BACKGROUND: Obesity is a risk factor for coronary artery disease (CAD), but adiposity may mimic symptoms of CAD and reduce the accuracy of diagnostic testing. METHODS: Patients from the PROMISE trial (n = 8889) were classified according to body mass index (BMI). We assessed relationships between BMI, physician's preference of functional test, test positivity, and results of invasive coronary angiography (Cath) using logistic regression models. RESULTS: Nearly half (48%) of the patients had BMI ≥ 30 kg/m2, and 20% had BMI ≥ 35. Providers were more likely to prefer nuclear myocardial perfusion imaging (MPI) over other functional tests as BMI increased. The rate of test positivity with coronary computed tomographic angiography (CTA) was not different (10% vs. 12%) in patients with BMI ≥35 vs. < 35. The same was true for stress echocardiogram and stress electrocardiogram (positivity 8-13%, P > 0.8 for both). In contrast, MPI was significantly more likely to be positive in those with BMI ≥35 vs. <35 (18% vs. 13%; P = 0.001). The likelihood of obstructive CAD at Cath did not differ with BMI ≥35 vs. <35 in patients having CTA (52% vs. 59%, P = 0.22), but among MPI patients, Cath positivity was only 29% with BMI ≥35 vs. 48% with BMI <35 (P = 0.005). Radiation exposure increased with higher BMI in both MPI and CTA groups. CONCLUSIONS: Increasing levels of obesity adversely affect the diagnostic yield of MPI more than CTA. The degree of obesity should be considered when choosing evaluation strategies for patients with chest pain. CLINICAL TRIAL REGISTRATION: PROMISE ClinicalTrials.gov number, NCT01174550.