RSS Feeds
Diagnosis and Treatment of Idiopathic Thrombocytopenic Purpura: Recommendations of the American Society of Hematology
- The American Society of Hematology ITP Practice Guideline Panel*. *For members of the American Society of Hematology ITP Practice Guidelines Panel, see the Appendix. Grant Support: In part by the American Society of Hematology. Requests for Reprints: James N. George, MD, The University of Oklahoma Health Sciences Center, Hematology-Oncology Section, PO Box 26901, Oklahoma City, OK 73190. Current Author Addresses: Dr. George: The University of Oklahoma Health Sciences Center, Hematology-Oncology Section, PO Box 26901, Oklahoma City, OK 73190.
Abstract
To develop guidelines for the diagnosis and management of idiopathic thrombocytopenic purpura (ITP) and to document the extent to which these guidelines are based on either scientific evidence or opinion, the American Society of Hematology established a panel composed of 13 hematologists with expertise in ITP, a clinical epidemiologist, and a practice guidelines methodologist.A comprehensive review was done of all published English-language studies that met explicit inclusion criteria and that evaluated the natural history of ITP or the effectiveness of testing and treatment options for ITP. The quality of each study was graded by two reviewers using formal methodologic rules. In subject areas for which data were inadequate, recommendations were based on opinion and were derived by using a formal scoring procedure. Confidential questionnaires were used to survey the hematologists on the panel about the appropriateness of testing and treatment options in hundreds of clinical scenarios. Practice recommendations were derived from the mean appropriateness scores for each indication. Voting was kept confidential to give each panel member an equal voice and to limit biases introduced by group dynamics. The recommendations were peer reviewed by eight outside experts.
This report focuses on data on and recommendations for adults with ITP.Little high-quality scientific evidence with which to assess the efficacy of diagnostic tests and treatments for ITP is available. The opinion of the panel was that most diagnostic tests are unnecessary in the routine work-ups of patients suspected of having ITP and that ITP accompanied by severe bleeding requires treatment with glucocorticoids, intravenous immunoglobulin, and other measures. However, treatment and hospitalization are often unnecessary when patients have only mild or moderate thrombocytopenia or minimal bleeding. Special therapeutic measures are sometimes indicated in pregnant women with ITP.
Idiopathic thrombocytopenic purpura [ITP], also known as primary immune or autoimmune thrombocytopenic purpura, is a common cause of thrombocytopenia and bleeding complications in children and adults. It may be confused with other causes of thrombocytopenia and is treated with agents that vary in efficacy, toxicity, and cost. In 1994, the American Society of Hematology established an expert panel to 1) review published data on the effectiveness of diagnostic tests and treatments for ITP and 2) develop evidence-based practice guidelines for the diagnosis and treatment of ITP. This article summarizes the panel's findings on adults with ITP, focusing on information of practical relevance to clinicians. The panel's evidence, recommendations, and methodology are described in detail elsewhere [1].
Methods
The 15-member panel comprised 13 hematologists, a clinical epidemiologist, and a practice guidelines methodologist. Nine of the hematologists were university based (6 had special expertise in adult ITP and 3 had special expertise in pediatric ITP; special expertise was defined by extensive publication in the subject area), and 4 were in private practice. At the initial meeting of the panel, each member provided a written statement disclosing the existence of any corporate relationships. At that time, one panel member was participating in and another had participated in pharmaceutical company-sponsored clinical trials on ITP that used intravenous immunoglobulin and anti-Rho (D). The panel determined that none of its members were disqualified. A comprehensive literature search, guided by an explicit evidence model, identified all relevant English-language studies that evaluated the natural history of ITP or the effectiveness of specific diagnostic and treatment options for ITP. The specific search terms used, the inclusion criteria, and the retrieval counts are described in the full report [1]. Case reports, case series that included fewer than five patients, editorials, and letters were generally excluded. Each study was reviewed independently by two panel members. Studies evaluating the effectiveness of treatment were rated using the “levels of evidence” approach outlined in Table 1. Only a subset of key references is cited here.
When it became clear that the literature on ITP lacked reliable studies with which to develop evidence-based recommendations, the panel turned to expert opinion to generate interim guidelines. The use of opinion was made explicit by using structured questionnaires, appropriateness scores, and consensus codes in a modified version of the approach described by Brook and colleagues [2]. Panel members completed questionnaires in which they rated, on a scale of 1 to 9, the necessity or appropriateness of tests or treatments in hundreds of clinical scenarios. The questionnaires were designed at panel meetings during which members were asked to identify the key diagnostic and treatment practices for which opinion would be assessed. The appropriateness of these practices was intentionally not discussed at meetings to avoid the influencing of responses by the opinions of more assertive panel members. Panel members were instructed to respond to questions in their areas of expertise; questions about adult patients were completed by 11 panel members. Tests or treatments that received a mean score of 7.0 to 9.0 were considered appropriate (or necessary), and those that received a mean score of 1.0 to 3.0 were considered inappropriate (or unnecessary) (Table 2). The panel recommended only tests and treatments that received a mean score of 7.0 to 9.0, and they recommended against those that received a mean score of 1.0 to 3.0. For each mean score, the panel assigned one of five letter codes (A through E) to rate the strength of consensus: the variance of panel members' opinions around the mean value (Table 2). Although it was based on methods that have been widely used [2], this way of assessing and reporting opinion was developed during this project and is unique to these guidelines. A complete listing of the appropriateness scores and consensus codes for the panel's recommendations is available in the full report [1].
The following recommendations represent the opinions of the panel because data on effectiveness were insufficient for developing evidence-based recommendations. Panel members discussed the design of the questionnaires and the meaning of the results, but voting was done confidentially to reduce bias and give each panel member an equal voice. The language used in the recommendations is keyed directly to the panel scores. “Should” statements (for example, “Testing should not be done …” or “Patients should be tested for human immunodeficiency virus …”) are used for tests or treatments that received a mean score of 1.0 to 3.0 for appropriateness or a mean score of 7.0 to 9.0 for necessity. Tests or treatments that received a mean score of 1.0 to 3.0 for necessity are described as “unnecessary,” and those that received a mean score of 7.0 to 9.0 for appropriateness are described as “appropriate.” The recommendations are silent about tests or treatments that received mean scores of 3.01 to 6.99 because these scores reflect a lack of consensus; thus, these recommendations do not address many clinical scenarios. For example, no explicit recommendations are given for patients with chronic refractory ITP because no data exist to support evidence-based recommendations and no consensus for opinion-based recommendations was reached. The full report [1] contains more details about the panel's methods.
Background Information
Definition
The panel defined ITP as isolated thrombocytopenia (low platelet count with otherwise normal results on complete blood count and peripheral blood smear) in a patient with no clinically apparent associated conditions or factors that can cause thrombocytopenia (such as infection with the human immunodeficiency virus [HIV], systemic lupus erythematosus, lymphoproliferative disorders, myelodysplasia, agammaglobulinemia, therapy with certain drugs, alloimmune thrombocytopenia, and congenital or hereditary thrombocytopenia). An abnormal blood count or peripheral blood smear due to a coexisting nonimmune condition (such as iron deficiency or thalassemia minor) does not, in itself, exclude the diagnosis of ITP.
Pathogenesis and Natural History
Primarily a disorder of increased platelet destruction, ITP is probably caused by the development of autoantibodies to platelet-membrane antigens [3]. The reported prevalence of ITP in adults and children is 1 to 13 per 100 000 persons [4]; the clinical presentation and course of ITP differ in children and adults. In children, ITP is usually an acute, self-limited disorder that resolves spontaneously; in adults, it is typically a chronic disorder with a more insidious onset. In about one third of adults with ITP, the condition is persistent and relatively resistant to most treatments [1]. Because adult patients with moderate to severe thrombocytopenia generally begin treatment immediately after diagnosis, data on the natural history of untreated disease are lacking. Data are also lacking on the natural history of adult patients who are incidentally discovered to have mild thrombocytopenia and are not treated. Available evidence suggests that only about 5% of adults with chronic ITP have spontaneous remission [1].
Health Consequences
If platelet counts are low enough, patients with ITP have an increased risk for clinically important and potentially life-threatening bleeding. The principal cause of death from ITP is intracranial hemorrhage. Older patients have hemorrhagic complications more often than do younger patients with the same platelet counts [5, 6]. Data indicate that about 5% of patients with ITP may die of hemorrhagic complications [1], but these data were obtained before current supportive care techniques were available.
Relatively few pregnant women with thrombocytopenia have ITP. About 75% have gestational thrombocytopenia (typically, mild thrombocytopenia that develops late in pregnancy and resolves spontaneously after delivery), and most of the others have low platelet counts related to preeclampsia [7]. Newborns of women with ITP (as opposed to newborns of women with gestational thrombocytopenia) may have thrombocytopenia and a slightly increased risk for intracranial hemorrhage at or after delivery. They have a reported risk of 10% to 13% for having a platelet count less than 50 × 10 (9/L), and they may have a risk of 1% for intracranial hemorrhage or death [1, 8]. Platelet counts may decrease further in the first week of life, generally reaching a nadir by day 1 or 2 after delivery [9].
Evidence of Effectiveness
Diagnostic Tests
Little scientific evidence is available about the accuracy or reliability of tests for ITP. The largely clinical diagnosis of ITP, which is made by confirming the presence of isolated thrombocytopenia and by excluding concurrent causes of thrombocytopenia, is best accomplished with a history, physical examination, complete blood count, and examination of the peripheral blood smear. Clinicians sometimes do additional tests, including those to rule out other causes of thrombocytopenia (such as HIV infection, thyroid disorders, and systemic lupus erythematosus); bone marrow examination; imaging studies; and tests for immunoglobulins, platelet antibodies, and lupus anticoagulant, but data on the predictive value or effectiveness of these tests in the typical patient are lacking.
No data support the need for special laboratory testing for pregnant women with ITP. Maternal platelet counts do not accurately predict fetal or neonatal platelet counts [1], and direct measurement of fetal platelet counts before delivery is problematic. Percutaneous umbilical blood sampling is done only in specialized centers and can induce fetal complications [10, 11]. Fetal scalp vein sampling can be done only after cervical dilatation and is less accurate than percutaneous umbilical blood sampling [1, 9].
Treatment
The principal therapeutic options for ITP include glucocorticoids, intravenous immunoglobulin, intravenous anti-Rho (D), and splenectomy. Other treatments have been used for refractory cases; these include azathioprine, cyclophosphamide, danazol, vinca alkaloids, ascorbic acid, colchicine, interferon-α, combination chemotherapy, protein A immunoadsorption, cyclosporine, ε-aminocaproic acid, plasma exchange, and accessory splenectomy. No direct evidence indicates that any of these treatments reduce bleeding complications or mortality from ITP. Most outcome studies rely on evidence of increased or normalized platelet counts as a surrogate measure of the effectiveness of treatment. The panel accepted low platelet counts as a valid surrogate measure because of studies that show a correlation between severe thrombocytopenia and clinically important bleeding [12]. The panel recognized the potential pitfalls of relying on this inference.
Even if platelet counts are accepted as a surrogate measure for clinical outcomes, relatively little scientific evidence is available on the effectiveness of treatments for ITP. A study with level II evidence, done in adults, reported no difference in outcomes between patients initially treated with glucocorticoids and those initially treated with intravenous immunoglobulin, but it may have lacked the statistical power necessary to detect an effect [13]. No direct evidence is available on the effectiveness of hospitalization or emergency treatment for severe manifestations of ITP. The effectiveness of other treatments (splenectomy, accessory splenectomy, anti-Rho [D], azathioprine, cyclophosphamide, vinca alkaloids, danazol, ascorbic acid, interferon-α, or plasma exchange) is supported only by studies with level V evidence [1]. The absence of control groups in these studies makes it unclear whether similar outcomes would have occurred with other regimens or in patients not receiving treatment. However, for splenectomy in adults, 36 studies with level V evidence have documented a sustained normalization of platelet counts in more than half of patients treated with splenectomy who were previously refractory to medical treatments for weeks or years [1]. All medical and surgical treatments for ITP carry a risk for potential adverse effects, but the probability and severity of complications vary [1].
Special issues arise in the treatment of pregnant women with ITP, but relevant data are limited. Certain medications used in ITP may harm the fetus, but few studies of these effects have been done. Splenectomy may increase the risk for preterm labor during the first trimester and can be technically difficult to do during the third trimester, but no data are available on the magnitude of risk. Pregnant women with low platelet counts theoretically have a higher risk for peripartum hemorrhage and bleeding complications from epidural anesthesia, but little direct evidence is available on the magnitude of these risks or the need for or effectiveness of preventive measures [14]. Little evidence is available on the threshold platelet count at which the fetus has an increased risk for traumatic intracranial hemorrhage. This risk is believed to be decreased by cesarean section, but data supporting this belief are lacking.
Recommendations
The following recommendations are based on the opinion of the panel and not on clinical outcome evidence. The inherent weakness of opinion-based recommendations is acknowledged; these recommendations should not be used as the basis for definitive decisions on health care policy or for inflexible rules about patient care. Further, these recommendations are not inclusive of all proper methods of care or of all methods of care that may achieve similar results. It is assumed in the recommendations that the patient does not present with atypical features or comorbid conditions.
Diagnosis
The diagnosis of ITP requires a history, physical examination, complete blood count, and examination of the peripheral smear. In the absence of findings that are atypical or suggest another diagnosis, further diagnostic studies in the routine work-up (Table 3) are not indicated. Patients with risk factors for HIV infection should be tested for antibodies to HIV. Bone marrow aspiration is appropriate to establish the diagnosis in patients older than 60 years of age and in patients for whom splenectomy is considered. Panel members felt that bone marrow aspiration in older patients may help to exclude myelodysplasia, which is more common with advancing age. It was also considered important to confirm the clinical diagnosis by bone marrow aspiration before recommending splenectomy. Preoperative thyroid function testing is appropriate to rule out occult hyperthyroidism or hypothyroidism before elective splenectomy. Additional testing is generally unnecessary and sometimes inappropriate.
Treatment
Recommendations for the initial treatment of patients presenting with ITP are summarized in Table 4. Patients with platelet counts greater than 20 × 109/L should not be hospitalized if they are asymptomatic or have only minor purpura. Patients with counts that exceed 50 × 109/L do not routinely require treatment; they should not routinely be given glucocorticoids or intravenous immunoglobulin as initial treatment. Intravenous immunoglobulin is also inappropriate as initial treatment in patients with platelet counts greater than 30 × 109/L who are asymptomatic or have only minor purpura.
Treatment is indicated in patients with platelet counts less than 20 to 30 × 109/L and in patients who have counts less than 50 × 109/L and have substantial mucous membrane bleeding (or risk factors for bleeding, such as hypertension, peptic ulcer disease, or the potential for substantial trauma to the body). Initial therapy with glucocorticoids, such as prednisone, is appropriate in such patients. Hospitalization is appropriate for patients with platelet counts less than 20 × 109/L who have substantial mucous membrane bleeding. Patients with severe or life-threatening bleeding should also be hospitalized and should receive conventional critical care measures along with treatment for ITP. Appropriate regimens include high-dose parenteral glucocorticoid therapy (for example, 1 g of methylprednisolone daily for 3 days), intravenous immunoglobulin, and platelet transfusions.
Splenectomy should not be done as initial therapy in patients who have no bleeding symptoms, patients who have minor purpura, or patients with mucous membrane or vaginal bleeding. In a patient who has had bleeding symptoms (for example, epistaxis or menorrhagia), splenectomy is often appropriate if platelet counts are still less than 30 × 109/L after 4 to 6 weeks of medical treatment. If elective splenectomy is planned, appropriate preoperative therapy includes prophylaxis with intravenous immunoglobulin or oral glucocorticoids for patients with platelet counts less than 20 × 109/L, provided that the patient is not known to be refractory to these therapies. Inappropriate preoperative prophylaxis includes intravenous immunoglobulin, oral or parenteral glucocorticoids, intravenous anti-Rho (D) when platelet counts exceed 50 × 109/L, and platelet transfusions when platelet counts exceed 10 × 109/L. The panel recognized that surgeons may require a platelet count exceeding 10 × 109/L before proceeding with surgery but felt that alternatives to platelet transfusions (such as therapy with intravenous immunoglobulin, anti-Rho [D], or glucocorticoids) were preferable and that bleeding complications from splenectomy appear to be minimal. (The panel endorsed the Advisory Committee on Immunization Practices' recommendation that patients receive Haemophilus influenza type B conjugate vaccine, polyvalent pneumococcal vaccine, and quadrivalent meningococcal polysaccharide vaccine at least 2 weeks before elective splenectomy [15].)
When symptoms of ITP persist after primary treatment with glucocorticoids and splenectomy, further therapy is recommended for patients with platelet counts less than 30 × 109/L who have active bleeding. When panel members were asked about their preferences for treatments in five categories of patients who had incomplete responses to prednisone and splenectomy (platelet counts between <10 × 109/L and 30 to 50 × 109/L, with or without bleeding symptoms), the most commonly recommended first-choice therapies included intravenous immunoglobulin, glucocorticoids, accessory splenectomy, and no additional treatment; other agents discussed in the full report [1] may also be appropriate. This difficult management area was not excluded from consideration, but no evidence on which to base recommendations was available. In addition, the opinions of panel members were too diverse to allow for consensus. Women with ITP who are of childbearing age and have platelet counts less than 10 × 109/L after splenectomy and other treatments should be discouraged from becoming pregnant.
Diagnosis in Pregnant Women
It is more difficult to diagnose ITP during pregnancy because it may be indistinguishable from gestational thrombocytopenia, but special laboratory testing is not required (Table 3). Blood pressure should be measured to rule out preeclampsia. Liver function testing is appropriate, and testing for antibodies to HIV should be done in patients with risk factors for HIV infection.
Treatment of Pregnant Women
Pregnant women with ITP and platelet counts greater than 50 × 109/L do not routinely require treatment and should not receive glucocorticoids or intravenous immunoglobulin as routine initial therapy. Women with platelet counts of 30 to 50 × 109/L in the first or second trimester also should not receive routine initial treatment with glucocorticoids or intravenous immunoglobulin. Treatment is required for women with platelet counts less than 10 × 109/L and women with platelet counts of 10 to 30 × 109/L who are in the second or third trimester or are bleeding. Intravenous immunoglobulin is appropriate initial treatment for women with platelet counts less than 10 × 109/L in the third trimester and for those with platelet counts of 10 to 30 × 109/L who are bleeding. In pregnant women with platelet counts less than 10 × 109/L who are bleeding and in whom glucocorticoid and intravenous immunoglobulin therapies have failed, splenectomy is appropriate in the second trimester. Splenectomy should not be done in asymptomatic pregnant women with platelet counts greater than 10 × 109/L.
As labor and delivery approach, women with ITP do not need testing for maternal platelet antibodies. Percutaneous umbilical blood sampling or fetal scalp vein sampling to measure the fetal platelet count and predict risk for neonatal bleeding are not necessarily required. Percutaneous umbilical blood sampling and fetal scalp vein sampling are unnecessary in pregnant women without known ITP, even if they have platelet counts of 40 to 75 × 109/L at term. Women with ITP should have cesarean section in some circumstances, which are outlined in the full report [1]. In general, assuming that the fetal platelet count (and the platelet counts of previous infants) is unknown, cesarean section is not indicated if the maternal platelet count exceeds 50 × 109/L. If the fetal platelet count is known, cesarean section is appropriate if the count is less than 20 × 109/L. A maternal platelet count greater than 50 × 109/L is considered sufficient to prevent complications due to excessive maternal bleeding during vaginal delivery or cesarean section. Prophylactic platelet transfusions before delivery are appropriate in women with counts less than 10 × 109/L who have a planned cesarean section or have epistaxis or other mucous membrane bleeding and are expected to deliver vaginally. These transfusions are unnecessary in women with platelet counts exceeding 30 × 109/L and no bleeding symptoms. Women with ITP should not be discouraged from breast feeding.
Priorities for Future Research
The exhaustive literature review done for this project underscored the limited data from rigorous clinical trials on which recommendations for the care of patients with ITP can be based. This lack of data affects almost every decision made by clinicians and patients. Major limitations of the current literature include lack of information on the clinical course (risk for clinically important bleeding or occurrence of spontaneous remission) of patients with untreated ITP, including patients who present with mild or moderate thrombocytopenia and no clinically important bleeding; the need for treatment in patients with ITP; the clinical course of chronic refractory ITP in adults; the relation of platelet counts to clinical outcomes (such as bleeding or mortality); the therapeutic efficacy of treatments related to clinical outcomes; and the relative cost-effectiveness of testing and treatment options. These issues should be addressed in prospective studies that have sample sizes and follow-up periods adequate to show effects and use end points that address clinical outcomes, such as bleeding and mortality, rather than relying solely on platelet counts.
Appendix
The members of the American Society of Hematology ITP Practice Guideline Panel were James N. George, MD (chair), University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma; Steven H. Woolf, MD, MPH, Medical College of Virginia, Fairfax, Virginia; Gary E. Raskob, MSc, University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma; Jeffrey S. Wasser, MD (co-chair), Watkins Center Hematology-Oncology, Manchester, Connecticut; Louis M. Aledort, MD, Mount Sinai Medical School, New York, New York; Penny J. Ballem, MD, University of British Columbia, Vancouver, British Columbia, Canada; Victor S. Blanchette, MD, University of Toronto, Toronto, Ontario, Canada; James B. Bussel, MD, Cornell University College of Medicine, New York, New York; Douglas B. Cines, MD, University of Pennsylvania, Philadelphia, Pennsylvania; John G. Kelton, MD, McMaster University, Hamilton, Ontario, Canada; Alan E. Lichtin, MD, Cleveland Clinic Foundation, Cleveland, Ohio; Robert McMillan, MD, Scripps Clinic and Research Foundation, La Jolla, California; John A. Okerbloom, MD, Heartland Oncology/Hematology, Council Bluffs, Iowa; David H. Regan, MD, Hematology-Oncology Group, Portland, Oregon; and Indira Warrier, MD, Wayne State University, Detroit, Michigan.
Dr. Woolf: 3712 Charles Stewart Drive, Fairfax, VA 22033.
Mr. Raskob: The University of Oklahoma Health Sciences Center, College of Public Health, Room 317, PO Box 26901, Oklahoma City, OK 73190.
Dr. Wasser: Watkins Center, Hematology-Oncology, 935 Main Street, Manchester, CT 06040.
Dr. Aledort: Mount Sinai Hospital, Box 1006, One Gustave I. Levy Place, New York, NY 10029.
Dr. Ballem: Women's Health Center, 4500 Oak Street, Vancouver, British Columbia V6S 1E2, Canada.
Dr. Blanchette: Hematology-Oncology Department, Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, Canada M5G 1X8.
Dr. Bussel: New York Hospital, Division of Pediatrics, 525 East 68th Street, New York, NY 10021.
Dr. Cines: 7 Founders, 3400 Spruce Street, University of Pennsylvania, Philadelphia, PA 19104.
Dr. Kelton: Medical Centre HSC-2N34, McMaster University, 1200 Main Street West, Hamilton, Ontario L8N 3Z5, Canada.
Dr. Lichtin: Hematology-Oncology, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195.
Dr. McMillan: Hematology Department, Scripps Clinic and Research Foundation, 10666 North Torrey Pines Road, La Jolla, CA 92037.
Dr. Okerbloom: Heartland Oncology/Hematology, One Edmundson Place, #100, Council Bluffs, IA 51503.
Dr. Regan: 5050 Northeast Hoyt Street, #256, Portland, OR 97213.
Dr. Warrier: Children's Hospital of Michigan, 3901 Beaubien Boulevard, Detroit, MI 48201.
- Copyright ©2004 by the American College of Physicians
