Data from consecutive patients admitted to the cardiac intensive care unit (CICU) following cardiac surgery at Wythenshawe Hospital (part of Manchester University NHS Foundation trust) were collected prospectively between January 2013 and May 2015. Wythenshawe Hospital is a tertiary centre for adult cardiac surgery, cardiothoracic transplantation and mechanical circulatory support as a bridge to cardiac transplantation or recovery. Patients requiring RRT preoperatively and those with no preoperative creatinine values were excluded as shown in Fig. 1. Patients who received mechanical circulatory support were excluded from length of stay (LOS) analyses as their CICU stay was prolonged while awaiting definitive treatment. All data were collected as part of the Vascular Governance North West (VGNW) database and processed according this project’s protocols and ethical approvals.

Flow chart for inclusion of patients in analyses. RRT = renal replacement therapy, sCR = serum creatinine result, MCS = mechanical circulatory support, PLOS = prolonged length of stay

Serum creatinine concentration was usually measured daily and all available results were extracted from the hospital’s pathology laboratory database. Our institution’s laboratory measures creatinine using techniques based on Jaffe chemistry with a total imprecision of < 6%. Every creatinine value for each patient was analysed and both relative and absolute increases in creatinine were used to classify AKI stages according to the KDIGO criteria (Table 1). The relative increases were calculated using the most recently recorded preoperative level as the baseline value. Urine output was recorded hourly on the CICU electronic patient record. Where the hourly value was recorded as none or zero, this value was accepted whereas when no value was entered for a given hour the next volume of urine recorded was divided equally by the number of blank hours prior to this recording. Whenever urine output fell below the thresholds in the KDIGO criteria, the time and appropriate stage of AKI was recorded. The need for RRT and postoperative LOS on CICU were identified from the electronic patient record. Serum creatinine concentration and urine output measurements recorded after initiation of RRT were not included in analyses as both are influenced heavily by RRT itself.
The hospital clinical governance database recorded 2-year all-cause mortality and the preoperative comorbidity, urgency and complexity of surgery as measured by the logistic EuroSCORE [16 (link)]. Prolonged LOS was defined as a CICU stay longer than 120 h for cardiac transplant patients or > 72 h for all other patients.
Free full text: Click here