During the study period, about 15,200 patients were admitted to participating ICUs. Among them, 757 patients were treated with RRT sometime during the ICU stay (Figure (Figure1).1). Of the 757 enrolled patients, thenthereby we excluded patients (n = 19) who received concomitant extracorporeal treatments (e.g. endotoxin adsorption) other than those specified in the methods of this report, and those with incomplete data (n = 117). The majority of incomplete data was due to one missing datapoint needed to calculate RRT dose on a specific day, such as percentage pre-dilution or actual start or stop time. Complete data on pre-specified outcomes were not available for 11 patients (1.4%).Among the remaining 553 AKI patients, 419 received CRRT only, 88 received IRRT only and 46 were treated with mixed RRT schedules (IRRT, CRRT, coupled plasmafiltration adsorption).
As patients in this last group crossed over from one RRT modality to another, delivered dose could not be calculated due to lack of clinically validated models, and they were excluded from the analysis. Among patients treated on only one RRT schedule (either continuous only or intermittent only), 82% received CRRT. This proportion represents current practice in Europe as previously reported. Out of 419 CRRT patients, 81 patients had at least one interruption of 18 hours or more, and then resumed CRRT [14]. The median interruption time was 49 hours (IQR = 29 to 113), predominantly due to filter clotting, disconnection for procedures and change in patient clinical status (e.g. CRRT not required in a window period).
As daily CRRT dose would appear artificially low in this situation, such patients were not included in the analysis. Eighty eight patients (18%) were treated exclusively with IRRT. One patient had only three IRRT sessions over a span of 146 days, and was excluded from analysis (Figure (Figure11).Characteristics of the study population are described in Table Table1.1. CRRT patients were younger, more likely to have sepsis, more likely to have been admitted directly into the ICU from the emergency room and less likely to be surgical patients. The mean serum creatinine at RRT initiation was 265 ��mol/L. Nearly 60% of all patients were in RIFLE class Failure at RRT initiation. A small minority of patients did not meet criteria even for Risk, and were labelled as a non-renal indication (e.g.
volume overload). Among the reasons cited to start RRT, azotaemia was significantly more common in the IRRT group, and oliguria in the CRRT group. Crude ICU mortality was 54% in the CRRT group, 22% in the IRRT group and 45% in the mixed group.Table 1Clinical characteristics of ICU patients receiving RRTPatient characteristics by RRT doseCRRTIn the CRRT group, the median delivered RRT dose was Brefeldin_A 27.1 ml/kg/hour (IQR = 22.1 to 33.9). Only 75 patients (22%) received more-intensive dose (�� 35 ml/kg/hour), while 262 (78%) received less-intensive CRRT.