Introduction

The primary systemic vasculitides (PSV) refer to a group of autoimmune rheumatic diseases (AIRDs) characterized by vascular inflammation, leading to ischemic events and end-organ damage [1]. Despite advances in the treatment of these conditions, infections remain an important cause of morbidity and mortality in people living with vasculitis. Mitigation strategies (e.g., antibiotic prophylaxis) are currently employed in clinical care; however, certain aspects such as duration, risk stratification, and optimization of these interventions remain unanswered. In this review, we will summarize recent findings on the epidemiology of infections in adults with PSV, as well as existing evidence on mitigation strategies. For the purpose of this review, and from here onwards, we will refer to severe infections (SI) as those leading to hospitalization, requirement of intravenous antibiotics, and/or death [2].

Epidemiology of Infections in Primary Systemic Vasculitides

Anti-neutrophil Cytoplasmic Antibody-associated Vasculitis

Anti-neutrophil cytoplasmic antibody (ANCA)–associated vasculitis (AAV) are the most common form of small vessel vasculitis. Recent longitudinal cohorts have reported high infection incident rates, particularly during the first year after diagnosis, with decreasing incidence during follow-up (Table 1) [3, 4]. A systematic review and meta-analysis that included pooled data from 21 studies (1,284 individuals with AAV) focused on the incidence of infections primarily during the maintenance period. The cumulative incidence of SI was 15.99% (95% CI: 6.95%-27.53%) during the total follow-up period (induction and maintenance) and 7.62% (95% CI: 4.43%-11.43%) in the maintenance phase [11]. Infections remain a common cause of mortality, accounting for 4.1-22.4% of deaths and the leading cause of excess mortality in individuals with AAV [3, 4, 12, 13] (Table 1).

Table 1 Cohort studies reporting incidence of infections and associated risk factors in individuals with primary systemic vasculitides
Full size table

Other Forms of Systemic Vasculitides

Infections in individuals with giant cell arteritis (GCA) are frequent, particularly during the first year after diagnosis (Table 1) [5, 71) [8]. For individuals with TAK, a systematic review and meta-analysis that included 517 patients on biologics, reported an incidence of infection of 6% (95% CI: 2%-10%) as the most common adverse event [14]. Compared to global infection rates of tuberculosis [15], a high prevalence of tuberculosis of 31.27% (95% CI: 20.48%-43.11%) has been reported in a meta-analysis of observational studies of patients with TAK [16]. Data on the incidence of infections in other forms of vasculitis is scarce (Table 1) [9, 10].

Patient-related Risk Factors for Infection

Age and Frailty

Older age is a risk factor for poor outcomes, including SI, in individuals with PSV [7, 10, 17,18,19]. In the PEXIVAS study, individuals with AAV who developed a SI within the first year after diagnosis were more likely to be older (68.2 ± 13.6 vs 59.7 ± 15.8 years, p =0.016). A recent multicentric study of 1,004 individuals with GCA showed that SI were more frequent in individuals aged 65-79 (OR 11.90 [95% CI: 1.69-83.63]) and ≥80 years (OR 13.71 [95% CI: 1.59-117.91]), when compared to individuals ≤64 years old [6]. Similar observations were made in a recent nationwide Danish study of 17,773 individuals with GCA. Age ≥70 years was associated with a higher incidence of overall infections, including SI, during the first year of treatment [5, 20]. Older age at disease diagnosis has also been associated with an increased risk of SI in a large cohort of individuals with anti-GBM disease [10].

Although multiple studies have utilized chronological age (i.e., age) as a predictor of outcomes, chronological age might not accurately reflect biological age, which is possibly a better predictor of morbidity and mortality [21]. This is even more important for older adults, as shown in studies of community-dwelling older adults. Frailty, a syndrome characterized by a decline in physiologic reserve and homeostasis with a subsequent increase in vulnerability to stressors, has been shown to provide a holistic assessment in older adults and better predict outcomes. In a large claims data analysis that included two different cohorts of individuals with GCA, moderate/severe frailty (as measured by a claims-based frailty index) had a two- and four-fold increased risk for SI when compared to pre-frail individuals (HR 2.51 [95% CI: 1.38-4.57] in federal and HR 4.54 [95% CI: 1.44-14.29] in commercial database, respectively) [7•]. Using the same claims-based frailty index, frailty was associated with an increased risk of SI (HR 8.62 [95% CI: 2.08-35.66] vs non-frail) in a large cohort of older adults with AAV and appeared to be a stronger risk factor than older age (age ≥75 years HR 3.27 [95% CI: 1.94-5.53], vs age 65-74) [22]. Frailty seems to be prevalent in individuals with multiple forms of vasculitis, even those <65 years old, and associated to a higher frequency of infections [23]. Although further data is needed to better understand the role, assessment, and incorporation of frailty in the care of individuals with AAV, these data highlight the potential use of frailty measurement in risk assessment and the development of personalized interventions toward risk mitigation.

Comorbidities

Individuals living with PSV are at increased risk of comorbidities, resulting as a complication of the disease itself or the associated treatments. Renal (i.e., chronic kidney disease) and pulmonary comorbidities (e.g., chronic obstructive pulmonary disease) have been associated with an increased risk of SI [4, 24, 25]. A higher burden of comorbidity, usually characterized by higher counts in cumulative scores such as the Charlson Comorbidity Index (CCI), is associated with an increased risk of SI [26]. In a recent multicentric retrospective study that included 162 individuals with AAV with a mean follow-up of 5.4 years, patients who developed a SI were more likely to have a higher composite Rheumatic Disease Comorbidity Index (median [IQR]: 1 [0 – 2] vs 1 [1–3], p =0.03) [4]. In an analysis of hospitalized patients with GCA due to SI, a CCI ≥2 (present in 50.7% of individuals) was associated with higher in-hospital mortality (OR 1.38 [95% CI: 1.09-1.76]) [8]. Higher CCIs were also associated with an increased risk of incident infections, including SI, in a large national Danish cohort of individuals with GCA [5, 20]. With the increased risk of multimorbidity in individuals with PSV, especially AAV, this data highlights the value of its assessment to identify patients at higher risk for SI and other poor outcomes [27].

Treatments and Risk Factors for Infection

Glucocorticoids

Glucocorticoids remain a main driver of infection, among other associated toxicities. Recent AAV trials have focused on the use of low glucocorticoid regimens, including the use of other immunosuppressive medications for glucocorticoid-sparing effect [28,29,30]. In the LoVAS trial [28•], individuals with AAV who received a reduced-dose glucocorticoid regimen alongside rituximab induction had a lower incidence (-12.8% [95% CI: -24.2%--1.3%]) of SI, compared to those randomized to high-dose glucocorticoid treatment. Similar observations were made in the PEXIVAS trial [29•]. PEXIVAS randomized 704 individuals with severe AAV through a 2-by-2 factorial design to a standard vs a reduced dose of glucocorticoids along with or without plasma exchange. No difference between the arms of the two different interventions was observed for the primary outcome (i.e., death or occurrence of end-stage renal disease [ESRD]); however, those randomized to the reduced-dose glucocorticoid arm experienced fewer infections during the 52-week study period (incidence rate ratio 0.69 [95% CI: 0.52-0.93]). Similar observations have been reported in recent contemporary observational AAV cohorts [4]. For GCA, a recent Danish national cohort study of 17,773 individuals with GCA reported that a greater cumulative dose of glucocorticoids (>6 g) increased susceptibility to overall infections (OR 1.47 [95% CI: 1.29-1.67]) and infection-related hospitalizations (OR 2.58 [95% CI: 2.02-3.30]), compared to a minor dose (<2 g) [5]. Glucocorticoids remain an important modifiable risk factor for infections in patients with PSV. Although a recent study has raised concern about the applicability of reduced-dose glucocorticoid protocols to some patients with AAV [31], glucocorticoid minimization should be attempted in most cases, as noted by recent treatment guidelines [32, 33].

Rituximab and Hypogammaglobulinemia

A meta-analysis that included 1,434 individuals with AAV receiving treatment with rituximab reported an incidence of SI of 6.5 cases per 100 PY [34••]. The most common infections were bacterial (9.4%) and mortality related to infection was 0.7%. Higher cumulative doses of rituximab were associated with an increase in the prevalence of infections. In a recent post-hoc analysis of the RAVE trial, 18/22 (81.8%) of the SI reported occurred during the first 6 months [35••]. Higher baseline serum IgM levels (HR 1.005 [95% CI: 1.002-1.009]) were associated with an increased risk of SI, while higher baseline CD19+ B cell count (HR 0.995 [95% CI: 0.991-0.9999]) and use of trimethoprim-sulfamethoxazole (TMP-SMX) (HR 0.232 [95% CI: 0.087-0.623]) with a decreased risk of SI.

The risk of hypogammaglobulinemia associated with rituximab, especially IgG levels <6 g/L at baseline and after treatment, constitutes one of the primary mechanisms for SI among AIRDs [36, 37]. Different thresholds have been utilized in the literature to define hypogammaglobulinemia. The incidence of hypogammaglobulinemia seems to be more frequent in AAV (<6 g/L in 40% and <5 g/L in 28.1%) [38, 39] compared to other AIRDs (RA: <7 g/L in 20%; systemic lupus erythematosus (SLE): <5g/L in 18.2%) [38, 40]. In a longitudinal observational study [39] of individuals with AIRDs, hypogammaglobulinemia was most frequently observed in individuals with AAV (40.0% vs 7.9%, p =0.008), as well as a greater reduction in IgG levels (11.3 [3.8 – 29.7] vs 2.3 [0–18.2] g/L, p =0.047). In an analysis of 657 individuals with newly diagnosed or relapsing AAV from the national Japanese registry (J-CANVAS), IgG <5 g/L levels were independently associated with the development of SI (HR 1.75 [95% CI: 1.03-3.00]) [41•]. It is important to note that only 37.3% received rituximab in the group of individuals developing hypogammaglobulinemia. These observations are in line with the recent report of the RITAZAREM trial that found no difference in the incidence of hypogammaglobulinemia between rituximab and azathioprine, highlighting the importance of monitoring IgG levels even when using non-B-cell depleting therapies (BCDT) [42]. Interestingly, individuals with sustained B cell-depletion (longer than median time to repopulation of 39 months) post rituximab therapy exhibit a higher risk of SI in the longer term [43]. Sustained B-cell depletion was associated with higher eGFR and female sex in this cohort.

Other Biologics

A systematic review and meta-analysis of the use of biologics (i.e., tocilizumab, abatacept, adalimumab, etanercept) along with glucocorticoids versus glucocorticoids alone for the treatment of individuals with GCA, showed no difference in severe adverse events (including SI) (OR 0.81 [95% CI: 0.54-1.20]) [44]. Reports of infections associated with biologic use in other PSVs are less common. In a trial of tocilizumab in TAK, although infections were more frequent in the treatment arm compared to placebo (9/18 [50%] vs 6/18 [33.3%]), no SI were reported. Few other studies have reported low rates of infections in individuals with TAK on biologic treatment [45, 46].

Plasma Exchange

The PEXIVAS trial [29•] has led to further discussion of the risks and benefits of plasma exchange (PLEX). Although PLEX still has a role in the management of specific subgroups (i.e., high risk of ESRD, severe diffuse alveolar hemorrhage), a recent meta-analysis [47] that included 1,060 participants of 9 clinical trials showed that PLEX increased the risk of SI at 12 months (RR 1.27 [95% CI: 1.08-1.49]). The increase in SI caused by PLEX occurs earlier in the disease (i.e., around time of the intervention) has been attributed to several factors, including the removal of immunoglobulins, the use of blood products, and the need for central venous access and an intensive care unit environment [48]. The increased risk of infections associated with PLEX needs to be weighed against the potential benefits, and this needs to be assessed on a case-by-case basis.

Avacopan

Avacopan, a C5a receptor inhibitor, has been recently approved as an adjunctive treatment in AAV [32]. An analysis combining data from the ADVOCATE, CLASSIC, and CLEAR trials showed that participants in the placebo arms had a higher rate of infections and a more significant decrease in white blood count when compared to those receiving avacopan [30, 49,50,51]. In a recent safety trial of 30 patients, SI were presumed to be secondary to avacopan in only 3.3% of patients [52]. Although emerging data seems favorable, further studies are needed to better delineate the incidence of infections associated with avacopan use.

Infection Risk Mitigation

Aside from risk assessment and identification of modifiable and non-modifiable risk factors, mitigation strategies include screening, vaccination, and the use of prophylaxis in selected individuals (Fig. 1).

Fig. 1
figure 1

Risk factors for infection and mitigation strategies in patients with primary systemic vasculitis (Created with Biorender.com). CKD = chronic kidney disease; PLEX = plasma exchange; IgG = immunoglobulin G; HBcAB = hepatitis B core antibodies; HBsAG = hepatitis B surface antigen; PJP = pneumocystis jirovecii pneumonia; TMP-SMX = trimethoprim-sulfamethoxazole; RTX = rituximab; CYC = cyclophosphamide; IVIG = intravenous immunoglobulin

Full size image

Infection Screening

Individuals with AIRDs receiving treatment with biologics therapies, other immunosuppressive agents, or medium-high dose glucocorticoid (>7.5 mg/day prednisone equivalent) are at higher risk for reactivation of hepatitis B (HBV) or hepatitis C virus (HCV) [53]. With the advent of immunosuppression, reactivation has also been reported in patients previously exposed to HBV or HCV. Use of BCDT (e.g., rituximab, obinutuzumab) carries a risk of reactivation of >20% in individuals who are HBsAg and anti-HBc positive [54]. Viral hepatitis screening should be performed in all patients, ideally prior to starting treatment. HBV vaccination should be given before initiation of treatment to patients with initial HBV testing negative, and adapted regimens have been suggested for those receiving immunosuppressive treatment [54, 55].

The screening of patients for latent and active tuberculosis prior to initiation of biologic or glucocorticoid-sparing immunosuppressive therapies is suggested for patients living in endemic areas or other higher-risk populations such as those with alcohol use disorder, tobacco use, or household TB contacts [53, 56]. Interferon-gamma release assay (IGRA) has been shown to have a lower false-negative and false-positive rate compared to tuberculin skin test (TST), in patients with rheumatic diseases before starting biologic therapy, particularly in patients treated with glucocorticoids [57]. Different cut-offs for TST are recommended for immunosuppressed individuals (e.g., >15 mg/day prednisolone equivalent for >1 month, use of tumor necrosis factor-alpha inhibitor [TNFi]). In the occurrence of latent or reactivated tuberculosis, patients should be treated before starting a biologic or glucocorticoid-sparing immunosuppressive therapy. Immunosuppressive treatment may be commenced after at least 1 month of anti-tuberculosis treatment with adequate monitoring every 3 months thereafter [53].

Vaccination

COVID-19

Adequate vaccination for people living with AIRDs remains challenging primarily due to differences in seroconversion observed in patients receiving different immunosuppressive drugs, especially rituximab and certain anti-metabolites (e.g., cyclophosphamide, methotrexate, azathioprine) [58•]. Vaccination should be encouraged in all patients with AIRDs, including PSV, although in several instances, including but not limited to life-threatening disease, AIRD’s treatment should be started swiftly, regardless of vaccination status. Despite the fact that flares of disease after vaccination have been reported, these are rarely severe, and the benefits of vaccination outweigh the potential risks, especially in this higher-risk population [58•]. The American College of Rheumatology (ACR) guidelines recommend that patients with AIRDs receive 3 doses of the primary COVID-19 vaccination series and 2 boosters (a total of 5 doses) [59•]. Holding off specific immunosuppressive treatments for 1-2 weeks after each COVID-19 dosage is recommended to enhance vaccine response, and in the case of rituximab, it is recommended that vaccine series are started 4 weeks before the next rituximab dose [59•].

Influenza

Yearly influenza vaccination with either high-dose or adjuvanted influenza vaccination is recommended for all patients with AIRDs 65 years or older and those 18 years or older on immunosuppressive treatments [59•]. If disease allows, methotrexate should be held for 2 weeks after vaccination to enhance vaccine response [60, 61]. For patients receiving rituximab, influenza vaccine should be given on schedule given its seasonal occurrence, and rituximab should be delayed for at least 2 weeks after vaccination. Rituximab is known to decrease the response to influenza vaccination [62, 63]. All other immunosuppressive treatments should be continued per the usual schedule [59, 64, 65]. Influenza vaccine should be recommended irrespective of disease activity [59•].

Pneumococcal Pneumonia

Pneumococcal vaccination is strongly recommended to all individuals with rheumatic diseases with less than 65 years of age on immunosuppressive treatments. All adults above 65 years of age, irrespective of comorbidities, are recommended pneumococcal vaccine [59•]. Current recommendations include a combination of different existing vaccines (i.e., PCV15 or PCV13 followed by PPSV23 or PSV20, or PCV20 followed by PPSV23) [65]. Methotrexate and rituximab may reduce vaccine response [62, 63], which is a concern in individuals with PSV such as AAV, who are at higher risk for infection due to immunosuppression and comorbidities (i.e., lung disease). A recent multicenter randomized label study compared two “reinforced” pneumococcal vaccine regimens to standard regimens in individuals with AAV receiving treatment with rituximab [66•]. Preliminary results of this study showed that one of the “reinforced” regimens that consisted of a double dose of PCV13 followed by a single dose of PPSV23 7 days after yielded a significantly improved antibody response. Although further studies are needed, these results in combination with those for COVID-19 vaccination highlight the need for innovative regimens in individuals with PSV on chronic immunosuppressive therapy.

Role of Prophylaxis Treatment

Pneumocystis Jirovecii

Individuals with PSV, especially AAV, have been associated with a high incidence of Pneumocystis jirovecii pneumonia (PJP) of up to 12.14 cases per 100 PY and a high associated mortality [67, 68,  69••]. More recent studies, including a recent systematic review and meta-analysis of 1,434 patients, have reported incidence rates as low as 1.1 per 100 PY [34••]. However, the risk of PJP might vary during different phases of treatment, even despite the use of maintenance medications, and this might be possibly associated with the use of high-dose glucocorticoids. In a recent retrospective study of 1,461 patients with AAV, only 10 cases of PJP were observed during the induction phase (incidence rate 1.50 cases per 100 PY) and 5 cases during the maintenance phase (0.21 cases per 100 PY) [70•]. Although glucocorticoid doses were not captured, dosing differences between induction and remission phases could explain the difference in PJP incidence. PJP prophylaxis use was only observed in 30% of rituximab or cyclophosphamide dosing episodes.

A retrospective multicenter study from South Korea reported that trimethoprim-sulfamethoxazole (TMP-SMX) was associated with a significant reduction in the incidence (HR 0.07 [95% CI: 0.01-0.53]) and mortality (HR 0.08 [95% CI: 0.0006-0.71]) in patients with rheumatic disease receiving prolonged, high-dose steroids. The number needed to treat (NNT) (52 [33–124]) for preventing one PJP case was lower than the number needed to harm (NNH) (131 [55–∞]), indicating that the benefit of TMP-SMX was greater than the risk of potential harm to the patient [68]. More recently, a retrospective study from South Korea that included 818 patients receiving treatment with rituximab reported a significant reduction in the risk of PJP infection associated with TMP-SMX (HR 0.11 [95% CI: 0.03-0.43]), giving an NNT of 146 [69••]. Although this number was higher than the NNH of 86, NNT decreased to 20 in those patients receiving high-dose glucocorticoids (i.e., prednisone ≥30 mg/day), aligning with the observations of differences in incidence rates during induction and maintenance phases. Finally, a recent post-hoc analysis of the Rituximab versus Cyclophosphamide for ANCA vasculitis (RAVE) study, not only showed a decreased risk of PJP infections associated with TMP-SMX use but also showed an overall lower frequency of SI (HR 0.30 [95% CI: 0.13-0.69]) [25].

The 2021 ACR and Vasculitis Foundation guidelines for the management of AAV conditionally recommend PJP prophylaxis for those patients who are treated with rituximab or cyclophosphamide and receiving moderate-dose glucocorticoids (>20 mg/day) or higher in combination with other immunosuppressive therapy [33]. The 2022 update EULAR recommendations for the management of AAV also recommends the use of TMP-SMX as PJP prophylaxis in patients receiving rituximab, cyclophosphamide and/or high-dose glucocorticoids [32]. This is also in line with prescribing information for rituximab that recommends prophylaxis for PJP for equal or more than 6 months after the last rituximab dose for patients with GPA or MPA [33].

While there is not enough information to guide the total duration of PJP prophylaxis, these observations and the risk of potential harm associated with TMP-SMX need to be taken into consideration to guide risk assessment and shared decision making. PJP prophylaxis during the first 6 months of therapy has been associated with a higher risk of adverse events including, rash, nephropathy (e.g., acute kidney injury, renal failure), and hyperkalemia [70•]. Although most of these studies have focused on TMP-SMX, alternative therapies such as dapsone, atovaquone, and pentamidine can also be associated with potential side effects. While benefit of PJP prophylaxis will certainly outweigh risk in certain individuals (e.g., ≥30 mg/day of prednisone equivalent, hypogammaglobulinemia, structural lung disease), the risk/benefit ratio needs to be reassessed in all patients [32].

For non-AAV PSV, the risk and benefit of PJP prophylaxis remain unclear. Existing studies have shown differences in the risk of PJP infection among systemic rheumatic diseases, and this difference might also apply to individuals with distinct PSV. Recent studies in individuals with GCA have shown a low risk of PJP infection, with an incidence rate from 0-0.08 cases per 1,000 PY, despite the concerns for older age and high doses of glucocorticoids. These studies show that the risk of PJP is not the same across individuals with PSV [71•].

Antivirals

Prophylactic antiviral therapy for HBV with tenofovir or entecavir is strongly recommended for all individuals positive for HBV core antibody (HBcAb) and HBsAg receiving immunosuppressive treatment [72]. Treatment should be started 1–2 weeks before starting therapy. Patients should be monitored for up to 12 months after cessation of anti-HBV therapy. If the patient is positive for HBcAb and negative for HBsAg, frequent monitoring every 1–3 months with ALT, HBV DNA, and HBsAg can be performed for evidence of reactivation. Risk of reactivation might be higher with rituximab compared to cyclophosphamide. Prophylactic therapy over frequent monitoring is favored for patients initiating rituximab who are HBcAb positive, regardless of HBsAg status [73]. Prophylaxis is not routinely recommended in individuals receiving other immunosuppressive treatments (i.e., methotrexate, mycophenolate mofetil, and azathioprine) due to lower risk of reactivation [74]. The exact duration of prophylaxis in patients with rheumatic diseases remains unclear; current guidelines recommend continuing it for at least 6 months (or for at least 12 for patients receiving anti-CD20 therapies) after cessation of immunosuppression [73]. If HBV reactivation occurs during antirheumatic therapy, immunosuppression may be continued if there is only mild hepatitis; however, for patients with moderate to severe hepatitis, immunosuppressive therapy may need to be held [75].

Intravenous immunoglobulin

Hypogammaglobulinemia has been observed in individuals with AAV, most commonly but not exclusively associated with rituximab use. Deficiency of IgG is a more important risk factor for infection, as compared with IgA or IgM deficiency [76•]. The determining factors for the development of clinically significant hypogammaglobulinemia during or BCDT include a pre-existing low IgG level and previous or concomitant immunosuppressive therapy [76•]. The decision to start immunoglobulin replacement therapy (IGRT) should be guided by the degree of hypogammaglobulinemia, serious or recurrent infections, demonstration of impaired antibody responses to polysaccharide antigens, and poor response to antibiotic prophylaxis. Immunoglobulin levels should be checked prior to initiation of BCDT and repeated every 6–12 months for the duration of BCDT and a minimum of one year after stopping treatment. Initial dosing of IGRT should be 0.4 g/kg/month, then it can be adjusted according to IgG trough levels, infection frequency, and clinical response, as well as the decision to continue IGRT should be reviewed annually based on clinical and laboratory parameters. Co-management with an immunologist is recommended [32, 76

Conclusion and Future Research Directions

Infection remains to be one of the main causes of morbidity and mortality among patients with PSV, highlighting the importance of risk assessment, and the implementation of mitigation strategies (Fig. 1). Glucocorticoid reduction, either through the use of reduced-dose regimens or adjunctive therapies (i.e., avacopan) is an important modifiable risk factor for infection. Future efforts should be focused on improving efforts for risk stratification and implementation of mitigation strategies in clinical practice, to better guide individualized shared decision-making process in the care of individuals with PSV. To date, most existing information on infections comes from the AAV literature, and information on other forms of PSVs is needed, given the limitations of extrapolating observations from AAV to individuals with other forms of PSV.