Acute kidney injury incidence, recovery, and long-term kidney outcomes among hospitalized patients with COVID-19 and influenza

Acute kidney injury incidence, recovery, and long-term kidney outcomes among hospitalized patients with COVID-19 and influenza

Retrospective cohort study of patients with COVID-19 hospitalized between March–May 2020 and historical controls hospitalized with influenza A or B between January 2017 and December 2019 within a large healthcare system. Cox proportional hazards models were used to compare the risk of AKI during hospitalization. Secondary outcomes included AKI recovery, mortality, new-onset chronic kidney disease (CKD) and ≥ 25% estimated glomerular filtration rate (eGFR) decline.

A total of 2425 patients were included; 1091 (45%) had COVID-19 and 1334 (55%) had influenza. Overall AKI rate was 23% and 13% in patients with COVID-19 and influenza, respectively. Compared to influenza, hospitalized patients with COVID-19 had an increased risk of developing AKI (adjusted hazard ratio, aHR 1.58, 95% CI 1.29–1.94). Patients with AKI were more likely to die during hospital when infected with COVID-19 vs. influenza (aHR 3.55, 95% CI 2.11–5.97). Among patients surviving to hospital discharge, the rate of AKI recovery was lower in patients with COVID-19 (aHR 0.47, 95% CI 0.36–0.62); however, among patients followed for ≥ 90 days, new-onset CKD (aHR 1.24, 95% CI 0.86–1.78) and ≥ 25% eGFR decline at last follow-up (aHR 1.36, 95% CI 0.97–1.90) were not significantly different between the cohorts.

Xie Y, Bowe B, Maddukuri G, Al-Aly Z. Comparative evaluation of clinical manifestations and risk of death in patients admitted to hospital with covid-19 and seasonal influenza: cohort study. BMJ. 2020;371:m4677. Epub 2020/12/17. doi: 10.1136/bmj.m4677. PubMed PMID: 33323357; PMCID: PMC7735416 at www.icmje.org/coi_disclosure.pdf and declare: support from the United States Department of Veterans Affairs, the Institute for Public Health at Washington University in Saint Louis, MO, USA, the American Society of Nephrology, and KidneyCure; no financial relationships with any organizations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.

There have been nearly 150 million cases of coronavirus disease 2019 (COVID-19) worldwide as of the end of April 2021 (). COVID-19 is a respiratory illness that ranges in severity from mild upper respiratory symptoms to respiratory failure and death. Severe COVID-19 can cause multi-organ injury (), and acute kidney injury (AKI) has emerged as a common complication of COVID-19 among hospitalized patients, affecting between 17 – 37% () and AKI is associated with a 9-fold increased risk of death during hospital stay ().

In the earliest autopsy series evaluating kidney histopathology in 26 patients who died of COVID-19 in China, the most common finding was prominent acute tubular necrosis (ATN) with a considerable number of patients also displaying endothelial injury and obstruction of peritubular and glomerular capillary lumens with erythrocytes (). Various studies have also demonstrated elevated rates of dipstick positive proteinuria and hematuria, and several glomerular diseases have been reported, including collapsing focal segmental glomerulosclerosis in patients with high-risk APOL1 risk alleles (). Significant subclinical kidney damage may occur even in those who do not have AKI, as kidney damage was discovered in autopsies of patients who did not have elevated creatinine at death (). Thus, it has been theorized that patients infected with COVID-19 might be at risk of developing chronic kidney disease (CKD).

There has been debate over whether AKI rates are higher in COVID-19 compared to other forms of sepsis, including influenza and community-acquired pneumonia (). AKI may also complicate cases of severe influenza; autopsy series have also shown ATN as the dominant kidney lesion in fatal cases of influenza (). We sought to compare the risk of AKI in hospitalized patients with COVID-19 compared to contemporary historical controls admitted with influenza. Additionally, among those surviving to hospital discharge, we sought to compare the rate of AKI recovery and new-onset CKD.

We conducted an observational, retrospective cohort study of hospitalized adults with COVID-19 within the Mass General Brigham (MGB) healthcare system located in the Boston region. Our institutional review board approved the protocol and waived the need for informed consent. Research was conducted in accordance with the Helsinki Declaration. We obtained clinical data from the Research Patient Data Registry (RPDR), which is the central data repository of MGB used for research and quality improvement purposes (). We included consecutive individuals infected with COVID-19 confirmed by RT-PCR between March 5, 2020 and May 31, 2020 who were hospitalized within two weeks of their first positive test. For our comparison group, we included consecutive individuals who were hospitalized within two weeks of a positive influenza A or B antigen or PCR test from January 1, 2017 and December 30, 2019. Hospitalizations were defined by the presence of a hospital admission encounter lasting at least 24 hours.

We excluded those who did not have at least one outpatient serum creatinine measured between 14 to 365 days prior to their first positive COVID-19 or influenza test in order to reduce any misclassification bias. We excluded patients with baseline estimated glomerular filtration rate (eGFR) below 15 mL/min/1.73m and those on dialysis. Patients were followed until death or 10 months after diagnosis of COVID-19 or influenza, they were censored at the time of death or their last serum creatinine value recorded.

Baseline demographics were obtained from the RPDR “demographics domain.” Race/ethnicity was defined by incorporating race data and primary language spoken, using an algorithm shown in Supplemental Table S1. Baseline creatinine was defined by the outpatient serum creatinine value closest to hospitalization occurring between 14 to 365 days prior to the first positive COVID-19 or influenza test. Baseline eGFR was calculated using the CKD-EPI equation (). Comorbidities were defined by the presence of at least two diagnosis codes using the “diagnoses domain” (Supplemental Table S2). Baseline medications were defined by at least one instance of a prescribed medication within one year of baseline obtained from the “medications domain” (Supplemental Table S3). Laboratory studies at the time of admission were defined as those performed on the day of admission. If laboratory studies were not available on the day of admission, the next closest laboratory value within 72 hours prior to or 24 hours following admission was recorded. Missing data for admission laboratory studies are shown in Supplemental Table S4.

The primary outcome was incidence of AKI during hospitalization, which was defined as a 1.5-fold or greater rise in serum creatinine from baseline. AKI stage 2 (defined as a ≥ 2-fold increase or greater in creatinine from baseline) and AKI stage 3 (defined as ≥ 3-fold increase or greater) were secondary outcomes. For the purposes of our models and analyses, AKI stages are presented as being inclusive of higher stages (e.g. AKI Stage 1 includes AKI Stage 1 or higher). Incident dialysis was not captured in the RPDR dataset. Additional outcomes included () recovery of serum creatinine to within 20% of baseline creatinine by hospital discharge, as was defined by Siew et al.), and () mortality during hospital stay among patients with AKI and the overall cohort. Among patients who survived hospital discharge and whose kidney function was followed for at least 90 days, we determined () new onset CKD defined by eGFR 60 mL/min/1.73m separated by at least 90 days occurring in a patient whose pre-hospitalization baseline eGFR was ≥ 60mL/min/1.73m, and () development of ≥ 25% GFR decline at last follow-up compared to pre-hospital baseline eGFR.

Baseline characteristics were described using means and standard deviations (SD) for normally distributed data and median and interquartile range (IQR) for non-normally distributed data. Counts and percentages were used for categorical variables. Univariate and multivariable Cox proportional-hazards models were performed to estimate the hazard ratios (HRs) of each outcome (AKI, AKI recovery by hospital discharge, in-hospital mortality, incidence of new-onset CKD, and ≥ 25% eGFR decline) by comparing COVID-19 and influenza cohorts. Covariates considered for confounding adjustment included age, gender, race, baseline creatinine, hypertension, diabetes mellitus (DM), cirrhosis, angiotensin-converting enzyme inhibitors / angiotensin II receptor blockers (ACEi/ARB), diuretics, and proton pump inhibitiors (PPIs).

Sensitivity analyses were performed to assess the robustness of our findings by propensity score (PS) adjustment: () matching patients with COVID-19 to those with influenza based on their PS, () inverse probability treatment weighting (IPTW) (). PS were determined based on a multivariable logistic regression model that estimated the probability of having a COVID-19 vs. influenza hospital admission. The covariates in the logistic regression model were age, gender, race/ethnicity, baseline creatinine, DM, chronic obstructive pulmonary disease, congestive heart failure, cirrhosis, human immunodeficiency virus, coronary artery disease, ACEi/ARB, loop diuretics, thiazide diuretics, potassium-sparing diuretics, PPIs, non-steroidal anti-inflammatory drugs, and immune-suppressants. We then matched the COVID-19 patients with those with influenza using 1:1 nearest neighbor greedy matching without replacement and a caliper of 0.2 standard deviartion (SD) of the PS. Standardized differences were calculated between the two cohorts in the original population and after matching or weighting (). The relative risk of developing AKI for COVID-19 patients was estimated by performing Cox regression stratified on matched pairs for the matched cohorts and adjusted for IPTW, repectively.

There were 8446 individuals who tested positive for COVID-19 between March 5, 2020 and May 31, 2020. After applying the exclusions shown in , we included 1091 hospitalized individuals infected with COVID-19 with an established baseline eGFR ≥ 15mL/min/1.73m. There were 8825 individuals who tested positive for influenza between January 1, 2017 and December 30, 2019. After applying the exclusions shown in , we included 1334 hospitalized individuals with influenza. There were 871 (80%) patients with COVID-19 and 1230 (92%) patients with influenza who were hospitalized on the day they tested positive; of those that were hospitalized in the days following a positive test (220 with COVID-19 and 104 with influenza), the median time between diagnosis and hospital admission was 4 days (IQR 2 – 7) for 220 patients with COVID-19 and 6 days (IQR 2 – 9) for patients with influenza. The mean length of stay is 9.9 days (SD 10.8) for COVID-19 and 5.7 days (SD 5.7) for influenza. The mean time from diagnosis to last creatinine value was an average of 117 days (SD 113) for patients with COVID-19 and 174 days (SD 110) for patients with influenza.

Baseline characteristics are shown in . Age, sex, and baseline creatinine were similar across both cohorts. Race and ethnicity differed between the two groups; in the COVID-19 cohort, 53% were white, 22% were Hispanic, and 17% were black; in the influenza cohort, 76% were white, 8% were Hispanic, and 9% were black. Baseline eGFR was higher in the COVID-19 cohort than in the influenza cohort (76 mL/min/1.73m vs 70 mL/min/1.73m). Comorbidities were more common in patients hospitalized with influenza compared to COVID-19, with a higher prevalence of congestive heart failure (20% vs. 10%), coronary artery disease (59% vs 41%), and chronic obstructive pulmonary disease (40% vs. 16%) than in the influenza cohort. Medication use was similar in the two cohorts, with the main difference being higher use of immunosuppressant medications in patients with influenza (39% in influenza vs. 17% in COVID-19). Standardized differences between baseline characteristics in non-propensity score matching, propensity score matching, and IPTW cohorts are shown in Supplemental Table S5.

Table 2. Laboratory studies at the time of admission were defined as those performed on the day of admission; if laboratory studies were not available on the day of admission, the next closest laboratory value within 72 hours prior to or 24 hours following admission was recorded. *Represents laboratory values that were significantly different between COVID-19 and influenza cohorts with a -value 0.05. Supplemental Table S4 shows the missing data for the time-of-hospitalization laboratory findings. Abbreviations: ALT = alanine aminotransferase; AST = aspartate aminotransferase; BUN = blood urea nitrogen; PT-INR = prothrombin time and international normalized ratio; PTT = partial thromboplastin time (PTT); WBC = white blood cell count.

Admission laboratory results among patients with AKI are shown in . We did not detect major differences between admission laboratory values between patients with COVID-19 and influenza. Inflammatory and coagulation markers such as C-reactive protein, interleukin-6, and D-dimer were elevated in patients with COVID-19; however, they were not routinely measured for patients with influenza, limiting any comparisons (Supplemental Table S6).

Association between COVID-19 vs. influenza and kidney outcomes. All models were adjusted for the following covariates: age, sex, baseline creatinine, race/ethnicity (white/Hispanic/black/other), COVID-19 vs. influenza. Models are shown in Supplemental Table S7A-G and Supplemental Table 8A-B. Abbreviations: COVID-19 = coronavirus disease 2019; AKI = acute kidney injury; HR = hazard ratios; CI = confidence interval; CKD = chronic kidney injury.

The incidence and severity of AKI in patients hospitalized with COVID-19 and influenza are shown . Unadjusted overall AKI Stage 1 or higher rates were 251 (23%) in COVID-19 cohort and 179 (13%) in influenza cohort ( 0.01); AKI Stage 2 or higher rates were 140 (13%) in COVID-19 cohort and 58 (4%) in the influenza cohort ( 0.01); AKI Stage 3 rates were 75 (7%) in COVID-19 cohort and 19 (1%) in the influenza cohort ( 0.01). The results of the univariate and multivariable models comparing incidence of AKI between influenza and COVID-19 are shown in . The median time to the first occurrence of AKI was 6 days (IQR 3 – 10) and 4 days (IQR 2 – 6) after baseline, respectively. The unadjusted hazard ratio for AKI stage 1 or higher was 1.46 (95% CI 1.20 – 1.77) ( and ). After adjusting for confounding factors, the risk of AKI was higher in patients with COVID-19 compared to patients with influenza, with the estimated aHRs for AKI stage 1 or higher, stage 2 or higher and stage 3 were 1.58 (95% CI 1.29 – 1.94), 2.09 (95% CI 1.50 – 2.91), and 2.67 (95% CI 1.56 – 4.58), respectively ( and Supplemental Tables S7A-B). Consistent findings were observed from sensitivity analyses using propensity score approaches (Supplemental Tables S8A-B).

Overall mortality rate was during hospital stay was 145 (13%) of patients with COVID-19 and 31 (2%) of patients with influenza ( 0.01); the aHR for death was 7.17 (95% CI 4.78-10.76) ( and Supplemental Table S7C).

In-hospital outcomes among patients with AKI are shown in . Compared to AKI patients with influenza, those in the COVID-19 cohort had a higher rates of in-hospital mortality (81 out of 251, 32% vs 19 out of 179, 11%, 0.01). After adjusting for confounders at baseline, the aHR for in-hospital mortality was 3.55 (95% CI 2.11 – 5.97) for patients with COVID-19-associated AKI vs. influenza-associated AKI ( and Supplemental Table S7D).

AKI recovery rates were lower in patients with COVID-19. Among the 251 patients with COVID-19-associated AKI, 140 (56%) recovered to within 20% of baseline creatinine while among the 179 patients with influenza-associated AKI, 126 (70%) recovered to within 20% of baseline creatinine ( 0.01). In the multivariable model, the aHR for AKI recovery prior to hospital discharge was 0.47 (95% CI 0.36 – 0.62) for patients with COVID-19 vs. influenza ( and Supplemental Table S7E).

We then examined long-term renal outcomes in patients who survived to discharge and were followed for at least 90 days. Ten months had elapsed since diagnosis of COVID-19 in all included patients so we censored follow-up at 10 months after baseline in all participants. Among patients with COVID-19, 525 (48%) survived to hospital discharge and were followed for ≥ 90 days. Among these patients, 60 (11%) had eGFR decline ≥ 25% from baseline. Among patients with influenza, 947 (71%) survived to hospital discharge and were followed ≥ 90 days. Among these patients, 103 (11%) had eGFR decline ≥ 25% from baseline ( = 0.81). Patients with COVID-19 had an insignificant trend toward higher risk of a long-term eGFR decline of ≥ 25% (aHR 1.36, 95% CI 0.97 – 1.90) compared to those with influenza ( and Supplemental Table S7F).

We also determined the risk of new-onset CKD (defined as eGFR 60 mL/min/1.73m separated by more than 90 days) in patients who did not have CKD at baseline. (). Among survivors, there were 51 (14%) cases of new-onset CKD among 374 in the COVID-19 cohort and 95 (17%) cases among 575 in the influenza cohort ( = 0.27). In the multivariable model, the aHR for new-onset CKD during the 10-month follow-up was 1.24 (95% CI 0.86 – 1.78) for patients with COVID-19 vs. influenza ( and Supplemental Table S7G).

In our study examining AKI incidence and outcomes among hospitalized patients with COVID-19 compared to a historical control group of patients who were hospitalized with influenza between 2017 – 2019 in a large healthcare system, we found that patients hospitalized due to COVID-19 were 1.58 times more likely to develop AKI compared to patients with influenza. For AKI Stage 2 or higher and AKI Stage 3 there was a 2.09-fold and 2.67-fold higher risk, respectively, in patients with COVID-19 compared to those with influenza. Our findings were robust, as similar findings were observed from the sensitivity analyses using propensity score approaches. Additionally, we found dramatically higher risk of overall mortality during hospital stay and higher mortality risk among patients with COVID-19-associated AKI compared to influenza-associated AKI. Patients with COVID-19-associated AKI were less likely to recover prior to hospital discharge; however, among patients who survived to hospital discharge and were followed for ≥ 90 days, the rates of progressive eGFR decline ≥ 25% from baseline and new-onset CKD were not significantly different. Our study is among the first papers to provide estimates of AKI recovery that extend beyond the index hospitalization and as well as providing long-term follow-up (10 months post-infection).

5 Presenting Characteristics, Comorbidities, and Outcomes Among 5700 Patients Hospitalized With COVID-19 in the New York City Area. 6 Mohamed MMB, Lukitsch I, Torres-Ortiz AE, Walker JB, Varghese V, Hernandez-Arroyo CF, Alqudsi M, LeDoux JR, Velez JCQ. Acute Kidney Injury Associated with Coronavirus Disease 2019 in Urban New Orleans. Kidney360.1(7). 7 Xie Y, Bowe B, Maddukuri G, Al-Aly Z. Comparative evaluation of clinical manifestations and risk of death in patients admitted to hospital with covid-19 and seasonal influenza: cohort study. BMJ. 2020;371:m4677. Epub 2020/12/17. doi: 10.1136/bmj.m4677. PubMed PMID: 33323357; PMCID: PMC7735416 at www.icmje.org/coi_disclosure.pdf and declare: support from the United States Department of Veterans Affairs, the Institute for Public Health at Washington University in Saint Louis, MO, USA, the American Society of Nephrology, and KidneyCure; no financial relationships with any organizations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work. 8 High rate of renal recovery in survivors of COVID-19 associated acute renal failure requiring renal replacement therapy. 9 Outcomes for Patients With COVID-19 and Acute Kidney Injury: A Systematic Review and Meta-Analysis. 9 Outcomes for Patients With COVID-19 and Acute Kidney Injury: A Systematic Review and Meta-Analysis. 44 AKI in Hospitalized Patients with and without COVID-19: A Comparison Study. 7 Xie Y, Bowe B, Maddukuri G, Al-Aly Z. Comparative evaluation of clinical manifestations and risk of death in patients admitted to hospital with covid-19 and seasonal influenza: cohort study. BMJ. 2020;371:m4677. Epub 2020/12/17. doi: 10.1136/bmj.m4677. PubMed PMID: 33323357; PMCID: PMC7735416 at www.icmje.org/coi_disclosure.pdf and declare: support from the United States Department of Veterans Affairs, the Institute for Public Health at Washington University in Saint Louis, MO, USA, the American Society of Nephrology, and KidneyCure; no financial relationships with any organizations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work. Our study highlights the difference between AKI in COVID-19 and influenza and adds to a growing body of literature that sheds light on the magnitude of the effect of COVID-19 on kidney function. In our study, 23% of hospitalized patients with COVID-19 experienced AKI, which is within the range found in other studies of AKI rates in hospitalized patients in the United States, estimated to be 17-37% (). Robbins-Juarez et al. conducted a metanalysis of 20 studies consisting of 13137 hospitalized patients in total, and found that AKI occurred in 17%, with a range of 0.5% to 80.3% (). Our cohort was more than 10 years older on average and tended to have more comorbidities compared to the patients in this prior metanalysis, with higher rates of baseline CKD and other comorbidities. Fisher et al. compared AKI rates in patients hospitalized with COVID-19 in another New York City area hospital system and found a 2.3 (95% CI 2.2 – 2.4) adjusted risk of AKI compared to a mix of hospitalized patients from the same hospital system one year earlier, defining AKI as a 0.3 mg/dl increase or > 50% increase in serum creatinine from the baseline creatinine to maximum creatinine during hospital stay (). Using patients with a similar reason for admission (influenza) makes our control group more homogenous and highlights the uniquely high risk associated with COVID-19 compared to another potentially life-threatening viral respiratory infection. Piroth et al. used the French national administrative database to compare outcomes in hospitalized patients with COVID-19 or influenza during the same time frame. They found an AKI rate of 6.4% among patients with COVID-19 and 4.9% among patients with influenza, which was statistically significant. Of note, in this study, AKI was defined by diagnosis codes, which likely explains the lower incidence. Xie et al. recently compared outcomes among hospitalized patients diagnosed with COVID-19 and influenza in the US Department of Veterans Affairs Database. They also found that AKI (defined as either a 0.3 mg/dL increase or ≥ 1.5 times baseline creatinine) was higher in COVID-19 than influenza, estimating a 1.52 (95% CI 1.37 – 1.69) increased risk (). Our study expands on this by demonstrating that the risk of higher stages of AKI is even more dramatically increased in patients with COVID-19. Additionally, we revealed a higher risk of mortality by end of hospital stay among patients with COVID-19 with AKI compared to patients with influenza who had AKI. Finally, because our study only included patients with known baseline creatinine within 14-365 days prior to admission, we were able to assess AKI recovery and long term kidney outcomes.

45 Outcomes of patients with end-stage kidney disease hospitalized with COVID-19. 8 High rate of renal recovery in survivors of COVID-19 associated acute renal failure requiring renal replacement therapy. 46 High rates of long-term renal recovery in survivors of COVID-19-associated acute kidney injury requiring kidney replacement therapy. There is no standardized definition for AKI recovery. Using a stricter definition than other recently published series in patients with COVID-19, we found that 56% of patients with COVID-19 recovered to within 20% of pre-hospital baseline by discharge, which was significantly lower than in patients with influenza. Ng and colleagues reported that 74% of hospitalized patients with COVID-19 had an improved serum creatinine by time of discharge, defined as a decline of 33% in discharge serum creatinine level from peak admission serum creatinine level (). Stevens et al. reported that 41% of patients with COVID-19 who had received renal replacement therapy for AKI achieved recovery, defined by cessation of dialysis treatment prior to discharge (). Gupta and colleagues studied patients hospitalized with COVID-19 in the intensive care unit who had received renal replacement therapy for AKI, and found that 143 (66%) were able to discontinue dialysis by discharge. Stockmann et al. revealed that patients with dialysis-dependent AKI who survived to discharge, 92% had discontinued dialysis and 62% had a full recovery of kidney function at a median of 151 days ().

To date there are no large reports of long-term kidney outcomes in patients who develop COVID-19. In our study we found similar rates of new onset CKD or ≥ 25% eGFR decline between the COVID-19 and influenza cohorts in survivors followed for up to 10 months post diagnosis of COVID-19 or influenza. This result is limited by the fact that approximately 50% of surviving patients did not have a creatinine measured ≥ 90 days after diagnosis of COVID-19; thus there is a large effect of missing data which may bias our estimates. Prospective studies that measure kidney function via blood and urine tests in large cohorts with COVID-19 will be needed to understand the true burden of kidney disease. Our study highlights the importance of close, post-discharge follow-up for patients with AKI.

47 Assessment of COVID-19 Hospitalizations by Race/Ethnicity in 12 States. Our study had several additional limitations. This is a single healthcare system and therefore limits generalizability. However, the COVID-19 cohort of patients was racially and ethnically diverse. As prior studies have noted, the burden of COVID-19 on minority populations is reflected by the fact that a substantially larger proportion of Hispanic and black individuals were hospitalized with COVID-19 compared to the historical controls that were hospitalized with influenza in our healthcare system (). Our study was unable to directly compare certain inflammatory markers between COVID-19 and influenza. Patients with COVID-19 had elevated markers of inflammation and blood coagulation, but we could not make a direct comparison to influenza because these markers were so infrequently checked in patients with influenza (Supplemental Table S6). We could not reliably capture urine output or renal replacement in our electronic databse, and thus could not use this criteria for our AKI definition which was based only on fold-creatinine changes. Since we only included patients who had a baseline creatinine, this ultimately led to the exclusion of considerable numbers of hospitalized patients (); however, this was necessary in order to accurately define AKI recovery and determine the rate of new-onset CKD. Additionally, by capturing a single baseline creatinine within one year of baseline, it meant that some patients had different follow-up intervals compared to others. Finally, our electronic health record database did not capture important markers of acuity of illness, including use of vasopressors, mechanical ventilation, incident dialysis, and intensive care unit stay; thus, we could not correct for severity among hospitalized patients. Nonetheless, we believe that disease severity among hospitalized patients may be a mediator of differential rates of AKI, and therefore our primary analysis is valuable for providers caring for these patients. Studies that evaluate illness severity markers on long term kidney function in patients with COVID-19 will be needed.

Our study confirms that the risk of AKI from COVID-19 exceeds the risk of AKI in with influenza in patients who are hospitalized. The risk of AKI Stage 2 or higher is greater than 2-fold in patients with COVID-19 than patients with influenza; additionally, those with COVID-19 who suffer AKI are 3.6 times more likely to die, and are nearly half as likely to recover by hospital discharge. However, among patients with COVID-19 that survive to hospital discharge and are followed for ≥ 90 days, there does not appear to be a heightend risk of new-onset CKD nor ≥ 25% eGFR decline within 10 months of baseline. Larger studies of inpatients and outpatients with longer follow-up will be needed to define long-term risks of COVID-19 on kidney function.