Chapter 35 Systemic Lupus Erythematosus and Infections

William R. Gilliland, MD , George C. Tsokos, MD

Despite improved treatment and overall survival of patients with systemic lupus erythematosus (SLE), infection remains a major cause of morbidity and mortality.¹^–³ Although many of the infections are attributed to common pyogenic organisms such as Staphylococcus sp and Escherichia coli, opportunistic pathogens such as uncommon bacteria, fungi, viruses, and parasites are increasingly being recognized in critically ill SLE patients.⁴ Features of SLE itself (including global dysregulation of the immune system, a hallmark of SLE) appear to play a role in the increased susceptibility of these patients.

In addition, immunosuppressive agents (most notably corticosteroids and cyclophosphamide) also significantly increase the risk for infections. Unfortunately, established predictors of impending infection or identifiers of a subgroup of patients prone to infections remain unclear. In lupus patients presenting with unexplained fever, confusion, or pulmonary infiltrates, differentiating between a disease flare and superimposed infections remains a clinical problem. This chapter reviews infections in SLE, including the impact of infections, possible pathogenic mechanisms, the spectrum of infectious agents, and diagnostic considerations.

IMPACT OF INFECTION IN SLE PATIENTS

Differences in study design make it difficult to determine and compare the mortality rates in various studies. However, as summarized in Table 35.1 it is clear that infection remains a major cause of death.^1–3,⁵^–¹⁵ As recorded in the studies listed in Table 35.1, infection and active disease are clearly the top two primary causes of death in lupus patients. When available, autopsy data is important and virtually always shows undiagnosed infections.^5,^9,¹⁶ Although the importance of atherosclerosis on late mortality is being emphasized, death due to infection tends to occur early in the disease but continues throughout the duration of the patient’s illness.¹⁷

TABLE 35.1 INFECTION AS A CAUSE OF DEATH IN SLE PATIENTS^a

In the latest multicenter European study of over 1000 patients for a 10-year period, active SLE and infections each accounted for 28.9% of deaths during the first five years of disease, whereas thrombosis (26.1%) was the leading cause of death during the last five years.¹⁸ Depending on the study cited, the frequency of major infections in lupus patients ranges from 14 to 77%.^19,²⁰ Attempts to identify risk factors for infections in lupus cohorts have often yielded conflicting results. The clinical predictors most frequently cited are summarized in Table 35.2. Interestingly, lymphopenia was not identified as a risk factor in any of these studies.

TABLE 35.2 RISK FACTORS FOR INFECTION IDENTIFIED FROM CLINICAL STUDIES

More recently, in a case-control study that investigated the risk factors associated with infection in lupus patients univariate analysis identified corticosteroid use at the time of or prior to infection, active renal disease, central nervous system involvement, and SLE disease activity index (SLEDAI) at the time of infection as risk factors. However, the use of corticosteroids was the only factor that remained statistically significant on multivariate analysis.¹⁸

In a monocentric cohort of 87 adults with SLE over a 37-year period (1960 to 1997), severe disease flares, renal disease, corticosteroid use, pulse cyclophosphamide, and/or plasmapharesis were identified as significant risk factor for infection. Multivariate analyses retained intravenous corticosteroids and/or immunosuppressants as independent risk factors for infection.³

Superimposed infections in lupus patients can also trigger disease exacerbation.^6,²² Induction^23,²⁴ and aggravation²⁵ of SLE have been reported with both parvovirus B19 and cytomegalovirus.²⁶ Superantigens released from certain common pathogens, such as mycoplasma species,²⁷ can independently activate B and T lymphocytes (causing SLE exacerbation).^28,²⁹ Polyclonal B-cell activation initiated by lipoploysaccharide from gram-negative bacteria can exacerbate autoimmune disease as well.³⁰

PATHOGENESIS OF INFECTION: DEFECTS OF IMMUNE DEFENSE IN SLE

Patients with SLE have numerous defects in both humoral and cellular immunity, which have been described in reviews.³¹^–³³ Several of these defects could partially explain the inadequacy of the immune defense in these patients (Table 35.3). Inherent defects of the immune effector cells may be supplemented by the possibility that the preexisting activation of the effector cells of immune defense in SLE render them refractory to any stimulation,^22,³⁴ and the plethora of circulating autoantibodies may interfere with various functions of the cellular and humoral arms of the immune system (Table 35.4). The resulting immune defects are not universal, and there is heterogeneity in their expression among lupus patients with variable susceptibility to different pathogens.

TABLE 35.3 IMMUNE ABNORMALITIES IN SLE THAT MAY RESULT IN DEFECTIVE IMMUNE DEFENSE ^a

PMN = polymorphonuclear cells, NK = natural killer, IgG = immunoglobulin G.

a. From reference 35.

TABLE 35.4 ANTIBODIES DIRECTED AGAINST CELLS AND CELLULAR COMPONENTS OF THE IMMUNE SYSTEM THAT MAYCONTRIBUTE TO IMMUNE DYSFUNCTION ^a

Cells	Surface Membrane Molecules
Neutrophils	CD11b/CD18 (Mac-1, CR3)
Neutrophils	HLA-I heavy chains
Monocytes/Macrophages	CR1, CR2 surface IgM and IgD
NK cells	FcγRI, FcγII, FcγIII, FcγIII HLA-DR framework epitope
T lymphocytes	β2-Microglobulin isoforms of CD45
B lymphocytes	IL-2 receptor

IgM = immunoglobulin M, IgD = immunoglobulin D, NK = natural killer, IL-2 = interleukin 2.

a. From reference 35.

Macrophage Defects

Multiple defects of the macrophage/monocyte system affect its antigen-presenting function. An important component of this defect is the diminished pahagocytic activity of lupus monocytes ³⁶ that does not increase upon stimulation in vitro with lipopolysaccharide.³⁷ Decreased tumor necrosis factor production by mononuclear cells may contribute to the deficient phagocytic ability and predisposition to bacterial infections.³⁸ Superoxide generation induced by phagocytosis by Fcγ receptor is also decreased in lupus patients.³⁹

Circulating IgG and IgM autoantibodies against this receptor may interfere with its function.⁴⁰ Furthermore, different patients may have autoantibodies directed against each of the three subclasses of Fcγ receptor, which may affect the phagocytic function of the macrophages and neutrophils.⁴¹ Last, monocytes from SLE patients have impaired capacity to adhere to plastic and ability to engulf apoptotic cells, which may indicate an intrinsic cellular defect.⁴²

Neutrophil Defects

Both quantitative and qualitative deficiencies may be seen in SLE (Table 35.5). Neutropenia, although not an American College of Rheumatology classification criterion, is a common finding in SLE.⁴³ This is at least partially immune mediated and has been correlated with the presence of complement-activating antineutrophil antibodies.⁴⁴ Antibodies to myeloid precursors have also been identified.⁴⁵

TABLE 35.5 NEUTROPHIL DISORDERS IN SLE ^a

Neutrophil Disorder	Effect on Immune Defense
Decreased chemotaxis	Decrease
Decreased phagocytic activity	Decrease
High neutrophil clustering activity	Possible decrease
Existence of anti-lactoferrin, anti-elastase, and anti-lysozyme antibodies
Increased spontaneous, and decreased after FMLP-stimulation, release of cytidine deaminase	?
Neutropenia	Decrease
Dysregulation of CD11b/CD 18 expression	Possible decrease

a. From reference 35.

The first component of neutrophil function, chemotaxis, is abnormal in SLE.⁴⁶ Several mechanisms have been identified to include reduced complement-derived chemotactic factors⁴⁷ and abnormal migration to a chemotactic stimulus.⁴⁸ Proximal white subungual onychomycosis, a rare nail infection, has been described in immunocompromised individuals (including SLE). It is associated with a defect in neutrophil chemotaxis.⁴⁹

Membrane recognition and attachment is also defective in SLE. Nived and colleagues ⁵⁰ found reduced opsonization of protein A containing Staphylococcus aureus by sera from lupus patients with active disease. Hartman and Wright⁵¹ demonstrated that some lupus patients have circulating autoantibodies directed against neutrophil adhesion glycoproteins (CD11b/CD18, Mac-1), and some of these antibodies blocked adhesion or opsonin receptor function of the Mac-1 proteins. However, the existence of these antibodies did not correlate with the presence of neutropenia.⁵² The clinical importance of other autoantibodies against cytoplasmic neutrophil components (ANCA), such as lactoferrin, is unknown.^53,⁵⁴

Defective phagocytosis by neutrophils has been noted in SLE patients since the 1970s ⁵⁵ and is more prominent in untreated than in treated patients.³⁸ Defective phagocytosis of apoptotic bodies leads to impaired disposal of autoantigens on dying cells that could enhance the autoimmune process.⁵⁶

Excess of circulating immune complexes is probably the main reason for the persistent activation of neutrophils during active diseases. Because prior neutrophil activation results in subsequent defective response to secondary stimuli, lupus patients may exhibit defective neutrophil function against superimposed infection during active phases of disease.³⁴

T-cell Defects

T cells display multiple abnormalities that are crucial in the pathogenesis and in the natural course of SLE. CD4+ T-cell lymphopenia is the most commonly observed disorder in untreated patients.³³ Lymphopenia correlates with disease flares⁵⁷ and may also contribute to the development of infections. Defective production of cytokines may also contribute to the increased rate of infections (Table 35.6). A complete description of the T-cell defects in SLE is presented elsewhere in this volume (see Chapter 10).

TABLE 35.6 THE EFFECT OF CYTOKINE ABNORMALITIES ON IMMUNE DEFENSES IN PATIENTS WITH SLE ^a

Cytokine	Abnormality in SLE	Effect on Immune Defense
IFN-gamma	Decreased production	Decrease
IL-1	Decreased production	Decrease
IL-2	Decreased production in certain patients	Decrease
IL-10	Increased production	Possible decrease
TNF	Decreased production	Decrease

a. From reference 35.

Natural Killer Cell Defects

Decreased numbers of NK cells have been reported in SLE patients that were more pronounced in patients with active disease.⁵⁸ Circulating antilymphocytic and anti-NK autoantibodies may contribute to the decreased NK cell activity.^57,⁵⁹

B-cell and Immunoglobulin Defects

Pronounced polyclonal B-cell activation and hyperglobulinemia are hallmarks of SLE. B cells seem to function adequately, as shown in several studies that reported normal antibody production and successful ^60,⁶¹ (or almost successful⁶²) immunization. However, B-cell immunologic disorders have been described in SLE (see Chapter 11). Some SLE patients have hypogammaglobulinemia,⁶³ IgG subclass deficiencies,⁶⁴ or IgA deficiency. SLE patients with IgA deficiency are especially susceptible to infections.⁶⁵

Complement Defects

Normal function of the complement system is essential for host defense. Congenital deficiency of the earlier complement proteins (C1q, C1r, C1s, and C4) has a high prevalence in SLE (75%), which is often severe. C2 deficiency is also associated with SLE, but less commonly so (30%). C3 deficiency is rarely associated with SLE.⁶⁶ The presence of normal-functioning early complement proteins may protect against SLE by allowing normal processing of immune complexes.⁶⁷ Although the vast majority of SLE patients do not have inherited complement deficiencies, consumption of complement proteins by circulating and fixed immune complexes limit the amount of complement available for host defense.⁶⁸^–⁷⁰

The number of complement receptors for C3b (CR1) on the surface membranes of erythrocytes is low in most patients with SLE, and this number is further decreased during disease flares.⁷¹ A decreased expression of CR1 has also been recognized on polymorphonuclear cells.⁷² Decreased expression of CR1 on polymorphonuclear cells resulted in an impaired recognition phase of phagocytosis.³⁸

Spleen/Reticuloendothelial System Defects

The spleen is the major component of the reticuloendothelial system (RES), and splenic dysfunction has been described in SLE patients. Several cases of functional asplenia with a high incidence of bacterial septicemia have been described.⁷³ In many cases, the functional asplenia subsides without treatment. Defective clearance of IgG-sensitized erythrocytes for the circulation by the RES correlates with disease activity.⁷⁴

Vascular Defects

Anatomic lesions in SLE patients, resulting from the impact of the primary disease or accelerated atherosclerosis, represent another risk factor for infection. Disseminated damage in the microcirculation has been found.⁷⁵ Small renal vessel injury and glomerular scarring may contribute to the increased susceptibility to urinary tract infections. Likewise, capillary vasculitis in the gastrointestinal mucosa facilitates transudation of pathogens (such as salmonella) into the bloodstream.⁷⁶ Synovitis decreases the resistance of synovium to penetration of macromolecules and consequently increases the risk for septic arthritis. The same pathogenic mechanism is probably responsible for the development of septic pericarditis with lupus pericarditis.⁷⁷ Lupus skin lesions provide an uncontrolled site of entrance for microbes.

DRUG THERAPY

Various immunosuppressive medications have been used over the past several decades to treat lupus patients. Along with corticosteroids, cyclophosphamide, azathioprine, and methotrexate remain the most commonly used medications. The impact of these agents on the immune system is fairly well established, particularly in those patients with lupus nephritis.^78,⁷⁹ Less is known about the potential infectious complications of newer immunosuppressive medications such as mycophenylate mofetil and TNFα blocking agents. Other drugs commonly used to treat SLE, such as nonsteroidal anti-inflammatory agents and antimalarial agents, are thought to have a lesser effect on immune defence.

It has been known for decades that SLE patients treated with immunosuppressive agents are more susceptible to infections than patients with other systemic rheumatic diseases treated comparably.⁸⁰ Numerous studies indicate that administration of corticosteroids and other immunosuppressive medications are also at least partially responsible for the high infection rate, although the extent of that increased risk is unclear (as evidenced in Table 35.7).

TABLE 35.7 CORTICOSTEROIDS, IMMUNOSUPPRESSIVE AGENTS, AND INFECTIONS IN SLE

Author	Findings and Comments
Ginzler et al., 1978 (6)	Corticosteroids predispose to infection. Opportunistic Infections only with high steroid dose. Azathioprine predisposes to herpes zoster infection.
Nived et al., 1985 (20)	Steroid-independent increase of bacterial infections in SLE in comparison to RA.
Rubin et al., 1985 (9)	Mean prednisone doses slightly higher in infection group (63 mg versus 50 mg).
Austin et al., 1986 (79)	Cyclophosphamide associated with localized herpes zoster infection.
Hellman et al., 1987 (16)	Fatal infections correlated with prednisone and cytotoxic therapy.
Duffy et al., 1991 (21)	Infection rate did not correlate with prednisone dose.
Oh et al., 1993 (81)	Pulse methylprednisolone and cytotoxics did not increase the risk of infection.
Janwityanuchit et al., 1993 (10)	Steroid therapy predisposed to opportunistic infections. Fatal infections were more common in cyclophosphamide-treated group.
Paton et al., 1996 (82)	Risk of major infection 20 times higher and incidence of minor infection 10 times higher in the month following a course of pulse methylprednisolone. No additional risk with azathioprine, oral or intravenous cyclophosphamide.
Pryor et al., 1996 (83)	Higher maximum corticosteroid dose (195 vs. 73 mg) in infection group. Infection occurred with equal prevalence in those patients treated with intravenous vs. oral cyclophosphamide.
Zonana-Nacach et al., 2001 (84)	Higher accumulative dose of prednisone and treatment with intravenous cyclophosphamide associated with more infections.
Noel et al., 2001 (3)	Intravenous corticosteroids and immunosuppressants were independent risk factors for infection.
Badsha et al., 2002 (85)	Low dose methylprednisolone pulse (< or = 1500 gm over 3 days) had fewer serious infection compared to high dose (> 3 gm over 3 days).

SLE = systemic lupus erythematosus, RA = rheumatoid arthritis.

Immunosuppressive medications have a dual effect on the immune system in SLE patients. Suppression of abnormally functioning cells may normalize some aspects of the immune system. For example, neutrophil migration is significantly depressed in untreated SLE patients but normal in the treated patients,⁸⁶ and treatment with a high dose of pulse methylprednisolone enhances Fcγ receptor-mediated mononuclear phagocyte function.⁸⁷ The increased infection rate in some clinical studies of patients receiving immunosuppressive therapy may be attributed to active or advanced disease.

High doses of corticosteroids and cytotoxics drugs are used almost exclusively in patients with organ-threatening disease.² When comparing the rate of infection between treated and untreated patients, the contribution of disease activity is often ignored. The studies from the National Institutes of Health (NIH)^78,⁷⁹ included only patients with lupus nephritis without end-stage renal disease, which may independently contribute to the increased infection rate.⁸⁸ Although these studies failed to demonstrate an additive effect of cytotoxics drugs and prednisone in increasing the infection rate, they showed that cyclophosphamide increases the incidence of localized herpes zoster infection.⁷⁹

Patients treated with only low doses of prednisone (<10 mg/day) also have an increased infection rate.^6,⁸⁹ Interestingly, the same dose did not increase the infection rate in patients with rheumatoid arthritis treated with similar doses.⁹⁰ However, Nived and colleagues concluded that when matched for corticosteroids dose rheumatoid arthritis patients had only a slightly reduced risk of infection rate compared to SLE patients.²⁰

Although the optimal treatment for lupus nephritis remains unclear, the use of mycophenylate mofetil (MMF, which selectively inhibits activated lymphocytes and renal mesangial cells) is being used more frequently as an agent for induction and maintenance in SLE nephritis. In a recent study of proliferative nephritis, Contreras and colleagues compared the infectious risks of pulse cyclophosphamide as an induction agent, followed by quarterly intravenous cyclophosphamide, daily azathioprine, or daily MMF. MMF and azathioprine maintenance therapy were associated with a significantly lower risk of infections (2% in each MMF and azathioprine, and 25% in cyclophosphamide).⁹¹ The safety and efficacy of using TNFα blockade is SLE is not yet known, although a recent small case series using infliximab suggests that it may be relatively safe.⁹²

THE SPECTRUM OF INFECTIONS

Although all major types of pathogens (including bacteria, mycobacteria, viruses, fungi, and parasites) are known to cause infections in lupus patients, it is difficult to determine the precise contribution of each. Several reasons may explain part of this difficulty. Isolation of pathogens may be difficult because of prompt initiation of broad-spectrum antibiotics. The empiric early use of broad-spectrum antibiotics that are effective against gram-negative organisms has changed the pattern of infections in immunosuppressed patients over the last decade.

Gram-positive infections are now more common than gram-negative ones, and fungal infections are the leading factor in the morbidity and mortality in immunocompromised patients with cancer.⁹³ It is also possible that older studies underestimated the accurate number of infections or at least the number of causative organisms. Futrell and colleagues⁹⁴ showed that autopsy in lupus patients with central nervous system disease frequently showed unsuspected systemic and brain infections.

In addition, newly recognized pathogens were not included in earlier studies. For example, Chlamydia and Legionella species, relatively common causes of pneumonia now, were not known pathogens of the respiratory system 30 years ago. Hellmann and colleagues ¹⁶ reported that only 10% of fatal opportunistic infections in lupus patients were diagnosed before autopsy. Therefore, the spectrum of infections in lupus patients has changed.

Although the methodology of studies investigating the site of infection in SLE patients varies widely, Table 35.8 provides evidence that the most common sites of infection in decreasing frequency are the lung, bladder, blood, skin, central nervous system (CNS), and vagina. Pneumonia is the most common major infection of hospitalized SLE patients.²¹ Bates and colleagues⁹⁵ showed that the determination of the pathogen is possible in only 50% of hospitalized and community-acquired infections, despite the use of modern laboratory techniques, invasive procedures, and autopsy. Futrell and colleagues showed that autopsy in SLE patients with CNS disease frequently shows unsuspected systemic and brain infections in most patients.⁹⁴ In addition, CNS infections frequently mimic signs and symptoms of SLE,⁹⁶ delaying treatment that may lessen morbidity and mortality.

TABLE 35.8 SITES OF INFECTION

Infections from Common Bacteria

Common bacteria cause most infections in SLE patients.¹⁹ Staphylococcus aureus and Escherichia coli are the leading bacterial pathogens. Studies evaluating the most common bacterial pathogens vary widely in terms of diagnostic evaluation and patient population. Table 35.9 provides an overview of the most common bacteria. Compared to outpatients, hospitalized patients presented with a broader spectrum of pathogens (with a predominance of Pseudomonas aeruginosa in major and E coli in minor infections).²¹ Aside from these typical bacterial pathogens in patients with defective immunity and frequent hospitalization, several bacteria are worthy of special consideration.

TABLE 35.9 PREVALENCE OF BACTERIAL ORGANISM IN SLE