JACC March 19,2003 ABSTRACTS - Cardiac Function and Heart Failure 19 1A CD (3.3%, p< 0.001). Conclusion: Cl and HRR are independent and incremental predictors of cardiac death. Score incorporated age, symptoms, estrogen status, and 6 other coronary risk factors. Using previously defined thresholds, women were placed into low, intermediate, and high probability groups. Women were followed for 2.6 + 1.5 years for determination of prog- nostic outcomes. Outcomes considered were death (D), infarction (Ml), stroke (SK), other vascular (OV) and revascularization (REV). Results: Composite outcome results are shown in table below. Numbers in () are percentagesConclusion: In this cohort of women undergoing coronary angiography with a low prevalence of coronary disease, the Pretest Score stratified women significantly concerning both hard and soft prognostic outcomes. 9:oo a.m. 838-3 Fitness Versus Activity for Predicting Mortality in Men Amir Kavkha, Jonathan N. Myers, Sheela George, Joshua Abella, Takuya Yamazaki, Victor F. Froelicher, VA Palo Alto Health Care System, Palo Alto, CA Background: Both physical fitness and daily physical activity panerns are inversely associated with mortality, but a comparison between the two has not been performed in the same population. Methods: Physical fitness was determined as METS calculated from speed and grade for 642 males (age 59ul2) referred for treadmill testing for clinical reasons. Adulthood recreational activity pattern, expressed in kcallweek, was quantified using a modified Harvard Alumni Questionnaire at the time of exercise testing. Subjects were followed for a mean of 5.s2.0 Years with all-cause mortalitv as the endooint. Results: Recreational energy expenditure showed a graded pattern, with the more active demonstratina a lower morialitv ihazard ratio for 9000 kcal/week = 0.53. o=O.O3). I However, adjusting for age, a Cox proportional hazards model showed that peak METS achieved was a stronger predictor of mortality than physical activity panern (hazard ratio for >5 METs=0.30, pcO.001). The Figure shows the reduction in risk for quartiles of fit- ness and activity we observed relative to the recent meta-analysis of Williams (Med Sci Sports Exert 2001; 33;754). Both fitness and activity levels in our subjects were similar to the meta-analysis in reducing risk. Conclusion: Both physical fitness and physical activity patterns were associated with survival, but fitness as estimated in METS from treadmill testing more powerfully predicts survival than activity from a questionnaire. 9:15 a.m. 838-4 Can a Pre-Exercise Test Score Predict Prognosis in Women With a Low Prevalence of Coronary Disease? The National Heart, Lung, and Blood institute- Sponsored Women’s ischemia Syndrome Evaluation (WISE) Study Anthonv Morise, Marian 8. Olson, C. Noel Bairey Merz, Sunil Mankad, William J. Rogers, Carl J. Pepine. Steven E. Reis, Barry L. Sharaf, George Sopko. Gerald M. Pohost. LeSlea J. Shaw, West Virginia University School of Medicine, Morgantown, WV Background: Recent guidelines for exercise testing suggest that a pretest score be used to stratify patients before exercise testing. A Pretest Score derived for use in women without known coronary disease has previously been shown to stratify women according to the prevalence of angiographic coronary disease. However, it has not been tested in a population with a low prevalence of coronary disease. Methods: To deter- mine whether this Pretest Score will stratify a separate cohon of women according to prognostic outcomes as well, we evaluated 563 women who underwent coronary angiog- raphy for suspected myocardial ischemia with an overall low (26%) prevalence of angio- graphic coronary disease, defined as xiO% stenosis in >l eoicardial vessel. The Pretest Prognostic Outcomes by Pretest Group Outcomes LOW Intermediate High p Value D, Ml 41164 (2.4) lo/245 (4.1) 121154 (7.6) 0.024 D. MI, SK, OV 61164 (4.9) 21/245 (6.6) 22/154(14.3) 0.004 D. MI. SK, OV, REV 15/164 (9.2) 40/245 (16.3) 46/154 (30) 0.001 9:30 a.m. 038-5 Should Age Be Part of the Duke Treadmill Score? Takuva Yamazaki Amir Kaykha. Jonathan Myers, Victor Froelicher, VA Palo Alto Health Care System, Palo Alto, CA Background: The Duke Treadmill Score is a validated means of estimating cardiovascu- lar (CV) mortality that has been recommended as part of the routine interpretation of treadmill tests. Age is a predictor of death but not included in the DTS. Methods: Analyses were performed on the first treadmill test perfoned on 6,352 con- secutive male veterans at the Palo Alto and Long Beach Veterans Affairs Medical Cen- ters since 1967. After removal of all CHF patients, 5,629 patients remained with a mean age of 59+11 years. The main outcome measure was CV mortality; during a mean follow- up of 6.6+4 years, there were 516 CV deaths. The DTS was calculated as METS - (4’[1 =angina occurred, P=reason for stopping])-(5’ amount abnormal ST depression). Results: Using Cox Hazard analysis, the DTS and age were found to have similar coeffi- cients but opposite sign, so a score was made as DTS-age. Using ROC analysis with CV mortality as the outcome, AUCs were calculated for the scores. DTS-age gave a signifi- cantly better discrimination than the DTS alone. Conclusion: The DTS minus age outperformed the DTS for predicting cardiovascular mortality. Afler validation in other populations, age should be subtracted from the DTS calculation as part of treadmill test interpretation. 9:45 a.m. 838-6 Association of Abnormal Heart Rate Recovery and Chronotroplc incompetence With Obesity in a Healthy Cohort Michael S. Lauer, David Yu. Claire E. Pothier, Eugene H. Blackstone, Cleveland Clinic Foundation, Cleveland, OH Background: Obesity is known to be associated with abnormalities of autonomic ner- vous system balance. We hypothesized that obesity is also associated with exercise hearl rate abnormalities, which are reflective of autonomic tone. Methods: We studied 3,071 adults (mean age 49; 65% men) without a history of cardiac disease or use of cardiovascular medications who referred for symptom-limited exercise stress testing. Exact height and weight were directly measured per routine protocol. Sub- jects were divided into sex-specific quintiles of body mass index (BMI) (quintile 5 > 30 kg/ n? in men, > 26 kg/m’ in women). An abnormal heart rate recovery (HRR) was defined as ( 12 beats per mswte during the first minute after exercise. Chronotropic incompe- tence (CRI) was defined as failure to use 60% of heart rate resewe. Results:There were 333 subjects (11%) who had an abnormal HRR, 222 (7%) who had CRI, and 169 (6%) who had impaired functional capacity for age and gender. Abnormali- ties of all 3 variables were particularly marked in the highest quintile of BMt (Figure).