Diagnostic Criteria Schemes for Multicentric Castleman Disease in 75 Cases

Bower, Mark MD, PhD*; Pria, Alessia Dalla MBBS*; Coyle, Christopher MBBS*; Nelson, Mark MBBS; Naresh, Kikkeri MBBS, PhD

JAIDS Journal of Acquired Immune Deficiency Syndromes: 1 February 2014 - Volume 65 - Issue 2 - p e80–e82
doi: 10.1097/QAI.0b013e3182a8338f
Letters to the Editor

Departments of *Oncology;

HIV Medicine; and

Histopathology, Chelsea and Westminster Hospital, London, United Kingdom

The authors have no funding or conflicts of interest to disclose.

Article Outline

To the Editors:

The diagnosis of HIV-associated plasmablastic multicentric Castleman disease (MCD) is based on the presence of histopathological features. However, MCD is a relapsing and remitting disease, and so, clinical correlates should also be present to confirm active disease.1 Although there are no evidence-based gold standard criteria for establishing a diagnosis of active MCD, both the French ANRS (Agence Nationale de Recherche sur le SIDA) 117 CastlemaB trial group and the National Cancer Institute (NCI) have described criteria to define an attack of MCD.2,3 The French ANRS definition requires raised serum C-reactive protein (CRP) (in the absence of any other cause), pyrexia, and at least 3 of 12 clinical features.2 The NCI scheme requires raised serum CRP, at least 1 clinical symptom, and 1 laboratory abnormality probably or definitely attributed to MCD.3 The serum CRP cutoff is higher in the French (>20 mg/L) than in the US (>3 mg/L) scheme. Neither system has been validated on an independent data series of patients.

We applied the 2 diagnostic schemes to our cohort of 75 newly diagnosed patients treated for MCD (including 61 previously described cases4). All patients had histologically confirmed MCD with IgM lambda restricted plasmablasts with positive immunostaining for Kaposi's sarcoma herpes virus (KSHV). Central histopathological review was completed for 64 cases by a single pathologist (K.N.). Full clinical and treatment details were prospectively collected since 2000 and retrospectively collated for 3 cases diagnosed before 2000. The publication of the ARNS definition of an attack of MCD in 20072 included 2 features not previously recorded (xerostomia and nasal obstruction); these were added to the prospective data collection in late 2007. Laboratory measurements were recorded at the time that the diagnosis was established. The median duration of symptoms at the time that the diagnosis was established was 3 months (maximum, 48 months).

The median age at MCD of the 75 patients (89% male) was 42 years (range, 23–69 years). The median CD4 cell count was 233 cells per cubic millimeter (range, 24–1429 cells/mm3), and the median interval between HIV diagnosis and MCD diagnosis was 2.5 years (range, 0–28 years). Two were diagnosed before 1996 (the date of routine use of combination antiretroviral therapy in our cohort) and 26 (35%) had an AIDS defining illness before MCD. Thirty-two (43%) were established on highly active antiretroviral therapy for at least 3 months at MCD diagnosis, of whom 18 (56%) had an undetectable plasma HIV viral load (<400 to <50 copies/mm3 depending on assay used) and a further 11 (34%) had low level HIV viremia <500 copies per cubic millimeter. The median duration of symptoms before the diagnosis of MCD was 3 months (range, 0.5–48 months). The median follow-up is 5.6 years (maximum, 19 years), and the overall 5-year survival for this cohort is 78% (95% confidence interval, 67% to 88%). At MCD diagnosis, 4 (5%) had histological evidence of “frank” plasmablastic lymphoma (n = 1) or “microlymphoma” (n = 3) coexisting with MCD.5,6

The ANRS scheme requires patients to fulfill all of the 3 criteria: 74/75 (99%) had a fever and 71/75 (95%) had at least 3 of the 12 clinical criteria; however, CRP measurement at the initial MCD diagnosis was only available for 68 patients, and only 55 (81%) had levels above 20 mg/L. Overall, 13 of the 68 patients with full data available failed to achieve compliance with all 3 criteria, giving a specificity of 81% for the ANRS scale (Table 1). Of the 13 patients who did not meet the 3 standards, only 9 failed because their serum CRP measurement was below 20 mg/L, whereas the remaining 4 patients had in addition the absence of fever (n = 1) or fewer than 3 clinical features (n = 3). The NCI criteria similarly requires fulfillment of all 3 categories, and 75/75 (100%) had at least 1 clinical criteria; 68/72 (94%) had at least 1 laboratory abnormality meeting the criteria, and 66/68 (97%) with serum CRP measurements available at diagnosis had levels above 3 mg/L. Overall, 6 of the 68 patients with full data available failed to achieve compliance with all 3 criteria, giving a specificity of 92% for the NCI scale (Table 1). Four of the 6 patients who did not meet the NCI criteria did not have the necessary abnormal laboratory levels, whereas 2 other patients had serum CRP values below 3 mg/L.

Although both systems categorize the majority of our patients as having active MCD, the looser criteria in the NCI scheme identified more of our cohort as having active MCD. The false-negative rates are 19% for ANRS and 8% for NCI. Both classification models were cumbersome and included redundancy; for example, every patient with jaundice, ascites, or edema already had more than 2 of the other clinical criteria in the ARNS classification, whereas every patient with fatigue and weight loss also had fevers or night sweats in the NCI scheme. Moreover, the systems allow interpretation that may introduce confusion. For example, many patients may have gastrointestinal symptoms that relate to antiretroviral therapy rather than MCD, and jaundice may be a consequence of atazanavir rather than MCD. This study has not attempted to establish the capacity of either scheme to correctly exclude patients without MCD, and so, the specificity and power of the 2 schemes cannot be established.

The combination of histopathological features, detectable plasma KSHV DNA, and B symptoms could provide a simpler system that could be widely applied without ambiguity. In our cohort, all 75 patients had fever, night sweats, or weight loss, and all had histopathological features of MCD, including IgM lambda–restricted KSHV LANA-positive plasmablasts. Plasma levels of KSHV were measured at diagnosis in 60 of the 75 patients and were detected in 58 (median, 237,000 copies/mL; range, 0–554,000,000 copies/mL). In 2006, a new assay for measuring plasma levels of KSHV was introduced, and since that time, all 47 new patients have had detectable levels.

The importance of diagnostic clinics criteria that complement histological findings remain for KSHV-negative idiopathic MCD, where the histological diagnosis is often less than straight forward and no serological markers for disease exist. However, in people living with HIV and KSHV associated MCD, the histological diagnosis has been made considerably easier by the emergence of immunostaining for KSHV and the presence of IgM lambda restriction analyses. Moreover, the detection of elevated KSHV DNA in the blood greatly assists the diagnosis. Under these circumstances, the value of clinical criteria is more limited, and the imprecision further limits their worth. The triad of histopathological findings, detectable plasma KSHV, and B symptoms should suffice.

Back to Top | Article Outline


1. Du MQ, Bacon CM, Isaacson PG. Kaposi sarcoma-associated herpesvirus/human herpesvirus 8 and lymphoproliferative disorders. J Clin Pathol. 2007;60:1350–1357.
2. Gerard L, Berezne A, Galicier L, et al.. Prospective study of rituximab in chemotherapy-dependent human immunodeficiency virus associated multicentric Castleman's disease: ANRS 117 CastlemaB trial. J Clin Oncol. 2007;25:3350–3356.
3. Uldrick TS, Polizzotto MN, Yarchoan R. Recent advances in Kaposi sarcoma herpesvirus-associated multicentric Castleman disease. Curr Opin Oncol. 2012;24:495–505.
4. Bower M, Newsom-Davis T, Naresh K, et al.. Clinical features and outcome in HIV-associated multicentric Castleman's disease. J Clin Oncol. 2011;29:2481–2486.
5. Dupin N, Diss TL, Kellam P, et al.. HHV-8 is associated with a plasmablastic variant of Castleman disease that is linked to HHV-8-positive plasmablastic lymphoma. Blood. 2000;95:1406–1412.
6. Dargent JL, Lespagnard L, Sirtaine N, et al.. Plasmablastic microlymphoma occurring in human herpesvirus 8 (HHV-8)-positive multicentric Castleman's disease and featuring a follicular growth pattern. APMIS. 2007;115:869–874.
© 2014 by Lippincott Williams & Wilkins