STRENGTH OF EVIDENCE
The outcome of our quality assessment is shown in Table 3 . In general, we determined that the strength of evidence was overall of low quality. The level of information provided about CRRT techniques was broadly inconsistent with few studies providing the level of information required to enable comprehensive interrogation or meta-analysis of the data. In one study [38] , demographics were only provided at aggregate level which prevented individualised pharmacokinetic analysis. Time spent on filter was only mentioned in one study [34]and while it could be assumed that haemofiltration was continuous, it was clear that at least some patients experienced filter clotting and pauses to treatment. The two prospective studies in our data set[37, 38] were judged to be of higher quality predominantly due to the more robust and comprehensive nature of their pharmacokinetic data.