Highly active antiretroviral therapy alone may be an effective treatment for HIV-associated multi-centric Castleman’s disease

Multicentric Castleman’s disease (MCD) is a rare lymphoproliferative disorder mainly seen in HIV-infected patients and associated with poor prognosis. Pre-HAART mortality was 70–85%.1 Proliferation of polyclonal but often monotypic plasmablastic cells is thought to be driven by Kaposi sarcoma herpes virus (KSHV) infection, present in all patients with HIV-associated MCD.2 There is no standard treatment. Interventions include rituximab, lymphoma-type treatment with chemotherapy and splenectomy,1,3–7 Many are used concomitantly making it difficult to ascertain which is of greatest efficacy.

We describe 4 cases of biopsy-proven HIV-associated MCD who showed a complete clinical and radiological response to MCD, and a reduction in KSHV viral load to HAART alone without additional therapeutic interventions. All patients are alive and relapse free 19–38 months later.

Figure 1.Composite histological figure showing features of MCD in lymph node biopsy. (A) Routine Hematoxylin and Eosin (H&E) section of a needle core biopsy of node showing an abnormal germinal center with irregular outline and scattered large plasmablastic cells (x110 original magnification). (B) High power H&E from a whole node biopsy showing a characteristic regressed germinal center with plasmablastic cells (x220 original magnification). (C) KSHV immunohistochemistry showing a target-like arrangement of KSHV-positive plasmablasts around a germinal center. (D) Lambda light chain immunohistochemistry showing cytoplasmic positivity in the plasmablastic cells in a germinal center.

Between February 1 2007 and October 1 2008, we identified 4 consecutive patients who received HAART alone for lymph node biopsy-proven HIV associated-MCD (Table 1). Median duration of HIV infection prior to MCD was two years (range 1–14). In all patients, nodes showed a mixture of reactive lymphoid follicles and follicles with irregular, regressed germinal centers. At the periphery of these regressed germinal centers were large plasmablastic cells with prominent nucleoli and basophilic cytoplasm, usually CD20 negative, and showing positive staining for KSHV, IgM and lambda light chains (Figure 1). All patients presented with fever, lymphadenopathy, hepatosplenomegaly and elevated C-reactive protein (CRP). Three patients investigated with FDG PET/CT scan showed FDG avid lymphadenopathy above and below the diaphragm. Three patients had started HAART less than 3–5 weeks before presentation with MCD; HAART had been initiated due to a combination of CD4 decline and constitutional symptoms. HAART was modified in one patient and interrupted in one patient due to renal and liver impairment. In all patients, resolution of constitutional symptoms occurred within three months of starting HAART. Patient 2 had a recrudescence of symptoms (fever, splenomegaly and elevated CRP) eight months after diagnosis of MCD which resolved spontaneously. Serum KSHV was 6,600 copies/mL during this flare. Three months previously KSHV had been 3,600 copies/mL. Three patients who initially demonstrated significant FDG avid lymphadenopathy now showed complete metabolic response on repeat PET/CT scan and the remaining patient CRu on repeat whole body CT scan. At time of MCD diagnosis, all samples from patients were positive for KSHV between 420 and 120,000 copies/mL; 3 patients had sustained undetectable KSHV viral loads following antiretroviral therapy. KSHV was detectable in 2 patients in whom retrospective sampling of stored blood was available (Patient 2: KSHV 2,050 copies/mL seven months before diagnosis; Patient 3: KSHV 180 copies/mL three months before diagnosis).

Table 1.Characteristics of 4 biopsy-confirmed cases of plasma cell variant Castleman’s disease and clinical, virological and radiological response to antiretroviral therapy.

Previous reports regarding the utility of HAART as treatment of MCD are conflicting. Anecdotal case reports describe clinical and radiological benefit from HAART in patients naïve to antiretroviral therapy.8–10 In one case, remission of MCD and reduction in KSHV viremia were attributed to rituximab.4 However, HAART had been initiated prior to improvement in both. Failure of HAART has been reported by De Jong et al. who describe the occurrence of MCD despite full suppression of HIV infection with HAART.11 Clinical deterioration despite HAART alongside promising clinical outcomes of newer therapies such as rituximab,5 has led many physicians to advocate the use of rituximab or chemotherapy as first-line treatment of MCD. However, benefit ascribed to newer agents may be confounded by the natural history of MCD and the co-prescribing of HAART.

The role of other therapies instead of, or in addition to HAART is not clear. Rituximab, an anti-CD20 monoclonal antibody, is increasingly being adopted for first-line treatment due to good tolerability and reported clinical activity, despite little understanding of its anti-MCD activity, as many of the KSHV infected plasmablastic cells do not express CD20.3–5,12 In a case series of 21 patients with HIV-associated MCD, rituximab resulted in clinical and radiological response. However, the temporal relationship between clinical response to rituximab and institution of HAART is not described.5 In one case, remission of MCD and reduction in KSHV viremia were attributed to rituximab.4 However, HAART had been initiated prior to this improvement. Based on our findings, the benefits of many anti-MCD treatments might be attributable to concurrently administered HAART that suppresses KSHV viremia, probably via restoration of immune function, as evidenced by increased CD4 counts and reduction of HIV viral loads. While optimal treatment for HIV-associated MCD remains undefined, our findings suggest HAART alone may represent an effective treatment for MCD patients, avoiding unnecessary splenectomy or chemotherapy with attendant toxicity and impeding immune reconstition/CD4 recovery.

In 3 of the 4 patients described, MCD was diagnosed 3–5 weeks after starting HAART. They presented with persisting constitutional symptoms with generalized lymphadenopathy and hepatosplenomegaly, resembling an immune reconstitution phenomena, an observation previously suggested by Zeitz.13 However, we demonstrated that continuation of HAART was associated with clinical improvement of MCD in our patients who remain disease free 23–38 months later, suggesting that the initial deterioration seen following HAART may be transient and not imply treatment failure. The remaining patient was antiretroviral naïve at MCD diagnosis, with CD4 count of 250 cells/mm and no organ failure. By comparison, the patient described by De Jong et al. was already taking HAART, initiated at a significantly lower CD4 count (15 cells/mm).11

KSHV was detected in 2 patients for whom sampling of stored blood was available. This observation raises the possibility that the incidence of HIV-associated MCD may be far higher than reported in the literature, because HAART given to many KSHV co-infected patients, presenting with constitutional symptoms and low CD4 counts but with undiagnosed MCD, may induce clinical remission before diagnosis of MCD. KSHV and FDG-PET/CT testing of HIV-infected HAART-naïve patients presenting with constitutional symptoms and enlarged glands may result in identification of more cases of MCD by performing biopsy for histological diagnosis.

These 4 patients demonstrate that HAART may be an effective treatment for HIV-associated MCD. Currently there is a lack of data to offer clear guidance to clinicians on treatment strategy. Multi-center clinical trials are needed to examine the role of HAART for first-line treatment of MCD.

Acknowledgments

Professor Lee is partly funded by the UCLH/UCL Comprehensive Biomedical Research Centre.