Acute kidney injury in patients with SARS-CoV-2 infection

Since December 2019, severe acute respiratory coronavirus 2 (SARS-CoV-2) has spread worldwide, causing more than 6.6 million cases and 390 000 deaths [1]. This pandemic has put unprecedented pressure on healthcare systems and especially on intensive care units (ICUs).

Acute Kidney Injury (AKI) is a frequent complication of severe SARS-CoV-2 infection but data are scarce in ICUs. AKI has been previously reported with an average incidence of 11% (8–17%) overall, with highest ranges in the critically ill (23%; 14–35%) [2‐4]. Different applications of the Kidney Disease Improving Global Outcomes (KDIGO) criteria for AKI, in particular different methods to estimate missing baseline creatinine and handling urinary output, can cause important variations of estimated incidence [5, 6] and may contribute to the discrepancies among these studies.

Multiple mechanisms are involved in COVID-19-associated AKI, ranging from direct viral infection of the kidney and secondary inflammation to complement activation and microthrombosis [7]. In particular, severe COVID-19 is associated with uncontrolled systemic inflammatory response with high levels of IL-6 [8] that could potentially lead to intrarenal inflammation and increased vascular permeability and share several features with hyperferritinemic syndromes such as macrophage activation syndrome [9]. Furthermore, unrestrained activation of complement leads to endothelial cell dysfunction and intravascular coagulation that could participate in COVID-19-associated AKI [10]. Both IL-6 [11] and complement [12] have been proposed as therapeutic targets and understanding their role in COVID-19-associated AKI is therefore a priority.

However, most of the studies performed to date gave little data regarding definition of AKI or influence of inflammation and complement markers on AKI.

In this study, we sought to describe the prevalence, risk factors and prognostic impact of AKI during COVID-19 in the ICU.

In this study, we describe the incidence of AKI in 100 COVID-19 patients admitted to the ICU, and the link between AKI, inflammation markers and complement levels. The main results of this case series are the higher than previously reported incidence of AKI and its lack of association with IL-6, ferritin or complement factors C3 and sC5B9. The importance of AKI in COVID-19 patients has been increasingly recognized. If initial reports from China reported rates of AKI as low as 2.9–8% in severe patients [2, 15, 16], incidences from later studies ranged between 15% [17] and 44% [18] in critically ill patients. Most reports lack a clear operational AKI definition [19‐21], but even in studies that used KDIGO serum creatinine criteria, diuresis was inconsistently taken into account, and none of them reported AKI stages. Our study is the largest published to date reporting incidence of AKI specifically in critically ill patients. With 81% of patients diagnosed with AKI, the rate of AKI in our study is much higher than previously reported.

Interestingly, one of the largest cohort from the United States reported a prevalence of AKI in COVID-19 patients requiring mechanical ventilation of 80% [22], close to our finding of 90%.

One strength of our study is the rigorous use of KDIGO criteria, including urinary output. Furthermore, given that many patients will present with AKI without a reliable baseline serum creatinine on record, estimation of the latter is of tremendous importance. Several methods to estimate missing baseline creatinine can lead to variations in incidence of AKI up to 15% [6]. Using complete KDIGO definition we found a 80% incidence of AKI during the first 7 days of ICU stay and a 90% incidence in patients requiring mechanical ventilation.

We also tested the hypotheses that COVID-19-associated-AKI is linked to the elevation of cytokines induced by SARS-CoV-2 infection [23], complement dysregulation [12] induced by SARS-CoV-2 infection or mechanical ventilation settings. High levels of IL-6 have been associated with the development of severe disease [24, 25] and acute respiratory distress syndrome [8] during COVID-19 infection, but the role of inflammation markers in COVID-19-induced-AKI remains speculative [7]. The deleterious role of IL-6 has been demonstrated in different models of AKI, including ischemic AKI, nephrotoxin-induced AKI and sepsis-induced AKI [26, 27]. In our study, IL-6 and ferritin levels correlated with severity but were not independently associated with AKI. The complement system represents the first response of the host immune system. It participates in the development of AKI [28] and has also been suspected to play a role in AKI in the context of SARS-CoV-2 infection [10, 12]. Recent studies showed a strong immunohistochemical staining of complement cascade components in the lungs [29] and kidneys [30] of severe COVID-19 patients. However, even if C3 levels were associated with AKI severity in univariate analysis, the association did not persist when forced into a multivariate model, and soluble C5b9 showed no association with AKI. C3 level was only independently associated with AKI stage 2 and 3 compared to no AKI or AKI stage 1. Overall, our results do not support a role for complement dysregulation in COVID-19-induced-AKI, even though complement dysregulation may be involved in the most severe forms of AKI.

Although our study did not include extensive coagulation explorations [31], the lack of difference in fibrinogen levels in patients with AKI when compared to patients without AKI does not support the evidence of a role of hypercoagulability in COVID-19-induced-AKI, suggested by the presence of thrombi in glomerular loops described by others [32, 33]. Last, high levels of PEEP in COVID-19 patients [34, 35] have also been suggested as a potential factor for increased AKI in severe COVID-19 patients [36]. In our study, PEEP levels were not independently associated with the development of AKI, but the lack of statistical power and longitudinal data on PEEP levels does not allow for any definite conclusion.

Our study also has limitations. First, despite being the largest focusing on COVID-19-induced-AKI in the ICU, the limited number of patients can translate into a lack of statistical power. Nevertheless, rate of AKI and sample size allow confirming a high AKI incidence with a reasonable level of confidence (81%; 95% CI 72–89). In addition, the monocenter design may have limited external validity of our findings. The proportion of missing baseline serum creatinine value (67%) may also be an issue. Missing baseline creatinine value is a common problem in research focusing on AKI, and although most studies do not report the proportion of missing baseline values, rates as high as 85% are common [22]. We provide sensitivity analyses (Additional file 1: Figure S3 and Table S2) to show that back calculation of missing baseline creatinine values unlikely results in a significant bias. Last, lack of association between inflammation biomarker, IL-6 or complement component do not preclude participation of these mechanisms to AKI. They only underline that in this setting and when adjusted to severity, these factors have little influence on AKI rate. Additional studies, assessing levels of proteinuria and hematuria [22, 37], pathological findings and translational research are needed to further explore different mechanisms that may participate to AKI during severe SARS-CoV-2 infection.

In conclusion, we did not find any association between complement activation or inflammatory markers and AKI. Our study suggests a tremendously high incidence of AKI in our cohort of critically ill COVID-19 patients, along with an independent association between AKI and outcome. Because of its marked influence on outcome, AKI should be identified promptly during SARS-CoV-2 infection.