Predictors of Glucocorticoid Responsiveness in Multicentric Castleman's Disease

Background: Multicentric Castleman's disease (MCD) is a rare lymphoproliferative disease accompanying various symptoms and multi-organ dysfunctions. Although anti-interleukin-6 monoclonal antibodies, tocilizumab and siltuximab are recommended as a therapeutic option, these treatments require patients to visit hospitals at least once a month for many years, and the drug costs are quite high. In Japan, where tocilizumab is the only biological agent covered by the medical insurance, glucocorticoid monotherapy is still an important treatment option. In fact, a part of MCD patients initially treated with glucocorticoids could be successfully controlled without adding or switching to other agents. Considering that emergence of neutralizing anti-drug antibodies which reduced therapeutic efficacy could also be a clinical problem in MCD, there would be a rationale for reserving newer agents for future relapse by using glucocorticoids as first-line treatment for potential responders. Here, we retrospectively analyzed clinical characteristics of MCD patients treated in our institution to explore factors predicting who could be successfully treated with glucocorticoid monotherapy.

Methods: We retrospectively reviewed histologically confirmed Castleman's disease patients who visited the Department of Hematology and Oncology of the University of Tokyo Hospital from January 2000 to March 2020. Based on the diagnostic criteria of Castleman Disease Collaborative Network (CDCN), MCD was diagnosed when enlarged lymph nodes were present in more than 2 lymph node stations and laboratory and/or clinical diagnostic criteria were met. Unicentric Castleman's disease (UCD) was diagnosed when only one single lymph node region was involved. MCD patients who were initially treated with prednisolone (PSL) were divided into two groups: patients who continued PSL until the latest follow-up (PSL-only group) and patients who received subsequent treatment (Second-line group).

Results: 8 patients with UCD and 27 patients with MCD were included in our analysis. With a median follow-up of 5.2 years, 21 MCD patients underwent first-line treatment (PSL, n=18; tocilizumab, n=3). Compared to patients who did not receive any treatment, patients who underwent treatment had marginally higher levels of CRP (7.15 mg/dl vs 4.17 mg/dl, respectively; p=0.066) and significantly lower levels of hemoglobin (9.5 g/dl vs 12.5 g/dl, respectively; p=0.036). Seven out of 18 patients initially treated with PSL had received subsequent treatments (tocilizumab, n=6; rituximab, n=1). Among the PSL-only group, biochemical responses at the latest follow-up could be assessed for 10 in 11 patients: two in complete response, five in partial response, two in stable disease, and one in progressive disease, based on the response criteria of CDCN. Compared to PSL-only group, patients classified into second-line group had higher levels of CRP (11.88 mg/dl vs 6.24 mg/dl, respectively; p=0.024) and lower hemoglobin levels (6.4 g/dl vs9.4 g/dl, respectively; p=0.033). Patients whose CRP levels were lower than 12 mg/dl before starting PSL had significantly longer time to next treatment (TTNT) (median not reached vs 0.88 years; HR, 0.078 [95%CI: 0.013-0.48], p=0.00044). Similarly, patients whose hemoglobin levels were more than 8 g/dl had marginally longer TTNT (median not reached vs 2.63 years; HR, 0.22 [95%CI: 0.040-1.17], p=0.054). In addition, patients with either Hb < 8 g/dl or CRP ≧12 mg/dl had significantly shorter TTNT (median not reached vs 2.12 years; HR, 0.090 [95%CI: 0.010-0.77], p=0.0074). The median PSL dose at the latest follow-up in PSL-only group was 3 mg per day [interquartile range: 0-5 mg per day], which would be comparably safe considering previous reports indicating that PSL dose of less than 5 mg per day was associated with lower incidence of adverse events.

Conclusion: Out data suggest that glucocorticoid monotherapy has a good potential to induce sustained disease control for MCD patients with higher Hb or lower CRP levels. Furthermore, PSL doses could be tapered to safer doses among patients who could continue PSL until the latest follow-up. In contrast, the efficacy of glucocorticoids for MCD patients with lower Hb or higher CRP levels were limited. Such patients would be good candidates for novel agents such as tocilizumab or rituximab.