Association of patients’ perception of health status and exercise electrocardiogram, myocardial perfusion imaging, and ventricular function measures

Association of patients’ perception of health status and exercise electrocardiogram, myocardial perfusion imaging, and ventricular function measures Jennifer A. Mattera, MPH,a Carlos Mendes de Leon, PhD,b Frans J. Th. Wackers, MD,d Christianna S. Williams, MPH,e Yongfei Wang, MS,c and Harlan M. Krumholz, MDa,d,e,f New Haven and Middletown, Conn, and Chicago, Ill

Background Patients’ viewpoint of their health status is increasingly used as an important outcome measure of the success of treatments. Because clinicians rarely formally measure patients’ health-related quality of life, the question arises whether noninvasive testing for ischemia can provide similar information regarding physical functioning and general health perception.

Methods We measured physical functioning and general health status with the Medical Outcomes Study Short Form (SF-36) survey in 195 consecutive patients (68% male, mean age 55.6 ± 11.1 years) referred for exercise testing with myocardial perfusion imaging. The survey was completed immediately before the exercise test.

Results In the multivariate analysis, the strongest predictor of physical functioning and general health perception was metabolic equivalents. However, the best model, including demographic, clinical, and test variables, predicted only 14% of the variation in physical functioning and 10% of the variability in general health perception.

Conclusions The variation in physical functioning and general health perception, as measured by the SF-36, among patients referred for exercise testing is not predicted well by the results of the test. As expected, several test results are significantly associated with physical functioning and general health perception; however, there was substantial overlap among individual patients, suggesting that the parameters are poor surrogates for the actual assessment of the domains. If these domains are deemed important to tracking patient outcomes, then they should supplement the current assessments of these patients. (Am Heart J 2000;140:409-18.)

See related Editorial on page 359. In patients with known or suspected ischemic heart disease, clinicians routinely use exercise electrocardiographic (ECG) testing, exercise myocardial perfusion imaging, and ventricular function tests to detect coronary artery disease, evaluate myocardial ischemia and the effectiveness of treatments, estimate prognosis, and determine functional capacity.1-5 Increasingly, the patient’s perception of his or her health status or healthrelated quality of life (HRQOL) is used as an indicator of the effectiveness of treatments.6-12 The HRQOL From the aCenter for Outcomes Research and Evaluation, Yale-New Haven Hospital; bRush Institute for Healthy Aging, Rush-Presbyterian-St. Luke’s Medical Center, Chicago; Yale University School of Medicine, cDepartment of Medicine and dSection of Cardiovascular Medicine, and the eDepartment of Epidemiology and Public Health, New Haven; and fQualidigm, Middletown. Submitted October 11, 1999; accepted March 21, 2000. Reprint requests: Harlan M. Krumholz, MD, Yale University School of Medicine, 333 Cedar St, Room IE.61 SHM, New Haven, CT 06520-8025. E-mail: [email protected] Copyright © 2000 by Mosby, Inc. 0002-8703/2000/$12.00 + 0 4/1/108518 doi:10.1067/mhj.2000.108518

refers to a patient’s perception of the impact of disease or treatments on his or her daily physical, functional, emotional, psychological, and social well-being.13,14 Because few invasive cardiovascular interventions have been proven to prevent death, the goal of many interventions is to improve the patient’s quality of life by reducing symptoms of angina and improving functional status and well-being. Because few clinicians formally measure HRQOL, the question arises whether information from the exercise testing and imaging can substitute for an assessment of HRQOL. A few studies have examined the association between exercise testing and HRQOL, but the results are mixed, with some studies reporting poor to modest correlation15-20 and others suggesting that exercise performance may be used as a surrogate for quality of life.11 One study examined the relation between the preoperative extent of myocardial viability, as measured by positron emission imaging, with improvement in patients’ functional status.21 We are not aware of other studies that have compared exercise myocardial perfusion imaging results with HRQOL. Accordingly, our specific objective was to determine the association of sev-

American Heart Journal September 2000

410 Mattera et al

Table I. PF and GHP scores by patient demographics (n = 195) PF

Age (y) <55 55-64 ≥65 Sex Female Male Race White Black Unknown Education
n (%)

Mean

SD

98 (50) 53 (27) 44 (23)

76.3 73.9 69.2

21.0 24.8 27.4

62 (32) 133 (68)

64.0 78.8

27.3 20.2

121 (62) 30 (15) 44 (23)

74.5 61.2 74.5

22.2 28.8 20.0

24 (13) 38 (20) 87 (45) 43 (22)

52.0 72.5 76.4 84.4

27.0 23.0 22.1 14.2

GHP P value

Mean

SD

59.8 61.0 68.3

21.3 22.1 17.8

60.5 62.8

21.4 20.8

62.8 50.2 68.1

20.3 24.2 17.5

51.8 61.6 62.6 68.2

20.1 22.6 20.7 16.9

.3

.08

<.0001

.5

.004

.004

<.001

eral cardiovascular noninvasive testing measures (exercise duration, metabolic equivalents [METS], exerciseinduced chest pain, exercise ECG result, myocardial perfusion imaging result, myocardial perfusion defect score, and left ventricular ejection fraction) with patient-reported physical functioning (PF) and general health perception (GHP). To address this objective, we evaluated these domains in 195 consecutive patients who were referred for exercise ECG treadmill testing with myocardial perfusion imaging.

Methods Study design A cross-sectional study was done to evaluate the patient’s perceived PF and general health status in consecutive patients immediately before exercise ECG treadmill testing with myocardial perfusion imaging and radionuclide left ventricular function assessment.

Study population The patient population consisted of a convenience sample of 238 consecutive in-patients and out-patients at Yale-New Haven Hospital referred for exercise ECG testing with myocardial perfusion imaging during 2 time periods: September 1993 to March 1994 and September 1994 to January 1995. Patients were referred for detection of coronary artery disease and evaluation of myocardial ischemia. Patients completed the Medical Outcomes Study Short Form (SF-36) health status survey immediately before testing. We were unable to get health status surveys in 22 (9%) patients because of refusal (n = 8), patients not having reading glasses (n = 6), and incomplete surveys (n = 8). We excluded 21 (9%) patients with significant comorbid disease (post–organ transplantation status, terminal cancer, and peripheral vascular disease) to minimize the heterogeneity of the population and to minimize the influence of other non-cardiac-related health problems on patient-

P value

.02

perceived health. The final study population consisted of 195 patients.

Assessment of PF and GHP The SF-36, a generic HRQOL survey that has demonstrated validity and reliability in a number of disease populations, including patients with angina, was used to assess patients’ perception of their overall health (GHP) and PF ability in daily life.22 The SF-36 survey covers 8 domains of health with 8 multi-item subscales: PF, physical role functioning, social functioning, bodily pain, mental health, emotional role functioning, vitality, and GHP. Each of the domains is scored separately. We chose the GHP domain as a measure of overall perceived health status and selected the PF index because it was most likely to be associated with exercise capacity on a treadmill and the presence of myocardial ischemia. In addition, we specifically wanted to address whether exercise duration could serve as a surrogate for patient-reported PF and overall health status perception. No attempt was made to help the patients complete the questionnaire. The SF-36 PF and GHP items were scored and transformed into a 0 to 100 continuous scale with 100% representing no impairment in daily PF or excellent GHP.23

Noninvasive cardiovascular testing All patients underwent exercise treadmill ECG testing, exercise and rest myocardial perfusion imaging, and resting firstpass radionuclide ventriculography. Exercise ECG treadmill test. The exercise ECG treadmill test was performed with a Bruce or modified Bruce protocol. Heart rate, 12-lead ECG, and blood pressure were obtained at baseline, during each 3-minute stage of exercise, and at 5 minutes after exercise. In addition, the patient reported any symptoms during exercise, such as chest pain and shortness of breath. The total exercise time, METS, and exercise ECG result were obtained in each patient. The exercise ECG result was obtained from the final test report and was categorized as

American Heart Journal Volume 140, Number 3

Mattera et al 411

Table II. PF and GHP scores by patient clinical characteristics (n = 195) PF

Body mass index (kg/m2) <25 25-29 ≥30 Risk factors Hypertension No Yes Smoking history No Yes Family history of CAD No Yes Hyperlipidemia No Yes Diabetes No Yes History of MI No Yes History of PTCA or CABG No Yes β-blocker therapy No Yes Calcium-channel blockers No Yes Type of chest pain Asymptomatic Nonanginal Atypical angina Typical angina

n (%)

Mean

SD

40 (21) 80 (41) 75 (38)

78.2 78.4 67.2

26.0 21.1 23.5

116 (59) 79 (41)

74.8 72.9

23.3 24.2

149 (76) 46 (24)

75.1 70.7

23.4 24.3

111 (57) 84 (43)

71.3 77.7

25.1 21.1

106 (54) 89 (46)

72.7 75.7

24.2 22.9

163 (84) 32 (16)

76.4 62.4

22.0 28.3

147 (75) 48 (25)

75.2 70.7

23.7 23.1

143 (73) 52 (27)

74.3 73.4

24.2 22.1

130 (67) 65 (33)

75.6 71.1

23.8 23.2

132 (68) 63 (32)

76.2 69.6

23.0 24.5

50 (26) 23 (12) 86 (44) 35 (18)

85.1 86.2 66.1 70.1

18.5 13.9 25.1 22.8

GHP P value

Mean

SD

63.0 67.0 56.3

24.5 18.6 20.2

64.2 58.9

20.2 21.7

64.3 54.8

19.9 22.9

61.3 63.0

20.2 22.0

63.5 60.3

20.5 21.4

64.1 51.8

19.9 23.6

63.8 56.8

20.9 20.6

63.7 57.4

20.7 21.1

64.3 57.5

21.7 18.7

64.3 57.2

19.4 23.4

72.8 70.7 53.8 61.2

16.3 14.7 21.3 21.1

.005

P value .004

.6

.09

.3

.001

.06

.6

.4

.3

.01

.002

.1

.05

.6

.06

.2

.03

.07

.04

<.0001

<.001

MI, Myocardial infarction; PTCA, percutaneous transluminal coronary angioplasty; CABG, coronary artery bypass graft surgery; CAD, coronary artery disease.

normal, nondiagnostic, or ischemic. The exercise ECG result was considered nondiagnostic if baseline ECG abnormalities were present that compromised interpretation of the exercise ECG or if the patient did not exercise sufficiently to achieve an adequate heart rate. The exercise ECG result was classified as ischemic if >1 mm of flat or downsloping ST-segment depression was present in 3 consecutive beats. Myocardial perfusion imaging. Myocardial perfusion imaging was performed in all patients after injection of 20 to 25 mCi of technetium 99m sestamibi at peak exercise. The patient underwent exercise and rest myocardial perfusion imaging with single photon emission computed tomography or the planar technique. The myocardial perfusion imaging result was categorized as normal, scar, or ischemia. Scar indicated the presence of a myocardial perfusion defect that remained unchanged or fixed from exercise to rest imaging. Ischemia was present if a myocardial perfusion defect was present at exercise and was either absent or smaller at rest.

The size of the myocardial perfusion abnormality was quantified with previously validated software24 and categorized empirically as absent (no quantitative defect), small, moderate, or large. Left ventricular ejection fraction. The resting injection of 99mTc sestamibi was used to assess left ventricular function by first-pass technique when possible (n = 166). The left ventricular ejection fraction was calculated from the left ventricular phase of the first transit of the radioactive tracer bolus through the heart with validated software on the SIM-400 system (Scinticor, Milwaukee, Wis). Clinical and demographic variables. The physician conducting the exercise test obtained demographic and clinical history from the patient and the medical record. The patient’s chest pain symptoms were categorized as asymptomatic, nonanginal chest pain, atypical angina, and typical angina, according to 3 historic criteria relative to location, precipitation, and relief of discomfort.25 Typical angina was considered if the patient had all 3 of the following criteria: (1) substernal

American Heart Journal September 2000

412 Mattera et al

Table III. PF and GHP scores by patient noninvasive cardiovascular testing parameters (n = 195) PF n (%) Exercise ECG treadmill parameters Exercise minutes >12 9-12 6-9 <6 METS ≥12.9 10.2-12.8 7.1-10.1 7.0 <7.0 Chest pain during exercise Absent Present Exercise ECG results Normal Nondiagnostic or equivocal Ischemia (≥1 mm ST depression) Myocardial perfusion imaging parameters Results Normal Scar Ischemia Exercise defect score None Small Medium Large Left ventricular ejection fraction Normal (≥55%) 45-54 35-44 <35 Missing

Mean

SD

36 (18) 95 (49) 48 (25) 16 (8)

91.0 78.3 62.2 46.1

10.2 20.0 24.1 23.3

18 (9) 58 (30) 68 (35) 42 (22) 9 (5)

91.7 83.5 73.9 61.8 36.7

9.4 16.6 21.8 25.7 17.9

155 (79) 40 (21)

77.6 60.2

22.0 24.8

76 (39) 71 (36) 48 (25)

77.8 66.9 78.8

22.0 25.5 20.8

119 (61) 25 (13) 51 (26)

76.2 64.8 73.7

24.0 22.5 22.5

84 (43) 57 (29) 21 (11) 33 (17)

76.4 73.3 70.6 71.5

24.0 22.1 27.4 22.9

118 (60) 33 (14) 8 (4) 2 (1) 34 (17)

74.2 75.2 86.3 95.0 68.2

24.2 21.9 14.8 0.0 24.3

GHP P value

Mean

SD

71.4 62.1 60.4 45.9

18.6 19.8 21.8 20.8

76.3 63.1 64.3 53.0 51.7

13.5 21.3 19.6 21.6 19.3

64.4 53.0

20.3 21.1

62.1 57.6 68.6

20.2 22.4 18.4

64.1 56.3 60.2

20.7 18.8 22.2

65.5 61.2 51.5 61.5

20.1 21.2 21.2 21.2

62.4 61.8 71.8 81.0 57.6

20.8 20.0 14.0 8.5 22.3

<.0001

P value

<.0001

<.0001

<.005

<.0001

.002

.005

.02

.03

.2

.6

.05

.2

.3

LVEF, Left ventricular ejection fraction.

discomfort, (2) discomfort precipitated by physical exertion, and (3) discomfort relieved within 10 minutes by rest or nitroglycerin. Atypical angina was considered if only 2 of the 3 above criteria were present. Nonanginal chest pain was considered if only 1 or none of the 3 criteria was present. Patients who denied discomfort above the level of the diaphragm were considered to be asymptomatic.

Statistical analysis Bivariate analysis. The bivariate analysis was used to test if individual noninvasive testing parameters, patient demographics, and clinical characteristics were associated with patientperceived PF and GHP. Continuous dependent variables (GHP and PF scores) were compared with ordinal or categoric independent variables by Student t test or analysis of variance for GHP and the Wilcoxon rank-sum test or Kruskal-Wallis test for nonnormal distributed PF scores. Simple linear regression was used to test for a linear trend between continuous or ordinal independent variables and GHP and PF scores. The Pearson

correlation coefficient was used to evaluate the strength of the linear relation between continuous noninvasive testing parameters and PF and GHP scores. An α level of .05 (2-tailed) was used for all tests and a P value <.01 was considered statistically significant, adjusting for multiple comparisons. Multivariate analysis. Multivariate analysis was used to evaluate how much of the variation in patient-reported PF and GHP was explained by the noninvasive testing measures after adjusting for patient demographic and clinical variables. Multiple linear regression was done with JMP version 3.2 software (SAS Institute, Cary, NC). The distribution of the PF scores was skewed, and the GHP scores were near normal distribution. Because PF scores were skewed, we checked the robustness of using the linear regression model by also testing the model with PF recoded into 4 groups as an ordinal variable based on quartile distribution. We then examined the residuals to ensure that they were normally distributed. Noninvasive measures were checked for collinearity with the Pearson or Spearman correlation coefficient. When 2 noninvasive measures were collinear (r > 0.35), such as exercise

American Heart Journal Volume 140, Number 3

Mattera et al 413

Table IV. Multivariate analysis of PF Model 2†

Model 1* Coefficient Exercise ECG treadmill variables METS Chest pain during exercise Exercise ECG result Normal Nondiagnostic Ischemia Myocardial perfusion imaging Normal Scar Ischemia Demographic variables Age Sex (female) Education level
SE

P value

Coefficient

SE

P value

.0001 .001

2.8048 –7.6657

0.6100 3.6443

.0001 .04

3.6862 –12.2408

0.5622 3.7143

— –4.5773 2.0863

— 3.3883 3.8749

— .2 .6

— –5.9036 –1.1261

— 3.1872 3.6401

— .06 .8

— –4.2630 4.5481

— 4.5196 3.5413

— .3 .2

— –4.0836 –0.6649

— 4.3397 3.6757

— .3 .8

–0.04003 –4.3083

0.1280 3.3813

.8 .2

— 11.4110 11.7107 16.4448

— 4.8070 4.4165 4.9763

— .02 .009 .001

— 0.0626 –14.3849 –4.4532

— 4.8036 3.5616 4.3999

— .9 .0001 .3

*Model 1 includes noninvasive test variables only (n = 195): degrees of freedom = 6, adjusted R2 = .3002, F ratio = 14.872, P = .0001. †Model 2 includes noninvasive test variables, demographic and clinical variables (n = 195): degrees of freedom = 14, adjusted R2 = .4116, F ratio = 10.694, P = .0001.

minutes and METS, the most clinically relevant measure (eg, METS) was included in the model. Separate multivariate models for PF and GHP were constructed by first including the 4 noninvasive measures (METS, chest pain during exercise, exercise ECG result, and myocardial perfusion imaging result) and then, in a stepwise fashion, adding patient demographic and clinical variables that were significant at an α level of ≤.10 in the bivariate analysis or based on clinical judgment. To determine the percent of variation in PF or GHP scores explained by the noninvasive cardiovascular testing parameters, we compared the coefficient of determination (R2) in the final model, adjusted for degrees of freedom, with the final model minus the noninvasive test variables.

Results Study sample The demographic characteristics of the study population are listed in Table I. Approximately 25% of the population was aged >65 years, and the majority was male (68%). The clinical characteristics of the patients are listed in Table II. Nearly 40% of the patients were obese, having a body mass index >30 kg/m2. Only 18% of the patients had typical angina, whereas nearly half were classified as having atypical angina. Approximately 25% of the patients had a history of myocardial infarction and prior

revascularization. The overall mean PF and GHP scores for the study population were higher than norms established by the SF-36 for patients with angina (74.0 vs 63.2 and 62.0 vs 52.0, respectively). However, the mean age for this study population was slightly younger compared with the SF-36 norms for patients with angina (55.6 years vs 59.7 years), and the percentage of women in this study was lower (32% vs 55%). In our study, women had significantly lower PF scores than men (P < .001); however, men and women did not differ in their GHP scores. Blacks had significantly lower PF and GHP scores than whites.

Noninvasive testing results The noninvasive testing results are shown in Table III. Two thirds of the patients exercised more than 9 minutes on the treadmill, and nearly 75% achieved greater than 7 METS. Nearly 40% of the patients had normal ECG treadmill tests and more than 60% of patients had normal exercise myocardial perfusion tests. Only one quarter of the patients demonstrated ischemia on the ECG treadmill test and by exercise myocardial perfusion imaging. Only a small percentage of the patients had a markedly abnormal resting left ventricular ejection fraction.

American Heart Journal September 2000

414 Mattera et al

Figure 1

Correlation between PF scores from SF-36 health status survey and exercise duration on ECG treadmill test.

Bivariate analysis of PF and GHP Demographic variables. The mean PF and GHP scores

were not statistically different between age groups (Table I); however, there was a trend for lower perceived PF with increasing age, whereas GHP increased with age. There was a significant linear trend of increasing PF and GHP scores with increasing education level. Clinical variables. Obesity (body mass index ≥30 kg/m2) and diabetes were associated with significantly lower PF and GHP scores (Table II). Anginal chest pain was also significantly associated with lower PF and GHP scores, the lowest scores occurring in patients with atypical angina rather than typical angina. A history of smoking, β-blocker therapy, and calcium-channel blocker medication was associated with lower GHP scores but not with PF scores. Noninvasive testing measures. The relations between noninvasive testing parameters and PF and GHP scores are shown in Table III. Total exercise minutes on the treadmill and METS were significantly associated with both PF and GHP scores; with increasing exercise time and METS there was a corresponding positive linear trend of PF and GHP scores (P < .0001). The correlation of exercise minutes and METS with PF and GHP scores was moderate (r = 0.52, 0.50 and 0.30, 0.28, respectively). Although there was a significant linear

relation between exercise minutes or METS and PF and GHP, there was wide variability on an individual patient level (Figures 1 and 2). Adjustment for age did not substantially reduce variability in PF and GHP scores as a function of exercise time (data not shown). The presence of chest pain during exercise was also associated with significantly lower PF and GHP scores. The exercise ECG result was associated with PF (P = .005) and GHP scores (P = .02); however, the lowest scores occurred in patients with a nondiagnostic ECG result rather than patients with ischemia. Patients with ischemia by exercise ECG testing had PF and GHP scores similar to or higher than patients with a normal study. Further analysis of the subgroup of patients with ischemia by exercise ECG testing showed that the presence of chest pain was associated with significantly lower PF and GHP scores; however, only one third of the patients had chest pain during exercise. Additionally, the patients with ischemia achieved the highest workload (11.0 METS) compared with patients with normal test (10.9 METS) and nondiagnostic test results (9.4 METS). The myocardial perfusion imaging result was associated with PF scores (P = .03) but not with GHP. Patients with a fixed defect or scar on the myocardial perfusion imaging study had the lowest PF and GHP scores.

American Heart Journal Volume 140, Number 3

Mattera et al 415

Figure 2

Correlation between GHP scores from SF-36 health status survey and exercise duration on ECG treadmill test.

Patients with ischemia by exercise myocardial perfusion imaging had slightly lower PF and GHP scores compared with normal patients. The size of the exercise myocardial perfusion defect was not significantly associated with PF or GHP. The rest left ventricular ejection fraction was not associated with PF or GHP; however, there were few patients with ejection fractions in the lower ranges.

Multivariate analysis of PF and GHP The results of the multivariate analysis including noninvasive cardiovascular testing measures alone (model 1) and with patient demographic and clinical variables (model 2) are shown in Tables IV and V. PF. The noninvasive cardiovascular testing parameters (model 1) significantly predicted PF scores (adjusted R2 = .30, P = .0001). The noninvasive measures explained 30% of the variance in PF. In model 2, after addition of patient demographic and clinical variables, all noninvasive measures together explained only 14% of the variation in PF. This finding suggests that a substantial part of the covariation between noninvasive measures and PF is accounted for by characteristics of the patient rather than the test results themselves.

Of the noninvasive measures, higher METS and the absence of chest pain during exercise were significant predictors of higher perceived PF. Ischemia detected on the exercise ECG and by myocardial perfusion imaging was not associated with lower perceived PF. GHP. The noninvasive cardiovascular testing parameters together (model 1) significantly predicted GHP scores (adjusted R2 =. 12, P = .0001); however, the measures explained only 12% of the variance in GHP. In model 2, after addition of patient demographic and clinical variables, the noninvasive measures together explained only 10% of the variance in GHP. METS and chest pain during exercise were also independent predictors of GHP. Ischemia by ECG treadmill test was an independent predictor of higher GHP.

Discussion The results of this study have shown that duration of exercise, METS, and the presence of chest pain during exercise ECG test are significantly associated with patients’ perception of their general health status and their PF ability in daily life. Although exercise minutes and METS achieved showed a highly statistically significant relation to PF and GHP (P < .0001), there was wide

American Heart Journal September 2000

416 Mattera et al

Table V. Multivariate analysis of GHP Model 2†

Model 1* Coefficient Exercise ECG treadmill measures METS Chest pain during exercise Exercise ECG result Normal Nondiagnostic Ischemia Myocardial perfusion imaging Normal Scar Ischemia Demographic variables Age Sex (female) Clinical characteristics Smoking Prior revascularization Chest pain type Asymptomatic Nonanginal chest pain Atypical angina Typical angina

SE

P value

Coefficient

SE

P value

1.3411 –10.5070

0.5599 3.7000

.02 .005

2.0700 –5.5821

0.5936 3.5741

.0006 .1

— –0.5764 9.5128

— 3.37511 3.8598

— .9 .01

— –0.5002 7.402

— 3.1101 3.5731

— .9 .04

— –6.1556 –2.3953

— 4.5020 3.5275

— .2 .5

— –3.3983 –3.1807

— 4.3916 3.6621

— .4 .4

0.3113 2.8598

0.1278 3.4208

.02 .4

–5.8783 –6.6869

3.2801 3.1954

.07 .04

— –7.4587 –18.1937 –11.4599

— 4.7066 3.4359 4.2628

— .1 .0001 .008

*Model 1 includes noninvasive test variables only (n = 195): degrees of freedom = 6, adjusted R2 = .1186, F ratio = 5.351, P = .0001. †Model 2 includes noninvasive test variables, demographic and clinical variables (n = 195): degrees of freedom = 13, adjusted R2 = .2867, F ratio = 6.997, P = .0001.

variability on an individual patient level. Interestingly, exercise-induced ischemia on the ECG or on the myocardial perfusion image was not associated with lower perceived PF and GHP. All the noninvasive measures together explained only 14% or less of the variation in PF and GHP after adjusting for patient demographic and clinical variables. This finding raises concern about the use of noninvasive exercise measures as substitutes for measures of HRQOL.

Comparison with other studies In our study, the correlation between exercise minutes and perceived PF ability was similar or higher than that reported by others using physical subscales of other quality of life instruments (r = 0.53 compared with 0.40,16 0.25,17 and 0.26 to 0.4126). The higher correlation in our study may be attributable to differences in the characteristics of the patient populations. Prior studies have included mostly men and selected cardiac populations (93 male patients with coronary artery disease and coronary artery bypass graft surgery,17 106 men and 12 women aged <55 years with prior myocardial infarction,18 182 Veterans Administration patients before and after percutaneous transluminal coronary angioplasty26). Our study included a more heterogeneous population with more women and patients with and without proven coronary artery disease. When we

analyzed a subset of our patients with known coronary artery disease (prior myocardial infarction and revascularization), we found a similar correlation between exercise minutes and PF and GHP.

Measures of ischemia and patient-reported health status Patients with ischemic test results by exercise ECG testing and myocardial perfusion imaging had the same or better perceived PF and GHP than patients with normal test results. This finding is somewhat surprising because presumably ischemia results in angina and would limit a patient’s functioning in daily living and negatively affect perceived health status. In this study, the presence of chest pain during exercise was associated with lower PF and GHP; however, only one third of patients with ischemia by ECG or myocardial perfusion imaging testing had chest pain during exercise. Additionally, patients with ischemia by ECG testing achieved the highest METS compared with patients with normal or nondiagnostic test results. In general, patients must exercise to a sufficient workload to elicit an ischemic response. Thus, in this patient population, myocardial ischemia may not have had much impact on the patient’s daily functioning and perception of health because patients generally had good exercise tolerance and the majority did not have symptoms of chest pain

American Heart Journal Volume 140, Number 3

during exercise. Furthermore, this study demonstrates that the presence of ischemia and chest pain on exertion is not always disabling from the patient’s viewpoint.

Left ventricular ejection fraction Resting left ventricular ejection fraction was not predictive of PF or GHP; however, only a small percentage of patients had an abnormal ejection fraction. The lack of association of left ventricular ejection fraction with patient-reported PF in a population with predominantly normal left ventricular ejection fractions was also found in the study by Permanyer-Miralda et al.17 The mean PF and GHP scores were lowest in the group of patients who had missing ejection fractions. These patients had either poor venous access in which to administer the radioactive tracer, or the study itself was not interpretable because of technical difficulties. Both of these factors may represent a less healthy group of patients and therefore negatively affect patients’ perception of their health.

Study limitations and strengths This study has several limitations. First, the results represent the patient population at one academic hospital laboratory and may not be representative of other patients and laboratories. Laboratories with different patient selection criteria may show different results. The patient population in this study was heterogeneous and overall was fairly healthy with very good exercise capacity; in addition, the majority had negative test results. This finding probably reflects a referral bias for patients undergoing exercise testing in this center. In this institution, patients who cannot exercise or are not likely to exercise to an adequate heart rate are generally referred for pharmacologic stress testing rather than exercise treadmill testing; therefore more physically functional patients are likely to undergo exercise treadmill testing. This study has several strengths. First, consecutive patients were included and are therefore representative of the patients that met inclusion criteria. Second, comprehensive noninvasive cardiovascular testing measures were obtained in each patient. We are not aware of another study that directly compared exercise ECG testing combined with myocardial perfusion imaging parameters and radionuclide ventriculography with patient-perceived functioning and health status. Lastly, we obtained the patient-reported health status surveys at the same time as testing occurred to improve comparability. Additionally, we obtained the surveys immediately before testing to eliminate any bias the test results may have had on the patients’ perception of their health.

Study implications The findings of this study demonstrate that when studying patient populations, traditional noninvasive

Mattera et al 417

measures of exercise performance, on average, are associated with patients’ perceptions of health and PF in daily life. However, on an individual level, patients’ perceptions of their health can be quite different than clinicians would predict based on test results. An exercise enthusiast may be more negatively affected by chest pain at a high exercise workload than a person who exercises poorly but is content leading a sedentary lifestyle. These findings suggest that the evaluation of the impact of disease and the success of medical treatments should be determined by traditional test results as well as the patients’ viewpoint of their health. The success of treatment in clinical practice as well as in clinical trials should include the patients’ perception of health status if we want to maximize outcomes from their perspective.

Conclusion In consecutive patients with known or suspected coronary artery disease referred for exercise ECG treadmill testing with myocardial perfusion imaging, METS achieved and chest pain during exercise were significantly associated with patient-reported PF and GHP. Conventional noninvasive testing measures explain only a small amount of the variation in patient-reported health status and therefore are of limited value as surrogates for HRQOL.

References 1. DeBusk RF. Specialized testing after recent acute myocardial infarction. Ann Intern Med 1989;110:470-81. 2. Guidelines for exercise testing. A report of the American College of Cardiology/American Heart Association Task Force on Assessment of Diagnostic and Therapeutic Cardiovascular Procedures (Subcommittee on Exercise Testing). J Am Coll Cardiol 1986;8:725-38. 3. Brown KA. Prognostic value of thallium-201 myocardial perfusion imaging: a diagnostic tool comes of age. Circulation 1991;83: 363-81. 4. Pryor D, Harrell F Jr, Lee K, et al. Prognostic indicators from radionuclide angiography in medically treated patients with coronary artery disease. Am J Cardiol 1984;53:18-22. 5. Greenberg H, McMaster P, Dwyer E. Left ventricular dysfunction after acute myocardial infarction: results of a prospective multicenter study. J Am Coll Cardiol 1984;4:867-74. 6. Peduzzi P, Hultgren H, Thomsen J, et al. Ten-year effect of medical and surgical therapy on quality of life: Veterans’ Administration cooperative study of coronary artery surgery. Am J Cardiol 1987;59:1017-23. 7. O’Brien BJ, Buxton MJ, Ferguson BA. Measuring the effectiveness of heart transplant programmes: quality of life data and their relationship to survival analysis [abstract]. J Chron Dis 1987;40:137S-53S. 8. Croog SH, Levine S, Testa MA, et al. The effects of antihypertensive therapy on the quality of life. N Engl J Med 1986;314:1657-64. 9. Stewart AL, Greenfield S, Hays RD, et al. Functional status and wellbeing of patients with chronic conditions: results from the Medical Outcomes Study. JAMA 1989;262:907-13. 10. Caine N, Harrison SCW, Sharples LD, et al. Prospective study of quality of life before and after coronary artery bypass grafting. BMJ 1991;302:511-6.

American Heart Journal September 2000

418 Mattera et al

11. Strauss W, Fortin T, Hartigan P, et al. Benefits in quality of life scores in patients treated with angioplasty vs. medical therapy [abstract]. J Am Coll Cardiol 1992;19:210A. 12. Krumholz HM, Cohen DJ, Williams C, et al. Health after coronary stenting or balloon angioplasty: results from the Stent Restenosis Study. Am Heart J 1997;134:337-44. 13. Patrick DL, Bergner M. Measurement of health status in the 1990s. Ann Rev Pub Health 1990;11:165-83. 14. Guyatt GH, Feeny DH, Patrick DL. Measuring health-related quality of life. Ann Intern Med 1993;118:622-9. 15. Guyatt GH, Thompson PJ, Berman LB, et al. How should we measure function in patients with chronic heart and lung disease? J Chron Dis 1985;38:517-24. 16. Wiklund I, Comerford M, Dimenas E. The relationship between exercise tolerance and quality of life in angina pectoris. Clin Cardiol 1991;14:204-8. 17. Permanyer-Miralda G, Alonso J, Anto JM, et al. Comparison of perceived health status and conventional functional evaluation in stable patients with coronary artery disease. J Clin Epidemiol 1991;44: 779-86. 18. Skinner JS, Albers CJ, Hall RJ, et al. Comparison of Nottingham Health Profile (NHP) scores with exercise duration and measures of ischaemia during treadmill exercise testing in patients with coronary artery disease. Eur Heart J 1995;16:1561-5. 19. Sjoland H, Wiklund I, Caidahl K, et al. Relationship betwen quality

20.

21.

22.

23. 24.

25.

26.

of life and exercise test findings after coronary artery bypass surgery. Int J Cardiol 1995;51:221-32. Sjoland H, Wiklund I, Caidahl K, et al. Improvement in quality of life and exercise capacity after coronary bypass surgery. Arch Intern Med 1996;156:265-71. DiCarli MF, Asgarzadie F, Schelbert HR, et al. Quantitative relation between myocardial viability and improvement in heart failure symptoms after revascularization in patients with ischemic cardiomyopathy. Circulation 1995;92:3436-44. Ware JE, Sherbourne CD. The MOS 36-Item short-form health survey (SF-36): Conceptual framework and item selection. Med Care 1992;30:473-83. Ware JE. SF-36 Health Survey: manual and interpretation guide. Boston: Nimrod Press; 1993. Liu Y, Sinusas A, Deman P, et al. Quantification of SPECT myocardial perfusion images: methodology and validation of the Yale-CQ method. J Nucl Cardiol 1999;6:190-203. Diamond G, Staniloff H, Forrester J, et al. Computer-assisted diagnosis in the noninvasive evaluation of patients with suspected coronary artery disease. J Am Coll Cardiol 1983;12:444-55. Strauss WE, Fortin T, Hartigan P, et al, and the ACME Investigators. A comparison of quality of life scores in patients with angina pectoris after angioplasty compared with after medical therapy: outcomes of a randomized clinical trial. Circulation 1995;92: 1710-9.