To the Editor:
We read with great interest the paper by Sun et al.1 and support their aim to reduce the occurrence of cerebral injury after cardiac surgery, since this is a feared and devastating complication. Overt stroke rate has been reported to occur in 1 to 2% of cases after cardiac surgery, whereas the frequency of covert injury detected by diffusion weighted magnetic resonance imaging has been reported to be more than 50%.2 In agreement with previous observations, Sun et al. report age, type of surgical procedure, preoperative hypertension, time on cardiopulmonary bypass (CPB), emergent operation, and occurrence of atrial fibrillation postoperatively as risk factors for stroke.1 The main result from their study is the observation that hypotension during surgery was a significant risk factor of stroke, in this setting the only modifiable risk factor. However, in the multivariable analysis, the risk of a low mean arterial pressure (MAP) was only statistically significantly associated during CPB. This clearly emphasizes the importance of the intraoperative phase and suggests that a low blood pressure should be treated, although a potential benefit can only be assessed in interventional trials and not based on retrospective data. Regarding the choice of intervention, there are two principally different approaches: one approach is to increase MAP by using vasoconstrictors and thereby increase the organ perfusion pressure, and an alternative approach is to increase pump flow during CPB. To better understand the contribution from each of these approaches, the study lacks information on the actual pump flow delivered during CPB, which we believe is a major shortcoming. Can the authors provide data on average flow during CPB in patients with and without stroke? Are there any associations between duration of low flow and the occurrence of stroke?
Even though CPB has been around for more than 60 yr, there is still no consensus on limits for cerebral autoregulation during CPB. Hori et al. published a study in 2017 using a combination of integrated MAP and transcranial ultrasound demonstrating very variable limits for cerebral autoregulation between patients. In this respect, there was no safe lower MAP level, but the product of duration and magnitude of MAP below lower individual limits of cerebral autoregulation was associated with an increased risk of stroke.3 This technique is not yet available on a commercial basis. However, what is worth noticing is the fact that whenever a patient was below the lower limit of cerebral autoregulation, they increased MAP by increasing flow on CPB, making the interpretation of a “sufficiently high” MAP more complex.
Cerebral monitoring has gained widespread interest, and one widely used technique is near infrared spectroscopy to monitor cerebral tissue oxygenation as a surrogate for cerebral blood flow. In a randomized study, patients were allocated either to a higher MAP target (70 to 80 mmHg) or a low MAP target (40 to 50 mmHg) during CPB with a fixed pump-flow of 2.7 (SD 0.1) l per min/m2. The high target MAP was achieved with vasopressors, mainly norepinephrine infusion. The high-target group had significantly lower mean cerebral tissue oxygenation levels and a higher accumulated desaturation load less than 10% from baseline.4 These data support a previous proof-of-concept study demonstrating that cerebral tissue oxygenation does not improve by a vasoconstrictor-induced increase in MAP; instead, vasoconstrictors led to a cerebral tissue oxygenation decrease. Only by increasing flow on CPB by 0.5 l · min · m2 could cerebral tissue oxygenation be increased in parallel with an increase in MAP.5 In conclusion, focusing exclusively on MAP as a single parameter without considering the concomitant flow delivery will only tell us half of the story.
Research Support
The Perfusion Pressure Cerebral Infarct trials were funded by the Danish Heart Foundation (Copenhagen, Denmark) and by Rigshospitalets Research Foundation (Copenhagen, Denmark).
Competing Interests
The authors declare no competing interests.