Incidence increasing of HIV-associated multicentric Castleman's disease, a relatively rare lymphatic cancer

Increasing numbers of cases of the relatively rare lymphatic cancer, multicentric Castleman's disease, are being diagnosed in HIV-positive individuals attending the UK’s largest HIV clinic, according to a presentation to the XVII International AIDS Conference in Mexico City earlier this month by Dr Mark Bower of London’s Chelsea and Westminster Hospital.

Castleman’s disease, first described 50 years ago by Dr Benjamin Castleman, is a disorder resulting from over-activity of lymph tissue that can develop into a type of lymphoma. Although rare, it is most often seen in HIV-positive individuals as multicentric Castleman's disease (MCD) which is characterised by swollen lymph nodes, fever, fatigue, weight loss, night sweats, blood disorders and sometimes liver and spleen irregularities.

Like Kaposi’s sarcoma (KS), MCD is caused by human herpesvirus-8 (HHV-8). Unlike KS, however, is it very difficult to diagnose and, as Dr Bower revealed during his presentation, MCD is neither associated with a prior AIDS diagnosis, nor are men disproportionately affected.

No information exists regarding the incidence and risk factors for MCD in HIV-positive individuals, so these data from the Chelsea and Westminster cohort – which includes 10,997 patients, prospectively followed for over 56,202 patient-years between 1983 and 2007 – provide the first overview of the impact of MCD on HIV-positive people.

Since MCD and KS are caused by the same virus, the investigators compared the incidence of both diseases over time. During 25 years of follow-up there were 1180 diagnoses of Kaposi's sarcoma in the Chelsea and Westminster cohort, compared to just 24 cases of multicentric Castleman's disease. This translates to a MCD incidence of 4.3 cases per 10,000 patient-years – about one-fiftieth of KS incidence.

However, when incidence was examined during three time periods – the pre-highly active antiretroviral therapy (HAART) era (1983-1996); the early HAART era (1997-2001); and the current HAART era (2002-2007) – incidence of multicentric Castleman's disease was found to have risen progressively, from 0.58 cases per 10,000 patient-years in the pre-HAART era; to 2.8 cases per 10,000 patient-years in the early HAART era; to 8.3 cases per 10,000 patient-years in the current HAART era.

Dr Bower conceded that, “there may be some diagnostic ascertainment bias” in the current era, since "we're more familiar with the diagnosis and more likely to make the diagnosis and search for it a bit harder."

However, he pointed out, rising MCD incidence in this cohort comes in stark contrast to declining KS incidence over the same three time periods. Multivariate analysis found that unlike the risk of developing KS – which is much greater in men, and much greater with a prior AIDS diagnosis – this was not the case for multicentric Castleman's disease. “There was no gender preference for developing Castleman's disease,” said Dr Bower.

The factors that were associated with an increased risk of developing multicentric Castleman's disease included:

• increasing age;
• non-Caucasian ethnicity – the risk was higher particularly in black African members of the cohort;
• a shorter duration since HIV diagnosis; and
• higher, rather than lower, nadir CD4 cell counts, unlike most HIV-related cancers – one third of MCD patients in the cohort had CD4 counts above 350 cells/mm³.

In addition, the analysis found that HIV treatment appears to protect against the development of multicentric Castleman's disease.

Although he said that MCD is "a diagnosis that we're going to be making more frequently", Dr Bower pointed out that “the diagnosis is very difficult to make” since clinical and pathological features of MCD overlap with a large number of other illnesses in people who are HIV-positive.

Although a lymph node biopsy is the most effective way of establishing a diagnosis, data from the Chelsea and Westminster suggest that measuring HHV-8 plasma viral load may be an important diagnostic tool.

In a study of 240 HIV-positive individuals, the vast majority (83%) of those with multicentric Castleman's disease had a detectable HHV-8 viral load at the time of active MCD whereas 65% of those with KS, and 97% of those with lymphoma from other causes had undetectable HHV-8 viral load (p = 0.0001). In addition, individuals with KS with detectable HHV-8 had far lower median viral loads compared with the individuals with multicentric Castleman's disease (3900 vs 41,000 copies/ml).

He added that other data suggested that using the HHV-8 viral load may be a helpful marker of MCD activity, since it will fall when the disease goes into remission, and rise again when the disease relapses. However, “as yet,” said Dr Bower, “this is an insufficient assay to make the diagnosis of Castleman's disease, but it certainly seems to help push towards undertaking a biopsy.”

Data published in the pre-HAART era suggested that the median survival following a diagnosis of HIV-associated multicentric Castleman's disease was 14 months. However, Dr Bower and colleagues have recently seen long-term remissions with a two-year survival rate of 95% and a five-year survival rate of 67%, following dramatic improvements in treatment that combines chemotherapy and immunotherapy. (Bower 2007)

Dr Bower concluded by reiterating that the incidence of multicentric Castleman's disease appears to be rising; and that, unlike KS, the risk of MCD is not associated with the degree of immunosuppression.

He added that plasma HHV-8 viral load may be used as a useful diagnostic marker, and that with an aggressive combination of chemotherapy, immunotherapy and splenectomy surgery, where necessary, a five-year survival rate of 67% can be achieved.

The British HIV Association (BHIVA) has recently published guidelines in HIV Medicine on the diagnosis and management of HIV-associated malignancies that includes a section on multicentric Castleman's disease. The guidelines can be downloaded from the BHIVA website.