The incidence, risk factors and prognosis of acute kidney injury in severe and critically ill patients with COVID-19 in mainland China: a retrospective study

Background The clinical correlates, prognosis and determinants of acute kidney injury (AKI) in patients with coronavirus disease 2019 (Covid-19) remain largely unclear. Methods We retrospectively reviewed medical records of all adult patients with laboratory-confirmed Covid-19 who were admitted to the intensive care unit (ICU) between January 23rd 2020 and April 6th 2020 at Wuhan JinYinTan Hospital and The First Affiliated Hospital of Guangzhou Medical University. Results Among 210 patients, 131 were males (62.4%). The median Age was 64 years (IQR: 56–71). Of 92 (43.8%) patients who developed AKI during hospitalization, 13 (14.1%), 15 (16.3%) and 64 (69.6%) were classified as being at stage 1, 2 and 3, respectively. 54 patients (58.7%) received continuous renal replacement therapy. Age, sepsis, nephrotoxic drug, invasive mechanical ventilation and elevated baseline serum creatinine levels were associated with the occurrence of AKI. Renal recovery during hospitalization was identified among 16 patients with AKI (17.4%), who had a significantly shorter time from admission to AKI diagnosis, lower incidence of right heart failure and higher ratio of partial pressure of oxygen to the fraction of inspired oxygen. Of 210 patients, 93 deceased within 28 days of ICU admission. AKI stage 3, critical disease, greater Age and the lowest ratio of partial pressure of oxygen to the fraction of inspired oxygen being < 150 mmHg were independently associated with death. Conclusions Among patients with Covid-19, the incidence of AKI was high. Our findings of the risk factors of the development of AKI and factors associated with renal function recovery may inform clinical management of patients with critical illness of Covid-19.


Background
Since December 2019, the outbreak of coronavirus disease 2019  has resulted in over 30.6 million laboratory-confirmed cases and 950,000 death cases worldwide as of September 20th 2020 [1]. The clinical manifestations of Covid-19 have been heterogeneous, ranging from asymptomatic viral carriers to critically ill cases [2,3]. The clinical outcomes of critically ill patients have been poor, with the mortality rate being 61.5% within 28 days [4]. Identification of the risk factors of death would be crucial to inform clinical decisions to guide early triage of patients for more intensive monitoring. Previous studies have documented that greater age and the presence of comorbidities correlated significantly with death in patients with Covid-19 [5]. Few studies, however, have focused on the systemic complications and the clinical outcomes of Covid-19.
Acute kidney injury (AKI) is a heterogeneous disease and a common complication in critically ill patients. AKI reportedly correlated with poor clinical outcomes in patients with Covid-19 [2-4, 6, 7]. Nonetheless, the clinical correlates, prognosis and determinants of AKI in patients with Covid-19 remain largely unclear. Early identification and intervention would help improve the clinical outcomes of patients with Covid-19 who were prone to develop AKI [8,9]. Therefore, we performed a retrospective study to evaluate the incidence, risk factors and prognosis of AKI in severe and critically ill patients with Covid-19.

Patients
We reviewed medical records of all adult patients (> 18 years) with laboratory-confirmed Covid-19 who were admitted to the intensive care unit (ICU) between January 23rd 2020 and April 6th 2020 at Wuhan JinYinTan Hospital and The First Affiliated Hospital of Guangzhou Medical University (designated hospitals for admitting patients with . The severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) RNA was detected by using reversetranscription polymerase chain reaction (RT-PCR) for laboratory diagnosis of Covid-19. The samples consisted of throat swabs, nasal swabs or endotracheal aspirate. All patients met the criteria of having severe or critical Covid-19 according to the Chinese guidelines [10]. The Ethics Committees of both hospitals approved the study protocol. Informed consent was waived due to the nature of the retrospective study.

Data extraction
The clinical data from 210 cases, including patient demographics, clinical symptoms and signs, laboratory findings, treatment [including respiratory supports, use of medications and continuous renal replacement therapy (CRRT)] and clinical outcomes, were extracted from the electronic records by two independent intensivists (LS and XZ) who subsequently crosschecked for data accuracy. The disagreement was further adjudicated by a third independent reviewer (YML). All data were entered into the computerized database for further statistical analyses.
The definition, severity staging and clinical management of AKI were based on the Kidney Disease: Improving Global Outcomes (KDIGO) classification [9]. The clinical management could also be amended as per the clinical needs at the discretion of the attending clinicians [11,12]. Recovery of AKI was defined based on Acute Disease Quality Initiative (ADQI) 16 Workgroup: (1) rapid sustained reversal: recovery from AKI within 48 h, (2) late sustained reversal: reversal after 48 h and sustained through 28 days after AKI diagnosis or hospital discharge, (3) relapsing AKI with complete recovery, (4) relapsing AKI without complete recovery, and (5) never recovered. The first three categories were classified as complete renal recovery while the last two categories were classified as renal non-recovery [13]. The recovery of AKI was monitored until patient's discharge from hospital.

Study outcomes
The study outcomes mainly consisted of the incidence of AKI and the use of CRRT. We also assessed the risk factors for AKI, the outcomes of renal diseases, and the impact of AKI on the clinical outcomes of Covid-19.

Statistical analysis
Continuous variables were expressed as the median and interquartile ranges (IQRs) and compared with Wilcoxon rank-sum test. Categorical variables were expressed as counts and percentages, and compared using chi-square test or Fisher's exact test as appropriate. In clinical practice, the elevation in serum creatinine levels preceded the development of AKI. Therefore, the risk factors for AKI were screened with univariate logistic regression model, and the variables with P value of 0.10 or less were considered as the potential risk factors and further imported into the multivariate logistic regression analysis. The AKI and survivor risk model was established by calculating the regression coefficient (β), The most abnormal data during hospitalization were presented among patients who did not develop AKI. PaCO 2 : arterial partial pressure of carbon dioxide; PaO 2 /FiO 2 ratio: the ratio between the arterial partial pressure of oxygen and the inspiratory concentration of oxygen; Data in bold indicated the comparisons with statistical significance odds ratio (OR) and 95% confidence interval. The significance threshold was set at a 2-sided P value of less than 0.05. All statistical analyses were performed using R version 3.5.1 software (the R Foundation, USA) or SPSS version 17.0 (SPSS Inc., USA).

Patient characteristics stratified by the severity of Covid-19
We have initially included 262 patients, and excluded 52 patients (20 patients discharged or died within 24 h, and 32 patients had missing data). Finally, we included 210 patients with complete data, of whom 131 were males (62.4%). The median age was 64 years (IQR: 56-71) ( Table 1). None of the patients had AKI at admission. The patient characteristics as stratified by the disease severity are shown in Table E1. 60 patients were categorized as having severe disease and the remaining 150 patients as having critical disease upon admission to ICU. The gender and the age were comparable. Diabetes was more common in critical cases than in severe cases (24.7% vs 11.7%, P = 0.037). There were, however, no significant differences for other comorbidities between severe and critical cases. Overall, the most common respiratory support for severe cases consisted of conventional oxygen therapy (COT) and high-flow nasal cannula (HFNC), A significantly larger proportion of patients with critical disease were initiated with invasive mechanical ventilation(IMV). There was, however, no significant difference in the use of non-invasive mechanical ventilation (NIV) between severe and critical cases. (Table E1).
Of the 210 patients, 194 were from Wuhan JinYinTan Hospital, and 16 patients were from The First Affiliated Hospital of Guangzhou Medical University. Right heart failure, hypoxemia, hypercapnia and higher levels of interleukin-6 were more common in patients from Wuhan JinYinTan Hospital (both P < 0.05). (Table E2).

Discussion
In this study, we sought to investigate the association between AKI and the clinical outcomes in severe and critical patients with  [14]. However, a recent study which enrolled 634 patients with ARDS demonstrated that an incidence rate of 68.3% for AKI after the onset of ARDS [15]. In another observational study on the critically ill patients with 2009 influenza A (H1N1) viral pneumonia, the incidence of AKI reached to 51% [16]. Our finding of the higher incidence of AKI (43.8%) in patients with Covid-19 therefore might have been attributable to the high incidence of ARDS during hospitalization.
Previous studies had demonstrated that greater age, sepsis and baseline serum creatinine levels was associated with the development of AKI in adult patients [10,15,[17][18][19][20][21], which was consistent with our results. Recently, the association between IMV and AKI in ARDS patients have become widely acceptable. Several studies had postulated that IMV might have played a key role in the development of AKI due to the mechanisms of lung-kidney cross-talk [22,23], and lung protective ventilation strategy including low tidal volume ventilation, limited plateau pressure, and suitable positive end-expiratory pressure level was postulated to decrease the risk of AKI [24]. However, hypercapnia was very common in our patients, especially the patients with AKI. It was challenging to correct for the hypercapnia even the patients had been intubated and receiving lung protective ventilation strategy. Physicians may adjust the tidal volume and plateau pressure in a very constrained environment, especially at the early stage of epidemic. This might explain why IMV was the strongest risk factors associated with AKI.
In our study, only 16 AKI patients (17.4%) achieved recovery of renal function during hospitalization, which was significantly lower than that reported previously (39.3%) [25]. Our analysis revealed that the delayed onset of AKI, the presence of right heart failure and lower PaO 2 /FiO 2 ratio were associated with an increased likelihood of having non-recovery of renal function. Patients with a delayed onset AKI might have a prolonged course of hospital management, including hemodynamic deterioration, use of nephrotoxic drugs and the adverse impacts associated with lung-renal cross-talk during the course of ARDS development [18,19]. Severe hypoxemia had been considered as a cause of chronic kidney disease and AKI [26][27][28].The severity and duration of AKI might depend on the severity and duration of hypoxemia. In our study, the overall incidence of right heart failure was similar with that reported in a previous study [29]. Persistent right heart failure would lead to elevated central venous pressure and peripheral edema, which were risk factors of AKI in critically illness [30].
Several previous studies had demonstrated that AKI, especially severe AKI was associated with poorer clinical outcomes [15,18,31]. The AKI-EPI study showed that patients with stage 1 AKI did not have a higher risk of mortality compared to patients without AKI [32]. However, in our study, only AKI KDIGO stage 3 was associated with ICU 28-day mortality, patients with KDIGO stage 1 and 2 AKI did not have a higher mortality compared to patients without AKI.
(See figure on previous page.) Fig. 1 Association between the minimal PaO 2 /FiO 2 ratio, time from admission to AKI diagnosis, right heart failure and the outcomes of renal disorders in patients with Covid-19 who had acute kidney injury during hospitalization. a. Association between the minimal PaO 2 /FiO 2 ratio, time from admission to AKI diagnosis and the outcomes of renal disorder and the severity of acute kidney injury in patients with Covid-19 during hospitalization. b. Association between the minimal PaO 2 /FiO 2 ratio, time from admission to AKI diagnosis and the outcomes of renal disorders and the presence of right heart failure in patients with Covid-19 who had acute kidney injury during hospitalization. KDIGO: Kidney Disease: Improving Global Outcomes. Min P/F: Minimal PaO 2 /FiO 2 ratio To our knowledge, this is the first and largest study which investigates the incidence, risk factors of AKI, and the association between AKI and clinical outcomes in severe and critical patients with Covid-19. However, we acknowledge a number of limitations of the current study, which are in part due to the nature of retrospective study. We did not collect the ventilator parameters which might play a key role in lung-kidney crosstalk. We also did not assess the fluid balance per day due to the very constrained environment. Therefore, findings of our study needed further validation.

Conclusion
In conclusion, among patients with Covid-19, the incidence of AKI was high. Age, sepsis, nephrotoxic drug, invasive mechanical ventilation and baseline serum creatinine were strongly associated with the development of AKI. Time from admission to AKI diagnosis, right heart failure and PaO 2 /FiO 2 ratio were independently associated with the potential of renal recovery. Finally, AKI KIDGO stage 3 independently predicted the risk of death within 28 days of ICU admission. Our findings of the risk factors of the development of AKI and factors associated with renal function recovery may inform clinical management of patients with critical illness of Covid-19.
Additional file 1: Table E1. Clinical characteristics of patients with Covid-19 when stratified by the disease severity. Table E2. Patient's characteristics between the two hospitals.   Fig. 2 The proportion of patients with Covid-19 when stratified according to the survival status and the staging of the acute kidney injury. AKI: Acute kidney injury; G: stage