In this study we retrospectively analyzed the renal outcome and prognostic value of a large cohort of hospitalized patients with a respiratory viral infection (COVID-19, Influenza, and RSV). We found that the incidence of AKI upon admission and within 7 days was the lowest among patients with COVID-19, on the contrary, the incidences of severe AKI, death or ventilation were highest among patients with COVID-19. Risk factors for the development of any AKI or stage 3 AKI were similar among the viruses, however only in patients with COVID-19 were desaturation and high ferritin levels independent risk factors for stage 3 AKI. Among all viruses, AKI and especially severe AKI were strong independent risk factors for adverse outcomes.

The reported incidence of AKI in hospitalized COVID-19 patients has shifted geographically and temporally since the outbreak of the pandemic (7–17% in China vs. 56% in New York [5, 39, 40]) (40% in March 2020 compared with 25% in November 2020 in the USA) [37]. Most recent reviews cite an incidence of 20% in European cohorts [41] which is substantially higher than the incidence we observed which was 11%. However, this may be related to the different patient population and perhaps to the fact that the early implementation of lockdowns and an efficient vaccination program in Israel prevented hospital insufficiency [42, 43] and enabled the admission of patients with COVID-19 with milder disease. Moreover, when reviewing reports from the U.K. and the USA, AKI rates were much lower among Caucasian patients, which make up the vast majority of individuals in Israel [4, 6, 37, 44]. The higher rates of AKI observed in patients with RSV reflect their baseline frailty rather than the severity of their disease, as is evident from the weaker association between AKI and general outcome compared to other viruses.

In COVID-19, the association between AKI during hospitalization and adverse outcome is well established. In most studies, however, the timing of AKI is generally defined within 7 days of admission, a relatively broad time frame. We examined the prognostic value of AKI at several clinically relevant time points: upon admission, after 48 h, and within 7 days. Across all viruses, we found that the development of AKI after 48 h is ominous, suggesting a four- to sixfold risk for ventilation or death. In fact, any rise in creatinine after 48 h is associated with an increased risk for adverse outcome in influenza and COVID-19. Thus, acute renal injury may express a more systemic and severe disease as implied by its association with a more adverse outcome. Additionally, we focused on a subset of patients, often overlooked in the day-to-day practice, those who develop AKI within normal range creatinine. These patients also had a three- and fourfold risk of adverse outcome for COVID-19 and Influenza, respectively. Special attention to renal function and its prognostic value at these well-defined time points can be useful for patient management and resource allocation (i.e., ventilation equipment), which may be crucial in the resource-limited environments encountered with the cumulative interference of a COVID-19 surge and flu and RSV season [45].

Several systemic causes (sepsis, hypoxia and medications) have been suggested to underlie the renal injury observed in COVID-19 [6, 21, 32, 46]. Nevertheless, a very large body of data suggests a direct injury to the kidneys by COVID-19 [20, 21, 47,48,49], including kidney tropism driven by cell entrance via the Angiotensin Converting Enzyme 2 receptor which is abundant in the kidney in general, and in the proximal tubule specifically [50]. Some of these findings, however, were partial or are grounds for debate [41]. In our study, patients with COVID-19 did not have a higher incidence of AKI, nor was COVID-19 an independent risk factor for AKI when compared with RSV or influenza in a logistic regression model. Our study was not able to determine whether there was a higher incidence of tubular injury in patients affected with COVID-19 compared to those with the other viruses, though rates of hematuria and proteinuria, which are usually associated with glomerular injury, were higher or equal in the COVID-19 patient group, and this is consistent with former reports [51,52,53]. Of note, COVID-19-related proteinuria has been shown to consist of low molecular weight “tubular proteins” α1 microglobulin which may account for a portion of the observed proteinuria [54].

Beyond the postulated tropism that ACE2 receptor expression underlies, COVID-19 is relatively unique with regard to the role the aberrant immune response plays in its pathogenesis [23, 55], evident in part by the unique prognostic value of NLR and ferritin [16, 36, 56, 57].

In the current study ferritin was a significant risk factor for stage 3 AKI in a logistic regression model (adjusted for age, sex, and comorbidities) only in the COVID-19 group. Indeed inflammatory processes causing tissue damage were observed specifically in the kidney tubules, which is consistent with both the histopathological findings of acute tubular injury and the predisposition of proximal tubular cells as targets of COVID-19 [49].

The proximal tubule reabsorbs nutrients and electrolytes including phosphate. Tubular dysfunction leading to phosphate wasting was suggested to underlie the association between hypophosphatemia and AKI in COVID-19 [58]. In our study, however, hyperphosphatemia upon admission was a very strong independent risk factor for AKI and stage 3 AKI across all viruses, which is in accordance with earlier reports [59]. This may reflect baseline kidney dysfunction with reduced GFR.

Two earlier studies compared the kidney injury in patients with COVID-19 to the kidney injury in patients with influenza. Bhasin et al. found a similar incidence of AKI in both groups (33%) in agreement with our results. In their study, however, COVID-19 was an independent risk factor for severe AKI compared to influenza, which was not the case in our study, and this may be due to the fact that the COVID-19 cohort and Influenza cohort had significant population differences which were not a covariate in their model [60].

Birkelo. et al. conducted a very large study that compared COVID-19 to Influenza-related AKI and found higher rates and more severe AKI in the COVID-19 group (41% vs. 29% patients with AKI and 26% vs. 6% for stage 3 AKI). Our results, as stated above, depict a different situation in which patients with COVID-19 have lower rates of AKI and equal rates of severe AKI. This disparity could stem, as discussed earlier, from population differences and baseline disease severity (30% mortality rates in Birkelo’s study compared to 18% in our study) [61].

Beyond the comparison of COVID-19 and Influenza our study includes a comparison with RSV, a common respiratory virus which presents clinically in the same manner as COVID-19 and influenza. The RSV cohort was significantly older and had more comorbidities than the two other groups. However, also after adjusting for these parameters, in this cohort, acute kidney injury had an attenuated ability to predict adverse outcome. This may reflect the higher occurrence of AKI in patients with RSV in this study. Interestingly, CRP levels were not associated with higher rates of AKI only in the RSV cohort, implying perhaps a weaker association between inflammatory and systemic processes and AKI. Of note, rat models of RSV-infection-induced kidney injury showed evidence of both direct kidney injury by RSV on the one hand, and immune dysregulation on the other [34].

Other strengths of our study are (1) The analysis of the prognostic value of kidney injury at specific, clinically relevant time points (upon admission, within 48 h, etc.). (2) The review of many laboratory inflammatory and renal parameters and their prognostic role in kidney injury.

Nevertheless, our study has several limitations, primarily the fact that it is retrospective in nature. The temporal formation of the three cohorts is asymmetrical; while the COVID-19 cohort was formed within a period of a year and a half, the RSV and influenza cohorts had accumulated for over 10 years. Moreover, during the early phase of the COVID-19 pandemic, even patients with mild symptoms were referred to the ER, while hospitalized patients were screened for COVID-19, which may have caused selection bias.

To conclude, our study compared the renal risk factors, the outcomes and the prognostic factors in patients affected by three respiratory viruses. To our knowledge, this is the first large scale study reporting on renal injury in hospitalized RSV patients. We found that, despite the growing evidence for COVID-19 tropism and direct kidney injury, the incidence of acute kidney injury is no higher than that observed with other viruses. COVID-19 is unique, however, in the prognostic value of inflammatory markers for predicting renal injury, which may pertain to the specific role of the aberrant immune response in kidney damage. We were further able to define the prognostic value of AKI at several time points on the overall outcome. In the near future, when influenza, COVID-19 and RSV may circulate simultaneously, our findings could provide a virus-specific evaluation allowing clinicians to estimate the expected renal and overall outcomes.