Introduction. Multicentric Castleman Disease (MCD) is an inflammatory disorder characterized by lymphadenopathy. The clinical course of MCD is variable; some patients have mild symptoms while others develop hypercytokinemia progressing to organ failure. Severely ill patients often have the TAFRO-variety of MCD characterized by Thrombocytopenia, Anasarca, Fever, Reticulin marrow fibrosis, and Organomegaly. We report on 76 MCD patients: the largest series from a single institution.

Methods. Data was collected from our institution and outside records via systematic chart review and standardized survey for self-reported symptoms. Laboratory data represent each patient's most severe episode. TAFRO- and non-TAFRO or "classical" iMCD subtypes were grouped (iMCD) for some analyses. Patients were classified as POEMS-associated MCD when they had MCD with full POEMS syndrome or clear features of POEMS.

Results. Forty-nine patients had classical iMCD and 12 TAFRO-iMCD. Fifteen patients had POEMS-associated MCD, of which 8 had coexistent POEMS syndrome. iMCD and POEMS-associated MCD had similar age and sex, but the hyaline vascular variant was more common in iMCD (44.3 vs 6.7%, p=0.006) (Table 1). iMCD patients were more symptomatic in terms of night sweats (61.0 vs 33.3%, p=0.001) and fever (50.8 vs 20%, p=0.031) compared to POEMS-associated MCD, but had similar laboratory characteristics and disease activity (CHAP) scores. TAFRO-iMCD patients were mostly male and had more fever (91.6 vs 42.9%, p=0.002), anasarca (100 vs 24.5%, p<0.0001), higher creatinine (2.6 vs 0.9 mg/dL, p=0.001), poorer CHAP scores (5.5 vs 2.0, p=0.01), and a trend to lower hemoglobin. Interleukin-6 (IL6) levels were not different among groups, but vascular endothelial growth factor (VEGF) was higher in TAFRO-iMCD (328.5 vs 79.5 pg/mL, p=0.034), accounting for anasarca.

Commonly-used treatment regimens for iMCD were the anti-IL6/IL6 receptor antibodies siltuximab and tocilizumab (n=41), rituximab ± steroids (n=23), steroids alone (n=13), and chemotherapy (n=7). There was a high rate of response to anti-IL6 therapy (97%) with a treatment failure (TF) rate of 5%. Two further relapses occurred in patients who discontinued therapy. Steroid monotherapy or rituximab + steroids yielded response rates of 46 and 82%, but were marred by high TF rates (85 and 43%). Chemotherapy responses were noted in 7 instances, but TF occurred twice. There were no significant differences in response to therapy comparing the classical and TAFRO-iMCD groups, but both patients failing anti-IL6 therapy had TAFRO-iMCD. Six TAFRO-iMCD patients received chemotherapy and 2 relapsed. Six patients with POEMS and CD underwent autologous stem cell transplant with major improvement in polyneuropathy and MCD; 2 experienced relapse of their MCD and one died of late myelodysplastic syndrome. One improved post irradiation of a plasmacytoma and 1 died of progressive disease. Most MCD patients with POEMS features received rituximab-based therapies; 6/7 were alive and 1 died of sepsis. There was no difference in overall survival (OS) among the three groups (p=0.35) with a median OS of 5.65 years (4 - 9.2, 95% CI) in the iMCD group, 7.9 years (4.5 - 11.8, 95% CI) in POEMS-associated MCD, and 5.85 years (3.2 - 8.7, 95% CI) in the TAFRO group. The 5-year OS was 91%, 94% and 100% for the iMCD, POEMS-associated MCD and TAFRO groups, respectively.

Conclusions. In a tertiary referral setting, excellent outcomes were achieved in all subgroups, likely attributed to highly selective therapeutic choices. Anti-IL6 therapy for iMCD was reserved for patients with active disease and laboratory evidence of inflammatory syndrome. Some TAFRO-iMCD patients achieved complete responses with anti-IL6 therapy, while others fared well with chemotherapy. Rituximab-based therapy was given to those with more indolent iMCD and those with POEMS-associated MCD. The polyneuropathy of POEMS benefitted most from transplant. Significant differences were noted in VEGF levels both in POEMS-associated MCD and TAFRO-iMCD compared to classical iMCD cases, with each cohort presenting a distinct clinical picture suggesting differing underlying etiologies and cytokine profiles. IL6 levels were not different among groups, though this may be confounded by effects of prior therapy. A better understanding of the pathobiology of the MCD continuum is essential to developing more targeted therapies.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution