Hemodilution in pediatric trauma: Defining the expected hemoglobin changes in patients with liver and/or spleen injury: An ATOMAC+ secondary analysis

Background: Many children with blunt liver and/or spleen injury (BLSI) never bleed intraperitoneally. Despite this, decreases in hemoglobin are common. This study examines initial and follow up measured hemoglobin values for children with BLSI with and without evidence of intra-abdominal bleeding. Methods: Children ≤18 years of age with BLSI between April 2013 and January 2016 were identified from the prospective ATOMAC+ cohort. Initial and follow up hemoglobin levels were analyzed for 4 groups with BLSI: (1) Non bleeding; (2) Bleeding, non transfused (3) Bleeding, transfused, and (4) Bleeding resulting in non operative management (NOM) failure. Results: Of 1007 patients enrolled, 767 were included in one or more of four study cohorts. Of 131 non bleeding patients, the mean decrease in hemoglobin was 0.83 g/dL (+/−1.35) after a median of 6.3 [5.1,7.0] hours, (p = 0.001). Follow-up hemoglobin levels in patients with and without successful NOM were not different. For patients with an initial hemoglobin >9.25 g/dL, the odds ratio (OR) for NOM failure was 14.2 times less, while the OR for transfusion was 11.4 times less (p = 0.001). Conclusion: Decreases in hemoglobin are expected after trauma, even if not bleeding. A hemoglobin decrease of 2.15 g/dL [0.8 + 1.35] would still be within one standard deviation of a non bleeding patient. An initial low hemoglobin correlates with failure of NOM as well as transfusion, thereby providing useful information. By contrast, subsequent hemoglobin levels do not appear to guide the need for transfusion, nor correlate with failure of NOM. These results support initial hemoglobin measurement but suggest a lack of utility for routine rechecking of hemoglobin. Level of evidence: Level II Prognostic Study.