Endothelium‐dependent vascular dysfunction in septic patients

Nitric oxide is considered an important contributory factor in the pathogenesis of septic shock. The sepsis-associated elevated expression of inducible nitric oxide-synthase and subsequent increase in nitric oxide largely accounts for the down-regulation of endothelial nitric oxide-synthase, resulting in an impaired vasodilatory response to substances such as substance P, bradykinin and acetylcholine. Given the overwhelming evidence in favour of endothelial dysfunction during sepsis, the finding of Kienbaum et al. [1] that acetylcholine-induced vasodilation is preserved in septic shock patients is surprising. This paper represents the first investigation of vascular reactivity in septic patients, for which we congratulate the authors, though we feel it is important to make some comments. In a previous study using the perfused forearm, conducted prior to and during experimental endotoxaemia in humans it was found that an infusion of acetylcholine comparable to that used by Kienbaum resulted in a 40% reduction in the vasodilatory response during endotoxemia compared to baseline response [2, 3]. Endothelial dysfunction was found 4–6 h after lipopolysaccharide administration, at the time that circulating cytokines are at their highest. In Kienbaum’s study, it is not clear in what phase of sepsis the patients were and it appears plausible that during later phases of septic shock, endothelial function may recover. In addition, in septic shock patients, paired observations are not possible and with a small sample size inter-individual differences may become important. Acetylcholine is highly unstable in blood and vasodilator responses to acetylcholine are dependent on aceylcholinesterase activity, sex, basal blood flow and forearm length [4]. Furthermore, it appears that Kienbaum et al. measured forearm blood flow in the infused arm only. In our opinion, and in previous studies, measurement of forearm blood flow in both forearms would be a better method. Measurements of the flow in the non-infused arm can be used as a contemporaneous time-control for the infused arm. In this way systemic effects can be observed and vasoactive effects can be expressed as a quotient of the infused and the non-infused arm [2, 5]. Kienbaum et al. report that the acetylcholine-induced decrease in forearm vascular resistance (forearm blood flow/mean arterial pressure) was similar between septic patients and controls, however, since septic patients had lower vascular resistance initially, decrease in vascular resistance may be less. In addition, since it was only feasible to infuse two acetylcholine concentrations, it cannot be determined at which part of the acetylcholine dose-response curve the measurements were made. It is conceivable that with higher infusion rates of acetylcholine or other endothelium-dependent agonists, like substance P and bradykinin, it would be possible to demonstrate endothelial dysfunction during early sepsis. Finally, the lack of a vasopressor response after intra-arterial infusion of the nitric oxide-synthase inhibitor L-N monomethylarginine in the controls is probably caused by the much lower infusion rates used compared to previous studies [2, 5]. In conclusion, in contrast to prior experimental and clinical work, endothelial dysfunction during systemic inflammation was not confirmed by Kienbaum et al. Although the ‘perfused forearm model’ enables us to investigate direct vascular effects of various pharmacological agents, allowing us to determine direct vasoactivity and endothelial function in various groups of patients, caution should be taken in the interpretation of the results obtained. In view of the above issues, additional studies are necessary before we can conclude that endothelial dysfunction is not present in septic patients.