Epidemiological profile of acute kidney injury in critically ill patients admitted to intensive care units: A Prospective Brazilian Cohort

Abstract Introduction Acute kidney injury (AKI) is a frequent syndrome affecting patients admitted to intensive care units (ICU), and it is associated with poor clinical outcomes. The aim of the present study was to understand the epidemiological profile of patients with AKI admitted to ICUs. Methods Prospective cohort study, carried out in three ICUs in the Federal District, Brazil. Between October/2017 and December/2018, 8,131 patients were included in the cohort. AKI was defined according to the KDIGO criteria. The main outcomes assessed were AKI development and mortality within 28 days of hospitalization. Results Of the 8,131 patients followed up, 1,728 developed AKI (21.3%). Of the 1,728 patients with AKI, 1,060 (61.3%) developed stage 1, while stages 2 and 3 represented 154 (8.9%) and 514 (29.7%), respectively. Of these, 459 (26.6%) underwent renal replacement therapy. The mortality was 25.7% for those with AKI, and 4.9% for those without AKI. Discussion Patients with AKI had higher mortality rates when compared to those without AKI. Likewise, among patients with AKI, higher disease stages were associated with higher death occurrences. AKI incidence (21.3%) and mortality (25.7%) in our study is in line with the largest meta-analysis ever conducted, in which incidence and mortality of 21.6 and 23.9% were observed, respectively. These findings confirm the importance of establishing the KDIGO guideline for the definition and management of AKI in Brazilian ICUs.

Of these, 459 (26.6%) underwent renal replacement therapy. The mortality was 25.7% for those with AKI, and 4.9% for those without AKI. Discussion: Patients with AKI had higher mortality rates when compared to those without AKI. Likewise, among patients with AKI, higher disease stages were associated with higher death occurrences. AKI incidence (21.3%) and mortality (25.7%) in our study is in line with the largest meta-analysis ever conducted, in which incidence and mortality of 21.6 and 23.9% were observed, respectively. These findings confirm the importance of establishing the KDIGO guideline for the definition and management of AKI in Brazilian ICUs.

IntRoductIon
Acute kidney injury (AKI) is a frequent syndrome in patients admitted to intensive care units (ICU), and it is associated with increased length of hospital stay, high costs to the healthcare system and high mortality rates. [1][2][3] Most epidemiological studies involving AKI have been carried out in North America, Europe and Asia; therefore, we have little information about it in Latin America. 4,5 It seems that the outcomes perceived in AKI patients in developed countries are different when compared to those from developing countries. 6 There are few studies in Brazil, with the few existing ones being retrospective in design or not bearing a very representative sample. 7,8 A better understanding of the epidemiological profile of AKI in Brazil could open the field for a series of investigations, in order to better understand its relationship with length of hospital stay and mortality, in addition to enabling cost reductions in the healthcare system. In this sense, our goal with present study was to better understand the epidemiological profile of patients with AKI admitted to ICUs in the Federal District, Brazil.

Study deSign
Observational prospective cohort study, carried out in critical patients admitted to three ICUs of tertiary hospitals in the Federal District during the period from October/2017 to December/2019. It was approved by the Research Ethics Committee of the Centro Universitário ICESP (nº 3.608.561). Eligible patients were those aged ≥18 years. Those receiving dialysis, with chronic kidney disease on renal replacement therapy (RRT), kidney transplant patients and with hospitalization <48 hours were excluded. We followed the STROBE recommendations (Strengthening the Reporting of Observational Studies in Epidemiology). 9

data ColleCtion
A previously trained team collected the data from the electronic records of the patients, and such team was led by a nurse specialized in clinical research (EAV). All data were collected at the time of admission to the ICU, daily. The collection included the baseline characteristics of the patients, comorbidities (hypertension, diabetes, heart failure, cancer, coronary artery disease, COPD, cirrhosis and others), clinical categories of ICU admission (pulmonary, sepsis, surgical, cardiovascular, gastrointestinal, infectious, neurological, orthopedic, oncological, acid-base and hydroelectrolytic disorders and others), serum creatinine and simplified acute physiological score (SAPS 3) to assess clinical severity and risk of mortality.
The following variables were investigated during follow up: development of AKI; need for RRT; use of mechanical ventilation; administration of vasoactive drugs or nephrotoxic antibiotics; length of stay in the hospital and ICU; and clinical outcome (death, discharge and hospital transference).

aki definition and StratifiCation
All admitted and eligible patients were followed for 28 days in the ICUs, in another inpatient unit or until some outcome, such as hospital discharge, death or hospital transference. AKI was defined according to the criteria proposed by Kidney Disease: Improving Global Outcomes (KDIGO), 10 of which: increase in serum creatinine ≥ 0.3 mg/dL within 48 hours or ≥ 50% in the first 7 days compared to creatinine levels upon admission to the ICU. We stratified the patients diagnosed with AKI according to the KDIGO: 1) increase in serum creatinine from 1.5 to 1.9 times or 0.3 mg/dL; 2) 2.0 to 2.9 times; and 3) ≥ 3.0 times the baseline value or serum creatinine ≥ 4.0 mg/dL or RRT onset. Urine output was not used for the diagnosis and stratification of AKI, because of the difficulties presented in the collection and appointment by the ICU technical personnel.

StatiStiCal analySiS
Descriptive statistics was conducted using the mean, standard deviation and frequency values. Kolmogorov-Smirnov test was used to check data normality. To compare continuous variables between groups with and without AKI, we used the independent Student's t-test. For categorical variables and frequencies, the chi-square test was applied. The level of significance was set at 95%. The analysis was perfomed using the Statistical Package for the Social Sciences (SPSS) program, version 26.0 (SPSS Inc., Chicago, IL, USA).

Results
A total of 8,131 patients were admitted and followed for a period of 9.3 ± 16.1 days in the ICUs. Table 1 shows the admission characteristics of the patients included in the cohort.
The number of patients who developed AKI in the first 7 days of hospitalization was 1,728 (21.3%). The creatinine level of patients upon admission who developed AKI was 1.55 ± 1.76 mg/dL; higher when compared to those who did not develop AKI. There were two or more comorbidities in 35.2% of the admitted patients. According to SAPS 3, patients who had AKI had a higher predictive mortality rate than those without AKI (25.8 ± 20.8 vs. 16.7 ± 15.8; <0.001). Table 2 shows the variables monitored and clinical outcomes during the ICU stay.
The patients who developed AKI had a higher prevalence of arterial hypertension (64.3%) and diabetes (35%), as well as greater use of vasoactive drugs (17.0% vs. 4.3%; p <0.001) and mechanical ventilation (27.8% vs. 6.5%; p <0.001). The length of ICU stay was longer for patients who Of the 1,728 patients with AKI, 1,060 (61.3%) developed stage 1, while stages 2 and 3 represented 154 (8.9%) and 514 (29.7%), respectively. A total of 459 (26.6%) underwent RRT, with the conventional modality being the most used (61.2%). In Figure 1, it is possible to identify mortality according to the stages of AKI.
Mortality increased as the stage of AKI increased. Mortality was significantly higher in those patients with AKI versus those without AKI (p <0.001).

dIscussIon
The findings of the present study demonstrate the clinical severity of critically ill patients admitted to ICUs who develop AKI; these patients were older, had a higher number of comorbidities and had sepsis and cardiovascular pathology as the main admission categories. In addition, patients with AKI spent more time in the hospital and in the ICU, used more nephrotoxic antibiotics and vasoactive drugs, as well as required mechanical ventilation, demonstrating greater severity when compared to those patients who did not develop AKI. It is important to notice that, among patients who developed AKI, more than 60% were using nephrotoxic antibiotics. Despite all the technological advances, hospital-acquired infection remains a cause of high rates of morbidity and mortality worldwide and, unfortunately, bacterial resistance resulting from the injudicious use of antibiotics is a serious problem. In this study we did not perform a specific analysis to investigate the reasons for the high use of these antibiotics and the impact of AKI development, which will be better explored in future publications.
Finally, confirming our hypothesis, patients with AKI had higher mortality, and, similar to Hoste et al 11 there was a gradual increase in mortality, with an increase in severity in AKI stages. AKI is a frequent and important problem in critically-ill patients, and its prevalence has increased considerably, both in developed and developing countries, to the point of becoming a public healthcare problem. 3,12 Most published epidemiological studies are from developed countries, but represent only 15% of the world population. 4 The worldwide incidence of AKI was analyzed in a meta-analysis with 312 studies. Of these, 154 (130 in adults and 24 in children) totaled  4 The KDIGO definition for AKI diagnosis has been adopted as a consensus established by experts worldwide. 4,13 Despite this, many Brazilian intensive care physicians continue to adopt other guidelines. 13 The incidence of AKI (21.3%) and mortality (25.7 %) in our study is in line with the meta-analysis conducted by the Acute Kidney Injury Advisory Group of the American Society of Nephrology, 4 with incidence and mortality of 21.6 and 23.9%, respectively. These findings confirm the importance of establishing the KDIGO guideline for the definition and management of AKI in Brazilian ICUs.
Based on our survey conducted in the literature, this is the largest prospective cohort study in Latin America to assess AKI occurrence in ICUs using the KDIGO criteria for diagnosis and classification. In the Federal District, where the sample was collected, there is a high concentration of inhabitants from other states and other regions of Brazil; therefore, we suggest that these findings may represent the reality of Brazil, especially in private ICUs; however, this may be different in public ICUs.
Our study has some characteristics that show a certain robustness in our findings, such as: i) large sample size; ii) prospective methodological design; therefore, we monitor the admitted patients daily, unlike the vast majority of cohorts, which have a retrospective design; and iii) large participation of inhabitants from other regions of Brazil, which gives the study a possible portrait of the epidemiological profile of AKI in private ICUs in Brazil.
The main limitations are: i) the fact that data collection occurred only in patients in ICUs, not representing the AKI acquired in the community. Another limitation was the fact that only three private hospitals were involved in the study, although it represents different regions of the Federal District. Finally, it was not possible to obtain serum creatinine results from at least the last three months prior to admission to the ICU, which led us to evaluate the serum creatinine upon admission to the ICU to monitor possible AKI occurrence.
We concluded, through our cohort, that the incidence of AKI was 21.3%. There was a higher mortality of patients with AKI, compared to those who did not develop it. Likewise, among patients with AKI, higher disease stages were associated with higher mortality rates.