Introduction
Acute kidney injury (AKI) in patients hospitalised with COVID-19 is common, and is associated with worse prognosis, especially among critically unwell patients. The aim of this study was to investigate the epidemiology, risk factors and impact of AKI on patients hospitalised with COVID-19 in a large UK tertiary centre.
Methods
We retrospectively collected and analysed data from electronic health records of all adult patients admitted with a clinical and laboratory-confirmed diagnosis of COVID-19 across both hospital sites of large UK tertiary centre from 1st March to 13th May 2020, during the first wave of the COVID-19 pandemic in the UK. Those with pre-existing end-stage kidney disease or possible hospital-acquired COVID-19 were excluded. Incidents of AKI were identified using KDIGO criteria and weekly incidence rates were calculated. We developed a logistic regression model to identify predictors of AKI, and a Cox regression model to investigate the impact of AKI on mortality.
Results
1248 inpatients were included, with a mean age of 69 years and male preponderance of 58.8%. Chronic kidney disease (CKD) defined as eGFR < 60 ml/min/1.73m2 was present prior to admission in 16.6%. AKI occurred in 39% (n=487) patients, including stage 1 51% (n= 248), stage 2 13% (n=64) and stage 3 36% (n=175). 109 patients (8.7% of total, 22% of all AKI) required renal-replacement therapy (RRT). Of those who developed AKI and were discharged alive, 219 (84.6%) had recovered renal function to baseline creatinine by time of discharge and none required on-going RRT. The incidence rate of AKI increased on a weekly basis initially, peaking at 2.19 per 100 person-days at weeks 5 to 6, before reducing at a similar pace to 1.41 per 100 person-days by the end the study period (week 9 to 10).
On multivariable logistic regression, male sex (OR 1.55, P<0.001), black ethnicity (OR 1.79, P<0.0005), baseline eGFR<60 ml/min/1.73m2 (OR 3.06, P<0.0001), hypertension (OR 1.73, P<0.0005) and in-patient diuretic usage (OR 1.69, P<0.005) were significantly associated with AKI. Of those with AKI, 42.1% (n=205) died during admission, compared to 16.8% (n=128) without AKI. AKI was a strong predictor of 30 day mortality (see Figure), even at stage 1 with HR 1.65 (P<0.001), which increased further by stage to HR 3.1 (P<0.0001) for AKI stage 2 and HR 3.89 (P<0.0001) for stage 3, adjusting for other variables by multivariable Cox regression.
Conclusions
AKI is common among patients hospitalised with COVID-19 and is associated with increased mortality, even at early stages. Inpatient diuretic use was significantly associated with AKI, and the reducing incidence of AKI over time following a peak may reflect a change in clinical practice as the pandemic progressed. Of those that survived, most recovered their renal function and none required ongoing RRT at discharge.
No conflict of interest