Design, rationale, and populations of an international outcomes and utilization study of pharmacologic stress SPECT myocardial perfusion imaging in contemporary practice

Design, rationale, and populations of an international outcomes and utilization study of pharmacologic stress SPECT myocardial perfusion imaging in contemporary practice James R. Johnson, PhD,a Richard J. Barrett, PhD,b Rory Hachamovitch, MD, MSc,c James E. Udelson, MD,d Joseph Massaro, PhD,e and Stephen A. Jenkins, MBAb Background. A prospective, international, multicenter, observational study was conducted to investigate patient and population characteristics; utilization of radiopharmaceuticals and pharmacologic stress (PS) agents; imaging protocols; clinical outcomes; the incidence, intensity, and time to onset of adverse events; and the prognostic value of single photon emission computed tomography (SPECT) myocardial perfusion imaging (MPI) procedures. The rationale, study methods, and data on presenting populations are described. Methods and Results. Investigators recorded the demographics, American College of Cardiology/American Heart Association pretest likelihood for coronary artery disease, cardiovascular risk factors, antianginal drug use, use of PS agents and associated adverse events, and radiopharmaceutical(s) and imaging protocol for each patient enrolled. SPECT images were reconstructed at each site; investigators assigned summed stress and summed rest scores using a 17-segment model (rating perfusion on a scale ranging from 0 to 4). Patients were followed up for 1 year for clinical outcomes of revascularization, nonfatal myocardial infarction, or death. Conclusion. The design offers a unique opportunity to study the characteristics of patients referred for SPECT imaging over a period of time consistent with the laboratories’ usual practices, provides an up-to-date PS safety registry, and allows assessment of the prognostic value of PS SPECT MPI across a wide number of covariables, as well as relationships between patient and population characteristics, SPECT MPI results, and clinical outcomes. (J Nucl Cardiol 2008;15:687-97.) Key Words: Pharmacologic stress • perfusion imaging • nuclear medicine • prognosis • outcomes Stress myocardial perfusion single photon emission computed tomography (SPECT) imaging (MPS) plays a central and widespread role in the assessment and management of patients with known or suspected coronary artery disease (CAD). Although extensive literature supports the clinical effectiveness and cost-effectiveness of MPS for guiding risk stratification and resource utiliza-

From the Premier Research Group, Limiteda and King Pharmaceuticals Research and Development, Cary, NCb; private practice, Los Angeles, Calif c; and Division of Cardiology, Tufts-New England Medical Center,d and School of Public Health, Department of Biostatistics, Boston University,e Boston, Mass. Received for publication Feb 27, 2008; final revision accepted April 30, 2008. Reprint requests: James R. Johnson, PhD, Global Biostatistics, Premier Research Group, 104 Lutterworth Ct, Cary, NC 27519-8682; [email protected]. 1071-3581/$34.00 Copyright © 2008 by the American Society of Nuclear Cardiology. doi:10.1016/j.nuclcard.2008.06.011

tion, a closer examination of the individual studies reveals a number of significant limitations.1-4 These studies are predominantly from a restricted number of academic, tertiary care centers with significant MPS expertise and thus suffer varying biases and methodologic issues. In 1994 80% of all MPS procedures were performed in this setting, but today, half are performed in private practices. Importantly, most prior studies consisted of patients who completed MPS before the introduction of significant therapeutic advances (eg, 3-hydroxy-3-methylglutaryl– coenzyme A reductase inhibitors), interventions (intravascular stents), and imaging methods (pharmacologic stress [PS] with low-level exercise, gated SPECT). Finally, given the challenges of evaluating noninvasive testing in a single-site setting, these studies are also fraught with diverse and varying biases and methodologic issues. To our knowledge, no study has examined prognostic and resource utilization issues in a widely heterogeneous population using a uniform MPS scor687

688

Johnson et al Outcomes and utilization study of pharmacologic stress SPECT MPI

Journal of Nuclear Cardiology September/October 2008

ing system and considering variability in practice settings, geography, and physician specialty. Currently, although almost half of stress MPS studies (46%) are performed using PS,5 PS MPS is significantly underrepresented in the prognosis and outcomes literature. Furthermore, although safety and adverse event (AE) data are available for each PS agent, no prospective study has examined and compared the incidence, intensity, and timing of AEs associated with adenosine, dipyridamole, and dobutamine, or the impact of newer protocols on AEs.6-10 We report an open-label, prospective, international, multicenter, observational, clinical study designed to characterize a cross section of patients presenting for adenosine, dipyridamole, or dobutamine stress MPS with respect to the nature, incidence, intensity, and timing of the AEs associated with each drug as well as post-MPS risk stratification, resource utilization, and clinical outcomes at 1 year of follow-up and the confounders of these endpoints. METHODS Study Objectives Objectives for which prespecified analysis strategies were designed include the following: 1. Overall and risk-stratified clinical endpoint event rates (revascularization, nonfatal myocardial infarction [MI], or all-cause mortality) for up to 1 year after PS or low-level exercise with PS (LEPS), as well as MPS procedure 2. Observed and risk-adjusted event rates both overall and as a function of MPS results for each PS agent 3. Distribution of patients by cardiovascular history and pretest likelihood for CAD 4. Distribution of patients undergoing adenosine, dipyridamole, or dobutamine stress; LEPS procedures; or exercise stress 5. Relationships of practice type (private practice, community hospital, or tertiary care center) and country to the distribution of patients 6. Proportion of patients who cannot complete an adenosine or dipyridamole stress MPS procedure because of a history of asthma or chronic obstructive pulmonary disease 7. Relationships of clinical endpoint events, as well as the timing of such events, to the nature and severity of perfusion defects, CAD history, pretest likelihood of CAD, demographic factors (age, gender, race), risk factors, country, and practice type 8. Incidence, intensity, and time to onset of AEs associated with each PS agent and LEPS 9. Relationships of patient outcomes/AEs and their timing to the following: A. Nature and severity of perfusion defects B. Pretest likelihood of CAD or prior CAD C. Demographic factors (age, sex, race, weight) D. Country E. Practice type

Figure 1. Study design and data flow.

10. Severity of perfusion defects observed in patient populations

Study Design The study was conducted from March to December 2003 and was approved by the institutional review boards at each participating center. Investigators designated their sites as private practice, community hospital, or tertiary care centers. The following data were acquired from all patients referred for MPS procedures over a period of 20 consecutive working days: gender, ethnic origin (white, Hispanic, black, Asian/Pacific Islander, American Indian/Alaskan Native, or other), and expected procedure (exercise stress, LEPS, or PS). Those patients aged greater than 18 years who were referred for a PS MPS study with adenosine, dipyridamole, or dobutamine or an LEPS study with one of these drugs were invited to participate. Patients provided written informed consent before conduct of any study procedures (Figure 1). Exclusion criteria included referral for an exercise MPS procedure (or a study in which drugs [eg, atropine] were used to enhance exercise protocols), pregnancy, presence of automated implantable cardiac defibrillator, known nonischemic cardiomyopathies, prior heart transplantation (or listed for transplantation), or known clinically significant valvular disease. Patients originally referred for an exercise stress study but who did not complete it were eligible. Patients were stratified before the MPS procedure as having known CAD (ie, history of MI, revascularization, or documented CAD by prior angiography) or no prior CAD with a very low, low, intermediate, or high pretest likelihood for

Journal of Nuclear Cardiology Volume 15, Number 5;687-97

Johnson et al Outcomes and utilization study of pharmacologic stress SPECT MPI

CAD according to American College of Cardiology (ACC)/ American Heart Association (AHA) definitions.11

Patient History and Demographics Patients also provided weight, age, and an abbreviated medical history including history of asthma, chronic obstructive pulmonary disease, congestive heart failure, angina, diabetes mellitus (insulin dependent or non–insulin dependent), left bundle branch block, family history of CAD, previous MI, coronary artery bypass graft or percutaneous coronary intervention, elevated cholesterol level, and smoking (current or previous). Patients provided contact data for themselves, an alternate contact, and a referring physician and agreed to respond to 6 telephone contacts during the subsequent year.

689

confirmed but not adjudicated. Patients completing revascularization after MPS were followed up through 1 year to capture occurrences of nonfatal MI or death. At the completion of the 20-day trial period, clinical sites submitted all CRFs and patient logs to a central repository for patient follow-up and data management. All data were single entered in a study-specific database. A 15% random sample of the patients in the study was selected for quality-assurance assessment. The study was not monitored, data entries were not queried, and data were not imputed. Occasional discrepancies in sums of patients and analyses result from missing data or because sparse data were not included (eg, data from the few American Indian/Alaskan Natives enrolled have not been segregated in Table 1).

Planned Statistical Analyses MPS Studies Enrolled patients completed a PS or LEPS MPS study. Images were acquired, processed, reconstructed, and interpreted at the sites per the laboratories’ usual practices. PS agents and radiopharmaceuticals used and their dosage were recorded. Investigators were trained during investigators’ meetings to score the stress and rest images using the AHA/ACC/ American Society of Nuclear Cardiology 17-segment, 5-point semiquantitative scoring system (0, normal perfusion; 1, minimally abnormal [not definitely abnormal]; 2; mildly abnormal [moderate reduction in counts, definitely abnormal]; 3, moderately abnormal [severe reduction in counts]; and 4, severely abnormal [absent uptake])12,13 and to record the summed stress score (SSS) and summed rest score (SRS) on the patient’s case report form (CRF) (Figure 2). The summed difference score (SDS) was calculated as the difference between SSS and SRS.

Adverse Events AEs that were deemed reasonably related to the PS agent and that occurred within 24 hours of drug administration were recorded by checking boxes on the CRF to indicate that a patient had 1 of the following 11 AEs: second/third degree atrioventricular block; supraventricular arrhythmias; ventricular arrhythmias; arm, back, or shoulder discomfort; throat, neck, or jaw discomfort; nausea and/or emesis; chest pain; dyspnea; flushing; dizziness or light-headiness; and palpitations. Sites were provided a list of verbatim terms (eg, “chest burning” and “queasy”) to ensure consistent coding of AEs across sites. The intensity (mild, moderate, or severe) and time to onset of each AE from the start of PS agent administration (⬍1 hour, 1-4 hours, or 4-24 hours) and all other data from the MPS procedure were recorded on a single-page CRF (Figure 2). Patients or designated caregivers were contacted by site staff 48 to 96 hours after the PS MPS procedure to assess late-occurring AEs and occurrence of clinical endpoint events. They or designated physicians were contacted by trained medical professionals at a central call center (Figure 1), working from a script, at 1, 2, 3, 6, and 12 months to assess occurrence of clinical endpoint events. Clinical events were

A minimum sample size of approximately 4,000 patients was estimated to detect a clinical endpoint event rate of 0.08 by use of a confidence interval approach.14 A population clinical endpoint event rate of 0.12 over a 1-year period was assumed for planning the study. Descriptive statistics will be used to summarize the overall study population, planned subgroups (eg, pretest CAD likelihood strata, demographics, country, and practice type), MPS test utilization, AE incidence, intensity, time to onset, and relationships between population descriptors and test utilization. The distribution of AE rates will be summarized overall, by population, by PS agent, by pretest CAD likelihood strata, and by nature and extent of perfusion defect by use of categorical analysis procedures (ie, ␹2, odds ratios, and tests for trends). Statistical models will be developed to fully characterize PS SPECT patient populations, relationships to clinical outcomes, and MPS results. Overall, stratified and risk-adjusted times to revascularization and all-cause mortality and their respective 95% confidence intervals will be evaluated by use of a Cox proportional hazards model with a log-rank test, as well as corresponding Kaplan-Meier survival plots to compare PS agents. Overall, stratified and risk-adjusted rates of all-cause mortality, combined all-cause mortality and nonfatal MI, and referral to revascularization and their respective 95% confidence intervals will be determined for the enrolled cohort as well as for each PS agent. These results will also be expressed as a function of MPS results. Image results will be categorized (as normal vs abnormal, normal vs fixed vs reversible, and so on) based on SSS, SRS, and SDS. Relationships between the timing of clinical endpoint events and (1) nature and extent of perfusion defects, (2) known CAD and pretest likelihood for CAD, (3) demographic factors (eg, age, gender, race, and weight), (4) country, and (5) practice type (eg, private, community, or tertiary care) will be assessed with multivariate regression models.

RESULTS Investigators at 89 private, community, and tertiary/academic imaging centers in 8 countries participated and logged all patients who presented for any

690

Johnson et al Outcomes and utilization study of pharmacologic stress SPECT MPI

Journal of Nuclear Cardiology September/October 2008

Figure 2. Study CRF. COPD, Chronic obstructive pulmonary disease; CABG, coronary artery bypass graft; PCI, percutaneous coronary intervention; MPI, myocardial perfusion imaging; LVEF, left ventricular ejection fraction; AV, atrioventricular.

SPECT myocardial perfusion imaging study over 20 consecutive working-day periods. (Appendix A lists sites, investigators, and collaborators.) Of 7,578 patients referred for PS or LEPS, 5,174 were enrolled. The majority of sites were in the United States (72/89 [81%]), and these sites provided 3,477 of the 5,174 PS patients enrolled (67%). The remaining 1,697 patients

were enrolled at 5, 7, and 5 sites in Canada, Latin America, and Europe, respectively. One high-volume site in Brazil and two in Canada account for a majority of patients from those regions. The distribution of patients enrolled by country is provided in Table 1. Of the 72 US sites, 45 (63%) enrolled fewer than 50 patients and 27 (38%) enrolled more than 50 patients

Journal of Nuclear Cardiology Volume 15, Number 5;687-97

Johnson et al Outcomes and utilization study of pharmacologic stress SPECT MPI

691

Table 1. Demographic data for all patients by practice type (in United States), region, and pretest likelihood for CAD

Presenting, all sites Exercise PS Enrolled, all sites PS alone PS ⫹ LEPS Pretest likelihood for CAD Very low Low Intermediate High Known CAD Practice type, all sites Community Private Tertiary care/academic Countries (No. of sites) Argentina (4) Belgium (1) Brazil (1) Canada (5) Chile (2) Italy (2) United Kingdom (2) United States (72)

n (%)

Female gender

Age [mean (SD)] (y)

9,075 (54.5%) 7,578 (45.5%)

39% 52%

58.6 (12.08) 66.2 (12.34)

4,231 (82%) 943 (18%)

43% 9%

157 (3.0%) 1,001 (19.3%) 1,335 (25.8%) 384 (7.4%) 2,142 (41.4%)

Race Weight [mean (SD)] (kg) White Black Hispanic Outpatient

78% 80%

11% 10%

6% 4%

— —

66.4 (11.84) 84.2 (21.67) 65.6 (11.95) 85.3 (21.24)

82% 86%

10% 7%

5% 4%

83% 92%

75% 57% 63% 48% 41%

51.1 (12.70) 64.9 (12.82) 65.5 (11.76) 68.8 (10.56) 68.2 (10.58)

88.0 (26.80) 83.1 (22.52) 84.4 (22.64) 82.7 (19.62) 84.9 (19.77)

80% 81% 81% 80% 86%

15% 12% 11% 9% 7%

4% 5% 5% 7% 4%

86% 85% 80% 80% 86%

1,369 (26.4%) 1,357 (26.2%) 2,448 (47.3%)

49% 54% 52%

66.5 (11.81) 85.4 (21.62) 68.2 (11.35) 87.5 (21.82) 65.1 (12.03) 82.2 (21.22)

81% 92% 78%

12% 5% 12%

3% 1% 7%

90% 99% 74%

173 (3.3%) 38 (0.7%) 352 (6.8%) 642 (12.4%) 67 (1.3%) 177 (3.4%) 248 (4.8%) 3,477 (67.2%)

51% 47% 51% 50% 51% 38% 48% 54%

65.0 (9.86) 62.7 (11.31) 64.4 (11.61) 66.6 (11.24) 69.2 (8.88) 63.7 (9.98) 63.7 (12.10) 66.7 (12.15)

56% 100% 81% 92% 14% 100% 84% 82%

— — 14% 1% — — 5% 13%

40% — — 4% 86% — — 2%

75% 48% 90% 84% 79% 40% 93% 87%

in the 20-day study period. Private practice sites tended to enroll fewer patients than tertiary care centers (Table 2). A total of 16,653 patients (9,174 men [55%] and 7,479 women [45%]) registered for a stress MPS procedure; 9,075 patients (55%) were referred for exercise stress, and 7,578 (45.5%) were referred for PS. Men were more likely than women to be referred for an exercise stress test (5,536/9,174 [60%] vs 3,539/7,479 [47%]). Of the 7,578 patients referred for PS, 5,174 (68%) consented to enroll, generating a study population of 3,269 men (63%) and 1,905 women (37%). In the United States slightly more women than men completed PS procedures (54% vs 46%), whereas the inverse was true at sites outside the United States (48% women vs 52% men). US patients tended to be heavier (mean, 87.9 kg) than those at sites outside the United States (Table 1). Of the 5,174

— —

77.3 (15.01) 76.5 (13.58) 72.6 (14.90) 80.1 (17.82) 73.4 (13.09) 76.6 (14.42) 78.4 (18.17) 87.9 (22.85)

enrolled PS patients, 82% (4,231/5,174) completed procedures with PS alone; low-level exercise was added in 484 men (9%) and 459 women (9%) (Table 1). Of the enrolled population, 41% had prior CAD whereas 59% had no prior CAD. Of the latter, few patients with a very low (5.5%) or high (13%) pretest likelihood of CAD were tested, whereas those with an intermediate likelihood and low likelihood comprise 46% and 35% of the remaining patients, respectively. More women than men were classified as having a very low (75%), low (57%), or intermediate (63%) pretest likelihood for CAD, and fewer women than men with known CAD (41% vs 59%) were tested. Enrollment by country, site type, pretest likelihood for CAD, and demographic data are presented in Tables 1 and 3. Patients referred for PS were older than those

692

Johnson et al Outcomes and utilization study of pharmacologic stress SPECT MPI

Journal of Nuclear Cardiology September/October 2008

Table 2. Distribution of enrollment volume by territory and by practice type in United States

No. of sites enrolling PS MPS patients during 20 consecutive days

No. of patients enrolled [mean (SD), median, range] United States (n ⫽ 72 sites) Private (n ⫽ 32 sites) Community (n ⫽ 21 sites) Tertiary care (n ⫽ 19 sites) Rest of world* (n ⫽ 17 sites)

<50 51–100 101–200 201–300 301–400 patients patients patients patients patients

48.3 (31.54), 40, 2–162 42.4 (31.25), 38, 2–162 45.4 (24.65), 40, 16–99 61.2 (36.33), 59, 17–153 99.8 (93.58), 58, 25–352

45 (62%) 24 (75%) 14 (67%) 7 (37%) 8 (47%)

22 (31%) 6 (19%) 7 (33%) 9 (47%) 3 (18%)

5 (7%) 2 (6%) 0 3 (16%) 3 (18%)

0 0 0 0 2 (12%)

0 0 0 0 1 (6%)

*Includes sites in Argentina (4), Belgium (1), Brazil (1), Canada (5), Chile (2), Italy (2), and United Kingdom (2).

Table 3. Distribution of ACC/AHA pretest likelihood for CAD for all patients and by region, gender, and practice type in United States

Pretest likelihood for CAD* Low

Intermediate

High

Known CAD†

157 (3.0%) ⬍1% 2.3% 110 (3.3%) 27.3% 72.7% 27.3% 27.3% 45.4% 47 (2.8%)

1,001 (19.3%) 8.2% 10.8% 585 (17.5%) 36.9% 63.1% 28.1% 41.5% 30.4% 416 (24.7%)

1,335 (25.8%) 9.5% 17.5% 818 (24.5%) 36.9% 63.1% 25.5% 30.7% 43.8% 517 (30.7%

384 (7.4%) 3.8% 3.2% 224 (6.7%) 45.1% 54.9% 24.6% 30.8% 44.6% 160 (9.5%)

2,142 (41.4%) 24.5% 16.5% 1,599 (47.9%) 57.0% 43.0% 27.1% 45.1% 27.8% 543 (32.3%)

19.1% 80.9% 29.8% — 70.2%

50.5% 49.5% 20.4% — 79.6%

36.2% 63.8% 24.7% — 75.3%

60.6% 39.4% 25.6% — 74.4%

65.9% 34.1% 26.1% — 73.9%

All patients Very low Overall Male Female US sites (n ⫽ 72) Male Female Community Private Tertiary care Sites in rest of world‡ (n ⫽ 17) Male Female Community Private§ Tertiary care

5,174 48% 52% 3,477 (67.2%) 46.8% 53.2% 26.6% 39.4% 34.0% 1,697 (32.8%) 51.2% 48.8% 24.4% — 75.6%

*ACC/AHA criteria (Gibbons et al11). † Prior history of MI, revascularization, or documented CAD by angiography. ‡ Includes sites in Argentina (4), Belgium (1), Brazil (1), Canada (5), Chile (2), Italy (2), and United Kingdom (2). § There were no self-declared private imaging clinics outside the United States.

referred for exercise (mean, 66.2 years vs 58.6 years). Most PS procedures (83%) were performed in outpatients; this frequency was 90% at private/community sites and 74% at tertiary care centers. Whereas only 13% of US procedures were conducted on inpatients, the proportions were 21%, 25%, 52%, and 60% in Chile, Argentina, Belgium, and Italy, respectively (Table 1).

Distribution by race reflected regional and cultural factors, in that 86% of the population enrolled in Chile, 40% in Argentina, and no Brazilians were classified as Hispanic; no black patients were tested in Argentina, Belgium, Chile, or Italy. Overall, the frequencies of presenting medical conditions were similar at US sites and sites outside the United States, although the frequency of congestive heart

Journal of Nuclear Cardiology Volume 15, Number 5;687-97

Johnson et al Outcomes and utilization study of pharmacologic stress SPECT MPI

693

Table 4. Frequencies of medical conditions and risk factors for all patients, by region, and by practice type in United States

US sites (n ⴝ 3,477 [67.2%] at 72 sites)

Medical history and risk factors Hypertension Elevated cholesterol level Smoking (present or former) Positive family history of CAD IDDM NIDDM CHF Angina Taking ␤-blocker* Taking CCB Taking nitrates Prior MI Prior CABG or PCI LBBB Asthma COPD

All sites (N ⴝ 5,174 at 89 sites) [n (%)]

Private (n ⴝ 1,357 [39.0%] at 32 sites) [n (%)]

Community (n ⴝ 957 [27.5%] at 21 sites) [n (%)]

Tertiary care (n ⴝ 1,163 [33.4%] at 19 sites) [n (%)]

Non-US sites (n ⴝ 1,697 [32.8%] at 17 sites) [n, %)]

3,602 (70%) 2,944 (57%) 2,200 (43%)

953 (70%) 810 (60%) 613 (45%)

686 (72%) 587 (61%) 400 (42%)

818 (70%) 643 (55%) 449 (39%)

1,145 (67%) 904 (53%) 738 (43%)

2,075 (40%)

553 (41%)

403 (42%)

485 (42%)

634 (37%)

484 (9%) 1,011 (20%) 463 (9%) 1,399 (27%) 663 (13%) 331 (6%) 351 (7%) 1,222 (23%) 1,483 (29%) 220 (4%) 391 (8%) 337 (7%)

131 (10%) 254 (19%) 123 (9%) 355 (26%) 169 (12%) 73 (5%) 76 (6%) 351 (10%) 471 (14%) 56 (2%) 111(3%) 125 (4%)

105 (11%) 178 (19%) 102 (11%) 251 (26%) 122 (13%) 59 (6%) 43 (5%) 211 (6%) 312 (9%) 29 (⬍1%) 77 (2%) 76 (2%)

144 (12%) 225 (19%) 122 (10%) 251 (22%) 137 (12%) 44 (4%) 75 (6%) 262 (8%) 351 (10%) 32 (⬍1%) 103 (3%) 87 (3%)

104 (6%) 354 (21%) 116 (7%) 542 (32%) 235 (14%) 155 (9%) 157 (9%) 398 (8%) 349 (7%) 103 (6%) 100 (6%) 89 (2%)

IDDM, Insulin-dependent diabetes; NIDDM, non–insulin-dependent diabetes; CHF, congestive heart failure; CCB, calcium channel blocker; CABG, coronary artery bypass graft; PCI, percutaneous coronary intervention; LBBB, left bundle branch block; COPD, chronic obstructive pulmonary disease. *Patients taking antianginal drugs chronically before presenting for MPS.

failure tended to be slightly higher and the frequency of angina slightly lower at US sites versus sites outside the United States. The frequencies of medical conditions were similar at US private, community, and tertiary care centers (Table 4). Predictably, the most commonly reported risk factors were hypertension, elevated cholesterol level, smoking, and family history. Of the 5,174 patients enrolled, 663 (13%) were taking ␤-blockers, 331 (6%) were taking calcium channel blockers, and 351 (7%) were taking nitrates at the time of their MPS study (Table 4). Only 181 subjects (35%) discontinued antianginal treatment before MPS. Most PS procedures (76%) were conducted with technetium 99m sestamibi alone (53%) or Tc-99m tetrofosmin alone (23%). Only 13% of PS procedures used thallium 201 alone, and most Tl-201– only studies were performed at non-US sites (20% of their patients) and at US tertiary care sites (17% of their patients). Only 11% of all patients completed dual-isotope procedures, and nearly all (96%) were conducted at US sites (Table 5).

DISCUSSION This study is the largest international, prospective registry of PS MPS investigating prognostic, resource utilization, and AE characteristics of this modality. The study was designed to address several specific limitations in our current knowledge of nuclear cardiology practice and PS MPS. First, little is known regarding the patterns of patient referral for SPECT MPS. This study examines this on several levels. Although only limited information will be available to compare patients referred to PS versus exercise stress, more in-depth analyses of patients referred to PS and to each stress agent are planned. These analyses include assessment of the role of geographic variability both within the US sites and between the US sites and sites outside the United States on these referral patterns. Similarly, the distribution of patients with known prior CAD and patients without prior CAD with various levels of likelihood of

694

Johnson et al Outcomes and utilization study of pharmacologic stress SPECT MPI

Journal of Nuclear Cardiology September/October 2008

Table 5. Radiopharmaceutical use by region

US sites

Radiopharmaceutical Tc-99m sestamibi Tc-99m tetrofosmin Tl-201 Tl-201 and Tc-99m

All sites* (N ⴝ 5,114) at 89 sites) [n (%)]

Private (n ⴝ 1,338) [n (%)]

Community (n ⴝ 9,947) [n (%)]

Tertiary care (n ⴝ 1,140) [n (%)]

Non-US sites (n ⴝ 1,698) [n (%)]

2,723 (53%) 1,176 (23%) 644 (13%) 571 (11%)

857 (64%) 317 (24%) 5 (⬍1%) 159 (5%)

367 (39%) 288 (30%) 98 (10%) 194 (20%)

601 (53%) 147 (13%) 199 (17%) 193 (17%)

898 (53%) 424 (25%) 342 (20%) 25 (1%)

*Sixty subjects had missing radiopharmaceutical information.

CAD within these referral patterns will be examined. Furthermore, the confounding effects of clinical and historical characteristics, practice setting (private, community hospital, or tertiary care), physician specialty, and comorbidities on these referral patterns will be examined. Finally, the relationship between the test results and the referral patterns can be determined and compared. Of the 9.1 million stress MPS procedures completed in the United States in 2005, stress was induced pharmacologically in 4.3 million patients (46%) with adenosine (61%), dipyridamole (31%), and dobutamine (8%).5 Hence it is important to have a better understanding of outcomes and resource utilization associated with these agents. Most prior reports of post-MPS outcomes and resource utilization were based on data from a single academic/tertiary care site or a few sites, usually in the United States. In an attempt to minimize the biases introduced by the reduction in patient event rates by revascularization procedures triggered by the MPS results, most prior studies removed or censored those patients who underwent revascularization in the first 60 to 90 days after MPS.15,16 This approach of considering only medically treated patients, however, has recently been shown to introduce as significant a bias as it was anticipated to correct.17 Although several studies since have included all patients, more information and current information regarding referral patterns, shortterm outcomes, and measures of accuracy of MPS with the 3 available PS agents are clearly needed. The current study will permit assessment of the independent and incremental prognostic value of PS MPS in a large, heterogeneous population, with sufficient power to also assess its prognostic value as a function of PS agent used, practice type, geographic location, and various patient demographic variables. The currently available PS agents induce an array of subjective AEs that are distressing to patients but

are generally managed well by experienced clinical staff. Adenosine and dipyridamole are contraindicated or are used with great caution in patients with histories of asthma and/or atrioventricular conduction disturbances, and dobutamine may cause or exacerbate ventricular ectopy.6-10 Although safety data for each individual drug are available, no single, prospective study has examined the incidence, intensity, and timing of AEs relative to administration of all 3 drugs; the impact of shorter infusion times or adjunctive low-level exercise on AEs; or the need for pharmacologic reversal of PS procedures. Finally, the proportion of patients unable to undergo vasodilator stress testing because of pulmonary contraindications is uncertain. The results from this study will also allow comparisons of the nature, intensity, and timing (onset and duration) of the AEs induced by the available PS agents, as well as determination of whether alterations in these variables (eg, fewer or less intense AEs or AEs of a shorter duration) are associated with different PS MPS protocols, patients, or other confounding factors. This study also provides a single, current PS agent safety registry that quantifies the intensity and duration of side effects and allows assessment of the prognostic value of PS MPS across a wide number of covariables. The study will allow for extensive model development and validation to assess patient populations in future imaging studies. The study was designed to acquire results that could be generalized across a wide population of PS MPS patients but that could also be interrogated to analyze possible relationships among a variety of variables while minimizing the sources of noise and variability. Clinical sites recorded all patients who presented to their laboratories over a 20-day working period of time that the laboratory was operating (Figure 1) and tested patients according to their usual practices and protocols but also recorded data in a simple format that sought to minimize variability.

Journal of Nuclear Cardiology Volume 15, Number 5;687-97

Johnson et al Outcomes and utilization study of pharmacologic stress SPECT MPI

Limitations Relatively few sites outside the United States (17/ 89) participated in this study, none declared themselves private practices, and the 3 highest-enrolling sites in the study were in Brazil (n ⫽ 352) and Canada (n ⫽ 272 and n ⫽ 201). Because of these imbalances in enrollment, the overall results are heavily weighted with American data, and limited inferences can be made with the data from outside the United States. The study and data management procedures were designed to allow simple data acquisition on a single case report page. For logistic reasons, the study was not monitored, and with few exceptions, data were not queried, although 15% of the data were subjected to quality assurance. Images acquired with a variety of equipment during many different PS and radiopharmaceutical protocols were evaluated by individual investigators trained during Web-based investigators’ meetings to grade SPECT images with the 5-point, 17-segment American Society of Nuclear Cardiology model. More rigorous procedures may be expected to reduce experimental variability. Finally, this study also has the limitations of all nonrandomized studies. It is possible that the pattern of recruitment at certain sites may have introduced an unrecognized bias. Similar to other registries, we are limited by the referral and recruitment patterns of our sites. Nonetheless, we believe that by use of a patient log at the sites, as well as careful scrutiny and modeling of the data, our results will be the most generalizable results to date and will likely reveal the extent of the generalizability of single-site studies. Conclusion This PS MPS utilization and outcomes study provides a unique opportunity to extend our knowledge of the patient populations, the conduct and clinical utilization at diverse clinical sites, the results of SPECT studies, the prognostic value of PS MPS, and the comparative safety and accuracy of the available PS agents. This report also provides quantitative data on the distribution of patients presenting for PS MPS and indicates that private, community, and tertiary laboratories image similar populations of patients with regard to pretest likelihood for CAD and medical history. Acknowledgment This study was funded by King Pharmaceuticals Research and Development. The investigators and/or their institutions were paid to conduct this research. Dr Barrett and Mr Jenkins are employees of King Pharmaceuticals Research and Development. Dr Johnson is a former employee of King Pharmaceuticals Research and Development and is currently employed by

695

Premier Research Group, Limited (Cary, NC). Drs Hachamovitch, Massaro, and Udelson were paid consultants.

References 1. Hachamovitch R, Berman DS. The use of nuclear cardiology in clinical decision making. Semin Nucl Med 2005;35:65-72. 2. Amanullah AM, Kiat H, Hachamovitch R, Cabico JA, Cohen I, Friedman JD, et al. Impact of myocardial perfusion single-photon emission computed tomography on referral to catheterization of the very elderly. Is there evidence of gender-related referral bias? J Am Coll Cardiol 1996;28:680-6. 3. Beller GA, Zaret BL. Contributions of nuclear cardiology to diagnosis and prognosis of patients with coronary artery disease. Circulation 2000;101:1465-78. 4. Shaw LJ, Hachamovitch R, Berman DS, Marwick TH, Lauer MS, Heller GV, et al. The economic consequences of available diagnostic and prognostic strategies for the evaluation of stable angina patients: An observational assessment of the value of precatheterization ischemia. Economics of Noninvasive Diagnosis (END) Multicenter Study Group. J Am Coll Cardiol 1999;33:661-9. 5. IMV Medical Information Division. 2005 Nuclear medicine census market summary report. Des Plaines (IL): IMV; 2006. 6. Cerqueira MD, Verani MS, Schwaiger M, Heo J, Iskandrian AS. Safety profile of adenosine stress perfusion imaging: Results from the Adenoscan multicenter trial registry. J Am Coll Cardiol 1994;23:384-9. 7. Lette J, Tatum JL, Fraser S, Miller DD, Waters DD, Heller G, et al. Safety of dipyridamole testing in 73,806 patients: The multicenter dipyridamole safety study. J Nucl Cardiol 1995;2:3-17. 8. Ranhosky A, Kempthorne-Rawson J. The safety of intravenous dipyridamole thallium myocardial perfusion imaging. Intravenous Dipyridamole Thallium Imaging Study Group. Circulation 1990; 81:1205-9. 9. Dakik HA, Vempathy H, Verani MS. Tolerance, hemodynamic changes, and safety of dobutamine stress perfusion imaging. J Nucl Cardiol 1996;3:410-4. 10. Johnston DL, Daley JR, Hodge DO, Hopfenspirger MR, Gibbons RJ. Hemodynamic responses and adverse effects associated with adenosine and dipyridamole pharmacologic stress testing: A comparison in 2,000 patients. Mayo Clin Proc 1995;70:331-6. 11. Gibbons RJ, Balady GJ, Bricker JT, Chaitman BR, Fletcher GF, Froelicher VF, et al. ACC/AHA 2002 guideline update for exercise testing: A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on Exercise Testing). American College of Cardiology Web site. 2002. Available from: URL: http://acc.org/qualityandscience/ clinical/guidelines/exercise/exercise.pdf. Accessed October 2003. 12. Klocke FJ, Baird MG, Bateman TM, Berman DS, Carabello BA, Cerqueira MD, et al. ACC/AHA/ASNC guidelines for the clinical use of cardiac radionuclide imaging. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/AHA/ASNC Committee to Revise the 1995 Guidelines for the Clinical Use of Cardiac Radionuclide Imaging). American College of Cardiology Web site. 2003. Available from: URL: http://acc.org/qualityandscience/clinical/guidelines/ radio/index.pdf. Accessed October 2003. 13. Port SC. Imaging guidelines for nuclear cardiology procedures, part 2. J Nucl Cardiol 1999;6:53-84. 14. Newman SC. Biostatistical methods in epidemiology. New York: John Wiley & Sons; 2001. 15. Pryor DB, Harrell FE Jr, Lee KL, Rosati RA, Coleman RE, Cobb FR, et al. Prognostic indicators from radionuclide angiography in

696

Johnson et al Outcomes and utilization study of pharmacologic stress SPECT MPI

medically treated patients with coronary artery disease. Am J Cardiol 1984;53:18-22. 16. Staniloff HM, Forrester JS, Berman DS, Swan HJ. Prediction of death, myocardial infarction, and worsening chest pain using thallium scintigraphy and exercise electrocardiography. J Nucl Med 1986;27:1842-8. 17. Hachamovitch R, Hayes S, Friedman J, Cohen I, Berman D. Stress myocardial perfusion SPECT is clinically effective and costeffective in risk-stratification of patients with a high likelihood of CAD but no known CAD. J Am Coll Cardiol 2004;43:200-8.

APPENDIX A. SITES, INVESTIGATORS, AND COLLABORATORS Argentina: O. Masoli, MD (Hospital Municipal Coseme Argerich Buenos Aires); A. Meretta, MD (Instituto Cardiovascular De Buenos Aires, Buenos, Aires); S. Molteni, MD (Hospital Israelita, Capital Federal); A. Orlandini, MD (Instituto Cardiovascular Rosario, Rosario) Belgium: T. Benoit, MD (C.H.R. Huy, Huy); P. Franken, MD (Nuclear Medicine/AZ VUB, Brussels); J. Vanoverschelde, MD (Cliniques Universitaires St. Luc, Brussels); D. Wyndaele, MD (Heilig Hartziekenhuig, Tienen) Brazil: C. Meneghetti, MD (Instituto do Coracao, Sao Paulo) Canada: N. Filipchuck, MD (Cardiology Plus, Calgary, Alberta); N. Filipchuck, MD (Foothills Hospital, Calgary, Alberta); M. Freeman, MD (St. Michael’s Hospital, Toronto, Ontario); M. Kiess, MD (St. Paul’s Hospital; Vancouver, British Columbia); T. Ruddy MD (University of Ottawa Heart Institute, Ottawa, Ontario); R. Taillefer, MD (Hotel-Dieu du CHUM, Montreal, Quebec) Chile: I. Godoy, MD (Hospital Clinico Universidad Catolica, Santiago); F. Lanas, MD (Clinical Alemana de Temuco; Temuco) Italy: G. Bisi, MD (Universita Di Torino, Torino); A. Cuocolo, MD (Centro per la Medicina Nuclear CNR, Napoli), R. Giubbini, MD (Spedali Civili-Brescia, Brescia); C. Marcassa, MD (Salvatore Maugeri Foundation, Veruno); P. Pieri, MD (M. Bufalini Hospital, Cesena); P. Zanco, MD (Medicina Nucleara-Ospedale, Vicenza) United Kingdom: M. Metcalfe, MD (Aberdeen Royal Infirmary, Aberden); L. Prvulovich, MD (Middlesex Hospital, London); A. Tweddle, MD (Royal Hull Hospital, London); R. Underwood, MD (Royal Brompton Hospital, London) United States O. Akinboboye, MD (St. Francis Hospital; Roslyn, NY); L. Altschul, MD (South Bay Cardiovascular Associates, West Islip, NY); J. Arrighi, MD (VA Connecticut Healthcare System, West Haven, CT); J. Baird, MD (Missouri Heart Center, Columbia, MO); B. Bart, MD (Hennepin County Medical Center, Minne-

Journal of Nuclear Cardiology September/October 2008

apolis, MN); S. Borges- Neto, MD, (Duke University Medical Center, Durham, NC); E. Botvinick, MD, (UCSF Nuclear Medicine, San Francisco, CA); R. Braastad, MD (Illinois Heart and Lung Associates, Normal, IL); C. Brown, MD, (The Heart Group, Mobile, AL); K. Brown, MD (Fletcher Allen Health Care, MCHV Campus, Burlington, VT); D. Calnon, MD (Midwest Cardiology Research Foundation, Columbus, OH); S. Carollo, MD (Bergen Cardiology Specialists, P.C., Omaha, NE); A. Chai, MD (Idaho Cardiology Associates, Meridian, ID); P. Colman, MD (Northern California Medical Associates, Santa Rosa, CA); L. Conway, MD (Mystic Cardiology Associates, Medford, MA); D. Courtade, MD (Cardiology Associates PSC, Edgewood, KY); R. Des Prez, MD (Oklahoma Heart Institute, Tulsa, OK); K. Desai, MD (Rhode Island Cardiovascular, Woonsocket, RI); N. Dhruva, MD (Cardiac Disease Specialists, Fayetteville, GA); E. Diltz (Cardiology Associates of East Tennessee, Knoxville, TN); F. Dixon, MD (Austin Cardiovascular Associates, Austin, TX); S. Edell, DO (Delaware SPECT Imaging Center, Newark, DE); S. Eisenberg, MD (Atlanta Heart and Vascular Research Group, Atlanta, GA); P. Farrell, MD (Jacksonville Heart Center, PA, Jacksonville, FL); E. Flores, MD (Georgia Heart Specialists, Covington, GA); J. Foster, Jr., MD (Cardiology Associates of Northern Mississippi, Tupelo, MI); R. Gal, MD (Milwaukee Heart Institute, Milwaukee, WI); M. Gerson, MD (University of Cincinnati Medical Center, Cincinnati, OH); A. Gomez, MD (Cardiovascular Research Center of South Florida, Miami, FL); T. Goraya, MD (Michigan Heart, Ypsilanti, Michigan); L. Gordon, MD (Medical University of South Carolina MUSC, Charleston, SC); D. Griffin, MD (Heart Clinic Arkansas, Little Rock, AR); C. Hansen, MD (Temple University School of Medicine, Philadelphia, PA); B. Haraden, MD (Louisville Cardiology Medical Group, Louisville, KY); H. Haronian, MD (Westerly Hospital, Westerly, RI); H. Haught, MD (The Heart Center, Huntsville, AL); G. Heller, MD (Hartford Hospital, Hartford, CT); L. Heller, MD (Cardiac Disease Specialists, Atlanta, GA); T. Hilton, MD (Jacksonville Heart Center, Jacksonville, FL); D. Hinchman, MD (Idaho Cardiology Associates, Boise, ID); R. Hoffman, MD (Northeast Cardiology Associates, Bangor, ME); T. Holly, MD (Northwestern University, Chicago, IL); D. Jain, MD (Drexel University School of Medicine, Philadelphia, PA); D. Joyce, MD (North Ohio Research, Ltd, Loraine, OH); S. Kapadia, MD (Cardiovascular Associates of Virginia, Richmond, VA); R. Karlsberg, MD (Access Clinical Trials, Beverly Hills, CA); M. Klapholz, MD (St. Vincent’s Hospital and Medical Center, New York, NY); C. Lim, MD (Asheville Cardiology, Asheville, NC); W. Martin, MD (St. Louis VA Medical Center, St. Louis, MO); B. Melek, MD (Tulane Univer-

Journal of Nuclear Cardiology Volume 15, Number 5;687-97

Johnson et al Outcomes and utilization study of pharmacologic stress SPECT MPI

sity Health Sciences Center, New Orleans, LA); J. Mieres, MD (North Shore University Hospital, Manhasset, NY); D. Miller, MD (St. Louis University, St. Louis MO); A. Movahed, MD (Pitt County Memorial Hospital, Greenville, NC); J. Murillo, MD (Cardiology Consultants, Norfolk, VA); M. Nathan, MD (Cardiovascular Consultants Medical Group, Walnut Creek, CA); R. Parmar, MD (Heart Place-Landry, Dallas, TX); C. Parrot, MD (Providence Hospital, Mobile, AL); J. Post, MD (Heart and Vascular Institute of Florida); F. Prigent, MD (Winthrop University Hospital, Mineola, NY); T. Rosamond, MD (Kansas University Medical Center, Kansas City, KS); J. Rosenblatt, MD (Maine Cardiology Associates, Portland, ME); G. Schuyler, MD (Heart and Vascular Institute of Florida, St. Petersburgh, FL); R. Schwartz, MD (University of Rochester, Rochester, NY); M. Shah, MD (Mount Carmel Center for Clinical Research, Columbus, OH); D. Sheps, MD (University of Florida, Gainsville, FL); D. Shonkoff, MD (Cardiovascular Group PC, Lawrenceville, GA); T. Sias, MD

697

(University Cardiovascular Services, and St. Mary’s, Huntington, WV); J. Smith, MD (Cardiovascular Associates, Louisville, KY); C. Sotolongo, MD (Diagnostic Cardiology, Jacksonville, FL); E. Spiegler, MD (St. Agnes Healthcare, Baltimore, MD); P. Tikemeier, MD (Miriam Hospital, Providence, RI); M. Traboulssi, MD (North Ohio Heart Center, Sandusky, OH); M. Treuth, MD (Delmarva Heart Research, Salisbury, MD); M. Vacante, MD (North Ohio Heart Center, Elyria, OH); R. Vaccarino, MD (Brooklyn Nuclear SPECT Imaging, Brooklyn,(1) Brooklyn,(2) and Rye Brook, NY); W. Van Decker, MD (Medical College of Pennsylvania, Philadelphia, PA); F. Wackers, MD (Yale University School of Medicine, New Haven, CT); M. Walsh, MD (The Care Group, LLC, Indianapolis, IN); B. Walters, MD (Baltimore Heart Associates, Randallstown, MD); F. Whittier, MD (United Health Network, Canton, OH); D. Wolinsky, MD (Albany Associates in Cardiology, Albany, NY); F. Wood, MD (Salus Clinical Research, Murrieta, CA); G. Yurow, MD (Mid-Atlantic Cardiovascular Associates, Baltimore, MD)