Kidney Injury in Critically Ill Patients with COVID-19 – From Pathophysiological Mechanisms to a Personalized Therapeutic Model

A series of factors such as advanced age, chronic kidney disease, cardiovascular diseases, and diabetes mellitus are associated with immune senescence and chronic inflammation that results in increased renal vulnerability to a “second-hit” injury such as the SARS-CoV-2 infection [41]. Equal consideration should be given to the chronic treatment that may limit the compensatory intrarenal hemodynamic mechanisms [31,42].

Imaging studies are essential in critically ill patients, and the use of contrast agents becomes inevitable in this already at-risk population. Their contribution in the genesis of AKI has been reviewed, with the latest evidence downgrading their role from a determinant one (i.e., CIAKI - contrast-induced acute kidney injury) to, at most, a contributory one (i.e., CAAKI - contrast-associated acute kidney injury) [43,44,45,46,47,48,49]. A recently published multi-site propensity-matched analysis led by Ehmann et al. found that contrast administration was safe and inconsequential even among patients with pre-existing AKI [50]. Consequently, the preventive strategy for contrast-induced nephropathy has been simplified and overlaps with the optimization of hemodynamics per se in critically ill patients, and, in the absence of other harmless alternatives (e.g., MRI, ultrasound), the diagnostic benefit takes precedence over the risk of AKI [43,51].

IPPV and cardiocirculatory failure conduce to an augmented sympathetic adrenergic tone and activation of the renin-angiotensin-aldosterone system (RAAS). These two systems have been proposed as the drivers of a generalized shock-induced endotheliopathy that ultimately involves a myriad of interorgan crosstalk signaling factors, including cytokines, growth factors, and damage-associated molecular patterns (DAMPs) [52]. In the pathogenesis of CARDS, AKI may arise from an intricate interplay between the kidneys and the lungs, as evidenced by renopulmonary crosstalk [53,54,55]. It is noteworthy that various other tissues, beyond the lungs, can also contribute to the release of DAMPs, which further fuel AKI development [56,57,58].

Cardiocirculatory failure in COVID-19 is propelled by multiple mechanisms that can act singularly or combined to disrupt renal inflow and outflow dynamics: 1) hypovolemia, 2) cardiac dysfunction, and 3) vasoplegia and peripheral vascular maldistribution.

In COVID-19, insensible water losses are common at admission to the intensive care unit (ICU) and directly proportional to clinical severity [15,23]. Volume depletion requires immediate therapeutic intervention to avoid a prerenal insult and to improve tissue oxygen supply. Conversely, fluid overload is equally harmful and, when associated with AKI, indicates an increased risk of mortality (adjusted RR 2.63 [1.30–5.30])[59,60].

It is universally accepted that judicious fluid administration, especially in the context of CARDS, breaks down to two macrohemodynamic principles: 1) fluid responsiveness and 2) fluid tolerance [61,62,63]. Recently, it has been demonstrated that macrohemodynamic fluid responsiveness does not guarantee microhemodynamic responsiveness, hence the importance of monitoring the microcirculatory perfusion [64]. Failure to consider this aspect can jeopardize the integrity of the glycocalyx, thereby exacerbating fluid therapy-associated complications[65,66]. Therefore, intensivists face the challenging task of finding a balance between two functional microhemodynamic extremes while simultaneously respecting these two macrohemodynamic principles: 1) limited convective flow that corresponds to an insufficient microcirculatory fluid filling associated with a low density of functional capillaries and 2) limited diffusion due to excessive fluid that leads to increased intercapillary distances and reduced density of functional capillaries [67]. The resolution of this dilemma remains a debated topic and is open for further research. The most recent solution proposes a cumulative parameter, the tissue red blood cell perfusion index (tRBCp), which incorporates both convective and diffusive components of tissue perfusion[68]. Although its quantification is automated and feasible at the bedside, additional studies are still needed to justify and describe its implementation in current practice.

Ultimately, the type of fluid administered can directly influence AKI. High chloride content and hydroxyethyl starch-based fluids were associated with an increased incidence of AKI[69,70]. Similarly, dextrans and gelatin were associated with an increased risk of bleeding, AKI, and mortality [71,72]. Therefore, current guidelines recommend balanced crystalloids over other fluids [30,73]. However, some recent trials and meta-analyses have indicated no discernible clinical advantage of balanced solutions over the utilization of 0.9% saline solutions [74,75]. Consequently, the rational approach to fluid selection aligns more with a strategy that tailors to the patient's biochemical profile and electrolyte imbalances, rather than a generic, one-size-fits-all strategy [76]. Additionally, based on recent trial data from patients with severe sepsis or septic shock, the administration of albumin has been suggested to potentially be associated with a trend towards reduced mortality [77,78].

Cardiorenal syndromes, commonly observed in COVID-19, can arise due to primary cardiac impairments such as myocarditis, in situ micro- and macro-immunothrombosis, pulmonary embolism, or the intricate interplay between the heart and lungs associated with IPPV [31,73]. Controversially, there was a proposition that the mechanical characteristics of CARDS encompass a wide spectrum that includes a subphenotype characterized by preserved lung elastance (referred to as L subphenotype) and another one characterized by increased lung elastance (referred to as H subphenotype) [79,80,81,82,83,84,85]. Similarly, Filippini et al. used latent class analysis to split CARDS subphenotypes based on their recruitment potential [86]. Additionally, compared to all-cause ARDS, Chiumello et al. showed that CARDS patients exhibited higher compliance and lung gas volume for the same oxygenation parameters, lower recruitment potential and higher blood flow redistribution [87]. The two subphenotypic extremes could very well represent different stages of disease progression. Ferrando et al. proposed that COVID-19 calls for a flexible and adaptable IPPV strategy, as the lung mechanics shift from an L phenotype found in the early stages to an H phenotype found in the late stages of COVID-19 [88]. The most recent ARDS guidelines incorporate and merge all recommendations for both non-CARDS and CARDS cases [89]. Regardless of the underlying cause, the alterations in pulmonary mechanical constants are accompanied by corresponding changes in the hemodynamic mechanisms associated with adequate renal perfusion. Overlooking this principle by employing high positive end-expiratory pressure indiscriminately was linked to a fivefold increase in the risk of COVID-AKI and elevated mortality, as evidenced in an observational study by Ottolina et al [90]. Consequently, the provision of circulatory support may necessitate an approach based on the physiological characteristics specific to the different stages of CARDS (i.e., early vs. late period) [91]. At all times, maintaining an optimal mean perfusion pressure and preventing central venous congestion is of utmost importance, as their breach has been associated with a heightened occurrence of AKI [92,93,94,95,96]. To this end, hemodynamic monitoring is crucial and must be personalized, combined, and comprehensive, incorporating both ultrasound and transpulmonary thermodilution methods, to characterize the functional cardiac reserve, and vigilantly avoid common pitfalls associated with the presence of acute cor pulmonale or low tidal ventilation [97,98] (see Table 3).

The kidney fulfills all the molecular prerequisites for direct involvement by SARS-CoV-2. This theoretical renal tropism is based on the following main aspects: 1) angiotensin-converting enzyme 2 (ACE-2), the receptor for SARS-CoV-2, is expressed in podocytes and the proximal tubule and 2) transmembrane serine protease 2, responsible for the cleavage of the Spike (S) protein of the novel coronavirus, is predominantly detectable in tubular cells and to a lesser extent in glomerular cells[99,100]. Additionally, CD147 has recently been described as a potential receptor for the S protein, and its abundance in the proximal tubule provides further evidence supporting the hypothesis of direct renal viral invasion [101,102].

Nevertheless, despite the elevated theoretical likelihood, the direct viral cytopathic effects on the kidney in COVID-19 have largely remained uncertain, given the conflicting histopathological findings. Several post-mortem studies demonstrated the presence of viral particles in tubular and podocyte cells under electron microscopy as well as SARS-CoV-2 viral RNA in glomerular cells [103,104,105]. The apparent clinical consequences of this renal tropism were an increased incidence of AKI and a higher risk of premature death [106]. On the contrary, real-time PCR analysis of renal autopsy samples did not detect SARS-CoV-2, raising doubts about the existence of COVID-19 nephropathy [107].

The underlying biological mechanisms of this potential direct viral invasion of renal tissue were explored by Dudoignon et al. who reported increased direct and indirect markers of RAAS activation in a cohort of 51 patients, particularly among those who developed AKI[108]. However, the specificity of this association with SARS-CoV-2 is difficult to assess due to the absence of: 1) a control arm consisting of AKI patients without COVID-19 and 2) a cytopathological diagnosis of direct viral aggression in patients infected with the novel coronavirus. In this regard, Puskarich et al. conducted a randomized clinical trial involving 13 hospitals to investigate the impact of SARS-CoV-2 on RAAS homeostasis. The study findings revealed that losartan, in contrast to the anticipated outcome, did not improve oxygenation after 7 days and resulted in a decreased number of vasopressor-free days compared to the placebo group[109].

Most recently, Perego et al. conducted a single-centre retrospective study that analyzed post-mortem kidney samples from critically-ill patients with COVID-19[110]. AKI was prevalent, affecting 55.8% of the population, and molecular biology analyses were performed on the renal tissues of 46% of the patients, detecting SARS-CoV-2 in only 20% of the samples, with no discernible difference between the AKI and non-AKI groups. Noteworthy, no evidence of direct viral damage, such as interstitial inflammatory infiltrate, was identified, suggesting that renal injury may be the consequence of multifactorial influences, with hemodynamic instability as a paramount contributor. In line with these data, Paranjpe et al., through the analysis of clinical and proteomic data, posited that while both acute and long-term kidney dysfunction associated with COVID-19 correspond with markers of tubular dysfunction, AKI is primarily driven by a multifaceted process encompassing hemodynamic instability and myocardial damage [111].

In general, while certain mechanisms of SARS-CoV-2 may contribute to the development of AKI, they do not constitute the primary focus of the targeted treatment strategies.

General preventive measures aim to avoid or mitigate the impact of risk factors, such as 1) contrast agents used in imaging studies, 2) antibiotics with renal excretion and metabolism, and 3) medication that may limit the compensatory intrarenal hemodynamic mechanisms such as the non-steroidal anti-inflammatory drugs, angiotensin-converting enzyme inhibitors, and angiotensin-receptor blockers (see Table 4).

Metabolic priorities in AKI coincide with those of critically ill patients, focusing primarily on two components: 1) intensive glycemic control and 2) nutritional support, with particular attention to protein intake.

The principles of cardiopulmonary support have been discussed in the subsection dedicated to non-specific pathophysiological mechanisms contributing to COVID-AKI development. These principles are universally applicable, with the key message that mean arterial pressure (MAP) and central venous pressure (CVP) are equally important in ensuring renal perfusion. An increase in CVP is associated with decreased glomerular perfusion, AKI, and death [112,113]. Legrand et al. reported a direct proportional relationship between CVP and the prevalence of AKI. Additionally, the same authors demonstrated no association between AKI (new or persistent) and classic macrohemodynamic parameters such as MAP, cardiac output (CO), and mixed venous oxygen saturation [114]. Another study revealed the ability of venous congestion to stratify the risk of AKI development and confirmed the lack of predictive value of CO in this regard [115]. In recent studies, there has been a growing focus on both poles of renal perfusion, namely MAP (i.e., renal preload) and CVP (i.e., renal afterload), demonstrating that the time-weighted average of mean perfusion pressure (MPP, calculated as MAP - CVP) is associated with an elevated risk of AKI and renal adverse events [116]. These findings underscore the importance of considering cardiocirculatory optimization strategies, as represented in Table 4 and adapted from the insights provided in references [11,30,31].

Ultrasound imaging has emerged as an invaluable tool in this context, offering a swift and non-invasive means of assessing diverse cardiocirculatory and pulmonary parameters [117]. Consequently, amid the COVID-19 pandemic, we devised an algorithm that relies on echocardiographic evaluation, specifically designed to promptly diagnose and address hemodynamic instability and shock in affected patients (see Figure 2).

Fig. 2.

The COVID-19 pandemic has presented unprecedented challenges for healthcare teams and patients alike. When there is improved access to RRT, the implementation of a coordinated local response by medical staff can significantly reduce the mortality rate. Efficient and fair allocation of limited medical resources can be achieved through various measures, including optimizing the modality and indications for RRT, adopting appropriate anticoagulation strategies, and carefully determining the dosage of RRT [31].

Assuming that human, technical, and material resources are not limiting factors, the guidelines for implementing RRT in COVID-19 align closely with the initiation of RRT in critically ill patients [16,17]. Nonetheless, it should be acknowledged that anticoagulation treatment protocols may require a more flexible implementation in individuals infected with SARS-CoV-2. Moreover, given the diminished fluid tolerance intrinsic to COVID-AKI, the process of fluid removal mandates an individualized approach, congruent with each patient's unique fluid tolerance [118]. Just as in resuscitation, the adoption of a functional hemodynamic algorithm during fluid removal could be superior to a generic approach [119,120,121,122,123,124,125]. This tailored strategy aims to prevent both under-resuscitation scenarios, characterized by residual vascular and extravascular congestion, and over-resuscitation scenarios, which may result in low cardiac output and hypotension (see Table 5).