by Thomas Kickler, MD
Professor, Division of Pathology
Johns Hopkins University School of Medicine.
Release Date: July 14, 2006
Expiration Date: July 14, 2008
Dr. Kickler has no significant financial interest or relationships to disclose.
Autoimmune thrombocytopenia is an interesting disease for hematologists. As you know, many things we take care of, such as leukemia, do not do very well, but with autoimmune thrombocytopenia, or ITP, we’ve really got something to offer. Today we’re going to discuss these questions:
- What is ITP?
- How do we diagnose ITP?
- What is the pathophysiology of ITP?
- How is ITP treated?
What is ITP?
In the late ’60s and early ’70s, we suspected this was an autoimmune, or at least an immune-based, disorder because we observed concomitant thrombocytopenia in both mother and child. Therefore, there was presumably something going across the placenta. Everyone assumed it was immunoglobulin, and they were right. In fact, some of the earliest clinical studies in any hematologic disorder (and perhaps many other types of disorders) were done in autoimmune thrombocytopenia. Investigators took plasma from patients who had thrombocytopenia, infused it into normal recipients, and provoked thrombocytopenia in the recipients. Subsequently, it was found that this factor was in the 7s fraction of plasma. For about 15 to 20 years, we were not sure whether it was an antibody or an immune complex. Platelets are very rich in Fc receptors, so they can bind things and lead to thrombocytopenia through activation mechanisms.
This is one disease that I like to call a “cognitive disorder”: We really have to use our thought processes to make the diagnosis. We do not need an angiogram; we do not need CT scans. We need a physical exam, history, and a blood smear. The treatment is primarily directed toward ameliorating and preventing the immune destruction of platelets in the reticuloendothelial system. Because we now know that there is actually suppression of thrombopoiesis in the bone marrow, newer approaches are directed at enhancing thrombopoiesis. But it can be a “surgical disease,” too; treatment often still involves a splenectomy.
Autoimmune thrombocytopenia is a relatively common disorder. Best estimates are 66 million cases a year in the United States. We are seeing more of it because everyone who gets a blood count today also gets a platelet count — which was not the case 25 years ago. This is primarily a disease of younger women: 70 percent of patients are women, and 72 percent of these women are under age 40. Also, it’s seen in association with other diseases, such as the autoimmune disorders of systemic lupus erythematosus or rheumatoid arthritis. We frequently see it as concomitant in patients with leukoproliferative disorders, and not uncommonly in patients with HIV infection. The acute form of this disease is a childhood disorder. Onset is abrupt, it usually follows an infectious illness, and it spontaneously remits in six months. The chronic form of autoimmune thrombocytopenia has duration greater than six months, and is an organ-specific autoimmune disease. Antibodies actually are directed to important platelet glycoprotein receptors, which not only leads to a decrease in platelet count but also to functional abnormalities.
How do we diagnose ITP?
When evaluating a patient with suspected autoimmune thrombocytopenia, there’s a diagnosis of exclusion. The consensus of the American Society of Hematology is that one can make a straightforward diagnosis by reviewing the history carefully, making sure there is no medication or toxin. A toxin we particularly inquire about is alcohol, as alcohol can drop the platelet counts in thrombopoiesis. A variety of medications can cause thrombocytopenia, either in an immunosuppressive or in an immune basis. Carefully perform a physical exam looking for splenomegaly. An enlarged spleen dissuades our thinking that this has an immune basis; it could be hypersplenism or some lymphoproliferative disorder. And we cannot emphasize enough the importance of looking at peripheral smear. For one thing, frequently in Hematology we are called to psychiatry or obstetrics and told the platelet count is low. But when we go down, the patient’s not bleeding, nothing is going on. We look at the blood smear and it is entirely normal; there are platelets on the smear. This is pseudothrombocytopenia due to phenamil EDTA-dependent agglutinins, leading to sparsely low platelet counts. Therefore, this is something that one always needs to be aware of. The other thing to consider with blood smears is that platelets are generally on the bigger side. Something that gets our attention rather quickly when we do a physical exam is what is call “wet purpura”.
If you see someone with this ugly-looking bleeding in gums, and ecchymosis, you really need to get moving on a treatment because this could be a warning of potential intercranial hemorrhage, which we always worry about.
What is the pathophysiology of ITP?
Plasma infusion studies (Harrington WJ, et al. J. Lab Clin Med. 1951;38:1-10) in the early 1950s examined the immune destruction of blood cells. The studies noted that when plasma was infused into non-splenectomized patients, the decrease in platelets was greater, whereas post-splenectomy, the plasma did not have as great an effect on dropping the platelet count. From this stemmed the earliest thought that autoimmune thrombocytopenia was primarily a hyperdestructive or consumptive form of thrombocytopenia due to destruction of the platelets in the spleen by monocytes phagocytizing these opsonized cells. Still, it was not entirely clear whether these were antibodies or immune complexes.
In the early 1980s, we made eluates from platelets with ITP (Van Leeuwen, et al. Blood. 1982;59(1):23-6). Serologically, we took these eluates and reacted them with normal platelets, and I said: “Well, why don’t we react them with some platelets that are abnormal?” — sort of Rh(null) red cells that pointed the way toward the Rh complex being the focus of the antibodies leading to autoimmune hemolytic anemia. So, we found patients with Glanzmann’s thrombocytopenia who are Glanzmann’s thrombasthenia and who lack glycoprotein IIb-IIIa, and the antibody eluates did not bind to those platelets.(Herman, J.H., Kickler, T.S., and Ness, P.M. Immunoblotting for the identification of platelet specific antibodies. Tissue Antigens 28(5):257 268, 1986) This was some of the earliest strong evidence that these antibodies were directed toward platelet glycoproteins, a true antigenic target.
To summarize, stimulation of the autoimmune process makes a variety of different types of antibodies to different glycoproteins: Ia/IIa, which is a collagen receptor; IbIX, which is a Von Willebrand receptor; IIb/IIIa, which is the fibrinogen receptor with glycoprotein IV receptor. All these types of antibodies can bind to platelets, leading to their coating the antibodies in activated phagocytosis by macrophages.
There are problems with this, though, in terms of consumptive process as much as phagocytosis. In fact, the removal of spleen is ineffective in about 30% to 40% of patients with autoimmune thrombocytopenia. That does not necessarily mean that it is not consumptive, but when you do a bone marrow, many patients do not have increased megakaryocytes. When we looked at labeled platelets with indium, they showed improved platelet survival after splenectomy, but the patient was still thrombocytopenic. Moreover, we know that the hallmark of the diagnosis of autoimmune hemolytic anemia is the reticulocyte — very important to measuring and sorting out immune destruction of red cells vs. underproduction. Reticulated platelets are not increased; there is something else going on. In fact, early studies (Ingram and Coopersmith. Br J Haematol. 1969;17(3):225-9) showed that platelets contain ribonucleic acid (RNA) after acute blood loss, and subsequent studies showed that increased RNA present in patients with consumptive thrombocytopenia can be measured by fluorocytometry (Ault et al. Am J Clin Pathol. 1992;98(6):637-46). This is not something we have been able to do very easily in the clinic, though, because we do not have a fluorocytometer.
Recently, I’ve become interested in the whole issue of reticulated platelets. With the cell counters that we have now, we can measure routinely, every day, when we do our CBC. We can stain platelets with dye for RNA; the software is set up arbitrarily to measure the highest 3% fraction of fluorescent intensity.
Comparing normals with a patient with ITP, you will note an elevated course in aplastic anemia. You do not see the reticulated platelets, but after recovery from chemotherapy, they reappear. We also looked at a series of different diagnostic criteria of patients (Kickler et al. Am J Clin Pathol. 2006;125:282-287).
In 37 patients with autoimmune thrombocytopenia, about 10 of them did not have elevated reticulated platelets. We went back and looked in medical records, and these, in fact, were patients who were not responding to steroids. Post-splenectomy, they still had thrombocytopenia. In other consumptive thrombocytpoenias disseminated intravascular coagulation (DIC), we also see patients in the normal range. This is going to be a useful, readily available tool to help us monitor patients’ response.
Until about 10 years ago, we did bone marrow on patients with suspected autoimmune thrombocytopenia. Now, the dogma is careful physical exam, history, a look at the blood smear, ensuring there is nothing to suggest lymphoma or monodysplasia, or hypergranular lymphocytes, or whatever — and there is no need to do a bone marrow. When we did bone marrows, and many parts of the world still do, we actually could see increased megakaryocytes (right panel) or absence of decrease of megakaryocytes (left panel).
Moreover, in in vitro cell culture studies using megakaryocytic lye, one can take plasma from a patient with autoimmune thrombocytopenia and see decreased suppression of cell proliferation.
Turning to immunopathogenesis: Antibodies show clonal restriction in light chain use, and antibodies from phage display show highly constrained Vh gene use. Adults have increased numbers of HLA-DR positive T-cells and cytokine profiles, suggesting activation of precursor helper T and type I helper cells. In one study, we looked at familial autoimmune disorders and associated thyroiditis and lymphoma. (Laster, A.J., et al. NEJM 307:1495 1498, 1982). Early on, the Immunogenetics Lab in the Department of Medicine carefully looked at HLA associations, at least in terms of class I and, I believe, class II polymorphisms. It did not find a consistent association. Others have shown that differences in response to therapy may be associated with different class II alleles. (Cines DB, Blanchette VS, Immune Thrombocytopenia. NEJM 346:995-1008, 2003.)
One thing that taught me a lesson is that you cannot always extend one type of thought process to another phenomenon; in other words, red cells and platelets are not equal. We should know that. Platelets are much more interesting — in terms of diagnosis, particularly. We’ve had a very reliable test to find increased immunoglobulin on the surface of a platelet. We thought that would be good, but there’s been a lot of debate in the literature about this. As I joked, platelets are much more interesting than red cells. They have an open funicular system that allows things from the plasma to get inside of them. In fact, immunoglobulin traverses platelets rather gleefully in getting to different granules. Of course, to do an anti-globulin test, you have to wash the platelets and centrifuge the dickens out of them. What happens is that you activate the platelets. The granules fuse immunoglobulin that was in them to appear on the surface. When you do a global Coobs test for platelets, you are measuring everything. People were measuring 190,000 molecules of antibody on a platelet. The other problem is the integrin molecule IIb/IIIa. These antibodies in autoimmune thrombocytopenia are primarily proactive against the disulfate, which bonds there. Also, the antibodies that are circulating the plasma are, for some reason, actually directed to the internal part of the molecule that does the outside-inside signaling for platelets. So, not only do we have non-specific binding of immunoglobulin platelets, but if you look in the plasma, you will find antibodies to epitopes that are not really revealed or exposed.
When should antibody testing of the platelets be done?
We do it when we are not quite sure of the disease, not quite sure what is going on. Patients with chronic lymphatic leukemia could have decreased production, or they might have an antibody. Therefore, there are complicated disorders or major antibodies. These antibodies do lead to functional abnormalities; they impair fibrinogen binding, Von Willebrand adhesion.
The bottom line clinically from these observations is that we might see some patients with a platelet count of 50,000 who have a normal bleeding time and not much problem. However, we see other patients who have platelet counts of 50,000 who really have a lot of bleeding. Part of the difference in the disease presentation is due to these qualitative abnormalities, these antibodies.
How is ITP Treated?
As for therapy, clearly corticosteroids and splenectomy are the mainstays. Other approaches include psychlophosphamide, intravenous gamma globulin (IV-IgG), anti-Rh(D), anti-CD40L, anti-CD20, cyclosporine, vitamin C, pectin, and monoclonal Fc blockade. Believe it or not, we still see patients coming in on high-dose Vitamin C. I really cannot say that it helps, and it is certainly nothing that I would recommend. Our goal with therapy is to achieve a platelet count of at least 50,000. But consider this: Many people do quite well with platelet counts of 20,000 to 30,000. Therefore, our goal is to ameliorate the petechiae and nosebleeds (epistaxis) and not necessarily get one’s hemostatic platelet count up to 100,000. Clearly, if you are going to send them for a brain biopsy or open-heart surgery, we need to do something a little more vigorous. However, in large part, our considerations for treatment of ITP center around the discussion with the patient, what the person is comfortable with. One algorithm, which was from the New England Journal of Medicine, focuses on patients with platelet counts of less than 20,000. (Cines DB, Blanchette VS, Immune Thrombocytopenia. NEJM 346:995-1008, 2003).
We start them on prednisone and generally follow them for three to six months. If the platelet count goes up to the 50,000 to 75,000 range, we start tapering down, seeing where we are. I am not reluctant to send someone these days for a splenectomy; it is done laparoscopically and, moreover, you can actually get the platelet counts up prior to splenectomy, by giving them intravenous hemoglobulin.
One point I want to emphasize is that, when you have someone who has a headache, who might be bleeding into their head, it really is a medical emergency. Again, therapy has improved a good bit in terms of steroids and intravenous hemoglobulin. I never hesitate to give platelet transfusions to someone in this situation. You might say, “Well, that particular platelet count does not go up” — but we know platelets provide phospholipids in the surface for the prothrombinase reaction to go on. Although the platelets may not be circulating and adhering, they are membranes upon which the clotting cascade is exerting and is able to generate some thrombin. The other amazing therapy we have for any form of thrombocytopenia when patients do not respond to platelet transfusion in oncology is recombinant factor VIIa. It really just makes a clot in the absence of platelets. I was totally incredulous because I always believed that we needed platelets to form a clot. But recombinant factor VIIa activates clotting cascade through the extrinsic system. By activating VIIa, you get a burst of thrombin generation that has a lot of other uses in hemophilia and in hemorrhagic stroke. Nevertheless, careful therapy is critical in these emergencies.
Patients who do not respond to prednisone and do not want a splenectomy sometimes end up getting IV-IgG monthly. Intravenous gamma globulin has many effects on the immune system other than decreasing immune destruction by RES system. One agent that has been popularized is intravenous anti-D. If you are Rh-positive and you have an intact spleen, administration of anti-D of 50-75 mg/kg can raise the platelet count within 48 hours. Pediatricians are using this approach more than adult hematologists. I have never been enthusiastic about promoting hemolysis that drops someone’s hematocrit, say, from 30 to 22, so we then have to give more red cells. Moreover, in the last year, there has been intravascular hemolysis in renal insufficiency RHD. Nonetheless, many subscribe to it. If this fails, do a splenectomy — remembering to immunize the patient for encapsulated organisms.
In general, the long-term results of splenectomy — long-term being defined as at least three to five years — show a platelet count of about 50,000 in about 50% of patients. We see an awful lot of patients referred for refractory autoimmune thrombocytopenia. They have been splenectomized already. Sometimes these cases are frustrating because most of the drugs already have been tried at other institutions. With these patients, we are using a fair amount of Rituximab®. There are no large controlled studies, though. Frequently, we just go back to prednisone and IV-IgG.
Oral cyclophosphamide, in doses not enough to promote neutropenia, also is being used, and often more widely than cyclosporine. I do not know about stem cell transplantation; however, an interesting area we hear more about is the experimental thrombopoietic factors.
There are other pathways you may not have considered.
- Several small studies (Emilia G, et al. Blood. 2001;97:812-814) showed that, after eradicating H. pylori infection with antibiotics, patients with autoimmune thrombocytopenia responded to their therapy.
- Particularly in older patients, and in men who present with thrombocytopenia, we need to make sure they do not have a lymphoproliferative disorder, by looking at the blood smear or just doing fluorocytometry on the lymphocytes. The fact that somebody has autoimmune thrombocytopenia with a lymphoproliferative disorder is not necessarily a bad sign or prognosis.
- People with autoimmune thrombocytopenia are often under 40 and female, so a frequent call is to Obstetrics. Most women will tolerate platelet counts of 20,000 to 30,000 during an entire pregnancy; at delivery, though, we do get the platelet count up to around 50,000. Of course, the anesthesiologist always wants a platelet count of 100,000 for an epidural. We do not always achieve that, but we usually achieve an increased elevation of platelets due to IV-IgG and anti-D. Anti-D does not appear to affect the baby. And because we are always concerned if the baby is Rh-positive, we might do hemolysis. Sometimes they are born with a positive Coombs’ test, which reiterates the rather dramatic increase in platelet counts. I try to avoid prednisone because of diabetes and fluid issues in pregnancy.
I have defined autoimmune thrombocytopenia, and discussed some of the diagnostic issues, emphasizing the use of cognitive skills. I emphasized that antibodies do lead to platelet destruction, they can lead to platelet function abnormalities, and we have to deal with impaired thrombopoiesis. I also gave you an approach to treatment emphasizing that we do not bring platelet counts up to 200,000, but we are happy with 50,000.