RESUMO
AIMS: To evaluate effects of caffeine on quantitative myocardial perfusion by positron emission tomography (PET) and associated major adverse cardiovascular events (MACE). METHODS AND RESULTS: Serum caffeine was measured for all 6087 PETs with 328 positive results (5.4%). Paired caffeine positive/negative PETs (84 patients for dipyridamole with median caffeine 1.6 mg/L, and additional 25 volunteers for regadenoson with median caffeine 7.4 mg/L) were compared for quantitative perfusion. Multivariate regression analysis for associations among caffeine, clinical/imaging variables, predicted caffeine probability was performed. MACEs were followed up to 9 years after PETs. For caffeine vs. no caffeine, respectively, stress flow was 1.74 ± 0.55 vs. 2.14 ± 0.53 for dipyridamole and 1.82 ± 0.61 vs. 2.33 ± 0.49 mL/min/g for regadenoson, and coronary flow reserve (CFR) was 2.26 ± 0.67 vs. 2.67 ± 0.72 for dipyridamole and 1.84 ± 0.33 vs. 2.31 ± 0.41 for regadenoson (all P < 0.001). Subjects were reclassified from high-risk CFR ≤2.0 with caffeine to low-risk CFR >2.0 without caffeine in 66.7% and 80% of dipyridamole and regadenoson caffeine-no-caffeine pairs, respectively. While relative images showed no differences, caffeine significantly altered coronary flow capacity (CFC) to false negative and false positive severity in 2.1% and 5.5% of the 328 caffeine positives, respectively (0.1% and 0.3% of 6087 PETs) but without change in severity guided management in most patients (92.4% of 328 caffeine or 99.6% of total 6087 PETs). CONCLUSION: Even low serum caffeine levels reduce quantitative perfusion during vasodilatory stress with false positive or false negative results minimized by empathic instruction, CFC analysis or repeat PET after strict caffeine abstention for definitive individualized risk stratification and management.