Risk factors for the recurrence of relapsing polychondritis

Background Although the survival rates of patients with relapsing polychondritis (RP) have increased remarkably, the high recurrence rate remains a significant concern for physicians and patients. This retrospective study aimed to investigate the risk factors for RP recurrence. Methods Patients with RP who presented to Kyoto University Hospital from January 2000 to March 2020 and fulfilled Damiani’s classification criteria were included. Patients were classified into recurrence and non-recurrence groups. Risk factors for RP recurrence were analysed using a Cox proportional hazards model, and Kaplan–Meier survival curves were drawn. Results Thirty-four patients were included. Twenty-five patients (74%) experienced 64 recurrences (mean: 2.56 recurrences per patient). The median duration before the first recurrence was 202 [55−382] days. The median prednisolone dose at the initial recurrence was 10 [5−12.75] mg/day. Tracheal involvement was significantly more frequent in the recurrence group at the initial presentation (44.0% vs. 0.0%, p=0.0172) than in the non-recurrence group, and pre-treatment C-reactive protein levels were significantly higher in the recurrence group than in the non-recurrence group (4.7 vs 1.15 mg/dL, p=0.0024). The Cox proportional hazards model analysis revealed that tracheal involvement (hazard ratio [HR] 4.266 [1.535−13.838], p=0.0048), pre-treatment C-reactive protein level (HR 1.166 [1.040−1.308], p=0.0085), and initial prednisolone monotherapy (HR 4.443 [1.515−16.267], p=0.0056) may be associated with recurrence. The median time before the initial recurrence was significantly longer in patients who received combination therapy with prednisolone and immunosuppressants or biologics (400 vs. 70 days, p=0.0015). Conclusions Tracheal involvement, pre-treatment C-reactive protein level, and initial prednisolone monotherapy were risk factors for recurrence in patients with RP. Initial combination therapy with prednisolone and immunosuppressants may delay recurrence.


Background
Relapsing polychondritis is a rare disease that causes inflammation of the chondrocytes of the ear, nose, and trachea, and it can affect any organ of the body [1].
Although corticosteroid therapy is the mainstream treatment for relapsing polychondritis (RP), immunosuppressive drugs, such as conventional synthetic disease-modifying anti-rheumatic drugs [2,3], biologics [4], and JAK inhibitors [5], have been used empirically in cases of treatment resistance or severe cases. The 5-year and 10-year survival rates of RP were 74% and 55%, respectively, in 1986, whereas higher rates (95% and 91%, respectively) were reported in 2016 [6,7]. This improvement is considered to be due to the early detection of RP, Open Access *Correspondence: yossii@kuhp.kyoto-u.ac.jp the availability of new immunosuppressive drugs, and intensified treatment by individual physicians [7].
However, achieving remission for RP remains challenging. A French study showed that only 19% of the patients achieved complete treatment response in the first 6 months, despite the use of biologics [4]. Furthermore, a recent study showed that most patients with RP have persistent disease activity, despite treatment [8]. In addition, the recurrence of symptoms that require intensive treatment is often encountered in clinical practice and poses a considerable concern for physicians and patients. Repeat recurrences of RP can lead to irreversible organ damage. Particularly, in patients with tracheal chondritis, chronic repetitive inflammation can cause irreversible destruction and fibrosis of the trachea, resulting in a high risk of mortality [9,10]. Therefore, it is crucial to understand the risk factors for RP recurrence, for preventive purposes. However, owing to the rarity of this disease, few reports are available in the literature. Thus, the present study aimed to retrospectively analyse the records of 34 patients with RP to determine the risk factors associated with RP recurrence.

Patients
Patients who presented with RP at Kyoto University Hospital from January 2000 to March 2020 and fulfilled Damiani's classification criteria [11] were included in this study. Patients who were followed up for < 1 year or those who had a history of glucocorticoid use for the treatment of other diseases (i.e. concomitant autoimmune diseases, such as systemic lupus erythematosus, mixed connective tissue disease, and dermatomyositis) before RP treatment were excluded. Auricular, nasal, and tracheal cartilage involvement were diagnosed based on physical examination, imaging (computed tomography and positron emission tomography-computed tomography), or pathohistological findings. RP relapse was defined as (1) worsening or newly developed symptoms related to RP, which were detected by physical examination or imaging studies and had led to the intensification of treatment or (2) the elevation of the levels of inflammatory biomarkers (C-reactive protein [CRP] and erythrocyte sedimentation rate [ESR]) from basal levels, which was considered to be due to the activity of RP and had led to the intensification of treatment.

Evaluation of clinical laboratory parameters
Blood tests included tests for CRP level (normal range ≤0.3 mg/dL), ESR (≤10/h), ferritin level (<280 ng/ mL), IgG level (820−1740 mg/dL), and white blood cell count, which were measured by standard methods. Antitype II collagen antibody level was measured via the enzyme-linked immunosorbent assay technique (Mayo Clinic Laboratories, cut-off > 25 EU/mL).

Statistical analysis
The chi-square test and Wilcoxon's rank-sum test were used to analyse categorical and numerical variables, respectively. Survival curves for treatment were drawn via the Kaplan-Meier method. Univariate analysis was performed to explore the risk factors for RP relapse. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using the Cox proportional hazards model. If no relapse was observed for more than a year, the data were time-censored in the first year. Statistical analyses were performed using JMP version 14 and SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). A p value <0.05 was considered significant.

Characteristics of patients who experienced recurrence
Of the 34 patients, 25 (74%) experienced recurrences (64 recurrences; mean: 2.56 recurrences per patient; minimum 1, maximum 9; 0.16 person-year) ( Table 2). Eleven recurrences (11/64, 17.2%) were major events, requiring hospitalization. The median age of the patients at the time of initial recurrence was 50 (40−67) years, and the median duration before the first recurrence was 202 (55−382) days. Six of the initial recurrences were major events (6/25, 24%). The median prednisolone dose at the time of initial RP recurrence was 10 (5-12.75) mg/ day. The initial symptoms recurred in 68% (17/25) of the cases, and one patient experienced the recurrence of encephalitis.

Risk factors for recurrence
To investigate the risk factors for recurrence, we divided the entire patient cohort (n= 34) into recurrence and non-recurrence groups. There were no significant differences in age, sex, disease duration, or observation period between the two groups before treatment (Table 1, middle and right columns). Although there was no significant difference in the number of symptoms at the initial presentation, occurrence of tracheal lesions at the initial presentation was associated with RP recurrence (44.0% vs. 0%, p=0.0172). Pre-treatment RPDAI tended to be higher in the recurrence group than in the non-recurrence group (median 30 vs. 23, p=0.0786). Serum IgG levels were also significantly higher in the recurrence group than in the non-recurrence group (median: 1704 vs. 1159 mg/dL, p=0.0325), although serum IgG levels remained within the normal range. Monocyte count tended to be higher in the recurrence group than in the non-recurrence group (median 468 vs. 378 /μL, p=0.0786). Positive results of the anti-type II collagen antibody test were not associated with RP recurrence (47.1% vs. 55.6%, p=1.0000). There were no significant differences in the initial prednisolone dose between the recurrence and non-recurrence groups (30 vs. 15 mg, p=0.2717). However, immunosuppressive agents (particularly methotrexate) were more commonly used in the recurrence group than in the non-recurrence group (72.2% vs. 22.2%, p=0.0168). Regarding prognosis, interventions for airway lesions, such as intubation, tracheostomy, and non-invasive positive pressure ventilation use (six patients), were recorded in the recurrence group alone, although no significant between-group difference was observed (24.0% vs. 0.0%, p=0.1622).
We performed a univariate analysis of major risk factors for relapse (Table 3). Pre-treatment CRP level (HR 1.166, 95% CI, 1.040−1.308, p=0.0085) was found to be a potential risk factor for RP recurrence. Monocyte count (HR 1.04, 95% CI, 1.000−1.008, p=0.0690) also tended to be a risk factor, although no significant between-group difference was observed. The hazard ratio for tracheal

Efficacy of combination therapy with immunosuppressants
To investigate the preventive effect of immunosuppressive drugs on relapse, we divided the entire cohort into two groups: PM group and prednisolone combined with immunosuppressants/biologics (PC) group. Table 4 shows the background of patients in each group. The PC group received the following immunosuppressants and biologics: methotrexate (5 patients), intravenous cyclophosphamide (3 patients), azathioprine (1 patient), infliximab (1 patient), and methotrexate, which was changed to tocilizumab (1 patient). There were no differences in age, sex, disease duration, or observation period between the groups. The rate of concomitant use of immunosuppressants was significantly higher when the initial symptom included tracheal involvement (63.6% vs 17.4%, p=0.0160). The dose of prednisolone at the initiation of treatment was not entirely different between the PM and PC groups (28 vs 40 mg, p=0.2566). The median prednisolone dose in both groups at relapse was 10 mg, and the rate of prednisolone dose reduction before the first relapse was not significantly different between the two groups (PM 0.1 vs PC 0.03 mg, p=0.2691). The median duration before initial RP recurrence was significantly longer in the PC group than in the PM group (400 vs. 70 days, p=0.0015).
Recurrence-free survival curve from the start of treatment is shown in Fig. 2. The hazard ratio for RP recurrence in the PM group versus the PC group was 2.718 (95% CI, 0.981−9.565, p=0.0547) ( Table 3). After adjusting for the influence of tracheal lesions, the hazard ratio was 4.443 (95% CI 1.515−16.267, p=0.0056) ( Table 3).

Discussion
In this study, we retrospectively assessed 34 patients with RP and investigated the risk factors for recurrence, which have not been reported before. The characteristics of the patients (age, sex, initial symptoms, and so on) in our study are essentially similar to those of patients described in previous reports [7,12]. We observed preceding infection in 29.4% of the total cases. Til et al. reported the cross-antigenicity of cartilage antigen and heat shock protein 60 from Mycobacterium tuberculosis, and William et al. reported the association between Propionibacterium acnes infection and RP [13,14]. Based on these reports, it is possible that infection is a trigger of RP.
There were two cases of death during the study period. However, the two cases of death were caused by stroke and senility, which were not considered to be related to RP. In a previous report published in 1986, RP was associated with a 5-year survival rate of 74% and a 10-year survival rate of 55% [6], whereas the corresponding values reported in a recent report published in 2016 were 95% and 91% [7]. In the present study, a similar improvement in survival was observed (median observation period, 4.5 years; survival rate, 95.1%).
The present study highlights a high recurrence rate of RP (74%). A recent study showed that RP had a more persistent active pattern than a relapsing-remitting pattern [8]. However, whether RP assumes a chronic active pattern or a relapsing-remitting pattern may depend on differences in treatment and the rate of therapeutic drug reduction. Moreover, further treatment improvement was considered necessary. We found that the median time before the first relapse was 202 days. The median prednisolone dose at the time of recurrence was 10 mg, consistent with an expert's opinion, which suggests that a 5−10-mg maintenance dose of prednisolone is suitable for RP [15]. In our analysis, 68% of the patients had similar symptoms at the initial recurrence, while the symptoms in 32% of the patients at the time of relapse were different from the initial symptoms. This suggests that various symptoms of RP should be considered at the time of recurrence. One patient had auricular chondritis as the initial symptom but had encephalitis symptoms at the time of RP recurrence. This finding is consistent with those of Shimizu et al. 's [16] cluster analyses, which showed the relationship between auricular chondritis and encephalitis.
In the analysis of the risk of recurrence, tracheal involvement, pre-treatment CRP level, and PM were identified as potential risk factors for recurrence. Tracheal involvement was associated with death in previous studies [17][18][19], and the possible mechanisms include ventilation insufficiency or concomitant infection due to airway narrowing or collapse [20,21]. The findings of the present study suggested that tracheal involvement is a risk factor for recurrence, and that repeat recurrences may lead to irreversible structural changes in the bronchi.
Pre-treatment RPDAI score was significantly higher in the recurrence group than in the non-recurrence group, although pre-treatment RPDAI was not a recurrence risk factor in the univariate analysis. The RPDAI is an index of activity that scores symptoms and examination findings in RP, using a scale of 1-24 points [22]. A relatively high score of 14 or 24 points is assigned to tracheal chondritis, depending on the presence or absence of acute respiratory failure. The recurrence group had a significantly higher median RPDAI score, which may have been influenced by tracheal involvement. A new assessment scale for assessing disease activity is required.
Pre-treatment CRP levels were also identified as risk factors for recurrence in the present study, although the results differ from those of several previous reports. Consistent with our study, which suggests that a high pretreatment CRP level is a risk factor for recurrence, a case  . 2 Recurrence-free survival curve of the prednisolone monotherapy (PM) and prednisolone combined with immunosuppressants/biologics (PC) groups. One-year relapse-free survival curves for patients who received prednisolone monotherapy as the initial therapy and for patients who received a combination of prednisolone and immunosuppressants. Wilcoxon's rank-sum test p=0.0372 (log-rank test p=0.0655)