Acute Kidney Injury in a National Cohort of Hospitalized US Veterans with COVID-19

Abstract

Background and objectives: Coronavirus disease 2019 (COVID-19) is associated with higher risk of AKI. We aimed to describe rates and characterize predictors and health outcomes associated with AKI in a national cohort of US veterans hospitalized with COVID-19.

Design, setting, participants, & measurements: In a cohort of 5216 US veterans hospitalized with COVID-19 identified through July 23, 2020, we described changes in serum creatinine and examined predictors of AKI and the associations between AKI, health resource utilization, and death, utilizing logistic regressions. We characterized geographic and temporal variations in AKI rates and estimated variance explained by key variables utilizing Poisson regressions.

Results: In total, 1655 (32%) participants had AKI; 961 (58%), 223 (13%), and 270 (16%) met Kidney Disease Improving Global Outcomes definitions of stage 1, 2, and 3 AKI, respectively, and 201 (12%) received KRT. Eight percent of participants had AKI within 1 day of hospitalization, and 47% did not recover to baseline serum creatinine by discharge. Older age, Black race, male gender, obesity, diabetes, hypertension, and lower eGFR were significant predictors of AKI during hospitalization with COVID-19. AKI was associated with higher mechanical ventilation use (odds ratio, 6.46; 95% confidence interval, 5.52 to 7.57) and longer hospital stay (5.56 additional days; 95% confidence interval, 4.78 to 6.34). AKI was also associated with higher risk of death (odds ratio, 6.71; 95% confidence interval, 5.62 to 8.04); this association was stronger in Blacks (P value of interaction <0.001). Hospital-level rates of AKI exhibited substantial geographic variability, ranging from 10% to 56%. Between March and July 2020, AKI rates declined from 40% to 27%; proportions of AKI stage 3 and AKI requiring KRT decreased from 44% to 17%. Both geographic and temporal variabilities were predominately explained by percentages of Blacks (31% and 49%, respectively).

Conclusions: AKI is common during hospitalization with COVID-19 and associated with higher risk of health care resource utilization and death. Nearly half of patients with AKI did not recover to baseline by discharge. Substantial geographic variation and temporal decline in rates and severity of AKI were observed.

Podcast: This article contains a podcast at https://www.asn-online.org/media/podcast/CJASN/2020_11_16_CJN09610620_final.mp3.

Keywords: Black race; COVID-19; Length of stay; acute kidney injury; diabetes; kidney function; mortality; obesity; racial disparities; respiratory failure.

Graphical abstract

Figure 1.

Trajectories of serum creatinine among a cohort of US veterans hospitalized with coronavirus disease 2019 who were discharged alive. (A) Fourteen and (B) 30 days. Baseline serum creatinine is included at time −1. Trajectories are colored by AKI status and severity. In those who received KRT, the serum creatinine trajectory was censored at the time of KRT. Bands represent the 95% confidence intervals.

Figure 2.

Risk of death by AKI status and by stage. (A) Association between AKI status (AKI versus no AKI) and death overall and effect modification of the association of AKI with mortality by effect modifier category. The P value indicates evidence of an interaction on the risk of mortality between AKI and no AKI (reference). (B) Association between AKI and death by AKI severity. Adjusted for age, race, sex, body mass index (BMI), smoking status, hypertension, diabetes, cardiovascular disease, angiotensin-converting enzyme inhibitor, angiotensin II receptor blocker, diuretics, anticoagulant, immunosuppressants, β-blockers, aspirin, and eGFR category. ADI, Area Deprivation Index; 95% CI, 95% confidence interval; OR, odds ratio.

Figure 3.

Map of the geographic distribution of participants hospitalized with coronavirus disease 2019 (COVID-19) in Veterans Affairs (VA) facilities. The size of the circle represents the absolute number of participants hospitalized with COVID-19 in a VA hospital system. The color gradient represents the rates of AKI among those hospitalized for COVID-19, expressed as a percentage. The histogram shows the distribution of rates across hospital systems, and descriptive statistics for this distribution are presented. Hospital systems with <20 participants with COVID-19 are excluded from the histogram. IQR, interquartile range.

Figure 4.

Rates of AKI in COVID-19 hospitalized veterans by calendar month. (A) Rates of AKI in COVID-19 hospitalizations. (B) Rates of AKI in COVID-19 hospitalizations by AKI stage. (C) Proportions of AKI stage among all patients with AKI during a COVID-19 hospitalization. Error bars represent 95% CIs.