CJASN ePress. Published on February 23, 2021 as doi: 10.2215/CJN.09130620 Article

Proteinuria and Clinical Outcomes in Hospitalized COVID-19 Patients A Retrospective Single-Center Study

Alexandre Karras ,1,2 Marine Livrozet,2,3 He´le`ne Lazareth,1,2 Nicolas Benichou,1,2 Jean-Se´bastien Hulot,2,3 Antoine Fayol,2,3 Sophie Chauvet,1,2 Anne-Sophie Jannot,2,4 Marie-Aude Penet,2,5 Jean-Luc Diehl,2,6 Anne Godier,2,7 Olivier Sanchez,2,8 Tristan Mirault ,2,9 Eric Thervet,1,2 and Nicolas Pallet2,10 Due to the number of contributing authors, Abstract the affiliations are Background and objectives Kidney involvement is frequent among patients with coronavirus disease 2019 listed at the end of (COVID-19), and occurrence of AKI is associated with higher mortality in this population. The objective of this this article. study was to describe occurrence and significance of proteinuria in this setting. Correspondence: Prof. Alexandre Design, setting, participants & measurements We conducted a single-center retrospective study to describe Karras, Department of the characteristic features of proteinuria measured within 48 hours following admission among patients with , Hoˆpital COVID-19 admitted in a tertiary care hospital in France, and to evaluate its association with initiation of dialysis, Europe´en Georges intensive care unit admission, and death. Pompidou, Assistance Publique-Hoˆpitaux de Paris, 20 rue Leblanc, Results Among 200 patients with available data, urine protein-creatinine ratio at admission was $1 g/g for 75015 Paris, France. 84 (42%), although kidney function was normal in most patients, with a median serum creatinine of 0.94 mg/dl Email: alexandre. (interquartile range, 0.75–1.21). Median urine albumin-creatinine ratio was 110 mg/g (interquartile range, [email protected] 50–410), with a urine albumin-protein ratio ,50% in 92% of patients. Urine retinol binding protein concentrations, available for 85 patients, were $0.03 mg/mmol in 62% of patients. Urine protein-creatinine ratio $1g/gwas associated with initiation of dialysis (odds ratio, 4.87; 95% confidence interval, 2.03 to 13.0; P,0.001), admission to the intensive care unit (odds ratio, 3.55; 95% confidence interval, 1.93 to 6.71; P,0.001), and death (odds ratio, 3.56; 95% confidence interval, 1.90 to 6.54; P,0.001).

Conclusions Proteinuriais very frequent among patients admitted forCOVID-19 and may precedeAKI. Low levels of suggest a predominant tubular origin, confirmed by the elevated levels of urine retinol binding protein. Urine protein-creatinine ratio $1 g/g at admission is strongly associated with poor kidney and patient outcome. CJASN 16: ccc–ccc, 2021. doi: https://doi.org/10.2215/CJN.09130620

Introduction reports suggested that acute kidney failure was rare Infection with severe acute respiratory syndrome (,5%) in the entire infected population (1,5), the coronavirus 2 (SARS-CoV-2) virus is responsible for incidence of this complication is much higher, up to coronavirus disease 2019 (COVID-19). Despite pre- 65%, among patients with severe disease who were dominant respiratory manifestations, COVID-19 has a admitted to the ICU (6) or those with a fatal outcome broad clinical spectrum encompassing asymptomatic (2,7–9). More recently, a retrospective study of 3993 infection, mild upper airways illness with fever, and patients with COVID-19 hospitalized in New York severe pneumonia with respiratory failure, as well as revealed that AKI occurred in 46% of patients and that thromboembolic complications, coagulopathy, acute 19% of patients with AKI required dialysis (10). An cardiac injury, and neurologic symptoms (1,2). Pa- early report from a large (n5701) cohort in China tients with critical presentations require admission to revealed that 43.9% of patients had proteinuria and intensive care units (ICUs) either for acute respiratory 26.7% had on admission (11). The same distress syndrome or for cytokine storm syndrome authors reported that among 333 patients hospitalized leading to multiple organ dysfunction (3). Mortality is in Wuhan with COVID-19 pneumonia, 75.4% had particularly important among older patients as well as transient abnormal urine dipstick tests (12). Among populations with comorbidities such as obesity, di- the 435 patients with AKI and available urine tests abetes, hypertension, chronic respiratory disease, or reported by Chan et al. (10), 84% had proteinuria, 81% previous immunosuppression (4). had hematuria, and 60% had leucocyturia. Several teams have reported kidney involvement In the first published autopsy series (13), postmor- in SARS-CoV-2–infected patients. Although initial tem examination of kidney was reported in www.cjasn.org Vol 16 April, 2021 Copyright © 2021 by the American Society of Nephrology 1 2 CJASN

26 cases, showing that the most common finding was patients during an outpatient clinic visit in the 2-month diffuse proximal acute tubular necrosis (ATN), due to period following hospital discharge. either kidney ischemia or to direct infection of kidney epithelial cells by SARS-CoV-2, as suggested by the pres- Laboratory Measures ence of viral-like particles on electron microscopy studies Urine biochemical parameters were obtained prospec- (13,14) and also by the presence of viral RNA in kidney cells tively on spot urine collection within 48 hours after fi (15). This tropism of the virus for the kidney was con rmed admission. Proteinuria was expressed as urine protein- by other authors detecting SARS-CoV-2 RNA in the kidney creatinine ratio (UPCR), and albuminuria was expressed as of 73% of patients with COVID-19 and AKI and also urine albumin-creatinine ratio (UACR). Tubular dysfunc- showing that patient-derived virus can replicate in kidney tion was evaluated in a subset of patients by assessment of tubular epithelial cells (16). urine retinol binding protein concentration, expressed as Notwithstanding, a comprehensive characterization of urine retinol binding protein-creatinine ratio. kidney injury associated with SARS-CoV-2 is lacking. We The total urinary protein concentration was quantified undertook this study to describe the clinical and biochem- by measuring pyrogallol red at 600-/800-nm absorbance ical features of AKI associated with SARS-CoV-2 in a large (urinary protein assay; Beckman Coulter), and the albumin cohort of patients. The aim of this study was to study the concentration was measured by immunoturbidimetry anal- prevalence of proteinuria among hospitalized patients with ysis (DIAgAM assay; Beckman Coulter) using a Beckman COVID-19 to describe its nature and to test the prognostic Coulter AU680 analyzer. The urinary levels of retinol value of proteinuria in this setting. binding protein were measured using a Siemens BN II Nephelometer Analyzer II. Urine protein concentrations were corrected for creatinine levels, which were measured Materials and Methods using a colorimetric assay (modified kinetic Jaffe method) Study Design, Setting, and Population on a Beckman Coulter DXC analyzer (serum) or a Beckman This retrospective, single-center, observational study Coulter AU680 analyzer (urine). was conducted in the Hôpital Européen Georges Pompi- dou, an academic hospital belonging to the Assistance Publique-Hôpitaux de Paris (APHP) hospital network and Definitions fi providing health care to the population of Paris area. The AKI was de ned by the Kidney Disease Improving local institutional review board (registration no. 00011928, Global Outcomes (KDIGO) AKI criteria (17). The KDIGO Comité d’Ethique de la Recherche Assistance Publique criteria for AKI were applied using only serum creatinine Hôpitaux de Paris 5) approved this study as noninterven- variations, as urine output data were not available for tional research using data collected for routine clinical patients developing AKI before entering ICU. Baseline fi practice. Patients included in this study were all informed serum creatinine value was de ned by the most recent that their medical data could be used for research purpose, value found in the centralized electronic medical records of in accordance with General Data Protection Regulation the Assistance Publique-Hôpitaux de Paris network pre- 2016/679. Our analysis follows recommendations pro- ceding the current admission. Whenever data were not vided by the Reporting of Studies Conducted Using available, we used as baseline value the initial serum Observational Routinely-Collected Health Data statement. creatinine level measured at the time of admission. eGFR After the publication of the first data concerning kidney was calculated using the Epide- miology Collaboration formula (18). CKD was defined as a injury during SARS-CoV-2 infection, local recommenda- 2 tions were made in our institution for clinicians in charge of previously known eGFR below 60 ml/min per 1.73 m patients with COVID-19 in the different units, such as preceding actual hospitalization or a former history of emergency rooms, internal , , vas- kidney transplantation. cular medicine, nephrology, and ICUs. In accordance with Indication for ICU admission was on the basis of the these recommendations, patients had early routine urine presence of acute respiratory distress syndrome requiring fl testing for proteinuria and, whenever possible, urine high- ow nasal oxygen , noninvasive mechanical albumin and retinol binding protein assessment. All pa- ventilation, or invasive mechanical ventilation. Hospitali- tients admitted in the hospital between March 17 and April zation in non-ICU departments was decided when patients fl 15, 2020, with proven SARS-CoV-2 infection and available required only nasal low- ow oxygen administration and laboratory proteinuria testing within 48 hours following standard medical monitoring. admission were included in this study. The SARS-CoV-2 Comorbidities, including diabetes mellitus, hyperten- infection was proven by a positive result on PCR testing of sion, or CKD, were retrospectively collected in individual fi a nasopharyngeal or oropharyngeal swab specimen, bron- medical les of the included patients. choalveolar lavage, or bronchial aspirates. The only exclu- sion criterion was kidney failure requiring long-term Statistical Analyses dialysis therapy as history. Clinical outcomes (AKI, ICU Categorical variables were summarized as percentages, admission, and death) were monitored until the end of and continuous variables were expressed as the mean 6 SD hospitalization through August 1, 2020, the end date of or median with interquartile range (IQR). Characteristics follow-up. Complete data were available for all patients of subgroups were compared with the use of a t test or a during hospitalization, retrieved from the institutional Wilcoxon rank-sum test for continuous measures and a chi- electronic medical records.Follow-updata,including squared test or a Fisher exact test for categorical variables. kidney function assessment, could be collected for most The effect of initial proteinuria on the occurrence of death CJASN 16: ccc–ccc, April, 2021 Proteinuria and Clinical Outcomes in COVID-19 Patients, Karras et al. 3

was examined by Kaplan–Meier analysis with the use of seven were kidney transplant recipients. Previous use of log-rank test. Multivariable logistic regression was per- antihypertensive therapy with angiotensin-converting en- formed to evaluate the relationship between UPCR, retinol zyme inhibitor (ACEi) or angiotensin II receptor blocker binding protein-creatinine ratio, and outcomes adjusted for (ARB) was reported for 60 (30%) patients, whereas 42 (21%) age, sex, body mass index (BMI), hypertension, diabetes, patients were receiving diuretics. The median delay be- and serum creatinine at admission. Adjusted analysis tween self-reported first signs of viral infection and excluded patients with missing data for one or more admission was 7 days (IQR, 5–9). covariates. Results are reported as odds ratios (ORs) with 95% confidence intervals (95% CIs). Statistical signif- icance was set at two-sided P50.05. Analyses were per- Proteinuria formed with either version 8.0.2 of JMP software (SAS Median UPCR at admission was 0.86 g/g (IQR, Institute) or version 1.2.5033 of the R Studio software (R 0.38–1.50). Proteinuria was ,0.2 g/g in 24 (12%) patients, Project for Statistical Computing; R Foundation). ,0.5 g/g in 72 (36%) patients, between 0.5 and 1 g/g in 44 (22%) patients, between 1 and 3 g/g in 68 (34%) patients, and above 3 g/g in only 16 (8%) patients. Despite the fact Results that urine specimens were collected within 48 hours of Study Population admission, 45 patients had these urine tests performed after Among the 322 patients admitted by April 15, 2020, 122 the AKI diagnosis was made on the basis of serum patients were excluded from analysis due to lack of data on creatinine elevation. proteinuria or to presence of preexisting kidney failure Baseline UACR was available for only 152 patients requiring long-term dialysis therapy. Figure 1 shows the (76% of the entire study population). Median UACR was flow diagram of the study. The characteristics of patients 0.11 g/g (IQR, 0.05–0.41). The median UACR-UPCR ratio without available proteinuria are detailed in Supplemental was 18% (IQR, 9.6–33.4). UACR-UPCR was above 50% for Table 1, revealing a population with fewer comorbidities only 8% of patients. and lower risk of admission to ICU or death. Baseline urine retinol binding protein was measured in Overall, 200 patients—143 men and 57 women—were 85 patients. All but two of them had UPCR and UACR included in this study (Table 1). Median age was 63 years measured on the same urine specimen. The urine retinol (IQR, 54–73), with 32 of 200 (16%) patients aged above binding protein-creatinine ratio was below the detection 80 years. BMI was available for 171 patients, with a median threshold (0.03 mg/mmol) in 32 (37%) patients. For the 53 value of 26.7 kg/m2 (IQR, 24.3–30.1). Of note, 66 (39%) (63%) patients with detectable urine retinol binding pro- patients had a BMI between 25 and 30 kg/m2, and 44 (26%) tein, the median value of urine retinol binding protein- patients had a value above 30 kg/m2.Formerhistoryof creatinine ratio was 1.35 (IQR, 0.63–4.81) mg/mmol. hypertension was reported in 103 patients (51%), and Among the 85 patients with available urine retinol binding diabetes mellitus was reported in 51 (25%) patients. Pre- protein assessment (Table 1), those with urine retinol viously known CKD was reported for 18 patients, of whom binding protein-creatinine ratio above the detection

Consecutive patients admitted for COVID-19 (n=322)

ESKD (n=4) Patients with available baseline No available baseIine UACR assessment proteinuria (n=118) (n=152) Patients with available baseline UPCR assessment (n=200)

Patients with available baseline In-hospital death (n=58) urine RBP assessment (n=85)

Discharged patients (n=142)

Lost to follow-up (n=7)

Discharged patients with follow-up data (n=135)

Figure 1. | Flow diagram. Depicts included and excluded patients, as well as numbers of patients available for each analysis. COVID-19, coronavirus disease 2019; RBP, retinol binding protein; UACR, urine albumin-creatinine ratio; UPCR, urine protein-creatinine ratio. 4 CJASN

Table 1. Baseline characteristics for patients admitted for coronavirus disease 2019 according to proteinuria level at admission

Urine Protein-Creatinine Urine Retinol Binding Protein- Ratio, g/g Creatinine Ratio, mg/mmol Characteristics All Patients, n5200 ,1, n5116 $1, n584 ,0.03, n532 $0.03, n553

Men 143 (71) 78 (67) 65 (77) 21 (65) 39 (76) Age, yr 63 (54–73) 60 (53–69) 67 (55–79) 59 (48–64) 63 (54–75) BMI, kg/m2a 26.7 (24.3–30.1) 26.5 (23.9–30.3) 27.4 (24.6–30.1) 26.1 (24.5–30) 27.6 (24.5–31.3) Hypertension 103 (51) 53 (46) 50 (59) 12 (37) 29 (55) Diabetes 51 (25) 25 (21) 26 (31) 7 (22) 16 (30) CKD 18 (9) 10 (9) 8 (9) 2 (6) 5 (9) Use of ACEi or ARB 60 (30) 25 (21) 35 (41) 7 (22) 20 (38) Use of diuretics 42 (21) 19 (16) 23 (27) 5 (16) 10 (19) First symptoms to admission, d 7 (5–9) 7 (6–9.7) 7 (4–8) 7 (4–9) 7 (5–10) UPCR, g/g 0.86 (0.38–1.50) 0.41 (0.25–0.72) 1.6 (1.3–2.1) 0.24 (0.15–0.51) 1.22 (0.75–1.8) UACR, mg/gb 110 (50–410) 60 (20–100) 480 (260–830) 98 (42–382) 250 (103–533) UACR-UPCR,b % 18 (9.6–33.4) 13 (6.7–20.5) 32 (20.9–43.3) 16 (8.5–27.6) 23 (13.2–36.6) URBPCR,c mg/mmol 0.31 (0–2.76) 0 (0–0.14) 2.8 (1–5.8) 0 (0–0) 1.35 (0.63–4.8) Serum creatinine, mg/dl 0.94 (0.75–1.21) 0.85 (0.69–1.12) 1.03 (0.85–1.36) 0.83 (0.63–1.11) 0.98 (0.77–1.30)

Quantitative values are given as medians (interquartile ranges), and qualitative values are given as number of patients/percentage. BMI, body mass index; ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; UPCR, urine protein-creatinine ratio; UACR, urine albumin-creatinine ratio; URBPCR, urine retinol binding protein-creatinine ratio. aData available for 171 patients. bData available for 152 patients. cData available for 85 patients.

threshold ($0.03 mg/mmol) had higher levels of baseline AKI was diagnosed in 88 (44%) patients, with 28 (14%), proteinuria and slightly higher serum creatinine 24 (12%), and 36 (18%) patients presenting stage 1, 2, or 3 at admission. AKI, respectively. KRT was required for 27 (13%) patients. In order to compare baseline characteristics, patients During hospitalization, 118 (59%) patients were admitted were divided according to their initial UPCR level (,1 to the ICU for severe COVID-19 pneumonia. At last follow- versus $1 g/g). The main characteristics of the two groups up, 58 (29%) had died after a median delay of 10 days are shown in Table 1. Patients with UPCR$1g/gwere (IQR, 6–20) from admission; 142 (71%) were discharged, older, more likely to have hypertension, and more likely to with a median length of stay of 11 days (IQR, 7–26) receive ACEi or ARB, but there was no difference regarding from admission. the sex ratio, the prevalence of diabetes, the prevalence of Long-term follow-up data after hospitalization for CKD, or the level of BMI when compared with patients COVID-19 were available for 135 of 142 surviving patients. with UPCR,1 g/g. Interestingly, high levels of proteinuria After a median follow-up period of 36 (IQR, 27–54) days, were observed in patients with or without former history of the median serum creatinine was 0.82 (IQR, 0.67–1) mg/dl. diabetes or CKD (Supplemental Tables 2–4). Among the ten surviving patients with last known serum Patients with UPCR$1 g/g at admission had signifi- creatinine .1.5 mg/dl, eight had preexisting CKD. cantly higher levels of UACR and UACR-UPCR ratio. However, only a minority (n530 of 84; 36%) of those fi patients had signi cant glomerular proteinuria with Association of Proteinuria with Outcomes . UACR 0.5 g/g, and only 11 of 86 (13%) had an UACR- The presence of a UPCR$1 g/g was associated with higher $ UPCR above 50%. Patients with UPCR 1g/galsohad peak median serum creatinine (1.79 versus 1.06 mg/dl; higher urine retinol binding protein levels (Table 1). P,0.001), but it was also associated with significant risk of Nevertheless, the predominantly nonglomerular origin of AKI (OR, 3.61; 95% CI, 2.02 to 6.58; P,0.001) and re- fi proteinuria was con rmed by the negative correlation quirement of KRT (OR, 4.87; 95% CI, 2.03 to 13.00; P,0.001), between UACR and urine retinol binding protein- as well as with risk of admission in the ICU (OR, 3.55; 95% creatinine ratio found at high levels of albuminuria CI, 1.93 to 6.71; P,0.001) and occurrence of death during (Supplemental Figure 1). hospitalization (OR, 3.56; 95% CI, 1.90 to 6.84; P,0.001) in unadjusted analysis (Table 2). When studying only the 138 patients with normal kidney function (eGFR.60 ml/min Outcomes per 1.73 m2) at admission, the criterion UPCR$1 g/g was Median serum creatinine level at admission was still associated with ICU admission (77% versus 44%; 0.94 mg/dl (IQR, 0.75–1.21). Only 29 (14%) patients had P,0.001), subsequent AKI (45% versus 18%; P50.004), initial serum creatinine .1.5 mg/dl at admission. Median need for KRT (22% versus 4.5%; P50.001), and occurrence peak serum creatinine was 1.19 mg/dl (IQR, 0.84–2.16), of death (43% versus 14%; P,0.001). Of note, the delay from with 79 (39%) patients having at least one value above first symptoms to admission was not statistically different 1.5 mg/dl during their hospitalization. between the two groups. CJASN 16: ccc–ccc, April, 2021 Proteinuria and Clinical Outcomes in COVID-19 Patients, Karras et al. 5

Table 2. Study outcomes (unadjusted analysis)

Urine Protein-Creatinine Ratio Urine Retinol Binding Protein-Creatinine Ratio

Odds Ratio Odds Ratio Outcomes ,1 g/g, $1g/g, (95% P ,0.03 mg/mmol, $0.03 mg/mmol, (95% P n5116 n584 Confidence Value n532 n553 Confidence Value Interval) Interval)

Peak serum 1.06 1.79 ,0.001 0.92 (0.77–1.19) 1.44 (0.99–1.44) 0.004 creatinine, (0.77–1.54) (1.13–3.79) mg/dl AKI, all stages 36 (31) 52 (62) 3.61 ,0.001 6 (19) 28 (52) 5.24 0.002 Stage 1 15 (13) 13 (15) (2.02 to 6.58) 4 (12) 11 (21) (1.95 to 15.98) Stage 2 10 (9) 14 (17) 1 (3) 8 (15) Stage 3 11 (9) 25 (30) 1 (3) 9 (17) KRT 7 (6) 20 (24) 4.87 ,0.001 0 (0) 6 (11) (2.03 to 13.00) ICU required 55 (47) 64 (76) 3.55 ,0.001 8 (25) 41 (77) 10.25 ,0.001 (1.93 to 6.71) (3.82 to 30.23) Death 21 (18) 37 (44) 3.56 ,0.001 1 (3) 21 (40) 22 ,0.001 (1.90 to 6.54) (4.21 to 406)

Quantitative values are given as medians (interquartile ranges), and qualitative values are given as number of patients/percentage. Odds ratios and P values are given for unadjusted analysis. ICU, intensive care unit.

Kaplan–Meier unadjusted survival curves according to weight proteins, which cannot be reabsorbed by the proximal baseline proteinuria (,1or$1g/g)areshowninFigure2, kidney tubule in this condition due to acute tubular damage. with a significant difference between groups as concerning Critically, in accordance with previous reports (19,20), we occurrence of death (P50.006). demonstrate that low-molecular weight proteinuria reflects A multivariable analysis was performed, including early tubular dysfunction sometimes preceding AKI as UPCR$1 g/g, age, sex, BMI, hypertension, diabetes, and suggested by the frequent normal-range serum creatinine serum creatinine at admission (Supplemental Table 2). The level at admission. criterion UPCR$1 g/g remained associated with need for The kidney histopathologic data published to date have KRT (OR, 4.84; 95% CI, 1.93 to 13.5; P50.001), admission to indeed described diffuse injury of proximal tubular epi- the ICU (OR, 3.66; 95% CI, 1.72 to 8.20; P50.001), occur- thelial cells with loss of brush border and vacuolar de- rence of death (OR, 3.03; 95% CI, 1.46 to 6.39; P50.003), and generation followed by ATN (13,21). Causes of ATN are the composite outcome of death or need for KRT (OR, 4.24; multiple in the setting of severe infection (22), as systemic 95% CI, 2.08 to 8.89; P,0.001). hypoxia, local hypoxia due to microvascular damage, drug The presence of a urine retinol binding protein-creatinine toxicity, or sepsis-induced rhabdomyolysis can all promote ratio $0.03 mg/mmol at admission was associated with a tissue injury. Nevertheless, electron microscopic examina- higher incidence of AKI (OR, 5.24; 95% CI, 1.95 to 15.98; tion of kidney biopsy from patients with COVID-19 has P50.002), admission to the ICU (OR, 10.25; 95% CI, 3.82 to revealed virus-like particles within the tubular epithelium, 30.23; P,0.001), and death (OR, 22; 95% CI, 4.21 to 406; suggesting direct evidence of tubular cells invasion by P,0.001) in unadjusted analysis (Table 2). The multivari- SARS-CoV-2 (14), and viral RNA has been found in kidney able analysis performed using the urine retinol binding tissue from patients with COVID-19–associated AKI (16). protein revealed that a urine retinol binding protein- Interestingly, it has been shown that proximal tubule creatinine ratio .0.03 mg/mmol was associated with occurrence of AKI (OR, 5.18; 95% CI, 1.18 to 29.9; P5 0.04), admission to the ICU (OR, 15.5; 95% CI, 4.35 to 100 UPCR < 1g/g 69.7; P,0.001), and death (OR, 20.9; 95% CI, 3.79 to 394; UPCR  1g/g P50.005) (Supplemental Table 3).

50

Discussion Survival (%) This study confirms that proteinuria is very frequent Log rank p = 0.006 among patients with symptomatic SARS-CoV-2 infection admitted to the hospital, with a UPCR above 1 g/g at 0 0 10203040 admission for 43% of them. Moreover, our results indicate Time from admission (days) that this proteinuria contains very small amounts of albuminuria, as the urine albumin-protein ratio is Subjects at risk Day 0 Day 10 Day 20 Day 30 Day 40 PCR<1 g/g 116 56 27 14 7 .50% for only 8% of patients, even among patients PCR1 g/g 84 56 33 18 11 with UPCR$1 g/g. In line with this, the urine levels of retinol binding protein are elevated in .50% of tested patients, an observation that strengthens the hypothesis Figure 2. | Patient survival with Kaplan–Meier analysis. Analysis of that COVID-19–associated proteinuria reflects low-molecular event-free survival according to admission proteinuria $1or,1g/g. 6 CJASN

kidney cells express the surface angiotensin-converting with COVID-19. Second, some clinical data are missing, enzyme 2 (ACE2), which is instrumental for the coronavi- notably regarding level of inflammatory markers, sever- rus to bind and penetrate target cells (13,23). More recently, ity of respiratory disease, and incidence of other com- proximal tubule dysfunction has been demonstrated in plications, such as coagulopathy or cardiovascular patients with COVID-19 attested by the presence of low- events. These parameters strongly affect patient survival molecular weight proteinuria (70%–80%), neutral amino- in this setting, and therefore, we cannot affirm that aciduria (46%), and defective handling of both uric acid proteinuria is indeed an independent predictor of patient (46%) and phosphate (19%), corresponding to a partial outcome. Third, the use of the UPCR should be inter- Fanconi syndrome (24). preted with caution in AKI, as urinary creatinine excre- Of note, viral-like particles have also been observed tion is variably decreased in this setting. Nevertheless, inside podocytes and endothelial cells, a finding that may the aim of this study was not to estimate the 24-hour promote glomerular damage. Although two recent cases of proteinuriabythisratio,asitisproposedinCKD,but collapsing glomerulopathy associated with COVID-19 was only to validate this simple test as a prognostic have been published (25,26), glomerulopathy due to marker during COVID-19. Fourth, urine retinol binding SARS-CoV-2 infection seems to be rare in Asian and protein—which was used as a marker for proximal European populations, an idea supported by the findings tubular dysfunction in our study—was available for of this study. Nevertheless, this complication may be only 42% of the study population, and we do not have especially prevalent among patients of African ancestry, data about other tubular biomarkers, such as , in particular owing to a specific genetic background and phosphaturia, or neutrophil gelatinase–associated lip- the presence of APOL1 genetic variants, which predis- ocalin, or kidney pathology data to confirm the hypoth- pose to the occurrence of collapsing glomerulopathy (25). esis of a proximal tubule injury. However, both proximal This severe form of FSGS could be secondary to a direct tubular lesions and presence of other features of Fanconi effect of the virus on podocytes or could be the conse- syndrome have recently been reported among patients quence of the SARS-CoV-2–induced cytokine storm, a with COVID-19 (24). To our knowledge, this study is the condition that has been associated with glomerular first large series to report on the composition of lesions (27). proteinuria during COVID-19—in contrast with pre- Our data confirm that overt proteinuria is more frequent vious publications based on dipstick analysis (10,11)— among older patients with history of hypertension. The use and to demonstrate that proteinuria in this setting of renin-angiotensin-aldosterone system inhibitors, such as contains mostly low-molecular weight proteins rather ACEi and ARB, was more frequent among patients with than albumin. high-grade proteinuria. Early reports had suggested that In conclusion, this study reveals that COVID-19 is these drugs could modify cellular mRNA expression and associated with early and frequent tubular proteinuria, protein level of ACE2, the SARS-CoV-2 receptor (28). which is associated with poor kidney outcome and higher Nevertheless, their use has not been associated with greater mortality among patients with symptomatic COVID-19. severity of disease, and recent studies have not confirmed The comprehension of the precise mechanisms underlying ACE2 overexpression in the kidney after renin-angiotensin- this tubular injury requires further investigations. aldosterone system inhibitor exposure (29). In addition, we demonstrate that the presence of low- molecular weight proteinuria may precede elevation of Disclosures serum creatinine and predict both AKI and KRT require- J.-L. Diehl reports receiving research funding from Alung, General ment, as it has been shown in non–COVID-19–associated Electrc Healthcare, and Xenios-Novalung (Fresenius Medical Care); ATN and AKI associated with sepsis (20,30). Although we receiving honoraria from Baxter and Fresenius Medical Care; and have no control non–COVID-19 group, we show that serving on the editorial board of the Annals of Intensive Care. UPCR$1 g/g and retinol binding protein-creatinine ratio A. Godier reports receiving honoraria from Aguettant, Bayer- .0.03 mg/mmol are associated with a higher risk of severe Healthcare, Boehringer-Ingelheim, Bristol-Myers-Squibb/Pfizer, COVID-19 that requires admission in the ICU and leads to and Sanofi. J.-S. Hulot reports consultancy agreements with Bayer higher mortality, even among patients with no significant and Bioserenity and receiving speaker, advisory board, or consul- elevation of serum creatinine at admission. tancy fees from Amgen, AstraZeneca, Bayer, Bristol-Myers Squibb, The prognostic value of kidney function impairment Novartis, and WeHealth. A. Karras reports receiving honoraria from during COVID-19 has already been suggested in other Abbvie, Amgen, Gilead, and Roche Pharmaceuticals. T. Mirault series (2,10,11,31), but our study is the first large series to reports receiving honoraria from Bayer Healthcare SAS and Incyte show that the outcome is associated with simple urine tests, Biosciences France and receiving nonfinancial support from Abbott such as UPCR and UACR, used in daily practice in contrast France, Alexion Pharma France, Amgen SAS, Bayer Healthcare SAS, with new biomarkers, such as neutrophil gelatinase– Boehringer Ingelheim France, Bristol-Myers Squibb, ICOMED, In- associated lipocalin or kidney injury marker-1 (22,32), cyte Biosciences France, MSD France, and Pfizer SAS. O. Sanchez that have been evaluated in other situations with high reports receiving research funding from Bayer, BMS, Daiichi risk of AKI but not yet in early phases of COVID-19. Sankyo, MSD, and Pfizer and receiving honoraria from Bayer, This study has some limitations. First, the study BMS, Boston Scientifics, Chiesi, MSD, Pfizer, and Sanofi Aventis. population probably included patients who have the E. Thervet reports consultancy agreements with Pfizer and Vifor; most severe forms of COVID-19, with a high proportion receiving research funding from Alexion and GSK; and receiving of patients requiring intensive care. In consequence, our honoraria from Pfizer and Vifor. All remaining authors have nothing results cannot be generalized to nonhospitalized patients to disclose. CJASN 16: ccc–ccc, April, 2021 Proteinuria and Clinical Outcomes in COVID-19 Patients, Karras et al. 7

Funding 10. Chan L, Chaudhary K, Saha A, Chauhan K, Vaid A, Zhao S, None. Paranjpe I, Somani S, Richter F, Miotto R, Lala A, Kia A, Timsina P, Li L, Freeman R, Chen R, Narula J, Just AC, Horowitz C, Fayad Z, Cordon-Cardo C, Schadt E, Levin MA, Reich DL, Fuster V,Murphy Supplemental Material B, He JC, Charney AW, Bo¨ttinger EP, Glicksberg BS, Coca SG, This article contains the following supplemental material online at Nadkarni GN; Mount Sinai COVID Informatics Center (MSCIC): http://cjasn.asnjournals.org/lookup/suppl/doi:10.2215/CJN. AKI in hospitalized patients with COVID-19. J Am Soc Nephrol 32: 151–160, 2021 09130620/-/DCSupplemental. 11. ChengY,LuoR,WangK,ZhangM,WangZ,DongL,LiJ,YaoY, Supplemental Table 1. Characteristics and outcome of admitted Ge S, Xu G: Kidney disease is associated with in-hospital patients with and without proteinuria data available at admission. death of patients with COVID-19. Kidney Int 97: 829–838, Supplemental Table 2. Multivariable logistic regression analysis 2020 of outcomes odds according to UPCR$1 g/g, age, sex, BMI, hy- 12. Pei G, Zhang Z, Peng J, Liu L, Zhang C, YuC, Ma Z, Huang Y,Liu W, Yao Y, Zeng R, Xu G: Renal involvement and early prognosis in pertension, diabetes, and initial serum creatinine. patients with COVID-19 pneumonia. J Am Soc Nephrol 31: Supplemental Table 3. Multivariable logistic regression analysis 1157–1165, 2020 of outcomes odds according to urine retinol binding protein- 13. Su H, Yang M, Wan C, Yi L-X, Tang F,Zhu H-Y,Yi F,Yang H-C, Fogo creatinine ratio $0.03 mg/mmol, age, sex, BMI, hypertension, di- AB, Nie X, Zhang C: Renal histopathological analysis of 26 abetes, and initial serum creatinine. postmortem findings of patients with COVID-19 in China. Kidney Int 98: 219–227, 2020 Supplemental Table 4. Characteristics of proteinuria in patients 14. Farkash EA, Wilson AM, Jentzen JM: Ultrastructural evidence for with and without diabetes and with and without CKD. direct renal infection with SARS-CoV-2 [published correction Supplemental Figure 1. Correlation between urine retinol binding appears in J Am Soc Nephrol 31: 2494, 2020 10.1681/ protein-creatinine ratio and UACR-UPCR, UACR, and UPCR. ASN.2020081117]. J Am Soc Nephrol 31: 1683–1687, 2020 15. Puelles VG, Lu¨tgehetmann M, Lindenmeyer MT, Sperhake JP, Wong MN, Allweiss L, Chilla S, Heinemann A, Wanner N, Liu S, References BraunF,LuS,PfefferleS, Schro¨derAS, EdlerC,GrossO,GlatzelM, 1. GuanW-J,NiZ-Y,HuY,LiangW-H,OuC-Q,HeJ-X, LiuL, ShanH, Wichmann D, Wiech T, Kluge S, Pueschel K, Aepfelbacher M, Lei C-L, Hui DSC, Du B, Li L-J, Zeng G, Yuen K-Y, Chen R-C, Tang Huber TB: Multiorgan and renal tropism of SARS-CoV-2. NEngl C-L, Wang T, Chen PY, Xiang J, Li S-Y, Wang J-L, Liang Z-J, Peng JMed383: 590–592, 2020 Y-X, Wei L, Liu Y,Hu Y-H, Peng P,Wang J-M, Liu J-Y,Chen Z, Li G, 16. Braun F, Lu¨tgehetmann M, Pfefferle S, Wong MN, Carsten A, Zheng Z-J, Qiu S-Q, Luo J, Ye C-J, Zhu S-Y, Zhong N-S; China Lindenmeyer MT, No¨rz D, Heinrich F, Meißner K, Wichmann D, Medical Treatment Expert Group for Covid-19: Clinical charac- Kluge S, Gross O, Pueschel K, Schro¨der AS, Edler C, Aepfelbacher teristics of coronavirus disease 2019 in China. N Engl J Med 382: M, Puelles VG, Huber TB: SARS-CoV-2 renal tropism associates 1708–1720, 2020 with acute kidney injury. Lancet 396: 597–598, 2020 2. Zhou F,YuT,Du R, Fan G, Liu Y,Liu Z, Xiang J, Wang Y,Song B, Gu 17. Kellum JA, Lameire N; KDIGO AKI Guideline Work Group: Di- X, Guan L, Wei Y, Li H, Wu X, Xu J, Tu S, Zhang Y, Chen H, Cao B: agnosis, evaluation, and management of acute kidney injury: A Clinical course and risk factors for mortality of adult inpatients KDIGO summary (Part 1). Crit Care 17: 204, 2013 with COVID-19 in Wuhan, China: A retrospective cohort study. 18. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF III, Lancet 395: 1054–1062, 2020 Feldman HI, Kusek JW, Eggers P, Van Lente F, Greene T, Coresh J; 3. Pedersen SF, Ho Y-C: SARS-CoV-2: A storm is raging. J Clin Invest CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration): 130: 2202–2205, 2020 A new equation to estimate glomerular filtration rate [published 4. Jordan RE, Adab P, Cheng KK: Covid-19: Risk factors for severe correction appears in Ann Intern Med 155: 408, 2011]. Ann Intern disease and death. BMJ 368: m1198, 2020 Med 150: 604–612, 2009 5. Wu C, Chen X, Cai Y,Xia J, Zhou X, Xu S, Huang H, Zhang L, Zhou 19. Norden AG, Scheinman SJ, Deschodt-Lanckman MM, Lapsley M, X, Du C, Zhang Y, Song J, Wang S, Chao Y, Yang Z, Xu J, Zhou X, Nortier JL, Thakker RV, Unwin RJ, Wrong O: Tubular proteinuria Chen D, Xiong W, Xu L, Zhou F,Jiang J, Bai C, Zheng J, Song Y: Risk defined by a study of Dent’s (CLCN5 mutation) and other tubular factors associated with acute respiratory distress syndrome and diseases. Kidney Int 57: 240–249, 2000 death in patients with coronavirus disease 2019 pneumonia in 20. Herget-Rosenthal S, Poppen D, Hu¨sing J, Marggraf G, Pietruck F, Wuhan, China [published correction appears in JAMA Intern Med Jakob H-G, Philipp T, Kribben A: Prognostic value of tubular 180: 1031, 2020 10.1001/jamainternmed.2020.1429]. JAMA proteinuria and enzymuria in nonoliguric acute tubular necrosis. Intern Med 180: 934–943, 2020 Clin Chem 50: 552–558, 2004 6. Fanelli V, Fiorentino M, Cantaluppi V, Gesualdo L, Stallone G, 21. Santoriello D, Khairallah P, Bomback AS, Xu K, Kudose S, Batal I, Ronco C, Castellano G: Acute kidney injury in SARS-CoV-2 in- Barasch J, Radhakrishnan J, D’Agati V,Markowitz G: Postmortem fected patients. Crit Care 24: 155, 2020 kidney pathology findings in patients with COVID-19. JAmSoc 7. Chen T,Wu D, Chen H, YanW, Yang D, Chen G, Ma K, Xu D, YuH, Nephrol 31: 2158–2167, 2020 Wang H, Wang T, Guo W, Chen J, Ding C, Zhang X, Huang J, Han 22. Ronco C, Bellomo R, Kellum JA: Acute kidney injury. Lancet 394: M, Li S, Luo X, Zhao J, Ning Q: Clinical characteristics of 113 1949–1964, 2019 deceased patients with coronavirus disease 2019: Retrospective 23. Pan X-W, Xu D, Zhang H, Zhou W, Wang L-H, Cui X-G: Identi- study. BMJ 368: m1091, 2020 fication ofapotential mechanismofacutekidneyinjuryduring the 8. DuY,TuL,ZhuP,MuM,WangR,YangP,WangX,HuC,PingR, COVID-19 outbreak: A study based on single-cell transcriptome Hu P,Li T, Cao F, Chang C, Hu Q, Jin Y,Xu G: Clinical features of analysis. Intensive Care Med 46: 1114–1116, 2020 85 fatal cases of COVID-19 from Wuhan: A retrospective ob- 24. Werion A, Belkhir L, Perrot M, Schmit G, Aydin S, Chen Z, servational study. Am J Respir Crit Care Med 201: 1372–1379, Penaloza A, De Greef J, Yildiz H, Pothen L, Yombi JC, Dewulf J, 2020 Scohy A, Ge´rard L, Wittebole X, Laterre P-F, Miller SE, Devuyst O, 9. Richardson S, Hirsch JS, Narasimhan M, Crawford JM, McGinn T, Jadoul M, Morelle J; Cliniques universitaires Saint-Luc (CUSL) Davidson KW, Barnaby DP, Becker LB, Chelico JD, cohen SL, COVID-19 Research Group: SARS-CoV-2 causes a specific dys- Cookingham J, Coppa K, Diefenbach MA, Dominello AJ, Duer- function of the kidney proximal tubule. Kidney Int 98: Hefele J, Falzon L, Gitlin J, Hajizadeh N, Harvin TG, Hirschwerk 1296–1307, 2020 10.1016/j.kint.2020.07.019 DA, Kim EJ, Kozel ZM, Marrast LM, Mogavero JN, Osorio GA, Qiu 25. Larsen CP, Bourne TD, Wilson JD, Saqqa O, Sharshir MA: Col- M, Zanos TP; Northwell COVID-19 Research Consortium: Pre- lapsing glomerulopathy in a patient with COVID-19. Kidney Int senting characteristics, comorbidities, and outcomes among Rep 5: 935–939, 2020 5700 patients hospitalized with COVID-19 in the New York City 26. Garandeau C, Figueres L, Renaudin K, Fakhouri F: The Case | A area [published correction appears in JAMA 323: 2098, neoplastic cause of acute kidney injury. Kidney Int 97: 621–622, 2020 10.1001/jama.2020.7681]. JAMA 323: 2052–2059, 2020 2020 8 CJASN

27. Thaunat O, Delahousse M, Fakhouri F, Martinez F, Stephan J-L, 31. Hirsch JS, Ng JH, Ross DW, Sharma P, Shah HH, Barnett RL, Noe¨l LH, Karras A: Nephrotic syndrome associated with hemo- Hazzan AD, Fishbane S, Jhaveri KD; Northwell COVID-19 phagocytic syndrome. Kidney Int 69: 1892–1898, 2006 Research Consortium; Northwell Nephrology 28. Ferrario CM, Jessup J, Chappell MC, Averill DB, Brosnihan KB, COVID-19 Research Consortium: Acute kidney injury in TallantEA,DizDI,GallagherPE:Effectofangiotensin-converting patients hospitalized with COVID-19. Kidney Int 98: enzyme inhibition and angiotensin II receptor blockers on car- 209–218, 2020 diac angiotensin-converting enzyme 2. Circulation 111: 32. Desanti De Oliveira B, Xu K, Shen TH, Callahan M, Kiryluk K, 2605–2610, 2005 D’Agati VD, Tatonetti NP, Barasch J, Devarajan P: Molecular 29. Wysocki J, Lores E, Ye M, Soler MJ, Batlle D: Kidney and lung nephrology: Types of acute tubular injury. Nat Rev Nephrol 15: ACE2 expression after an ACE inhibitor or an Ang II receptor 599–612, 2019 blocker: Implications for COVID-19. J Am Soc Nephrol 31: 1941–1943, 2020 Received: June 8, 2020 Accepted: January 14, 2021 30. Bagshaw SM, Langenberg C, Bellomo R: Urinary biochemistry and microscopy in septic acute renal failure: A systematic review. Published online ahead of print. Publication date available at Am J Kidney Dis 48: 695–705, 2006 www.cjasn.org.

AFFILIATIONS

1Department of Nephrology, Hoˆpital Europe´en Georges Pompidou, Assistance Publique-Hoˆpitaux de Paris, Paris, France 2Universite´ de Paris, Paris, France 3Centre d’Investigation Clinique1418 and Departement Medico-Universitaire, CArdiologie Rein Transplantation nEurovasculaire, Hoˆpital Europe´en Georges Pompidou, Assistance Publique-Hoˆpitaux de Paris, Paris, France 4Department of Bioinformatics, Hoˆpital Europe´en Georges Pompidou, Assistance Publique-Hoˆpitaux de Paris, Paris, France 5Department of , Hoˆpital Europe´en Georges Pompidou, Assistance Publique-Hoˆpitaux de Paris, Paris, France 6Department of , Hoˆpital Europe´en Georges Pompidou, Assistance Publique-Hoˆpitaux de Paris, Paris, France 7Department of Anaesthesiology and Intensive Care, Hoˆpital Europe´en Georges Pompidou, Assistance Publique-Hoˆpitaux de Paris, Paris, France 8Department of Pneumology and Intensive Care, Hoˆpital Europe´en Georges Pompidou, Assistance Publique-Hoˆpitaux de Paris, Paris, France 9Department of Vascular Medicine, Hoˆpital Europe´en Georges Pompidou, Assistance Publique-Hoˆpitaux de Paris, Paris, France 10Department of , Hoˆpital Europe´en Georges Pompidou, Assistance Publique-Hoˆpitaux de Paris, Paris, France Supplemental Material

Supplemental Table 1: Characteristics and outcome of admitted patients with and without proteinuria data available at admission

Supplemental Table 2: Multivariable logistic regression analysis of outcomes odds according to UPCR ≥ 1 g/g, age, sex, BMI, hypertension, diabetes, initial serum creatinine

Supplemental Table 3: Multivariable logistic regression analysis of outcomes odds according to urine retinol-binding protein to creatinine ratio ≥ 0.3 mg/g, age, sex, BMI, hypertension, diabetes, initial serum creatinine

Supplemental Table 4: characteristics of proteinuria in patients with/without diabetes, with/without chronic kidney disease (CKD)

Supplemental Figure 1: Correlation between urine retinol-binding protein to creatinine ratio and UACR/UPCR, UACR, UPCR.

1 Supplemental Table 1 : Characteristics and outcome of admitted patients with and without proteinuria data available at admission.

Patients with Patients with NO available proteinuria available proteinuria (n=200) (n=122) Male sex 143 (71.5) 75 (61.5)

Age (years) 64±13 65±18

BMI (kg/m2) 27.5±3.9 27.0±5.1

Initial sCreat 1.11±0.41 1.29±0.89 (mg/dl) Initial eGFR 76±24 77±33 (ml/min/1.73m2) Hypertension 103(51) 45 (37)

Diabetes 50 (25) 16 (13)

CKD 18 (9) 14 (11) ESKD 0 (0) 4 (3)

ICU admission 98 (49) 19 (16)

Death 47 (23) 17 (14)

Quantitative values are given in Mean±SD, qualitative values are given as number of patients/%. BMI (Body Mass Index), CKD (Chronic Kidney Disease), ESKD (End-Stage Kidney Disease), sCreat (Serum Creatinine), eGFR (estimated Glomerular Filtration rate), ICU (Intensive Care Unit).

2 Supplemental Table 2: multivariable logistic regression analysis of outcomes odds according to UPCR ≥ 1 g/g, age, sex, BMI, hypertension, diabetes, initial serum creatinine a.

KRT ICU admission Death KRT or Death n = 27/200 n = 109/169 b n = 51/169 b n = 58/169 b OR 95%CI p-value OR 95%CI p-value OR 95%CI p-value OR 95%CI p-value UPCR>1 g/g (Y/N) 4.84 1.93- 0.001 3.66 1.72- 0.001 3.03 1.46- 0.003 4.24 2.08-8.89 < 0.001 13.5 8.20 6.39

Age 0.98 0.70- >0.9 1.00 0.74- >0.9 1.35 0.98- 0.069 1.41 1.03-1.95 0.037 (per +10 years) 1.38 1.34 1.87

Male sex (Y/N) 1.39 0.48- 0.6 1.38 0.59- 0.5 1.22 0.51- 0.2 1.00 0.42-2.42 > 0.9 4.69 3.18 3.06

BMI 1.12 1.03- 0.008 1.06 0.98- 0.2 1.10 1.02-1.20 0.015 1.22 1.15

Hypertension 0.53 0.20- 0.2 0.41 0.15- 0.078 1.04 0.36- > 0.9 0.95 0.42-2.13 0.9 (Y/N) 1.35 1.09 2.87

Diabetes (Y/N) 0.74 0.31- 0.5 0.72 0.30- 0.4 0.59 0.25-1.34 0.2 1.79 1.65

Initial sCreat 1.39 0.82- 0.2 1.46 0.79-3.3 0.3 1.21 0.73- 0.4 1.32 0.81-2.34 0.3 (per +1 mg/dl) 2.50 2.02

OR (Odds Ratio), CI (Confidence Interval), KRT (Kidney Replacement Therapy), ICU (Intensive Care Unit), UPCR (urine Protein-to-Creatinine Ratio), BMI (Body Mass Index), sCreat (serum Creatinine), Yes/No (Y/N). n = number of event / number of patients included in analysis aFor the outcome KRT, only 5 variables (UPCR, age, sex, hypertension, sCreat) were included in the multivariable model analysis due to small number of events. bComplete case analysis were performed : patients with missing data for one or more covariates were excluded.

3

Supplemental Table 3: multivariable logistic regression analysis of outcomes odds according to urine retinol-binding protein to creatinine ratio ≥ 0.3 mg/g, age, sex, BMI, hypertension, diabetes, initial serum creatinine a.

AKI ICU admission Death KRT or Death n = 30 / 72 b n = 43 / 71 b n = 23 / 85 n = 24 / 85 OR 95%CI p-value OR 95%CI p-value OR 95%CI p-value OR 95%CI p-value URBPCR>0.3 5.18 1.18- 0.041 15.5 4.35- <0.001 20.9 3.79- 0.005 5.68 1.60-24.70 0.011 (Y/N) 29.3 69.7 395

Age 1.72 0.87- 0.14 0.97 0.58- >0.9 1.11 0.73- 0.6 0.91 0.59-1.40 0.7 (per +10 years) 3.75 1.65 1.69

Male sex (Y/N) 0.70 0.11- 0.7 0.66 0.13- 0.6 0.51 0.14- 0.3 0.18 0.04-0.76 0.024 4.76 3.00 1.82

BMI (kg/m2) 1.15 0.99- 0.09 1.13 0.98- 0.11 1.37 1.33

Hypertension 1.08 0.21- >0.9 0.11 0.01- 0.063 (Y/N) 5.25 0.96

Diabetes (Y/N) 0.90 0.16- 0.9 0.94 0.20- >0.9 4.70 4.52

Initial sCreat 2.00 1.41- <0.001 2.54 0.39- 0.4 1.67 0.39- (per +1 mg/dl) 3.17 21.5 7.54 0.5 293 26.66-6647 < 0.001

OR (Odds Ratio), CI (Confidence Interval), AKI (Acute Kidney Injury), ICU (Intensive Care Unit), URBPCR (urine retinol-binding protein to creatinine ratio), BMI (Body Mass Index), sCreat (serum Creatinine), Yes/No (Y/N). n = number of event / number of patients included in analysis a Only 4 variables (URBPC, age, sex, sCreat) were included in the multivariable model analysis for the death and KRT or death outcomes, due to small number of events. bComplete case analysis were performed : patients with missing data for one or more covariates were excluded.

4 Supplemental Table 4 : characteristics of proteinuria in patients with/without diabetes, with/without chronic kidney disease (CKD).

Diabetes CKD Yes (n=51) No (n=149) Yes (n=18) No (n= 182)

UPCR (g/g) 1 0.52-1.47 0.76 0.36-1.52 1 0.65-3.63 0.83 0.37-1.49

UPCR >1 g/g (%) 50% 38% 45%% 54%%

UACR (mg/g) 248 (93-561) 101 (26-349) 135 (46-353) 110 (39-390)

UACR/UPCR (g/g) 0.23 (0.13-0.37) 0.16 (0.08-0.29) 0.39 (0.21-0.52) 0.17 (0.08-0.29)

CKD (Chronic Kidney Disease), UACR (urine albumin to creatinine ratio), UPCR (urine protein to creatinine

5

Supplemental Figure 1 : Correlations between urine retinol-binding protein to creatinine ratio (URBPCR) and urine albumin to protein ratio (UACR/UPCR) (A), urine albumin to creatinine ratio (UACR) (B), urine protein to creatinine ratio (UPCR) (C), by using quadratic (second order polynomial) regression. These analyses indicate that URBPCR is negatively correlated with albuminuria at high levels of alb/prot ratio (UACR/UPCR) or albuminuria (UACR), thus not due to glomerular proteinuria.

6