.05 (0.70-1.58) | NS | 1.19 (0.92-1.54) | NS |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
