Haemodynamic and hyperaemic effects of adenosine in patients with atrial fibrillation undergoing quantitative myocardial perfusion cardiovascular magnetic resonance

Abstract

Aims: Patients with atrial fibrillation (AF) are thought to have an attenuated response to adenosine during vasodilator stress testing. We sought to investigate the haemodynamic and hyperaemic effects of adenosine in patients with AF undergoing adenosine-stress cardiovascular magnetic resonance (CMR) assessment.

Methods and results: We retrospectively examined 318 patients referred for clinical adenosine-stress CMR (AF n = 158, sinus rhythm [SR] n = 160). Baseline and peak heart rate (HR) and quantitative myocardial perfusion were compared between groups. At peak stress, the haemodynamic response was blunted in patients with AF (HR increase 7 ± 10bpm vs. 17 ± 11bpm in SR, P < 0.001). Fewer patients in AF met the threshold for a satisfactory HR response ≥10bpm (40% in AF vs. 76% in SR, P < 0.001). There were no intergroup differences in hyperaemic myocardial blood flow (1.52 ± 0.65 mL/min/g in AF vs. 1.55 ± 0.65 mL/min/g in SR, P = 0.670) or myocardial perfusion reserve (2.66 ± 1.11 in AF vs. 2.66 ± 1.08 in SR, P = 0.981). AF (odds ratio [OR], 0.29 [0.17-0.50], P < 0.001) and left ventricular ejection fraction (OR 1.03 [1.00-1.05], P = 0.023) were independently associated with achieving a satisfactory HR response on multivariable analysis, but only ejection fraction (OR 1.05 [1.02-1.09], P = 0.003) predicted a satisfactory hyperaemic response.

Conclusion: The heart rate response during adenosine-stress CMR is blunted in AF patients. Despite this, the majority of patients with AF generate a sufficient hyperaemic response with a standard adenosine-stress protocol. Further work is needed to determine the diagnostic accuracy of adenosine-stress CMR in patients with AF.

Keywords: adenosine; atrial fibrillation; cardiovascular magnetic resonance; hyperaemic myocardial blood flow; left ventricular systolic dysfunction; stress perfusion.

Graphical abstract

AF, atrial fibrillation; BPM, beats per minute; CAD, coronary artery disease; CI, confidence interval; CMR, cardiovascular magnetic resonance; LGE, late gadolinium enhancement; LVEDVi, indexed left ventricular end-diastolic volume; LVEF, left ventricular ejection fraction; MBF, myocardial blood flow. Case 1 A 77-year-old male with persistent atrial fibrillation and recurrent chest pain with absent haemodynamic response to 170µg/kg/min of adenosine (baseline; heart rate, 94bpm, blood pressure, 127/85mmHg; and peak stress, 91bpm, 123/80mmHg). No perfusion defects are appreciated on qualitative assessment of raw stress perfusion images (Panel B), but pixel-wise myocardial blood flow perfusion maps (Panel A) and polar plot (Panel C) confirm adequate hyperaemia at all three left ventricular levels. Case 2 A 67-year-old male with persistent atrial fibrillation and severe left ventricular systolic dysfunction referred for investigation of recurrent angina: absent haemodynamic response to high-dose (210µg/kg/min) adenosine (baseline; heart rate, 85bpm; blood pressure, 129/78mmHg; peak stress, 85bpm, 124/63mmHg). Pixel-wise hyperaemic maps (Panel A) and raw perfusion sequences (Panel C) show a perfusion defect in the right coronary artery territory with a corresponding regional reduction in hyperaemic myocardial blood flow (Panel B) with evidence of adequate hyperaemia in the unaffected myocardial segments. The perfusion defect significantly exceeded the areas of infarction detected on phase-sensitive inversion recovery late gadolinium enhancement images (Panel D). Invasive coronary angiography demonstrating an occluded mid-right coronary artery with collateralisation from the left coronary system (Panel E).

Figure 1

ROC analyses for ability of the heart rate response to predict a satisfactory hyperaemic response (stress myocardial blood flow >1.43 mL/min/g in ≥1 myocardial segment as the reference standard).