Author + information
- William E. Boden, MD⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. William E. Boden, Buffalo General Hospital, 100 High Street, Buffalo, New York 14203
Among the most common diagnostic tests performed in adults who present with chest pain symptoms is a noninvasive evaluation of suspected obstructive coronary artery disease (CAD), which is performed annually in nearly 20 million patients (1). The options for testing include the use of the standard exercise treadmill test (ETT), or the use of ETT combined with single-photon emission computed tomography (SPECT) myocardial perfusion imaging (MPI) or, in patients unable to exercise, the use of MPI with pharmacological vasodilator stress. Despite the widespread clinical use of ETT, there is a paucity of randomized trials comparing the effectiveness of the standard ETT with that of an imaging-guided strategy as an initial diagnostic strategy and for assessing clinical outcomes of symptomatic women (1). Current evidence-based clinical practice guidelines recommend an ETT as the preferred procedure for the initial evaluation of women with suspected myocardial ischemia (2). Yet, frequent reports detail significant limitations in the accuracy of the electrocardiogram in women, including a high false-positive rate and the inability to accurately quantify high lateral or posterolateral ischemia on the ETT (2). As a result, current practice patterns are variable and often include direct referrals of women to higher-cost procedures, such as exercise MPI.
It is generally well recognized that physicians frequently underestimate ischemic heart disease (IHD) in women, despite the fact that women with IHD have more adverse outcomes as compared with men (3,4). It is also noteworthy to emphasize that women represent only 30% of patients enrolled in clinical trials that form the evidence base used to support the 2007 American Heart Association guidelines for cardiovascular disease prevention in women. Even more challenging is the recognition that IHD is the most common cause of death worldwide among adult women, and that data derived from the World Health Organization estimates that 80% of IHD deaths occur in developing nations (5). Such observations raise legitimate concerns regarding the availability and choice of various diagnostic approaches that would be best deployed to address the growing problem of IHD worldwide and, in particular, in women with suspected CAD.
Against this backdrop, Cerci et al. (6) report in this issue of iJACC that in their single-site study from the Quanta Nuclear Diagnostic Clinic in Curitiba, Brazil, SPECT-MPI added significant incremental prognostic information to clinical and left ventricular functional variables while enhancing the ability to classify this population of Brazilian women into low- and high-risk categories of all-cause mortality. This was an ambitious undertaking by the Brazilian investigators who, between March 2004 and October 2007, studied 2,427 consecutive female patients age 55 to 75 years who underwent SPECT-MPI for evaluation of chest pain and suspected CAD, surveillance of known IHD, investigation of left ventricular (LV) systolic dysfunction, follow-up nuclear cardiac imaging for initially ambiguous or inconclusive treadmill exercise testing, or for additional diagnostic evaluation for diabetes mellitus or pre-operative assessment of suspected myocardial ischemia. A total of 202 women (8.3% of the observational cohort) had missing or incomplete follow-up data, and were excluded from the analysis, leaving a study population of 2,225 women who underwent SPECT-MPI and in whom mortality, as the primary outcome measure, was subsequently evaluated during a mean 3.7-year follow-up.
There are notable strengths of this paper, for which the authors should be commended. The current report details the prognostic accuracy of myocardial perfusion SPECT-MPI in a relatively large cohort of women referred clinically for nuclear cardiac noninvasive testing in Brazil. Their study deals with 2 important populations of patients (i.e., women with suspected IHD and subjects from a developing country) that have been relatively under-represented in clinical trials, and hence understudied scientifically as relating to the role of MPI in the diagnosis and management of IHD. The strengths of this analysis are the prospective nature of patient enrollment and data collection as well as the relatively long length of follow-up (mean: 3.7 years) and appreciable event rate (6.2% cumulative all-cause mortality), which was the primary outcome measure.
The main study findings appear consistent with prior published results in the United States and Europe and provide validation of prior findings to non-U.S. cohorts. As such, the main study findings can be considered largely confirmatory of the known clinical benefits associated with SPECT-MPI as a prognostic tool among those patients who exhibit an abnormal scan. In this single-site observational study, the death rate in Brazilian women with an abnormal MPI scan was 13.1% as compared with 4.0% in those with a normal scan (p < 0.001). Further subgrouping of abnormal MPI studies as “reversible” (mild/moderate vs. severe; complete vs. partial) and “fixed defects only” revealed differential survival rates that decreased from 89.8% to 80.3% and 78.8% in the reversible subgroups, respectively, as compared with 81% in the “fixed defects” subgroup. After adjustment for age, risk factors for ischemic heart disease, and left ventricular ejection fraction (LVEF), the stepwise multivariable model showed that the presence of any perfusion defect (summed stress score > 4) was an independent predictor of mortality. Additionally, when a Cox proportional hazards analysis was employed, the authors were able to demonstrate that, after adjusting for known IHD, age, smoking, and LVEF, SPECT-MPI significantly improved the model discrimination (or the integrated discrimination index = 0.009; p = 0.02), added incremental prognostic information with an increase in global chi-square from 87.7 to 127.1; p < 0.0001, as well as improved overall risk prediction (net reclassification improvement = 0.12; p = 0.005. On the surface, these appear to be impressive findings of incremental clinical benefit for SPECT-MPI in the cohort of Brazilian women studied.
However, there are several limitations of the present study that warrant comment. First, because this was a single-site observational study, there are obvious concerns that both selection/sampling bias and ascertainment bias were important factors in the results the authors report. Second, there are methodological concerns pertaining to SPECT-MPI data acquisition and interpretation; no attenuation correction or scatter correction was employed, and final results of scans were interpreted by a consensus of 2 experienced nuclear cardiology readers as opposed to blinded interpretations of independent experts to resolve MPI discrepancies, which could have influenced the study findings. Third, the results of the study were restricted to Brazilian women aged 55 to 75 years, so it is unclear whether the results of the current findings can be generalized to younger or older women. In fact, the authors never clarify why women younger than age 55 years or older than age 75 years were excluded. Fourth, a sizeable percentage of patients (n = 202; 8.3%) were lost to follow-up and, despite a post hoc sensitivity analysis, one cannot be certain to what degree this—among other unmeasured factors—introduced significant confounding to the current findings. Fifth, a large percentage of the study cohort were asymptomatic (45%), and had preserved systolic function, with a mean LVEF of 64 ± 11%, where only 20% had a history of known IHD. This, again, raises concerns about the validity and external generalizability of the study findings to the larger universe of patients (women or men; developed or undeveloped nations) with suspected IHD for whom physicians would contemplate diagnostic testing with SPECT-MPI.
In particular, if we examine the entire study population (n = 2,427) of those who underwent SPECT-MPI, less than one-quarter (n = 548; 22.6%) had an abnormal scan, of which 443 (81% of the abnormal group) had a completely reversible perfusion defect. Importantly, however, in terms of the severity of the ischemic defect, 382 of 443 completely reversible scans (86%) exhibited a mild/moderate defect. Thus, 382 of 2,427 patients (15.7%) had a mild/moderate defect, and only 166 patients (6.8%) had either a severe reversible perfusion defect (n = 61), a fixed perfusion defect (n = 53), or a partial reversible perfusion defect (n = 52). In more closely examining the 3-year risk prediction models of the incremental utility of adding LVEF to that of clinical risk, and the SPECT-MPI perfusion model added to the clinical risk + LVEF model (Table 3 in Cerci et al. ), we see that for those with a 3-year risk of <3% (37% of all patients), there is no incremental benefit of either LVEF or the perfusion models over and above the clinical model, whereas we see the greatest incremental benefit of the addition of SPECT-MPI in the subsets of patients with 3% to 6% and >6% 3-year risk of all-cause mortality, as would be expected.
Event-free survival at a mean 3.7 years was 96% in the ∼75% of patients with a normal scan (∼1.1% annualized mortality), whereas event-free survival was ∼90% in the ∼16% with a mild/moderate reversible perfusion defect (∼2.8% annualized mortality), which comprised only about 16% of the entire cohort. The clearest and most significant incremental benefit associated with the addition of SPECT-MPI to either the clinical or clinical + LVEF model could be demonstrated in the 3 small subsets of patients (n = 166 patients) with either a severe reversible perfusion defect, a partial reversible perfusion defect, or a fixed perfusion defect—yet this comprised only ∼7% of the entire population. Thus, although there clearly is incremental benefit associated with SPECT-MPI in assessing adverse outcomes, it would appear from these data in Brazilian women that MPI defines a serious signal of mortality risk only in a small minority of patients. This raises the important consideration as to whether, for population screening purposes, the added complexity, capital acquisition costs of imaging hardware, and associated recurring expenditures (e.g., ongoing costs associated with personnel, software, institutional overhead, etc.) associated with SPECT-MPI warrant this as a first-line approach to the detection of IHD and assessing prognosis.
Clearly, a compelling need exists for higher levels of scientific evidence, including more randomized trials, in the area of diagnostic testing and has resulted in increased attention toward developing targeted, prospective comparative effectiveness research to more rigorously address the outpatient diagnostic evaluation of patients with suspected CAD (1,3). For developing countries such as Brazil and many other emerging nations, one has to confront the more fundamental question of what is the preferred noninvasive diagnostic test to undertake in patients with suspected IHD, those with multiple cardiac risk factors for premature CAD, those with myocardial systolic dysfunction, or those who require pre-operative risk stratification or screening for occult CAD—regardless of gender. This would seem to be an even more compelling concern in developing societies such as Brazil where one might expect that economic constraints and resource utilization would be important factors in determining what diagnostic approaches for detecting suspected IHD in women, as well as men, are the most clinically appropriate and cost effective (7,8).
In the paper by Cerci et al. (6), we cannot discern whether MPI added significant, incremental diagnostic utility over and above standard ETT. The stress testing method employed for the SPECT-MPI studies was treadmill exercise testing in 1,626 women (73.1%), pharmacological vasodilator stress in 567 women (25.6%), and low-level treadmill exercise combined with pharmacological stress in only 1.4% of patients. Unfortunately, the authors provide no data regarding electrocardiographic findings or Duke Treadmill Score among the almost three-quarters of patients who underwent ETT. One might expect this to be an important consideration, since standard ETT is far more prevalent diagnostically, is less costly, and would logically be expected to be the preferred initial diagnostic approach to evaluate women and men with suspected IHD—regardless of whether this clinical decision by physicians is contemplated in developed, or developing, nations.
As noted in the preceding text, the evaluation of women with suspected CAD remains a conundrum due to their atypical symptom presentation that often results in varied diagnostic practice patterns. A recent study, known as the WOMEN (What Is the Optimal Method for Ischemia Evaluation in Women) trial, was designed to evaluate whether a diagnostic strategy using ETT versus exercise MPI in women with an intermediate pre-test likelihood of CAD affects 2-year post-test outcomes for major adverse cardiovascular events (9). In this trial, a total of 824 women from 43 centers were randomized to either standard ETT or ETT + MPI; 88% of enrolled patients were classified as low-intermediate risk. For women randomized to ETT alone, test results were normal in 64.1%, equivocal in 15.8%, and abnormal in 20.2%. By comparison, the exercise MPI results were normal in 90.5%, mildly abnormal in 3.3%, and moderate to severely abnormal in 6.2% of women. At 2 years, there was no difference in major adverse cardiovascular events (98.0% for standard ETT and 97.7% for exercise MPI, p = 0.59) (9).
For women randomized to ETT alone, the rate of crossover to MPI was 17.7% as compared with repeat MPI in 9.3% for those who underwent an initial exercise MPI (p < 0.0001). Overall, 6.2% of women underwent coronary angiography during follow-up. As compared with ETT alone, index testing costs were higher for the exercise MPI group (p < 0.001), whereas downstream procedural costs were slightly lower (p = 0.008). Overall, the cumulative diagnostic cost savings was 48% for the ETT-alone group as compared with the exercise MPI group (p < 0.001) (9). These findings, quite comparable to the study population of the Brazilian cohort, show that in low- to intermediate-risk women capable of exercise at baseline, a nonimaging strategy employing standard ETT was as effective in predicting clinical outcomes and more cost effective than a strategy of initial exercise MPI. Thus, the principal finding of the WOMEN trial is that there was no incremental benefit for an initial diagnostic strategy of exercise MPI as compared with standard ETT only in symptomatic women with suspected CAD and preserved functional capacity who were able to exercise. Additionally, we could not demonstrate that there was any incremental prognostic benefit of exercise MPI in predicting clinical outcomes at 2 years as compared with an ETT-only strategy.
On the basis of the study by Cerci et al. (6), and in light of the WOMEN trial findings (9), what conclusions might be drawn from these studies, and what are the implications for clinical practice? First, the Brazilian investigators have demonstrated the incremental value of SPECT-MPI above and beyond clinical parameters and LVEF in predicting 3- to 4-year mortality in women, particularly those with severe reversible, partially reversible, and fixed perfusion defects. This is an important scientific contribution. Second, the data derived from the WOMEN trial suggest that, for women capable of performing exercise, standard ETT without imaging appears a reasonable (and justifiable) first-line test resulting in equivalent 2-year outcomes as compared with ETT + MPI. These results could have profound implications for the estimated 2 to 3 million women undergoing exercise MPI testing each year, and would likely result in sizeable cost savings (i.e., 48%) within the noninvasive diagnostic evaluations for suspected myocardial ischemia which, in turn, would lead to reductions in the societal economic burden of medical imaging. Current clinical practice patterns often rely on direct imaging strategies for women. In fact, use of MPI has been one of the fastest growing Medicare procedures and routinely ranks as one of its most costly expenditures (1). The more judicious use of exercise MPI as a first diagnostic procedure may decrease exposure to ionizing radiation for many women, particularly those at low risk for CAD (10,11), even though the preponderance of evidence supports the fact that lifetime attributable risk of MPI is low (12).
In summary, SPECT-MPI remains a valuable risk assessment and prognostic tool in patients with IHD. Whether this should be the first-line approach in women with low to intermediate risk remains unclear. Performing MPI in developing nations may also result in additional incurred costs that, in light of the above information, need to be justified in future studies designed prospectively to measure and quantify cost effectiveness and resource utilization in subsets of patients with low, intermediate, and high risk for IHD.
Dr. Boden has reported that he has no relationships relevant to the contents of this paper to disclose.
↵⁎ Editorials published in JACC: Cardiovascular Imaging reflect the views of the authors and do not necessarily represent the views of JACC: Cardiovascular Imaging or the American College of Cardiology.
- American College of Cardiology Foundation
- Shaw L.J.,
- Min J.K.,
- Hachamovitch R.,
- et al.
- Mieres J.H.,
- Shaw L.J.,
- Arai A.,
- et al.
- Bairey Merz C.N.,
- Shaw L.J.,
- Reis S.E.,
- et al.
- Mosca L.,
- Banka C.L.,
- Benjamin E.J.,
- et al.
- World Health Organization
- Shaw L.J.,
- Marwick T.H.,
- Berman D.S.,
- et al.
- Marwick T.H.,
- Shaw L.,
- Case C.,
- Vasey C.,
- Thomas J.D.
- Mieres J.H.,
- Shaw L.J.,
- Hendel R.C.,
- et al.
- Gerber T.C.,
- Gibbons R.J.