An update on risk factors for relapse in antineutrophil cytoplasmic antibody-associated vasculitis

Summary

Ongoing therapeutic advances in antineutrophil cytoplasmic antibody-associated vasculitis (AAV) have significantly reduced the risk of death in AAV, but 30%–50% of patients still relapse. Relapse is a major problem in these diseases, leading to increased morbidity and mortality. It is, therefore, necessary to find predictors of relapse at the end of the remission induction and maintenance phases in order to personalize treatment.

AAV is a chronic relapsing disease that leads to increased morbidity and mortality. It is essential to identify relapse predictors and implement personalized treatment to reduce relapse and mortality. This review summarizes the current predictors of relapse of AAV, including ANCA types, serological characteristics, organ involvement, and treatment options, to better guide clinical practice.

Graphical abstract

Graphical Abstract.

Introduction

Antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) is a group of autoimmune diseases characterized by inflammation and fibrinoid necrosis in small and medium-sized blood vessels. Based on pathological and clinical features, AAV is classified into granulomatosis with polyangiitis (GPA), microscopic polyangiitis (MPA), eosinophilic granulomatosis with polyangiitis (EGPA), and renal-limited vasculitis [1]. Serological markers of AAV are classified as proteinase 3 ANCA (PR3-ANCA) disease and myeloperoxidase ANCA (MPO-ANCA) disease [2]. Therapeutic advances have changed AAV from a highly fatal disease to a chronic disease with a relapsing course. Clinical relapse is common, with more than 50% of patients experiencing a relapse within 5 years [3–5]. However, long-term maintenance therapy often leads to unnecessary treatment and increases the risk of serious adverse events, which may outweigh the benefits of preventing relapse. Adverse events, particularly infections, have been reported to be the leading cause of death in AAV patients within the first year [6, 7]. Therefore, a better understanding of risk factors for relapse would help to stratify patients; more treatment for those at high risk may prevent the development of relapse, while minimizing treatment for those at low risk may reduce the risk of adverse events, particularly infectious complications, and drug toxicity. Many factors have been found to be associated with the relapse of AAV in various cohort studies and long-term follow-up studies. This article provides an overview of the risk factors for relapse in patients with AAV to better guide clinical practice.

Definition of relapse

The relapse was defined as the reoccurrence due to active inflammation recommended by the European League Against Rheumatism (EULAR), and it was classified as a minor or major relapse [8]. A major relapse was defined as reoccurrence of potential organs-threatening or life-threatening disease activity after remission, which could not be treated with an increase of glucocorticoid alone; a minor relapse as a disease activity that was neither organs-threatening nor life-threatening. In order to reduce heterogeneity and improve the comparability of results, the use of uniform definitions in clinical trials could be beneficial.

Known risk factors for relapse

Types of ANCA

There are two ways to test for ANCA, using indirect immunofluorescence (IIF) and enzyme-linked immunosorbent assay (ELISA) [9, 10]. The serotype of ANCA is divided into cytoplasmic ANCA (c-ANCA) and perinuclear ANCA (p-ANCA) by IIF and into MPO-ANCA or PR3-ANCA by ELISA respectively. In a prospective cohort of 16 GPA patients with PR3-ANCA positive, it was suggested that the predictive values of ≥50% for relapse within 6 or 12 months for 4-fold rises in C-ANCA titer by IIF or >75% increase in PR3‐ANCA level by direct ELISA. The study also indicated that PR3-ELISA might be superior to IIF in predicting relapses. Of those, it used identical criteria for relapse, a standardized interval for sequential ANCA measurements, and measuring two sequential samples in the same assay [9]. Accumulating studies have shown that GPA patients have a lower mortality rate and a higher relapse rate, whereas MPA patients have a higher mortality rate and a lower risk of relapse after remission [11]. EGPA has unique clinical features and is a rare form of AAV. Only 30%–40% of patients with EGPA were ANCA positivity, and more than 90% of these cases were MPO-ANCA positivity [12]. This disease was different in many aspects from the other vasculitis and was often excluded from AAV studies. EGPA patients often experience relapses during glucocorticoid tapering [13].

It is a consistent finding from various cohort studies and long-term follow-up of clinical trials that PR3-ANCA-positive or c-ANCA-positive patients, as compared with MPO-ANCA-positive or p-ANCA-positive patients, were associated with a higher risk of clinical relapse (Table 1) [38–40]. In the cohort by Córdova-Sánchez et al. [25], 62 patients with renal biopsy-proven AAV were analyzed, and compared with patients with MPO-ANCA, patients with PR3-ANCA had a significantly increased risk of relapse (hazard ratio, 2.93; 95% confidence interval [CI], 1.20–7.17; P = 0.019). In a study of ANCA-positive retroperitoneal fibrosis (RPF), all patients who relapsed were PR3-ANCA-positive [41]. Long-term follow-up results from the Wegener’s Granulomatosis-Entretien (WEGENT) study showed that PR3-ANCA positivity (hazard ratio, 1.99; 95% CI, 1.29–3.08; P = 0.002) was independently associated with relapse rate in multivariate analysis [24]. A study from China found no predictive value for PR3-ANCA, but only 9.5% of patients in this cohort were PR3-ANCA-positive, making it difficult to assess the relationship between PR3-ANCA and relapse [20]. A previous report found that c-ANCA titer at 24 months (relative risk, 2.6; 95% CI, 1.2–5.0; P = 0.02) was significantly associated with relapse within 5 years of diagnosis [42]. The possible mechanisms may be related to persistent PR3 expression on the membrane of neutrophils. Rarok et al. [43] demonstrated that patients with PR3-ANCA-associated vasculitis had an increased expression of PR3 on the surface of resting neutrophils compared with healthy controls, and the increased expression of PR3 on the membrane of neutrophils was associated with relapse. Persistent granulomata could be another reason for relapses in GPA patients [44]. Granulomata could be a potential lymphoid tissue to maintain autoantibody production; meanwhile, granulomata provided an environment for Staphylococcus aureus colonization, S. aureus superantigen could promote the PR3-affinity B cells maturation and differentiate to plasma cells, leading to PR3-ANCA production [45]. As GPA was more frequently associated with PR3-ANCA than with MPO-ANCA, while the opposite was true for MPA [46], the question was whether the ANCA specificity or the clinical phenotype was related to the relapse risk. Numerous studies have indicated that the serotype of ANCA was a better predictor of treatment response and relapse than the clinicopathological phenotype.

Table 1.

The risk factors for relapse in ANCA-associated vasculitis trials

Author information	Follow-up time, months	Study design	Type of vasculitis (%)	Risk factors	Risk of relapse
Girard et al. [14]	24	Prospective	GPA	ANCA positive	OR = 18.00 (95% CI, 4.00–87.00; P < 0.010)
Hogan et al. [15]	49	Retrospective	GPA (17.0) MPA (58.0) RLV (25.0)	PR3-ANCA positive Disease of the lung Upper respiratory tract	HR = 1.87 (95% CI, 1.11–3.14; P = 0.022) HR = 1.71 (95% CI, 1.04–2.81; P = 0.034) HR = 1.73 (95% CI, 1.04–2.88; P = 0.030)
Pavone et al. [16]	3.8	Retrospective	GPA (63.9) MPA (2.7) EGPA (44.4)	Nasal Staphylococcus aureus in CSS Gastrointestinal involvement Kidney disease	HR = 4.45 (95% CI, 0.01–1.15; P = 0.066) HR = 2.83 (95% CI, 1.28–8.33; P = 0.013) HR = 0.39 (95% CI, 0.18–0.87; P = 0.026)
Pouchot et al. [17]	50	Prospective	GPA	Specific cardiac involvement C-ANCA positive Higher age Severe renal insufficiency	HR = 2.90 (95% CI, 1.30–6.50; P = 0.008) HR = 2.10 (95% CI, 1.10–4.30; P = 0.040) HR = 1.40 (95% CI, 1.10–1.70; P = 0.010) HR = 0.40 (95% CI, 0.20–0.80; P = 0.010)
Comarmond et al. [18]	66.8 ± 62.5	Retrospective	EGPA	Lower eosinophil count at diagnosis	HR = 0.99 (95% CI, 0.99–0.99; P < 0.01)
Thai et al. [19]	70	Retrospective	GPA	PR3-ANCA negative Lung involvement	HR = 0.60 (95% CI, 0.39–0.92; P = 0.020) HR = 1.77 (95% CI, 1.09–2.87; P = 0.021)
Li et al. [20]	26	Retrospective	MPA (67.9) GPA (25.5)	A lower serum creatinine level	HR = 0.93 (95% CI, 0.87–0.92; P = 0.009)
Cao et al. [21]	36	Retrospective	GPA (25.5) MPA (74.5)	PR3-ANCA positive Lung involvement	HR = 1.31 (95% CI, 1.01–5.35; P = 0.001) HR = 1.87 (95% CI, 1.12–4.35; P = 0.014)
Kemna et al. [22]	39	Prospective	GPA (76.1) MPA (13.7) EGPA (10.2)	ANCA titers rise ANCA titers rise in RLV	HR = 5.84 (95% CI, 3.44–9.92; P < 0.001) HR = 11.09 (95% CI, 5.01–24.05; P < 0.001)
Yamaguchi et al. [23]	41	Retrospective	MPA (78.6) GPA (6.3) RLV (19.1)	Lung involvement Increased ANCA levels	HR = 2.29 (95% CI, 1.13–4.65; P = 0.022) HR = 17.4 (95% CI, 8.42–36.0; P < 0.001)
Puéchal et al. [24]	141	Prospective	GPA (77.0) MPA (23.0)	PR3-ANCA positive	HR = 1.99 (95% CI, 1.29–3.08; P < 0.002)
Córdova-Sánchez et al. [25]	40.5	Retrospective	MPA (54.8) GPA (32.3) RLV (12.9)	PR3-ANCA positive	HR = 2.93 (95% CI, 1.20–7.17; P = 0.019)
Karras et al. [26]	48	Prospective	GPA (47.0) MPA (53.0)	ANCA positivity at randomization	OR = 2.57 (95% CI, 1.16–5.68; P = 0.017)
Morgan et al. [27]	48	Prospective	ANCA-positive ANCA negative	ANCA negative PR3-ANCA positive	HR = 0.63 (95% CI, 0.42–0.95; P = 0.026) HR = 1.99 (95% CI, 1.30–3.04; P = 0.002)
Tsurikisawa et al. [13]	98 ± 69	Prospective	EGPA	Myocardial involvement Gastrointestinal tract involvement	OR = 5.805 (95% CI, 1.901–17.729; P < 0.010) OR = 3.961 (95% CI, 1.251–12.536; P < 0.050)
Terrier et al. [28]	60	Prospective	AAV	PR3-ANCA positive	HR = 2.04 (95% CI, 1.06–3.91; P = 0.032)
Marco et al. [29]	43.2	Retrospective	GPA(17.1) MPA (82.9)	PR3-ANCA positive Gastrointestinal tract involvement	HR = 2.09 (95% CI, 1.05–4.16; P = 0.037) HR = 13.12 (95% CI, 1.63–105.84; P = 0.016)
Saku et al. [30]	56	Retrospective	EGPA	High IgE levels at onset AZA maintenance therapy High eosinophil counts at onset	HR = 1.025 (95% CI, 1.011–1.039; P < 0.001) HR = 0.261 (95% CI, 0.079–0.870; P = 0.029) HR = 0.946 (95% CI, 0.907–0.987; P = 0.011)
Hong et al. [31]	21	Retrospective	AAGN	PR3-ANCA positive	HR = 3.57 (95% CI, 1.34–9.49; P = 0.011)
Huang et al. [32]	21	Retrospective	GPA (3.8) MPA(91.85) RLV (4.3)	Lung involvement Cardiovascular involvement Serum globulin	HR = 4.60 (95% CI, 1.27–16.60; P = 0.020) HR = 3.69 (95% CI, 1.24–11.0; P = 0.019) HR = 0.88 (95% CI, 0.81–0.95; P = 0.002)
Aljuhani et al. [33]	64	Retrospective	MPO-ANCA positive (55.0) PR3-ANCA positive (45.0)	Increase in ANCA titers Reappearance of ANCA	HR = 8.10 (95% CI, 1.6–40.0; P = 0.009) HR = 17.87 (95% CI, 3.1–103.0; P = 0.001)
Bantis et al. [34]	24	Prospective	MPO-ANCA positive (55.0) PR3-ANCA positive (45.0) ANCA negative	PR3-ANCA positive at time of diagnosis	38.9% vs 11.8% vs 10.3% (P = 0.006)
Duran et al. [35]	47	Retrospective	EGPA	Elevated Ig E level	OR = 6.50 (95% CI, 1.09–38.63; P = 0.040)
Jiang et al. [36]	22.8	Retrospective	MPO-ANCA positive PR3-ANC positive ANCA negative	ANCA persistent positive ANCA positive conversion Infections	HR = 3.35 (95% CI, 1.36~9.59; P = 0.010) HR = 4.76 (95% CI, 1.78~11.35; P < 0.001) HR = 4.68 (95% CI, 2.51~24.17; P < 0.001)
Casal Moura et al. [37]	54	Retrospective	MPO-ANCA-associated vasculitis with GN	MPO-ANCA reappearance ENT involvement	HR = 2.82 (95% CI, 1.11–3.29; P = 0.020) HR = 2.24 (95% CI, 1.28–29.01; P = 0.020)

Abbreviations: AAV: ANCA-associated vasculitis; MPA: microscopic polyangiitis; GPA: granulomatosis with polyangiitis MPO: myeloperoxidase; RLV: renal-limited vasculitis; EGPA: eosinophilic granulomatosis with polyangiitis; AAGN: ANCA-associated glomerulonephritis; ENT: ear-nose-throat; HR: hazard ratio; OR: odds ratio; CI: confidence interval.

ANCA levels

ANCA measurements were established for the diagnosis of disease in AAV. However, their contribution and role in managing AAV follow-up remains controversial. The published data on serial ANCA measurements were heterogeneous which may affect the interpretation of results, such as types of ANCA testing assays used, time intervals, and definitions of rises in ANCA titer. The major limitation was that different studies had an inconsistency in the time intervals of ANCA measurements. A meta-analysis suggested that the serial ANCA measurements and their detection techniques had wide heterogeneity in AVV, but it was limited to predicting relapse and guiding therapeutic decisions for patients with AAV during the remission period [47]. These methodological differences between the studies may be contributed to the different results, and make it difficult to assess the relapse value of serial ANCA measurement in the follow-up of patients with AAV.

Whether persistently positive or increasing ANCA levels are predictive of relapse is also controversial. A retrospective study evaluated the relationship between ANCA subtypes, titers, and disease relapse in AAV patients over 30 years, a total of 202 patients (111 MPO-ANCA-positive and 91 PR3-ANCA-positive) were included, and relapse was higher in PR3-ANCA-positive than in MPO-ANCA-positive patients, especially in double-positive subtypes. Of those, an increased ANCA titer at the time of relapse or the last follow-up was the strongest predictor (hazard ratio, 8.1; 95% CI, 1.6–40; P = 0.009); among patients who had a serological remission, an increase in ANCA at follow-up (reappearance of ANCA) was associated with a higher risk of relapse (hazard ratio, 17.87; 95% CI, 3.1–103; P = 0.001) [33]. In a study by Oristrell et al. [48], 99 patients were included if they had high titers of anti-MPO or anti-PR3 antibodies at least once, and the patients’ ANCA patterns over time were classified as monophasic, remitting, recurrent, or persistent. They found that recurrent or persistent ANCA patterns were correlated with a higher risk of clinical relapse (hazard ratio, 3.7; 95% CI, 1.29–8.79; P = 0.01 and hazard ratio, 4.31; 95% CI, 1.78–10.42; P = 0.001, respectively). In the study by Kemna et al. [22], which included a total of 166 patients, 104 of whom had renal involvement, ANCA elevations during follow-up were associated with relapses in patients with renal involvement (hazard ratio, 11.09; 95% CI, 5.01–24.55; P < 0.001) and patients without renal involvement (hazard ratio, 2.97; 95% CI, 1.30–5.98; P = 0.01).

In addition, a retrospective study of 147 patients receiving rituximab (RTX) as maintenance, found that persistent ANCA positivity became a strong predictor of relapse 12 months after the last RTX (hazard ratio, 2.73; 95% CI, 1.56–4.80; P < 0.001) [40]. A single-center cohort study including 110 cases of ANCA-positive patients investigated how the B-cell response was associated with relapses in AAV patients treated with RTX as remission induction (RI), and found that persistent PR3-ANCA positivity and PR3-ANCA recurrence were associated with more relapses compared to PR3-ANCA negativity (37% vs 3%, P = 0.002 and 50% vs 3%, P = 0.002); patients with B-cell repopulation had significantly more relapses as compared to patients without B-cell repopulation after RTX treatment (41% vs 15%, P = 0.03), ANCA and B-cell status could predict the majority of relapses [49]. Monitoring PR3-ANCA and B cells could guide therapeutic decisions to prevent relapses in AAV patients treated with RTX. The dynamic change in ANCA titer is a predictor of relapse, it is necessary to have serial ANCA measurements, but whether there is an absolute value that provides a reference is inconclusive. There was evidence that PR3-ANCA levels >10 U/ml at 18 months (relative risk, 2.7; 95% CI, 1.1–4.3; P < 0.001) and 24 months (relative risk, 4.6; 95% CI, 1.2–6.3; P < 0.001) were predictive of relapse within 5 years [42]. Future studies are needed to assess whether this could provide a reference value.

Several studies have shown that immunological characteristics of MPO-ANCA, including IgG subclass, epitope specificity, affinity, and titer, contribute to the development of AAV [50, 51]. A study of propylthiouracil (PTU)-induced MPO-ANCA-positive AAV found that linear epitopes of MPO molecules may be closely associated with PTU-induced AAV, particularly P and H4 fragments, which are distinct from primary AAV patients, and also found that PTU withdrawal resulted in clinical remission and a significant decrease in the avidity of PTU-induced MPO-ANCA [52]. Lin et al. [53] investigated the changes in MPO-ANCA avidity in patients with primary AAV at the stages of initial onset, remission, and relapse. During remission, the avidity of MPO-ANCA did not decrease significantly, suggesting that chronic repeated antigen stimulation had not been eliminated, which may be the reason for relapses. The epitopes of MPO-ANCA are remarkably diverse, including both conformational and linear epitopes. MPO-ANCA has been shown to recognize mainly conformational epitopes on MPO molecules, and different epitopes of MPO-ANCA may lead to different disease phenotypes (GPA or MPA) [54]. Linear epitopes of MPO may also contribute to the pathogenesis of MPO-ANCA. Gou et al. [55] analyzed the relationship between the epitope pattern of MPO-ANCA and clinicopathological features of AVV patients and found that the sera of five individuals from six relapsed patients recognized the linear fragment. The linear fragment at relapse was similar to that at baseline, suggesting a role for immune memory in triggering relapse. These observations suggest that differences in the immunological characteristics of MPO-ANCA may contribute to clinical and pathological heterogeneity (Fig. 1). Wojcik et al. [57] showed that a decrease in sialylation was associated with relapse in GPA patients with PR3-ANCA-positive and it could be detected before relapse occurs. The reduced sialylation indicated an inflammatory environment, it has been reported that sialylation on IgG had anti-inflammatory potential through a shift in IgG conformation [58, 59].

Figure 1.

The role of epitopes in relapse of MPO-ANCA-associated vasculitis. (A) The epitopes of MPO molecules include conformational epitopes and linear epitopes. MPO-ANCA has been shown to recognize mainly conformational epitopes on MPO molecules. (B) Linear epitopes may be closely associated with drug-induced MPO-ANCA-associated vasculitis, such as PTU. (C) MPO-ANCA is bound to conformational epitopes or linear epitopes on MPO molecules. (D) MPO-ANCA activates neutrophils. (E) Neutrophil activation triggers the production of reactive oxygen species and the release of the proteolytic enzymes. (F) Different epitopes of MPO-ANCA lead to different disease phenotypes. The differences epitopes of MPO-ANCA may contribute to the different disease phenotypes (GPA and MPA) and clinical and pathological heterogeneity. GPA and MPA had many differences in pathophysiological characteristics and clinical manifestations. Abbreviations: MPO: myeloperoxidase; :ENT: ear-nose-throat; PTU: propylthiouracil; GPA: granulomatosis with polyangiitis; MPA: microscopic polyangiitis. ROS: oxygen species. Adapted by permission from BMJ Publishing Group Limited [56].

The British Society for Rheumatology guidelines do not recommend making treatment decisions based on changes in the ANCA titer alone [60]. An increase or persistent positive ANCA during complete remission is associated with an increased risk of relapse, particularly in patients with renal involvement. Regular ANCA measurements during remission remain a necessary clinical practice, and increasing ANCA levels may require more careful follow-up, but do not guide treatment decisions.

Organ involvement

The specific target organ involvement was associated with relapses (Table 1). Hogan et al. [15] found that disease of the lung (hazard ratio, 1.71; 95% CI, 1.04–2.81; P = 0.034) or upper respiratory tract (hazard ratio, 1.73; 95% CI, 1.04–2.88; P = 0.030) was independently associated with relapse. The precise mechanism of the lung and upper respiratory tract involvement related to relapse in AAV was unclear; one of the biologically plausible hypotheses is infection. The presence of the upper respiratory tract and lung involvement from AAV may foster an environment for infection, especially S. aureus infection. Several studies have suggested that renal function was associated with relapses. In a single-center cohort of Chinese, lower serum creatinine levels were independently associated with an increased risk of relapse [20]. A possible mechanism for elevated serum creatinine and reduced risk of relapse may be a state of immunosuppression caused by impaired immune function, such as some uremic toxins [61]. Renal involvement is common in MPO-ANCA-positive patients and is associated with more severe renal damage, which may account for the lower risk of relapse than in PR3-ANCA-positive patients. However, it has been observed in clinical work that relapse can occur even when patients are on maintenance dialysis, and future studies are needed to assess the effect of renal function on relapse.

There are few studies on renal relapse. Hematuria is not only an indicator of AAV activity with renal involvement but also one of the most important indicators of AAV relapse. A study by Rhee et al. [62] included a total of 149 patients with AAV (42% with persistent hematuria and 43% with persistent proteinuria), and persistent hematuria after 6 months was associated with a significantly higher risk of renal recurrence (adjusted hazard ratio, 3.99; 95% CI, 1.20–13.25; P = 0.02), and the manner in which the greater the cumulative duration of hematuria and the higher the risk of renal recurrence. This study emphasizes the importance of regular monitoring of urinalysis in patients with AAV even in the absence of renal involvement. A study of ANCA-associated glomerulonephritis (AAGN) evaluated the predictive value of histopathological classification at diagnosis for renal relapse. The risk of renal relapse was four times higher in patients without interstitial infiltration than in those with mild infiltration (Fine and Gray’s hazard ratio: 0.09; 95% CI, 0.02–0.39; P = 0.001). Cox model analysis showed that the focal class had a 10.8-fold lower hazard rate of renal relapse than the sclerotic class, and the crescentic class had a 4.8-fold lower hazard rate of renal relapse than the sclerotic class (cause-specific hazard ratio, 0.10; 95% CI, 0.02–0.60; P = 0.01 and cause-specific hazard ratio, 0.21; CI, 0.07 to 0.62; P = 0.004, respectively) [63]. This study suggests that the baseline histopathological class of AAGN is a risk factor for renal relapse.

As reported in previous studies, many non-lung or nonrenal disease factors were associated with the risk of relapse including cardiac and gastrointestinal involvement [29, 64]. Huang et al. [32] found that pulmonary and cardiovascular involvement were independently associated with a higher risk of relapse (hazard ratio, 4.595; 95% CI, 1.272–16.599; P = 0.020 and hazard ratio, 3.689; 95% CI, 1.237–11; P = 0.019, respectively). The results of these studies suggest that AAV relapse may be related to the specific target organs rather than the overall severity or aggressiveness of AAV.

Infection

Reactivation of AAV requires a second hit, and infection is an important contributing factor. In vitro studies have shown that isolated ANCA may not be the only factor in disease reactivation and that a second attack may be required. The importance of the second hit was proposed in a mouse model of anti-MPO IgG-induced glomerulonephritis using bacterial lipopolysaccharide (LPS) as a proinflammatory stimulus. Systemic administration of LPS increased anti-MPO IgG-induced kidney injury in a dose-dependent manner, suggesting that ANCA and proinflammatory stimuli act synergistically to induce vascular inflammatory disease [65]. Possible candidates for proinflammatory stimuli include microorganisms, environmental factors, and/or other autoantibodies [66, 67]. A study by McGregor et al. [68] showed that the cumulative incidence of infection after 2 years of follow-up was 58%. Of those, lung and upper respiratory tract infections were the most common, and 41% of the pathogens were S. aureus. During follow-up, the infection events were associated with relapse of AAV in the first 24 months (P < 0.005).

CpG motifs in bacterial DNA have been shown to induce ANCA production in vitro by B lymphocytes from patients with active AVV. A study by Tadema et al. [69] found that ANCA-producing B lymphocytes are also present in the peripheral blood of AAV patients during remission and that CpG-ODN can stimulate ANCA production in vitro, especially in PR3-ANCA-positive patients, contributing to the development of AAV relapses. This is consistent with an increased risk of relapse in patients with PR3-ANCA-positive AAV. Current hypothesized mechanisms for infection-induced ANCA production include autoantigen complementarity, epigenetic silencing, molecular mimicry, and the formation of neutrophil extracellular traps [70]. GPA often involves the nasal cavity, which increases susceptibility to infection. One study reported that 63% of GPA patients were chronic nasal carriers of S. aureus [71]. Glasner et al. [72] found that several genetic loci in S. aureus are associated with either PR3-ANCA or MPO-ANCA-positive AAV and that specific virulence genes of S. aureus isolated from AAV patients contribute to disease relapse. Conversely, Zycinska et al. [73] showed that the use of prophylactic antimicrobial therapy with co-trimoxazole reduced the incidence of relapses in patients with GPA, 16 patients were assigned to co-trimoxazole and 15 to placebo, during 18 months of follow-up, 75% of patients in the co-trimoxazole group remained in remission compared with 55% of patients in the placebo group (hazard ratio, 0.4; 95% CI, 0.12–0.69; P = 0.003). However, in a prospective multicenter review based on two randomized controlled trials from the European Vasculitis Study Group, prophylactic use of co-trimoxazole was associated with a reduction in chronic S. aureus carriage but not with a reduced risk of relapse [74]. Therefore, S. aureus carriers may only play a secondary pathogenic role in relapse, and further studies on the pathogenesis are needed. Mechanistically, the role of S. aureus carriage in a relapse of PR3-ANCA-associated vasculitis can be explained in different ways (Fig. 2). In patients with GPA who are chronic S. aureus carriers, prophylactic antimicrobial therapy may not only reduce infectious complications but may also reduce the risk of relapse.

Figure 2.

The role of S. aureus in relapse of PR3-ANCA-associated vasculitis. (A) PR3-ANCA-associated vasculitis often involves the lung and upper respiratory tract, which increases the susceptibility to infection. (B) Staphylococcus aureus infections are the most common. (C) Staphylococcus aureus infections cause upregulation of endothelial adhesion molecules (E-selectins, ICAM-1, and VCAM-1), causing increased neutrophil–vessel wall adherence and transmigration. (D) Staphylococcus aureus infection primes the neutrophil by activating monocytes to produce proinflammatory cytokines such as IL-1 and TNF-α. Subsequently, neutrophil priming upregulates adhesion molecules and expression of PR3 on the cell surface of PMN, and PMN adheres to the vessel wall. (E) Phagocytes engulf staphylococci resulting in PR3 release. (F) Bacterial DNA and bacterial LPS are involved in ACNA production and granuloma formation. (G, I) Bacterial DNA activates monocytes and PR3-specific B cells to enhance ANCA production by bridging SAg-reactive helper T cells in the presence of PR3, ANCA activates neutrophils that adhere to endothelial cells. Neutrophil activation triggers the production of reactive ROS and the release of the proteolytic enzymes to damage the vessel wall. (H, J) SAg can bridge auto-specific T cells to other APC cells leading to autoreactive T-cell activation, which can lead to granuloma formation. (K) Endothelial detachment and lysis lead to vasculitis. Abbreviations: ICAM-1: intercellular adhesion molecule-1; VCAM-1: vascular cell adhesion molecule-1; IL-1: interleukin 1; TNF-α: tumor necrosis factor α; PMN: polymorphonuclear neutrophil; LPS: lipopolysaccharide; ROS: oxygen species; SAg: Staphylococcal superantigens; APC: antigen-presenting.

Treatment options

There are some differences in the efficacy of the different treatments, such as choice of drug, drug use, and discontinuation time, which lead to different risks of relapse (Table 2).

Table 2.

The efficacy of different treatment in reducing relapse rates in ANCA-associated vasculitis trials

Author information	Follow-up time, months	Study design	Type of vasculitis (%)	Drug options	Result	Risk of relapse
Aasarød et al. [75]	41.5	Retrospective	GPA	Intravenous CTX vs. oral CTX	Greater relapse with i.v. CTX	RR = 2.9 (95% CI, 1.4–5.8; P < 0.001)
Koldingsnes et al. [4]	42.5	Retrospective	GPA	Pred 0.5–1 mg/kg and oral CTX 2 mg/kg body weight/i.v. Pred 0.5–1 mg/kg and i.v. CTX 15 mg/kg body weight every second week	Greater relapse with < 10 g CTX during the first 6 months Pred > 20 mg/day for < 2.75 months	RR = 2.83 (95% CI, 1.33–6.02; P < 0.007) RR = 2.41 (95% CI, 1.12–5.21; P < 0.030)
Jayne et al. [76]	18	Prospective	GPA (61.3) MPA (38.7)	CTX vs. AZA	Same relapse	13.7% vs.15.5% (P = 0.650)
Slot et al. [77]	12	Retrospective	AAV	CTX vs. AZA PR3-ANCA vs MPO-ANCA	Greater relapse with PR3-ANCA-positive treated with AZA	RR 2.6 (95% CI, 1.1–8.0; P = 0.040)
De Groot et al. [78]	18	Prospective	GPA (93.7) MPA (6.3)	CTX vs. MTX	Greater relapse with MTX	46.5% vs 69.5% (P = 0.023)
Metzler et al. [79]	21	Prospective	GPA	MTX vs. LEF	Greater relapse with MTX	25% vs 3.8% (P = 0.037)
Pagnoux et al. [64]	29.2 ± 13.3	Prospective	GPA (76.2) MPA (23.8)	AZA vs. MTX	Same relapse	HR = 0.92 (95% CI, 0.52–1.65; P = 0.780)
de Groot et al. [80]	18	Prospective	GPA (37.6) MPA (47.6) RLV (17.8)	Intravenous CTX vs. oral CTX	Same relapse	HR = 2.01 (95% CI, 0.77–5.30; P > 0.050)
Zycinska et al. [73]	18	Prospective	GPA	Co-trimoxazole vs. placebo group	Greater relapse with the placebo group	HR = 0.40 (95% CI, 0.12–0.69; P = 0.003)
Jones et al. [81]	12	Prospective	GPA (50.0) MPA (36.4) RLV (13.3)	RTX/CTX vs. CTX/AZA	Same relapse	42% vs. 36% (P = 0.700)
Hiemstra et al. [82]	39	Prospective	GPA (68.7) MPA (31.3)	MMF vs. AZA	Greater relapse with MMF	HR = 1.80 (95% CI, 1.10–2.93; P = 0.020)
Harper et al. [38]	53	Retrospective	AAV	Intravenous CTX vs. oral CTX	Greater relapse with i.v. CTX	HR = 0.50 (95% CI, 0.26–0.93; P = 0.029)
Guillevin et al. [83]	28	Prospective	GPA (75.7) MPA (20.0) RLV (4.3)	RTX vs. AZA	Greater relapse with AZA	HR = 6.61 (95% CI, 1.56–7.96; P = 0.002)
Jones et al. [84]	24	Prospective	AAV	RTX vs. CTX/AZA	Same relapse	42% vs.36% (P = 1.000)
Puéchal et al. [24]	141	Prospective	GPA (77.0) MPA (23.0)	AZA vs. MTX	Same relapse	HR = 0.83 (95% CI, 0.55–1.25; P > 0.050)
Thiel et al. [85]	36	Retrospective	EGPA	RTX vs. CTX	Same relapse	HR = 2.01 (95% CI, 0.77–5.30; P＜0.050)
Puéchal et al. [86]	24	Prospective	EGPA(53.7) MPA (26.3) PAN (20.0)	AZA vs. placebo	Same relapse	OR = 1.14 (95% CI, 0.28–4.74; P = 0.860)
Wechsler et al. [87]	13	Prospective	EGPA	Mepolizumab vs. placebo	Greater relapse with placebo	OR = 0.20 (95% CI, 0.09–0.41; P < 0.001)
Karras et al. [26]	48	Prospective	GPA (47.0) MPA (53.0)	AZA/Pred to 48 months vs. AZA/Pred to 24 months	Greater relapse with AZA/Pred to 24 months	OR = 5.96 (95% CI, 2.58–13.77; P < 0.001)
Charles et al. [88]	18	Prospective	GPA (72.2) MPA (28.8)	RTX infusion when CD19 + B lymphocytes or ANCA had reappeared or ANCA titer rose vs. fixed 500 mg RTX infusion	Same relapse	17.3% vs. 9.9% (P = 0.230)
Terrier et al. [28]	60	Prospective	AAV	AZA vs. RTX	Greater relapse with AZA	HR = 2.72 (95% CI, 1.55–4.76; P < 0.001)
La-Crette et al. [89]	56	Retrospective	AAV	Intravenous CTX vs. oral CTX	Same relapse	HR = 1.00 (95% CI, 0.40–2.60; P = 1.000)
Jones et al. [90]	18	Prospective	GPA (65.0) MPA (35.0)	MMF vs. CTX	Greater relapse with MMF	RR = 1.97 (95% CI, 0.96–4.23; P = 0.049)
Jayne et al. [91]	24	Prospective	GPA (79.0) MPA (21.0)	AZA vs. Belimumab	Same relapse	HR = 0.88 (95% CI, 0.29–2.65; P = 0.821)
Smith RM et al. [92]	36	Prospective	AAV	AZA vs. RTX	Greater relapse with AZA	HR = 0.41 (95% CI, 0.27–0.61; P < 0.001)

Abbreviations: AAV: ANCA-associated vasculitis; MPA: microscopic polyangiitis; GPA: granulomatosis with polyangiitis MPO: myeloperoxidase; RLV: renal-limited vasculitis; EGPA: eosinophilic granulomatosis with polyangiitis; AAGN: ANCA-associated glomerulonephritis; PAN: polyarteritis nodosa; ENT: ear-nose-throat; i.v.: intravenous; CTX: cyclophosphamide; Pre: prednisolone; AZA: azathioprine; MTX: methotrexate; LEF: leflunomide; RTX: rituximab; MMF: mycophenolate mofetil; HR: hazard ratio; OR: odds ratio; RR relative risk; CI: confidence interval.

Either RTX or cyclophosphamide (CTX) in combination with high-dose glucocorticoids are the first-line treatments for severe AAV and have similar relapse rates. The Rituximab in ANCA-Associated Vasculitis (RAVE) trial and Rituximab Versus Cyclophosphamide in ANCA-Associated Vasculitis (RITUXVAS) trial showed that RTX or CTX were equally effective in inducing remission and that RTX may be superior in relapsed disease in the trial [84, 93]. In a retrospective study of 60 patients with renal-involved AAV who underwent an ANCA elevation during remission with oral or intravenous CTX or RTX and another treatment option [methotrexate (MTX) or mycophenolate mofetil (MMF)], the risk of relapse was higher in patients who did not receive CTX or RTX during active disease (hazard ratio, 3.48; 95% CI, 1.60–7.59; P = 0.002) [94]. In a randomized controlled trial in which 140 patients were assigned to MMF or pulsed CTX on the same oral glucocorticoid regimen and switched to azathioprine (AZA) after remission, the MMF group had more relapses after remission than the pulsed CTX group (rate ratio, 1.97; 95% CI, 0.96–4.23, P = 0.049) [90]. The Nonrenal Wegener’s Granulomatosis Treated Alternatively with MTX (NORAM) trial demonstrated that MTX was associated with more relapses than the CTX regimen after termination of treatment by 12 months [78]. In patients with relapsing AAV, Smith et al. [92] compared the efficacy of repeat-dose RTX with daily oral AZA for relapse prevention. They found that the repeat dose of RXT was associated with a lower risk of relapse. RTX is an anti-CD20 monoclonal antibody. It can deplete B lymphocytes through a variety of mechanisms. RTX has become a promising drug in the induction and maintenance treatment of AAV [93]. It has been reported that B cells play a role in the pathogenesis of AAV, given its ANCA-producing plasma cells and their function as antigen presentation cells and T-cell costimulation [95]. Compared to conventional immunosuppression, B-cell depletion therapy is more specific. The open-label randomized controlled trial, International Mycophenolate Mofetil Protocol to Reduce Outbreaks of Vasculitis (IMPROVE) trial compared the effectiveness of MMF and AZA on the prevention of relapses in patients with AAV, relapses were more common in the MMF group [82]. These studies showed that a milder treatment therapy, such as MTX and MMF, induced a higher relapse compared with CTX or RTX. It remains unknown what underlies these differences, but it may give us a clue to predict relapse.

Whether different immunosuppressive agents are associated with the different relapses in EGPA remains unclear. Patients with EGPA were not included in randomized clinical trials such as RAVE and RITUXVAS. Thiel et al. [85] investigated the efficacy and safety of RTX and CTX in EGPA as induction therapy, patients had no difference in relapse between the two groups. A systematic literature review investigated the effects of RTX in patients with EGPA, and scheduled RTX maintenance therapy significantly reduced the relapse rate compared to RTX administered on demand [96]. Saku et al. [30] investigated the risk factors for relapse in EGPA, and they found that AZA maintenance therapy was associated with a lower risk of relapse. A randomized controlled trial of nonsevere vasculitis, including EGPA, MPA, and polyarteritis nodosa, showed that AZA was not superior to placebo in preventing relapse [86]. In a multicenter, double-blind, parallel-group, phase 3 trial, a total of 136 participants with relapsing or refractory EGPA received mepolizumab (300 mg, every 4 weeks) or placebo, mepolizumab reduced the relapse compared to the placebo. In the study, only 20% of the participants had a relapse classified as vasculitis and 54% had a relapse categorized as vasculitis in combination with asthma or sinonasal relapse [87]. Mepolizumab is a humanized monoclonal antibody that binds to and inactivates interleukin‐5 (IL‐5), leading to the depletion of eosinophils within the blood and tissues [97]. EGPA is characterized by eosinophil-rich, and the inflammation in EGPA is mainly composed of eosinophils. Eosinophils can lead to tissue and vascular infiltration and inflammation through a variety of mediators. Glucocorticoids can reduce eosinophil, and glucocorticoid induction therapy leads to successful remission in most patients with EGPA [98].

A study suggested that the dose of maintenance therapy was a crucial factor that affected the relapse rates [99]. Recent studies suggest that intravenous pulsed CTX has a higher relapse rate than oral CTX. In this study, 148 patients were recruited to evaluate the long-term outcomes, and oral CTX had a significantly lower risk of relapse than pulsed CTX (hazard ratio, 0.50; 95% CI, 0.26–0.93; P = 0.029) [38]. The difference may be due to the greater cumulative dose of oral CTX compared with intravenous pulsed CTX. The risk for relapse was inversely associated with the cumulative CTX dose [100]. de Groot et al. [80] compared intravenous pulse CTX with daily oral CTX as induction remission, the median follow-up was 18 months, but there was no difference in relapse rates between the two groups. However, this study aimed to compare the efficacy of two different induction regimens in AAV and the patients with relapse were few, so the conclusion was not sufficient. Khoudour et al. [101] showed that low RTX plasma concentration (CM3 < 4 μg/ml) at 3 months was an independent risk factor for severe relapse. A study showed that scheduled RTX therapy significantly reduced relapses compared with RTX on demand, but discontinuation of RTX after 2 years of maintenance increased the risk of relapse, especially after B-cell reconstitution with rising ANCA levels [102]. However, in the rituximab versus azathioprine in ANCA-associated vasculitis 2 (MAINRITSAN2) trial, no significant differences were found between individually tailored and fixed schedule RTX regimens [88]. RTX can be used for relapsed and refractory AAV, and it is recommended that all patients should receive induction therapy for at least 2 years, with routine monitoring of ANCA titers and lymphocyte counts [49]. Treatment duration is critical, as early discontinuation increases the risk of relapse. For example, in a prospective randomized trial, 117 patients were in stable remission after cyclophosphamide/prednisolone-based induction followed by azathioprine/prednisolone maintenance therapy. They were randomized (1:1) to continue azathioprine/prednisolone for 48 months (continuation group) or to discontinue azathioprine/prednisolone after 24 months (withdrawal group), withdrawal group was associated with a significantly higher rate of relapse than in the continuation group (odds ratio, 5.96; 95% CI, 2.58–13.77; P < 0.001) [26]. It is unclear which maintenance therapy is better at preventing relapse, but these different induction regimens may give us a clue as to what might trigger a relapse.

Clinimetrics

There are a number of different evaluation scales to assess the activity and irreversible chronic damage of AAV, such as Birmingham vasculitis activity score (BVAS), vasculitis damage index (VDI), disease extent index(DEI), and five-factor score (FFS). The BVAS and DEI have been used to assess the disease activity. The BVAS, which was based on organ involvement and disease activity in each organ, was the most widely used. A study has shown that individuals who entered the trial with major BVAS items at diagnosis were less likely to experience a disease relapse during the trial, it was consistent with previous observations of lower relapse risk with worse renal vasculitis [103]. BVAS consisted of 59 items grouped into nine organ systems, and the score of renal systems was the highest. VDI was used to assess the reversible chronic organ damage of AAV, and it was associated with therapy resistance (no complete remission) [4, 104]. In a retrospective study in Japan by Kitagawa K et al, the risk factors for relapse were determined in a cohort of 52 patients diagnosed with MPA. The result showed that the value of VDI was high for the relapse group and VDI value was associated with relapse [105]. Data from 535 patients from 4 European Vasculitis Study Group trials after a long-term follow-up also found that high levels of VDI at diagnosis were independently associated with an increasing number of relapses [106]. FFS was a measure of prognosis in AAV, the current criteria included gastrointestinal involvement, cardiac involvement, renal insufficiency, age > 65 years, and lack of evidence of involvement of the ear, nose, and throat sites. Oh et al. [107] showed that BVAS ≥ 13.5 and FFS ≥ 1 at the time of diagnosis could predict relapse of MPA. One study investigated the risk factors for relapsing EGPA in Japanese patients. It was found that FFS ≥ 1 at diagnosis was an independent predictor of relapse in EGPA [108]. This was consistent with a higher relapse rate of gastrointestinal involvement and cardiac involvement. However, in another study, FFS was not associated with relapse in GPA [109]. This may be related to the criteria of FFS including lack of involvement of the ear, nose, and throat sites, which was the most common organ involvement in GPA.

Others

Several studies have found that patients with a history of relapses are more likely to relapse [40, 110]. One study found that different seasons were associated with subsequent relapses, with ANCA increase followed by a relapse occurring more frequently in autumn than ANCA increase occurring in other seasons, a possible reason for this is that lower vitamin D levels lead to a proinflammatory tendency of the immune system, and patients may benefit from vitamin D supplementation during autumn and winter [94]. One study found that elevated serum S100A8/A9 levels were associated with a higher risk of relapse between baseline and month 2 or 6 in PR3-ANCA patients treated with RTX [111]. The Prognostic Nutritional index (PNI) reflects immune-related nutritional status and is calculated from serum albumin and peripheral blood lymphocyte count. A study performed by Ahn et al. [112] found that PNI ≤ 36.75 was a predictor of disease relapse during the follow-up period. The Delta Neutrophil Index (DNI) represents the number of immature granulocytes associated with neutrophil depletion, and a DNI ≥ 0.65% has been correlated with relapse in patients with GPA and MPA [113]. A study in rat models suggested that urinary sCD163 may be an early predictor of renal relapse. Villacorta et al. [114] found that there was a consistent increase in urinary soluble CD163 levels during relapse compared to those in the remission period. Saku et al. [30] investigated the risk factors for relapse in EGPA, and they found that high eosinophil counts at onset were associated with a lower risk of relapse, and that high IgE levels at onset were risk factors for relapse. However, there was no significant difference in relapse after asthma exacerbation between the two groups. Asthma exacerbation may be an independent event from vasculitis. Further research is needed to confirm the value of these predictors.

Lymphocytes

The promising biomarker for subsequent relapse is the CD8⁺T cells subset, but its predictive value is controversial. A previously published report found that there was a correlation between CD8⁺ T cells and AAV relapse, which needs to be confirmed in prospective studies [115]. Relapse after RTX cessation was also associated with B-cell reconstitution as mentioned earlier in this review. RTX effectively induced B-cell depletion in patients, but almost all patients showed peripheral blood B-cell reconstitution at 18–24 months. Berti et al. [116] evaluated the changes and associations with relapse of the circulating autoreactive B-cell pool following RTX treatment in PR3-ANCA-positive, and the higher plasmablast frequency within the PR3⁺ B-cell pool was associated with relapse. In the RITAZAREM trial, the risk of relapse was increased when B cells were reconstituted, and the risk of relapse was lower in the absence of B cells and ANCA [38, 117]. A single-center cohort study aimed to investigate the relationship between ANCA B-cell status and relapses in AAV patients receiving RTX as RI, and found that incomplete B-cell depletion or B-cell repopulation had significantly more relapses than B-cell depletion or no B-cell repopulation [49]. Other candidates with potential ability to predict the relapse after RTX are circulating CD19⁺B cells and CD5⁺ B cells [118, 119]. Whether repeated B-cell depletion with RTX can prevent relapse after remission remains unknown. Saito et al. [120] found that increased numbers of CD4⁺ T cells producing IL-25 in the peripheral blood may increase EGPA relapse. A prospective multicenter biomarker study aimed to predict renal relapse by quantifying urinary CD4⁺ T cells, and they concluded that urinary CD4⁺ T-cell counts could be associated with a high risk of renal relapse within 6 months [121]. The published data are limited regarding biomarkers to predict relapse in AAV, so the findings should be interpreted with caution.

Conclusion

AAV is a relapsing disease and it is important to assess the risk of relapse and stratify patients to optimize treatment by minimizing toxic therapy and using more targeted therapy. Although many factors have been associated with relapse in AAV, there is no necessary correlation between each factor and relapse. The risk factors for relapse in AAV are summarized in Fig. 3. Further prospective studies are needed to confirm the real value of these factors in clinical management. In clinical practice, comprehensive assessment and stratification are needed to develop individualized treatment strategies to reduce relapse and mortality in patients with AAV.

Figure 3.

Risk factors for relapse in antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis. Abbreviations: GPA: granulomatosis with polyangiitis; AZA: azathioprine; MTX: methotrexate; PNI: Prognostic Nutritional index; DNI: Delta Neutrophil Index; eGFR: estimated glomerular filtration rate.

Glossary

Abbreviations:

ANCA

antineutrophil cytoplasmic antibody

AAV

antineutrophil cytoplasmic antibody-associated vasculitis

AAGN

ANCA-associated glomerulonephritis

AZA

azathioprine

APC

antigen-presenting

BVAS

Birmingham vasculitis activity score

c-ANCA

cytoplasmic ANCA

CTX

cyclophosphamide

CI

confidence interval

DEI

disease extent index

DNI

delta neutrophil index

EGPA

eosinophilic granulomatosis with polyangiitis

EULAR

European League Against Rheumatism

ELISA

enzyme-linked immunosorbent assay

ENT

ear-nose-throat

eGFR

estimated glomerular filtration rate

FFS

five-factor score

HR

hazard ratio

IIF

indirect immunofluorescence

IL-1

interleukin 1

IMPROVE

International Mycophenolate Mofetil Protocol to Reduce Outbreaks of Vasculitis

IL‐5

inactivates interleukin‐5

LEF

leflunomide; i.v: intravenous

ICAM-1

intercellular adhesion molecule-1

LPS

lipopolysaccharide

MPA

microscopic polyangiitis

MPO-ANCA

myeloperoxidase ANCA

MTX

methotrexate

MMF

mycophenolate mofetil

MAINRITSAN2

rituximab versus azathioprine in ANCA-associated vasculitis 2

NORAM

nonrenal Wegener’s granulomatosis treated alternatively with MTX

OR

odds ratio

PR3-ANCA

proteinase 3 ANCA

p-ANCA

perinuclear ANCA

PTU

propylthiouracil

PNI

Prognostic Nutritional index

PAN

polyarteritis nodosa

Pre

prednisolone

PMN

polymorphonuclear neutrophil

RLV

renal-limited vasculitis

RPF

retroperitoneal fibrosis

RTX

rituximab

RI

remission induction

RITUXVAS

rituximab versus cyclophosphamide in ANCA-associated vasculitis

rave

rituximab in ANCA-associated vasculitis

RR

relative risk

ROS

oxygen species;

SAg

staphylococcal superantigens

TNF-α

tumor necrosis factor α

VDI

vasculitis damage index

VCAM-1

vascular cell adhesion molecule-1

WEGENT

Wegener’s Granulomatosis-Entretien

Conflict of interest:

All authors have declared no conflicts of interest.

Funding

This study was supported by a grant from the National Natural Science Foundation of China (81970631 to W. L.).

Data availability

The manuscript does not contain any original data, as it is a review.

Author contributions

Han Zhou: investigation, data curation, visualization, writing of original draft. Wei Liang: conceptualization, supervision, review and editing. Guohua Ding: supervision, review and editing. Hongtu Hu, Zikang Liu, Fan Chu: review and editing.