Objectives To study the efficacy of rescue treatment strategies and outcomes in patients with Wegener's granulomatosis (WG) and microscopic polyangiitis (MPA) not achieving remission with first-line induction with corticosteroids (CS) and intravenous cyclophosphamide (CYC).
Methods 159 eligible patients in the Wegener's Granulomatosis-Entretien (WEGENT) trial newly diagnosed with systemic or renal WG or MPA with ≥1 poor prognosis factors were included in this prospective study. Rescue treatment strategies and outcomes in patients with induction-refractory disease were analysed and patient characteristics at diagnosis were compared with those of induction-responders.
Results Most patients (n=126, 79.2%) achieved remission; 1 stopped induction because of allergy and 32 were induction-refractory (24 WG and 8 MPA); 11 died rapidly within a median of 2.5 months, 6 of uncontrolled disease, 1 of an infectious complication and 4 of both. Treatment was discontinued in 1 patient with MPA with end-stage renal disease. Induction was switched to oral CYC in 20 patients, combined with infliximab in 1; 15 (75%) achieved remission or low disease activity state, 3 subsequently died of uncontrolled disease and 2 entered remission using several other agents including biological agents. Alveolar haemorrhage and a creatinine level >200 μmol/l were independently associated with induction-refractory disease. Among patients with induction-refractory disease, massive alveolar haemorrhage was associated with higher mortality.
Conclusion Switching to oral CYC can be an effective rescue treatment for patients with systemic forms of WG or MPA who fail to achieve remission with first-line CS and intravenous CYC. However, a more rapidly effective regimen remains to be identified for most severely affected patients whose outcomes can be rapidly fatal.
Statistics from Altmetric.com
A combination of corticosteroids (CS) and cyclophosphamide (CYC) remains the standard treatment for inducing remission of Wegener's granulomatosis (WG) and microscopic polyangiitis (MPA), but there is a potential risk of side effects, particularly with long-term CYC.1 2 Because intravenous CYC has been shown to be as effective as oral CYC in achieving remission of antineutrophil cytoplasm autoantibody (ANCA)-associated vasculitides and is less toxic,3,–,6 it is now widely considered an alternative to oral CYC for remission induction.
Although 75–90% of the patients achieve remission with oral or intravenous CYC,7,–,10 few data are available on therapeutic strategies for patients with disease refractory to this first-line treatment. To date, only one study on four patients has focused on the efficacy of oral CYC as a rescue treatment for patients refractory to intravenous CYC.11 Moreover, factors associated with induction-refractory vasculitis have mainly been studied in patients with predominant renal disease and little is known about these factors in patients with more systemic forms of the disease.12 13
We report the characteristics, rescue therapies and outcomes of patients with WG and MPA who failed to enter remission with the standard CS and intravenous CYC induction regimen given during the pre-randomisation phase of the prospective multicentre randomised open-label Wegener's Granulomatosis-Entretien (WEGENT) trial.14
Patients and methods
WEGENT trial inclusion criteria
Patients included in this study were those screened for the WEGENT trial (ClinicalTrials.gov number NCT00349674) which was designed to compare methotrexate with azathioprine as maintenance therapy for patients with WG or MPA who had achieved complete remission on the trial induction regimen. This trial was conducted from November 1998 to February 2005 in university and general hospitals in France and Belgium. The study protocol has been described elsewhere14; it was developed by the Investigator Advisory Committee of the French Vasculitis Study Group (FVSG) and received ethics committee and review board approvals. All patients gave informed consent to participate. Eligible patients were >18 years old and had (1) newly diagnosed WG fulfilling American College of Rheumatology classification criteria15 with renal disease, involvement of ≥2 organs or systems or involvement of one organ or system in association with constitutional symptoms (fever >38°C, weight loss >3 kg within the past month, diffuse arthralgias and/or myalgias); or (2) MPA fulfilling the Chapel Hill nomenclature definition16 with ≥1 factor of poor prognosis according to the 5-factor score.17 Eligibility also required histological confirmation of the diagnosis, positive ANCA serology by indirect immunofluorescence and/or ELISA specificity for myeloperoxidase and/or proteinase-3.
The planned induction regimen applied to all patients consisted of CS (daily intravenous methylprednisolone (15 mg/kg) for 3 days followed by oral prednisone (1 mg/kg/day) for 3 weeks, then progressively tapered) and intravenous CYC pulses (3 pulses (0.6 g/m2) administered at 2-week intervals, then every 3 weeks (0.7 g/m2) until remission followed by three additional consolidation pulses).14
Definition of induction-refractory disease
Patients who failed to achieve remission under the trial induction regimen according to the treating physician's impression (no switch to maintenance because of persistent active disease) and those who died during induction were considered to have induction-refractory disease. These patients had not been randomised in the WEGENT trial because of persistent disease activity. Moreover, the files of all patients considered for this study were reviewed to assure that they met the EUVAS/EULAR definitions of refractory disease18—that is, disease activity unchanged or increased after 4 weeks of treatment, a 50% reduction in the disease activity score not observed within 6 weeks after treatment initiation and/or chronic persistent disease after >12 weeks of treatment.
Management of patients with induction-refractory disease
No protocol-defined rescue treatment strategy was planned for patients with induction-refractory disease in the trial and their therapeutic management was left to the discretion of their treating physicians, usually after contacting the trial coordinators (LG, CP, PC or AM).
All patients were evaluated at diagnosis and then followed prospectively. At diagnosis, clinical manifestations and biological parameters were recorded using a standardised form for the trial, as described elsewhere.14 Disease activity was measured at diagnosis using the Birmingham Vasculitis Activity Score (BVAS).19 Characteristics, rescue therapies (drugs, doses and durations) and outcomes of patients with induction-refractory disease were analysed prospectively.
Remission (absence of disease activity (BVAS=0) attributable to active disease, qualified as the need for ongoing stable maintenance immunosuppressive therapy), low active disease state (persistence of minor symptoms such as arthralgias or myalgias that responded to a modest increase in the CS dose and did not require treatment intensification) and relapse (recurrence or new onset of disease attributable to active vasculitis) were defined according to the EUVAS/EULAR definitions of disease activity states.18 For patients who achieved remission or a low disease activity state, the doses of CS and immunosuppressant were recorded.
The results are expressed as numbers and/or percentages for categorical variables and mean±SD or median (IQR) for continuous variables, unless stated otherwise.
To identify parameters associated with induction-refractory disease, the patient characteristics at diagnosis in patients with induction-refractory disease and induction-responders were compared. The demographic characteristics and other variables to be assessed as potential predictors of non-response were determined before the study. For creatinine, three thresholds were tested: 140 µmol/l,17 500 µmol/l20 and the upper quartile of the overall population distribution. Univariate analyses compared qualitative variables using a χ2 test or, when appropriate, the Fisher exact test, and quantitative variables were compared with the Student t test. Variables selected in univariate analysis were entered into a multivariate logistic regression model.
Among patients with induction-refractory disease, survival analyses were conducted to search for independent determinants of death. Survival time was calculated as the time from the date of diagnosis to the date of censoring or the time of death. Patient survival was assessed by life table analysis using the Kaplan–Meier method.21 22 Survival was evaluated as a function of the parameters recorded at diagnosis with Cox proportional hazards models.23 Tests for the proportional hazards assumption were obtained for each covariate and the model as a whole. Because of the small number of patients, the number of variables included in the multivariate model was limited to 3. The results are expressed as HRs with 95% CI.
For all multivariate analyses, variables with p<0.20 in univariate analysis were entered into the model and selected using a stepwise regression. For all statistical analyses, p≤0.05 was considered significant. Statistical analyses were conducted using SAS software Version 9.1 (SAS Institute, Cary, North Carolina, USA).
Among the 159 patients with WG or MPA recruited to the WEGENT trial, 33 (20.8%) failed to achieve remission with the trial induction regimen. One patient with MPA stopped induction because of allergy and was switched to mycophenolate mofetil; he had exclusive renal involvement which progressed to end-stage renal disease. Thirty-two patients (20.1%) did not achieve remission using the trial induction regimen, 24/120 (20%) with WG and 8/38 (21.1%) with MPA. Their median follow-up was 60.5 (3.6–84.9) months. For the 126 randomised patients, the median duration of induction therapy was 6.2 (5.5–7.3) months, corresponding to a median of 10 (9–12) CYC pulses.
Characteristics of patients with induction-refractory disease
The characteristics of the 32 patients with induction-refractory disease are shown in table 1. Thirteen patients had alveolar haemorrhages, including 5 in whom the haemorrhages were considered massive (defined as associated with acute respiratory distress syndrome and/or a haemoglobin level <8 g/dl with a decrease of 2 g/dl) and 13 patients (including 7 with alveolar haemorrhage) had raised creatinine levels >200 μmol/l; thus, 7 patients (21.9%) had both manifestations. Thirty were ANCA-positive and 17 had biopsy findings supporting their vasculitis diagnoses. Twenty-seven (84.4%) of the 32 patients satisfied the EUVAS/EULAR definitions for refractory disease: 11 had unchanged or increased disease activity after 4 weeks of treatment, 8 had not obtained a 50% reduction in the disease activity score within 6 weeks after treatment initiation and 8 had chronic persistent disease after >12 weeks of treatment. Their refractory systemic manifestations were alveolar haemorrhage in 9 patients, glomerulonephritis in 10, excavated lung nodules in 5, ocular involvement in 2, new onset of peripheral neuropathy in 2, central nervous system involvement in 5 and severe gastrointestinal manifestations in 2. All of the 5 patients who did not meet the EUVAS/EULAR definitions of refractory disease initially improved with intravenous CYC but rapidly worsened before they could be randomised and switched to maintenance therapy.
Outcomes and rescue treatment of the 32 patients with induction-refractory disease
Eleven patients (34.4%) died during first-line induction therapy (figure 1) after a median of 2.5 (1.3–3.3) months; two patients died despite intravenous immunoglobulins and/or plasma exchanges. Six deaths were attributed to uncontrolled vasculitis, one to an infectious complication and four to both causes. Interestingly, all five patients with massive alveolar haemorrhage died.
Only one patient had kidney-limited disease, with end-stage kidney disease and no other systemic involvement. The treating physician decided to stop intravenous CYC after eight pulses and the patient remained on low-dose prednisone (10 mg/day). He underwent transplantation 1 year later and 6.5 years later he has suffered no relapse with an antirejection regimen of ciclosporin and mycophenolate mofetil.
The 20 remaining patients were switched to oral CYC (2 mg/kg/day), associated with infliximab in one, after a median of 7 (4–9) pulses. Fifteen (75%) of them achieved remission (10 patients) or low disease activity state (5 patients) according to the EUVAS/EULAR definitions after a median of 15.5 (11.3–19.5) months after diagnosis. The low disease activity states corresponded to persistent lung nodules in three patients, tracheal stenosis in one and ‘grumbling’ ear, nose and throat manifestations in one. Oral CYC was continued for a median of 12.0 (6.0–14.0) months. One patient was rapidly switched to azathioprine because of CYC-induced neutropenia. CS doses had been tapered to ≤10 mg/day in six patients and to ≤7.5 mg/day in two of them, and discontinued in eight. All but four were able to discontinue their immunosuppressive maintenance therapy.
Of the five patients who did not achieve remission with oral CYC, two entered remission after several rescue therapies and three died. One death was attributed to cardiac insufficiency due to vasculitis, one to refractory pulmonary, cerebral and gastrointestinal vasculitis and one to severe septicaemia with refractory disease.
Overall, the primary mortality of patients with induction-refractory disease was 43.8% (14/32) for nine patients with WG and six with MPA (p=0.42). The 18 survivors were followed for a median of 81.2 (70.4–94.9) months. Ten patients (58.8%) relapsed (10 with WG vs 0 with MPA, p=0.15), including four who suffered ≥2 relapses. These relapses occurred a median of 20.5 (17.3–31.9) months after the first remission. Four patients relapsed while receiving maintenance therapy with azathioprine (2 patients), methotrexate (1 patient) or mycophenolate mofetil (1 patient) and six relapsed after maintenance discontinuation, within only a few weeks for four of them. One of these four patients died rapidly from uncontrolled disease and severe septicaemia while receiving oral CYC to treat the relapse.
Identification of factors associated with induction-refractory vasculitis and predictors of mortality
Compared with induction-responders in the WEGENT trial, the only parameters associated with non-response to induction (table 1) in multivariate analysis were alveolar haemorrhage (p=0.03; adjusted OR 2.5, 95% CI 1.1 to5.8) and creatinine level >200 μmol/l (p=0.03; adjusted OR 2.6, 95% CI 1.2 to 6.1).
Among patients with induction-refractory disease, alveolar haemorrhage (p=0.02)—especially massive alveolar haemorrhage (p<0.0001)—was associated with a significantly higher risk of death (figure 2). Early need of haemodialysis was significantly associated with death (p=0.03) whereas a creatinine level >140 µmol/l was not. In contrast, the presence of lung nodules was associated with a lower mortality risk (p=0.01). In the multivariate model including massive alveolar haemorrhage, the need for haemodialysis and lung nodules, only massive alveolar haemorrhage was retained as being associated with higher mortality (p=0.007; HR 9.18, 95% CI 1.86 to 46.37), while the protective effect of having lung nodules did not reach statistical significance (p=0.10; HR 0.34, 95% CI 0.09 to 1.23).
First-line CS and intravenous CYC failed to induce remission in approximately 20% of patients with WG or MPA with systemic involvement.3 4 A switch to oral CYC as intensified second-line therapy was effective for 75% of the patients who did not die very rapidly. However, the relapse rate in these survivors was high (∼60%). For this study population, alveolar haemorrhage and a creatinine level >200 µmol/l were independent factors associated with induction-refractory disease, and massive alveolar haemorrhage was associated with a high risk of death.
Most studies on ANCA-associated vasculitides have focused on the efficacy of first-line induction3,–,5 24 or maintenance regimens7 14 in newly diagnosed patients or the treatment of patients with multiple relapses and/or who failed to respond to several successive therapeutic lines.20 25,–,28 Notably, only a few studies have investigated treatment strategies for patients with vasculitis refractory to first-line treatment, and there are no standardised or consensual therapeutic guidelines for patients with induction-refractory disease. Almost all our patients received oral CYC as the second-line agent, based on the results of a previous study from our group on four patients refractory to intravenous CYC and CS induction11 which showed that three of them achieved remission with oral CYC rescue. These new findings again indicate that oral CYC rescue may represent valuable second-line therapy for patients in whom this treatment is not contraindicated and after dose adjustment for patients with altered glomerular filtration rate. Even though intravenous and oral CYC have been shown in several studies3,–,6 to have similar efficacies as first-line therapy for remission induction, the higher cumulative dose received when switching from intravenous CYC to oral CYC, with its well-known higher toxicity, may be considered therapeutic intensification.
Other strategies have also been used in patients refractory to induction with CYC and CS, some of which were effective. Intravenous immunoglobulin alone or combined with an immunosuppressant has been reported to induce remission,27 29,–,32 but the response in three of the four patients treated with intravenous immunoglobulin in our study was poor, possibly because of the extreme severity of their disease. Only one of our patients underwent plasma exchange. Notably, our study was conducted before the MEPEX study results showed that plasma exchange in patients with severe renal insufficiency (creatinine >500 µmol/l) clearly improved renal outcome.20 Today more patients would probably be treated with plasma exchange, including those with severe renal insufficiency33 34 and also those with alveolar haemorrhage,34 35 even though this practice is only experience-based. In addition, because the WEGENT trial was conducted before biological agents were used in patients with vasculitis, only one patient received biological agents as second-line rescue treatment. Since then, the results of several studies have indicated potentially promising effects for some of these biological agents on relapsing/refractory vasculitides,25 36 37 while other studies have suggested that some biological agents such as etanercept might be harmful or favour the development of cancer.24 The results of recent randomised controlled trials on biological agents and particularly with rituximab, including cost-effectiveness studies, will clearly help to define their place in the therapeutic armamentarium for systemic vasculitides.
For all patients with WG and MPA screened for the WEGENT trial, induction-refractory and randomised baseline characteristics were rigorously and homogeneously recorded on a standard trial form. They have since been prospectively followed. This standardised evaluation helped to identify predictors of induction-refractory disease (eg, alveolar haemorrhage and serum creatinine >200 µmol/l). Unlike elevated creatinine level, alveolar haemorrhage had not previously been reported to be associated with refractory disease.12 13 38 We also identified a subset of patients with fulminating and rapidly fatal vasculitides who died before a switch could be made to second-line rescue therapy. As previously demonstrated39 or suggested40 for MPA, massive alveolar haemorrhage at the time of diagnosis was an independent factor associated with a ninefold higher risk of mortality. In contrast, the presence of lung nodules (eg, granulomatous involvement) was associated with a lower risk of death according to our univariate analysis. Owing to the small sample size, these findings must be interpreted cautiously but are in accordance with the previously reported hypothesis that patients with a granulomatous form of WG (ie, with lung nodules and/or ear, nose and throat involvement) have a better prognosis than those with a vasculitic form (ie, with alveolar haemorrhage and/or glomerulonephritis),13 41 even if granulomatous forms are often associated with persistent grumbling but not life-threatening manifestations.
Patients with induction-refractory disease suffered a high relapse rate (>50%), especially those with WG, as could be expected.10 42 Despite the heterogeneity of maintenance regimens in our study, this observation is in accordance with previous reports of relapse rates of 30–50% with various maintenance regimens.7 8 14 43,–,45
Oral CYC is a valuable rescue treatment for patients with disease refractory to induction with CS and intravenous CYC. Patients with induction-refractory disease are more likely to have severe renal insufficiency and alveolar haemorrhage than those who respond to induction. Reserving its use for these refractory patients seems reasonable and offers the advantage of limiting treatment toxicity, which is inherent in the high cumulative doses of CYC administered orally. However, some patients with refractory vasculitis have rapidly fatal outcomes which do not allow time to initiate any rescue therapy. It is probable that these patients with fulminant vasculitis would also have been refractory to CYC given orally. Early identification of these patients and the use of a more rapidly effective regimen—possibly a combination of treatments—might contribute to lowering the mortality rate and improving their outcomes.
The French Vasculitis Study Group (FVSG) physicians who actively contributed to the FVSG cohort and followed patients included in the present study are listed in the appendix which is available online.
The authors thank Janet Jacobson for editorial assistance and the Société Nationale Française de Médecine Interne for assistance with the recruitment of patients.
Competing interests All the authors were employees of the Assistance Publique–Hôpitaux de Paris or their respective hospitals at the time the study was conducted and the article was written. No funding or grants from pharmaceutical or commercial companies were received in relation with this study.
Ethics approval This study was conducted with ethical committee approval for the WEGENT trial.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.