Biomarkers of disseminated intravascular coagulation in pediatric intensive care unit in Thailand

Interpretable machine learning model for early prediction of disseminated intravascular coagulation in critically ill children

Disseminated intravascular coagulation (DIC) is a thrombo-hemorrhagic disorder that can be life-threatening in critically ill children, and the quest for an accurate and efficient method for early DIC prediction is of paramount importance. Candidate predictors encompassed demographics, comorbidities, laboratory findings, and therapy strategies. A stepwise logistic regression model was employed to select the features included in the final model. Six machine learning algorithms-logistic regression (LR), extreme gradient boosting (XGB), random forest (RF), support vector machine (SVM), decision tree (DT), and k-nearest neighbor (KNN)-were employed to construct predictive models for DIC in critically ill children. Models were then evaluated by using area under the curve (AUC), accuracy, specificity, sensitivity, positive predictive value (PPV), negative predictive value (NPV), precision, recall and decision curve analysis (DCA). Interpretation of the optimal model was conducted using shapley additive explanations (SHAP). A total of 6093 critically ill children were encompassed in this study, of whom 681 (11.2%) developed DIC. The RF model exhibited the highest levels of accuracy (0.856), sensitivity (0.866), Kappa (0.472), NPV (0.423), and recall (0.866). However, the XGB model outperformed RF, LR, SVM, DT, and KNN in terms of AUC (0.908), specificity (0.859), PPV (0.978), and precision (0.969). Decision curve analysis (DCA) confirmed the superior clinical utility of the XGB model. Overall, the XGB model demonstrated superior clinical utility compared to RF, LR, SVM, DT, and KNN. We named the final model Alfalfa-PICU-DIC. SHAP analysis identified D-dimer, INR, PT, TT, and PLT count as the top predictors of DIC. Machine learning models can be a reliable tool for predicting DIC in critically ill children, which will facilitate timely intervention, thereby reducing the burden of DIC on patients in the pediatric intensive care unit (PICU).

Predicting relapsed/refractory disease in childhood hemophagocytic lymphohistiocytosis based on clinical features at diagnosis: A 13-year single-institute retrospective study in Thailand

Hemophagocytic lymphohistiocytosis (HLH) is a life-threatening disease. Relapsed/refractory disease is the main cause of death. This study aims to determine the prognostic indicators for relapsed/refractory disease in childhood HLH (R/R HLH). Infants and children under 18 years of age who were diagnosed with HLH according to HLH-2004 criteria, MAS-HLH criteria for rheumatologic diseases, or H-score undergoing treatment in Chiang Mai University hospital between 2010 and 2022 were included. Demographic data, clinical characteristics, and laboratory parameters were retrospectively reviewed. Out of 86 childhood HLH cases, 30 patients (34.9%) experienced R/R HLH. All patients with primary HLH developed R/R HLH. The most common form of secondary HLH was infection-associated hemophagocytic syndrome (IAHS), comprising 43 cases. Of these, 37.2% had relapsed or refractory disease. Univariable analysis identified several potential risk factors for R/R HLH, including younger age, severe disease status, higher HLH-2004 criteria scores, higher H-scores, overt DIC, higher pSOFA scores, and increased levels of aspartate aminotransferase, total bilirubin, and direct bilirubin. Multivariable logistic regression analysis revealed that a pSOFA score of ≥ 8 and age < 3 years were independent risk factors for R/R HLH, with adjusted odds ratios of 6.35 (95% confidence interval [CI], 1.18-34.19; P = 0.032) and 3.62 (95% CI, 1.04-12.63; P = 0.044), respectively. Children with HLH who have a pSOFA score of ≥ 8, or are younger than 3 years, are at a higher risk of relapsed or refractory disease. Further evaluation of management strategies in this context is warranted.

Rotational Thromboelastometry and Clot Waveform Analysis as Point-of-Care Tests for Diagnosis of Disseminated Intravascular Coagulation in Critically Ill Children in Thailand

OBJECTIVES

This study aimed to determine the test performances of rotational thromboelastometry (ROTEM) and activated partial thromboplastin time-based clot waveform analysis (aPTT-CWA) compared with the International Society on Thrombosis and Hemostasis disseminated intravascular coagulation (ISTH-DIC) score for diagnosis of overt disseminated intravascular coagulation (ODIC) in critically ill children. Prognostic indicators of DIC complications were also evaluated.

DESIGN

A prospective cross-sectional observational study was conducted. ROTEM and aPTT-CWA were assessed alongside standard parameters based on the ISTH-DIC score and natural anticoagulants. Both conventional and global hemostatic tests were repeated on days 3-5 for nonovert DIC.

SETTING

PICU of the Department of Pediatrics, Faculty of Medicine, Chiang Mai University, Chiang Mai, Thailand.

SUBJECTS

Infants and children who were admitted to PICU with underlying diseases predisposed to DIC, such as sepsis, malignancy, major surgery, trauma, or severe illness, were included in the study between July 1, 2021, and November 30, 2022.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Sixty-four children were enrolled in this study. The prevalence of ODIC was 20.3%. Regarding ROTEM parameters, using EXTEM clot formation time (CFT) cutoff of greater than 102 seconds provided sensitivity and specificity of 90.9% and 80.9%, respectively, for diagnosing ODIC, with the area under the curve (AUC) of 0.86. In the case of aPTT-CWA performance, no biphasic waveform was observed, whereas both maximum coagulation acceleration (Min2) of less than 0.35%/s 2 and maximum coagulation deceleration of less than 0.25%/s 2 demonstrated identical sensitivities of 76.9% and specificities of 79.6%. Combining two global hemostatic tests significantly improved the diagnostic performance (INTEM CFT + EXTEM CFT + Min2 AUC 0.92 [95% CI, 0.80-1.00] vs. EXTEM CFT AUC 0.86 [95% CI, 0.75-0.96], p = 0.034). Bleeding was the most common consequence. In multivariable logistic regression analysis, Min2 of less than 0.36%/s 2 was an independent risk factor for bleeding complications, with an adjusted odds ratio of 15.08 (95% CI, 1.08-211.15, p = 0.044).

CONCLUSIONS

ROTEM and aPTT-CWA were valuable diagnostic tools in critically ill children who might require point-of-care tests. Min2 showed significant clinical implications for predicting bleeding events in this population.

Prevalence and risk factors of disseminated intravascular coagulation in childhood acute lymphoblastic leukemia

BACKGROUND

Few studies have examined disseminated intravascular coagulation (DIC) in childhood acute lymphoblastic leukemia (ALL). Our aims were to evaluate the prevalence, risk factors and outcomes of DIC at ALL presentation and during induction chemotherapy.

METHODS

The medical records of ALL patients aged <15 years were retrospectively reviewed. Logistic regression analysis was used to identify risk factors. The Kaplan-Meier method was used to depict survival.

RESULTS

Of the 312 patients, 48 (15.4%) and 76 (24.4%) had DIC at presentation and during induction chemotherapy, respectively. Risk factors for DIC at presentation (OR and 95% CI) were antibiotics prior to admission 2.34 (1.17-4.89), white blood cell count ≥100 × 10⁹/L 2.39 (1.04-5.72), platelets <100 × 10⁹/L 5.44 (1.84-23.4) and high National Cancer Institute (NCI) risk 2.68 (1.08-6.62). Risk factors for DIC during induction chemotherapy were antibiotics prior to admission 1.86 (1.07-3.27), high peripheral blasts 1.01 (1.00-1.02) and transaminitis 2.02 (1.18-3.48). Five-year overall survival of patients who had DIC was significantly lower than those who did not (45.0% vs. 74.1%, p <0.001).

CONCLUSION

Antibiotics prior to admission, hyperleukocytosis, thrombocytopenia and high NCI risk were risk factors of DIC at presentation. Antibiotics prior to admission, high peripheral blasts and transaminitis were risk factors of DIC during induction chemotherapy.

IMPACT

There are only two studies, both published before 2000, evaluating risk factors of DIC in pediatric ALL patients without reporting outcomes. DIC was associated with lower remission and survival rates in pediatric ALL patients. We identified the risk factors of DIC at presentation as antibiotics prior to admission, hyperleukocytosis, thrombocytopenia and high NCI risk. The risk factors of DIC during induction chemotherapy were antibiotics prior to admission, high peripheral blasts and aspartate transaminitis. Pediatric ALL patients who have the aforementioned risk factors should be closely monitored for DIC secondary to infection, and early treatment with appropriate antimicrobial agents is recommended.

Disseminated intravascular coagulation: new identity as endotheliopathy-associated vascular microthrombotic disease based on in vivo hemostasis and endothelial molecular pathogenesis

Disseminated intravascular coagulation (DIC) can be correctly redefined as disseminated intravascular microthrombosis based on "two-path unifying theory" of in vivo hemostasis. "DIC" is a form of vascular microthrombotic disease characterized by "microthrombi" composed of platelets and unusually large von Willebrand factor multimers (ULVWF). Microthrombotic disease includes not only "DIC", but also microthrombosis occurring in thrombotic thrombocytopenic purpura (TTP), TTP-like syndrome, and focal, multifocal and localized microthrombosis. Being a hemostatic disease, microthrombotic disease occurs as a result of lone activation of ULVWF path via partial in vivo hemostasis. In endothelial injury associated with critical illnesses such as sepsis, the vascular damage is limited to the endothelial cell and activates ULVWF path. In contrast, in intravascular traumatic injury, the local damage may extend from the endothelial cell to subendothelial tissue and sometimes beyond, and activates both ULVWF and tissue factor (TF) paths. When endotheliopathy triggers exocytosis of ULVWF and recruits platelets, ULVWF path is activated and promotes microthrombogenesis to produce microthrombi composed of microthrombi strings, but when localized vascular damage causes endothelial and subendothelial tissue damage, both ULVWF and TF paths are activated and promote macrothrombogenesis to produce macrothrombus made of complete "blood clots". Currently, "DIC" concept is ascribed to activated TF path leading to fibrin clots. Instead, it should be correctly redefined as microthrombosis caused by activation of ULVWF path, leading to endotheliopathy-associated microthrombosis. The correct term for acute "DIC" is disseminated microthrombosis-associated hepatic coagulopathy, and that for chronic "DIC" is disseminated microthrombosis without hepatic coagulopathy. TTP-like syndrome is hematologic phenotype of endotheliopathy-associated microthrombosis. This correct concept of "DIC" is identified from novel theory of "in vivo hemostasis", which now can solve every mystery associated with "DIC" and other associated thrombotic disorders. Thus, sepsis-associated coagulopathy is not "DIC", but is endotheliopathy-associated vascular microthrombotic disease.