Sepsis-associated acute kidney injury: is COVID-19 different?

Editor’s Note.

This is one of several articles we think you will find of interest that are part of our special issue of Kidney International addressing the challenges of dialysis and transplantation during the COVID-19 pandemic. Please also find additional material in our commentaries and letters to the editor sections. We hope these insights will help you in the daily care of your own patients.

When faced with a new disease, clinicians understandably resort to what they know. For example, we know severe viral pneumonia secondary to influenza, so when we began seeing coronavirus disease 2019 (COVID-19), our first impression was that it was a similar disease, so we thought we knew how to manage it. However, it soon became clear that COVID-19 was strikingly different. First, there was the high transmission rate—including spread within hospitals and health care workers¹—but this could simply have been due to a new disease in a naive population. After all, influenza is also quite contagious. Second, the numbers of patients with severe disease also seemed significantly greater, but this too could simply have been a function of the overall number of infected patients.

We now know that (i) at least 80% of patients with COVID-19 have mild or even asymptomatic disease and (ii) the majority of patients with COVID-19 who develop critical illness are older than 70 years, have an underlying disease, or both. Still, we are seeing cases of severe COVID-19 across all age groups and occasionally in the previously healthy. Overall case-fatality rates from COVID-19 have ranged from 1% to 10%, making it much worse than seasonal influenza but not nearly as deadly as the 2 other novel Betacoronavirus infections from the past: severe acute respiratory syndrome and Middle East respiratory syndrome.

However, other distinctions between COVID-19 and diseases with which we are more familiar also began to emerge. COVID-19 can cause severe hypoxemia in patients who do not appear to be that “sick” and, although fulfilling the Berlin criteria for acute respiratory disease syndrome, their illness often does not behave in a way that we expected from our “usual” practice.² Anecdotal reports from emergency responders in hard-hit cities featured dramatic themes, such as progression from “talking to unconscious” in minutes and intensive care physicians finding themselves with patients who were profoundly hypoxemic without otherwise altered vital signs—features that are not part of severe bacterial pneumonia and equally unusual with other pathogens. The lack of shock (as measured by lower blood pressure or arterial lactate levels) in the majority of patients may also have contributed to the sense that COVID-19 is not like other severe infections. Then came the coagulation abnormalities. Many patients with severe COVID-19 develop hypercoagulability that manifests as increased venous thrombosis, elevated plasma d-dimer concentrations, and clotting of extracorporeal circuits. The syndrome is distinct from disseminated intravascular coagulation or thrombocytopenic purpura (i.e., it is not characterized by changes in the international normalized ratio or ADAMTS-13), and the platelet count is usually normal.

Finally, early reports of severe, even fatal, COVID-19 stated that kidney injury was absent.³ This was surprising because extrapulmonary manifestations of COVID-19 appeared to be rather broad with involvement of the heart, brain, and gastrointestinal tract.^{1 , 4} Furthermore, these early reports showed evidence of systemic inflammation reminiscent of sepsis and the multiorgan failure followed this familiar pattern. Although the news media embraced the term “cytokine storm,” reports from China suggested a “drizzle” at best.⁴ However, cytokine assays are notorious for not being standardized and results from one assay system are not comparable to another. Conversely, standardized laboratory tests for C-reactive protein and ferritin clearly showed severe inflammation in the sickest patients.⁴ Thus, a hyperinflammatory syndrome—like the hypercoagulation—appears to be very much a part of severe COVID-19. This finding made the absence of acute kidney injury (AKI) all the more remarkable—if only it were true. By late March it became apparent that AKI was very much a feature of severe COVID-19, and by mid-April news media in the United States were calling the associated renal failure a “second surge.”

Much of the contradictory information on AKI in COVID-19 came from the rapid, and often unreliable, reporting as investigators raced to be the first to report their findings and journals competed to be the first to publish. Amidst this frenzy, much was overlooked and “secondary outcomes” such as AKI were not rigorously tracked. As the dust began to settle, more exacting reports specifically on AKI emerged and the findings were at the same time sobering and predictable. Overall AKI occurs in approximately 25% to 35% of patients with COVID-19 who are admitted to hospitals; the rate is nearly twice that for patients admitted to intensive care units.⁵ These results are completely consistent with the published literature on bacterial pneumonia⁶ and as such we might reasonably ask 2 related questions. (i) How different is AKI in the setting of COVID-19 than in other forms of sepsis? And (ii) how applicable is the evidence from non-COVID sepsis–associated AKI to AKI in patients with COVID-19?

We believe it is important to ask both questions, because if we ignore the first, the only way to answer the second is to test each and every recommendation specifically in patients with COVID-19. For example, if we recommend caution in the use of vancomycin for patients with bacterial sepsis, can we recommend the same caution in treating patients with COVID-19? We believe that we can because vancomycin is frequently prescribed and because there is no evidence to suggest that patients with COVID-19 are less susceptible to its toxicity.

Importantly, we use the term “COVID-19 sepsis” because many patients with severe COVID-19 meet the diagnostic criteria for sepsis and many of them develop AKI in the same way as do patients with bacterial sepsis. However, there are important differences between sepsis-associated AKI in COVID-19 compared with bacterial infection (Table 1 ) that may directly impact how we apply evidence across the two syndromes. Nevertheless, the answer to the first question would appear to be “not very different.” So it follows that most—if not all—our recommendations for management of sepsis-associated AKI should also apply to COVID-19–associated AKI.^{7 , 8} Importantly, though, all of the uncertainty regarding management of sepsis-associated AKI—like AKI in general—is also applicable. For example, we do not know precisely when to initiate extracorporeal kidney support or, even more controversially, if and when to use other forms of blood purification therapy. We also do not know exactly how to manage various trade-offs between potentially nephrotoxic therapies (e.g., drugs, procedures) and the risks of not providing these therapies. In bacterial sepsis, these generally relate to nephrotoxic antibiotics and contrast agents for imaging. In COVID-19, we might also add the use of diuretics to reduce lung edema but possibly potentiating other nephrotoxic effects of drugs or indeed the disease itself. Finally, there are specific questions for COVID-19. (i) What effect, if any, will low-dose corticosteroids and anticoagulation have on AKI? And (ii) will any antivirals offer any protection from direct viral tropism in the kidney? In the meantime, though, our standard approach to sepsis-associated AKI—which includes careful fluid and hemodynamic management, avoidance (or cautious use) of potentially nephrotoxic agents, and personalized application of extracorporeal kidney support⁹—should be followed. We should not forget what we do know about AKI, this common complication, even as it arises in a disease we know relatively little about.

Table 1.

Comparison of bacterial and COVID-19 sepsis

Etiology/mechanism	Bacterial sepsis	COVID-19 sepsis	Comments
Systemic inflammation affecting multiple organs	+++	++	Corticosteroids have been shown to improve survival in COVID-19, whereas the effect in bacterial sepsis is only on reversal of shock.
DAMPs released from injured tissue	+++	+++	Injured lung and also remote organs (e.g., muscle) may contribute DAMPs.
PAMPs released from microorganisms	+++	?	Extensive involvement of the GI tract in some patients raises concern for translocation of bacteria/bacterial products.
TMA	+	++	The TMA in COVID-19 appears to be unique from DIC, TTP, and HUS.
Nephrotoxic drugs	+++	+++	Different agents are used in the 2 syndromes.
Direct viral infection of tubular epithelial cells	n/a	±	The full extent of this pathobiologic mechanism is still unknown.
Cardiac dysfunction and/or reduced preload	+	++	COVID-19 may directly affect the heart and high PEEP ± volume depletion may impair venous return.
Macrophage activation syndrome	+	+	Generally more common in viral infection and in children.

+, weak evidence/involvement; ++, moderate evidence/involvement; +++, strong evidence/involvement; ±, evidence/involvement is equivocal; ?, unknown; COVID-19, coronavirus disease 2019; DAMPs, danger-associated molecular patterns; DIC, disseminated intravascular coagulation; GI, gastrointestinal; HUS, hemolytic uremic syndrome; n/a, not available; PAMPs, pathogen-associated molecular patterns; PEEP, positive end-expiratory pressure; TMA, thrombotic microangiopathy; TTP, thrombocytopenic purpura.