Partial splenic embolization preceding ... - Wiley Online Library

0 downloads 0 Views 147KB Size Report
Feb 13, 2008 - case of refractory immune thrombocytopenic purpura. To the Editor: Immune thrombocytopenic purpura (ITP) is an autoimmune disorder ...
Brain natriuretic peptide level as marker of cardiac function in imatinib—treated chronic myeloid leukemia patients: no evidence of cardiotoxicity of imatinib therapy To the Editor: In the last year, the issue of cardiotoxicity of imatinib mesylate (IM) was on focus. After the original work from Kerkela¨ et al. [1], various papers reported a low incidence of severe cardiac adverse events in large groups of patients treated with imatinib for chronic myeloid leukemia (CML) [2–5] and gastrointestinal stromal tumor (GIST) [6]. Safety issues are particularly important, as imatinib is the frontline treatment of CML [7], and patients are expected to receive treatment for indefinite time.

TABLE I. Clinical Characteristics of 49 CML Patients Tested for BNP Levels While on Treatment with Imatinib Mesylate, at the Time of Analysis Male/Female Median age, years (range) Concomitant cardiovascular risk factors Hypertension Hyperlipemia Diabetes Imatinib daily dose (mg): Mean (SD) Median (range) 400 mg Imatinib duration (months) Median (range)

34/15 59 (24–82) 17 (35%) 3 (6%) 2 (4%) 406 (121) 400 (200–800) 12 32 5 39 (2–81)

B-type natriuretic peptide (BNP) is released by the heart in response to myocardial tension [8] and is considered an accurate test for the diagnosis of heart failure. The measurement of BNP in the serum is a rapid and easy tool for evaluation of left ventricular ejective function (LFEV), also in asymptomatic patients [9]. We have measured BNP level in 49 consecutive patients with chronic phase CML during imatinib therapy and in five newly diagnosed CML patients that were tested before IM start and after 1, 2, and 3 months of therapy. BNP level was measured using a direct chemiluminescent sandwich immunoassay: the analytical range extends from 0 to 5,000 pg/ml, with a sensitivity 22 pg/ml (mean 37.2 ± 31.4), compared to only 1/25 (4%) in the cohort 65%). In the five consecutive patients (Table II) tested for BNP before start of IM and monthly during treatment there was no significant change in BNP levels during imatinib therapy (Fig. 1). Mean BNP value before start of IM was 20.9 ± 10.6, after three months of therapy was 14.8 ± 13.8. With the limits of a relatively small study population, our data indicates that imatinib therapy does not cause an increase in BNP, a rapid and reliable marker for the assessment of ventricular function. The role of BNP in monitoring heart function of CML patients treated with this drug deserves confirmation in larger groups of patients.

TABLE II. Characteristics of the Five CML Patients Tested for BNP Levels Before Start of Imatinib and During Treatment No. 1 2 3 4 5

Sex

Age

Hypertension

Hyperlipemia

Diabetes

History

F M M F M

23 63 50 60 33

No Yes No No No

No No No No No

No No No No No

No No No No No

MARIO TIRIBELLI1 ANTONIO COLATUTTO2 LUCIANA MARIN1 GIUSEPPE BARBINA2 UGO QUALIZZA2 DANIELA DAMIANI1 ELEONORA TOFFOLETTI1 MARTA MEDEOT1 ANNA CANDONI1 ELIO TONUTTI2 PIERGUIDO SALA2 RENATO FANIN1

History: history of cardiac disease.

1 Division of Hematology and Bone Marrow Transplantation, Department of Medical and Morphological Researches, Udine, Italy 2 Department of Laboratory Diagnostics, ‘‘Azienda OspedalieroUniversitaria’’, Udine, Italy Published online 29 January 2008 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.100/ajh.21157

References

Figure 1. Values of BNP in five imatinib-treated CML patients at baseline (before start of imatinib) and after 1, 2, and 3 months of therapy. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

1. Kerkela¨ R, Grazette L, Yacobi R, et al. Cardiotoxicity of the cancer therapeutic agent imatinib mesylate. Nat Med 2006;12:908–916. 2. Hatfield A, Owen S, Pilot PR. In reply to ‘‘Cardiotoxicity of the cancer therapeutic agent imatinib mesylate.’’ Nat Med 2007;13:13. 3. Gambacorti C, Tornaghi L, Franceschino A, et al. In reply to Cardiotoxicity of the cancer therapeutic agent imatinib mesylate. Nat Med 2007;13:13–14. 4. Rosti G, Martinelli G, Baccarani M. In reply to ‘Cardiotoxicity of the cancer therapeutic agent imatinib mesylate’. Nat Med 2007;13:15.

C 2008 Wiley-Liss, Inc. V

American Journal of Hematology

517

http://www3.interscience.wiley.com/cgi-bin/jhome/35105

5. Atallah E, Durand JB, Kantarjian H, Cortes J. Congestive heart failure is a rare event in patients receiving imatinib therapy. Blood 2007;110;1233–1237. 6. Verweij J, Casali PG, Kotasek D, et al. Imatinib does not induce cardiac left ventricular failure in gastrointestinal stromal tumours patients: Analysis of EORTC-ISG-AGITG study 62005. Eur J Cancer 2007;43:974–978. 7. Baccarani M, Saglio G, Goldman J, et al. Evolving concepts in the management of chronic myeloid leukemia: Recommendations from an expert panel on behalf of the European LeukemiaNet. Blood 2006;108:1809–1820. 8. Liang F, Wu J, Garami M, Gardner D. Mechanical strain increase expression of the brain natriuretic peptide gene in rat cardiac myocytes. J Biol Chem 1997;272:28050–28056. 9. Wang T, Larson M, Levy D, et al. Plasma natriuretic peptide levels and the risk of cardiovascular events and death. N Engl J Med 2004;350:655–663. 10.Redfield M, Rodeheffer R, Jacobsen S, et al. Plasma brain natriuretic peptide concentration: Impact of age and gender. J Am Coll Cardiol 2002;40: 976–982.

Mutations are no substitutes for clotting, chromogenic and immunological assays in patients with congenital prothrombin deficiency

References 1. Kling SJ, Jones KA, Rodgers GM. A second case of prothrombin Puerto Rico I in the United States. Am J Haematol 2007;82:661–662. 2. Lefkowitz JB, Weller A, Nuss R, et al. A common mutation, Arg 457 Gln links prothrombin deficiencies in the Puerto Rican population. J Thromb Haemost 2003;1:2381–2388. 3. Denson KW, Barrett R, Biggs R. The specific assay of prothrombin using the Taifen snake venom. Br J Haematol 1971;21:219–223. 4. Girolami A, Patrassi G, Virgolini L, Zucchetto M. The effect of several viper venoms on prothrombin. Padua Blut 1975;31:155–160. 5. Girolami A, Patrassi G, Toffanin F, Saggin L. Chromogenic substrate (e-2238) prothrombin assay in prothrombin deficiencies and abnormalities. Am J Clin Pathol 1980;74:83–87. 6. Shapiro SS, Maldonado NI, Fradere J, McLord S. Prothrombin San Juan: A complex mew dysprothrombinemia. J Clin Invest 1974;13:73a. 7. Shapiro SS. Prothrombin San Juan: A complex new dysprothrombinemia. In: Hemker HC, Veltkamb JJ, editors. Prothrombin. Leiden University Press, Leiden, The Netherlands; 1975. p 205. 8. Girolami A, Scarano L, Saggiorato G, et al. Congenital deficiencies and abnormalities of prothrombin. Blood Coagul Fibrin 1998;9:557–569.

Response to Girolami et al. We have read with interest the paper by Kling et al. on a ‘‘Second case of prothrombin Puerto Rico in the United States’’ [1]. The observation is surely important since it represents the confirmation of previous data [2]. However, we were surprised to see that no prothrombin activity and antigen assays are included in the article. These are the only tests which are capable of differentiate between different forms of the defect. Clotting assays using different viper venom are also very useful [3,4]. The same has been demonstrated to be true for chromogenic assays [5]. Mutations are unable to do so, since, so far, no specific mutations have been identified as responsible for ‘‘true’’ (type I) deficiency or for ‘‘dys’’ forms (type II). These observations are even more important if one takes into account that allocation of Prothrombin Puerto Rico I to the type I or type II defects is still unclear [2]. Among the cases presented by Lefkowitz et al. [2], five patients belonging to three families were compound heterozygotes between the common Arg 457 Gln mutation in exon 12 together with another molecular alterations. The other mutations were: exon 5/6 splice junction for family 1; 5 base pair deletion in exon 13 for family 2; and Glu 16 Gln mutation for family 3. Only two probands in family 4 and 5 were homozygotes for the Arg 457 Gln mutation. The two homozygote patients described by Lefkowitz et al. [2] show activity levels varying from 14.8% to 18.9%, whereas the antigen levels varies between 27.6% and 39.5% of normal. The pattern is, therefore, not fully uniform and additional information would be welcome. The two cases (brother and sister) presented by Shapiro et al. and known as prothrombin San Juan [6,7] show the activity level of 15–20% of normal whereas antigen levels varied between 64 and 94% of normal. These patients were interpreted to be cases of prothrombin abnormality [6]. By comparing the results presented by Lefkowitz et al. [2] and those presented by Shapiro et al. [6,7], it is clear that prothrombin activity level was practically identical whereas antigen levels in Shapiro’s cases were slightly higher. The significance of this discrepancy remains undetermined. Unfortunately, no molecular analysis is available for Shapiro’s patients. However, because of the common geographical area of origin and the similar laboratory features, it is likely that the same mutation (Arg 457 Gln) was present. It would have been very interesting to know the activity and antigen levels of this new proband and of her available family members. It is a pity this was not the case since such an interesting and rare congenital defect deserves always a complete investigation [1,8].

ANTONIO GIROLAMI PAMELA SCARPARO EMANUELE ALLEMAND Department of Medical and Surgical Sciences and Northeastern Italy Association for the Study of Coagulation Disorders, University of Padua Medical School, Padua, Italy Published online 29 January 2008 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ajh.21161

518

To the Editor: The prothrombin activity level of our patient [1] was 20% of normal in 1965, as measured by coagulation assay. Prothrombin antigen was not tested. Other factor assays were normal. The prothrombin activity levels of the patient’s siblings and parents were normal also. The 2006 patient evaluation revealed a prothrombin activity of 7–9% of normal (also measured by coagulation assay). By this time, the patient, having been treated on numerous occasions for bleeding with plasma and red cells, had developed hepatitis C-induced cirrhosis. The lower prothrombin activity value in 2006 was likely due to progressive liver disease superimposed upon congenital prothrombin deficiency; other factor levels were now low (factor V ¼ 41%, factor VII ¼ 42%, factor X ¼ 59%), but factor VIII activity was elevated at 267%. Because of her progressive liver disease, the patient underwent liver transplantation. Her factor levels [PTT 29 sec (nl 26–37 sec), PT 14 sec (nl 12– 15.5 sec), and prothrombin activity 107% (nl 86–150%)] have subsequently returned to normal, precluding further studies. In the original reference, Lefkowitz et al. [2] described prothrombin Puerto Rico I as having characteristics of both hypoprothrombinemia (Type I deficiency or decreased antigen) and dysprothrombinemia (Type II deficiency or decreased activity). Prothrombin Puerto Rico I consists not only of an Arg457Gln missense mutation but also a Thr122Met missense mutation, among other consensual synonymous or noncoding mutations. The Thr122Met mutation in the kringle 1 domain may in fact decrease the stability of prothrombin, particularly under--carboxylated prothrombin [3], leading to the Type I deficiency component. However, the adjacent kringle 2 domain binds FV, part of the prothrombinase complex; alterations in the kringle 1 domain may affect the efficiency of FV binding to kringle 2 thus affecting the activation of prothrombin to thrombin [4] and leading to the Type II deficiency component. Without finding carriers of only one of the two missense mutations co-segregating in prothrombin Puerto Rico I, it would be difficult to assess which missense mutation is responsible for the reduced antigen and which for the reduced activity. Recently, one group in Italy [5] has found a family with the Thr122Met mutation (along with a previously reported Arg382Hys mutation) and suggests the Thr122Met mutation results in dysprothrombinemia. We look forward to their in vitro functional analysis.

STEPHEN KLING GEORGE RODGERS Division of Hematology, University of Utah Health Sciences Center, Salt Lake City, Utah Published online 13 February 2008 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ajh.21169

References 1. Kling SJ, Jones KA, Rodgers GM. A second case of prothrombin Puerto Rico I in the United States. Am J Hematol 2007;82:661–662. 2. Lefkowitz JB, Weller A, Nuss R, et al. A common mutation, Arg457?Gln, links prothrombin deficiencies in the Puerto Rican population. J Thromb Haemost 2003;1:2381–2388.

American Journal of Hematology

3. Wu W, Bancroft JD, Suttie JW. Structural features of the kringle domain determine the intracellular degradation of under-g-carboxylated prothrombin: studies of chimeric rat/human prothrombin. Proc Natl Acad Sci USA 1997;94:13654– 13660. 4. Kotkow KJ, Deitcher SR, Furie B, Furie BC. The second kringle domain of prothrombin promotes factor Va-mediated prothrombin activation by prothrombinase. J Biol Chem 1995;270:4551–4557. 5. Prothrombin deficiency, 2008. CEINGE—Biotecnologie Avanzate. 23 January 2008 http://www.ceinge.unina.it/index.php?option¼com_content&task¼view&id¼ 60&Itemid¼26&lang¼en.

Clinical JAK2 V617F mutation testing: Limited utility for general hospital patients with venous and arterial thromboses in common locations To the Editor : It is well known that polycythemia vera (PV) and essential thrombocythemia (ET) increase a patient’s risk of arterial and venous thromboses, and that this risk can manifest years before the disorders become clinically apparent in some cases. Although recent reports have suggested the possibility of latent myeloproliferative disease by finding JAK2 V617F mutations in patients with splanchnic vein thromboses (SVT, such as portal and mesenteric vein thrombosis [1] and the Budd-Chiari syndrome [2]) the usefulness of routine JAK2 V617F testing has seldom been assessed in patients with thromboses in common locations. In a retrospective analysis of 295 patients, the majority of whom were recruited for a prior study of risk factors for venous thromboembolism, only one was found to be positive for JAK2 V617F [3]. While provocative, the applicability of these results to a general hospital population is uncertain. In a more recent study of 210 patients undergoing workup for non-SVT thrombosis in the absence of overt myeloproliferative disorder, using a very sensitive research-based test, four patients (2%) were found to harbor the JAK2 V617F mutation [4]. Another analysis from this same group found the prevalence of JAK2 V617F in patients undergoing workup for stroke to be less than 1% [5]. To our knowledge, the results of these two studies have not been confirmed in another cohort of general hospital patients undergoing workup for arterial or venous thrombophilia and utilizing a standard clinical test for JAK2 V617F. We identified all patients who had undergone molecular diagnostic tests for the factor V Leiden (FVL) mutation and the prothrombin gene mutation (PT G20210A) over the course of 1 year at the Brigham and Women’s Hospital and the Dana-Farber Cancer Institute. For patients who were negative on both tests, we then determined, by medical record review, the nature of the thrombotic events for which they were being evaluated, and tested them for JAK2 V617F. Our clinical JAK2 V617F assay, modified from the method of Jones et al. [6] reliably detects >5% mutated alleles in any given sample with sufficient DNA quantity and quality; a prior pilot study in our laboratory using our test showed JAK2 V617F to be present in 8 of 9 (89%) patients with a known diagnosis of PV. Patients without a documented thrombosis were excluded, as were patients with a history of malignancy, myeloproliferative disease, or heparin-induced thrombocytopenia (HIT) within 1 year of thrombosis. For patients with documented thromboses, other predisposing laboratory and clinical risk factors were also assessed.

From February 1, 2006, to January 31, 2007, 111 patients had negative tests for both FVL and PT G20210A at our institutions. Of these, our medical record review yielded 66 who had documented thromboses without a recent history of malignancy, myeloproliferative disease, or HIT. Only one patient was found to have the JAK2 V617F mutation (1.5%, 95% CI [1.4, 4.5]), and this patient was subsequently found to have undergone stem cell transplantation for idiopathic myelofibrosis more than 10 years prior. This patient presented with SVT and was eventually found to have relapsed myeloproliferative disease. Positive laboratory hypercoaguability studies by thrombosis type are summarized in Table I (these are from medical record review; note that outside of JAK2 V617F, all tests were not performed on all patients). Some of the patients in our cohort had one or more documented clinical risk factors for thrombosis such as current smoking (25.8%, 95% CI [15.2, 36.3]), surgery (27.3%, 95% CI [16.5, 38.0]) or pregnancy (11.1%, 95% CI [0.8, 21.4] within the prior month, or use of oral contraceptives (19.4%, 95% CI [6.5, 32.4). Interestingly, 30.3% (95% CI [19.2, 41.4]) of our general hospital patients with arterial or venous thrombosis and negative tests for FVL and PT G20210A had no other clinical or laboratory thrombophilic risk factors found through medical record review. In summary, we were not able to detect JAK2 V617F mutations in the overwhelming majority of general hospital patients with arterial or venous thromboses in our cohort. Especially in the context of the aforementioned studies, this finding has important clinical implications for the workup of patients with thrombophilia. Specifically, outside of special circumstances such as splanchnic vein thromboses or thrombophilic patients with evidence of myeloproliferative disease (eg, elevated hemoglobin or platelet count), JAK2 V617F mutation testing should not be part of the routine thrombophilia workup.

GREGORY A. ABEL1 DANIEL J. DEANGELO1 JEAN M. CONNORS2 LYNETTE M. SHOLL3 RONALD P. MCCAFFREY2 JANINA A. LONGTINE3 1

Division of Hematologic Oncology, Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA 2 Hematology Division, Department of Medicine, Brigham and Women’s Hospital, Boston, MA 3 Department of Pathology, Brigham and Women’s Hospital, Boston, MA Published online 29 January 2008 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ajh.21159

References 1. Colaizzo D, Amitrano L, Tiscia GL, et al. The JAK2 V617F mutation frequently occurs in patients with portal and mesenteric venous thrombosis. J Thromb Haemost 2007;5:55–61. 2. Patel RK, Lea NC, Heneghan MA, et al. Prevalence of the activating JAK2 tyrosine kinase mutation V617F in the Budd-Chiari syndrome. Gastroenterology 2006;130:2031–2038.

TABLE I. JAK2 V617F and Other Laboratory Parameters in 66 General Hospital Patients with Documented Venous or Arterial Thromboses and Negative Tests for Factor V Leiden and the Prothrombin Gene Mutationa Protein C deficiency

Protein S deficiency

Antithrombin III deficiency

Antiphospho-lipid Abs

Elevated homocysteine

JAK2 V617 mutation

None

0% 11% 29% 0%

20% 20% 29% 25%

0% 2% 0% 0%

10% 4% 0% 0%

0% 13% 14% 25%

0% 0% 14%b 0%

70% 60% 29% 50%

CVA/stroke (n ¼ 10) DVT/PE (n ¼ 45) SVT (n ¼ 7) Other arterial thrombosis (n ¼ 4)

CVA, cerebrovascular accident; DVT/PE, deep venous thrombosis/pulmonary embolism; SVT, splanchnic vein thrombosis. Note: Except for JAK2V671F, not all tests were performed on all patients; numbers reflect percentage of patients with positive tests documented in the medical record. b Note: This represents one patient who had a history of a distant prior successful stem cell transplantation for idiopathic myelofibrosis; the patient subsequently had relapsed myeloproliferative disease. a

American Journal of Hematology

519

3. Remacha AF, Estivill C, Sarda MP, et al. The V617F mutation of JAK2 is very uncommon in patients with thrombosis. Haematologica 2007;92:285–286. 4. Pardanani A, Lasho TL, Schwager S, et al. JAK2V617F prevalence and allele burden in non-splanchnic venous thrombosis in the absence of overt myeloproliferative disorder. Leukemia 2007;21:1828–1829. 5. Pardanani A, Lasho TL, Morice WG, et al. JAK2V617F is infrequently associated with arterial stroke in the absence of overt myeloproliferative disorder. J Thromb Haemost 2007;5:1784–1785. 6. Jones AV, Kreil S, Zoi K, et al. Widespread occurrence of the JAK2 V617F mutation in chronic myeloproliferative disorders. Blood 2005;106:2162–2168.

Partial splenic embolization preceding splenectomy, in a case of refractory immune thrombocytopenic purpura To the Editor: Immune thrombocytopenic purpura (ITP) is an autoimmune disorder characterized by peripheral destruction of opsonized platelets in the reticuloendothelial system, particularly the spleen, which results in a markedly reduced platelet count and bleeding. Steroids are the accepted front-line treatment for ITP, with 80–90% of patients achieving a raised platelet count above 50  109/L. However, thrombocytopenia recurs in most cases [1]. In steroid-refractory ITP, many therapeutic approaches have been proposed, including danazol, azathioprine, plasmapheresis, anti-CD20 monoclonal anti-

bodies (Rituximab), orally active thrombopoietin receptor agonists, and combined therapies (immunoglobulin, methylprednisolone, vinca alkaloids) [2]. Surgical approach by splenectomy effectively prolongs platelets survival, providing a definitive ‘‘cure’’ in more than 50% of patients [3], but it is burdened by the risk of major bleeding events if thrombocytopenia is severe. Partial splenic embolization (PSE) is a nonsurgical procedure developed to treat hypersplenism due to hepatic diseases and has recently been attempted as an ITP treatment option [4]. We report the case of a 21-year-old female ITP patient who was refractory to steroids (prednisone, 1mg/kg daily for 30 days) and high-dose immunoglobulin (400 mg/kg daily for 5 days). She was subsequently treated with Rituximab (375 mg/sqm intravenously once weekly for 4 weeks), and steroids (3 cycles of Dexamethasone, 40 mg daily for 4 days after a 15-day interval), without benefit. Because of the persistently low platelet count (below 10  109/L), PSE was performed prior to splenectomy. Splenic artery embolization was performed by the right femoral access, using a 4 French Cobra-shaped catheter (C1) to minimize bleeding risk at the site of arterial approach. The catheter tip was placed in the distal portion of the splenic artery and an angiographic run was acquired to define the arterial anatomy, the parenchymal perfusion and the portal venous phase. Angiography demonstrated a normal splenic vascular anatomy and the absence of parenchymal focal lesions or portal thrombosis (Fig. 1A). After selective catheterization of main splenic vessels, arterial occlusion was performed by slow injection of a biodegradable large reabsorbable embolic agent (Gelfoam pledgets mixed with iodinate contrast), which obtained a complete recanalization. The procedure was carried out under fluoroscopic guidance, until a slowing flow was visualized without contrast reflux in the splenic artery proximal part (Fig. 1B). Five hours later, the platelet count was 56  109/L and laparoscopic splenectomy was performed without complications. The platelet count rose to 190  109/L, 1000  109/L, and 486  109/L after one, seven, and thirty days from splenectomy. Twelve months after splenectomy, platelet count remains normal. To date, a small but significant fraction of ITP patients undergo splenectomy with severe thrombocytopenia. Complete or PSE has been proposed as an alternative treatment in ITP patients not eligible for splenectomy, but thrombocytopenia recurred in about 30% of initially responding patients. In the present case, PSE was not intended to be curative, but preparatory to laparoscopic splenectomy, reducing the intraoperatory risk. The combination of PSE and splenectomy was safe and successful, with the patient alive and in complete remission 12 months after splenectomy. This approach seems to be promising, and its safety and efficacy shall find confirmation in studies of larger cohorts of ITP patients.

NICOLA VIANELLI1 FRANCESCA PALANDRI1 LUCIA CATANI1 LUCA BOSCHI2 GILBERTO POGGIOLI2 EMANUELA GIAMPALMA3 MICHELE BACCARANI1 1

Institute of Hematology and Medical Oncology ‘‘Sera`gnoli,’’ Policlinico S.Orsola-Malpighi, University of Bologna, Bologna, Italy 2 General Surgery Department, Policlinico S. Orsola-Malpighi, University of Bologna, Bologna, Italy 3 Department of Radiology, Policlinico S. Orsola-Malpighi, University of Bologna, Bologna, Italy Published online 13 February 2008 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ajh.21172

References

Figure 1. Angiographic control of partial splenic embolization. Selective catheterism of the splenic artery (Fig.1A) and embolization of splenic vessels, with complete interruption of distal flux (Fig. 1B).

520

1. Ruggeri M, Fortuna S, Rodeghiero F. Heterogeneity of terminology and clinical definitions in adult idiopathic thrombocytopenic purpura: A critical appraisal from literature analysis. Pediatr Blood Cancer 2006:47:653–656. 2. Cines DB, Bussel JB. How I treat idiopathic thrombocytopenic purpura (ITP). Blood 2005;106:2244–2251. 3. Kojouri K, Vesely SK, Terrell DR, George JN. Splenectomy for adult patients with idiopathic thrombocytopenic purpura: a systematic review to assess longterm platelet count responses, prediction of response, and surgical complications. Blood 2004;104:2623–2634. 4. Kimura F, Itoh H, Ambiru S, et al. Long-term results of initial and repeated partial splenic embolization for the treatment of chronic idiopathic thrombocytopenic purpura. AJR Am J Roentgenol 2002;179:1323–1326.

American Journal of Hematology