Abstract
Background and Aims
Acute kidney injury (AKI) is one of the most serious complications of patients admitted to intensive care units (ICUs). It is associated with high short- and long-term mortality and resource utilization. The definition of AKI has been established by the KDIGO guidelines based on changes in serum creatinine, urine output or both. However, in clinical practice physicians may ignore the standard criteria and rely on clinical judgement. We therefore aimed to assess the degree of physicians’ compliance with the KDIGO guidelines in diagnosis of AKI.
Method
We collected data (demographic, clinical, and biochemical) in a multicenter prospective cohort study from all adults admitted to ICUs (10 surgical and 8 medical) units at Alexandria University Teaching Hospitals from February 1st, 2016 till August 1st, 2016. Alexandria Teaching Hospitals cover four governorates of Northern Egypt and serve approximately 14 million people. Doctors were preliminarily instructed to apply KDIGO criteria for the diagnosis of AKI. Personal and clinical experience data were collected from the treating physicians. We followed patients for thirty days from study entry until discharge, death or study end. Written informed consent was obtained from all participants. AKI was defined and classified based on KDIGO 2012 criteria. In parallel, we registered the actual clinical diagnosis made by the treating physicians. We used frequencies and means for qualitative and quantitative variables as appropriate.
Results
The study included 532 patients who were on average 46 year old (±18), 41.7% were males, 23.5% with smoking, 23.1% had diabetes, 34.8%, were hypertensive, 11.3 % with pre-existing chronic kidney disease, and 30.1% had cardiovascular diseases. There were 140 physicians responsible for treating the enrolled subjects, with mean age 30 ±3 years, 57% were males, 20% were nephrologists, and the median years of experience was 3 years (inter-quartile range: 2-4years). The AKI incidence was 62.2% according to KDIGO criteria versus 49.9% based on the clinical diagnosis of treating physicians. Among those not reported to have AKI by the treating physicians; 19.1% were in stage 1, 26.4% in stage 2, and 12.9% in stage 3 AKI based on KDIGO. About 24% of patients who had AKI at ICU admission and 15% of those who developed AKI after ICU admission were not appropriately identified as AKI patients according to the physicians. There was a significant association between the physician speciality (nephrology vs other specialties) and the correct AKI diagnosis based on KDIGO criteria (X2=47.06, p<0.001).
Conclusion
To streamline a correct and timely identification of AKI, treating physicians in ICUs at a large hospital in North Africa, like the Alexandria University Teaching Hospitals in Egypt, need well focused training and knowledge verification post training on KDIGO guidelines for identifying AKI patients. Implementation of electronic alerts could help in proper diagnosis and management.
Predictive capacity of prognostic scores for kidney injury, dialysis, and death in intensive care units