Acute Kidney Injury in Hospitalized Patients with COVID-19: Risk Factors and Serum Biomarkers

Background. AKI is one of the COVID-19 complications with high prognostic significance. In our research, we studied the prognostic role of several biomarkers that could help us understand AKI pathogenesis in patients with COVID-19. Methods. We evaluated the medical data of 500 patients hospitalized with COVID-19 in Tareev Clinic from 5 October 2020 to 1 March 2022. The diagnosis of COVID-19 was confirmed with positive RNA PCR in nasopharyngeal swabs and/or typical radiological findings on CT scans. Kidney function was assessed in accordance with KDIGO criteria. In the selected 89 patients, we evaluated serum levels of angiopoetin-1, KIM-1, MAC, and neutrophil elastase 2 and their prognostic significance. Results. The incidence of AKI in our study was 38%. The main risk factors for kidney injury were male sex, cardiovascular diseases, and chronic kidney disease. High serum angiopoetin-1 levels and a decrease in blood lymphocyte count and fibrinogen level also increased the risk of AKI. Conclusions. AKI is an independent risk factor for death in patients with COVID-19. We propose the prognostic model of AKI development, which includes the combination of serum levels of angiopoetin-1 and KIM-1 on admission. Our model can help to prevent AKI development in patients with coronavirus disease.


Introduction
In March 2020, the WHO announced a COVID-19 pandemic caused by the novel coronavirus SARS-CoV-2.
Clinical manifestations of the infection range from an asymptomatic course to severe pneumonia, requiring mechanical ventilation and associated with high mortality [1]. One of the serious complications of COVID-19 is acute kidney injury (AKI), which also determines the prognosis of the disease [2].
According to several researchers, the incidence of AKI in hospitalized patients varies from 0.5% to 39% [3][4][5]. Earlier studies have suggested an important impact of cytokine storm and acute inflammatory response to acute kidney injury in COVID-19 [6]. However, the use of glucocorticosteroids and immunobiological therapy did not significantly affect the incidence of AKI, and the leading mechanisms of renal damage are still unknown [7]. The direct cytopathic effect of the virus on the renal tubular epithelium, hyperinflammation factors ("cytokine storm"), endothelial dysfunction, activation of neutrophil traps and immunothrombosis, and the contribution of various concomitant diseases, drugs, and other factors are discussed [8][9][10].
The aim of our study was to assess the frequency and risk factors for AKI, including those revealing the underlying pathological mechanisms of kidney damage in COVID-19. We examined the serum concentration of endothelial dysfunction marker (angiopoietin-1),

The Biomarkers Levels Assessment
10 mL samples of blood samples of patients (n = 89) were collected at admission in dry plastic test tubes and centrifuged at room temperature for 15 min with a rotational frequency of 3000 rpm. The serum was immediately frozen and stored in a refrigerator at the temperature of −80 • C.

Statistical Analysis
Results are presented in medians and interquartile ranges (25th; 75th percentile). The differences between the groups were analyzed by nonparametric methods, the Mann-Whitney and Kruskal-Wallis U-test. We used the χ 2 and Fisher tests to compare the qualitative variables. To determine the correlation relationships between the variables, we used the Spearman correlation (rs).
Statistical analysis was performed using the IBM SPSS v.23.0 program (SPSS: An IBM Company, Armonk, NY, USA).

Subject Characteristics
In the examined cohort (n = 500), AKI was developed in 190 (38%) patients. In the majority of them (79.5%), AKI stage 1, in 13% AKI stage 2, in 14 patients (7.4%), AKI stage 3 was developed (Table 1). In the AKI group, older age and males were prevalent. Arterial hypertension, CKD stages 3-4, AF, coronary artery disease, and 3-4 functional classes of CHF, according to NYHA, were more frequent comorbidities in the AKI group than in the group with no AKI. Patients in the AKI group had significantly lower oxygen saturation on admission and more severe pulmonary lesions according to chest CT compared with the group without AKI. These groups also differed in most laboratory parameters of inflammation and bloodclotting activation (Supplementary Table S2). Patients with AKI were significantly more likely to receive antibiotic therapy and immunobiological therapy than patients with no AKI. Patients in this group were more likely to be transferred to the ICU, needed respiratory and inotropic support, and had a longer length of hospital stay. Mortality in patients with AKI was also significantly higher compared to patients with no AKI (35.3% vs. 5.5%, p < 0.001) ( , p = 0.02 in AKI stage 3. Since the risk of death was not significantly different in AKI stages 2 and 3, these two subgroups were combined for further analysis. The main cohort and the special group were comparable in sex, age, percent of lung parenchyma involvement, incidence of CKD, and other comorbidities. There were no significant differences in the main laboratory parameters at admission (D-dimer, leukocytes, lymphocytes, ferritin, glucose, LDH, fibrinogen, proteinuria). The median time from blood sampling to the development of AKI was 6 [1; 11.3] days.

Serum Angiopoetin-1, KIM-1, Neutrophil Elastase 2, MAC Levels in Patients with COVID-19
AKI developed in patients with a significantly higher level of angiopoietin-1 at admission compared to the patients with no AKI (Figure 1). There was a positive correlation of the serum angiopoietin-1 level with the peak levels of inflammation markers CRP and LDH (Rs = 0.258, p = 0.024 and Rs = 0.334, p = 0.003, respectively), as well as with the level of D-dimer (Rs = 0.338, p = 0.003), and negative correlation with the lymphocyte (Rs = −0.338, p = 0.003) and platelet counts (Rs = −0.411, p = 0.0001) and the level of fibrinogen (Rs = −0.229, p = 0.047).
KIM-1 serum levels on admission were lower in patients who developed AKI ( Figure 2). The level of another factor, MAC, in the blood serum at admission did not differ significantly in patients with AKI and with no AKI (Figure 3). A significant increase in neutrophil elastase 2 was detected in patients with more severe AKI stages 2-3 ( Figure 4). There were significant correlations of the neutrophil elastase 2 level with markers of inflammation: CRP (Rs = 0.254, p = 0.016), LDH (Rs = 0.255, p = 0.016), fibrinogen levels (Rs = 0.351, p = 0.045), and proteinuria (Rs = 0.306, p = 0.046).
AKI developed in patients with a significantly higher level of angiopoietin-1 at admission compared to the patients with no AKI (Figure 1). There was a positive correlation of the serum angiopoietin-1 level with the peak levels of inflammation markers CRP and LDH (Rs = 0.258, p = 0.024 and Rs = 0.334, p = 0.003, respectively), as well as with the level of D-dimer (Rs = 0.338, p = 0.003), and negative correlation with the lymphocyte (Rs = −0.338, p = 0.003) and platelet counts (Rs = −0.411, p = 0.0001) and the level of fibrinogen (Rs = −0.229, p = 0.047). KIM-1 serum levels on admission were lower in patients who developed AKI ( Figure  2). The level of another factor, MAC, in the blood serum at admission did not differ significantly in patients with AKI and with no AKI (Figure 3). A significant increase in neutrophil elastase 2 was detected in patients with more severe AKI stages 2-3 (

Uni-and Multifactorial Regression Analysis
According to unifactorial logistic regression analysis, the male sex increased the risk of AKI in patients 1.7 times [95% CI 1.2-2.5, p = 0.003]. Among the concomitant diseases, the AKI risk was significantly increased by cardiovascular diseases: arterial hypertension, AF, coronary artery disease with a history of MI, and CHF of 3-4 functional classes. A history of CKD increased the risk of AKI by 1.8 times [95% CI 1.2-2.7, p = 0.008].
Among the laboratory parameters, an increased risk of AKI was associated with proteinuria, the severity of lymphopenia, high D-dimer levels, and a decrease in fibrinogen levels and platelet counts. Levels of acute phase proteins (CRP, ferritin) did not contribute to the increased risk of AKI (Table 2). Most medications in the hospital affect the risk of kidney damage. In contrast, metformin reduced the risk of AKI 1.6 times [OR 0.6 95% CI 0.3-0.9), p = 0.022] ( Table 2) In a multifactorial logistic regression model, male sex, detection of proteinuria, systemic hypoxia associated with respiratory and heart failure, lymphopenia, and a decrease in fibrinogen levels had an independent effect on the risk of AKI. Moreover, an increased risk of AKI was associated with nephrotoxic antibiotics (OR 1.9 [95% CI 1.1-3.3]) (amikacin, vancomycin). In contrast, treatment with metformin reduced the risk of AKI by 2.5 times (Table 2).

Uni-and Multifactorial Regression Analysis
According to unifactorial logistic regression analysis, the male sex increased the risk of AKI in patients 1.7 times [95% CI 1.2-2.5, p = 0.003]. Among the concomitant diseases, the AKI risk was significantly increased by cardiovascular diseases: arterial hypertension, AF, coronary artery disease with a history of MI, and CHF of 3-4 functional classes. A history of CKD increased the risk of AKI by 1.8 times [95% CI 1.2-2.7, p = 0.008].
Among the laboratory parameters, an increased risk of AKI was associated with proteinuria, the severity of lymphopenia, high D-dimer levels, and a decrease in fibrinogen levels and platelet counts. Levels of acute phase proteins (CRP, ferritin) did not contribute to the increased risk of AKI (Table 2). Most medications in the hospital affect the risk of kidney damage. In contrast, metformin reduced the risk of AKI 1.6 times [OR 0.6 95% CI 0.3-0.9), p = 0.022] ( In a multifactorial logistic regression model, male sex, detection of proteinuria, systemic hypoxia associated with respiratory and heart failure, lymphopenia, and a decrease in fibrinogen levels had an independent effect on the risk of AKI. Moreover, an increased risk of AKI was associated with nephrotoxic antibiotics (OR 1.9 [95% CI 1.1-3.3]) (amikacin, vancomycin). In contrast, treatment with metformin reduced the risk of AKI by 2.5 times (Table 2).

ROC Analysis
In the ROC analysis of the risk of AKI, we noted a moderate value of each individual serum biomarker: angiopoietin-1, neutrophil elastase 2, and KIM-1 (Table 3). Table 3. ROC analysis of AKI serum biomarkers (angiopoetin-1, KIM-1, neutrophil elastase 2). In order to increase the predictive value of these factors in AKI risk, the predicted value of the markers combinations was determined. Among the possible options, the model including angiopoietin-1 and KIM-1 (AUC 0.833 [95% CI 0.7-0.9], p = 0.001, sensitivity 83.3%, specificity 76.9%) had the highest significance in determining the risk of AKI. The addition of neutrophil elastase 2 did not significantly change the AUC (0.808 [95% CI 0.6-0.9], p = 0.003) ( Figure 5).

Discussion
The incidence of AKI in our cohort of patients hospitalized with COVID-19 was 38%. AKI increased the risk of death by 3.7 times, which is consistent with other studies. In the research of Chan L et al., the incidence of AKI was 43.5% [13]. According to Feng X et al., acute kidney damage was observed in 30.7% of patients with COVID-19 [14]. In our cohort, AKI stage 1 increased the risk of an adverse outcome by 1.4 times, while severe kidney damage (stages 2-3) increased the risk of an unfavorable outcome by 5.4 times. These data are in agreement with the data of Cheng Y et al., who showed that AKI stage 1, stage 2, and stage 3 increased the risk of death by 1.9, 3.5, and 4.7 times, respectively [15].

Discussion
The incidence of AKI in our cohort of patients hospitalized with COVID-19 was 38%. AKI increased the risk of death by 3.7 times, which is consistent with other studies. In the research of Chan L et al., the incidence of AKI was 43.5% [13]. According to Feng X et al., acute kidney damage was observed in 30.7% of patients with COVID-19 [14]. In our cohort, AKI stage 1 increased the risk of an adverse outcome by 1.4 times, while severe kidney damage (stages 2-3) increased the risk of an unfavorable outcome by 5.4 times. These data are in agreement with the data of Cheng Y et al., who showed that AKI stage 1, stage 2, and stage 3 increased the risk of death by 1.9, 3.5, and 4.7 times, respectively [15].
Cardiovascular diseases (arterial hypertension, atrial fibrillation, CHF), as well as CKD with a decrease in GFR of less than 60 mL/min, were the comorbidities that influenced the renal injury risk in our cohort. Similar results were obtained by other authors. In the work of Chan L et al., CHF was also identified as one of the prognosis-determining comorbidities in the group of patients with AKI (13% of patients) [13]. In the other prospective study, Sullivan M et al. found that the patient's history of CKD increased the risk of developing AKI by 1.6 times [16]. According to Hirsch JS et al., among the concomitant diseases, the independent risk factors for the development of AKI were arterial hypertension and diabetes mellitus [4].
Among the comorbidities, heart failure had an independent effect on the risk of AKI in a multifactorial regression model in our cohort. The initial impairment of intrarenal hemodynamics in this category of patients increased the risk of cardiorenal syndrome and acute decline of glomerular filtration rate, as shown in a previous study by Fang et al., congestive CHF increased the risk of developing AKI by 14.8 times [17]. Kolhe N et al. showed that the presence of congestive CHF in a patient was also an independent factor of renal damage and AKI by 1.7 times [18].
Although many risk factors for AKI are widely discussed, the underlying mechanisms of renal damage are not fully understood. This is also evidenced by the maintenance of the previous incidence of AKI, despite the use of corticosteroids and immunobiological drugs to suppress the systemic inflammatory response.
In our study, laboratory parameters for assessment were chosen in accordance with many studies dedicated to the evaluation of the severity and prognosis of the novel coronavirus infection [19]. Ng et al. discuss markers of hyperinflammation in patients with COVID-19 and AKI and refer to secondary hemophagocytic lymphohistiocytosis. Cardinal markers of this syndrome are IL-2,6,7, ferritin, and some others, which are associated with COVID-19 severity and AKI incidence [20]. Moreover, CRP, LDH, ferritin, fibrinogen, and D-dimer levels were evaluated by Yildirim et al., Serum levels of these biomarkers were significantly higher in patients with AKI than in patients with preserved renal function [21].
According to our study, independent risk factors for AKI in patients hospitalized with COVID-19 were male sex, activation of endothelial dysfunction, and mechanisms of immunothrombosis, as well as respiratory and circulatory hypoxia due to severe forms of CHF. Markers of inflammation have not demonstrated a significant impact on AKI development. The earliest mechanism of renal damage is endothelial dysfunction. Angiopoietin-1 is involved in blood vessel differentiation, growth, and remodeling [22]. In the acute phase of inflammation, it is involved in the regulation of the vascular wall tone and permeability. In our study, the serum concentration of angiopoietin-1 at admission was significantly higher in patients who had AKI at follow-up than in patients with no AKI during the in-hospital stay. In addition, angiopoietin-1 was one of the risk factors for AKI, which showed its significance in the multifactorial regression model. Given the significant correlations of angiopoietin-1 with the levels of acute phase reactants, as well as the degree of lymphopenia, it is possible to assume the effect of proinflammatory cytokines and the SARS-CoV-2 itself on the endothelium with its damage and consequent endothelial dysfunction. The correlation of the level of angiopoietin-1 with the level of D-dimer, a decrease in the number of platelets, and the level of fibrinogen indicates the activation of the processes of microthrombi formation and consumption of coagulation factors following endothelial dysfunction.
Our results are consistent with the data of other authors, who revealed a high predictive value of the serum angiopoietin-1 level in relation to 28-day mortality in patients with sepsis in the ICU departments [23].
Based on the evidence of the important role of NETosis in kidney damage in COVID-19, we established a significantly higher baseline serum level of neutrophil elastase 2 in patients with AKI stages 2 and 3 at a late period of hospitalization. Neutrophil elastase 2 also correlated with inflammatory markers (CRP and LDH) and proteinuria. In the Henry B et al. study, neutrophil elastase 2 levels were shown to be significantly higher in patients with AKI compared to patients without AKI, and it correlated with another surrogate marker of NETosis, extracellular DNA circulating in the blood [24]. The obtained results allow us to discuss the contribution of the NETosis process to the development of severe AKI. NETosis in circulating blood neutrophils and possibly in tissue neutrophils, which infiltrate the kidneys and lungs, plays a role in prolonging inflammation, the development of a cytokine storm, and on the other hand, can cause activation of blood clotting by the internal pathway and the process of immunothrombosis [25]. The increase in neutrophil elastase 2 in our study correlated with proteinuria, which appears to have inflammatory and ischemic genesis in COVID-19 patients.
Our study showed a tendency for patients with particularly severe stages of AKI to have lower serum levels of KIM-1 on admission. The established pattern is not entirely clear and requires further investigation. In the works of other authors, KIM-1 in patients with COVID-19 was studied in the urine, and its level was significantly increased in patients who developed AKI [26]. In addition, KIM-1 in urine is a well-studied marker of tubule damage in AKI of ischemic and toxic genesis [27].
In the serum of COVID-19 patients, KIM-1 levels were assessed by Kerget B et al., but kidney damage was not studied in this study [28]. According to the authors, serum KIM-1 was elevated in patients with extremely severe novel coronavirus infection compared to patients with a moderate course. In experimental studies, it was found that KIM-1, in addition to ACE, is a receptor for SARS-CoV-2 [29]. The KIM-1 molecule is expressed in many tissues of the body, including the tubular epithelium cells in kidneys, in response to a damaging factor, for example, hypoxia, from where it can enter the bloodstream, cleaved from the extracellular domain by metalloproteinases [30]. The KIM-1 receptor is able to bind particles of the virus, "blocking" it and preventing its further spread throughout the body. We believe that in the process of active KIM-1 binding to the epitopes of the virus, its amount in the blood may temporarily decrease, for example, immediately before the development of AKI.
The role of complement activation as one of the leading mechanisms of renal damage is discussed, but in our study, the level of MAC in the blood serum on admission did not differ significantly in patients with the development of AKI and without it, which indicates that there is no direct effect of complement activation on the processes of kidney tissue damage at least at an early stage of the disease [31].
According to the ROC analysis, serum angiopoietin-1 and KIM-1 levels were the most informative combination for predicting AKI development within a week from the day of hospitalization, indicating a predominant activation of endothelial dysfunction and the virus's consumption of KIM-1 receptor determinants.
Metformin had a protective effect on AKI and an unfavorable outcome, reducing the risk of AKI by 2.7 times and the risk of death by 3.2 times in our study. These data are consistent with data from other studies. For example, in a study of 136 patients with COVID-19 and DM2, metformin reduced the rate of ICU admission and death by 2.9 and 5.5 times, respectively [32]. In the study by Do et al., patients with COVID-19 and DM2 taking metformin had a significantly lower incidence of AKI than patients taking other glucose-lowering drugs (1.7% vs. 3.3%, p = 0.021, respectively) [33] Luo et al., in their study, showed, that metformine significantly decreased in-hospital mortality in patients with COVID-19 and DM2. The death rate in patients taking metformine was 2.9% (n = 3) vs. 12.3 (n = 22) in patients, who did not take metformin, p = 0.01 [34]. In the work of Bell S et al., it was shown that metformin significantly increased 28-day survival in patients with AKI [35].
The pleiotropic effects of metformin, such as anti-inflammatory and antiviral actions, are discussed in addition to the hypoglycemic effect. The anti-inflammatory effect is mediated by AMPK modulation, which decreases the expression of proinflammatory genes, decreasing levels of the inflammatory markers in patients with and without DM2 [36,37]. Moreover, metformin has a favorable influence on the endothelium, preventing its damage and decreasing the risk of thrombosis by inhibiting the platelet activation factor [38]. Metformin is thought to reduce the activity of NETosis. In the study by Menegazzo et al., serum levels of NETosis biomarkers (cell-free dsDNA, histones, elastase, proteinase-3, neutrophils gelatinase-associated lipocalin (NGAL), and lactoferrin) were measured in 91 patients with pre-diabetes (n = 18), DM2 with bad glucose control (n = 73). It was shown that metformin significantly reduced serum levels of NETosis markers such as elastase, proteinase-3, histones, and double-strand DNA. Moreover, metformin prevented DNA release and membrane translocation of PKC-βII and activation of NADPH oxidase in neutrophils [39]. By AMPK activation, metformin also interferes with binding to the SARS-CoV-2 receptor ACE2 by its phosphorylation and also reduces the expression of proinflammatory genes by inhibiting the mTOR pathway [40].
Our study has limitations. The number of samples for the study of angiopoietin-1, neutrophil elastase 2, KIM-1, and membrane attack complex was relatively small. These parameters were assessed only at the time of admission. The dynamics of serum angiopoetin-1, ELA 2, KIM-1, and MAC levels, as well as urinary KIM-1 levels, were not evaluated.

Conclusions
The clinical course of COVID-19 was complicated by AKI in 38% of cases. The independent risk factors of AKI in patients with COVID-19 are male sex, cardiovascular diseases, and chronic kidney disease, as well as high serum angiopoetin-1 levels and a decrease in blood lymphocyte count and fibrinogen levels. We propose the prognostic model of AKI development, which include the combination of serum levels of angiopoetin-1 and KIM-1 on admission.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/biomedicines11051246/s1, Table S1: Clinical and demographic characteristics of hospitalized patients with COVID-19 in the special sample (n = 89). Table S2: Laboratory parameters of hospitalized patients with COVID-19 (n = 500). Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The datasets used during the current study are available from the corresponding author upon reasonable request.

Conflicts of Interest:
The authors declare no conflict of interest.