Objectives: COVID-19 is a great threat to the modern world and significant threat to immunocompromised patients, including patients with chronic renal failure. We evaluated COVID-19 incidence among our hemodialysis patients and investigated the most probable immune mechanisms against COVID-19.
Materials and Methods: Başkent University has 21 dialysis centers across Turkey, with 2420 patients on hemodialysis and 30 on peritoneal dialysis. Among these, we retrospectively evaluated 602 patients (257 female/345 male) with chronic renal failure receiving hemodialysis as renal replacement therapy; 7 patients (1.1%) were infected with SARS-CoV-2. We retrospectively collected patient demographic characteristics, clinical data, and immunological factors affecting the clinical course of the disease. We divided patients into groups and included 2 control groups (individuals with normal renal functions): group I included COVID-19-positive patients with normal renal function, group II included COVID-19-positive hemodialysis patients, group III included COVID-19-negative hemodialysis patients, and group IV included COVID-19-negative patients with normal renal function. Lymphocyte subsets in peripheral blood and typing of human leukocyte antigens were analyzed in all groups, with killer cell immunoglobulin-like receptor genes analyzed only in COVID-19-positive patients and healthy controls.
Results: No deaths occurred among the 7 COVID-19-positive hemodialysis patients. Group I patients were significantly older than patients in groups II and III (P = .039, P = .030, respectively) but not significantly different from group IV (P = .060). Absolute counts of natural killer cells in healthy controls were higher than in other groups (but not significantly). Activated T cells were significantly increased in both COVID-19-positive groups versus COVID-19-negative groups. Groups showed significant differences in C and DQ loci with respect to distribution of alleles in both HLA classes.
Conclusions: Although immunocompromised patients are at greater risk for COVID-19, we found lower COVID-19 incidence in our hemodialysis patients, which should be further investigated in in vitro and molecular studies.
Key words : Chronic renal failure, Pandemic, SARS-CoV-2
The 2019 novel coronavirus disease (COVID-19) outbreak became a great threat to the modern world. This insidious and highly contagious novel coronavirus is prone to cause clusters of outbreaks. Quarantine and social distancing continue to be the basic principles to prevent the rapid spread. As more information is collected about the pandemic, geographical differences of incidence rate and severity of the disease have become evident. Certain regions have lower incidence and mortality rates. COVID-19 is also reported to be a significant threat for immunocompromised patients, including those with chronic renal failure. However, epidemiologic data for this group of patients are limited.1 In this study, we evaluated the incidence of COVID-19 among our patients on hemodialysis (HD) and investigated the most probable immune mechanisms against COVID-19.
Materials and Methods
Başkent University has 21 dialysis centers across Turkey, with 2420 patients on HD and 30 patients on peritoneal dialysis. Among these, we retrospectively evaluated 602 patients (257 female, 345 male) with chronic renal failure who were receiving HD as renal replacement therapy in our center. The demographic and clinical data (comorbid diseases, medications, and vaccination history for influenza, tuberculosis, and hepatitis A virus) of HD patients were retrospectively collected. Immunologic factors that affect the clinical course of the disease were also investigated. Among the patients who were receiving HD, 2 study groups were designed; we also included 2 control groups (that consisted of individuals with normal renal function; non-HD) for comparison. Study groups were as follows: group I included COVID-19-positive non-HD patients, group II included COVID-19-positive HD patients, group III included COVID-19-negative HD patients, and group IV included COVID-19-negative non-HD patients. The diagnosis of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was confirmed with a positive polymerase chain reaction (PCR) test, which was identified from throat swab samples and/or the characteristic appearance in thoracic high-resolution computed tomography scans.
Lymphocyte subsets in peripheral blood and HLA typing were analyzed in all study groups; however, killer cell immunoglobulin-like receptor (KIR) genes were only analyzed in COVID-19-positive patients and healthy controls (that is, not in group III patients).
Peripheral blood lymphocyte subsets
Lymphocyte subsets were determined in peripheral blood samples with EDTA using flow cytometer (FACS Canto system; Becton Dickinson, Franklin Lakes, NJ, USA). As per the manufacturers’ instructions, kits compatible with the devices were used for these analyses. Complete blood count tests were performed by using an automated haematology analyzer (Cell-Dyn Ruby, Abbott Diagnostics, Santa Clara, CA, USA), and standard hematologic parameters were counted for calculation of absolute counts of lymphocyte subsets.
Analyses of HLA alleles and killer cell immunoglobulin-like receptor genes
The typing of HLAs and KIR genes were done in peripheral blood samples with EDTA. A low-resolution typing method by PCR amplification with sequence-specific primers (PCR-SSP) was used for this purpose. Commercial HLA-C SSP kits and KIR were provided by Olerup SSP (Saltsjöbaden, Sweden). Target gene amplifications were performed using the Corbett Research Thermo Cycler (Qiagen, Hilden, Germany). After PCR products were run on 2% agarose gel by electrophoresis, the separated DNA bands were displayed under ultraviolet light and their photographs were taken. Genetic analyses were done according to instructions provided by the kit manufacturer.
We conducted statistical analyses using SPSS software (IBM SPSS Statistics for Windows, version 21.0, IBM Corp., Armonk, NY, USA). Kolmogorov-Smirnov test was used for normality tests. Kruskal-Wallis test was used for multiple comparisons of numerical nonparametric data, and chi-square test was used for multiple comparisons of categorical data. An appropriate post hoc test (Mann-Whitney U test or chi-square test) was used for comparing 2 independent groups. In addition, relationships among all variables were investigated by the Spearman rank test. All directional P values were 2-tailed, and significance was assigned to values < .05. Univariate logistic regression analysis was conducted to evaluate the risk factors.
Clinical and demographic data of hemodialysis patients
The study involved 602 patients (345 male and 257 female) on HD with a median age of 64 years (range, 18-93 y). Of these, 554 HD patients (92%) had at least 1 comorbid condition (Table 1). The most prevalent comorbidities were hypertension (56%) and diabetes mellitus (35%). Fifty-two HD patients (8%) had history of solid-organ transplant. The rates of cancer and smoking were 6% and 16%, respectively. The rate of patients taking angiotensin-converting enzyme inhibitor was 14.1%, the rate of patients taking antiaggregant and/or oral anticoagulant was 37%, and 4% of patients received low-molecular-weight heparin. In our patient groups, 65% had Bacillus Calmette-Guérin (BCG) vaccine scar and 97% had hepatitis A virus (HAV) seropositivity.
In the previous year, the influenza vaccination rate was 29% (Table 2). Among 602 HD patients, only 7 (1.1%) were infected with SARS-CoV-2. The disease progressed from mild to moderate in severity, but all patients recovered. Comparisons of demographic and clinical data between SARS-CoV-2 PCR-negative and PCR-positive patients are provided in Table 1. There were no significant differences between these 2 groups in terms of age (P = .113) and sex (P = .999). The median duration of HD was similar. The differences between COVID-19-positive and COVID-19-negative groups in terms of comorbid conditions were not statistically significant (P > .05). Inhaler steroid use was nearly 4 times higher in the COVID-19 PCR-positive group, with the difference being statistically significant (P = .023; 95% confidence interval, 1309-37 553). Oral antidiabetic and anticoagulant use was also higher in COVID-19 patients; however, the differences were not significant (P = .103 and P = .133, respectively). Positivity for SARS-CoV-2 PCR was higher in smokers.
Clinical and demographic data of study groups undergoing immunologic
Although patients in group I (COVID-19-positive non-HD patients) were significantly older than patients in groups II and III (P = .039, P = .030, respectively), the difference was not significant versus group IV (P = .060) (Table 3). There were no differences in the number of men versus women among the study groups. The HD duration was not different between COVID-19-positive and -negative HD patients.
Immunologic and molecular analyses of study groups Table 3 and Figure 1 show that the absolute counts of natural killer (NK) cells in healthy controls were higher than in the other study groups, although not significantly. Although some significant differences in NK cell subsets were found between study groups, percentages of NK cell subsets bearing CD16bright and CD56dim (CD3-CD16bright56dim) in COVID-19-positive non-HD and COVID-19-positive HD patients were lower than percentages for COVID-19-negative HD patients and healthy controls. Similarly, the absolute counts of CD3-CD16bright56dim in both COVID-19-positive study groups were lower than counts for healthy controls, although not different from COVID-19-negative HD patients. In addition, there were no differences between both COVID-19-positive study groups in terms of either percentages or absolute counts of CD3-CD16bright56dim cells.
Percentages of NK cell subsets bearing CD16dim and CD56bright (CD3-CD16dim56bright) in COVID-19-positive non-HD and COVID-19-positive HD patients were higher than percentages in healthy controls, although these percentages were not different from COVID-19-negative HD patients. None of the study groups were different from each other in terms of percentage of CD3-CD16dim56bright cells. Absolute counts of CD3-CD16dim56bright cells in COVID-19-positive groups were higher than counts in the COVID-19-negative groups, although not significantly different from counts in COVID-19-positive HD patients. No differences were shown between COVID-19-negative groups and COVID-19-positive HD patients with respect to CD3-CD16dim56bright cells.
Percentages of activated T cells (CD3+HLA-DR+) were found to be significantly increased in both COVID-19-positive groups compared with the COVID-19-negative groups. COVID-19-positive non-HD patients also had higher percentages of CD3+HLA-DR+ cells than COVID-19-positive HD patients. No differences were observed between the COVID-19-negative groups in terms of percentage of CD3+HLA-DR+ cells.
Groups showed significant differences in C and DQ loci with respect to the distribution of alleles in both classes of HLA (class I and II) (Table 4). However, when HLA-C alleles were evaluated as 2 groups of antigens, C1 and C2 according to KIR ligands (Table 5), C2 alleles were more common in COVID-19-positive groups than in COVID-19-negative groups (66.7% vs 38.5%; P = .136). In addition, the genes encoding KIR 2DL2 molecules with inhibitory activity were more common in COVID-19-positive patients than in healthy controls (P = .047) (Table 6).
COVID-19 has quickly become a worldwide pandemic following the announcement of the first case on December 31, 2019. As of May 10, 2020, more than 4.1 million people have been infected across the globe, and the disease has shown a mortality rate of 36 per 1 million people. In Turkey, as of May 10, 2020, 138 657 people were infected and 3486 people have died due to COVID-19.2 So far, it appears that the incidence rate and the severity of disease vary by geography. Although the number of patients per 1 million is between 2045 to 5561 in developed countries, the frequency is lower in developing ones (ranging between 734 and 1626). Most notably, mortality rates in western countries are 10 to 40 times higher than in eastern countries (13 deaths per 1 million infections in Russia vs 569 deaths per 1 million infections in Spain). Although the diagnostic tests also vary by region, it is increasingly difficult to explain the differences in the severity of the disease by geography solely based on patient demographics. These differences perhaps could be explained by the underpinned global infections by different mutative strains across different geographies, the impact of the virus based on human genotype, and immunities developed against past infections, which could be providing partial immunity against COVID-19. When assessed based on the third mechanism, geographies with lower COVID-19 incidence rates seem to have higher rates of HAV and tuberculosis. Accordingly, HAV seroconversion and BCG vaccinations are common practices in these regions.
For immunocompromised patients, including HD patients, this novel coronavirus pandemic can have a significant threat. In addition to compromised immune systems, HD patients are often older and have multiple comorbidities (including diabetes and chronic cardiac and pulmonary diseases). Moreover, because these patients need regular HD to maintain survival, the frequency of their medical visits and the long hours spent at dialysis centers expose them to additional risks. In Turkey, the incidence of COVID-19 among HD patients (0.26%) was twice as high as the incidence among the general Turkish population (0.16%). In previous studies, COVID-19 incidence in HD patients ranged from 2% to 30%.3,4 Therefore, the rate of COVID-19 among our HD patients was surprisingly lower than expected. Also of note, whereas the mortality rate among HD patients was expected to be high (due to existing comorbidities and compromised immune systems in patients), in reality the mortality rate among HD patients (we had no mortality) was lower than that shown in the general Turkish population (0.0045%). None of the 7 patients required intensive care during follow-up. Six of the 7 patients recommenced regular HD sessions after 21 to 30 days of follow-up. The mild prognosis within our HD patients may be explained by the patients’ epidemiological characteristics. Among the group, the use of inhaler steroids, probably pointing out chronic obstructive pulmonary disease along with allergic asthma as underlying diseases, seemed to be the only statistically significant factor associated with COVID-19. Smoking and oral antidiabetic and oral anticoagulant use were also higher in COVID-19 patients; however, the differences were not statistically significant.
The protective effects of BCG and HAV seropositivity against COVID-19 have been questioned in the context of the geographical distribution of COVID-19. There is experimental evidence from both animal and human studies that the BCG vaccine has nonspecific effects on the immune system, but its clinical relevance is unknown. There is not yet evidence that the BCG vaccine protects people against infection with COVID-19. Two clinical trials addressing this question are underway.5
The overlap between high HAV seropositivity rates and low COVID-19 incidence among various countries raises questions about the protective effect of HAV antibody. A recent report concluded that the stimulation of the immune system with a HAV vaccine may be beneficial to prevent further fatalities.6 The study also reported that the patients’ existing immune responses to HAV and BCG had no impact on the incidence of COVID-19. However, the positive impact of this immunity on prognosis of COVID-19 needs to be investigated with further studies.
Our main important findings were identified during immunologic analyses, which could potentially clarify the immune pathogenesis of COVID-19. We observed that COVID-19-positive non-HD patients and COVID-19-positive HD patients had some differences with respect to rate and/or absolute counts of NK cell subsets circulated in peripheral blood. A detailed analysis reported some defects in the molecular mechanisms that provide their cytotoxic activities, with number and activity of NK cells and Tc cells found to be influenced with decrease in renal function.7,8 In a study from Vacher-Coponat and associates, diminishing NK cell numbers were found to be associated with decreasing renal function in uremic patients, and dialysis duration was identified as a factor influencing NK cell function.9 Although there is no recent study on the effects of chronic renal failure and/or HD on NK cells, it is reported that HD suppresses NK cell function and the membranes used play an effective role in this circumstance.9 In the present study, we found lower percentages and absolute counts of NK cells with cytotoxic features (CD3-CD16bright56dim) in COVID-19-positive non-HD and HD patients. However, percentages and absolute counts of NK cells with immunoregulatory features (CD3-CD16dim 56bright) in COVID-19 non-HD and HD patients were higher than those shown in COVID-19-negative patients. The decrease in cytotoxic subgroups of NK cells in COVID-19-positive patients and an increase in favor of immunosuppressive subgroups of NK cells support that these cells may play an important role in the immunopathogenesis of this disease.7-9 From another point of view, the difference in distribution of NK cell subsets may be a reason or consequence in the severe course of COVID-19.
Interestingly, patients with end-stage renal disease receiving HD have lower risk of graft rejection than patients receiving peritoneal dialysis because of premature aging of the immune system especially playing a role in the cellular arm of immunity.10 This finding supports that the dialysis modality in patients with end-stage renal disease may harm T-cell immunity. In our study, although we did not compare patients on HD versus patients on peritoneal dialysis, we observed no differences among T-cell subsets with respect to their percentages and absolute counts in peripheral blood. However, increased HLA-DR expression on T cells in COVID-19-positive patients was found, supporting an immune activation effort against this infection.
Some studies have reported an association between severe acute respiratory syndrome (SARS) and HLA. For instance, Lin and associates suggested that HLA-B*4601 might be a risk factor for developing severe SARS infection.11 In another study, Hajeer and colleagues demonstrated an association between severe Middle East respiratory syndrome and HLA-DRB1*11:01 and HLA-DQB1*02:02.12 According to our data, HLA-C and HLA-DQ allele distributionS were different from one another in COVID-19-positive versus -negative groups. However, although not significant, the HLA-C alleles belonging to the C2 group were observed to be more frequent in the COVID-19-positive study groups. As mentioned by Lin and colleagues,11 we did not find any individual who had HLA-B*46 in our study groups.
The distribution of HLA alleles exhibits some racial and geographical differences.13 Considering the ability of HLA molecules to present microbial antigens and stimulate immune response, different populations may be expected to be more susceptible to certain infectious diseases. Indeed, compared with other countries, Turkey has been less affected by COVID-19 in terms of both spread and fatality. Together, these clues indicate that a more detailed investigation on HLA factors must be performed as soon as possible to clarify the role of these molecules in the immunopathogenesis of COVID-19.
No novel or older studies have investigated associations between KIR genes and coronaviruses. However, associations between KIR genes and some viral infections, including HIV, hepatitis C virus, and influenza, have been reported.14 Our molecular analyses showed that there was a dominance in favor of the inhibitory KIR (especially in KIR2DL2) in COVID-19-positive patients compared with healthy controls. It is well known that KIR molecules expressed on surfaces of NK cells have dual roles (inhibiting and activating) according to the length of their cytoplasmic tails.15,16 The balance between the expression patterns of these functionally opposite molecules on cells is genetically controlled. Of course, the interaction of KIR molecules with their own ligands results in either activation or inhibition of NK cells according to this balance. Because HLA-C antigens are the main ligands of KIRs, it is inevitable that receptor-ligand interactions between these molecules have a role in the immunopathogenesis of COVID-19 and the severe course of disease.17
Importantly, the quantitative and functional deficiencies of NK cells with high cytotoxic capacity might negatively affect the course of COVID-19. In addition, differences detected in the gene level of KIRs and their ligands (HLA) that regulate the activity of cytotoxic cells are important in terms of clarifying the immunopathogenesis of the disease. We believe that more comprehensive analyses with more cases will clearly demonstrate the role of NK cells and KIR-HLA interactions in this infection, perhaps leading to identification of points to be targeted in treatment.
Although immunocompromised patients have more risk for COVID-19, we found lower rates of COVID-19 incidence in our HD patients. We believe that this issue should be tested further in in vitro and molecular studies.
Volume : 18
Issue : 3
Pages : 275 - 283
DOI : 10.6002/ect.2020.0194
From the 1Department of Infectious Diseases, the
2Department of Immunology, the 3Department of General
Surgery, Division of Transplantation, the 4Department of Nephrology,
the 5Department of Pulmonary Diseases, and the 6Department
of Radiology, Baskent University,Ankara, Turkey
Acknowledgements: The authors have no sources of funding for this study and have no conflicts of interest to declare.
Corresponding author: Ebru H. Ayvazoglu Soy, Department of General Surgery, Division of Transplantation, Baskent University, Ankara, Turkey
Phone: +90 312 2036868
Table 1. Demograhic and Clinical Data of Patients on Hemodialysis
Table 2. Vaccination and Hepatitis A Serology of Hemodialysis Patients
Table 3. Absolute Counts and Percentages of Peripheral Blood Lymphocyte Subsets in the Study Groups
Table 4. Distribution HLA Alleles
Table 5. HLA-C Alleles Belonging to C1 and C2 Groups
Table 6. Distribution of Killer Cell Immunoglobulin-Like Receptor Gene Frequencies in COVID-19-Positive Patients and Healthy Controls
Figure 1. Comparisons of Lymphocyte Subsets in Peripheral Blood Samples Among Groups