Coronavirus disease (COVID-19) has become one of the worst pandemics in recent human history, with dramatic short- and long-term clinical outcomes, more than 250 million people infected, and more than 5 million deaths around the world.1 Over time, the international scientific community has made an unprecedented effort in modern history to understand better the nature and pathophysiology of this disease.

The presentation of COVID-19 varies considerably. Although it manifests primarily as a respiratory disease, it exacerbates and progresses in some individuals due to abnormal immune/inflammatory responses with generalised endothelial damage and alterations in blood coagulation, generating a much more serious spectrum of the disease.2 A substantial proportion of patients with severe COVID-19 show signs of kidney damage. Clinically, renal involvement ranges from mild/moderate proteinuria and haematuria up to acute kidney injury (AKI), with the need for renal replacement therapy (RRT) in some cases. The pathophysiological mechanisms of kidney damage in patients with COVID-19 remain unclear, but they are known to be multifactorial. Current knowledge implies direct effects dependent on SARS-CoV-2 on renal cells (tubular epithelial cells and podocytes) and indirect effects through the systemic impact of the viral infection secondary to critical lung disease and secondary to the management of the disease itself.2,3 Kidney involvement is a determinant factor in higher in-hospital mortality in all published series.4–7

The reported incidence of AKI is highly variable, ranging between 0.5 and 56% depending on the series.8 The population studied, and the region of the world from which the series originates directly impact these differences between the published data. The initial cohorts of Asian origin reported an incidence lower than 7%9,10 while Western series indicate an incidence between 11 and 46%,4–6 reaching up to 76–80% in patients admitted to the ICU.11–13 The factors related to the development of kidney injury during COVID-19 also differ widely between publications, having been related to demographic characteristics (age over 60 years, male gender or African descent), comorbid situations (arterial hypertension (HBP), diabetes mellitus (DM), cardiovascular disease (CVD), obesity, chronic obstructive pulmonary disease (COPD) or chronic kidney disease (CKD), as well as the severity of the patient's condition (haemodynamic instability or need for mechanical ventilation).8,11–15

The mainstay of treatment continues to be the prevention of kidney damage and, if necessary, RRT. In this aspect, the literature reports large fluctuations in the frequency of the need for RRT, ranging from 0.8% to 31%.8 At the time, many doubts arose about the approach to RRT in these patients: whether it should be approached in the same way as in other acute processes or differently, whether any modality of RRT could be superior for the elimination of cytokines, whether adsorptive therapy, repeatedly postulated in the treatment of sepsis, could provide advantages, or whether low-efficiency continuous therapies could be more efficient. Over the course of the pandemic, numerous studies aimed to address concerns that have arisen in the nephrology community in different aspects. However, drawing meaningful conclusions from these studies has often been difficult due to barriers posed by data limitations and the use of uncertain methodology that carries numerous biases.

Since the beginning of the pandemic, the Spanish Society of Nephrology (SEN) was very sensitive to the situation, actively participating in different initiatives that helped in clinical decision-making. The SEN has repeatedly highlighted that kidney injury can be a severe complication of COVID-19, highlighting the importance of providing more in-depth knowledge to this entity, as well as establishing the necessary interventions to improve the prognosis of patients with COVID-19. In this context, the SEN COVID-AKI Registry emerged, where clinical data of patients with AKI secondary to COVID-19 were collected from May 2020 to November 2021, including variables related to SARS-Cov-2 infection, kidney involvement and the need for RRT in our country. A committee of SEN experts chose a minimum set of variables that would give us a perspective of the pandemic's impact and the need for RRT in acute patients throughout our country. Given the large discrepancies found in the available studies, the exploitation of the SEN COVID-AKI Registry represents an opportunity to characterise patients with this injury in our environment within the context of the great heterogeneity of the published cohorts, as well as to know how these patients were managed in our environment, identify the profile of patients who have required RRT, the replacement treatment modalities used and the results obtained in terms of recovery of kidney function and mortality.

This is a retrospective observational cohort study in which all data collected in the Spanish COVID-AKI Registry from May 2020 to November 2021 was included. The COVID-AKI Registry presented an online structure with access through a website (SEN website: www.senefro.org) where the members of this society, after identifying themselves through a previously registered user, were able to add hospitalised patients with AKI and SARS-CoV-2 infection confirmed by real-time polymerase chain reaction.

The study complied with the guidelines established by the Declaration of Helsinki and the Declaration of Istanbul. The research was approved by the Regional Institutional Ethics Committee of Andalusia. Obtaining informed consent was not necessary given the retrospective nature of the study.

Data were collected from 730 patients with documented SARS-CoV-2 infection, who were hospitalised as a consequence of COVID-19 and who developed an episode of AKI during admission. The patients came from 30 hospitals distributed throughout Spain. In Appendix B, Annexes 1 and 2 detail all the nephrology units that reported cases, as well as the geographic coverage by the autonomous community.

The average age of these patients was 69 years, 71.9% were men, 70.1% had hypertension, 32.9% had diabetes, 33.3% had a history of CVD, 11.8% had a history of oncology, 17.3% COPD, 13.4% immunosuppressed, 21.1% obese, and 12.4% were smokers. 23.9% of the patients (171 patients) had some degree of CKD prior to admission (Table 1). Most patients with CKD had stage G3 (135 patients), with advanced stages of disease being less common (G4: 30 patients; G5: 6 patients).

In 91.2% of the cases, the form of presentation of the infection was typical of respiratory semiology, and up to 94.6% of the reported patients developed pneumonia. Three hundred eighty-four patients (54.2%) required ventilatory support, 312 (44.1%) were admitted to the ICU and 15 required ECMO. The median time from the onset of COVID symptoms until the patient required admission to the ICU was 8 days (IQR: 5–12).

A 44.6% of the reported cases developed KDIGO III, followed in frequency by KDIGO I, which was reported in 37.1% of the cases. The remaining 18.3% presented grade II. Renal biopsy was performed in only 4 patients, so the data are not included in this study. At the time of admission, there is only data regarding urinary parameters in approximately half of the cases (373 for albuminuria and 374 for haematuria), with the presence of albuminuria positive in 30.5% of the cases collected (113 patients) and haematuria in 24.3% (90 patients).

The analytical values 7 days after the onset of symptoms are shown in Table 1, presenting very high mean levels, above the normal range in all the parameters analysed (CPR; D-dimers, IL6, ferritin), as well as lymphopaenia.

We present a descriptive study of the clinical profile of patients who developed AKI during hospitalisation for COVID-19 in our country based on the data collected in the SEN COVID-AKI Registry (May 2020–November 2021). The variables collected did not allow us to calculate the incidence that this entity has presented, but they allow us to adequately characterise this population. Overall, our patients were mostly men, with a high average age (70 years) and a considerable percentage of them had hypertension. (69,8%); DM, CKD, CVD and obesity were frequently observed comorbidities. All these comorbid factors have been previously identified as predisposing factors, both to greater severity of the infection and to the appearance of AKI among patients requiring hospital admission.4,8,11–15 It is striking that the prevalence of CKD prior to admission was 23.9%, much lower than other series collected in our environment. Portolés et al.6 reported a prevalence of CKD in 43.5% of patients admitted for COVID-19 who developed AKI during the first wave of the pandemic in a hospital in Madrid. In our work, data were collected over a long period of time, and this difference in prevalence may reflect the moderation of the pandemic's effect over time among the most fragile patients, on whom the first wave had a major impact.

Previous publications report that the factors with the greatest impact on the risk of acute kidney injury in the course of COVID-19 are related to the severity of the patient's condition, the need to use catecholamines, septic shock and/or the need to use artificial ventilation due to respiratory failure.4,8,13 Our cohort coincided in these terms, with kidney injury appearing in patients with severe COVID and a slow progression of the infectious disease: almost all had pneumonia, half of them required ventilatory support, and the mean numbers of inflammatory markers were very high globally.

A 32.9% of the analysed cohort required RRT. Hassanein et al.,8 in a review of the literature, reports great variation in the frequency of the need for renal replacement therapy in these patients. Without a doubt, differences in the incidence of AKI, clinical management of the patient in each hospital, as well as the saturation of health systems and the availability of technical and human resources, have been determining factors in these oscillations. Several meta-analyses that investigate the overall frequency of need for RRT in this type of patients estimate it varies between 3.6 and 4.3%,18,19 reaching figures of 20–40% in series of patients admitted to the ICU.20 We must consider the selection bias that our Registry presents, due to the fact that the data come from patients evaluated by the nephrology services. It is reasonable to think that we have followed the most serious cases, hence our results are more similar to those obtained in the series of serious patients admitted to intensive care.

The presence of haematuria and proteinuria in hospitalised COVID patients has been identified as predictors of kidney injury and mortality during admission.6,11 In our data these urinary parameters were only recorded in approximately half of the cases. Even so, the detection of haematuria and proteinuria was significantly higher in patients requiring RRT. These urinary alterations probably reflect the development of severe kidney injury. In the literature, the most frequently reported histological lesion was acute tubular necrosis,21 although interstitial nephropathy, thrombotic microangiopathy, collapsing glomerulosclerosis or even direct invasion by SARS-CoV-2 of kidney tissue have also been described in the context of a large viral load2,3,20; in all of them, it is possible to find associated haematuria and proteinuria.

An interesting aspect of this work is that the clinical characteristics of a significant number of patients who developed AKI without needing RRT were collected. In our environment, where access to replacement therapy was practically guaranteed, the fact that these patients were an average 6 years older than those who received RRT is especially striking. Furthermore, milder AKI predominated in this group (KDIGO I 59.8%) and, both the average time of appearance of AKI and recovery was shorter (5 and 4 days, respectively). The need for ICU admission, ventilatory support, and COVID treatment was lower in this group, indicating, overall, a lower severity of the course of the infection. This type of AKI, which appeared early in the course of the infection and resolved without the need for RRT, was the most frequently recorded in our series, constituting 2 out of every 3 reported cases. There are series of patients in our environment that identify the prerenal cause as predominant in patients with COVID-19 who developed AKI.22,23 Although the Registry did not collect the clinical impression of nephrologists on the origin of AKI, the rapid appearance and recovery of the condition leads us to think that the prerenal cause probably justifies renal dysfunction in an elderly, fragile population, with high comorbidity, with greater risk of kidney dysfunction during the first days of infection. We only found a similar description in the literature by Peng et al.24 They described 2 patterns of AKI in a total of 285 cases with COVID: one earlier and another later which occurs when there is already dysfunction of other organs. This later kidney injury was accompanied by significantly higher levels of systemic inflammatory parameters, a higher percentage of proteinuria and glycosuria, lower mean arterial pressure on admission, higher rates of lymphopenia, thrombocytopenia and coagulation disorders on admission, and higher mortality. All of this would coincide with our data: later AKI occurs in the context of greater severity of the infection, reaches greater severity and is more likely to require RRT. The time elapsed from the onset of symptoms to the appearance of AKI was identified in our analysis as an independent predictor of the need for RRT, such that each day of delay in the onset increases the risk by 13%, which would confirm this proposed hypothesis. And, along the same lines, age acts as a protective factor, which would be justified by the fact that older patients are those who have developed milder AKI, as it was previously explained.

Other factors independently related to the need for RRT during hospitalisation for COVID-19 were, in addition to the serum creatinine level, the need for ventilatory support (OR: 20.20) and smoking habit (OR: 3.41). Hypoxaemia has already been related in other studies to RRT in AKI due to COVID-19.25 In fact, hypoxaemia can compromise cardiac output and renal perfusion, as well as stimulate signalling pathways that determine greater thrombogenicity. In turn, aggressive ventilation strategies can lead to neurohormonal alterations and the release of systemic inflammatory mediators, which impact intrarenal blood flow and further deteriorate renal perfusion, inducing more severe kidney injury through all these mechanisms.

Another fact that the data reveal is the absence of clear indications and multiple questions regarding which modality of RRT could be most effective. eEach centre has opted for different treatment modalities depending fundamentally on the experience and availability of resources in each. In this manner, CRRT, intermittent haemodialysis, daily haemodialysis, expanded haemodialysis and haemodiafiltration have been used, with the first two being the most widely represented, in many cases being used in combination. It has been suggested in the literature, although there are no randomised controlled studies, that the use of adsorption membranes, special membranes or convective dialysis modalities could be useful to eliminate circulating inflammatory molecules, as well as SARS-CoV-2 molecules in selected patients.26 All of them are represented in this Registry to a greater or lesser extent, without having found a clear benefit in favour of any of them in terms of results. To date, the data available on the results of RRT in these patients remain limited. The largest series, which included 3,099 critically-ill COVID-19 patients admitted to the ICU in 67 hospitals in the United States, showed that the most commonly used initial modality was CRRT (52.4%), followed by intermittent haemodialysis (30,0%) with wide interhospital variability both in the frequency of use and in the modality used.25

All published works agree that the coexistence of AKI during hospitalisation for COVID-19 behaves as an independent predictor of mortality.14,18,19 Mortality in our series was also high, 38.6% in the entire cohort, reaching 59.1% in patients who required RRT. Factors related to greater severity of infection were those that had the greatest impact on mortality; thus, the development of pneumonia and the need for ventilatory support tripled the risk of death. The development of more severe AKI in the context of infection requiring RRT doubled the risk of mortality. On the other hand, the patient's age and the presence of CKD prior to admission were host-dependent factors, which were associated with a higher risk of mortality in our series. The age of the patients has been a factor clearly related to the worse outcome and higher mortality of COVID-19 in the literature.5,26,27 Frailty and age-related comorbidity make these patients especially vulnerable to COVID-19 infection.

Unlike other series, we have not found that the presence of CKD is independently related to greater severity of AKI with higher RRT requirements than the rest of the population. However, a direct impact of CKD on mortality was detected (OR: 2.21). CKD has been identified by the ERACODA Group as one of the 4 comorbidities associated with the highest risk of mortality from COVID-19. Specifically, the risk associated with CKD stages 4 and 5 is higher than the risk associated with diabetes mellitus or chronic heart disease.28 The publication Global Burden of Disease identified that, worldwide, CKD has been the most common risk factor for severe COVID-19.29 There is likely a bidirectional relationship between COVID-19 and CKD: CKD increases the risk of severe COVID-19, and the greater severity of COVID-19 leads to greater risk of acute renal dysfunction, in turn leading to greater mortality.

Much has been discussed in recent months about the role played by ACEIs and ARBs in the course of COVID-19 infection. Although most studies have attributed a neutral role to them, in our analysis, chronic treatment with ARB was identified as a protective factor in the mortality of those patients who developed AKI. Given that ARBs act by direct blockade of AT1R and do not competitively inhibit ACE, the use of ARBs could be beneficial in this way. Furthermore, ARBs, by blocking AT1R, can also increase the availability of ACE2 bound to the cell membrane, inhibiting its internalisation, thus also contributing to a lower aggressiveness of SARS-Cov2.30,31

Like any retrospective registry-based study, our study has limitations. Firstly, the entire population that suffered AKI in our country was not included, presenting the bias that the patients registered were those treated by nephrologists, so the cohort may represent a group of patients with a more torpid evolution. Data collection was limited and follow-up data are lacking. Finally, the evolution of the pandemic and the application of vaccines have altered its virulence and transmission, which influences the severity and mortality of the disease, so the results would probably not be applicable today. As strengths, it is worth highlighting that this analysis shows a fairly global view of the evolution of patients admitted for COVID-19 who developed AKI in the early phases of the pandemic in our country, not previously available in subjects with different evolution of both the infection as from the AKI. On the other hand, although the development of AKI has been an important complication of COVID-19, data on RRT in these patients are limited. Fortunately, our health system was not overwhelmed in most cases, and we can evaluate the impact without major resource limitations. We can also review the indication of RRT in a large cohort, not only limited to critically ill patients, but also comparing the different modalities used and the results obtained, an aspect that has been scarcely discussed in the literature to date.

In conclusion, patients who presented AKI during hospitalisation for COVID-19 had a high average age, with associated comorbid conditions, and presented with severe infection. We differentiate 2 clinical patterns of kidney injury in these patients: the first one occurs early in older patients, with a greater burden of comorbidity, and resolves in a few days without the need for RRT; and the second one, is more severe and with greater requirements for RRT, appeared later in the course of the disease and was related to greater severity of the disease. Early and more mild presentation was the most frequent, constituting 2 of every 3 reported cases. It is also reflected in this analysis that different modalities of RRT were used without a clear benefit of any of them over the others. The severity of the infection, age, and presence of CKD prior to admission were determining factors in the mortality of these patients, identifying chronic treatment with ARB as a protective factor for mortality.

The authors declare that they have no conflicts of interest.