The Incidence of HIV-Associated Multicentric Castleman's Disease

Mark Bower: Thank you very much. Castleman's disease is a rare lymphoproliferative disorder that was first described just over 50 years ago, by Benjamin Castleman, who was the pathologist at Massachusetts General Hospital.¹ In fact, the first report of Castleman's disease appeared as a case record in the New England Journal of Medicine -- you know, that bit in the middle, where they have very difficult diagnoses that you can never work out. Clearly, nobody could have guessed that it was an entirely new illness.

Professor Mark Bower

These are the images from that first-ever case, in which a woman with a mediastinal mass, in fact, had a unicentric and hyaline vascular variation of Castleman's disease. Since that time, the same disease has been called loads of different things, which makes it rather difficult to follow in the literature. But increasingly, it's now being honed down and focused on, using the term Castleman's disease, which makes things a little bit easier.

There are, in fact, four forms of Castleman's disease. There are anatomical differences. It can either be localized or multicentric. Then there are histological differences. It can either be hyaline vascular or plasmablastic. As far as people with HIV are concerned, they tend to get plasmablastic multicentric Castleman's disease, rather than any of the other variants. That's pretty much the focus of the rest of this talk.

Histologically, the only way to make the diagnosis of Castleman's disease currently is by histology. On the microscope, the lymph nodes on the spleen have these hypoplastic follicles with expanded mantle zones in between them and this -- what, as far as histopathologists are concerned, is rather attractive, apparently -- onionskin arrangement of mantle cell zones. Those cells in that onionskin world around a central vessel are, in fact, monotypic, but polyclonal, plasmablasts. That's to say that they are not all from the same origin, like a monoclonal proliferation, but they, for some reason, have light chain restriction. Now, it's not clear why they are IgM lambda light chain restricted, but it's quite helpful in establishing the diagnosis histologically.

Those very light chain restricted plasmablasts are infected with human herpesvirus-8; that's the same virus as Kaposi's sarcoma herpesvirus. They are synonymous names for the same virus. So this is the same virus that's associated with the development of Kaposi's sarcoma.

As far as the incidence information that I've got for you, this is based on the Chelsea and Westminster cohort, which includes nearly 11,000 HIV seropositive patients, followed for over 52,000 patient-years, between 1983 and 2007, prospectively.²

First of all, what about just the overall incidence? We've compared here the incidence of Castleman's disease and the incidence of Kaposi's sarcoma in the same cohort, because they are both caused by the same virus. There were over 1,000 cases of Kaposi's sarcoma in our cohort, but only 24 cases of Castleman's disease, with an incidence about 1/50 that of Kaposi's sarcoma in the whole cohort population.

We then looked at initially univariate and then multivariate analysis, but I'm not going to bore you too much with the statistics, partly because I don't understand them terribly well. But the first thing to notice is, the risk of developing Kaposi's sarcoma, of course, as you know, is much greater in men, and is much greater if you had a prior AIDS diagnosis. However, this was not the case for Castleman's disease. There was no gender preference for developing Castleman's disease. Similarly, you weren't more likely to develop Castleman's disease if you'd had a prior AIDS-defining diagnosis.

The factors that were associated with an increased risk of developing Castleman's disease included: increasing age (the older you were, the more likely you were to develop Castleman's disease); being non-Caucasian in terms of your ethnicity, so that the risk was higher particularly in black African members of our cohort; and the shorter the interval between being known to be HIV positive and the event. So in fact, this seems to be an issue that occurs relatively early in the sequence of HIV infection. It's not a late event, as you'd expect for Kaposi's sarcoma, for example. It's a relatively early event, in terms of timing, relative to your seroconversion.

Similarly, it's associated with higher, rather than lower, nadir CD4+ cell counts -- quite the opposite from most HIV-related malignancies. The final point is that HAART [highly active antiretroviral therapy] does seem to protect against the development of Castleman's disease by reducing the risk of Castleman's disease.

This [slide] is just comparing, again, the risk of Castleman's disease and Kaposi's sarcoma by nadir CD4+ cell counts. You're all very familiar with the fact that the risk of Kaposi's sarcoma rises as the CD4+ cell count falls, and that's basically the same data shown in our particular cohort. You'll notice that on the Y-axis, it's a log scale; however, quite the opposite for multicentric Castleman's disease. You can see there is no increase in the risk of developing Castleman's disease as your CD4+ cell count falls.

What we have noticed in the cohort is that the incidence of multicentric Castleman's disease appears to be rising in different calendar periods. So we've split it into three calendar time periods: the pre-HAART era from 1983, when the cohort was established, to 1996, when HAART became widely available to members of the cohort; and we then divided the post-HAART era into early HAART era, up to 2001, and the true post-HAART era of 2002 to 2007, when basically the criteria for starting HAART were pretty much similar to those currently. During those three different calendar periods you can see that the incidence of Castleman's disease has progressively risen. There may be some diagnostic ascertainment bias within that. Maybe we're more familiar with the diagnosis and more likely to make the diagnosis and search for it a bit harder. But nonetheless, there does appear to be a rising incidence of Castleman's disease. That's, again, a completely different issue within our cohort for Kaposi's sarcoma, which is progressively declining in incidence, a particularly large decline from the pre-HAART era to the HAART era, and perhaps a modest further decline between the early and the late post-HAART eras.

In complete contrast, multicentric Castleman's disease incidence appears to be rising in those three calendar periods -- a diagnosis that we're going to be making more frequently, in other words. As I mentioned, the diagnosis is very difficult to make. The clinical and pathological features of Castleman's disease overlap with a large number of other illnesses in people who are HIV positive, and the only real way of establishing the diagnosis is by undertaking a biopsy. However, increasingly there are other criteria that may help in establishing the diagnosis and pushing you towards undertaking a biopsy -- in particular, by measuring the amount of Kaposi's sarcoma herpesvirus free in the plasma. Just like an HIV viral load, you can measure the HHV-8 in the viremia.

As you can see from this study, we did quantitative PCRs [polymerase chain reactions] looking for HHV-8 viral loads in the plasma of 240 HIV-positive patients. The vast majority of patients with Castleman's disease had a positive HHV-8 viral load at the time of their active Castleman's disease, whereas the majority of other HIV patients with other diagnoses had undetectable HHV-8 in their blood.

Patients with Kaposi's sarcoma placed somewhere in the middle, with about a third of those patients having detectable HHV-8 viremia in their plasma. But if we look a little bit further at the value of the quantitative nature of this assay, you can see that patients with Kaposi's sarcoma who had a positive HHV-8 had far lower viral loads of HHV-8 than patients with multicentric Castleman's disease.

As yet, this is an insufficient assay to make the diagnosis of Castleman's disease, but it certainly seems to help push towards undertaking a biopsy. There's some further data that I haven't got or presented for you today suggesting that using the HHV-8 viral load can be a helpful marker of disease activity. It will go down when the disease goes into remission, and rises again when the disease relapses.

I just thought I would, very briefly, in the last couple of slides, tell you about the management experience that we have with Castleman's disease using a combination of rituximab [brand name: Rituxan] with etoposide and splenectomy where necessary. The overall outcome for patients in the current era with Castleman's disease suggests that a five-year survival of 67% can be achieved in this way.

My final conclusion slide is just to remind you: that it does appear that the incidence of multicentric Castleman's disease is rising; that, unlike Kaposi's sarcoma, the risk of Castleman's is not associated with the degree of immunosuppression; that the plasma HHV-8 viral load can be used as a useful diagnostic marker, and probably an ongoing tumor marker for multicentric Castleman's disease; and that with an aggressive combination of chemotherapy, immunotherapy and splenectomy surgery where necessary, a five-year survival of 67% can be achieved for these patients with HIV-associated multicentric Castleman's disease. Thank you very much. [Applause]

If you like, I'll answer some questions.

Man 1: Thank you very much. That was a great presentation on a much-neglected disease in AIDS, I think. But is there not some literature about the use of ganciclovir [brand name: Cytovene] in this disease? I certainly have had a patient in remission for a long time, to which I attributed oral ganciclovir as being the reason.

Mark Bower: You're absolutely right that there are. Unlike Kaposi's sarcoma, the HHV-8 in multicentric Castleman's is reproducing. It's replicating. So there are a number of strategies that have been attempted using antivirals as a treatment for Castleman's disease. There is a small, a very small, series on ganciclovir being an effective treatment.³ But increasingly, people are considering the use of valganciclovir [brand name: Valcyte], particularly as maintenance after achieving remission. There's a large study going on, I believe, in Seattle, which I think is somewhere in America, that is being used for maintaining remission of Castleman's disease.⁴

Man 2: I'm [a doctor] from Dudley, Birmingham. My question is: If you have Castleman's disease, is it more in favor of an AIDS diagnosis, or should it be included as an AIDS-defining illness?

Mark Bower: My personal view about what illnesses should be included, and particularly what cancers should be included as AIDS-defining illnesses, is that they should be, as the original reason for using an AIDS-defining illness, a surrogate of advanced immunosuppression. They should kind of designate that your immune system is pretty vulnerable and, of course, they should be an indication for commencing antiretrovirals if you're not already on antiretrovirals.

Castleman's disease is unassociated with the CD4+ cell count. About a third of our patients presenting with Castleman's disease have got CD4+ cell counts above the 350 level at which in the U.K. we would initiate therapy.⁵ So I don't think Castleman's disease should be an AIDS-defining diagnosis, personally.

Man 3: [I am] from Germany, Berlin. I know that in Japan antibody against interleukin-6 is approved for the treatment of Castleman's disease,⁶ and that Roche didn't develop it for Europe.⁷ Do you have any explanation for this? Or do you have any suggestion as to whether there will be a next step for this drug?

Mark Bower: The antibody is against the interleukin-6 receptor. Now, interleukin-6 levels are very high in the majority of people with Castleman's disease. Indeed, many of the symptoms of Castleman's disease may be attributed to the high levels of interleukin-6. So in theory, blocking the receptor blocks that, and that's the proposed mechanism. The drug has been developed, as you say, in Japan, and the use in Japan is exclusively on HIV-negative patients, with either unicentric or multicentric Castleman's disease. The experience of using this drug in patients with HIV is extremely limited. We have actually used it in one patient, back door, about eight years ago. In my view, fortunately, the guy got much better, and returned to his native Zambia, where he subsequently died. The drug company got very irate with us because we couldn't tell them whether he had active Castleman's at the time at which he died, and refused to ever give us any of that agent again. At that time it wasn't Roche, I have to say. But I can't even remember the name of who it was, so that's probably a good thing.

OK, I need to stop there, if that's all right. [Applause.]

This transcript has been lightly edited for clarity.

References:

1. Castleman B, Towne VW. Case records of the Massachusetts General Hospital: case no. 40231. N Engl J Med. June 10, 1954;250(23):1001-1005.
2. Krell J, Tuthill M, Campbell V, et al. The incidence of HIV-associated multicentric Castlemans disease. In: Program and abstracts of the XVII International AIDS Conference; August 3-8, 2008; Mexico City, Mexico. Abstract MOAX0102.
3. Stebbing J, Pantanowitz L, Dayyani F, Sullivan RJ, Bower M, Dezube BJ. HIV-associated multicentric Castleman's disease. Am J Hematol. June 2008;83(6):498-503.
4. US National Institutes of Health. Valganciclovir to treat HHV-8 associated multicentric Castleman's disease. ClinicalTrials.gov. Accessed August 5, 2008.
5. Gazzard BG, on behalf of the BHIVA Treatment Guidelines Writing Group. British HIV Association guidelines for the treatment of HIV-1-infected adults with antiretroviral therapy. HIV Medicine. In press.
6. Actemra -- world's first drug for Castleman's disease -- launched in Japan. Medical News Today. June 13, 2005.
7. Roche, Chugai file for EU approval of arthritis drug Actemra. AFX News Ltd. December 3, 2007.