Restricted accessResearch articleFirst published online 2025-6
A Combined Model of Vital Signs and Serum Biomarkers Outperforms Shock Index in the Prediction of Hemorrhage Control Interventions in Surgical Intensive Care Unit Patients
Distinguishing surgical intensive care unit (ICU) patients with ongoing bleeding who require hemorrhage control interventions (HCI) can be challenging. Guidelines recommend risk-stratification with clinical variables and prediction tools, however supporting evidence remains mixed.
Methods
This retrospective study evaluated adult patients admitted to the surgical ICU with concern for ongoing hemorrhage under our institution's “Hemorrhage Watch” (HW) protocol and aimed to derive a clinical prediction model identifying those needing HCI with serial vital signs (VS) and serum biomarkers. The HW protocol included ICU admission followed by a 3-h observation period with VS monitoring every 15 min and hourly biomarkers. The primary outcome was the need for HCI (operative and endovascular interventions) within nine hours of ICU arrival. Secondary outcomes included in-hospital mortality, blood transfusions, and ICU and hospital length-of-stay. A clinical prediction model was developed by utilizing the variables most associated with HCI in a best subsets regression, which was subsequently internally validated using a Bootstrap algorithm.
Results
305 patients were identified for inclusion and 18 (5.9%) required HCI (3 operative, 15 endovascular). The median age was 70 years (IQR 54, 83), 60% had traumatic injuries, and 73% were enrolled from the emergency department. Blood product transfusion and mortality were similar between the HCI and no-HCI groups. Our analysis demonstrated that a model based on the minimum hemoglobin (9.9 vs 8.1 g/dL), minimum diastolic (57 vs 53 mm Hg) and systolic blood pressures (105 vs 90 mm Hg), and minimum respiratory rate (15 vs 18) could predict HCI with an area under the Receiver Operating Characteristics curve (AUROC) of 0.87, outperforming the Shock Index (SI) (AUROC = 0.64).
Conclusions
In this study of surgical ICU patients with concern for ongoing bleeding, a prediction model using serial VS and biomarkers outperformed the SI and may help identify those requiring HCI.
KauvarDSLeferingRWadeCE. Impact of hemorrhage on trauma outcome: An overview of epidemiology, clinical presentations, and therapeutic considerations. J Trauma Acute Care Surg. 2006;60(6):S3. doi:https://doi.org/10.1097/01.ta.0000199961.02677.19
2.
RoshanovPSEikelboomJWSesslerDI, et al.Bleeding independently associated with mortality after noncardiac surgery (BIMS): An international prospective cohort study establishing diagnostic criteria and prognostic importance. Br J Anaesth. 2021;126(1):163‐171. doi:https://doi.org/10.1016/j.bja.2020.06.051
RossaintRAfshariABouillonB, et al.The European guideline on management of major bleeding and coagulopathy following trauma: Sixth edition. Crit Care. 2023;27(1):80. doi:https://doi.org/10.1186/s13054-023-04327-7
5.
OgawaTInoueSInadaMKawaguchiM. Postoperative intensive care unit admission does not affect outcomes in elective surgical patients with severe comorbidity. Med Intensiva Engl Ed. 2020;44(4):216‐225. doi:https://doi.org/10.1016/j.medine.2019.01.012
6.
TranAMatarMLampronJSteyerbergETaljaardMVaillancourtC. Early identification of patients requiring massive transfusion, embolization or hemostatic surgery for traumatic hemorrhage: A systematic review and meta-analysis. J Trauma Acute Care Surg. 2018;84(3):505. doi:https://doi.org/10.1097/TA.0000000000001760
7.
ParkYHRyuDHLeeBKLeeDH. The association between the initial lactate level and need for massive transfusion in severe trauma patients with and without traumatic brain injury. Acute Crit Care. 2019;34(4):255‐262. doi:https://doi.org/10.4266/acc.2019.00640.
BrookeMYeungLMiraflorEGarciaAVictorinoGP. Lactate predicts massive transfusion in hemodynamically normal patients. J Surg Res. 2016;204(1):139‐144. doi:https://doi.org/10.1016/j.jss.2016.04.015.
10.
ShresthaMPBorgstromMTrowersEA. Elevated lactate level predicts intensive care unit admissions, endoscopies and transfusions in patients with acute gastrointestinal bleeding. Clin Exp Gastroenterol. 2018;11(eCollection):185‐192. doi:https://doi.org/10.2147/CEG.S162703
11.
LarsenJBHvasAM. Predictive value of whole blood and plasma coagulation tests for intra- and postoperative bleeding risk: A systematic review. Semin Thromb Hemost. 2017;43(7):772‐805. doi:https://doi.org/10.1055/s-0037-1602665
12.
MadbakFPriceDSkarupaD, et al.Serial hemoglobin monitoring in adult patients with blunt solid organ injury: Less is more. Trauma Surg Acute Care Open. 2020;5(1):e000446. doi:https://doi.org/10.1136/tsaco-2020-000446
13.
MadsenTDawsonMBledsoeJBossartP. Serial hematocrit testing does not identify major injuries in trauma patients in an observation unit. Am J Emerg Med. 2010;28(4):472‐476. doi:https://doi.org/10.1016/j.ajem.2009.01.034
14.
ZehtabchiSSinertRGoldmanMKapitanyanRBallasJ. Diagnostic performance of serial haematocrit measurements in identifying major injury in adult trauma patients. Injury. 2006;37(1):46‐52. doi:https://doi.org/10.1016/j.injury.2005.09.015
15.
ShahiVShahiVMowerWR. Using serial hemoglobin levels to detect occult blood loss in the early evaluation of blunt trauma patients. J Emerg Med. 2018;55(3):307‐312. doi:https://doi.org/10.1016/j.jemermed.2018.06.017.
16.
BrunsBLindseyMRoweK, et al.Hemoglobin drops within minutes of injuries and predicts need for an intervention to stop hemorrhage. J Trauma. 2007;63(2):312‐315. doi:https://doi.org/10.1097/TA.0b013e31812389d6
17.
ThorsonCMRyanMLVan HarenRM, et al.Change in hematocrit during trauma assessment predicts bleeding even with ongoing fluid resuscitation. Am Surg. 2013;79(4):398‐406.
18.
TongletML. Early prediction of ongoing hemorrhage in severe trauma: Presentation of the existing scoring systems. Arch Trauma Res. 2016;5(4):e33377. doi:https://doi.org/10.5812/atr.33377
19.
BrockampTNienaberUMutschlerM, et al.Predicting on-going hemorrhage and transfusion requirement after severe trauma: A validation of six scoring systems and algorithms on the TraumaRegister DGU®. Crit Care. 2012;16(4):R129. doi:https://doi.org/10.1186/cc11432
20.
YangSMackenzieCFRockP, et al.Comparison of massive and emergency transfusion prediction scoring systems after trauma with a new bleeding risk Index score applied in-flight. J Trauma Acute Care Surg. 2021;90(2):268‐273. doi:https://doi.org/10.1097/TA.0000000000003031
21.
SchrollRSwiftDTatumD, et al.Accuracy of shock index versus ABC score to predict need for massive transfusion in trauma patients. Injury. 2018;49(1):15‐19. doi:https://doi.org/10.1016/j.injury.2017.09.015
22.
CarsettiAAntoliniRCasarottaE, et al.Shock index as predictor of massive transfusion and mortality in patients with trauma: A systematic review and meta-analysis. Crit Care. 2023;27(1):85. doi:https://doi.org/10.1186/s13054-023-04386-w
23.
GianolaSCastelliniGBiffiA, et al.Accuracy of risk tools to predict critical bleeding in major trauma: A systematic review with meta-analysis. J Trauma Acute Care Surg. 2022;92(6):1086. doi:https://doi.org/10.1097/TA.0000000000003496
24.
RobertsDJBobrovitzNZygunDA, et al.Indications for use of damage control surgery in civilian trauma patients: A content analysis and expert appropriateness rating study. Ann Surg. 2016;263(5):1018. doi:https://doi.org/10.1097/SLA.0000000000001347
25.
RobertsDJBobrovitzNZygunDA, et al.Indications for use of damage control surgery and damage control interventions in civilian trauma patients: A scoping review. J Trauma Acute Care Surg. 2015;78(6):1187. doi:https://doi.org/10.1097/TA.0000000000000647
26.
WatsonJJJNielsenJHartK, et al.Damage control laparotomy utilization rates are highly variable among level I trauma centers: Pragmatic, randomized optimal platelet and plasma ratios findings. J Trauma Acute Care Surg. 2017;82(3):481. doi:https://doi.org/10.1097/TA.0000000000001357
27.
KushimotoSAraiMAiboshiJ, et al.The role of interventional radiology in patients requiring damage control laparotomy. J Trauma Acute Care Surg. 2003;54(1):171.
28.
KaoFCChangYCChenTSLiuPHTuYK. Risk factors for unplanned return to the operating room within 24 h: A 9-year single-center observational study. Medicine (Baltimore). 2021;100(49):e28053. doi:https://doi.org/10.1097/MD.0000000000028053
McIntyreLKSchiffMJurkovichGJ. Failure of nonoperative management of splenic injuries: Causes and consequences. Arch Surg. 2005;140(6):563‐569. doi:https://doi.org/10.1001/archsurg.140.6.563
31.
McKinneyW. Data Structures for Statistical Computing in Python. In: ; 2010:56-61. doi:10.25080/Majora-92bf1922-00a.
32.
R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing; 2018.
33.
De PasqualeMMossTJCeruttiS, et al.Hemorrhage prediction models in surgical intensive care: Bedside monitoring data adds information to lab values. IEEE J Biomed Health Inform. 2017;21(6):1703‐1710. doi:https://doi.org/10.1109/JBHI.2017.2653849.
34.
VirtanenPGommersROliphantTE. Scipy 1.0: Fundamental algorithms for scientific computing in python. Nat Methods. 2020;17(3):261-272. doi:https://doi.org/10.1038/s41592-019-0686-2
35.
KangSParkCLeeJYoonD. Machine learning model for the prediction of hemorrhage in intensive care units. Healthc Inform Res. 2022;28(4):364‐375. doi:https://doi.org/10.4258/hir.2022.28.4.364.
36.
LeeSWKungHCHuangJF, et al.The clinical application of machine learning-based models for early prediction of hemorrhage in trauma intensive care units. J Pers Med. 2022;12(11):1901. doi:https://doi.org/10.3390/jpm12111901.
37.
LauzierFArnoldDRabbatC, et al.In medical-surgical ICU patients, major bleeding is common but independent of heparin prophylaxis. Crit Care. 2012;16(1):P435. doi:https://doi.org/10.1186/cc11042.
38.
LukiesMGipsonJTanSYClementsW. Spontaneous retroperitoneal haemorrhage: Efficacy of conservative management and embolisation. Cardiovasc Intervent Radiol. 2023;46(4):488‐495. doi:https://doi.org/10.1007/s00270-023-03359-4.
39.
KohnJRDildyGAEppesCS. Shock index and delta-shock index are superior to existing maternal early warning criteria to identify postpartum hemorrhage and need for intervention. J Matern Fetal Neonatal Med. 2019;32(8):1238‐1244. doi:https://doi.org/10.1080/14767058.2017.1402882.
40.
KimMJLeeJGLeeSH. Factors predicting the need for hemorrhage control intervention in patients with blunt pelvic trauma: A retrospective study. BMC Surg. 2018;18(1):101. doi:https://doi.org/10.1186/s12893-018-0438-8.
41.
DeMuroJPSimmonsSJaxJGianelliSM. Application of the shock index to the prediction of need for hemostasis intervention. Am J Emerg Med. 2013;31(8):1260‐1263. doi:https://doi.org/10.1016/j.ajem.2013.05.027.
42.
LiuNTHolcombJBWadeCE, et al.Development and validation of a machine learning algorithm and hybrid system to predict the need for life-saving interventions in trauma patients. Med Biol Eng Comput. 2014;52(2):193‐203. doi:https://doi.org/10.1007/s11517-013-1130-x.
43.
LeeSMLeeGKimTK, et al.Development and validation of a prediction model for need for massive transfusion during surgery using intraoperative hemodynamic monitoring data. JAMA Netw Open. 2022;5(12):e2246637. doi:https://doi.org/10.1001/jamanetworkopen.2022.46637.
44.
PustavoitauARizkallaNAPerlsteinB, et al.Validation of predictive models identifying patients at risk for massive transfusion during liver transplantation and their potential impact on blood bank resource utilization. Transfusion (Paris). 2020;60(11):2565‐2580. doi:https://doi.org/10.1111/trf.16019.
45.
de HondAAHLeeuwenbergAMHooftL, et al.Guidelines and quality criteria for artificial intelligence-based prediction models in healthcare: A scoping review. Npj Digit Med. 2022;5(1):1‐13. doi:https://doi.org/10.1038/s41746-021-00549-7
46.
MeyerAZverinskiDPfahringerB, et al.Machine learning for real-time prediction of complications in critical care: A retrospective study. Lancet Respir Med. 2018;6(12):905‐914. doi:https://doi.org/10.1016/S2213-2600(18)30300-X
47.
BondeAVaradarajanKMBondeN, et al.Assessing the utility of deep neural networks in predicting postoperative surgical complications: A retrospective study. Lancet Digit Health. 2021;3(8):e471‐e485. doi:https://doi.org/10.1016/S2589-7500(21)00084-4
48.
ShillanDSterneJACChampneysAGibbisonB. Use of machine learning to analyse routinely collected intensive care unit data: A systematic review. Crit Care. 2019;23(1):284. doi:https://doi.org/10.1186/s13054-019-2564-9
49.
CrochemoreTGörlingerKLanceMD. Early goal-directed hemostatic therapy for severe acute bleeding management in the intensive care unit: A narrative review. Anesth Analg. 2024;138(3):499‐513. doi:https://doi.org/10.1213/ANE.0000000000006756
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
