S908 ASSESSMENT OF BLEEDING SEVERITY IN THROMBOCYTOPENIC PATIENTS UNDERGOING MINOR INVASIVE PROCEDURES IS HIGHLY VARIABLE, A SYSTEMATIC REVIEW

Abstract

Background:Platelet concentrates are frequently administered to thrombocytopenic patients to prevent bleeding prior to invasive procedures, such as central venous catheter (CVC) placement, diagnostic punctures or biopsies. Studies analyzing bleeding risk of such procedures strongly depend on the method of documenting bleeding complications.Aims:The objective of our study was to systematically review what methods and definitions are used to assess bleeding severity in clinical studies in the setting of minor invasive procedures in thrombocytopenic patients.Methods:We performed a systematic review to identify studies in which thrombocytopenic patients underwent various minor invasive procedures (CVC, liver biopsy (LB), bone marrow biopsy or lumbar puncture) and in which bleeding was described as outcome. Studies were retrieved from MEDLINE and platelet transfusion guidelines. Randomized controlled trials (RCT's) and observational studies (prospective and retrospective) were included. Case reports/series and pediatric studies were excluded. All relevant articles were reviewed to determine the methods used for bleeding assessment. A Mann‐Whitney U‐test was used to find any difference in bleeding incidence between prospective and retrospective studies, which we reported as median (interquartile range).Results:Forty‐nine articles met the predefined inclusion criteria (consisting of 2 RCT's and 47 observational studies; including 17 prospective and 31 retrospective studies, in 1 study the design was unclear). In all studies a different approach to assess bleeding severity was used. Sixteen studies used a categorical bleeding scale (4 used an existing scale and 12 designed their own scale), 32.7% of studies used a non‐categorical bleeding definition (12 according to their own design, 4 were derived from existing scales), 34.7% of studies provided no definition of bleeding.The bleeding incidence was highly variable, even between studies in comparable patient populations undergoing the same intervention (table 1). E.g.: 0.24% ‐ 6.49% bleeding in 5 studies of hematologic patients undergoing ultrasound guided CVC placement and 0% ‐ 51% bleeding in 2 studies of coagulopathic patients undergoing plugged percutaneous LB. In none of the studies a standard operating procedure was described and only in 1 study the protocol was published alongside the article.For the comparison of prospective and retrospective studies with respect to bleeding incidence 47 studies were included in our analysis. (2 studies were excluded because of: a mixed prospective/retrospective design (1) and an unclear design (1)). The median bleeding incidence was 3.9% (0.6%>6.8%) in prospective studies (n = 17) and 0.9% (0.02%>3.9%) in retrospective studies (n = 30), p = 0.08.Summary/Conclusion:The high variability in methods of bleeding assessment makes it impossible to compare clinical studies on bleeding risk following minor invasive procedures. Differences in bleeding definition and study design are the main cause of variability in bleeding incidences between the studies analyzed. In a considerable part of the studies, it was even unclear how bleeding complications were defined and recorded, making studies difficult to interpret and reproduce while these results are used to determine clinical transfusion thresholds. Retrospective studies appeared to less adequately capture minor bleeding and may therefore tend to report lower bleeding incidences. There is a strong need of consensus in bleeding definition in the field of transfusion medicine, for which we provide suggestions.image