Kinetics of Plasma Cytokines During and After Two Different Modalities of Extracorporeal Blood Purication in Critically Ill COVID-19 Patients

Cytokine storm syndrome (CSS) has been documented in coronavirus disease 2019 (COVID-19) since the rst reports of this disease. In the absence of vaccines or direct therapy for COVID-19, extracorporeal blood treatment (EBT) could represent an option for the removal of cytokines and may be benecial to improve the clinical outcome of critically ill patients. Intermittent haemodialysis (IHD), using high ux (HF) or high cut-off membranes, and continuous renal replacement therapy (CRRT) could be used for blood purication in COVID-19 patients with CSS. To the best of our knowledge, cytokine kinetics during and after different types of EBT on COVID-19 patients have never been studied. In this study, we describe cytokine variation and removal during and after IHD and CRRT in COVID-19 patients with acute kidney injury (AKI). Methods: Patients with COVID-19-related AKI according to Kidney Disease Improving Global Outcomes (KDIGO) criteria and admitted at Intensive Care Unit (ICU) were studied. Blood samples were collected at the start and end of both IHD using HF membranes (10 patients) and continuous venovenous haemodialtration (CVVHDF: 10 patients) in two sessions for measuring 13 different plasma interleukins and calculating the cytokine removal rate. Results: We evaluated cytokine removal in patients with COVID-19-related AKI undergoing either prolonged IHD (10 patients) or CRRT (CVVHDF: 10 patients). There was no difference between the IHD and CVVHDF groups regarding mechanical ventilation, vasoactive drug use, age or prognostic scores. Patients treated by CRRT presented higher levels of IL-2 and IL-8 than patients treated by prolonged IHD at the start of dialysis. Cytokine removal ranged from 9–78%. Patients treated by CRRT presented higher cytokine removal rates than those treated by prolonged IHD for IL-2, IL-6 IL-8, IP-10 and TNF. The removal rates of IL-4, IL-10, IL-1β, IL-17A, IFN, MCP-1 and free active TGF-B1 were similar in the two groups. After one session of CVVHDF (24 h) the IL-2 and IL-1β levels did not vary signicantly, whereas IL-4, IL-6, IL-8, IL-10, IL-17A, TNF, IFN, IP-10, MCP-1, IL-12p70 and free active TGF-B1 decreased by 33.8–76%, and this decrease was maintained over the next 24 h. In the prolonged IHD groups, IL-2, IL-6, TNF, IP-10 and IL-1β levels did not decrease signicantly whereas IL-4, IL-8, IL-10, IL-17A, IFN, MCP-1, IL-12p70 and free active TGF-B1 decreased by 21.8–72%. However, all cytokine levels returned to their initial values after 24 h, despite their removal. during CSS-related evaluated cytokine removal in patients with COVID-19-related AKI undergoing prolonged IHD using CVVHDF. Our study showed that cytokine removal ranged from 9–78%. Patients treated by CRRT presented higher cytokine removal rates for IL-2, IL-6, IL-8, IP-10 and TNF than patients treated by prolonged IHD. The removal rates of IL-4, IL-10, IL-1β, IL-17A, IFN, MCP-1 and free active TGF-B1 were similar in the two groups.

The aim of acute RRT is to support native kidney function in controlling acid-base and electrolyte derangements, as well as uid overload, and to reduce the effects of AKI on non-renal organs, with the long-term goals of patient survival and renal recovery. Choice of the initial RRT modality is based on local availability and experience with a speci c treatment, the patient's clinical status (haemodynamic stability, catabolic state, need for removal of large amounts of uid, presence of life-threatening complications or acute brain injury) and the nancial cost of therapy.
Based on these data, the Kidney Disease Improving Global Outcomes (KDIGO) Clinical Practice Guidelines for AKI recommend considering continuous and intermittent RRT modalities as complementary techniques [5]. Hybrid treatments such as prolonged intermittent haemodialysis (IHD) incorporate the advantages of both continuous RRT (CRRT) and IHD and are used worldwide in many ICUs [6][7][8]. They may be considered for haemodynamically unstable patients in situations where other forms of CRRT are not available, but data on comparative e cacy and harm are limited [8][9][10][11][12].
Extracorporeal blood treatment (EBT) has also been proposed as an approach to remove cytokines in patients with sepsis [13] and potentially could be bene cial in critically ill patients with COVID-19 [14]. The rationale for the use of EBT is that cytokine removal could prevent organ damage due to cytokine release syndrome. Among the EBT therapies both IHD, using high ux (HF) or high cut-off membranes (characterized by larger pore size and more effective clearance for medium-molecular-weight molecules such as cytokines), and CRRT can be considered.
In this paper, we performed a prospective study that aimed to describe and compare cytokine variations and removal during and after using IHD with HF membranes and CRRT in COVID-19 patients with AKI.

Patients
A prospective cohort study of 20 hospitalized patients diagnosed with AKI related to COVID-19 undergoing RRT (IHD or HDF) was realized in the ICUs of a public and a tertiary university hospital in São Paulo, Brazil, from June to August 2020. Diagnosis of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was con rmed using reverse transcription polymerase chain reaction and the AKI diagnosis and classi cation were performed according to KDIGO criteria [6]. The need for dialysis was indicated by the gap between demand and capacity and also the presence of cytokine storm syndrome (CSS: continuous fever > 39ºC). Patients were treated with either IHD or CRRT according to local availability and the patient's clinical status. Patients with chronic kidney disease (stages IV and V), those needing kidney transplants and individuals under 18 years old were excluded.
This study was registered in the Brazilian Registry of Clinical Trials (ReBEC: number RBR-62y3h7) and approved by the Research Ethics Committee of Botucatu School of Medicine (CAAE: 30451520.6.0000.5411). All research was performed following current regulations and written informed consent was obtained from all participants or their legal guardians.

Ebt Procedures
Vascular access was obtained using 12-Fr dual-lumen catheters in an internal jugular vein.

Prolonged Ihd
Sessions lasting 6 h were performed with blood ows of 200 ml/min and dialysate ows of 300 ml/min. We used a Fresenius 4008S ARrT Plus dialysis machine (Fresenius Medical Care AG, Bad Homburg, Germany) and HF membranes (polyethersulfone lters, surface area 1.7 m 2 ; Nipro, Germany). The HF membranes were de ned according to an estimated pore size of 5 nm, resulting in an in vitro cut-off point of at least 40 kDa (at least 30 kDa for in vivo) and an ultra ltration coe cient of > 20 mmHg.
During the sessions, patients were anticoagulated with an initial 50-100 IU/kg bolus dose of heparin and then at 500-1000 IU/h over the following hours. In cases of contraindication to anticoagulation, the system was washed with 50 ml of 0.9% sodium chloride every 30 min throughout the entire procedure.
The concentrations of bicarbonate (26-36 mEq/l), potassium (1-3 mEq/l), sodium (140-145 mEq/l) and calcium (2.5-3.5 mEq/l) in the dialysis bath were adjusted according to examination and the individual needs of the patients. The ultra ltration rate did not exceed 500 ml/h and the bath temperature was 35-35.5 °C.

CRRT
CRRT was operated as continuous venovenous haemodia ltration (CVVHDF) using an AN69 membrane (ST 100, Baxter) through a Prisma ex CRRT system (Baxter). Due to the clinical conditions, blood ow rates were kept in the range 140-180 ml/min and dialysis doses were 30 ml/kg/h. Patient uid removal is tailored to the individual's volume status (100-250 ml/h) and the lter was changed every 48 h.
A regional citrate anticoagulation regimen was adopted, with anticoagulant doses adjusted individually according to the patients' coagulation state, and a sterile citrate-containing solution without calcium was infused in predilution mode. In postdilution, a sterile saline solution containing calcium chloride was infused to maintain the post lter (Ca 2+ = 0.25-0.35 mmol/l) and arterial line (Ca 2+ = 1-1.2 mEq/l).

Cytokine removal during dialysis
We measured plasma concentrations of cytokines (IL-2, IL-4, IL-6, IL-8, IL-10, IL-17A, TNF, IFN, IP-10, IL-12p70, MCP and free active TGF-B1) at the initial (T 0 ) and nal (T f ) dialysis session (6 h for prolonged IHD and 24 h for CRRT) over 2 days. Blood samples and dialysate were collected in non-heparinized tubes, immediately centrifuged at 4000 rpm for 10 min (4 °C) and then subsequently stored at −70 °C until assayed. Quanti cation of cytokines was determined by enzyme-linked immunosorbent assay (ELISA) using using LEGENDplex™ Human Essential Immune Response Panel (13-plex San Diego, CA, USA) according to the manufacturer's instructions. Assay sensitivity limits were 0.22 pg/mL for IL-2, 2.84 pg/mL for IL-4, 2.0 pg/mL for IL-6, 2.50 pg/mL for IL-8, 3.14 pg/mL for TNF-α, 1.25 pg/ mL for IL-10, 6.32 pg/mL for IP-10, 2.27 pg/mL for IL-1b, 1.8 pg/mL for MCP-1, 1.36 pg/mL for IL-17A, 3.0 pg/mL for INF, and 1.89 pg/mL for IL-12p70. All measured cytokines had molecular weight between 8 and 35 kDa We calculated cytokine removal according to the formula C f − C 0 / C f x 100, where C 0 is the cytokine concentration at the initial dialysis session and C f is the cytokine concentration 2.3 Statistical Analysis SPSS 19.0 software was used for statistical analysis. Frequency measures were calculated for categorical variables; measurement of continuous variables conforming to a normal distribution were expressed as mean ± standard deviation and those not conforming to a normal distribution were expressed as median with interquartile range. Comparison of data before and after treatment was performed by paired t-test or non-parametric test for continuous variables and by chi-squared test for categorical variables.

Results
We evaluated cytokine removal in 20 patients with COVID-19-related AKI undergoing prolonged IHD (10 patients) or CRRT (CVVHDF). There was no difference between the IHD and CVVHDF groups regarding mechanical ventilation, vasoactive drug use, age or prognostic scores (APACHE II and SOFA). On dialysis, these patients presented elevated levels of cytokines measured at the start of the dialysis session, as shown in Table 1. Patients treated by CRRT presented higher levels of IL-2 and IL-8 than patients treated by prolonged IHD at T 0 (before the start of dialysis).  Table 2 provides details of cytokine removal during the dialysis session. Cytokine removal ranged from 9-78%. Patients treated by CRRT presented higher cytokine removal rates than those treated by prolonged IHD for IL-2, IL-6, IL-8, IP-10 and TN. The removal rates of IL-4, IL-10, IL-1β, IL-17A, IFN, MCP-1 and free active TGF-B1 were similar in the two groups. After one session of CVVHDF (24 h) the IL-2 and IL-1β levels did not vary signi cantly whereas IL-4, IL-6, IL-8, IL-10, IL-17A, TNF, IFN, IP-10, MCP-1, IL-12p70 and free active TGF-B1 decreased by 33.8-76%, and this decrease was maintained over the next 24 h, as shown in Fig. 1. In the prolonged IHD groups, IL-2, IL-6, TNF, IP-10 and IL-1β did not decrease signi cantly but IL-4, IL-8, IL-10, IL-17A, IFN, MCP-1, IL-12p70 and free active TGF-B1 decreased by 21.8-72%. However, all cytokine levels returned to their initial values, despite their removal by prolonged IHD (Fig. 2).

Discussion
COVID-19 is an epidemic of global concern and previous studies have pointed out CSS as a possible cause of aggravation [16]. CSS is a systemic in ammatory reaction that releases a series of cytokines, including TNF-α, IL-1β, IL-2, IL-6, IFN-α and MCP-1. These cytokines induce immune cells to release large amounts of free radicals, which are the leading cause of multiple organ failure, including AKI [17].
EBT is considered an important means for removing cytokines from patients with sepsis and also the main advocated technology of organ support therapy for patients with severe COVID-19. Prolonged IHD (using HF or high cut-off membranes) and CVVHDF are two different modalities of EBT that have the functions of supporting renal function, adsorbing endotoxins and removing in ammatory mediators [18].
Prolonged IHD is a hybrid treatment that incorporates the advantages of both CRRT and IHD and is used worldwide in many ICUs but mainly in developing countries [6][7][8] where other forms of CRRT are not available due to the high costs. A systematic review and meta-analysis including 17 studies during 2000-2014 (7 randomized controlled trials and 10 observational studies involving 533 and 675 patients, respectively) focused on the impact of prolonged intermittent RRT and CRRT on mortality and renal recovery [12] and showed no difference in mortality between the two modalities.
We evaluated cytokine removal in patients with COVID-19-related AKI undergoing prolonged IHD using HF membranes or CVVHDF. Our study showed that cytokine removal ranged from 9-78%. Patients treated by CRRT presented higher cytokine removal rates for IL-2, IL-6, IL-8, IP-10 and TNF than patients treated by prolonged IHD. The removal rates of IL-4, IL-10, IL-1β, IL-17A, IFN, MCP-1 and free active TGF-B1 were similar in the two groups.
After one session of CVVHDF (24 h) the IL-2 and IL-1β levels did not vary signi cantly whereas IL-4, IL-6, IL-8, IL-10, IL-17A, TNF, IFN, IP-10, MCP-1, IL-12p70 and free active TGF-B1 decreased by 33.8-76%, and this decrease was maintained over the next 24 h. These ndings are in accordance with previous studies [10,19]. The application of EBT is helpful for the removal of cytokines and may be bene cial for improving the clinical outcome of critically ill patients. Commonly used EBT therapies are plasma exchange and high-dose CRRT. Previous studies have suggested that therapeutic plasma exchange and CRRT play an important role in the treatment of severe haemophagocytic syndrome and sepsis in critically ill children and adult patients with AKI by reducing the levels of cytokines (such as TNF-α and IL-6) and improving the SOFA (sequential organ failure assessment) score after the application of CRRT [19][20][21][22].
CRRT with the recently developed CytoSorb and oXiris (commercially available haemoadsorption and haemodia lters that utilize EBT) has been designed to reduce systemic cytokine burden. Numerous case reports and case series have suggested improved clinical outcomes with CytoSorb in patients with septic shock [23].
Recently, Ronco et al. [23] showed that EBT with oXiris decreased the serum IL-6 level, attenuated systemic in ammation, improved multi-organ dysfunction and reduced the expected ICU mortality rate in 37 patients with COVID-19. This shows that CRRT has become a multi-organ support rather than pure renal replacement among these patients, due to the development of new medical and membrane technology, and provides opportunity and conditions for further treatment that may improve patient outcome.
In our study, we observed that prolonged IHD using HF membranes only partially and transiently removed plasma IL-4, IL-8, IL-10, IL-17A, IFN, MCP-1, IL-12p10 and free active TGF-B1, and did not remove IL-2, IL-6, TNF, IP-10 or IL-1β. The kinetics of plasma cytokine levels after IHD in patients with COVID-19-related AKI have not been reported previously and studies on sepsis-related AKI are scarce. Mayeur et al. [24] evaluated cytokine removal in 10 patients with sepsis-related AKI treated with IHD using HF membranes (polymethymethacrylate haemodialyser). During IHD, the IL-6 level did not vary signi cantly whereas IL-8 and IL-10 were reduced by 31.8 ± 21.2% and 36.3 ± 26%, respectively. After 3 h, both IL-8 and IL-10 returned to their initial values.
The rebound phenomenon, which can be applied to the shift of soluble molecules from tissues to the intravascular compartment through a concentration gradient until a new equilibrium is reached, can explain the transient cytokine reduction. Cytokines are heavier and less diffusive molecules than urea or potassium, therefore this rebound could re ect a cytokine concentration gradient between tissue and vascular compartments that appeared during IHD. After IHD, a progressive release of cytokines from dialysis-induced hypoperfused tissue to the intravascular compartment may occur until a new equilibrium is reached.
There were several limitations to this study: it was a descriptive research study without a control group due to the speci city of the disease; the small sample size may have led to deviation of the results; we did not analyse the cytokine levels in dialysate and thus did not calculate their clearance; and we did not evaluate the association between cytokine removal and stabilization of haemodynamic status and improvement of organ function.

Conclusion
Despite these limitations, our results show that cytokine removal is lower by prolonged IHD using HF membranes than by CVVHDF. Furthermore, prolonged IHD allows a transient and selective decrease in cytokines that can be correlated with mortality during septic shock or CSS-related COVID-19.
The hypothesis that reduction of overexpressed cytokines through EBT can induce a signi cant decrease in leukocyte adhesion and plasma extravasation, which leads to improvement of the microcirculation disorder [19] and contributes to improve the haemodynamic status, needs to be explored further in larger studies.

Declarations
Ethics approval and consent to participate This study was registered in the Brazilian Registry of clinical trials (ReBEC) under number RBR-62y3h7 and was approved by the Research Ethics Committee of Botucatu School of Medicine (CAAE 30451520.6.0000.5411).