Oral candidiasis is a significant predictor of subsequent severe infections during immunosuppressive therapy in anti-neutrophil cytoplasmic antibody-associated vasculitis

Background Several studies have identified predictors of severe infections in antineutrophil cytoplasmic antibody-associated vasculitis (AAV). However, the development of oral candidiasis (OC) as a predictor of subsequent severe infections has not been evaluated. The aim of this study was to assess the association between OC and subsequent severe infection requiring hospitalization during immunosuppressive therapy in AAV. Methods This single-center retrospective cohort study included 71 consecutive patients with newly diagnosed AAV from Aichi Medical University Hospital, Japan, starting immunosuppressive therapy between March 2013 and December 2018. The relationships between OC and subsequent severe infections were assessed using multivariate Cox proportional hazards models, adjusted for clinically relevant factors. Results During the follow-up period (median, 23 months; interquartile range, 11–51 months), 25 severe infectious episodes occurred in 19 patients (26.8%) and OC occurred in 17 patients (23.9%). A log-rank test showed that the OC group was significantly associated with severe infection (P <  0.001). Multivariate Cox proportional hazards models identified lower serum albumin (per 1 g/dl adjusted hazard ratio (HR)  = 0.38, 95% confidence interval (CI): 0.15–0.85; P  =  0.018), use of methylprednisolone pulse (adjusted HR  =  5.44, 95% CI: 1.54–20.0; P  =  0.010), and OC (adjusted HR  = 5.31, 95% CI: 1.86–15.8; P  =  0.002) as significant predictors of severe infection. Furthermore, a significant effect modification of the use of methylprednisolone pulse on OC was observed (P <  0.001). Conclusions OC is one of the predictors of subsequent severe infections. The results suggest the importance of prolonging infection surveillance, especially for patients who developed OC under strong immunosuppressive therapy. Electronic supplementary material The online version of this article (10.1186/s12879-019-4300-0) contains supplementary material, which is available to authorized users.

Although several studies have reported that older age, smoking, worsened kidney function, low level of CD4+ T cells, glucocorticoid, and CYC therapy are significant predictors of severe infection [15][16][17], other predisposing factors remain unidentified. To the best of our knowledge, no previous studies have focused on the association between oral candidiasis (OC) and subsequent severe infections occurring with AAV therapy and have investigated the incidence of OC in patients undergoing AAV therapy. The occurrence of OC might be a sign of cell-mediated immune decline; thus, we hypothesized that OC might be a predictor of subsequent severe infection in AAV. The aim of this study was to assess the association between OC and subsequent severe infection during immunosuppressive therapy in AAV.

Patients
Our retrospective cohort study included patients aged > 20 years diagnosed with AAV including GPA, MPA, and EGPA on the basis of the European Medicines Agency algorithm [18] with a consensus methodology for the classification of the AAV between March 2013 and December 2018 at Aichi Medical University Hospital. The exclusion criteria were as follows: patients who had been started on immunosuppressive therapy for AAV at another hospital or patients receiving no immunosuppressive therapy. The study protocol was approved by the Ethics Committees of Aichi Medical University (approval number 2018-H350). The requirement for informed consent was waived given the retrospective nature of the study.

Measurements
The clinical characteristics at the time of starting immunosuppressive treatment were used as baseline,  [20], organ involvement, anti-MPO and anti-PR3 ANCA titers, OC, and immunosuppressive treatment; induction immunosuppressive therapy; use of methylprednisolone pulse therapy (0.5 or 1.0 g/d for 3 consecutive days), glucocorticoid monotherapy, intravenous CYC and RTX, maintenance immunosuppressive therapy, glucocorticoid monotherapy, oral CYC, azathioprine, methotrexate, and RTX; point and cumulative prednisolone (PSL) dose; concurrent use of other immunosuppressants at 0, 3, 6, 12, and 24 months after initial immunosuppressive therapy; and adverse events including severe infection during the follow-up period.

Outcomes
In this study, the main exposure was the development of OC. OC was defined as clinical signs of oral thrush treated with antifungal drugs, such as itraconazole or fluconazole. The main outcome of interest was the development of severe infection. Severe infection was defined as infection requiring hospitalization. Remission was defined as the absence of clinical signs and symptoms of active vasculitis (BVAS = 0) for more than 2 months. A relapse was defined as clinical signs of vasculitic activity in any organ system following remission [22]. The patients were divided into those who developed OC (OC group) and those who did not (non-OC group). Patients were followed until September 2018 and censored at the time of death (if before primary outcome).

Statistical analysis
Differences in clinical characteristics between the OC group and the non-OC group were compared by using the Wilcoxon rank-sum test or Fisher's exact test. To evaluate predictors of severe infection, univariate and multivariate Cox proportional hazards (CPH) models were constructed, including age, sex, serum albumin level, serum creatinine level, methylprednisolone pulse, and OC.
To assess whether an association between methylprednisolone therapy and outcome was different in the OC and non-OC groups, the effect modification between methylprednisolone pulse therapy and the OC group was assessed by the inclusion of interaction terms in the multivariate Cox proportional hazards models.
The proportional hazards assumption for covariates was tested using scaled Schoenfeld residuals. For continuous variables, the Wilcoxon rank-sum test was performed to evaluate the significance of intergroup differences. Categorical variables were expressed as percentages and compared using Fisher's exact test. The cumulative probability of the development of first severe infection was calculated using the Kaplan-Meier method and log-rank test. The level of statistical significance was set at P < 0.05. All statistical analyses were performed using JMP version 14.0.0 (SAS Institute, Cary, NC, USA) and STATA version 13.0 (StataCorp LP, College Station, Texas).

Baseline characteristics
From a total of 77 such patients, we excluded 5 (6.5%) patients who had been on immunosuppressive therapy in another hospital and 1 (1.3%) patient with a lack of immunosuppressive therapy. Finally, 71 (92%) consecutive patients with newly diagnosed AAV who were on immunosuppressive therapy were included in this study. The study population consisted of 71 AAV patients, including those with MPA (n = 58), GPA (n = 1), and EGPA (n = 12). During a median observation period of 23 months (interquartile range, 11-51 months), 17 (23.9%) patients developed OC (OC group) after median of 13 days (interquartile range: 11-19 days) of starting immunosuppressive therapy, and 54 (76.1%) had never developed OC (non-OC group). The clinical characteristics of the two groups (OC-group and non-OC group) are summarized in Table 1.
Baseline characteristics were not different significantly between the OC and non-OC groups. Methylprednisolone pulse therapy was administered in 15 (27.8%) and 9 (52.9%) patients in the non-OC and OC groups, respectively (P = 0.078).
During the observation period, 2 (3.7%) patients in the non-OC group and 2 (11.8%) patients in the OC group developed end-stage renal disease requiring permanent dialysis therapy. Furthermore, 4 (7.4%) patients in the non-OC group and 1 (5.9%) patient in the OC group died, and the causes of death were infection (n = 3), CVD (n = 1), and unknown (n = 1).

Oral candidiasis and severe infection
During the observation period, a total of 25 severe infection episodes occurred in 19 patients (26.8%), wherein some patients had several infection episodes. The median time from immunosuppressive therapy initiation to the occurrence of first severe infection was 13 months (interquartile range: 7-26 months) in the OC group and 16 months (interquartile range: 1-49 months) in the non-OC group. In the OC group, all severe infections occurred after an OC episode at a median of 2.5 months (interquartile range: 0.3-15 months).
During the observation period before the occurrence of severe infection (or censoring without outcome), no significant difference was observed between the OC and non-OC groups in terms of the point and cumulative PSL dose, and using immunosuppressants at 0, 3, 6, 12, and 24 months after initiating immunosuppressive therapy (Table 3). These data suggest that differences in the intensity of immunosuppressive therapy between OC and non-OC groups did not explain the differences in the occurrence of severe infection.

Discussion
This study revealed that OC was significantly associated with subsequent severe infection, after adjusting for important preventable risk factors, suggesting that OC might be a significant predictor of subsequent severe infection. Furthermore, significant effect modifications  [1-5, 11-17, 22-27], and the main injured organ due to infection was the lungs [23]. Several risk factors for severe infections have been identified, including older age at diagnosis, severe kidney involvement at diagnosis, low level of CD4 + T cell, and immunosuppressive treatment using high-dose glucocorticoid and CY [11][12][13][14][15][16][17][22][23][24][25][26][27]. Although our result was compatible with those of previous studies, which showed that patients who use high-dose glucocorticoids such as methylprednisolone pulse therapy were at a higher risk for severe infection, no previous studies have focused on the association between OC and subsequent severe infection in AAV.  OC is a common opportunistic infection of the oral cavity caused by an overgrowth of Candida species, and the most common is Candida albicans [28]. The reported risk factors for OC were broad-spectrum antibiotics, immunosuppressive drugs, smoking, diabetes, Cushing's syndrome, immunosuppressive conditions such as human immunodeficiency virus infection, malignancies such as leukemia, and nutritional deficiencies [28]. In this study, we considered that in AAV patients, OC might be an important sign of decreased cellular immune function, predisposing the patient further to subsequent severe infections. This study also showed that glucocorticoid dose and use of immunosuppressive treatment during the observation period were clinically comparative between the OC and non-OC groups, suggesting that differences in the intensity of immunosuppressive therapy between OC and non-OC did not explain the differences in the occurrence of severe infections. Interestingly, this study also showed an interaction between OC and methylprednisolone pulse therapy, suggesting that patients who developed OC under strong immunosuppressive therapy might have an increased risk of severe infection.
Regarding OC treatment, although all patients were prescribed oral antifungal drugs such as itraconazole or fluconazole for systemic effects, we consider that OC treatment itself does not directly influence the outcome of severe infection.
As immunosuppressive therapy, glucocorticoid monotherapy was frequently used for the treatment of AAV in the present study based on previous Japanese studies [29,30]. Although a recent study showed that glucocorticoid monotherapy is considered less effective than combination therapy with an immunosuppressive agent, such as CYC or RTX, in other countries, it is unknown whether elderly MPA patients who might be at a high risk of infection should use these aggressive immunosuppressive therapies. In our cohort, glucocorticoid monotherapy was frequently used for elderly patients considering the risk of infection; therefore, we could not evaluate the relationship between IVCY or RTX therapy and severe infection because the number of patients administered with IVCY or RTX was insufficient to evaluate it. This should be considered when interpreting our results.
This study also showed that lower serum albumin level was a significant risk factor for severe infection in AAV patients, as previously reported [24]. The present study suggests that patients with AAV presenting hypoalbuminemia should be carefully managed for severe infections.
Our study has some limitations. First, given the retrospective nature of this study, we should consider that unmeasured factors associated with the treatment may not be included in the model. Second, this study has a single-center small cohort design and the observation period was short; therefore, our results should be validated in other multicenter large cohorts with longer follow-ups. Third, in Japan, most AAV patients were elderly MPA patients; thus, the subjects in this study may not be representative of all AAV patients. Fourth, several studies recommend CYC or RTX therapy for AAV patients as induction therapy [10][11][12][13][14]. However, few patients in our study were treated with RTX or CYC. Therefore, we could not assess the influence of RTX or CYC on the outcome; a larger cohort using the treatment should be included to re-evaluate our results. Thus, we advise caution when interpreting and generalizing our results. Despite these methodological issues, to the best of our knowledge, this study is the first to describe the relationship between OC and subsequent severe infection in AAV patients.

Conclusions
OC is one of the predictors of subsequent severe infections. The results suggest the importance of prolonging infection surveillance, especially for patients who developed OC under strong immunosuppressive therapy.