Can Modification of Renal Replacement Therapy Improve the Outcome of Patients with Systemic Inflammatory Response Syndrome?

fi ltration within the fi rst few hours of ICU admission. Looking at the time course of the infl ammatory response then, there is a sequence of events leading to the activation of innate immunity through the Toll-like receptors and the production of pro-infl ammatory cytokines and mediators, followed later by a corresponding increase in anti-infl ammatory cytokines and mediators. Thus those studies, which used the very early initiation of high-volume haemofi ltration, might have impacted on patients in the early pro-infl ammatory phase of their illness, and so potentially altered their clinical course. Unfortunately, by the time most patients arrive in the ICU, or those who develop ARF in the ICU setting, the initial pro-infl ammatory phase has passed, and patients have activation of both proand anti-infl ammatory systems. In this phase of the illness, removal of cytokines and other mediators will be non-selective and may explain why the later start of high-volume haemofi ltration has not been shown to improve patient outcomes [5] . High-volume haemofi ltration, using standard commercially available sterile replacement fl uids, is a complex treatment and is usually only performed for limited periods due to problems with circuit clotting, electrolyte, acid-base and thermal balance, and micronutrient losses. To try and achieve the same goals, other centres have developed high-volume exchanges based on haemodiafi ltration using on-line ultrapure dialysis fl uids, and this has been combined with daily plasma exchange [6] . This techThe mortality of patients, both admitted to the intensive care unit (ICU) with acute renal failure (ARF), and those that develop ARF within the ICU, remains high. In a recent multinational survey of clinical practice in ICUs, the highest survival for ICU ARF was reported from Australia, which predominantly uses continuous haemofi ltration [1] . Haemofi ltration, a predominantly convective technique, removes more larger hydrophilic molecular weight compounds than diffusion-based haemodialysis. In the ICU setting, haemofi ltration leads to greater cytokine removal, by a combination of membrane adsorption and convection [2] , but does not remove bacterial endoand exotoxins. If haemofi ltration techniques can remove some infl ammatory mediators, then the question arises as to whether cytokine removal could infl uence patient outcome. In everyday clinical practice the amount of cytokines and other infl ammatory mediators removed by standard haemofi ltration systems is limited, dependent not only upon molecular size but also protein binding. To try and effect cytokine and other infl ammatory mediator clearance, two small clinical studies used very high-volume haemofi ltration exchanges, and reported a marked improvement in patient survival [3, 4] . A further trial by the Amsterdam group did not show any effect of the dose of haemofi ltration delivered, or the timing of treatment initiation on patient survival [5] . Re-analysis of Honore’s original study [4] showed that survival was greatest in those patients who were started on high-volume haemoPublished online: March 3, 2006