Quality Assessment in Dobutamine Stress Echocardiography: What are the Clinical Predictors Associated With a Non-Diagnostic Test?

BACKGROUND Non-diagnostic dobutamine stress echocardiography (ndDSE, failure to achieve 85% of maximal predicted heart rate (HR) without evidence of inducible ischemia) is an important limitation affecting quality of DSE testing. The objectives of this study were to identify the clinical variables associated with a non-diagnostic Dobutamine Stress Echocardiogram (ndDSE) and further evaluate the patterns of subsequent testing for myocardial ischemia. METHODS Consecutive DSE's over a 17 month period (January 2008 to June 2009) were studied. Baseline demographics, medical history, and vital signs were collected. Subsequent testing was determined for up to 6 months after the initial DSE. Univariate and multivariate logistic regression analysis was performed to identify clinical factors associated with ndDSE. RESULTS Of 467 total DSE, 314 (67%) were negative for ischemia, 69 (15%) were positive, and 84 (18%) were ndDSE. Of those recommended for further nuclear MPI testing 12 (14%) had an ndDSE compared to 16 (4%) patients with a diagnostic DSE (P = 0.001). Fifty percent of the ndDSE nuclear MPI tests were positive for ischemia. In the univariate analysis, Diabetes Mellitus (DM; P = 0.003), calcium channel antagonist (CCA) use (P = 0.047), Hypertension (HTN; P = 0.06), low baseline HR (P < 0.001), and younger age group (P = 0.02) were predictive of ndDSE. Of these, all except CCA use remained independent predictors of ndDSE in multivariate analysis. A 4 variable model for predicting ndDSE was developed from the multivariate logistic regression displayed in Table 1 (age and baseline HR were categorized and scored 0-2; DM and HTN were scored as 0 (absent) or 1 (present)). Figure 2 demonstrates how risk of ndDSE correlated with a higher score, with each increment having an odds ratio of 2.1 (P < 0.001). CONCLUSIONS DM, HTN, younger age, and lower baseline HR affect the quality of DSE testing, resulting in non-diagnostic tests. A model combining these factors can identify patients most likely to have this outcome. Identification of this cohort may improve referral patterns and improve the quality of stress testing.