Short summary

Acute kidney injury is serious and common complication (~30%) following cardiac surgery, and even minor episodes are associated with progression to chronic kidney disease as well as increased mortality. We will develop a machine learning algorithm based on biochemical data and clinical data from the Western Danish Heart Registry, which can identify patients at risk of developing chronic kidney disease and enable a personalized medicine approach with early initiation of renal protective medicine.

Rationale

Acute kidney injury was previously considered a self-limited and reversible condition, but even minor episodes of kidney injury are associated with progression to chronic kidney disease, potentially end-stage kidney disease, as well as increased mortality. There is currently no preventive medication in order to avoid kidney injury. However, new medicine (e.g. SGLT2-inhibitors) have shown to be effective in slowing long-term decline in kidney function. Thus, early identification of patients developing declining kidney function is likely to benefit from treatment before incident chronic kidney disease is established.

We will develop a machine learning algorithm (QLattice) provided in collaboration with the AI company Abzu based on biochemical data and clinical data from the Western Danish Heart Registry - a seminational, multicentre-based registry with longitudinal registration of detailed patient and procedure data since 1999. The algorithm can be used as a clinical decision support system to identify patients at risk of developing chronic kidney disease following cardiac surgery. Such an identification, instead of the current use of renal markers, will enable a personalized medicine approach with initiation of renal protective medicine in future patients, thereby preserving the kidney function longer and avoid end-stage renal disease and dialysis or transplantation ultimately.

Description of the cohort

Patients (≥18 years) undergoing cardiac surgery at the department of Cardiac Surgery (isolated coronary artery bypass graft (CABG), single and multiple valvular procedures, combined CABG and valvular surgery, and others), at Odense University Hospital in the period 2000-2022 is included.

Exclusion criteria is; pre- or postoperative serum creatinine not available or preoperative end-stage kidney disease (defined as receipt of dialysis).

Data and biological material

Information on age, sex, BMI, comorbidity, date of surgery, procedure, peri-operative interventions, length of stay, complications and date of death are reported from Western Danish Heart Registry. Routine blood samples (e.g. creatinine, hemoglobin, ALAT etc.) are obtained from the regional laboratory databases.

Collaborating researchers and departments

Department of Anaesthesiology and Intensive Care, Odense University hospital

Department of Thoracic surgery, Odense University Hospital

Department of Cardiology, Odense University Hospital

MeCiSu - frontline centre in Mechanical Circulatory Support, Odense University Hospital

Abzu Copenhagen, Denmark