Abstract 3107: Arm Exercise Stress Testing Variables Predict Survival and Adverse Clinical Outcomes

Valuable prognostic and clinical information from treadmill exercise testing (GXT) includes exercise capacity (METs), heart rate, and electrocardiographic (ECG) responses. However, little or no prognostic data are available for arm ergometer stress testing (AXT). To determine whether AXT variables predict survival, myocardial infarction (MI), or coronary revascularization (CVASC), we performed AXT from 1997 to 2002 in 359 patients, mean age 63 +/− 11 (SD) years, referred for clinical reasons but unable to perform GXT, and followed for 63 +/− 24 months, during which 98 deaths occurred (27%). Average annual mortality, MI, CVASC, and combined adverse event rates were 5.2%, 1.7%, 2.2%, and 7.1%, respectively. Student’s t-tests were used to assess differences between outcome groups. Cox regression models were employed to determine hazard ratios (HR) and 95% confidence intervals (CI). Kaplan-Meier survival models were used to compare survival curves among AXT groups. AXT METs was highly predictive of survival after adjustment for age and beta blocker treatment (p < 0.001; when stratified by tertiles; death HR 0.47, CI 0.22– 0.71 middle vs. lowest; HR 0.61, CI 0.28 – 0.94 highest vs. middle). A greater delta (peak-rest) heart rate was associated with survival (p = 0.0003) and/or event-free outcome as were faster % age-predicted peak heart rate (death HR 0.58, CI 0.36 – 0.80 for >70% vs.> 70%), higher exercise systolic blood pressure (SBP) (p = 0.002) and peak heart rate x SBP product (PRPP) (p = 0.0006). A positive (+) AXT ECG was observed in 22% of deaths and 10% of survivors, 27% of MI and 12% with no MI, and 32% of CVASC versus 11% with no CVASC. A+AXT ECG was a powerful predictor of adverse outcome, even after accounting for peak heart rate, peak SBP and PRPP (death HR 2.2, CI 1.94 –2.43; MI HR 2.9, CI 2.48 –3.30; CVASC HR 4.1, CI 3.73– 4.43; combined events HR 2.8, CI 2.55–2.98). Sensitivity, specificity, positive and negative predictive values of a +AXT ECG in prognosticating adverse outcomes ranged from 22–31%, 88 –92%, 18 – 61%, and 62–92%, respectively. Thus, in veterans who are older and have more comorbidities than most other study populations based on adverse event rates, AXT is an alternative to GXT for predicting clinical outcome in patients with lower extremity disabilities.