Secondary Immunodeficiency Related to Kidney Disease (SIDKD)—Definition, Unmet Need, and Mechanisms

Abstract

Kidney disease is a known risk factor for poor outcomes of COVID-19 and many other serious infections. Conversely, infection is the second most common cause of death in patients with kidney disease. However, little is known about the underlying secondary immunodeficiency related to kidney disease (SIDKD). In contrast to cardiovascular disease related to kidney disease, which has triggered countless epidemiologic, clinical, and experimental research activities or interventional trials, investments in tracing, understanding, and therapeutically targeting SIDKD have been sparse. As a call for more awareness of SIDKD as an imminent unmet medical need that requires rigorous research activities at all levels, we review the epidemiology of SIDKD and the numerous aspects of the abnormal immunophenotype of patients with kidney disease. We propose a definition of SIDKD and discuss the pathogenic mechanisms of SIDKD known thus far, including more recent insights into the unexpected immunoregulatory roles of elevated levels of FGF23 and hyperuricemia and shifts in the secretome of the intestinal microbiota in kidney disease. As an ultimate goal, we should aim to develop therapeutics that can reduce mortality due to infections in patients with kidney disease by normalizing host defense to pathogens and immune responses to vaccines.

Cardiovascular disease and infections are predominant causes of death in patients with kidney disease. Whereas cardiovascular disease is a primary focus of many research activities in nephrology, awareness, and funding, research on aberrant immune function and infections in patients with kidney disease is sparse. Recently, the coronavirus disease 2019 (COVID-19) pandemic garnered more attention for kidney disease as a risk factor for severe and lethal COVID-19 and that kidney disease can impair the immune response to vaccines.¹ However, the molecular mechanisms underlying the increased susceptibility to severe infections and impaired vaccine responses remain unclear, as do possible therapeutic interventions to interfere with the effect of kidney disease on the immune system.

Kidney disease is associated with a state of secondary immunodeficiency (SID), here referred to as SID related to kidney disease (SIDKD). Notably, most chronic organ failures, but also malnutrition, aging, chronic infections, and numerous immunosuppressive drugs, can cause SID (Figure 1). In this review, we highlight the unmet needs relating to SIDKD in terms of awareness, a new definition, epidemiology, pathophysiologic mechanisms, and a call for targeting SIDKD with specific interventions to improve outcomes in patients with kidney disease.

Figure 1.

Factors leading to secondary immmunodeficiency. Multiple factors contribute to SIDKD, including HIV, most forms of chronic organ failure (such as heart, liver, and kidney failure), malnutrition, aging, cancer, dysbiosis, gut disorders, chronic infections (such as tuberculosis), and various immunosuppressive drugs.

Toward a New and Uniform Definition of SIDKD

A definition of SIDKD has not been established but would be essential for epidemiologic studies, preclinical science, clinical research, and teaching. Ideally, such a definition would match those used for other forms of immunodeficiency (ID). The European Society for Immunodeficiencies working definitions for the heterogeneous inborn errors of immunity, and definitions of SID used in other medical contexts, provide some guidance.^2,3 In general, such definitions include a clinical component indicating increased susceptibility to infections, represented by recurrent or severe infections and/or insufficient vaccine responses. A second component refers to abnormal results of immunophenotyping. For SIDKD, any form of kidney disease, but particularly CKD, would be a conditio sine qua non. Hence, we propose the a definition of SIDKD in Table 1, which is shown in a simplified way in Figure 2A. Of note, kidney disease can also evolve from primary ID. For example, genetic complement deficiencies and defects in phagocytosis or lymphocyte apoptosis may first lead to systemic autoimmunity, followed by immune complex GN. Furthermore, in patients with primary ID, systemic infections, kidney infections, or recurrent use of nephrotoxic antimicrobial drugs may trigger kidney disease (Figure 2B).

Table 1.

Definition of SIDKD

Domain	Criterion
Clinical
Kidney disease	CKD (as defined by KDIGO)
AND any of the following
Infection	Severe infection requiring antimicrobial treatment and hospitalization
	OR Repetitive infections requiring repetitive antimicrobial treatment or permanent antimicrobial prophylaxis
	OR Persistent infection requiring permanent or repetitive antimicrobial treatments
	OR Opportunistic infections
OR Impaired response to vaccines	Lack of developing an antigen-specific immune response to vaccines that produce a high rate of immune response in healthy individuals
OR Attenuated sterile inflammation	For example, declining activity of autoimmune diseases or lower incidence of acute gouty arthritis than expected given the profound and persistent hyperuricemia
Laboratory
AND/OR Humoral	Subnormal serum levels of total Ig or Ig subclasses or IgG subclasses
	OR Deficiency in complement factors
AND/OR Cellular	OR Deficiency of any other humoral effector element of immunity
	Subnormal white blood count, neutropenia, lymphopenia, or any specific lymphocyte subset, such as B cells or CD4 T cells
	OR Abnormal results in immune cell activation assays and phagocytosis

KDIGO, Kidney Disease Improving Global Outcomes.

Figure 2.

Definition and the path toward SIDKD. (A) Kidney disease, i.e., CKD, is associated with SIDKD by ultimately increasing the susceptibility to infections, as represented by recurrent or severe infections and/or insufficient vaccine responses. Such abnormalities also contribute to defective humoral and cellular immune responses. (B) Primary ID can also cause kidney disease, e.g., genetic complement deficiencies and defects in phagocytosis or lymphocyte apoptosis that may lead to systemic autoimmunity and immune complex GN. In patients with primary ID, systemic or kidney infections and recurrent use of nephrotoxic antimicrobial drugs may trigger kidney disease. Thus, all risk factors contributing to SIDKD are associated with increased susceptibility to infection, impaired vaccine response, and attenuated sterile inflammation.

Epidemiology of SIDKD

Without a uniform definition of SIDKD, definite epidemiologic data have remained vague. One approach is to assess infection rates in CKD/ESKD patient cohorts, e.g., by using hospitalization and mortality data.

Impaired Host Defense to Infections

The global influence of COVID-19 has raised awareness of the vulnerability of patients with CKD/ESKD to lethal viral infection.^4–6 Cohort studies observed a COVID-19 mortality rate of 17%–44% among the CKD/ESKD population, particularly in those requiring dialysis across countries such as Italy,^7,8 Spain,^9,10 Korea,¹¹ Turkey,¹² Germany,¹³ the United States,^14,15 Sweden,¹⁶ and Australia/New Zealand.¹⁷ Kidney transplant recipients had an even higher mortality of COVID-19 (37.5%; hazard ratio, 3.36; 95% confidence interval [95% CI], 1.19 to 9.50, P=0.022) compared with patients with ESKD on dialysis (11.3%).¹⁸

Infection-associated mortality rates have also been reported in large CKD outcome trials (Table 2). For example, the Dapagliflozin in Patients with Chronic Kidney Disease trial reported an infection-related mortality rate of 18.6% in >2000 patients with CKD who did and did not have diabetes.¹⁹ In the Atherosclerosis Risk in Communities Study, involving >9000 patients in the United States from 1996 to 2011, CKD stage G5 was associated with an adjusted hazard ratio of 3.76 (95% CI, 1.48 to 9.58) for infection-related death.²⁰ Retrospective cohort studies before the pandemic reported a higher infection-related mortality in patients on dialysis compared with kidney transplantation recipients from the United States (adjusted death rate of 18.6 versus 12.8 per year; age 60–64)²¹ and the European Renal Association–European Dialysis and Transplant Association registry (82-fold in patients on dialysis versus 32-fold in transplant recipients).²² Data from the United States in 2016 report an incidence of 614 hospitalizations per 1000 person-years in patients with CKD,²³ which places infections as the second leading cause for hospitalization. A lower eGFR is associated with an increased one- to three-fold risk of hospitalization as CKD stages progress.^20,24,25 In >230,000 Canadian patients, the risk for hospitalization due to community-acquired pneumonia increased with decreasing eGFR.²⁶ The influenza A virus subtype H1N1 (swine influenza),²⁷ severe acute respiratory syndrome,^28,29 and tuberculosis^30,31 are associated with a longer duration in the hospital and a more aggressive clinical course or death in patients receiving dialysis, compared with the general population (Table 2). Similar observations have been documented for parasitic infections, e.g., Plasmodium falciparum or Plasmodium vivax³² and protozoa³³ in patients with CKD in developing countries.

Table 2.

Selected epidemiologic studies of SIDKD

Consequences of SIDKD	Clinical study	Mortality, Incidence, or Prevalence	Reference
Infection
General
	DAPA-CKD study	22% of non-CV death without TDM2 and 17.8% with TDM2 (HR, 0.64; 95% CI, 0.36 to 1.16)	19
	Stockholm CREAtinine Measurement Project	Community-acquired infection in patients with eGFR <30ml/min per 1.73 m2 (IRR, 1.53; 95% CI, 1.39 to 1.69) and eGFR 30-59/min per 1.73 m2 (IRR, 1.08; 95% CI, 1.01 to 1.14)	202
	ASN 2003 USRDS	CKD infection rate per 100 patients per year: sepsis, 4; UTI, 22; pneumonia, 14; any infection, 33; two-fold higher in patients with ESKD than in those with CKD	203
	CanPREDDICT prospective cohort study (Canada)	Risk of infection in 24.3% patients with eGFR 15–45 ml/min per 1.73 m2 (HR of mortality, 3.39; 95% CI, 2.65 to 4.33)	204
	The Atherosclerosis Risk in Communities Study	HRs for infection-related death: eGFR 60–89 ml/min per 1.73 m2, 0.99 (95% CI, 0.80 to 1.21); eGFR 30–59 ml/min per 1.73 m2, 1.62 (95% CI, 1.20 to 2.19); eGFR 15–29 ml/min per 1.73 m2, 3.76 (95% CI, 1.48 to 9.58)	20
	Propensity score–matched retrospective cohort study (Canada)	Increased risk for infection-related hospitalization PD versus HD with HR of 1.52 (95% CI, 1.38 to 1.68)	205
	Observational cohort study (Australia, New Zealand)	Incidence rates of infectious mortality for PD 2.8 and for HD 1.7 per 100 patients per year (IRR PD versus HD, 1.66, 95% CI, 1.47 to 1.86)	206
COVID-19	Mexican Open Registry of patients with COVID-19	Mortality 20.8% higher in patients with diabetes and CKD versus those with CKD only	207
	Retrospective study (UK Transplant Center)	Overall infection-related mortality, 0.4%	208
	ERACODA retrospective cohort study	Mortality in kidney transplant recipients (16.9%) and patients on HD (23.9%) within 28 days; HR, 1.78 (95% CI, 1.22 to 2.61)	209
	Systemic review and meta-analysis (Taiwan)	Overall mortality rate of 22.4% in patients on HD with COVID-19 (95% CI, 17.9% to 27.1%)	210
	Retrospective cohort study (United States)	Overall mortality of 24.9% in patients on HD with COVID-19	211
	Multicenter retrospective observational study (Turkey)	Overall mortality of 16.3% in patients on HD with COVID-19	212
	Retrospective cohort study (Spain)	Mortality of 36% in patients with CKD stage G5D	10
Tuberculosis	Tuberculosis register (United Kingdom)	TB incidence among patients with CKD per year: CKD stage G1–3, 80 per 100,000; CKD stage G4–5, 128 per 100,000; CKD stage G5D, 256 per 100.000 (95% CI, 183 to 374)	213
	Population-based study (British Columbia)	Relative risk of TB in patients on HD is 25.3% (95% CI, 22.86 to 31.49)	214
	Systemic review and meta-analysis	OR of 3.62 for TB in patients on HD (95% CI, 1.79 to 7.33)	215
	Retrospective cohort study (Japan)	Adjusted OR for TB-related death: eGFR <30 ml/min per 1.73 m2, 2.99 (95% CI, 1.20 to 7.51)	216
	Retrospective cohort study (Taiwan)	Adjusted OR of TB 1.45-fold higher in patients with CKD versus those without CKD (95% CI, 1.27 to 1.64)	217
Herpes zoster	Taiwan Longitudinal Healthy Insurance Database	Risk for herpes zoster in patients with CKD: HR of 1.6 (95% CI, 1.41 to 1.81)	218
Clostridium difficile infection	Single-center, retrospective, case-control study (Korea)	Increased risk for CDI and higher mortality in CKD stage G4–5 (OR, 2.9) and ESKD G5D (OR, 3.34)	219
	Case-control study (United States)	OR of CDI-associated diarrhea was 2.60 (95% CI, 1.25 to 5.41) in ESKD versus 1.07 (95% CI, 0.65 to 1.77) in CKD stage G3–5	220
	Case-control study (Mexico)	3-month mortality rate 10.4%; CDI in 29.6% of patients with CKD	221
	Case-control study (Poland)	CDI in 67% of patients with CKD stage G5D versus 5.7% in those with CKD stage G1	222
	Case-control study (Korea)	OR of 4.44 (95% CI, 2.19 to 8.99) in CKD stage G5D	223
Pneumonia	Community-based study Alberta Kidney Disease Network (Canada)	Adjusted HR for hospitalization with pneumonia: eGFR 45–59 ml/min per 1.73m2, 3.23 (95% CI, 2.40 to 4.36); eGFR 33–44 ml/min per 1.73 m2, 9.67 (95% CI, 6.36 to 14.69); eGFR <30 ml/min per 1.73 m2, 15.04 (95% CI, 9.64 to 23.47)	26
	Cardiovascular Health Study (United States)	Infection-related hospitalization: eGFR 60–89 ml/min per 1.73 m2, 16% eGFR 45–59 ml/min per 1.73m2, 37% eGFR 15–44 ml/min per 1.73 m2, 64%	25
	Retrospective cohort study (United Kingdom)	IRR for pneumonia: eGFR 45–59 ml/min per 1.73 m2, 0.95 (95% CI, 0.89 to 1.01); eGFR 30–44 ml/min per 1.73 m2, 1.19 (95% CI, 1.11 to 1.28); eGFR 15–29 ml/min per 1.73m2, 1.73 (95% CI, 1.57 to 1.92); eGFR <15 ml/min per 1.73 m2, 3.04 (95% CI, 2.42 to 3.83). IRR for LRTI: eGFR <15 ml/min per 1.73 m2, 1.47 (95% CI, 1.34 to 1.62)	224
	Nationwide population-based study (Taiwan)	Adjusted HR of pneumonia: HR of 1.97 (95% CI, 1.89 to 2.05) in patients with CKD, HR of 1.4 (95% CI, 1.31 to 1.49) in outpatient pneumonia, 2.17 (95% CI, 2.07 to 2.29) in inpatient pneumonia	225
	Observational study (Spain)	15.8% mortality in patients with CKD versus 8.3% in those without CKD (OR of 0.38 in CKD and pneumonia, 95% CI, 0.04 to 3.05)	226
	Retrospective cohort study (China)	Risk of empyema in 66.7% of patients with CKD and 52.5% of those with ESKD	227
Sepsis	Retrospective cohort study (France)	70% mortality rate in patients with CKD after 28 d versus 50% in those without CKD	228
	Retrospective cohort study (United Kingdom)	IRR for sepsis: eGFR <60 ml/min per 1.73 m2, 5.56 (95% CI, 3.90 to 7.94)	224
	Observational cohort study (Germany)	HR of 2.25 (95% CI, 1.46 to 3.46) in patients with CKD versus those without CKD	229
Dialysis catheter– and dialysis process–related infections	EPIBACDIAL multicenter prospective study (France)	Incidence of bacteremia in patients on HD of 0.93 episode per 100 patient-months; vascular access RR of 7.6 (95% CI, 3.7 to 15.6)	38
	Matched cohort study (United States)	Decreased all-cause mortality in patients on home HD versus those on PD; HR of 0.80 (95% CI, 0.73 to 0.87)	230
	Retrospective cohort study (Algeria)	Central venous catheter–related infection in patients on HD: 16.6 infections/1000d	231
	Systemic review of cohort studies	Higher RR using catheters compared with fistulas in mortality (RR, 1.53, 95% CI, 1.41 to 1.67) and infections (RR, 2.12, 95% CI, 1.79 to 2.52)	232
	Retrospective cohort study (Germany)	Infection risk of permanent HD catheters: 17.9% (systemic infection, 2.26 episodes per 1000 catheter days; local infection, 0.6 episodes per 1000 catheter days)	233
	Retrospective cohort study of outpatient dialysis center (United States)	Peritonitis due to peritoneal dialysis solution in 14 of 22 patients on PD	234
	Retrospective observational cohort study using NHSN surveillance data (United States)	Bloodstream infection due to buttonhole cannulation to access arteriovenous fistulas in 37% of patients on dialysis; adjusted RR, 2.6, 95% CI, 2.4 to 2.8	235
	The Network 9 peritonitis and catheter survival studies (United States)	Infection-related cause of catheter removal in 68% of patients; peritonitis and exit/tunnel infections in 13.4% of patients	236
	HEMO study; prospective multicenter study (United States)	35% infection-related hospitalization due to vascular access in 21% of cases; infection-related death, 23% (RR of infection-related death with age 1.47, 95% CI, 1.29 to 1.68) and with comorbidity 1.46, 95% CI, 1.21 to 1.76)	40 , 237
	Observational cohort single-center study (United States)	Patients on PD and HD with similar overall infection rates (0.77 for HD versus 0.86 for PD per year); PD versus HD: RR, 1.3, 95% CI, 0.93 to 1.8); HD catheters increased rate of bacteremia by 67%	41
	Case-control study (United Kingdom)	Incidence of endocarditis in patients with HD and PD due to AVF, 41.3%; DLTC, 37.9%; PTFE, 10.3%; DLNTC, 4%	238
Staphylococcus aureus	Meta-analysis of methicillin-resistant Staphylococcus aureus colonization	Dialysis modality of infection in 7.2% of patients on HD (95% CI, 4.9% to 9.9%) and 1.3% in patients on PD (95% CI, 0.5% to 2.4%)	239
	Nationwide study on the Danish National Registry on Regular Dialysis and Transplantation (Denmark)	Infection in 12.8% of patients on dialysis (RR, 7.42, 95% CI, 5.63 to 9.79)	240
Oral yeast colonization	Retrospective cohort study (Turkey)	Incidence of oral yeast colonization was 32.1% in RTR, 40% in HD, 20.9% in CAPD, and 18% in healthy controls	241
Drug-related infection	Retrospective cohort study in patients with AAV (China)	Pulmonary infection after rituximab treatment: 20.9 per 100 person-years (HR, 1.493, 95% CI, 1.017 to 2.191)	242
Attenuated sterile inflammation
Gout	NHANES studies of gout in patients with CKD and marked hyperuricemia (United States)	Prevalence of gout: CKD stage G1, 4%; CKD stage G2, 6%–10%; CKD stage G3, 11%–13%; CKD stage >G4, 30% (95% CI, 1.94 to 5.24 to gout patients without CKD)	64
	GCKD study of gout in patients with CKD and marked hyperuricemia (Germany)	Prevalence of gout: eGFR ≥60 ml/min per 1.73 m2, 16%; eGFR <30 ml/min per 1.73 m2, 35.6% (prevalence ratio 1.46, 95% CI, 1.21 to 1.77)	65
Impaired vaccine response
COVID-19 vaccine	Cohort study (Israel)	Lower antibody levels in patients on HD versus controls (OR of 2.7, 95% CI, 1.13 to 7.51)	50
	Cohort study (Germany)	4.92% nonhumoral responders and 28.4% noncellular responders in patients on HD who were fully vaccinated with BTN162b2	45
	Cohort study (United States)	Impaired humoral immune response to the COVID-19 Ad26.COV2.S vaccine in 33.3% of patients on HD	46 (preprint)
	Cohort study (Germany)	Impaired humoral immune response to the COVID-19 BTN162b2 vaccine in patients on HD and kidney transplant recipients	47
	Cohort studies (Germany)	Impaired humoral and cellular immune response after BTN162b2 vaccination in kidney transplant recipients	48 , 49
Hepatitis B vaccine	Cohort study (Brazil)	Vaccine response to HBV was 70% in patients on HD (anti-HBs seroconversion, OR, 5.239, 95% CI, 1.279 to 21.459)	54
	Case-control study (India)	Seroconversion rates after 3 doses of HBV vaccine: 87.5% in mild CKD, 66.6% in moderate CKD, 35.7% in HD	55
	Prospective cohort study (Spain)	1 month after vaccination, 77.5% of patients had seroconverted, 72.5% achieved high antibody response, whereas 22.5% were nonresponders	56
	Placebo-controlled, randomized, double-blind trial (France)	Immune response observed in 60% of patients on HD after vaccination	57
Diphtheria and tetanus vaccines	Prospective cohort study (Germany)	Anti-diphtheria antibody level after 5 years in 32% of patients on HD; anti-tetanus antibody level after 5 years in 65% of patients on HD	59
	Prospective cohort study (Germany)	Overall protection rate against diphtheria: 22%	60
Pneumococcal vaccine	Prospective cohort study (Germany)	Postvaccine antibody titers in 83% of patients with CKD after 1 month, 68% after 6 months, 48% after 1 year	61
Influenza vaccine	Prospective cohort study	66% protection rate against A-H3N2, 25% against A-H1N1, and 27% against B strain in patients on HD; booster injection had low effect	62

DAPA-CKD, Dapagliflozin in Patients with Chronic Kidney Disease; CV, cardiovascular; TDM2, type 2 diabetes mellitus; IRR, incidence rate ratio; ASN, American Society of Nephrology; USRDS, United States Renal Data System; UTI, urinary tract infection; CanPREDDICT, Canadian Study of Prediction of Death, Dialysis and Interim Cardiovascular Events; HR, hazard ratio; ERACODA, European Renal Association COVID-19 Database; TB, tuberculosis; OR, odds ratio; CDI, Clostridium difficile infection; LRTI, lower respiratory tract infection; RR, risk ratio; NHSN, National Healthcare Safety Network; HEMO, Hemodialysis Study; AVF, arteriovenous fistula; DLTC, dual-lumen tunneled catheter; PTFE, polytetrafluoroethylene; DLNTC, dual-lumen nontunneled catheter; RTR, renal transplant recipients; CAPD, continuous ambulatory peritoneal dialysis; AAV, ANCA-associated necrotizing vasculitis; NHANES, National Health and Nutrition Examination Survey; GCKD, German Chronic Kidney Disease; HBV, hepatitis B virus.

Dialysis-related infections can occur upon recurrent insertion of arteriovenous fistulas/grafts or dialysis catheters.^34,35 The type of dialysis access matters in this context.^36,37 Patients undergoing dialysis via catheters have a two-fold higher risk for bacteremia or sepsis^38,39 and infection-related death⁴⁰ compared with those with arteriovenous fistulas/grafts. In addition, patients on hemodialysis (HD) are more susceptible to bacteremia, whereas patients on peritoneal dialysis (PD) have an increased risk for peritonitis.^36,41,42 Furthermore, contamination of dialysis equipment and dialysate^43,44 with diverse virulent microorganisms^42,43 can underlie SIDKD in this population.

Impaired Immune Responses to Vaccines

Another clinical indicator of SIDKD is impaired vaccine response. Data from the ongoing COVID-19 pandemic show that patients on HD and PD may have a humoral antibody response against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) after receipt of the mRNA vaccine BNT162b2.^{45,46(preprint)–49} However, the levels of anti-S1 IgG antibodies and SARS-CoV-2 surrogate neutralizing antibodies were substantially lower in patients on dialysis after the first and second vaccination compared with patients without CKD.^50,51 Within 12 weeks surrogate neutralizing antibodies start to decrease from 93% to 85% in patients on PD and from 87% to 79% in patients on HD, respectively.^52–53 Consistently, diminished seroconversion rates after immunization have been reported for hepatitis B virus (no difference in responsiveness between patients on PD and HD),^54–58 diphtheria and tetanus,^59,60 pneumococcal,⁶¹ and influenza⁶² vaccines compared with the general population. As vaccine-induced antibody levels and humoral immunity decrease, patients with ESKD benefit from booster vaccinations.⁶³

Sterile Forms of Inflammation

SIDKD also implies an attenuation of noninfectious forms of inflammation, such as the autoimmune disease activity that declines as CKD progresses. In addition, patients with CKD/ESKD who have significant hyperuricemia show an unexpectedly low rate of acute gout attacks.^64,65

Although the lack of a common definition of SIDKD makes definite epidemiologic data scarce, numerous clinical observations suggest that SIDKD is prevalent in patients with CKD and ESKD; therefore, mechanistic explanations of SIDKD are required.

Immunophenotype of Patients with Kidney Disease

Biomarkers for SID in clinical use include white blood cell count, total Ig levels, measurement of antibodies to specific pathogens or vaccine responses, and fluid-phase complement factors, but also specific assays of neutrophil and T cell function. Immunophenotyping of patients with CKD/ESKD reveals signs of chronic inflammation and impaired immune effector functions (Figure 3A).⁶⁶

Figure 3.

Immunophenotype in kidney disease. (A) SIDKD is associated with an abnormal immunophenotype, characterized by chronic inflammation as a result of persistent immune cell activation, and simultaneously with immune paralysis due to impaired immune effector functions. (B) Mechanisms that are involved in SIDKD include suppressed innate immune responses, e.g., impaired leukocyte and platelet function, reduced antigen-presenting ability of macrophages and dendritic cells to T and B cells, and impaired adaptive immune responses (e.g., T lymphocyte maturation and activation, reduced antibody production by plasma cells).

Impaired Function of Innate Immunity

In patients with CKD/ESKD, biomarkers of oxidative stress, such as advanced oxidation protein products and myeloperoxidase, and for inflammation, such as high sensitivity C-reactive protein and IL-6, are inter-related.^67,68 Thus, uremia leads to chronic low-grade inflammation, in which permanent immune cell activation and secretion of inflammatory cytokines, together with uremic solutes/metabolites, reduce the ability of neutrophils and monocytes to migrate to the inflammation site (Figure 3B).^69–71 Mechanistically, selectin-induced slow leukocyte rolling and transmigration are abolished,^72,73 and β₂ integrins are downregulated.⁷⁴ Moreover, neutrophils from patients with ESKD are more likely to undergo apoptosis, which decreases the ability of neutrophils to form neutrophil extracellular traps ^75,76 and to phagocytose and kill pathogens,⁷⁷ mechanisms that are associated with an impaired host defense in SIDKD. A number of uremic solutes/metabolites that affect neutrophil oxidative burst have been identified, such as phenylacetic acid, resistin, and methylglyoxal.^78,79

Monocytes from patients on PD are hyporeactive to LPS stimulation and release fewer proinflammatory cytokines in comparison to control subjects.^80,81 Monocytes and monocyte-derived dendritic cells display decreased endocytosis and impaired maturation when cultured in uremic serum or when obtained from patients with ESKD.⁸² Moreover, dendritic cells and macrophages have a reduced antigen-presenting capability to activate T cells, which might be due to an abnormal expression of Toll-like receptor 4⁸³ and costimulatory molecules CD80 and CD86.⁸⁴ An increased intracellular uptake of uric acid, related to an impaired renal clearance of uric acid uptake, contributes to uremia-related monocyte dysfunction.⁷⁴

In patients on HD, higher serum levels of mannose-binding lectin are associated with an increased risk of severe infection.^85,86 Mannose-binding lectin can bind to carbohydrates on the surfaces of bacteria and viruses (e.g., SARS-CoV-2⁸⁷) and initiate complement-driven pathogen lysis.^88,89 However, uncontrolled complement activation also contributes to hyperinflammation, cell injury, immunothrombosis, and multiorgan failure during COVID-19 infection.⁹⁰ In addition, monocytes/macrophages from patients with ESKD have increased expression and activity of the macrophage scavenger receptors SR-A and CD36,^91–93 but decreased inducible nitric-oxide synthase expression⁹⁴ and phagocytic capability, e.g., of peritoneal macrophages in chronic PD and PD-induced peritonitis^95,96 and potentially also in alveolar macrophages in the context of pulmonary infections. ESKD reduces the cell number and cytotoxicity of natural killer (NK) cells in association with downregulation and modulation of ligand expression for activating receptors on NK cells.^97,98 Thus, uremia impairs the functional properties of neutrophils, monocyte/macrophages, and NK cells that are important to maintain a sufficient host defense.

Impaired Platelet Function

Beyond their important role in homeostasis, platelets are important cellular contributors to host defense⁹⁹ and essential for sufficient neutrophil activation and recruitment into peripheral organs under different inflammatory conditions. However, gut microbial uremic toxins, which accumulate during CKD,¹⁰⁰ and altered HDL levels contribute to altered homeostasis and immune effector functions in SIDKD (Figure 3).¹⁰¹ This dichotomy may be explained by the fact that critical factors determining the effective formation of platelet-leukocyte aggregates, e.g., plasma fibrinogen levels involved in bond formation between integrins, GPIIbIIIa on platelets, and Mac-1 on neutrophils, are increased in patients with CKD.¹⁰² Furthermore, platelets release more α-granules on stimulation, leading to increased incorporation and expression of P-selectin on the platelet surface, which, in turn, binds to PGSL-1, which is expressed on various leukocyte subsets, including neutrophils.¹⁰³ Thus, uremic platelet dysfunction not only promotes bleeding complications¹⁰⁴ but also immunothrombosis in CKD.^105,106

Impaired Function of Adaptive Immunity

Impaired vaccine responses indicate that CKD/ESKD also suppresses adaptive immunity (Figure 3B).¹⁰⁷ This suppression involves aberrant T-cell activation caused by altered dendritic cell and macrophage antigen presentation and increased apoptosis of effector CD4+ and CD8+ T cells.^98,108,109 Studies performed in vitro show decreased T helper (Th) CD4 cell proliferation in the uremic milieu.^110,111 HD and PD have different effects on Th cell phenotypes and proliferation.^112,113 In patients on PD, the maturation of both effector Th1 and Th2 CD4 cell subsets is impaired, and there are increased numbers of central memory T cells and CD8+ naive T cells compared with controls and patients on HD.¹¹³ In contrast, patients on HD seem to have sustained Th cell maturation, but the immune response is mainly driven via Th1 CD4 (e.g., production of IFN-γ) rather than Th2 CD4 cells.^114,115 A possible explanation for the increased Th1/Th2 ratio in patients on HD could be increased IL-12 production.¹¹⁵ Moreover, the different clinical course of infection, e.g., hepatitis B virus, between patients on dialysis and the general population relates to dysfunctional CD8 cytotoxic T cells, which are needed to eliminate infected cells, and dysfunctional CD4 T cells, which sustain CD8 cytotoxic T cells and induce antibody production from B and plasma cells.¹⁰³

Patients with CKD/ESKD show a defective or repressed humoral and vaccination response.¹¹⁶ B-cell lymphopenia, due to increased B-cell apoptosis and insufficient B-cell proliferation, is common in patients on HD and controls.^117–119 This has been related to decreased Bcl-2 expression and a resistance to IL-7 and B cell–activating factor BAFF/BLyS.^117,120 Patients on HD have fewer B1 cells secreting IgM and IgA and B2 cells producing IgG.^106,120 Patients on dialysis and kidney transplant patients have poor antibody, memory B cell, and plasmablast responses to vaccines, such as the COVID-19 (BNT162b2),⁴⁷ hepatitis B virus, and influenza vaccines compared with the general population (Table 2)^121,122. Therefore, these patients may not be sufficiently protected against the respective pathogens. CKD and ESKD also impair adaptive immune responses, which are important for host defense but are also implicated in auto- and alloimmunity.

Pathogenic Mechanisms of SIDKD

SIDKD due to Intestinal Barrier Dysfunction

Evidence indicates that the elevated circulating endotoxin/LPS levels and markers of systemic inflammation in patients with CKD stage G3–5 and 5D are caused by a leaky intestinal barrier and enhanced endotoxin translocation from the intestinal lumen into the circulation.^123,124 LPS levels were the highest in patients on HD/PD; levels were similar to those in patients with severe liver disease, irradiation-induced intestinal damage, and decompensated heart failure.¹²³ Elevated LPS levels are considered a strong and independent predictor of mortality.¹²³ Wang et al.¹²⁵ demonstrated that experimental uremia in rats increases bacterial translocation from the gut into mesenteric lymph nodes, liver, and spleen, which was associated with higher levels of serum IL-6 and C-reactive protein.¹²⁶ The aberrant humoral and cellular immunity in CKD is partially reversible by treatment with antibiotics that eliminate the intestinal flora.¹²⁷ Endotoxin tolerance and the see-saw phenomenon in sepsis posit that persistent exposure to bacterial endotoxins induces negative regulators of innate immunity that paralyze the innate and adaptive immune system.^127,128 In accordance with this concept, CKD-related intestinal dysbiosis, barrier dysfunction, and bacterial translocation account for the state of persistent systemic inflammation in CKD (Figure 4).¹²⁹

Figure 4.

Pathomechanisms of SIDKD. Kidney disease is associated with an impaired urinary clearance of immunoregulatory metabolites, increased production of immunoregulatory proteins and persistent immune activation, extrinsic immunosuppressors, and a leaky gut and shift in the secretome of the intestinal microbiota—all of which contribute to SIDKD. The pathogenic mechanisms of SIDKD include dysregulation of innate and adaptive immune responses, such as a decreased phagocytic capability of immune cells to clear pathogens, reduced respiratory burst (ROS production) to kill bacteria and viruses, enhanced immune cell activation (upregulation of cell surface receptors) and release of proinflammatory cytokines (e.g., IL-1β, IL-6, TNFα) by innate immune cells (e.g., neutrophils, monocytes, macrophages, dendritic cells), increased neutrophil apoptosis but decreased neutrophil extracellular trap (NET) formation, and reduced antigen-presenting ability of macrophages and dendritic cells to activate T and B cells, which results in decreased T-cell proliferation and antibody production by plasma cells. Although leukocyte/lymphocyte interactions with endothelial cells are increased, the ability of leukocytes/lymphocytes to migrate is impaired due to a downregulation of, e.g., β₂ integrins. Subsequently, the dysregulation of the innate and adaptive immune system contributes to bacterial overgrowth, infections, and attenuated sterile inflammation.

SIDKD due to Persistent Inflammation

Persistent inflammation is an important trigger of immune paralysis and SID. Indeed, dendritic cells, neutrophils, monocytes, macrophages, and mast cells remain unresponsive to a secondary endotoxin challenge due to an induced suppression of signaling pathways contributing to inflammation. This see-saw phenomenon of immune activation is an important, evolutionarily conserved mechanism that avoids a potentially lethal cytokine storm and septic shock. In patients with permanent triggers of inflammation, this see-saw phenomenon begins at the single-cell level.¹³⁰ The flood of new immune cells produced daily results in a state of concomitant systemic inflammation and immune paralysis at the organismal level.

The endotoxin tolerance–like concept explains the immunosuppressive state in sepsis, cystic fibrosis, pancreatitis, and tuberculosis.¹³¹ In the early phase, cytokine storm–like hyperimmunoreactivity predominates and may be lethal. At later stages, and in chronic disorders, hypoimmunoreactivity (as a sign of SID) prevails, leading to potentially fatal secondary infections.¹³² Cellular mechanisms that underlie immunosuppression include unresponsiveness of monocytes to LPS challenge ex vivo;¹³³ reactivation of adaptive immune cells, such as T regulatory cells and myeloid-derived suppressor cells^134,135; but loss of CD4+ and CD8+ T cells (Figure 4).^136,137 However, this induced immune paralysis is not restricted to sepsis or cystic fibrosis, but is also present in other forms of sterile inflammation, e.g., in hepatic and kidney ischemia, coronary occlusion, acute coronary syndromes, and even cancer.^138–141

An impaired intestinal barrier is a possible source of persistent exposure to bacterial endotoxins, as indicated by increasing endotoxin serum levels at later stages of CKD.^123,127

Shifts in the Secretome of the Intestinal Microbiota

The secretome of the intestinal microbiota is an integral component of human physiology and changes in the microbiota composition, which affect the secretome, disturb physiology and homeostasis.^142–144 Kidney disease–related alkalosis, intestinal wall congestion, azotemia, and other metabolic disturbances alter the composition of the intestinal microbiota and hence its secretome.^127,145 Among many other aspects of physiology, a uremia-related altered intestinal secretome implies changes in numerous immunoregulatory metabolites, including a reduction in short-chain fatty acids, such as butyrate and acetate, that usually suppress inflammation (Figure 4).^146,147 Other gut-derived metabolites that affect the immune system include indoxyl sulfate, p-cresyl sulfate, indole-3-acetic acid, and phenylacetylglutamine.^147,148 Mechanistically, indoxyl sulfate and p-cresyl sulfate are associated with systemic inflammation and an altered host defense in patients with CKD, e.g., by inducing glutathione peroxidase, TNFα, and IL-6 release in monocytes¹⁴⁹; impairing respiratory burst activity and phagocytosis in neutrophils; and causing eryptosis (programmed death of red blood cells) partly by increasing cytosolic calcium.¹⁵⁰

A high salt diet can reduce the intestinal population of Lactobacillus murinus, which decreases indole-3-lactate levels, affecting Th17 cell function in a way that can aggravate experimental hypertension and multiple sclerosis.¹⁵¹ It is possible that CKD or CKD-related medications, dietary restrictions, and use of antibiotics alter the intestinal microbiota,¹⁵² leading to concomitant systemic inflammation and ID. Moreover, fecal transplantation in CKD can induce metabolic changes that lead to sarcopenia and immune dysregulation. Mice receiving a fecal transplant from CKD patients had significantly higher plasma trimethylamine N-oxide levels and a different gut microbiota composition than mice receiving a fecal transplant from non-CKD patients.¹⁵³ These findings suggests that CKD modifies the composition of the intestinal microbiota and its secretome, which, together with barrier dysfunction and bacterial translocation, might explain the persistent systemic inflammation and defective host defense associated with CKD.

Impaired Urinary Clearance of Immunoregulatory Metabolites

Low-molecular-weight solutes involved in both systemic and organ-specific metabolism may impair urinary clearance and exhibit increased production, or decreased breakdown, in kidney disease (Figure 4). For example, p-cresyl sulfate induces an increase in oxidative burst and phagocytosis in human macrophages, but decreases their antigen-presenting ability to T cells by downregulating HLA-DR and CD86.¹⁵⁴ Moreover, indoxyl sulfate promotes leukocyte-endothelial interactions through the upregulation of vascular endothelial adhesion molecules, such as E-selectin, via JNK- and oxidative stress–dependent pathways¹⁵⁵; can induce the expression of adhesion molecules, such as MAC-1; and increase production of reactive oxygen species (ROS) in a manner dependent on NADPH oxidase and p38 MAPK in monocytes.¹⁵⁶ In patients with common variable ID that present with defective B cell to plasma cell development, several studies report a strong association between trimethylamine N-oxide and asymmetric dimethylarginine with systemic inflammation, immune cell activation, and gut microbial abundance.^157–159

Whereas the above-mentioned metabolites drive systemic inflammation, CKD-related hyperuricemia instead suppresses innate immunity.⁷⁴ This phenomenon occurs through the intracellular uptake of soluble uric acid into CD14+ monocytes via selective urate transporters, such as the glucose transporter 9 (SLC2A9), which inhibits Toll-like receptor signaling, production of proinflammatory cytokines, and migration of CD14+ monocytes in patients with CKD/ESKD.⁷⁴ Hyperuricemia also attenuates β₂ integrin activation and integrin trafficking in neutrophils, which impairs neutrophil recruitment to sites of DAMP/PAMP-driven sterile inflammation (own unpublished data). This effect highlights a previously unexpected immunoregulatory role of asymptomatic serum UA levels of 7–12 mg/dl in those with CKD/ESKD. Further studies are needed to clarify the pathophysiologic effects of asymptomatic hyperuricemia in SIDKD, especially whether urate-lowering therapy can improve host defense in CKD/ESKD.

Immune Paralysis due to Increased Production of Immunoregulatory Proteins

A decline in kidney excretory function is associated with overexpression of certain immunoregulatory proteins (Figure 4). For example, fibroblast growth factor 23 (FGF23) levels drastically increase with CKD progression, which has a direct effect on neutrophil functions, including selectin-mediated slow rolling, chemokine-induced adhesion and arrest under dynamic flow, and ROS release. The resulting neutrophil recruitment defect impairs pathogen control in bacterial pneumonia in mice.⁷¹ Mechanistically, FGF23 binding to its counter-receptor FGFR2 on neutrophils counteracts selectin- and chemokine-triggered β₂ integrin activation by activating protein kinase A and suppressing the activation of the small GTPase Rap1. Reduced neutrophil rolling and arrest on activated endothelium, associated with elevated FGF23 blood levels, was also observed in whole blood samples from patients with CKD stage G3–5.⁷¹

Levels of the middle-molecular-weight proteins leptin,^160,161 resistin,^161,162 and modified ubiquitin⁷⁰ increase in patients with CKD/ESKD compared with healthy controls,¹⁶² which contributes to neutrophil dysfunction. Leptin can inhibit neutrophil migration by impairing neutrophil locomotion via MAPK- and Src kinase–related F-actin polymerization.¹⁶⁰ In contrast, resistin¹⁶² and p-cresol¹⁶³ decrease the respiratory burst activity of neutrophils, whereas complement factor D and angiogenin overexpression inhibit neutrophil degranulation^164–166 and glucose-modified proteins (glycated proteins) promote neutrophil apoptosis.¹⁶⁷

SIDKD due to Extrinsic Immunosuppressors

Extrinsic factors contribute to SIDKD (Figure 4). For example, there is a concern that iron administration for renal anemia could decrease host defense. Indeed, iron can decrease the phagocytic capacity of neutrophils and macrophages.^168,169 Iron can also impair the proliferation of T cells, lower IL-2 and IFN-γ production, and reduce the antibody response of B cells.^170–172 Infection may be a long-term complication of iron therapy in patients with CKD and may play an important role in suppressing immunity, especially in cases of iron overload.¹⁷³ A meta-analysis of randomized controlled trials showed that intravenous iron therapy is associated with a 30% greater risk of infection compared with oral iron and no iron therapy.¹⁷⁴ However, modern intravenous iron formulations do not seem to increase infection rate in patients on HD.¹⁷⁵

Steroids and other immunosuppressant drugs are frequently prescribed to patients with CKD/ESKD and kidney transplant recipients for their anti-inflammatory and immunosuppressive effects. However, many clinical trials^176–180 (including the STOP-IgAN¹⁸¹ and TESTING^182,183 trials) assessing the use of steroids, e.g., corticosteroids, prednisolone (initially 20–60 mg/d, then tapered, for 4–24 months), reported a high rate of serious adverse outcomes (STOP-IgAN, 55%; TESTING, risk difference, 11.5% [95% CI, 4.8% to 18.2%]), including an increased risk of infections and certain opportunistic infections (STOP-IgAN, 25%; TESTING, 8%). Another cohort study reported similar results, with a 40% higher risk of serious adverse events (46.1 per 1000 person-years; relative risk, 1.4 [95% CI, 1.3 to 1.6]) in patients with CKD treated with corticosteroids most of them infections¹⁸⁴ Steroids are an independent risk factor for SID because they also increase the risk of infection in patients without CKD who have rheumatoid arthritis, asthma, vasculitis, or inflammatory bowel disease.^185–189 However, infectious complications are not increased in those receiving a daily dose <10 mg or a cumulative dose of >700 mg of prednisone.¹⁸⁹ The extent to which SID is more severe in patients treated with steroids who have CKD versus those without CKD has not been established and will certainly also affect the many confounding factors contributing to SID. Mechanistically, these agents inhibit the release of proinflammatory cytokines (e.g., IL-1β, IL-6, TNFα) in monocytes/macrophages, the recruitment and respiratory burst in neutrophils, and cause lymphopenia and impair lymphocyte function.^190–193 Although these drugs are sometimes needed to control the underlying kidney disease,¹⁹⁴ balancing protective and dysregulated host immune responses can be intricate in patients who are critically ill and experience reinfection, e.g., with COVID-19.^1,51

The removal of metabolic/uremic waste products by passing blood through synthetic tubing and dialyzer membranes during HD triggers the activation of monocytes, neutrophils, platelets, and the complement system; the release of inflammatory cytokines and ROS; and apoptosis.^195,196 Neutrophil activation with decreased β₂ integrin expression¹⁹⁷ and neutrophil degranulation,¹⁹⁸ along with the neutrophil extracellular trap formation associated with the release of DNA, myeloperoxidase, histones, and S100 proteins, can trigger endothelial activation,¹⁹⁹ which ultimately contributes to vascular inflammation and cardiovascular disease.²⁰⁰ This may explain, at least in part, the cell-mediated immunosuppression and immune paralysis seen in patients with ESKD who are on HD.

Summary and a Call for Action

More awareness is needed for the unmet need of infection-related morbidity and mortality and the other clinical consequences of SID in patients with kidney disease. The ongoing pandemic provides a unique opportunity in this context, with kidney disease a major risk factor for lethal COVID-19 and a poor vaccine response. As a starting point, we provide a practical definition of SIDKD to endorse focused research on the epidemiology of this important cause of death in patients with kidney disease. As a community of kidney researchers, we have to improve patient outcomes in this domain as we do in CKD-related cardiovascular disease. To develop better preventive and therapeutic strategies, we first need to understand the pathophysiology of SIDKD (Table 3). Some mechanisms are discussed here but, undoubtedly, there are more to discover. SIDKD offers plenty of research opportunities for established and next-generation researchers, and opportunities for industry to invest in a naive domain for therapeutic interventions. The unmet needs of SIDKD require more attention at all levels.

Table 3.

Approaches to improve SIDKD

Approaches to Identify and Prevent SIDKD	Potential Diagnosis/Intervention
Primary prevention of kidney disease	Pregnancy counseling to avoid pregnancy complications, optimized nutrition of pregnant women, minimized exposure to nephrotoxins, healthy lifestyle, early diagnosis and treatment of systemic disorders (diabetes, amyloidosis, SLE) or infections (viral hepatitis, malaria, HIV, bacterial tonsillitis)
Recognizing primary ID as a cause of kidney disease	Clinical awareness, diagnosis, monitoring, and treatment of primary ID
Secondary prevention of progression of kidney disease	Disease-specific therapies; healthy lifestyle; smoking cessation; control of diabetes, BP, body weight, and metabolic acidosis; avoiding nephrotoxins and RAS/SGLT2 inhibitors; mineralocorticosteroid receptor antagonism (in diabetes); avoiding infections (vaccination) and rigorous treatment of infections
Identifying predominant pathomechanisms of SIDKD	Targeting specific pathomechanisms
Clinical trials testing potential interventions using primary end points

RAS, renin-angiotensin system; SGLT2, sodium-glucose transporter 2.

Disclosures

H.-J. Anders reports receiving consultancy or lecture fees from AstraZeneca, Bayer, Boehringer Ingelheim, Eleva, GlaxoSmithKline, Janssen, Kezar, Lilly, Novartis, Otsuka, and PreviPharma; and serving as a scientific advisor for, or member of, JASN and Nephrology Dialysis Transplantation. S. Steiger reports receiving research funding from Eleva. A. Zarbock reports serving as a scientific advisor for, or member of, AM Pharma, Anesthesia & Analgesia, Guard Therapeutics, Journal of Immunology, Novartis, and Piaon; and receiving consulting fees, unrestricted research grants, or lecture fees from AM Pharma, Anomed, Astellas, Astute Medical, Baxter, BiMerieux, Braun, Deutsche Forschungsgemeinschaft, Else-Kröner Fresenius Stiftung, Fresenius, Guard Therapeutics, La Jolla Pharmaceuticals, Novartis, and Ratiopharm. The remaining author has nothing to disclose.

Funding

S. Steiger was supported by the Deutsche Forschungsgemeinschaft grants STE2437/2-1 and STE2437/2-2 and the Ludwig-Maximilians-Universität München Excellence Initiative. H.-J. Anders was supported by the Deutsche Forschungsgemeinschaft grants AN372/14-4, 20-2, 27-1, and 30-1 and the Volkswagen Foundation grant 97-744. A. Zarbock was supported by the Deutsche Forschungsgemeinschaft grants ZA428/17-1, ZA428/18-1, and SFB1009A05. J. Rossaint was supported by the Deutsche Forschungsgemeinschaft grant RO4537/4-1 and the Interdisziplinäres Zentrum für Klinische Forschung, Universitätsklinikum Würzburg grant Ross2/010/18.