AKI is a rapid and usually reversible impairment of kidney function that is life-threatening in the short term well described by the "Kidney Disease: Improving Global Outcomes - KDIGO" classification of 2012. Whatever etiology of acute renal failure, drug iatrogeny still has its place. Hospital data from the information systems medicalization program (PMSI) can be used for epidemiological research. No study has yet been performed on these data to assess drug-related AKI. However, it should be remembered that these databases were not originally designed for research purposes but for reimbursement of care. Therefore, before conducting a large-scale study, it remains important to determine the validity and representativeness of the codes used for coding the studied events. The objective of this project is therefore to validate the use of hospital coding to identify AKI.

Acute kidney injury (AKI) is a rapid and usually reversible impairment of kidney function that is life-threatening in the short term; the grades of which are defined by the KDIGO classification of 2012. Whatever etiology of acute renal failure, drug iatrogeny still has its place. Data literature indicate that AKI affects more than one in five adults worldwide while hospitalized, with an associated mortality rate above 20%. It constitutes a non-negligible part of hospitalizations when it is acquired in the community, and also causes prolongations of hospitalizations, and therefore an increase in the cost for the hospital. Hospital data from the information systems medicalization program (PMSI) is a device that is part of the reform of the French health system with the aim of reducing inequalities in resources between health establishments. Since 2007, it has been possible to link all discharge summaries for one patient. The diagnoses identified during hospitalization are coded according to the 10th edition of the International Classification of Diseases (ICD-10). The data from PMSI coding can be used for epidemiological research. The majority of studies were related to patients with chronic kidney disease, especially dialysis patients. No study has yet been performed on these data to assess drug-related AKI. However, it should be remembered that these databases were not originally designed for research purposes but for reimbursement of care. Therefore, before conducting a large-scale study, it remains important to determine the validity and representativeness of the codes used for coding the events to be studied.

Inclusion Criteria:

• Patients over > = 18 years old
• Patients hospitalized at the CHU Amiens Picardie during 6 months
• patients for which at least two plasma creatinine measurements were taken during the hospitalization

Exclusion Criteria:

• Patients under <18 years old
• Patients on dialysis for chronic renal failure
• Hospitalization for a kidney transplant
• Postpartum AKI

Patients hospitalized at the CHU Amiens Picardie during 6 months and for which at least two plasma creatinine measurements were taken during the hospitalization