Ability of hemostatic assessment to detect bleeding disorders and to predict abnormal surgical blood loss in children: a systematic review and meta-analysis.

Background: Systematic preoperative coagulation testing is still widely used in children scheduled for surgery, although current guidelines recommend that a bleeding history should be the first choice for hemostatic assessment. We performed a systematic review with meta-analysis to evaluate the pertinence of bleeding questionnaire and screening laboratory testing to detect bleeding disorders (BDs) in children and to predict abnormal surgical blood loss.

Methods: A search was conducted in PubMed, EMBASE, MEDLINE(R), Cochrane Central Register of Controlled Trials, Health technology Assessment, and all EBM Reviews (Cochrane DSR, ACP Journal Club, DARE, CCTR, CMR, HTA, and NHSEED and EBM Reviews) up to October 22, 2013. Prospective trials containing 20 children or more and any tests evaluating either the ability of the test to detect a congenital BD or the ability of the test to predict increased surgical blood loss were retained. The quality of the study was judged with the Cochrane Collaboration Tool and two investigators extracted data independently. Data were combined to calculate the pooled diagnostic odds ratio (DOR) and their 95% confidence intervals (CI 95%). I(2) statistics were used to assess statistics heterogeneity.

Results: Data could be extracted from 16 studies. Best results for detecting a congenital abnormality at potential risk for increased surgical blood loss were obtained with the PFA-100 (DOR = 113.0; 95% CI, 22.6-566.2; I(2) = 0%) in two studies, followed by the bleeding time in two other studies (DOR = 110.7; 95% CI, 24.4-502.3; I(2) = 0%). With a high amount of heterogeneity, questionnaires showed disappointing performances (DOR = 7.9; 95% CI: 3.5-17.5; I(2) = 72.6%).

Conclusion: Current evidence does not identify a tool that adequately predicts BDs and/or abnormal surgical blood loss in children. Questionnaires currently available do not perform well. In the setting of a pediatric coagulation clinic, the PFA-100 has the highest chance of detecting a BD. This meta-analysis highlights the weakness of the literature regarding the prediction of perioperative bleeding due to congenital hemostatic disorders in children.