Etiology and pathogenesis of rheumatoid vasculitis

INTRODUCTION — Rheumatoid vasculitis (RV) refers to a destructive, inflammatory process that is centered on the blood vessel wall itself. The condition almost always occurs in patients with longstanding, severe rheumatoid arthritis (RA). RV is often associated with substantial potential morbidity, requires intensive immunosuppressive therapy, and leads to a significantly higher mortality than RA itself. Within a given patient with RV, clinical features of both medium- and small-vessel disease may be found. RV leads to necrosis, blood vessel occlusion, and tissue ischemia in a manner that resembles other forms of systemic vasculitis, particularly polyarteritis nodosa (medium-vessel disease) and cutaneous small-vessel vasculitis. (See "Overview of and approach to the vasculitides in adults" and "Clinical manifestations and diagnosis of polyarteritis nodosa in adults" and "Overview of cutaneous small vessel vasculitis".)

RA is not the only rheumatologic condition associated with the development of vasculitis. Systemic lupus erythematosus (SLE) patients are at increased risk as well, particularly in individuals who have onset in childhood, associated Raynaud phenomenon, hematologic manifestations, and higher disease activity [1].

Understanding of the precipitating factors for these extremes of blood vessel inflammation in RA is limited. Although the decreased life expectancy and early cardiovascular mortality in RA are well-recognized, RV of the coronary vessels is thought to be very rare. There is no evidence that atherosclerotic vascular disease itself leads to RV, despite many common risk factors for these two RA complications. Atherosclerotic coronary heart disease and its relationship to RA are discussed in detail separately. (See "Coronary artery disease in rheumatoid arthritis: Implications for prevention and management".)

The etiology and pathogenesis of RV are reviewed in this topic. The epidemiology, clinical manifestations, diagnosis, and treatment of this disorder are discussed separately. (See "Clinical manifestations and diagnosis of rheumatoid vasculitis" and "Treatment of rheumatoid vasculitis".)

HISTOPATHOLOGY AND CLASSIFICATION — Rheumatoid vasculitis (RV) has been classified among the forms of vasculitis associated with systemic disease, rather than as a small-, medium-, or variable-vessel form of vasculitis (see "Overview of and approach to the vasculitides in adults"). The first series of patients with RV was published in 1951, describing transmural inflammation of large arterioles and small arteries in 5 of 55 patients (9 percent) who underwent skeletal muscle biopsy [2]. A classification of vasculitis published in 1978 focused on the skin manifestations of RV and placed it as a subcategory of hypersensitivity angiitis associated with immune complex deposition primarily in small venules [3]. Although this type of blood vessel involvement occurs in RV, the most devastating clinical features of the condition stem from vasculitis affecting medium-sized arteries in a manner resembling polyarteritis nodosa. (See "Clinical manifestations and diagnosis of polyarteritis nodosa in adults".)

Placing RV in any classification of vasculitis is difficult because of the variability both in the size of the vessel involved and in the histologic findings. Three major vasculitis classifications published in the 1990s chose not to address "secondary" forms of vasculitis that occur in patients with known "connective tissue disorders" (systemic autoimmune rheumatic diseases) [4-6]. However, a group of vasculitis experts convened in 2011 for the second International Chapel Hill Consensus Conference updated the previous classification and chose to include four new categories of vasculitis, one of which was "vasculitis associated with systemic disease" [7]. The conference participants were reluctant to broadly divide all the different vasculitic diseases into "primary" and "secondary" forms of vasculitis, due to the expectation that ongoing research might be expected to eventually move certain conditions from the primary into the secondary category as the pathogenesis of these diseases becomes better understood. (See "Overview of and approach to the vasculitides in adults".)

The histological findings in RV may include 1) leukocytoclastic vasculitis associated with immune complex deposition in venules, capillaries, and arterioles; and 2) pauci-immune lesions (ie, inflammation associated with sparse deposition of immunoreactants) in medium-sized arteries and renal glomeruli. One paradoxical finding in RV is that, although glomeruli are considered small blood vessels (the renal equivalent of capillaries), the well-documented cases of renal involvement in RV are generally associated with pauci-immune glomerulonephritis.

The pathology of aortitis in patients with RV is a granulomatous pattern with lymphoplasmacytic infiltration, necrosis of the medial smooth muscle, and elastic fiber loss. These findings are generally similar to other large vessel vasculitides in which aortitis is more common, such as giant cell arteritis and Takayasu arteritis [8]. In addition to aortic inflammation and aortic wall destruction, aortitis lesions in rheumatoid arthritis (RA) patients can also frequently involve areas of atherosclerotic change [9]. While RV may occur in the coronary arteries, RA is a significant risk factor for development of atherosclerosis. (See "Coronary artery disease in rheumatoid arthritis: Pathogenesis, risk factors, clinical manifestations, and diagnostic implications".)

Blood vessel inflammation is a central feature of RA, but the vascular inflammatory changes in the synovium are generally not considered as rheumatoid vasculitis. Synovial membrane inflammation is characterized histologically by mononuclear cell cuffing of postcapillary venules. However, vascular inflammation is considered a primary event in the formation of rheumatoid nodules [10]. In nodule formation, small-vessel vasculitis leads to fibrinoid necrosis that forms the core of the lesion, surrounded by fibroblastic proliferation. (See "Rheumatoid nodules".)

Practical considerations may preclude a formal identification of RV in certain presentations of severe extraarticular complications of RA, although vasculitis may contribute to the findings. As examples, scleritis and episcleritis are usually vasculitic in nature. By contrast, most cases of pulmonary disease, uveitis, myositis, and ischemic cardiomyopathy in RA have no demonstrable vasculitis. RA patients with acute coronary syndromes have significantly higher mortality compared with non-RA patients, but arteritis affecting the coronary vessels is a rare manifestation of RV (see "Coronary artery disease in rheumatoid arthritis: Pathogenesis, risk factors, clinical manifestations, and diagnostic implications"); it remains possible that RV may be an under-diagnosed underlying cause.

ETIOLOGY AND RISK — The etiology of rheumatoid vasculitis (RV) is unclear. Proposed triggers have included viral infections and medications used to treat rheumatoid arthritis (RA), including glucocorticoids, conventional nonbiologic disease-modifying antirheumatic drugs (DMARDs), and tumor necrosis factor (TNF) inhibitors; however, none of these has been consistently and strongly implicated (see 'Etiology' below). Genetic characteristics and cigarette smoking have also been suggested as risk factors for RV. (See 'Genetic risk factors' below and 'Cigarette smoking' below.)

There has been a decline in the occurrence of RV since the 1990s [11-13]. The reasons for this are uncertain but likely relate to overall improved management of RA. (See "Treatment of rheumatoid vasculitis", section on 'Prognosis'.)

As an example, in a cohort of 86 patients at a single tertiary referral center, with RV diagnosed between 2000 and 2010, risk factors associated with the development of RV, after adjusting for age and disease duration, included current smoking at time of RA diagnosis, coexistent vascular disease (both peripheral vascular disease and cerebrovascular disease), and severe RA (defined by the presence of radiographic erosions, rheumatoid nodules, or a requirement for joint surgery) [14]. Patients who had used biologic agents were at increased risk, while those who had received hydroxychloroquine (HCQ) or low-dose aspirin were at reduced risk.

Etiology — Inadequately treated rheumatoid arthritis (eg, due to lack of access to health care) may be associated with an increased risk of rheumatoid vasculitis [15]. Other suggested causes and related observations include:

●Viral infections – Viral infections have been proposed as a cause for RV, but there is little evidence to support these hypotheses. Although longstanding hepatitis C virus infections are often associated with the development of systemic vasculitis (mixed cryoglobulinemia), the clinical manifestations and pathologic findings of cryoglobulinemia differ from those of RV. Moreover, the majority of patients with RV are hepatitis C-negative and do not have cryoglobulins. (See "Mixed cryoglobulinemia syndrome: Clinical manifestations and diagnosis".)

Two patients have been described with features suggestive of RV in whom Epstein-Barr virus (EBV) RNA was identified in infiltrating lymphocytes from the vessel walls of ulcerative cutaneous lesions [16]. In both cases, the pathology bore a resemblance to EBV-associated lymphomatoid granulomatosis. The ulceration regressed in both patients after discontinuation of methotrexate (MTX). Another patient treated with MTX and prednisone developed ulcerated and non-ulcerated EBV-staining subcutaneous nodules on all extremities that resolved with discontinuation of the MTX [17]. A vasculitis-like EBV-associated mucocutaneous ulceration, characterized by large atypical B-lymphocytes in oral ulcerative lesions from MTX-treated RA patients, has also been described [18]. Neither child [19] nor adult [20] arthritis patients with chronic EBV infection, treated with MTX, have been shown more generally to experience increases in EBV viral loads.

●Glucocorticoids and nonbiologic DMARDs – Numerous medications used to treat RA have also been proposed as triggers of RV, partly because of some similarities between drug-induced small vessel cutaneous vasculitis and RV. In particular, the treatment of RA with glucocorticoids has been suggested as having a role in the development of RV. This association is partly circumstantial, because the first cases of RV were appreciated shortly after the discovery of cortisone [21]. In addition, the putative association between glucocorticoid use and the development of RV almost certainly reflects the greater use of glucocorticoids in patients with severe RA, precisely the subset of patients that is predisposed to developing RV. The same considerations apply to the associations reported between the development of RV and treatment with intramuscular gold, oral penicillamine, and azathioprine [22,23]. (See "Overview of cutaneous small vessel vasculitis".)

●TNF inhibitors – There is no indication that the use of TNF inhibitors predisposes patients to the development of RV, although anti-TNF therapies have been linked to the development of cutaneous vasculitis in a tiny minority of RA patients treated with these agents. The forms of vasculitis reported with TNF inhibitors are generally mild, resemble a small-vessel cutaneous vasculitis, and usually resolve with discontinuation of the drug, but lesions may recur in a minority of such patients upon resumption of a TNF inhibitor [24]. The use of TNF inhibitors has been associated with a variety of other cutaneous adverse events as well, including psoriasiform eruptions and granulomatous lesions. (See "Tumor necrosis factor-alpha inhibitors: Induction of antibodies, autoantibodies, and autoimmune diseases", section on 'Vasculitis' and "Tumor necrosis factor-alpha inhibitors: An overview of adverse effects", section on 'Cutaneous reactions'.)

Genetic risk factors — Several findings imply a role for major histocompatibility complex (MHC) molecules in the pathogenesis of RV. In addition to the overrepresentation of human leukocyte antigen (HLA)-C*03 in the context of the contribution of cell-mediated immunity to RV (see 'Cell-mediated immunity' below), other genetic studies have suggested a possible linkage between RV and the DRB1*0401 locus [25,26]. In one series, for example, at least one DRB1*0401 allele was present in 71 percent of patients with RV compared with 58 percent of those with RA without vasculitis [26]. The risk for vasculitis was increased in patients who were homozygous at this locus. However, the results were somewhat different in a case-control study in which the DRB1*0401 allele was associated with a 10- to 20-fold increase in risk for cutaneous vasculitis but no increase in risk for systemic vasculitis [27].

A 2004 meta-analysis involving 14 studies and a total of 129 patients with RV found certain genotypes of the shared epitope to be associated with RV [28]. The carriage of the shared epitope, even in two copies, was not significantly associated with the development of RV. However, three specific combinations of shared epitope-containing alleles (DRB1*0401/*0401, *0401/*0404, or *0101/*0401) were associated with an increased risk of vasculitis (odds ratios [OR] of 6.2, 4.1, and 4, respectively). Other studies have identified a variety of other genetic loci that may be associated with RV [29-32].

Cigarette smoking — Cigarette smoking is the only environmental factor that has been shown to increase risk for RV. Smoking was identified as a risk factor for more severe RA before it was recognized as a risk factor for RA itself. In a cohort of 256 RA patients from the United Kingdom, smoking within the past year was significantly more common in men with coincident vasculitis, particularly with associated nailfold lesions, compared with men without vasculitis (74 versus 39 percent), and a similar but statistically nonsignificant trend was also seen in women [33]. Subsequent studies suggested that smoking was strongly associated with an increased risk of developing RA [34,35] and with an increased risk of more severe disease [36].

A study of patients with RA from the United States and Sweden found a statistically significant increase in the risk of RV in smokers with RA (adjusted OR 2.02, 95% CI 1.01-4.02) [37]; the degree of risk was comparable to that from the presence of the HLA-DRB*04 shared epitope, though not as strong as the presence of the HLA-C3 alleles (adjusted OR 3.70, 95% CI 1.83-7.48). There was no synergy in this study between HLA-C3 and smoking for development of RV, in contrast with the strong gene-environment interaction evident between smoking and the shared epitope for development of RA [38].

There are insufficient data to establish whether RV patients have a higher risk for cardiovascular disease (CVD) than other patients with RA and whether RV and CVD in RA have a shared pathogenesis. The elevated CVD risk in RA appears to be greater than that predicted by smoking and other traditional heart disease risk factors [39,40]. (See "Coronary artery disease in rheumatoid arthritis: Pathogenesis, risk factors, clinical manifestations, and diagnostic implications".)

PATHOPHYSIOLOGY — Immune mechanisms of vascular injury have been implicated in the pathogenesis of rheumatoid vasculitis (RV). (See 'Mechanisms of vascular injury' below.)

Mechanisms of vascular injury — Postulated mechanisms of vessel wall destruction with RV include:

●Autoantibody targeting of the vessel (see 'Endothelial cell antibodies' below)

●Incidental inflammation due to deposition of immune complexes (see 'Immune complex deposition' below)

●Collateral damage due to a local antigen-driven cellular immune response (see 'Cell-mediated immunity' below)

●Cytokine- and free radical mediated endothelial injury (see 'Cytokine- and free radical-mediated endothelial injury' below)

Multiple components of the immune system could play roles in each of these proposed mechanisms.

Endothelial cell antibodies — Antibodies directed against the surface of endothelial cells are present in approximately 75 percent of patients with RV compared with 15 to 20 percent of those with rheumatoid arthritis (RA) alone [41,42]. The enhanced presence of these antibodies suggests that antibody-dependent cellular cytotoxicity (ADCC) or antibody-dependent complement-mediated cytotoxicity may destroy the vessel wall [41,43]. In theory, this damage may be accelerated by immunoglobulin G (IgG) rheumatoid factors, which are ubiquitous in those with RV (see 'Immune complex deposition' below), the level of which correlates with the consumption of complement [44]. Despite some data supporting a role for anti-endothelial antibodies in RV, little progress has been made in defining antigenic targets within the endothelium for such antibodies, and their true importance is unknown.

Antineutrophil cytoplasmic antibodies — Antineutrophil cytoplasmic antibodies (ANCA) may be seen in patients with RA and have pathogenic implications in RV, although ANCA are classically associated with three types of small vessel vasculitis: granulomatosis polyangiitis (GPA), microscopic polyangiitis (MPA), and eosinophil-associated granulomatous disease. Patients with RA frequently test positive for autoantibodies demonstrating an atypical (non-myeloperoxidase [MPO]) perinuclear (P) ANCA pattern [45]. RV is more highly associated with ANCA reactivity, with approximately half demonstrating a P-ANCA pattern, though only a fifth or fewer show positive anti-proteinase (PR)-3 or anti-MPO [46]. ANCAs have been implicated directly in the pathogenesis of small vessel vasculitis [47], so their presence in a high proportion of RA patients who develop vasculitis may implicate autoantibody-activated neutrophils in the pathogenesis of RV.

In the classic ANCA-associated vasculitides, using indirect immunofluorescence with neutrophil targets, these antibodies mostly produce patterns of cytoplasmic ANCA (C-ANCA) when binding to neutrophil PR-3, and P-ANCA when binding to neutrophil MPO. The perinuclear pattern can also be seen with antibodies binding to neutrophil proteins like lysozyme, cathepsin G, elastase, and lactoferrin, and has been classically associated with inflammatory bowel disease and autoimmune hepatitis [48].

Immune complex deposition — A relatively nonspecific form of vascular damage can occur with the deposition of circulating immune complexes, which may be detectable in the subendothelium of affected vessels in RV [49]. Circulating IgG immune complexes are a very sensitive marker for RV, being present in virtually all patients. The presence of IgA immune complexes in sera, although slightly less sensitive, may be more specific [50]. Assays for IgA immune complexes are not available routinely for clinical use, however, and their clinical utility would require further investigation.

Immune complex deposition may be found in uninvolved areas of skin in patients with RV [51]. Whether deposited immune complexes cause vessel destruction is, therefore, likely dependent upon additional pathogenic processes. Factors that affect the pathogenicity of immune complexes are likely to include:

●Differential deposition of complement components – Patients with RV demonstrate C3 and C4 aggregated with circulating immune complexes, a finding that is highly specific for vasculitis and that appears to correlate with recurrent disease [52]. Proinflammatory effects ensue once these complement components are activated. Abnormalities relating to complement processing also markedly affect the metabolism of immune complexes (see below). (See "Overview and clinical assessment of the complement system".)

●The quantity and type of rheumatoid factor – Patients with RV tend to have significantly higher levels of serum IgM, IgA, and IgG rheumatoid factors than patients with RA alone [44,53]. Higher levels of rheumatoid factor may augment immune complex formation and may decrease their clearance [54], thereby favoring deposition of immune complexes in peripheral vessels.

●Increased prevalence of IgA-containing immune complexes – Whereas IgG and IgM primarily trigger the classical complement pathway, IgA-containing immune complexes activate complement through the alternative pathway (see "Complement pathways"). In lower primates, IgA immune complexes have an increased propensity for deposition in organs other than the liver and spleen, in part because they do not bind to red blood cells as well as IgG immune complexes [55]. Individuals with predominant IgA responses to an intravascular antigenic stimulus may, therefore, form immune complexes that bypass the reticuloendothelial system of liver and spleen and that deposit in sensitive vascular beds, activating the alternative complement pathway and accelerating vascular inflammation.

●Decreased ability to process circulating immune complexes – Monocytes from patients with RV and active RA have decreased ability to bind and degrade IgG-containing immune complexes, compared with monocytes from those with their RA in remission [56]. This may be due, in part, to decreased expression of complement receptors and a decreased ability to degrade activated C3b [57]. More significantly, patients with RV have a high incidence of antibodies against C1q [58]. These antibodies may interfere with normal complement opsonization of immune complexes. (See "Overview and clinical assessment of the complement system" and "Regulators and receptors of the complement system".)

Cell-mediated immunity — A more direct role in vessel wall destruction has been postulated for activated lymphocytes than just regulation of antibody production. Circulating T cells with a unique phenotype are present in increased numbers in a large proportion of patients with RV but in a smaller proportion of patients with uncomplicated RA [59]. A large percentage of circulating CD4+ T cells in patients with RV lack expression of the important costimulation molecule, CD28. Such T cells are almost never detectable in normal individuals. These CD4+ CD28-null T cells share several features with cytotoxic lymphocytes (CTLs) and natural killer (NK) cells, releasing large amounts of interferon-gamma. They are also capable of lysing appropriate target cells through the release of cytolytic granules.

Like NK cells, these unusual lymphocytes express killer cell Ig-like receptors (KIR) activated by interaction with distinct class I major histocompatibility complex (MHC) molecules on endothelial and other cells. In particular, virtually all CD4+ CD28-null clones generated from patients with RV express the KIR2DS2 stimulatory receptor, but not the family of related KIR inhibitory receptors [59]. Expression of the KIR2D gene family on chromosome 19q13.4 is highly variable between individuals, but virtually all of 30 patients with RV expressed the KIR2DS2 gene (odds ratio [OR] of 7.96 for RV patients versus uncomplicated RA). The odds ratio for a large number of unselected RA patients without RV versus healthy controls was 0.7, suggesting that the KIR2DS2 gene is highly over-represented in RV patients.

The KIR2D receptors are selective for certain alleles at the human leukocyte antigen (HLA)-C locus of the MHC. In particular, HLA-C*03 is overrepresented in RV compared with uncomplicated RA [37]. This association appears not to represent linkage disequilibrium with the DRB1*0401 allele, another established risk factor (see 'Genetic risk factors' above). KIR2DL3 was identified as possibly having a protective effect against RV, being present at comparable levels among 23 Brazilian patients with RV and healthy controls, but not in RA patients without vasculitis [32]. Thus, the presence of a KIR2D class I-recognizing receptor and its cognate class I MHC allelic ligand represent independent risk factors for development of severe RV, setting the stage for further work to evaluate the specific role of the CD4+ CD28-null lymphocyte in mediating vascular damage [59].

CD8 T cells may predominate in cutaneous RV and other forms of vasculitis. CD8 T cells that are predominantly involved in the innate immune response have been identified that express a unique phenotype characterized by the presence of granzyme B and a lack of CD25 (interleukin [IL]-2-receptor) and programmed cell death protein 1 (PD-1). One study of 28 cutaneous vasculitis patients, including four RV patients but excluding antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis, found that the number of CD8 T cells infiltrating the vessel wall intima was significantly higher than the number of CD4 T cells [60]. In this study, antigen-activated granzyme B/CD25-double-positive cells were only rarely seen, and antigen-activated PD-1-positive CD8 T cells were not present at all, suggesting that cutaneous RV often involves a cytotoxic CD8 T-cell-mediated process with cells that are low expressors of CD25 and non-expressors of PD-1, consistent with antigen-independent killer T-cell activation.

Cytokine- and free radical-mediated endothelial injury — A potential for tumor necrosis factor (TNF) and other cytokines associated with the disease process and disease activity in RA to contribute to vascular injury is suggested several observations. Elevated levels of TNF-alpha and other cytokines cause endothelial cells to express leukocyte adhesion molecules, resulting in a capacity to bind and activate circulating leukocytes. TNF-alpha itself induces increased vascular permeability, while inducing an activated endothelial cell phenotype that includes triggering of nitric oxide (NO) synthase activity. The NO produced by these TNF-activated endothelial cells further interacts with reactive oxygen species to accelerate endothelial damage [61].

In addition to the counterregulatory role played by TNF receptors (TNF-R1 and -R2), the sympathetic nervous system may play a role in protecting against TNF-mediated vascular injury. The sympathetic nervous system is significantly activated in RA patients and may have a part in maintaining the integrity of the endothelium through production of mediators such as chromogranin A (CgA). One study found higher levels of CgA associated with severe extraarticular manifestations of RA, including RV, and demonstrated an in vitro protective effect of CgA on human microvascular endothelial cells to TNF-alpha [62].

Other factors — The role, if any, of other factors remains uncertain. Elevated IgG4 antibodies appear to be common in RA patients [63], but it is unknown whether they are involved in the pathogenesis of RA or RV. Although elevations in serum IgG4 levels have been associated with IgG4-related aortitis, this syndrome differs from rheumatoid aortitis in a number of respects both histologically and clinically. (See 'Histopathology and classification' above and "Clinical manifestations and diagnosis of rheumatoid vasculitis", section on 'Aortitis'.)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Vasculitis".)

SUMMARY

●Classification – Rheumatoid vasculitis (RV), which typically occurs in patients with longstanding, severe rheumatoid arthritis (RA), has been classified among the forms of vasculitis associated with systemic disease. Placing RV in any classification of vasculitis (eg, as a small-, medium-, or variable-vessel vasculitis) is difficult because of the variability both in the size of the vessel involved and in the histologic findings. (See 'Histopathology and classification' above.)

●Histology – The histological findings in RV may include leukocytoclastic vasculitis associated with immune complex deposition in venules, capillaries, and arterioles; and pauci-immune lesions (ie, inflammation associated with sparse deposition of immunoreactants) in medium-sized arteries and renal glomeruli. The pathology of aortitis in patients with RV is a granulomatous pattern with lymphoplasmacytic infiltration, necrosis of the medial smooth muscle, and elastic fiber loss; findings which are generally similar to other large vessel vasculitides in which aortitis is more common. (See 'Histopathology and classification' above.)

●Risk factors – The etiology of RV is unknown. Proposed triggers have included viral infections and medications used to treat RA, including glucocorticoids, conventional nonbiologic disease-modifying antirheumatic drugs (DMARDs), and tumor necrosis factor (TNF) inhibitors; however, none of these has been consistently and strongly implicated. (See 'Etiology' above.)

●Genetics – Genetic characteristics, including products of the class II major histocompatibility complex (MHC) region, may contribute to the risk of RV. Cigarette smoking is the only environmental factor that has been implicated. (See 'Genetic risk factors' above and 'Cigarette smoking' above.)

●Pathogenesis – Immune mechanisms of vascular injury have been implicated in the pathogenesis of RV, including autoantibody targeting of the vessel; incidental inflammation due to deposition of immune complexes; collateral damage due to a local antigen-driven cellular immune response; cellular activation of CD8 T cells that may be antigen-independent; and cytokine- and free radical-mediated endothelial injury. (See 'Immune complex deposition' above and 'Cell-mediated immunity' above and 'Cytokine- and free radical-mediated endothelial injury' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges John Stone, MD, who contributed to an earlier version of this topic review.