Introduction
The pandemic of Coronavirus Disease 19 (COVID-19) from the new Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), originating from China on December 2019, is still menacing with huge social and financial consequences [
1]. When severe respiratory failure becomes apparent, patients present with a complex immune dysregulation presenting features of both hyper-immune activation and immunoparalysis [
2,
3]. Elderly people and patients with comorbidities like type 2 diabetes mellitus, chronic obstructive pulmonary disease, coronary artery disease, obesity and chronic kidney disease (CKD) are considered populations at risk for unfavorable outcome [
4,
5]. The risk is more prevailing among patents at End-Stage Renal Disease (ESRD) on chronic hemodialysis (CHD) [
6]. CHD in an independent state of immune paralysis and immune activation [
7,
8] often described as “inflammaging” and this may explain the vulnerability to unfavorable outcome.
Patients at CHD present with delayed clearance of the virus and a defective response to vaccination [
9]. However, there are still several aspects of the immune response to SARS-CoV-2 in CHD which remain unclear: a) which are the cytokines and white blood subsets which are implicated in the immune response; b) can CHD modulate inflammatory mediators so as to offer protection from an exacerbated inflammatory reaction; and c) can the early immune responses define outcomes? In order to provide replies to these questions we used a study design which has not been applied so far; patients with COVID-19 on stable CHD were compared to matched comparators on CHD. Patients with severe COVID-19 already in need of mechanical ventilation (MV) or non-invasive ventilation (NIV) were excluded in order to decipher the role of early immune responses to 28-day outcomes.
Patients and methods
Study design
This is an observational prospective cohort study conducted in the Hemodialysis Units of three University Hospitals in Greece (Attikon University General Hospital, University General Hospital of Patras and AHEPA University General Hospital of Thessaloniki) between March 2021 and February 2022. The study protocol and informed consent form were approved by the Institutional Review Boards (approvals 48/24.02.2021; 161/01.04.2021 and 199/30.3.2021 respectively).
Inclusion criteria were: 1) age 18 years or more; 2) estimated glomerular filtration rate (eGFR) less than 15 ml/min/1.73m
2estimated by the CKD Epidemiology Collaboration (CKD-EPI) creatinine Eq [
10]. and at least three consecutive months on CHD; 3) infection by SARS-CoV-2 defined by positive RT-PCR of the nasopharyngeal swab; and d) hospitalization without need of MV or NIV. During the study period, all CHD patients infected by SARS-CoV-2 were hospitalized in public hospitals irrespective of COVID-19 severity since they were considered high risk patients. Hospitalization at private hospitals was not allowed.
Exclusion criteria were: 1) age less than 18 years; 2) infection by the human immunodeficiency virus; 3) absolute neutrophil count less than 1,000/mm3; 4) treatment with biological agents targeting cytokines or cell receptors during the last one month; 5) chronic intake of corticosteroids defined as oral or intravenous daily intake of prednisolone (or equivalent) dose 0.4 mg/Kg the last month. This one-month period is considered in most randomized clinical trials adequate wash out from the immunomodulating effect of any previous corticosteroid or biological treatment.
One comparator group was enrolled using the following criteria of matching: 1) age and gender; 2) at least 3 months on CHD; and 3) Charlson’s Comorbidity Index (CCI). Comparators were not infected by SARS-CoV-2 as defined by negative RT-PCR of one nasopharyngeal swab. The same exclusion criteria applied for comparators as for patients with COVID-19.
The following information was recorded: demographics; reason of primary renal disease; CCI; COVID-19 severity according to WHO; frailty index [
11]; history of vaccination and number of vaccine doses; white blood cell count, C-reactive protein, procalcitonin, ferritin, urea and creatinine on hospital admission; type of administered therapy; and WHO-CPS (World Health Organization Clinical Progression Scale) at days 14 and 28 [
12]. The WHO-CPS is an ordinal score ranging from 0 (fully recovered) to 10 (dead) reflecting the different stages of severity of COVID-19. For patients who were discharged before day 28, the WHO-CPS was captured after phone contact.
All patients were receiving standard-of-care treatment according to the WHO guidance including supplementary oxygen to maintain oxygen saturation above 93%; intravenous 6 mg dexamethasone once daily for 10 days or until discharge [
13] and low molecular weight heparin. Intake of remdesivir and antibiotics was at the discretion of the attending physicians. HD sessions were performed in an isolated room of the study hospital, regularly three times per week, or more as per patient needs. HD was done using low or high flux filters at the discretion of the attending physician and recorded. Low and high flux filters were discriminated by predefined characteristics as follows: a) ultrafiltration coefficient greater or less than 12 mL/h/mm for high and low flux respectively; or b) beta‐2 microglobulin clearance greater than 20 mL/min or less than 10 mL/min for high and low flux respectively.
Blood was collected on the first time of HD session in the study hospital and always less than 48 h after hospital admission. Eight ml of peripheral whole blood was collected immediately before the start of HD under sterile conditions. Another eight ml was collected at the end of HD. From the collected samples, five ml was poured into sterile tubes (Vacutainer, BD) and used for cytokine measurements; and another three ml was poured into one EDTA-coated tube (Vacutainer BD) and used for flow cytometry. Five ml of the HD effluent dialysate was collected 30 min after the start of the HD session into one sterile tube (Vacutainer, BD) and used for cytokine measurements.
Laboratory investigation
Blood was centrifuged and serum was stored in -80 °C until assayed. Levels of biomarkers were quantified by using commercially available kits of enzyme immunosorbent assays from Bio-Techne (R&D Systems, Minneapolis, MN) according to the manufacturer’s instructions. The lower limits of detection were: 16 pg/ml for tumor necrosis factor-alpha (TNFα); 40 pg/ml for IL-6; 31 pg/ml for IL-10; 62 pg/ml for IL-38; 156 pg/ml for IFNγ; and 313 p/ml for platelet-derived growth factor (PDGF)-A.
White blood cells were incubated for 15 minutes in the dark with the monoclonal antibodies anti-CD14 FITC, anti-HLA-DR-PE, anti-CD45 PC5 (Beckman Coulter, Marseille, France). White blood cells were also incubated for 15 minutes in the dark with anti-CD3 FITC, anti-CD4 FITC and anti-CD19 FITC (fluorescein isothiocyanate, emission 525nm, Beckman Coulter); with anti-CD4 PE, anti-CD8 PE, and anti-CD(16+56) PE (phycoerythrin, emission 575nm, Beckman Coulter); and with anti-CD45 PC5 (emission 667nm, Beckman Coulter). Fluorospheres (Beckman Coulter) were used for the determination of absolute counts. Cells were analyzed after running through the CYTOMICS FC500 flow cytometer (Beckman Coulter Co, Miami, Florida). Isotypic IgG controls stained also with anti-CD45 were used for each patient. Results of HLA-DR on CD14/CD45-cells were expressed as mean fluorescence intensity (MFI).
Study endpoints
The primary study endpoint were the differences between cytokines and white blood cell subsets between patients on CHD and COVID-19 and comparators on CHD.
The secondary study endpoint was to identify the best predictor of the 28-day outcome. Pre-defined outcomes were WHO-CPS of 6 or more (interpreted as hospitalized with severe disease or dead) at day 28 and WHO-CPS 3 or less at day 28 (interpreted as ambulatory disease).
Statistical analysis
The calculation of the study power was done with the assumption that at least two of the measured variables would differ statistically between patients and comparators at a p value less than 0.0001 and that the true difference would be 150 with 210 standard deviation. To demonstrate this difference with 80% power at the 10% level of significance with patient: comparator ratio 3:1, 54 experimental subjects and 16 control subjects were needed.
Results are presented as medians and distribution. Comparisons between patients and comparators were done by the Mann Whitney U test. Paired comparisons before HD and at the end of HD were done by the Wilcoxon ranked sum test. Patients were split into subgroups based on their outcome at day 28. The subgroups were: those with WHO-CPS ≥ 6 and WHO-CPS < 6; and those with WHO-CPS ≤ 3 and WHO-CPS > 3. Comparison of cytokines and white blood cell subsets between subgroups were done by the Mann–Whitney U test. The cytokine or cell subset for which significant differences were found was analyzed through ROC (receiver operator characteristics) curve to define a cut-off which defines outcome. The cut-off was found after applying the Youden index. The sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of the cut-off for the outcome was calculated; the odds ratio (OR) and 95% confidence intervals were defined by the Mantel–Haenszel statistics. Any value of p less than 0.05 was considered significant.
Discussion
The current study showed that patients under CHD hospitalized for moderate COVID-19 have increased circulating levels of TNFα, IFNγ and PDGF-A, decreased CD14-lymphocytes, decreased CD19-lymphocytes and decreased expression of HLA-DR on CD14-monocytes. The filter of HD did not change the circulating concentrations of the measured mediators. The main predictors of 28-day outcome, as this is expressed by the WHO-CPS, were CD19-subsets and the expression of HLA-DR. More precisely, patients with defective expression of HLA-DR on CD14-monocytes were prone to unfavorable outcome with WHO-CPS 6 or more i.e. severe disease or dead. On the contrary, patients with more than 40 CD19-lymphocytes/mm3 were at greater likelihood for favorable outcomes with WHO-CPS 3 or less i.e. discharge home with or without symptoms.
Existing observational studies report increased mortality of patients under CHD from COVID-19 [
14,
15], although some studies suggest that dialysis patients are rather protected from the severe form of COVID-19 [
16] probably because they are unable to develop pro-inflammatory responses similar to the non-CHD patients [
17]. Our results are in general agreement with the results of another study of 32 patients with COVID-19 on CHD [
18]. However, that study did not report on the expression of HLA-DR on CD14-monocytes whereas no measurements were done before and at the end of the HD session.
Patients at CHD present immmunosenescence expresse through lymphopenia and low-grade chronic inflammation [
18]. Indeed, following HD CD8-lymphocytes, NK-cells, NKT-cells, CD19-lymphocytes and the expression of HLA-DR on CD14-monocytes were increased in the comparator group. Our findings suggest that when patients are acutely infected by SARS-CoV-2 they present further attenuation of T-cell and B-cell responses and defective HLA-DR presentation on monocytes. In these patients the HD filter could not absorb circulating cytokines. This probably implies that the lack of retention of cytokines by the HD filter does not influence outcome. However, controversial findings are published on the retention of circulating cytokines by the HD filter in COVID-19. In an observational study of 60 patients, there were no differences regarding the need for admission in the intensive care unit and the incidence of death between patients under low-flux HD and patients under medium flux HD [
19]. Special emphasis should be given to three studies. In the first study 74 patients under CHD infected by SARS-CoV-2 were enrolled. The use of filters by asymmetric cellulose triacetate (ATA) was compared to filters by polymethylmethacrylate (PMMA). Using a multivariate logistic regression model, the authors concluded that ATA favoured the retention of IL-6 [
20]. In the second small-scale randomized trial, 15 patients underwent HD using medium cut-off (MCO) membranes and were compared to 14 patients who underwent HD using filters by PMMA. Results showed that following the session of HD using MCO, circulating IL-8 and IL-10 decreased. IL-8 was also decreased after the HD session using PMMA [
21]. In the third study, nine patients under CHD with COVID-19 were subject to HD, using MCO filter. IL-6 levels before and after dialysis were not uniform in all patients. However, IL-6 was increased between serial HD session among non-survivors and overall circulating IL-6 was higher in non-survivors [
22].
One fully novel finding was the increase of IL-10 in the effluent dialysate of HD. This can be partly explained by the greater anti-inflammatory burden in patients under CHD. Although considered a major anti- inflammatory cytokine, early rise of IL-10 is associated and possibly involved in severe COVID-19 [
23]. Similar increases were shown for PDGF-A levels. Increased levels of PDGF-A have been described in moderate COVID-19 and showed inverse correlation with circulating D-dimers [
24]. PDGF-A is a marker of endothelial dysfunction and COVID-19 is a state of endotheliitis [
25]. The overall favourable outcome of the patients generates the question if this increase is protective from immune exhaustion.
Despite the enthusiasm generated at the early stages of the pandemic that COVID-19 is an IL-6-driven disease, it soon became evident that not all severe patients have increased IL-6 at hospital admission [
17,
26]. This seems to be the case of patients under CHD. Another feature of the immune response in these patients is the decrease of the expression of HLA-DR on monocytes. This denotes defect of antigen-presentation which contributes to the immune dysregulation of COVID-19 hallmarked by B-cell lymphopenia and defective antibody production. On the contrary, increased B-lymphocyte counts in infected patients under CHD are associated with better outcomes.
The major strengths of the present study are the participation of patients from three different departments and the study design providing cytokine measurements before and at the end of the HD session and from the dialysate. Several limitations of the study need to be acknowledged. These limitations are: i) the small number of patients; ii) the lack of serial day measurements; iii) the inclusion of comparators vaccinated with more vaccine doses than patients. Medical history of vaccination may induce false positive increase of cytokine levels and white blood cell subpopulations. Since vaccine dosing was more intense in the comparators than the patients, this false positive effect if limited; iv) the lack of ferritin data of patients before they were infected with SARS-CoV-2. This does not allow to fully address COVID-19 as the single reason of hyper-ferritinemia. However, increase of ferritin is so common among non-CHD patients [
2,
3] that SARS-CoV-2 infection may be considered the reason of ferritin increase with relative certainty; and v) the lack of information on viral variants.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.