Friday, January 30, 2009

Should I have a transplant 2?

Here is the attraction of stem cell transplants: they cure people. Here is the problem with stem cell transplants in CLL: they kill people. Full blown transplants for any condition are of dubious value; they work best for people under the age of 20. Hardly anybody has CLL under the age of 20 – my youngest patient was 21. The older you get the worse the problem – the total body irradiation and the high dose chemotherapy are so toxic that even the best centers in the world struggle to keep anyone over the age of 50 alive after that insult, and even 50 is stretching it. The one disease where standard transplants were absolutely the treatment of choice was chronic myeloid leukemia (CML) – until imatinib (Glivec or Gleevec) came along. When Glivec became available for CML hardly anyone continued to transplant patients over 30, even though it seemed that 10% of patients had relapsed after Glivec after the first year. Nowadays CML doctors like to confine transplants to patients under 20.

Since half of patients with CLL are aged 70 or over when they present, you can see that there was very little call for transplants in CLL. There are some younger patients of course, but the toxicity of the treatment in these was still very great. Treatment-related-mortality (defined as deaths in the first 100 days) was 40% even in the best centers.

Then along came reduced-intensity-conditioning (RIC). The idea works like this: if you have an autograft – that is your own bone marrow stem cells are taken out and frozen down and this is followed by strong chemo- or radiotherapy, after which the stem cells are returned – you get the same dose of chemo- or radiotherapy as when you are having an allograft (where the stem cells come from somebody else. Theoretically this should be better (I certainly thought so in the 1980s) since there is no bother with the immunological consequences of someone else’s bone marrow growing inside you. However, when you look at the long term consequences, for most diseases patients having allografts survive for longer than those having autografts.

There must be something special about an allograft that an autograft hasn’t got. And what could it be except that those very immunological effects that we were worried about? These immunological effects weren’t rejection (which is the problem with kidney transplants or heart transplants), but quite the reverse. The graft tries to reject the patient! This graft-versus-host disease (GVHD) can be very unpleasant. Mild acute forms just have a bit of a rash, particularly on the palms and soles, which fades after a while, but severe forms have a rash that scales and won’t heal, diarrhea that won’t stop and severe liver damage. It also damages the immune system making rotten infections more likely. With chronic GVHD the skin becomes fixed to the underlying tissue so that its like walking around in a suit of scar tissue, with loss of hair and teeth. I’ve seen an 18-year old looking like an old man of 80. I have also had patients kill themselves because their condition was intolerable. But within the GVHD there is also a graft-versus-leukaemia (GVL) effect and it is this effect that can really cure disease that doesn’t respond to anything else.

In fact it’s not the chemotherapy that cures you with an allograft, but the GVL effect. So why have the chemotherapy? Rainer Storb at the Hutch in Seattle was the person who realized this and put it into practice. Using immunosuppressive drugs rather than marrow ablative drugs an allograft can successfully be transplanted, and because the damaging chemo- or radiotherapy is not used you can give it to older patients. But you still have the GVHD problem to contend with. Usually this is done with immunosuppressive drugs after the transplant like ciclosporin and methotrexate, but one thing that has been tried for a long time is T depletion. This simply means removing the T cells from the graft. It is a very effective way of getting rid of GVHD, though if too many are removed the patient rejects the graft. The problem is that getting rid of GVHD often means getting rid of GVL, so although such grafts are safer, they are usually less effective and patients start relapsing. Theoretically, Campath ought to be a more efficient way of T-depleting because not only does it get rid of the T cells, it is also an anti-CLL drug and will get rid of any residual disease. As we saw yesterday, Campath does exactly this. Much less GVHD, but more relapses than not using it. Nonetheless, still pretty good long term survivals.

As it happens, another paper was published in the Journal of Clinical Oncology in October last year from Seattle. 82 patients aged between 42 and 72 who had CLL no longer responding to fludarabine were treated with a RIC allograft between 1997 and 2006. The conditioning regime was low dose total body irradiation (200 centiGray) with or without fludarabine. 78 patients still had measurable disease when transplanted; only four has no detectable disease. So what happened to them? Let’s take the four already in remission: one died of relapsed leukemia, two died of treatment complications while in complete remission and on remains in complete remission. But no-one should read into this that if you are in CR and have a RIC transplant you have a 75% chance of dying. The numbers are just too small to know. The next 50 patients might all survive.

What about the other 78? 55% got a complete remission and 15% a partial remission. Remember these were patients who did not respond to fludarabine, so getting any sort of remission is remarkable. There is a sort of randomization going on here, between those who have a matched sibling donor and those who don’t. Interestingly, the chance of getting a complete remission was significantly greater in those who didn’t. This suggests that a matched unrelated transplant has greater immunological power at killing the CLL than a sibling transplant.

Of the 41 patients who achieved a complete remission, 30 are still alive in CR, 8 died of the complications of the transplant and 3 relapsed (one of these still alive the other two now dead). Of the 13 who achieved PR, 4 are still alive and in PR, 3 died of the complications of the transplant and six developed progressive disease (4 of these dead, 2 still alive in PR after treatment with Campath or Rituximab). In addition there were four who died of transplant complications before their response could be assessed, three who had stable disease, two of whom died of transplant complications and one who remains alive with stable disease. 17 patients had disease progression despite the transplant; 14 of these have died of progressive disease and three are alive after antibody treatment.

To summarize all this: 23% died of the complications of transplant, 38% progressed, and the actuarial survival at 5 years was 50%, while progression-free survival was 39%.

There were no significant differences for any of these 4 measurements between those who had a sibling and those who had a matched unrelated transplant, nor between those who had their transplants at Seattle compared to those who had theirs with the same protocol at a sister organization. However, the numbers are relatively small and ‘no significant difference’ doesn’t mean ‘no difference’; it just means there were not enough numbers to demonstrate a difference. In any case even if there were differences, unless the trial was designed to look for differences, any that were found would be examples of data dredging. Data dredgers always find differences one in 20 dredges because ‘significance’ is defined as p<0/05.

Grade 3 and 4 GVHD was found in 16% of those having a sibling transplant and 23% of those with matched unrelated transplants. However, bad chronic GVHD was worse, occurring in 49% of related and 53% of unrelated transplants. The transplant-related causes of death were as follows: acute GVHD 5, chronic GVHD 6, bleeding from the lungs 1, stroke 1, cardiac arrest 2, lung cancer 1, sepsis 2, multi-organ failure after cardiac surgery 1.

A lot of factors did not affect the outcome including age, donor type, CD34+ cell dose, CD3+ cell dose, CD38 positivity, cytogenetics, splenomegaly, or time between diagnosis and transplant. The following did influence outcome: lymph nodes more than 5 cm in diameter and co-morbidities. In a univariate analysis heavier marrow infiltration, higher beta-2 microglobulin levels and higher white counts added to the prognosis but these were subsidiary to the large lymph nodes in a multivariate analysis.

What we learn from today’s posting is that RIC transplants extend the age range where transplants become a possibility, and the toxicity of the chemo-radiotherapy is removed, but people still die and/or suffer from having a transplant. The major reason for this is GVHD.

What we learned from yesterday’s posting is that we can reduce GVHD by removing the T cells with Campath. But when we reduce the graft versus host disease we also reduce the effectiveness of the transplant.

As I said at the beginning; the attraction of a transplant is that it cures. The problem with a transplant is that it kills.


Brian Koffman said...

Dr. Hamblin,
Thank you for your reviews, again
I am so confused.
You said what the data show: "removing the T cells from the graft. It is a very effective way of getting rid of GVHD, though if too many are removed the patient rejects the graft."
Wouldn't ATG or Campath reduce the likelihood of rejecting the graft by depleting host T cells that would lead the rejection?
I thought I might have been less likely to have rejected my graft (I had FCR conditioning) if I had received ATG for my MUD RIC HSCT which is what they do at MDACC and other centers in patients who at getting MUD and/or have not had recent Fludarabine.
Thanks for your help.

Terry Hamblin said...

As I understand it some T cells must be present to avoid graft failure. I'm not sure that I understand the mechanism for this.

I will have to do some more reading around the subject.

Anonymous said...

Dr. Hamblin,

Your excellent summary only serves to confirm what we already intuitively know...much more data needs to be acquired to dissect out those parameters which may influence the likelihood of success and minimize that of GVHD.

I was wondering if you might care to comment on your opinion of the future of related techniques of inducing "GVL" effect while avoiding or minimizing GVHD, such as so called "NK cell transplants" and other studies utilizing techniques to "awaken" the patients own T-cells, causing them to recognize the CLL cells within their midst as "enemies of the body".



Terry Hamblin said...

I think there is a lot more unknown than known about these effects. I saw yesterday a reference to minor compatability antigens on CD19 which might explain why MUDs are more effective than sibling transplants for B cells tumors. I'lll keep educationg myself.

Wayne said...

Dr. Hamblin,

Your overview of transplants may be quite useful for me to think about given my disease progression.

I was wondering if the data reflects the quality of HLA matching and how that might affect GVL, GVHD or relapse?


Anonymous said...

And this is the best chance we have for a 'cure'? How barbaric. How tragic of a decision.

The more Dr. Hamblin writes of transplants, the less interested I become.

What we need is the anti-leukemia effect without a transplant. We know the immune system is the key to any cancer cure. The fact that a few transplants work, proves the case.

bob larkin said...

I think there is another category between "cure" or "kill", called "change". I was no longer responding to treatment (including Campath) just prior to transplant in April '05. Even though I relapsed a year after the transplant, the good news is I just finished a nice year-plus response from subQ campath. It seems my CLL changed or the partial chimesrism left over from my donor brother helped me respond to campath, I'm not sure which.

Terry Hamblin said...

Matching is being refined, but a full 10/10 sibling match is very close. 10/10 volunteer matchimg is not quite so good. A slightly worse match increases the risk of GVHD but diminishes the risk of relapse.
Clearly the extremes don't inculde every patient and there are some in between.