Kidney Outcomes in Critically Ill Patients With and Without COVID-19

Background: Case series have described high rates of AKI in critically ill persons with covid-19. However, no study has directly compared kidney outcomes in similarly ill persons with and without covid-19 that are contemporaneously enrolled. Method(s): We assessed 346 participants from a study of covid-19 in critical illness enrolled from University of Washington from April 2020 to May 2021. Patients in the ICU were recruited if symptoms of lower respiratory tract infection prompted covid testing. 2/3 of the cohort were covid positive;the remaining 1/3 had another cause of respiratory illness and served as controls. We defined major adverse kidney events (MAKE) as doubling of serum creatinine, dialysis, or death during hospitalization. Among 186 patients with available urine samples, we also assessed kidney injury molecule-1 (KIM-1), epidermal growth factor (EGF), and Cr. We used inverse probability of treatment weighting with propensity scores to increase similarity between the comparison groups. Result(s): Mean age was 55 years;64% were male, and mean admission serum Cr was 1.3 +/- 1.0 mg/dL. Baseline characteristics, including APACHE III and SOFA scores, were similar between groups after propensity weighting. Among Covid-19 patients the incidence of MAKE was 45% and in non-covid-19 patients was 26%. Covid-19 positivity was associated with a 55% greater incidence of MAKE and each component of MAKE was numerically higher in the covid-19 positive group (figure). The covid positive group had lower urine EGF levels (indexed to urine Cr) over time. Urine KIM-1 levels were similar. Conclusion(s): The incidences of clinical kidney outcomes are numerically higher in critically ill patients with covid-19 compared with similarly ill control patients without covid-19. Urinary concentrations of EGF are lower in covid positive patients.

In ICU Patients With COVID-19, Estimating Glomerular Filtration Rate From Creatinine and Cystatin C

Background: Accurate estimation of the glomerular filtration rate (eGFR) in critical illness is essential for judging the severity of kidney injury and informing medication dosing. Many drugs used to treat COVID-19 require dose adjustment (baricitinib) or are withheld (remdesivir) based on kidney function. Cystatin C provides more accurate and precise estimation of GFR than serum creatinine in outpatient settings, but the relationship of creatinine and cystatin C in critically ill patients with and without COVID-19 is less known. Method(s): We prospectively enrolled 253 ICU patients, including 176 (70%) patients with COVID-19 and 77 (30%) patients without COVID-19. We collected plasma on days 1, 3, 7, 10 and 14 leading to a total of 643 samples. Plasma creatinine was measured using the modified Jaffe method and plasma cystatin C was measured using an immunoturbidimetric assay (Gentian AS) on a Beckman DXC Unicell clinical analyzer. Result(s): Among 486 plasma samples in 176 unique COVID-19 patients, plasma cystatin C and creatinine were correlated (r=0.79) but calculated eGFR differed. Values of eGFRcr were, on average, 16-41 ml/min/1.73m2 higher than those of eGFRcys during hospitalization (p <0.001) (Figure 1). Each 15 ml/min/1.73m2 lower eGFRcys was associated with an estimated 11% greater risk of hospital mortality (RR =1.11, 95% CI, 1.03 to 1.21) compared to an estimated 6% greater risk of hospital mortality with eGFRcr (RR = 1.06;95% CI: 0.97 to 1.16). Among non-COVID-19 patients there was a similar trend with higher eGFRcr than eGFRcys but the difference on ICU admission was nonsignificant. Conclusion(s): In COVID-19 ICU patients eGFRcr was consistently higher than eGFRcys, and eGFRcys may more strongly associate with clinical outcomes. Our findings suggest that in COVID-19, calculating eGFR using creatinine or cystatin C could have implications on which treatments are available to patients.

Identifying ARDS using the Hierarchical Attention Network with Sentence Objectives Framework

Acute respiratory distress syndrome (ARDS) is a life-threatening condition that is often undiagnosed or diagnosed late. ARDS is especially prominent in those infected with COVID-19. We explore the automatic identification of ARDS indicators and confounding factors in free-text chest radiograph reports. We present a new annotated corpus of chest radiograph reports and introduce the Hierarchical Attention Network with Sentence

Rates of aki in ICU patients admitted under suspicion for COVID-19 do not differ by SARS-CoV 2 status or SARS-CoV-2 genomic load

Rationale: Since the onset of coronavirus disease 2019 (COVID-19), the disease caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), studies have suggested a high incidence of acute kidney injury (AKI) among patients with COVID-19. However, these studies lack contemporaneously enrolled critically ill patients to understand whether high rates of AKI are unique to COVID-19. It is also unknown whether the risk of AKI is related to SARS-CoV-2 genomic load. Methods: We prospectively enrolled a cohort of patients admitted to the ICU with suspicion of COVID-19 (persons under investigation) from April to September 2020. Of these patients, 78 (46%) tested positive for SARS-CoV-2 (COVID-19) and 91 (54%) tested negative (non-COVID-19). AKI was defined as an increase ≥ 0.3 mg/dL in 48 hours or ≥ 50% increase in serum creatinine (sCr) measured during hospitalization compared to a 'baseline' sCr measured at study enrollment. New dialysis was defined as initiation of dialysis during hospitalization. SARS-CoV-2 qRT-PCR was performed across four different platforms with comparable cycle threshold (Ct) values. Ct values were a semiquantitative measure of genomic load with an inverse relationship of Ct to genomic load. We used relative risk regression to determine if there was an increased risk of AKI in COVID-19 compared to non-COVID-19 and whether SARS-CoV-2 genomic load was associated with AKI. Analyses were adjusted for age, sex, body mass index, and APACHE III scores. Results: Rates of AKI and new dialysis were similar in COVID-19 compared to non-COVID-19 (AKI: n=23 (29%) vs n=24 (26%) and Dialysis: n=8 (10%) vs n=6 (6%). Unadjusted and adjusted analyses demonstrated a non-significant difference in risk of AKI (adjusted RR = 1.04 (95% CI: 0.65-1.66) or new dialysis (adjusted RR = 1.55 (95% CI 0.58-4.12) in COVID-19 compared to non-COVID-19. We had Ct values available prior to ICU admission in 47 patients. In unadjusted and adjusted analyses, a 10-unit decrement in Ct values was not associated with AKI (adjusted RR = 0.40 (95% CI: 0.14-1.14) or new dialysis (adjusted RR = 0.96 (95% CI: 0.23-2.69) (Figure 1). Conclusions: Our study demonstrates that rates of AKI and new dialysis in ICU patients with COVID-19 are similar to rates in non-COVID-19 ICU patients. Moreover, the lack of association between Ct values and AKI in COVID-19, suggests that immune and host response to SARS-CoV-2 may contribute more to risk of AKI in ICU patients rather than the pathogen itself.