Ischaemic stroke is a major complication of cardiac surgery. The optimal strategies for operating on patients with obstructive carotid and/or intracranial artery disease (CIAD) are controversial. We aimed to clarify whether single-photon emission computed tomography (SPECT) with acetazolamide, to quantify the cerebral perfusion reserve, could predict the risk of haemodynamic ischaemic stroke during cardiac surgery. The incidence of stroke related to obstructive CIAD and the corresponding autoregulatory reserve were prospectively assessed in 514 consecutive patients who underwent elective cardiac surgery with cardiopulmonary bypass (n = 484) and off-pump coronary artery bypass grafting (n = 30) between 2009 and 2013. Preoperative cerebral blood flow and its reactivity to acetazolamide were quantitatively determined in patients (n = 88) with obstructive CIAD, diagnosed by carotid ultrasonography and/or magnetic resonance angiography. An impaired cerebral perfusion reserve was identified in 1 (1.1%) of the 88 patients. This patient underwent prophylactic superficial temporal artery to middle cerebral artery anastomosis 1 month before coronary artery bypass surgery. Subsequently, the patient underwent conventional coronary artery bypass surgery, without experiencing perioperative stroke. Seven (1.4%) patients died in-hospital mortality and 5 (1.0%) experienced perioperative stroke. However, no patients experienced perioperative haemodynamic ischaemic stroke. It is unusual for CIAD to affect the incidence of haemodynamic ischaemic stroke during cardiac surgery. Brain perfusion SPECT with acetazolamide is effective for narrowing down patients at high risk of ischaemic stroke during cardiac surgery. Meanwhile, the application of brain perfusion single-photon emission tomography should be confined only to patients with obstructive CIAD because it is an expensive examination tool.