Disease Activity, Prognosis, and Damage

Scores

Various scores have been designed to monitor disease activity and assess damage, primarily for therapeutic trials and cohort studies. These scores are easily accessible online (e.g., http://www.canvasc.ca/tools.htm). The need to use the scores and their usefulness in routine practice is debatable. However, the items contributing to the scores can help physicians assess patients in a more systematic manner and not forget some important symptoms and signs. The Birmingham Vasculitis Activity Score (BVAS) and the Vasculitis Activity Index (VAI) include clinical and biological factors to assess the degree of activity of systemic vasculitis, not specifcally GPA [112]. The BVAS/GPA (WG) was developed specifcally for GPA [113]. The Damage Extent Index (DEI) is a rarely used activity score, and the result is also related to prognosis [114].

The Five-Factor Score (FFS) is a prognostic score that was initially designed and validated only for polyarteritis nodosa, microscopic polyangiitis, and eosinophilic granulomatosis with polyangiitis [115]. After a subsequent analysis of the French Vasculitis Study Group database, the FFS was revised and can also be used for GPA [116]. Five factors, assessed at the initial diagnosis of GPA, are associated with a poor prognosis (lower survival): a serum creatinine level ≥150 μmol/L, specifc GPA-related cardiomyopathy, age >65 years, severe and specifc GPA-related gastrointestinal involvement, and the absence of ear, nose, or throat manifestations. The 5-year mortality for an FFS = 0 is <10% and is 60% for an FFS ≥ 2 [116]. This revised FFS is not an activity score to be repeatedly calculated during the course of the disease, must not be used to determine treatment choices (as opposed to the original FFS, which can be used to guide treatment for patients with polyarteritis nodosa, microscopic polyangiitis, or eosinophilic granulomatosis with polyangiitis), and has not been validated for GPA relapse.

Although the BVAS, BVAS/GPA (WG), and DEI can quantify some of the persistent clinical signs and damage, while the disease is no longer active, assessment of diseaseand/or treat- ment-related damage can be more accurate with the Vasculitis Damage Index (VDI) or the Combined Damage Assessment Index (CDAI), which was developed later than the VDI and is more comprehensive but also takes longer to complete [117].

The European League Against Rheumatism (EULAR) recommends use of the BVAS, VDI, and DEI for all vasculitis as well as the Medical Outcomes Survey Short Form 36 (SF-36) to assess quality of life [118]. The Outcome Measures in Rheumatoid Arthritis Clinical Trials (OMERACT) 10 recommends the BVAS (original and/or version 3/2003 and/or BVAS/GPA (WG)) for assessing disease activity, the VDI for evaluating damage (Combined Damage Assessment Index remaining under validation), and the SF-36 [119].

Treatment

Without treatment, GPA was almost always fatal within 6–12 months in historical reports and series [73]. Since the introduction of glucocorticoids in the early 1960s for vasculitis, and the use of glucocorticoids combined with cyclophosphamide since the early 1970s, survival has greatly improved. In the 1992 study by Hoffman et al., more than 87% of patients achieved remission and were alive at 8 years [5]. Patient outcomes have continued to improve, with most of the advances made in optimizing treatment strategies, mainly to limit treatment-­related toxicity. A major advance in the early 2000s was the demonstration that rituximab was equally effective as cyclophosphamide in remission, which increased the therapeutic armamentarium, and, then, it was also proven superior to conventional azathioprine for remission maintenance. New agents have been studied and more are still being investigated, which may further alter how patients with GPA can be optimally treated.

Treatment has two phases: induction therapy, currently with a combination of glucocorticoids and a second immunosuppressive agent and, then, once remission is achieved, maintenance therapy (to maintain remission).

Induction Treatment for Systemic/Severe/ Generalized Forms

Glucocorticoids

Glucocorticoids can rapidly improve the symptoms and signs of active GPA, but if prescribed alone, they will not induce sustained remission [73]. The initial dose of glucocorticoids is 1 mg/kg/day prednisone equivalent, sometimes preceded by 1–3 intravenous (IV) boluses of methylprednisolone (7.5– 15 mg/kg/day) [120]. After the frst couple of weeks of treatment, the dose of glucocorticoids is usually reduced by about 10% every 1–2 weeks to achieve a half dose (0.5 mg/kg/day) at about the second or third month of treatment. The optimal tapering regimen and duration of treatment with glucocorticoids are still unclear and controversial. The recent results of the PEXIVAS trial have suggested that a rapid tapering schedule, as of Week 1, was as effective as a slower, more conventional one and associated with a reduction in infections [121]. In the United States, some centers consider glucocorticoids beyond 6–9 months of little beneft or even with an increased risk of long-term toxicity [122, 123]. Most of the other centers prescribe glucocorticoids for a longer duration but at low doses (5–10 mg/day) for 1 year or more. An analysis of several trials conducted prior to 2010 suggested that low-dose prednisone beyond 6 and 12 months may be associated with a low risk of relapse [124]. Trials are ongoing to determine whether this is true or not.

Cyclophosphamide

The combination of glucocorticoids and cyclophosphamide had been the frst-line standard treatment for severe GPA until around 2010 [125]. Cyclophosphamide induces remission in >80% of patients [126]. It can be administered as intravenous “pulses” (boluses at regular intervals) or continuous (daily) oral tablets. Oral daily cyclophosphamide is prescribed at a dose of 2 mg/kg/day (not exceeding 200 mg/ day). Intravenous boluses of cyclophosphamide are administered every 14 days for 1 month at 0.6 g/m2 or 15 mg/kg (at Days 0, 14, and 28) and then at 0.7 g/m2 or 15 mg/kg every 3 weeks [126]. The dose of each pulse should not exceed 1200 mg. Pulsed oral administration is another option but is rarely used (the total dose of each pulse—15 mg/kg/day—is usually divided and administered over 3 days; i.e., 5 mg/kg/ day for 3 consecutive days every 2 weeks for the frst month and then every 3 weeks). Proper hydration should be ensured during the administration of cyclophosphamide. If cyclophosphamide is administered intravenously at >600 mg, then mesna should also be prescribed to reduce the bladder toxicity with cyclophosphamide [127, 128]. The dose of oral or intravenous cyclophosphamide should be adjusted (i.e., lowered by 25–50%, with a minimal dose of 500 mg per IV pulse) in patients aged >65 years old, with renal impairment, low leukocyte count (white blood cell count <4 G/L), and low hemoglobin level [106].

Several studies have established that the two routes of cyclophosphamide administration (intravenous bolus or continuous oral) are comparable in achieving remission and time to achieve remission. However, the cumulative dose of cyclophosphamide is higher with the continuous oral route and has been associated with an increased frequency of neutropenia [129] and, in some (but not all) studies, with infections [126]. The risk of infertility and/or late complications (i.e., cancers, mainly bladder cancer, and also late lymphomas or skin cancers) is directly associated with the cumulative dose of cyclophosphamide. Therefore, in France and most European countries, the intravenous route is most often used frst and the oral route represents an alternative for patients without remission as escalation therapy. However, the long-­ term follow-up (median 4.3 years) in one study comparing pulsed intravenous and oral continuous routes for induction (followed by azathioprine maintenance) suggested that oral continuous cyclophosphamide (i.e., eventually a higher cumulative dose of about 16 g, as compared with 8 g for patients who achieved remission with the intravenous regimen) is associated with a lower subsequent relapse rate (20% instead of 40%) [130].

Irrespective of the route of administration, the 2003 CYCAZAREM study showed that oral continuous cyclophosphamide could be stopped as soon as the patient achieves clinical remission (defned as the absence of clinical disease activity, confrmed by a BVAS score of 0), thus possibly after

only 3 months in most patients [125]. Thereafter, cyclophosphamide can be switched to a less toxic immunosuppressive agent for maintenance. This format has also been used with intravenous bolus cyclophosphamide, even though not proven directly; thus, the prescription of three additional boluses to consolidate the achieved remission is no longer seen as needed [131]. Once remission is achieved, usually after 6–9 boluses, therapy can also be switched to a less toxic immunosuppressive therapy. By limiting the duration of exposure and cumulative doses of cyclophosphamide, the risk of subsequent malignancies, as reported in earlier studies, is greatly reduced [5].

Rituximab

Rituximab is a chimeric monoclonal anti-CD20 antibody and a cornerstone agent for treating lymphomas, which was frst tested in the early 2000s in GPA for treating refractory disease. It has been subsequently evaluated in two randomized trials (RAVE and RITUXVAS) as an alternative to cyclophosphamide to induce remission and was offcially approved in April 2011 by the US Food and Drug Administration (FDA) for treating severe forms of ANCA-­ positive GPA (and microscopic polyangiitis) in adults, combined with glucocorticoids [41, 43], and, subsequently, in late 2019, in children aged 2 years and older. The treatment of severe ANCA-negative GPA, which is a rarer form, should follow the same therapeutic approach, but access to the drug may be more diffcult in some countries. In both the RAVE and RITUXVAS studies, rituximab was not inferior to cyclophosphamide in inducing remission at 6 months. Tolerance to rituximab was good, but the infection rate was (disappointingly) comparable in the two arms and mainly consisted of community-acquired upper and lower respiratory tract infections. At 18 months, the rates of relapse (around 30%) and adverse events were the same in both arms [132]. Neither the RAVE or the RITUXVAS trial involved maintenance therapy after the induction courses of rituximab.

The doses of rituximab used in these studies were four infusions of 375 mg/m2 at 1-week intervals. Data from various cohorts support that a different protocol of two infusions of 1 g at a 2-week interval, as used for rheumatoid arthritis, has comparable effectiveness and is now used very often [106].

In practice, the choice of cyclophosphamide or rituximab for induction is based on several factors, including several of the disease and patient characteristics, patient’s plans for pregnancy, comorbidities, and the higher cost of rituximab. Rituximab use as a frst-line induction agent, combined with glucocorticoids, has increased over the past decade, after initially being mostly limited to patients with contraindication to cyclophosphamide, those with frequent relapses, and/or those who had already received large cumulative doses of cyclophosphamide (>20 or 30 g, but consensus is lacking on

the threshold dose at which the risk of cancer becomes unacceptable). Being female and of childbearing age is an important consideration (as well as males wishing to be a father), even though the risk of infertility in a 20-year-old patient after receiving a cumulative dose of 6–9 g of cyclophosphamide is probably low. Other factors in uence the therapeutic choice. The response to rituximab appears to differ also depending on the form of the disease, with a poorer and/or slower response for granulomatous than vasculitic manifestations and some better results with rituximab than cyclophosphamide in PR3-ANCA-positive patients, especially in those with relapsing disease [132–134]. Of note, rare cases of progressive multifocal leukoencephalopathy have been reported in patients receiving rituximab for other conditions (lymphoma, lupus), as well as potentially serious allergic reactions, and, rarely, interstitial “immunoallergic” pneumonias.

Other Current Induction Approaches

The combined use of glucocorticoids and either cyclophosphamide or rituximab represents the current standard of care for the treatment of severe GPA. The combined use of rituximab plus cyclophosphamide could be considered in severe or refractory cases, although most experts do not favor such an option because the added effcacy of this combination is likely negligible, whereas the risk of complications, especially infectious, would increase. However, some groups are investigating this combination therapy with study designs aimed at limiting the use of glucocorticoids. Recent preliminary results have shown that this combination allowed for good disease control, with highly limited use (1–2 weeks only) of glucocorticoids or even none, with less infections as a result [135, 136].

Plasma exchange, avacopan, and a few other therapies rarely used at present or still under investigation are further discussed in section on “Other Treatments in GPA”. Avacopan may lead to radical changes in the treatment of GPA, by allowing for an effective and safe steroid-free option, but, it is not yet readily accessible in every country.

Maintenance Therapy for Systemic/Severe/ Generalized Forms

Maintenance therapy follows induction therapy once remission is achieved. The continuation of oral cyclophosphamide in early studies, in patients induced with cyclophosphamide, found a relapse rate as low as 13% at 5 years [126]. Continuing intravenous boluses of cyclophosphamide in gradually longer intervals were not as effective, with a relapse rate of about 60% at 5 years [126, 137]. However, the cumulative toxicity of cyclophosphamide, administered

orally or intravenously, precludes its prolonged use, and, thus, use of this agent must remain limited to induction therapy (3–6 months).

Other immunosuppressive agents, including azathioprine or methotrexate, have frst been found as effective as continuing oral cyclophosphamide in maintaining remission but with less toxicity. More recently, rituximab has been found to be superior to those agents, thereby gradually becoming the most frequently used and recommended agent for maintenance, whether the induction includes rituximab or cyclophosphamide [106].

After therapy induction with cyclophosphamide, maintenance therapy can be initiated as early as remission is achieved, usually after 3–6 months of cyclophosphamide [125]. After therapy induction with rituximab, maintenance therapy can be considered (and started) at around 4–6 months after the last rituximab infusion [138, 139]. Whereas maintenance appears mandatory after cyclophosphamide-based induction therapy, a few authors suggested that patients in remission after rituximab-based induction, especially (or exclusively, for some more prudent authors) those with a new diagnosis of GPA, could be simply monitored and only retreated with rituximab (induction doses) if a clinical relapse occurs, as was done for the RAVE study patients [132]. The latter approach encompasses risks, including that of a major, acute, and life-threatening are, but might be an option for some specifc, highly selected patient subsets. MAINRITSAN, a prospective study of the French Vasculitis Study Group, the results, of which were published in 2014, showed rituximab infusions (500 mg at the time of remission, then at Day 15 and at months 6, 12, and 18) to be superior to azathioprine for maintenance therapy [139]. In this study, patients had newly diagnosed or relapsing GPA (or microscopic polyangiitis) and were treated with IV cyclophosphamide and glucocorticoids for remission induction. The rate of major relapses at 28 months post-enrollment was 5% with rituximab compared to 29% with azathioprine. Based on these results, in October 2018, the FDA approved the use of rituximab for “follow-up” (i.e., maintenance) in GPA (or microscopic polyangiitis). In the subsequent similarly designed international study (RITAZAREM), patients with relapsing GPA (or microscopic polyangiitis), who achieved remission with a rituximab-based regimen this time, were randomized to receive maintenance azathioprine or 1 g rituximab infusions every 4 months for 2 years. Again, rituximab proved superior, with a lower relapse rate (major and minor combined) at 20 months after randomization (13% compared to 38% in the azathioprine group—the preliminary results were presented as Abstract in November 2019) [140].

Mainly because of the high cost of rituximab, approaches other than systematic rituximab reinfusions have been inves-

tigated. The MAINRITSAN 2 study investigated the value of an “on-demand, tailored” maintenance treatment, with repeat 500 mg infusions of rituximab only if a change in the ANCA status from negative to positive, an increase in the ANCA titer by at least twice the previous value, and/or a CD19+/ CD20+ B-cell repopulation occurred (all measured serially every 3 months). Such a strategy was associated with a statistically comparable rate of (major and minor) relapses at 28 months (17.3% vs. 9.9% with the fxed schedule reinfusions of 500 mg every 6 months) [141]. This “tailored” approach requires repeat ANCA testing and CD19+ B-cell monitoring every 3 months and thus can be more cumbersome in practice than reinfusions systematically scheduled every 6 months for 2 years. A post hoc analysis of MAINRITSAN 2 also showed that the second dose of 500 mg of maintenance rituximab at Day 15 was not required, thus rendering the MAINRITSAN regimen simpler (500 mg at Month 6, then every 6 months, for at least three more doses) [142].

When rituximab is not available, contraindicated, or not the preferred agent for some reason, azathioprine can be administered for maintenance at a dose of 2 mg/kg/day orally or methotrexate at the dose of 0.3 mg/kg/week (up to 25 mg/ week), orally or subcutaneously [125, 131, 143]. These two drugs are equally effective and safe, at least at a 2-year follow­ -­up. However, methotrexate should not be used for patients with renal insuffciency or a serum creatinine level ≥2 mg/day/L (>175 μmol/L). With these two agents, the relapse rate (following a cyclophosphamide-based induction) is around 16% at 18 months, 37% at 25 months, 52% at 32 months, and 51–64% at 7 years, with a relapse-free survival rate of only 42% at 5 years [125, 131, 137, 144, 145]. Both azathioprine and methotrexate can increase the risk of opportunistic infections and cause liver toxicity and/or myelosuppression. Methotrexate-induced lung hypersensitive pneumonia is a rare adverse event. Pharmacogenetic and/or genotypic study to measure the activity of thiopurine methyltransferase may be performed before prescribing azathioprine, to identify the rare patients at an increased risk of hematological toxicity [146]. Close laboratory monitoring is anyhow mandatory after the start of these maintenance drugs.

In a single randomized controlled study, terminated prematurely because of the higher-than-expected relapse rate in the control methotrexate arm, the rate of severe relapse with le unomide was only 5% at a 2-year follow-up [147]. Le unomide (at the standard dose of 20 mg/day, possibly increased to 30 or 40 mg/day after 3–6 months) could thus be another option for maintenance therapy. However, side effects were frequent (especially a dose >20 mg/day) and included respiratory infections, arthralgia, high blood pressure, liver toxicity, diarrhea, and, rarely, peripheral neuropathy. The frst open studies of mycophenolate mofetil for

maintenance therapy reported a relapse rate of only 11% [148], but the following studies reported higher rates, up to 43–48% [149]. The results of the European IMPROVE trial, comparing azathioprine and mycophenolate mofetil as maintenance therapy, showed mycophenolate mofetil to be inferior to azathioprine: at 4 years, the relapse rate was 55% and 38%, respectively [150].

The optimal duration of all these maintenance treatments remains unknown, and several prospective studies have been conducted, with some still ongoing, to determine when and for whom it is safe to stop therapies. The results of several completed controlled trials agree that the total duration of treatment (induction + maintenance) must not be <18 months. The results of the European REMAIN trial (4 years of maintenance therapy with azathioprine and low-dose prednisone as compared with discontinuing therapy at 2 years) showed that the continuation of therapy for 4 years was associated with fewer relapses (22% versus 63% at Month 48 post-­enrollment), especially for patients with persistent ANCAs at remission. The MAINRITSAN 3 trial compared rituximab maintenance (500 mg at Day 1, Day 15, and then every 6 months) for 18 months (a total of fve infusions) vs. 46 months (a total of none infusions) and showed a greater major relapse-free survival rate of 100% at 56 months with the longer treatment vs. (an “only slightly lower”) 87% when stopping after the 18-month infusion [151]. As mentioned previously, glucocorticoids should not be unduly prolonged, but maintaining lowdose glucocorticoids of about 5 mg/day over an additional 6–12 months may reduce the risk of relapse [124]. Two studies are ongoing to further address this latter issue (ClinicalTrials. gov Identifers: NCT01933724 and NCT03290456).

Treatment of Localized/Limited/Early

Systemic GPA

In patients with localized GPA, mainly with ear, nose, and throat manifestations, cyclophosphamide should be avoided, with other less toxic treatment strategies used as frst-line therapy. Glucocorticoids alone can improve disease in half of such patients but are unlikely to achieve or sustain remission [96, 143]. Few patients have achieved treatment success with cotrimoxazole (at 2 tablets/day, double-strength) alone or most often combined with glucocorticoids [97, 98, 152]. However, the results with this cotrimoxazole-based treatment have been disappointing in the few prospective studies, some of which had to be stopped earlier than planned because of frequent disease progression and worsening [153]. In the European NORAM trial, methotrexate (0.3 mg/kg/week for 12 months) led to remission as often and as quickly as cyclophosphamide with early-systemic GPA [94]. The remission rate with glucocorticoids and methotrexate was 90% (as