COVID-19 and Acute Kidney Injury: A Systematic Review and Meta-Analysis

Background: coronavirus disease 2019 (COVID-19), caused by severe acute respiratory syndrome—coronavirus-2 (SARS-CoV-2)—is an ongoing pandemic with high morbidity and mortality rates. Preliminary evidence suggests that acute kidney injury (AKI) is uncommon in patients with COVID-19 and associated with poor outcomes. Study aims and design: we performed a systematic review of the literature with a meta-analysis of clinical studies to evaluate the frequency of AKI and dialysis requirement in patients who underwent hospitalization due to COVID-19. The incidence of AKI according to the death risk was calculated in these patients. The random-effects model of DerSimonian and Laird was adopted, with heterogeneity and stratified analyses. Results: thirty-nine clinical studies (n = 25,566 unique patients) were retrieved. The pooled incidence of AKI was 0.154 (95% CI, 0.107; 0.201; p < 0.0001) across the studies. Significant heterogeneity was found (p = 0.0001). The overall frequency of COVID-19-positive patients who underwent renal replacement therapy (RRT) was 0.043 (95% CI, 0.031; 0.055; p < 0.0001); no publication bias was found (Egger’s test, p = 0.11). The pooled estimate of AKI incidence in patients with severe COVID-19 was 0.53 (95% CI, 0.427; 0.633) and heterogeneity occurred (Q = 621.08, I2 = 97.26, p = 0.0001). According to our meta-regression, age (p < 0.007) and arterial hypertension (p < 0.001) were associated with AKI occurrence in hospitalized COVID-19 positive patients. The odds ratio (OR) for the incidence of AKI in deceased COVID-19 positive patients was greater than among survivors, 15.4 (95% CI, 20.99; 11.4; p < 0.001). Conclusions: AKI is a common complication in hospitalized COVID-19 positive patients. Additional studies are under way to assess the risk of AKI in COVID-19 patients and to deepen the mechanisms of kidney injury.


Introduction
The newly discovered Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2, previously named as 2019 novel coronavirus or 2019-nCoV) has been identified as the pathogen of Coronavirus Disease 2019 . The novel virus identified in Wuhan, China, in December 2019 and has spread rapidly all over the world. The World Health Organization (WHO) declared COVID-19 a pandemic in March 2020 [1]. Typical patient clinical manifestations included fever, unproductive cough, dyspnea, fatigue, normal or lowered white blood cell count, and imaging evidence of pneumonia. The clinical course of infection by SARS-CoV-2 is widely unpredictable and variable, ranging from asymptomatic infection to multi-organ systemic failure and death.

End-Points of Interest
Primary outcomes of interest were the pooled incidence of AKI and the requirement of renal replacement therapy (RRT), in hospitalized patients with COVID-19. An additional end-point was the AKI occurrence in hospitalized patients with severe COVID-19. The impact of AKI on the death risk of COVID-19 patients was addressed by calculating the summary estimate for unadjusted or adjusted death risk. As detailed below, the adjusted death risk was generated by multivariate analysis in a subset of reports. The adjusted relative risks (aRR) of all-cause mortality was calculated in each study.

Statistical Methods
The summary estimate of the incidence of AKI and the need of RRT was calculated. We computed fixed and random effect estimates and the random-effects model of Der Simonian and Laird was adopted if moderate to severe heterogeneity occurred [8]. To assess the between-study heterogeneity, we used Cochran's Q test (p > 0.10 for statistical significance) and I 2 test [9]. To further explore the origin of heterogeneity, we restricted the analysis to subgroups of studies defined by study characteristics such as the country of origin (China, United States of America), and study design (retrospective or not), among others. We made a funnel plot to detect a publication bias in the relation exposure at hand; publication bias was calculated by Egger's test. Meta-regression was carried out to assess the independent effect of continuous covariates on the incidence of AKI in hospitalized patients with COVID-19. We adopted the odds ratio (OR) with 95% CI for the dichotomous outcomes. In a subset of reports, a summary estimate of the adjusted RR of all-cause mortality among hospitalized COVID-19 patients who developed AKI compared with those who did not was generated by weighting the study-specific RR's (by the inverse of the variance). The aRR was calculated by multivariate analysis (i.e., after adjustment for potential confounders such as comorbidities and complications). Ri (the proportion of total variance due to between studies variance) was adopted to take into account the heterogeneity. All the statistical analyses were performed using Rev Man (Review Manager) 5.0, The Cochrane Collaboration (2020), Comprehensive meta-analysis (CMA 1.0), and HEpiMA, version 2.1.3 [10]. The 5% significance level was adopted for alpha risk. Every estimate was recorded with 95% confidence intervals (CIs).

Literature Review
Our electronic and manual searches identified 432 full-text articles that were considered potentially relevant and selected for full-text review. A complete list of the 432 full-text articles reviewed is available from the authors on request. We excluded 393 full-text articles, as detailed in Figure 1.
We included a total of 39 reports giving information on 25,566 patients who had been admitted to tertiary hospitals all over the world and were diagnosed with COVID-19 ( Figure 1). There was a 100% concordance between reviewers with respect to the final inclusion and exclusion of studies based on the predefined exclusion criteria.

Patient Characteristics
According to the design of the study, two sets of reports were identified. The first set was constituted by studies listed in Tables 1 and 2 and Supplementary Table S1 (n = 22; n = 8792 patients) which evaluated the frequency of AKI based on the severity of COVID-19 . The frequency of chronic obstructive pulmonary disease (COPD) ranged between 1.1 and 13%, and the chronic liver disease (CLD) rate between 0.9% and 11.8%. The second set included reports shown in Tables 3-4 and Supplementary Table S2 (n = 20, n = 16,774 patients) which assessed the impact of AKI development upon the outcomes (death rate) of COVID-19 patients [33][34][35][36][37][38][39][40][41][42][43][44][45][46][47][48][49]. Some (n = 3) studies gave information on both relationships [16,24,28]. The frequency of COPD varied between 1.9 and 12%, and the rate of CLD between 0.52% and 6.3%. Tables 1 and 3 report the list of studies evaluated, the countries where they were carried out, the reference year and some demographic data. All studies were conducted between January and June 2020. As listed in Tables 1-4, the majority of reports were from China (n = 30), and the others from the USA (n = 7) and South Korea (n = 2), respectively. The frequency of male patients ranged from 38.8% to 75%, and the mean age from 47 to 69 years. Comorbidities (arterial hypertension, diabetes mellitus, chronic kidney disease (CKD), cardiovascular disease) have been recorded in Tables 2 and  4 and Supplementary Tables S1 and S2. The majority of the studies adopted the definition of AKI according to the 2012 Kidney Disease: Improving Global Outcomes (KDIGO) clinical practice guidelines, where AKI is defined as any of the following: increase in serum creatinine by >0.3 mg/dL (>26.5 μmol/L) within 48 h; or increase in serum creatinine to >1.5 times baseline, which is known or presumed to have occurred within the prior 7 days; or < urine volume 0.5 mL/kg/h for 6 h [50]. In some papers, the definition of AKI was not mentioned [16,36,38,40,46].

Patient Characteristics
According to the design of the study, two sets of reports were identified. The first set was constituted by studies listed in Tables 1 and 2 and Supplementary Table S1 (n = 22; n = 8792 patients) which evaluated the frequency of AKI based on the severity of COVID-19 . The frequency of chronic obstructive pulmonary disease (COPD) ranged between 1.1 and 13%, and the chronic liver disease (CLD) rate between 0.9% and 11.8%. The second set included reports shown in Tables 3 and 4 and Supplementary Table S2 (n = 20, n = 16,774 patients) which assessed the impact of AKI development upon the outcomes (death rate) of COVID-19 patients [33][34][35][36][37][38][39][40][41][42][43][44][45][46][47][48][49]. Some (n = 3) studies gave information on both relationships [16,24,28]. The frequency of COPD varied between 1.9 and 12%, and the rate of CLD between 0.52% and 6.3%.   Tables 1 and 3 report the list of studies evaluated, the countries where they were carried out, the reference year and some demographic data. All studies were conducted between January and June 2020. As listed in Tables 1-4, the majority of reports were from China (n = 30), and the others from the USA (n = 7) and South Korea (n = 2), respectively. The frequency of male patients ranged from 38.8% to 75%, and the mean age from 47 to 69 years. Comorbidities (arterial hypertension, diabetes mellitus, chronic kidney disease (CKD), cardiovascular disease) have been recorded in Tables 2 and 4 and Supplementary Tables S1 and S2. The majority of the studies adopted the definition of AKI according to the 2012 Kidney Disease: Improving Global Outcomes (KDIGO) clinical practice guidelines, where AKI is defined as any of the following: increase in serum creatinine by ≥0.3 mg/dL (≥26.5 µmol/L) within 48 h; or increase in serum creatinine to ≥1.5 times baseline, which is known or presumed to have occurred within the prior 7 days; or < urine volume 0.5 mL/kg/h for 6 h [50]. In some papers, the definition of AKI was not mentioned [16,36,38,40,46]. Table 5 shows that the summary estimate for the occurrence of AKI across the identified trials was 0.154 (95% CI, 0.107, 0.201). Significant heterogeneity was found (Table 5) (p = 0.0001). The Egger's regression intercept shows that there was publication bias (p = 0.025) (Figure 2).  Table 5 shows that the summary estimate for the occurrence of AKI across the identified trials was 0.154 (95% CI, 0.107, 0.201). Significant heterogeneity was found (Table 5) (p = 0.0001). The Egger's regression intercept shows that there was publication bias (p = 0.025) (Figure 2).

AKI Incidence: Primary and Stratified Analysis
Stratified analyses were undertaken to explain the heterogeneity across studies ( Table 5). The analysis by the fixed-effects model yielded very similar findings to the random-effects model (data not shown).   Stratified analyses were undertaken to explain the heterogeneity across studies ( Table 5). The analysis by the fixed-effects model yielded very similar findings to the random-effects model (data not shown).
The overall estimate for the frequency of COVID-19 positive patients who had AKI and underwent RRT during their hospital stay was 0.043 (95% CI, 0.031; 0.055) (Figure 3  * small studies (studies with size < 100 pts).

AKI Incidence and Death Rate in COVID-19 Positive Patients
The pooled OR of AKI incidence among deceased COVID-19 positive patients was greater than among survivors, 15.4 (95% CI, 20.99; 11.4). Test for heterogeneity was significant (p = 0.00001) ( Figure  5). Publication bias occurred (Egger's regression, p = 0.0016) (Figure 6).   Some authors (n = 5, n = 5435 unique patients) evaluated the association between AKI and death risk by multivariate analysis. As shown in Table 6, the link between AKI and death risk remained significant in many comparisons. The results of meta-regression are reported in Table 7. Age (p = 0.007) and arterial hypertension (p = 0.001) are significantly associated with the frequency of AKI. The independent and significant relationship between the frequency of AKI and arterial hypertension according to meta-regression is shown in Figure 7.  Some authors (n = 5, n = 5435 unique patients) evaluated the association between AKI and death risk by multivariate analysis. As shown in Table 6, the link between AKI and death risk remained significant in many comparisons. The results of meta-regression are reported in Table 7. Age (p = 0.007) and arterial hypertension (p = 0.001) are significantly associated with the frequency of AKI. The independent and significant relationship between the frequency of AKI and arterial hypertension according to meta-regression is shown in Figure 7.

Discussion
Controversy exists about kidney involvement in COVID-19-positive patients. Preliminary evidence indicated that the frequency of kidney disease in the COVID-19 population was negligible and limited interest has been given to the incidence of AKI in patients with COVID-19 [3]. Additional studies have highlighted the frequency of kidney abnormalities in patients with COVID-19 [51]. The

Discussion
Controversy exists about kidney involvement in COVID-19-positive patients. Preliminary evidence indicated that the frequency of kidney disease in the COVID-19 population was negligible and limited interest has been given to the incidence of AKI in patients with COVID-19 [3]. Additional studies have highlighted the frequency of kidney abnormalities in patients with COVID-19 [51]. The current systematic review of the scientific literature with a meta-analysis of clinical studies indicated that the incidence of AKI in patients with COVID-19 during their hospital stay was common (around 15%). The frequency of AKI among patients with severe COVID-19 was much greater (around 50%). We noted important heterogeneity that could be explained by numerous factors such as patient characteristics, severity of illness, differences in daily clinical practice (regarding fluid management, ventilation options and medications, among others).
According to our meta-regression analysis, some comorbidities (age and arterial hypertension) were significantly related to AKI occurrence and this is in keeping with the evidence on the development of AKI in patients without COVID-19.
The pathophysiological mechanisms which are responsible for COVID-19-related AKI are yet to be discovered [51]. Unspecific mechanisms exist including comorbidities (diabetes mellitus, arterial hypertension, and others) which confer vulnerability to kidneys, nephrotoxic drugs or contrast media, hypovolemic conditions and subsequent pre-renal AKI. Multiorgan involvement is common in patients with COVID-19 including damage to kidneys, heart, and gastrointestinal tract; this mirrors the presence of the ACE2 receptors in various organs which serve as an entrance door for SARS-CoV-2. It has been hypothesized that the development of AKI in COVID-19 patients include viral cytopathic activity, hypoperfusion, cytokine storm, and microvascular thrombosis [51]. Alternatively, patients with severe SARS-CoV-2 infection frequently show acute respiratory distress syndrome (ARDS); severe hypoxemia or high intra-thoracic pressures have been linked to AKI in the ARDS population [52]. COVID-19-specific mechanisms include the entry of SARS-CoV-2 into the kidneys and the binding of SARS-CoV-2 with the ACE2 receptor on the cell membrane of the host cells; in the kidneys, the ACE2 receptor is expressed in the apical brush borders of the proximal tubules as well as podocytes [53]. In addition, COVID-19 promotes an imbalanced activation of the renin-angiotensin-aldosterone system (RAAS), which induces the downregulation of the membrane-bound ACE2 receptor that promotes the accumulation of angiotensin II by lowering its degradation into angiotensin 1-7. Imbalanced RAAS activation leads to inflammation, vasoconstriction and fibrosis at the kidney level [54]. Some studies have suggested that ACE inhibitors and angiotensin receptor blockers (ARBs) may improve ACE2 expression and therefore increase the susceptibility of patients to SARS-CoV-2 infection [55]. High levels of inflammatory cytokines have been noted in severe COVID-19 patients and may participate to AKI in these patients [35].
The current meta-analysis is flawed by numerous issues. Most studies included in this study had retrospective design; there were no RCTs. It is well known that prospective studies having data at baseline and over follow-up provide better evidence. Second, individual data from each study (e.g., 'patient-level data') were not available; thus, it was impossible to make our own adjustments. An additional limitation is given by the occurrence of publication bias: negative studies are less likely to be published. In addition, an enormous body of data is rapidly accumulating on COVID-19 patients, including those with kidney disease, and this clearly makes difficult the retrieval of the whole evidence on the subject. We have not adopted the studies published as abstracts or letters as information presented in this format as these are not of high quality.
In conclusion, this meta-analysis of clinical studies shows that AKI is common in COVID-19-positive patients during their hospital stay. The frequency of AKI was much greater in patients with severe disease. There is a consistent relationship between the development of AKI and unsatisfactory outcomes (death rate) in hospitalized patients with COVID-19.
Funding: This paper was not funded.

Conflicts of Interest:
All the authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript.