The systemic review of subgroup analysis on the incidence of acute kidney injury(AKI) in patients with COVID-19

Background: Acute kidney injury (AKI) occurs among patients with Coronavirus disease-19 (COVID-19) and has also been proven to be associated with in-hospital mortality. Remdesivir has been authorized for the treatment of COVID-19 patients. We conducted a systematic review to evaluate the incidence of AKI in hospitalized COVID-19 patients. The incidence of AKI in different subgroups was also investigated. Method: A thorough search was carried out to nd relevant studies in PubMed, Web of Science, medRxiv and EMBASE from 1 Jan 2020 until 1 June 2020. All systemic reviews of proportions were performed using the meta package for R project (4.0.1). Results: A total of 16199 COVID-19 patients were included in our systematic review. The pooled estimated incidence of AKI in all hospitalized COVID-19 patients was 10.0% (95% CI: 7.0-12.0%). The pooled estimated need for continuous renal replacement therapy (CRRT) in COVID-19 patients was 4% (95% CI: 3-6%). According to our subgroup analysis, the incidence of AKI could be associated with the age, disease severity and the ethnicity of the patients. The incidence of AKI in hospitalized COVID-19 patients being treated with remdesivir was 7% (95% CI: 3-13%) in a total of 5 studies. Conclusion: We found that AKI was not rare in hospitalized COVID-19 patients. The incidence of AKI could be associated with age, disease severity and ethnicity. Remdesivir probably not induced AKI in COVID-19 patients. Our systemic review provides evidence for future studies that AKI might be closely associated with SARS-CoV-2 infection.


Background
Coronavirus disease-19 , caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has led to more than 60 million infections and over 1 million deaths worldwide [1]. The mortality of COVID-19 is particularly high among older patients with chronic diseases, including hypertension, diabetes, obesity, chronic kidney disease and cardiac disease [2]. In 2003, the incidence of acute kidney injury (AKI) in patients with SARS was reported to be 6.7%, and 91.7% of patients who died were diagnosed with AKI as a complication [3]. Recent studies have suggested that the incidence of AKI during hospitalization in patients with COVID-19 has a wide range and that AKI is associated with a poor prognosis [4][5][6]. Continuous renal replacement therapy (CRRT) is usually required for critically ill COVID-19 patients, not only for the treatment of AKI but also to effectively eliminate the cytokine cascade [7]. The requirement of CRRT in COVID-19 cases also needs to be summarized.
Given the current ongoing pandemic of COVID-19, there is a need to identify safe and effective treatment options. Remdesivir, a broad-spectrum antiviral agent, has been shown to have antiviral activity against several RNA viruses, including MERS-CoV and Ebola virus disease (EVD) [8,9]. As remdesivir effectively inhibits SARS-CoV-2 in vitro and in a mouse model [10,11], it has been authorized for the treatment of COVID-19 patients in some countries, including the United States [12]. The incidence of AKI in COVID-19 patients being treated with remdesivir is still uncertain. Overall, the exact rate and characteristics of AKI associated with COVID-19 patients are not well understood. In this study, we performed a systemic review of the incidence of AKI in hospitalized patients with COVID-19.

Search Strategy
A systematic literature search was performed using PubMed, Web of Science, medRxiv and EMBASE from 1 Jan 2020 until 1 June 2020 to summarize the data of AKI among patients hospitalized with COVID-19 and being treated with remdesivir. Two authors independently carried out systematic literature searches employing the terms "kidney" OR "renal" OR "acute kidney injury" OR "acute renal failure" AND "COVID-19" OR "SARS-COV-2" to obtain the data on the AKI incidence in patients hospitalized with COVID-19. No language restrictions were applied.

Inclusion and Exclusion Criteria
Studies were included if they met the following criteria: 1) observational studies that reported the incidence of AKI in all hospitalized patients with COVID-19, and 2) observational studies or randomized, placebo-controlled trial (RCT) studies that reported the incidence of AKI in hospitalized patients with COVID-19 being treated with remdesivir.
Studies that were 1) editorials, review articles or case reports, 2) preprint article, 3) studies with incomplete information about AKI, and 4) studies that did not utilize the 2012 KDIGO criteria to de ne AKI were excluded.

Quality assessment
The methodological quality of the retrospective cross-sectional studies was assessed independently by two reviewers (Chen and Xu) using the method of the Agency for Healthcare Research and Quality (AHRQ) (http://www.ncbi.nlm.nih.gov/books/NBK 35156). An item would be scored 0 if it was answered NO or UNCLEAR; if it was answered YES, then the item was scored 1. Studies achieving an 8 or above were considered high quality. At the same time, the randomized controlled trials (RCTs) in our study were analysed using Cochrane Collaboration's tool (http://handbook-5-1.cochrane.org/). It can be divided into groups A, B and C. Studies that achieved an "A" were considered high quality.

Statistical Analysis
All systemic reviews of proportions were performed using the meta package for R project (4.0.1). The incidence of AKI in COVID-19 patients (proportion variables) was used in our study. The incidences and their 95% CIs are presented as forest plots by the Metaprop function. Statistical heterogeneity between studies was assessed using the I 2 statistic. The random-effects model was used if there was heterogeneity between studies (I 2 <50%); otherwise, the xed-effects model was adopted. Rate consolidation was conducted using ve methods (untransformed, log transformation, logit transformation, arcsine transformation, and Freeman-Tukey double arcsine transformation), and the logit transformation that yielded the results with the lowest I 2 was selected for inclusion in our study. Sensitivity analysis was performed by one-by-one exclusion. Peter's test was performed for publication bias, and signi cance was considered if P<0.05.

Literature Search and Study Characteristics
A total of 1852 papers were identi ed according to our search criteria. After the exclusion based on titles and abstracts, two authors independently looked through 204 papers. Of them, 159 publications were unrelated to AKI and therefore excluded from the study. Forty-ve papers received a full-article review, and 23 were excluded according to the exclusion criteria. The ow diagram of the selection process is shown in Fig. 1. Finally, 22 studies including 16199 COVID-19 patients met the prede ned inclusion criteria and were used to determine the incidence of AKI in COVID-19 patients. 5 of 22 studies including 972 patients were used to determine the incidence of AKI in COVID-19 patients being treated with remdesivir. Table 1 show the characteristics of the studies in this systemic review. All studies in our systemic review showing the incidence of AKI were retrospective cross-sectional studies, and most of them were of high quality (12/19). The RCTs included in our study were also of high quality.

Sensitivity analysis and publication bias
In the sensitivity analysis, we used the one-by-one exclusion method ( Supplementary Figures 4 and 5) and found similar results to our main study. Peter's test was performed to evaluate publication bias (Table 2), and no signi cant difference was detected in the analysis of the incidence of AKI in COVID-19 patients.

Discussion
In this systemic review, the results from 22 retrospective cross-sectional studies including 16199 patients hospitalized with COVID-19 from 1 January 2020 to 1 June 2020 demonstrated that AKI was not rare in COVID-19. The incidence of AKI might be associated with the age, disease severity and ethnicity of the patients in our subgroup study.
COVID-19 infection is primarily a respiratory disease, but other organs, including the kidneys, are often involved. SARS-CoV-2 enters cells via the angiotensin-converting enzyme 2 (ACE2) receptor and is highly homologous to SARS-CoV [33]. High ACE2 expression in proximal tubular epithelial cells may be a potential target for kidney injury [34]. Renal abnormalities, such as proteinuria, haematuria, and AKI, occur in patients with COVID-19 [35]. AKI is characterized by a rapid increase in serum creatinine, a decrease in urine output, or both [36]. The currently widespread AKI de nition was developed by the Kidney Disease Improving Global Outcomes (KDIGO) group in 2012 [37]. The most common causes of AKI were septic shock, post major surgery, cardiogenic shock, drug toxicity and hypovolemia [38]. The cause of AKI in COVID-19 is likely to be multifactorial, including a direct attack by SARS-CoV-2 (COVID-19associated acute kidney injury: consensus report of the 25 th Acute Disease Quality Initiative (ADQI) Workgroup) or haemodynamic instability, microcirculatory dysfunction, tubular cell injury, renal congestion, microvascular thrombi and endothelial dysfunction [39], which are commonly found in critically ill patients. Pathology from autopsies of patients with COVID-19 with renal failure revealed that the kidneys showed the presence of viral particles within both the tubular epithelium and the podocytes on electron microscopy [40], varying degrees of acute tubular necrosis (ATN), diffuse proximal tubule injury with the loss of the brush border, nonisometric vacuolar degeneration, haemosiderin granules and pigmented casts [40,41].
We found that the incidence of AKI in COVID-19 patients was 10%. A similar AKI incidence in COVID-19 patients (10.8%) was also reported in other studies [34]. The diversity of patients included in our systemic review caused heterogeneity. According to the subgroup analysis, the estimated AKI incidence of patients with an average age greater than 60 years old was 12%, meanwhile, that of patients with an average age less than 60 years old was 6%. Many reports on COVID-19 have highlighted age-related differences in health outcomes, and the mortality of COVID-19 is particularly high among older patients [42,43]. Age is also an important risk factor for AKI [44]. The pooled estimated AKI incidence in the Asian subgroup was 7%. But in the non-Asian subgroup, it was 15%. Black ancestry is also a risk factor for AKI [45]. In a large cohort study of hospitalized COVID-19 patients, 76.9% of the patients who were hospitalized with COVID-19 and 70.6% of those who died were black, whereas blacks comprised only 31% of the population [46]. There might be a difference between the criteria for hospital admission in Asian and non-Asian COVID-19 patients. A European study showed that 190/1457 (13%) of COVID-19 patients were diagnosed with AKI on arrival [47]. The incidence of AKI in ICU patients with COVID-19 was particularly high, ranging from 8%-62% [14, 17, 22-24, 26, 27]. In our subgroup analysis, we found that the incidence of AKI was 26% in ICU patients. Critically ill patients hospitalized with COVID-19 who stayed in the ICU were more likely to develop AKI [5]. Lin L proved that disease severity was associated with the incidence of AKI in COVID-19 patients [34].
The incidence of CRRT was 4% in COVID-19 patients according to our investigation. CRRT has been practiced in many sepsis patients complicated with AKI [48]. There has been growing evidence suggests that patients with severe COVID-19 infection may have cytokine storm syndrome [49,50]. CRRT can remove in ammatory factors, thus blocking the cytokine storm syndrome and ultimately reducing the damage in icted on multiple organs [51]. However, the timing of CRRT in severe COVID-19 remains controversial [49]. Additional research is needed to determine whether early CRRT could improve the prognosis of COVID-19 patients with AKI.
The introduction of antiviral drugs is a common cause of drug-induced AKI [52,53]. As shown in Figure 4, the incidence of AKI in hospitalized COVID-19 patients being treated with remdesivir was 7%. In the clinical studies of remdesivir, AKI was the most frequent adverse event leading to drug discontinuation [29,31]. Antiviral drugs cause AKI by many mechanisms, including direct renal tubular toxicity, allergic interstitial nephritis (AIN), and crystal nephropathy [54,55]. However, in animal models, remdesivir treatment was effective against MERS-CoV and did not show any side effects like AKI [56]. According to a recently published multicentre matched cohort study on remdesivir, remdesivir was not signi cantly associated with an increased incidence of AKI in COVID-19 patients, even with patients who had baseline eCrCl<30 mL/min [57]. Based on our study, we also did not get evidence for remdesivir associated AKI in COVID-19 patients. More RCTs should be studied for cogent evidence in the future.

Limitations
Our systemic review had some limitations. First, most of the studies included were retrospective cross-sectional studies, although most of them (65%) were high quality. Second, the systemic review of proportions was collected from studies with a single group. Compared with a study with two groups, heterogeneity was more common. There was statistically signi cant heterogeneity in the systemic review of AKI incidence in COVID-19 patients. The diversity of the included studies involving different disease stages or activities, ages, races and sexes might also be associated with heterogeneity. Although we performed a subgroup study, the results still had signi cant heterogeneity. As a new and unknown infectious disease, our study only summarized the studies that have already been published on this topic. The potential bias of COVID-19 patients reported may not represent all of the patients hospitalized in the pool of total COVID-19 worldwide patients. Third, there were limited original studies (n<10) for the systemic review of the incidence of AKI in hospitalized COVID-19 patients being treated with remdesivir. Finally, since the investigation of COVID-19 is ongoing, additional clinical data is expected to be published.

Conclusion
According to our study, AKI was common in hospitalized COVID-19 patients. The incidence of AKI could be associated with age, disease severity and ethnicity. Remdesivir probably not induced AKI in COVID-19 patients. Our systemic review demonstrated the clinical characteristics of AKI in COVID-19, providing evidence for future studies that AKI might be closely associated with SARS-CoV-2 infection.

Availability of data and materials
The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.

Competing interests
All of the authors declared no competing interests.