#6104 CLINICAL IMPACT OF UNDIAGNOSED HOSPITAL ACUTE KIDNEY INJURY

Abstract

Abstract Background and Aims Hospital Acute kidney injury (H-AKI) is associated with an increased risk of complications and mortality. Despite attempts to uniform the diagnosis of AKI and point attention to this condition, in clinical practice, the recognition of AKI is often limited. In this study, we tried to evaluate the impact of underdiagnosed AKI in hospitalized patients. Method We collected data of patients hospitalized at IRCSS Policlinico San Martino, Genoa, Italy, between 1/1/2106 and 31/12/2019. We considered clinical data, serum creatinine (sCr), length of hospital stay, death, comorbidities, and primary diagnoses as codified in the hospital discharge form-HDF. Patients with CKD stage 4-5 were excluded. In the selected patients, we defined and graded AKI according to the KDIGO criteria, by comparing the peak sCr to the lowest sCr during hospitalization under the assumption that the lowest sCr would represent baseline kidney function. Then, we divided this cohort into 3 groups: 1) patients with no AKI at all, 2) AKI calculated by sCR changes and formally codified in HDF (diagnosed AKI), 3) AKI calculated by sCR changes but not codified in HDF (i.e., undiagnosed AKI). Finally, we compared the clinical characteristics and outcomes of these groups. Results We included 56,820 pts. The incidence of AKI was 24.5% (n = 13,920), evaluating the distribution among the 3 groups, we noticed that a small percentage of AKI was reported in HDF (Fig. 1). Only in 27% of cases in medical wards, 13% in surgical wards, and 19% in Intensive care Units AKI was formally reported in HDF, while in the emergency department, AKI was diagnosed in 78% of cases. Compared with No AKI patients, those with AKI (both diagnosed and undiagnosed) had a higher prevalence of comorbidities (diabetes mellitus, heart failure, atrial fibrillation) and incidence of myocardial ischemia and sepsis (Table 1). Moreover, patients with AKI had a significantly longer hospitalization and major mortality risk, at cox regression (HR 2.6, IC 2.4-2.8, p0.000) and Kaplan Meier (Fig. 2). The mortality risk augmented in all patients with AKI analyzed by both logistic regression univariate (OR 7.1, IC 6.7-7.6, p<0.001) and multivariate analysis corrected by age, gender, and comorbidities (OR 4,5, IC 4,2-4,9, p<0.001). Interestingly, these findings were confirmed also considering only the group of patients with undiagnosed AKI (univariate: OR 5.9, IC 4.2-4.9, p<0.001; multivariate: OR 4.6, IC 4.2-4.9, p<0.001). Conclusion Undiagnosed AKI is still very common in hospitalized patients, even if it identifies a category of patients with a high risk of complications and mortality. Proper recognition of H-AKI remains a problem to face.