Christopher Sigouin

Evaluation of pretest clinical score (4 T's) for the diagnosis of heparin-induced thrombocytopenia in two clinical settings.

Summary.  Background: Heparin‐induced thrombocytopenia (HIT) is a prothrombotic adverse drug reaction caused by heparin. As thrombocytopenia is common in hospitalized patients receiving heparin, it would be useful to have a clinical scoring system that could differentiate patients with HIT from those with other reasons for thrombocytopenia. Aim: To compare prospectively the diagnostic utility of a clinical score for HIT in two different clinical settings. Methods: The pretest clinical scoring system, the ‘4 Ts’, was used to classify 100 consecutive patients referred for possible HIT in one hospital (Hamilton General Hospital, HGH) into high, intermediate, and low probability groups. This system was also used to classify likewise 236 patients by clinicians in Germany referring blood for diagnostic testing for HIT in Greifswald (GW). The clinical scores were correlated with the results of laboratory testing for HIT antibodies using the serologic criteria for HIT with high diagnostic specificity. Results: In both centers, patients with low scores were unlikely to test positive for HIT antibodies [HGH: 1/64 (1.6%), GW: 0/55 (0%)]. Patients with intermediate [HGH: 8/28 (28.6%), GW: 11/139 (7.9%)] or high scores [HGH: 8/8 (100%), GW: 9/42 (21.4%)] were more likely to test positive for clinically significant HIT antibodies. The positive predictive value of an intermediate or high clinical score for clinically significant HIT antibodies was higher at one center (HGH). Conclusions: A low pretest clinical score for HIT seems to be suitable for ruling out HIT in most situations (high‐negative predictive value). The implications of an intermediate or high score vary in different clinical settings.

Systematic Review: Efficacy and Safety of Rituximab for Adults with Idiopathic Thrombocytopenic Purpura

Idiopathic thrombocytopenic purpura (ITP) is a common hematologic disorder characterized by platelet autoantibodies, low platelet counts, and bleeding. Rituximab is a chimeric, monoclonal anti-CD20 antibody that targets B lymphocytes and causes Fc-mediated cell lysis (14). It is currently indicated for the treatment of lymphoma (58), but because of its ability to deplete autoantibody-producing B lymphocytes and its favorable toxicity profile (9), it has been used in patients with various autoimmune diseases (1012), including ITP. In some patients with ITP, rituximab has been associated with a reduction in specific platelet-associated autoantibodies and an increase in platelet count (13). Early success with rituximab in ITP has lead to its widespread use and incorporation into recent treatment algorithms (14, 15). However, the evidence to support the use of rituximab in ITP is uncertain. We performed a systematic review of the literature to evaluate the efficacy and safety of this treatment. Methods Search Strategy One hematologist and one internist independently searched the literature in June 2005 and updated the search in April 2006. The electronic databases of MEDLINE (from 1966) and EMBASE (from 1980) were searched by using the explode function for the Medical Subject Heading (MeSH) terms antibodies, monoclonal and purpura, thrombocytopenic, idiopathic and the textwords rituximab, rituxan, mabthera, anti CD20, anti CD20 antibody, immune thrombocytopenic purpura, and idiopathic thrombocytopenic purpura. The MEDLINE database was also searched with the PubMed search engine by using the MeSH term purpura, thrombocytopenic, idiopathic and the textwords rituximab and rituxan. The Cochrane Registry for Controlled Trials was searched by using the terms rituximab, immune thrombocytopenic purpura, and ITP. Scientific abstracts were identified by searching the electronic databases of the American Society of Hematology and the American Society of Clinical Oncology from 1997 (the year of licensure of rituximab) to 2005 by using the search terms ritux*, thrombocytopenic, and ITP. Bibliographies of relevant articles and reviews were manually searched, and authors were canvassed for additional citations. Eligibility Criteria and Study Selection Exclusion criteria were secondary causes of thrombocytopenia, including splenomegaly, hepatitis B or C virus infection, HIV infection, lupus, antiphospholipid antibody syndrome, bone marrow failure syndromes, and drug-induced thrombocytopenia; malignancy, including chronic lymphocytic leukemia and lymphoma; the Evan syndrome; and rituximab re-treatments. Children (<16 years of age) were excluded because the biology and natural history of ITP in children were believed to differ considerably from those in adults. There was no restriction on study design or language of publication. Reports published only in abstract form were eligible. Where duplicate or redundant publications were uncovered, the latest and most informative version was retained. Initially, titles and abstracts of all articles were evaluated independently by 2 reviewers. Full-text articles were retrieved when they were judged by at least 1 reviewer to possibly contain relevant original data. Final article selection was done independently by both reviewers, and disagreements were resolved by consensus in all cases. Data Extraction The following data were collected in duplicate: proportion of patients with complete, partial, or minimal platelet count responses (and their definitions); time to platelet count responses; duration of platelet count responses; dose and schedule of rituximab administration; toxicities; previous ITP treatments; baseline platelet count; duration of ITP before rituximab treatment; study design and use of controls; and sources of funding. Individual-patient data were used where possible. Assessment of Methodologic Quality Study quality was assessed independently by 2 hematologists with expertise in research methods. Reviewers evaluated 4 key design features for each study: prospective data collection, consecutive patient enrollment, a clearly stated duration of follow-up, and a description of losses to follow-up. Assessors were blinded to study author, journal, publication date, and main results. Disagreements were resolved by independent adjudication. Statistical Analysis Patient demographic characteristics and platelet count responses were analyzed only from those studies enrolling 5 or more patients because we felt that smaller studies may be subject to extreme reporting bias. To determine estimates of response, we defined complete response as the achievement of a platelet count greater than 150109 cells/L; partial response as a platelet count between 50 and 150109 cells/L; and overall response as a platelet count greater than 50109 cells/L. These definitions were chosen to reflect the most common criteria used in primary reports. Toxicities were considered from all studies, including those enrolling fewer than 5 patients each, to provide the most thorough description of safety. We determined estimates of effect of rituximab by calculating the weighted mean proportion by using a random-effects model. This model estimated the between-study variance by using the method of moments and assumed that the proportion from each study was sampled from the normal distribution, with variance calculated from the data. Continuous variables, including time to response, response duration, and follow-up, were summarized with medians, minimum and maximum values, and interquartile ranges assuming a normal distribution of the data. Unweighted chance-corrected values were used to assess agreement between reviewers for study selection (16). Role of the Funding Source This systematic review had no external source of funding. The organizations that fund the individual authors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; or preparation, review, and approval of the manuscript. Results Study Selection We identified 599 citations through our comprehensive literature search, of which 60 were retrieved for detailed review (Figure 1). Agreement between reviewers for initial study inclusion was excellent (= 0.87). After exclusion of ineligible studies, redundant or duplicate publications, and reports that did not contain original data, 31 reports were included. Nineteen studies (313 patients) enrolled at least 5 patients each and were included in the efficacy analysis (13, 1734), and 29 studies (306 patients) reported toxicity data (13, 1719, 2128, 3046). Of the 19 reports describing efficacy outcomes, 9 were published in abstract form only. Abstracts were carefully scrutinized, and authors were contacted when necessary to ensure that redundant publications were excluded. Figure 1. Article selection. Results of article search and selection conducted in accordance with guidelines on reporting systematic reviews of observational studies (56). ITP = idiopathic thrombocytopenic purpura. Study Designs and Sources of Funding There was 1 dose-finding phase II study (28) and 18 single-arm cohort studies (13, 1727, 2934). Source of funding was not reported in 26 of 31 reports; of the remaining 5, 1 was industry-sponsored (19), 3 were funded by nonprofit organizations (21, 31, 41), and 1 reported that it had no funding information to disclose (32). Description of Patients Patients were 16 to 89 years of age, had had ITP for 1 to 360 months, and had a platelet count that ranged from 1 to 89109 cells/L before rituximab treatment (Table 1). Nearly all (99.0%) patients had received corticosteroids, and 158 (50.5%) had had splenectomy. Other previous treatments were immunosuppressants, including cyclosporine, azathioprine, or mycophenylate (n= 26); cyclophosphamide (n= 12); vinca alkaloids (n= 18); and danazol (n= 17). The number of previous treatments varied between and within reports. Table 1. Characteristics of Patients with Idiopathic Thrombocytopenic Purpura in Rituximab Studies Enrolling 5 or More Patients Each (n= 313)* Rituximab Dose and Schedule Rituximab was administered as a weekly infusion of 375 mg/m2 for 4 consecutive weeks in 16 of 19 studies. Of the remaining 3 studies, 1 did not report the dosing schedule (30); 1 used a schedule of 1 to 8 infusions of 325 mg/m2 per dose (29); and 1 used a low dose of rituximab (50 mg/m2 on day 1, then 150 mg/m2 on days 8 and 15), an intermediate dose (150 mg/m2 on day 1, then 375 mg/m2 weekly for 3 weeks), and a standard dose (28). Platelet Count Response In most reports, complete response and partial response were defined according to the achievement of predefined platelet count thresholds; however, these thresholds varied. Certain reports used additional criteria to define a response, including the discontinuation of steroid therapy (32) and the resolution of bleeding symptoms (26). One report defined complete response as the achievement of a platelet count that was adequate for hemostasis (25); in 2 reports, neither complete response nor partial response was defined (22, 27). The timing of platelet count measurements in the definitions of a response was specified in 2 reports: 12 weeks after the first rituximab infusion (29) and 2 weeks after the last infusion (30). In 1 report, a response was considered only if it lasted at least 30 days (21), and in another, at least 4 months (13). Where reporting of studies included homogenous criteria to define platelet count responses to therapy, treatment with rituximab resulted in a complete response (platelet count> 150109 cells/L) in 46.3% of patients (95% CI, 29.5% to 57.7%), partial response (50 to 150109 cells/L) in 24.0% (CI, 15.2% to 32.7%), and overall response (>50109 cells/L) in 62.5% (CI, 52.6% to 72.5%). Rates of complete, partial, and overall response were based on 191, 284, and 313 eligible patients, respectively (Table 2). In a sensitivity analysis that exclud

Quantitative interpretation of optical density measurements using PF4‐dependent enzyme‐immunoassays

Summary.  Background: Many laboratories test for heparin‐induced thrombocytopenia (HIT) using a PF4‐dependent enzyme‐immunoassay (EIA). An advantage of the EIA is its simplicity; a disadvantage is that it only indirectly detects heparin‐dependent, platelet‐activating antibodies (‘HIT antibodies’). Objectives: To determine whether the magnitude of a positive EIA result, expressed in optical density (OD) units, predicts risk of HIT antibodies, defined as a strong‐positive platelet serotonin‐release assay (SRA) result (≥50% serotonin release). Patients/methods: We determined the risk of a strong‐positive SRA result for five categories of OD reactivity (<0.40, 0.40–<1.00, 1.00–<1.40, 1.40–<2.00, and ≥2.00 OD units) using two EIAs (commercial anti‐PF4/polyanion IgG/A/M and in‐house anti‐PF4/heparin–IgG). Results: For patient sera investigated for HIT antibodies, a weak‐positive result (0.40–<1.00 OD units) in either EIA indicated a low probability (≤5%) of a strong‐positive SRA; the risk increased to ∼90% with an OD ≥ 2.00 units. Quantifying the EIA–SRA relationship for 1553 referred patient sera, we found that for every increase of 0.50 OD units in the EIA–IgG, the risk of a strong‐positive SRA result increased by OR = 6.39 [95% confidence interval (CI), 5.13, 7.95; P < 0.0001]. For every increase of 1.00 OD units in the EIA–IgG, the risk increased by OR = 40.81 (95% CI, 26.35, 63.20; P < 0.0001). Conclusions: The probability of HIT antibodies (strong‐positive SRA result) inferred by a positive PF4‐dependent EIA varies considerably in relation to the magnitude of the EIA result, expressed as OD values. In our laboratory, the probability of HIT antibodies being present reached ≥50% only when the OD level was ≥1.40 units.

Methodologic issues in the use of bleeding as an outcome in transfusion medicine studies

BACKGROUND: Prophylactic platelet transfusions are given to thrombocytopenic patients to prevent bleeding. The benefit of platelet transfusions has frequently been assessed by measuring the count increment; however, more recently, an assessment of bleeding has been used because it is a more clinically relevant outcome measure. The purpose of this study was to identify platelet transfusion trigger studies that used bleeding as an outcome measure, compare and contrast methods used to document bleeding and analyze bleeding outcomes, and identify and discuss methodologic issues to consider when bleeding is used as a study outcome.

Utilization of platelet transfusions in the intensive care unit: indications, transfusion triggers, and platelet count responses.

BACKGROUND: A description of current platelet (PLT) transfusion practice in the intensive care unit (ICU) is needed.

Methods for the analysis of bleeding outcomes in randomized trials of PLT transfusion triggers

BACKGROUND:  A number of methodologic challenges arise in the analysis of bleeding data from clinical trials of PLT transfusion triggers. It is important to understand the assumptions and role of the various methods of analysis to interpret published trials and to design future studies appropriately.

In vivo recovery and survival of apheresis and whole blood‐derived platelets: a paired comparison in healthy volunteers

BACKGROUND: Methods of platelet preparation may alter the recovery and survival characteristics of platelets following transfusion. As suggested by a recent clinical trial, platelet recovery may be better preserved with apheresis platelet preparations than with platelets prepared from whole blood by the platelet‐rich plasma (PRP) method.

In vitro evaluation of prestorage pooled leukoreduced whole blood–derived platelets stored for up to 7 days

BACKGROUND: Advantages to storing whole blood–derived platelets (PLTs) as a pool for 7 days would include operational efficiencies and facilitation of bacterial testing and pathogen inactivation. The in vitro quality of pre‐storage pooled PLTs stored for up to 7 days was assessed.

Assessing the effectiveness of whole blood–derived platelets stored as a pool: a randomized block noninferiority trial

BACKGROUND: Prestorage pooling of whole blood–derived platelets (PLTs) would simplify bacterial detection. This study evaluated the in vivo effect of the prestorage pooling of PLTs stored for up to 5 days, by assessing the corrected count increment (CCI) 18 to 24 hours after transfusion of the product.