CONTENTS:
The Spleen in Immune Thrombocytopenia
By Julia T Geyer MD1 and James B Bussel MD2
1Department of Pathology and Laboratory Medicine, Weill Cornell Medicine, New York, NY, USA
2Department of Pediatrics, Division of Hematology/Oncology, Weill Cornell Medicine, New York, NY, USA
The spleen in ITP is as central an organ in the pathogenesis and in the understanding of ITP as is the bone marrow. Studies from many years ago demonstrated that not only is the spleen the major source of destroying platelets but also that it is a very important site where antiplatelet antibodies are made. Since that time, there have not been major advancements in understanding the role of the spleen in the pathophysiology of ITP. In this review we will consider serologic and anatomic reasons that the spleen plays a role in ITP and proceed with an analysis of response to splenectomy and prediction of response to splenectomy. Under serologic features, we will consider platelet antibodies and the contributions of complement. Under prediction we will look at platelet lifespan, sequestration studies and exploration of the predictive role of previous response or non-response to other treatments. In exposition of the anatomy, we will consider both the anatomy of blood flow through the spleen as well as its relationship to the creation and/or especially perpetuation of the immune response of the spleen including the role of antiplatelet antibodies synthesized by the spleen. Finally, we will try to analyze carefully the role and effects of splenectomy in ITP.
Primary immune thrombocytopenia is an autoimmune disorder in which platelet destruction is a consequence of both B- and T-cell dysregulation. The spleen is a major site of platelet destruction and of production of autoantibodies in ITP. Autoantibodies not only accelerate platelet destruction but also impair platelet production by megakaryocytes in bone marrow. Autoreactive antibodies bind to platelets, which are then sequestered by splenic macrophages via FcγReceptors and cleared. Antibody production in ITP appears to be driven by CD4-positive helper T cells reacting to platelet surface glycoproteins, likely involving CD40:CD40L co-stimulation1. These cells appear to correspond to T follicular helper cells (one subtype of CD4-positive T cells), which are reported to be expanded within the spleens of ITP patients2. IgG and/or IgM reactive to the glycoprotein complexes of the platelets have been identified on the platelets or in the plasma of many ITP patients.
It was originally thought that the mechanism of action of splenectomy was simply removal of the site of platelet destruction. However, antibody-coated platelets and cytotoxic T cells can persist or wax and wane after splenectomy. One study analyzed the distribution and phenotypic characteristics of B-cell subsets in non-splenectomized and splenectomized patients with ITP and demonstrated decreased frequencies of memory B cells in the splenectomized individuals with a decline of CD27+IgD+ and CD27+IgD− and CD27−/IgD− cells 3. Another study reported that the frequencies of circulating GPIIb/IIIa-reactive T and B cells were significantly decreased after splenectomy in patients with a complete response but were unchanged in non-responders, suggesting that GPIIb/IIIa-reactive T- and B-cell interactions that induce anti-platelet antibody production in patients with ITP occur primarily in the spleen4. In ITP, the spleen is usually mildly enlarged (weight <200g). The morphology of the spleen in untreated ITP shows evidence of active antibody production with well-developed germinal centers. The red pulp shows histologic evidence of antibody-coated platelets within cordal macrophages. The number of splenic macrophages is increased and they have abundant foamy cytoplasm. Splenic macrophages appear to be the major antigen-presenting cells5. By virtue of the tortuous blood vessels in the spleen the flow is slower and facilitates phagocytosis by adjoining macrophages. In turn, the phagocytosis in the spleen leads to antibody production. While normally these antibodies might be anti-bacterial, in ITP they are directed against platelet glycoproteins.
The clinical role of splenectomy starts back in 1916 when a medical student named Kaznelson in Czechoslovakia “convinced” a surgeon to remove the spleen in a woman with ITP6.
Things that seem as if they could be apocryphal are:
1) why did Kaznelson think at the time that the spleen should be removed at all;
2) that a surgeon at that time now a little more than 100 years ago would actually listen to a medical student, and
3) why a surgeon would have to be “talked into” doing an operation.
This splenectomy initiated the idea that splenectomy might be useful in patients who had at that time what was considered “idiopathic” thrombocytopenia even though its effects were not always positive in subsequent cases. Subsequently splenectomy became a regular feature of management of ITP in part because of the lack of “simpler” or “effective” options. As of 1980, splenectomy was one of the two most important treatments of ITP and the dominant second line therapy. It is worth noting that the only medical therapy in any regular use now that was available then was prednisone.
More recently, for a number of reasons not all of which are rational, the use of splenectomy declined substantially over the last 20 or more years particularly in the United States, Western Europe, and China. Reasons behind this reduced use of splenectomy are varied and not completely certain but the most obvious is the relatively large number of other available treatments. n particular, these include the thrombopoietic agents, rituximab, and immunosuppressive agents like azathioprin. A second reason is that it became appreciated that patients, even adults, could get better over time and that this is not rare. This lead to the going recommendation in adults (not really based on data) to wait for one year from onset of ITP before performing splenectomy. Third, the response to splenectomy is considered to be approximately 80% of all patients, with one quarter of those (or 20% of the original group) relapsing almost always within 2 years. Given that there are now multiple options, patients are increasingly reluctant to undergo splenectomy when they know that 40% of the time it may not be curative. I would find it hard personally to want to undergo surgery with the idea that it might not be successful a significant percentage of the time (even if it would be successful most of the time).
This uncertainty in response drives the concept of trying to predict the success of splenectomy. The only well-established test, given that there is debate in published reports, is splenic sequestration testing using autologous indium-labeled platelets. It is important that it be performed by people who are experienced and can perform the testing without damaging the platelets so that they are cleared according to the patient’s disease not because they are damaged in the handling. This test is still debated as to its utility. A recent preliminary set of data suggested that glycosylation of platelets would matter but this requires prospective study in larger numbers of patients. Another predictor that has been heavily considered in the past has been previous response to therapies but these studies have all been equivocal when looked at as a group i.e. response to IVIG, response to steroids and even response to IV anti-D. In the absence of data, there is some consensus that in the patient in whom might most want to perform splenectomy, i.e. the patient who responds to nothing else, is probably not very likely to respond to splenectomy but this has never been confirmed. It is also important to note that there is no diagnostic test that unequivocally says “this is ITP”. Therefore some cases of apparent refractory ITP may in reality not be ITP at all; this is most frequently either inherited thrombocytopenia, a marrow failure state or MDS.
Another issue with response to splenectomy that could be looked at either negatively or positively, is secondary ITP. There are a number of causes of secondary ITP, especially infectious ones, in which if an ongoing viral infection could be identified and then treated, the patient could improve substantially and therefore might not require splenectomy. In addition, it is possible that the patient might not have responded to splenectomy if they had an ongoing viral infection. The only substantial data is in HIV in which patients did respond to splenectomy but this required some level of viral control. It has been our recent practice to send patients (if they can afford the trip) who are considering splenectomy to St. Bart’s in London to undergo a splenic sequestration scan to determine whether they undergo splenectomy or not.
Studies of mouse autoimmune hemolytic anemia by Alan Schreiber explored the role of complement. These studies demonstrated that mice who had IgG alone on their red cells would have them cleared in the spleen whereas those with red cells attacked by IgM who then had complement bound to their red cells (the IgM would have dissociated from the red cell) would not respond to splenectomy because their cells would be largely destroyed in the liver. Whether this translates to platelets and ITP in humans remains to be proven however the role of complement in hematologic diseases, including ITP, is assuming a more important role.
Another key issue is response to splenectomy overall. The data that suggests a 60% long-term response rate is mostly taken from the time when most splenectomies were performed in the first 3 months of ITP and usually within the first 6 months of disease. If one makes the assumption that there are people who get spontaneously better after that period of time (that is known to be true although the number remains inexact), that is one reason why the current approach of waiting longer prior to proceeding to splenectomy may lessen the overall response rate. There is also speculation that auto-reactive B cells and plasma cells making antibody will originally develop in the spleen for the reasons explained. However these cells may over time migrate out of the spleen e.g. to the bone marrow. That remains a hypothesis but would be another reason why splenectomy later would not be as good as splenectomy earlier.
What are the adverse effects of splenectomy? The single most agreed upon unequivocal issue is the small but potentially devastating risk of overwhelming post splenectomy sepsis. Minimization of this involves vaccination to the most important organisms e.g. pneumococci, liberal use of oral antibiotics (prophylaxis of adultis variably used), and inviolate rule of getting immediate urgent parenteral antibiotics if there are fever or chills. A second highly concerning adverse event is thrombosis. Longterm studies using national databases, especially from Denmark, have shown an increased risk of stroke. Other thromboembolic events occur with venous ones more substantial than arterial ones. The third type of adverse events involve the surgical complications of splenectomy. These include infections, bleeding and thromboembolic events. These are usually minor; there is a debate as to whether anticoagulation should be standard in part because of the very high rate of splenic vein thrombosis after it is tied off. More recent studies have attempted to discover other issues such as pulmonary hypertension as has been reported in hemolytic disease but this has not been described in ITP.
Overall, splenectomy remains an important option in ITP. Prediction of anticipated success (or of vulnerability to specific complications) would greatly advance its use. Also if the likelihood of spontaneous improvement could be anticipated, this would be very helpful as well. Similarly better ability to predict response to other treatments would similarly help to determine which patients might be best served by splenectomy. The first stage might be whole exome sequencing but this is unlikely alone to provide all the future information that would optimize care. Preliminary medical decision analyses suggest that splenectomy is a reasonable option for management of chronic ITP with perhaps lower cost but higher mortality than other options.
References:
- 1 Patel VL, Schwartz J, Bussel JB. The effect of anti-CD40 ligand in immune thrombocytopenic purpura. British journal of haematology 2008;141:545-8.
- 2 Audia S, Rossato M, Santegoets K, et al. Splenic TFH expansion participates in B-cell differentiation and antiplatelet-antibody production during immune thrombocytopenia. Blood 2014;124:2858-66.
- 3 Martinez-Gamboa L, Mei H, Loddenkemper C, et al. Role of the spleen in peripheral memory B-cell homeostasis in patients with autoimmune thrombocytopenia purpura. Clinical immunology (Orlando, Fla) 2009;130:199-212.
- 4 Kuwana M, Okazaki Y, Kaburaki J, Kawakami Y, Ikeda Y. Spleen is a primary site for activation of platelet-reactive T and B cells in patients with immune thrombocytopenic purpura. Journal of immunology (Baltimore, Md : 1950) 2002;168:3675-82.
- 5 Kuwana M, Okazaki Y, Ikeda Y. Splenic macrophages maintain the anti-platelet autoimmune response via uptake of opsonized platelets in patients with immune thrombocytopenic purpura. Journal of thrombosis and haemostasis : JTH 2009;7:322-9.
- 6 Stasi R, Newland AC. ITP: a historical perspective. British journal of haematology 2011;153:437-50.