Perfusion thresholds in human cerebral ischemia: historical perspective and therapeutic implications.

After middle cerebral artery occlusion (MCAO) in the laboratory animal, the ischemic penumbra has been documented as a severely hypoperfused, functionally impaired, but still viable cortex which can regain its function and escape infarction if it is reperfused before a certain time has elapsed. The penumbra surrounds the ischemic core of already irreversibly damaged tissue, and is progressively recruited into the core with increasing MCAO duration. In the animal, the threshold of cerebral blood flow (CBF) below which neuronal function is impaired and the tissue is at risk of infarction is around 22 ml/100 g/min (approximately 40% of normal) in the awake or lightly anesthetized monkey, and around 30--35 ml/100 g/min in the cat and the rat. The threshold of CBF below which the tissue becomes irreversibly damaged and will progress to infarction depends on the duration of ischemia, and is around 10 ml/100 g/min for 1--2 h (approximately 20% of normal) and around 18 ml/100 g/min for permanent ischemia in the monkey. Mildly reduced CBF down to the 40% threshold (termed 'oligemia') is normally well tolerated, and the affected tissue is not at risk of infarction under uncomplicated conditions (in the animal, however, selective neuronal death may occur even with only mildly reduced CBF values, but this sequela of stroke seems an exceptional encounter in man). Classic studies with carotid artery clamping in man have provided estimates for the penumbra threshold at around 20 ml/ 100 g/min for the whole brain, but only recently have imaging studies allowed to document the existence of the penumbra in acute stroke and given estimates of local CBF thresholds. With PET, the penumbra is characterized by a reduced CBF, an increased oxygen extraction fraction, and a relatively preserved oxygen consumption (CMRO(2)). In a series of PET studies performed 5--18 h after stroke onset, we have determined the threshold for penumbra to be around 20 ml/100 g/min, and documented that the extent of neurological recovery is proportional to the volume of penumbra that eventually escaped infarction. Within this time interval, the thresholds for irreversible damage were around 8 ml/ 100 g/min for CBF and around 0.9 ml/100 g/min for CMRO(2). Recent studies with diffusion-weighted and perfusion MR have reported similar relative thresholds for CBF of about 50 and 18% for penumbra and core, respectively. Although it is likely that the threshold for irreversibility will be lower with shorter duration since clinical onset, this has not been documented thus far. Because saving the penumbra will improve clinical outcome, it should constitute the main target of acute stroke therapy. We found evidence of penumbra in about one third of the cases studied between 5 and 18 h after onset, and as late as 16 h after symptom onset in occasional patients, suggesting the therapeutic window may be protracted in at least a fraction of the cases; similar experience has recently accrued from diffusion-weighted MR and perfusion MR. In the remaining patients, there was evidence of early extensive damage or early spontaneous reperfusion, which would make them inappropriate candidates for neuroprotective therapy. Recent evidence from PET studies of relative perfusion performed within 3 h of onset suggests that early thrombolysis indeed saves the tissue with CBF below a critical threshold of 12 ml/ 100 g/min, with a correlation between the volume of such tissue escaping infarction and subsequent neurological recovery. Thus, mapping the penumbra in the individual patient with physiologic imaging should allow to formulate a pathophysiological diagnosis, and in turn to design a rational management of the stroke patient and to increase the sensitivity of drug trials by appropriate patient selection.