Coagulation abnormalities as predictors of mortality, disseminated intravascular coagulation, and multiple organ dysfunction in children with septic shock

¹Pediatric Intensive Care Unit, Nghe An Obstetrics and Pediatrics Hospital, Vinh City, Nghe An, Vietnam, ²Pediatrics Department, Thai Nguyen University of Medicine and Pharmacy, Thai Nguyen City, Thai Nguyen Province, Vietnam, ³Pediatrics Department, Haiphong University of Medicine and Pharmacy, Haiphong, Vietnam, ⁴Clinical Department, Vietnam National Children's Hospital, Hanoi, Vietnam, ⁵Department of Adolescent Health, Vietnam National Children's Hospital, Hanoi, Vietnam

^&Corresponding author
Ngo Anh Vinh, Department of Adolescent Health, Vietnam National Children's Hospital, Hanoi, Vietnam

Introduction    

Pediatric septic shock remains a major global health challenge, contributing substantially to childhood mortality, long-term morbidity, and health-care resource utilisation worldwide. Pooled global estimates indicate that the case-fatality rate of pediatric severe sepsis and septic shock is approximately 25%, with a marked disparity between developing and developed countries (31.7% vs. 19.3%) [1]. Septic shock confers a significantly higher risk of death compared with severe sepsis alone, with an adjusted odds ratio of 1.47 [1]. Multicenter studies from diverse settings have reported mortality rates ranging from 19.8% in Brazilian pediatric intensive care units to over 40% in high-risk populations such as children with cancer or those developing shock within general sepsis cohorts [2-4]. Beyond in-hospital mortality, pediatric septic shock is associated with persistent organ dysfunction and long-term impairment in quality of life among survivors, highlighting its broad clinical and societal impact [5].

Clinical severity in pediatric septic shock is closely linked to the extent of organ dysfunction and the intensity of cardiovascular support. Systematic evaluations by the Pediatric Sepsis Definition Taskforce have consistently identified multi-organ dysfunction, PELOD and PELOD-2 scores, PRISM, and the requirement for vasoactive or inotropic agents as strong predictors of mortality [6]. The Vasoactive-Inotropic Score (VIS) has emerged as a pragmatic marker of cardiovascular failure, with VIS values above 20 repeatedly associated with increased mortality across pediatric septic shock cohorts [6,7]. Each additional organ failure further increases the risk of death, particularly in vulnerable subgroups such as children with complex chronic conditions or malignancies [4,7]. These severity variables reflect the systemic nature of septic shock and provide an essential clinical context for risk stratification and outcome prediction.

Sepsis-associated coagulation and fibrinolysis abnormalities represent a critical pathophysiological axis linking inflammation, endothelial injury, and organ failure. Evidence synthesis indicates that lower platelet count and reduced fibrinogen levels are consistently associated with increased mortality in pediatric sepsis and septic shock [6]. Observational studies have demonstrated that prolonged activated partial thromboplastin time, elevated INR, and low fibrinogen identify children at particularly high risk of death, while fibrinogen levels below critical thresholds are associated with extremely poor outcomes [2]. Disseminated intravascular coagulation is a frequent and severe complication, with pooled mortality estimates exceeding 40% among DIC-positive septic patients, including pediatric septic shock cohorts [8]. Routine coagulation parameters-platelet count, PT/INR, aPTT, fibrinogen, and D-dimer-are therefore increasingly recognised as accessible and clinically meaningful prognostic tools for predicting mortality and organ dysfunction in sepsis [6,9,10].

Despite the growing international evidence base, data on the prognostic value of coagulation parameters in pediatric septic shock remain limited in low- and middle-income countries, including Viet Nam. Differences in health-care resources, disease epidemiology, and patient characteristics may influence both the prevalence of sepsis-associated coagulopathy and its relationship with clinical outcomes. Generating local evidence is therefore essential to inform context-appropriate risk stratification and clinical decision-making. This study aims to evaluate the prognostic value of basic coagulation indices and D-dimer for mortality, disseminated intravascular coagulation, and multiple organ dysfunction among Vietnamese children with septic shock, addressing a critical evidence gap in regional pediatric critical care.

Methods     

Study design: this was an observational ambispective cohort study combining retrospective and prospective data collection to evaluate coagulation abnormalities and clinical outcomes in pediatric patients with septic shock.

Setting: the study was conducted in the Pediatric Intensive Care Unit of Nghe An Obstetrics and Pediatrics Hospital, Vietnam, over 24 months from September 2023 to August 2025. All clinical and laboratory assessments were performed as part of routine intensive care management within the unit.

Participants: pediatric patients aged 1 month to 15 years who were diagnosed with septic shock and treated in the unit during the study period were eligible for inclusion. Septic shock was defined according to the Phoenix Sepsis Criteria as sepsis with cardiovascular dysfunction manifested by severe age-adjusted hypotension, blood lactate >5 mmol/L, or requirement for vasoactive or inotropic support. Patients were excluded if they had received anticoagulant therapy before or during hospitalisation or had a known congenital coagulation disorder. Consecutive eligible patients were enrolled using a convenience sampling approach, resulting in a total cohort of 59 pediatric patients with septic shock.

Variables: the primary outcomes were in-hospital mortality, disseminated intravascular coagulation (DIC), and multiple organ dysfunction syndrome (MODS), defined as dysfunction of more than two organ systems. Covariates included demographic characteristics, comorbidities, number of organ failures, vasoactive-inotropic score at 24 hours, and Pediatric Logistic Organ Dysfunction-2 (PELOD-2) score. Coagulation variables comprised platelet count, prothrombin time, international normalised ratio, activated partial thromboplastin time, fibrinogen concentration, and D-dimer level.

Data sources and measurement: for retrospectively included patients, clinical and laboratory data were extracted from medical records, whereas prospectively enrolled patients were evaluated using a standardised case report form. Laboratory parameters were obtained from routine coagulation testing performed in the hospital laboratory. DIC status and organ dysfunction were determined according to routine clinical and laboratory assessments conducted in the intensive care unit. To improve consistency between retrospective and prospective phases, the same diagnostic criteria, laboratory methods, and variable definitions were applied throughout the study period.

Bias: selection bias was minimised by enrolling all consecutive eligible patients during the study period and applying standardised diagnostic criteria for septic shock. Information bias was reduced through the use of routine laboratory measurements and predefined clinical scoring systems. However, the ambispective design may have introduced potential bias related to differences in data completeness, timing of measurement, and documentation quality between retrospectively and prospectively collected cases. In particular, retrospectively collected data depended on the accuracy and completeness of medical records, whereas prospective data collection allowed more standardised assessment using predefined case report forms. The single-centre setting may also limit the generalizability of the findings.

Study size: the study sample comprised all eligible pediatric septic shock cases admitted during the study period without prior sample size calculation, yielding a final analytical cohort of 59 patients.

Quantitative variables: continuous variables, including laboratory parameters and clinical scores, were summarised as medians with interquartile ranges due to non-normal distribution, while categorical variables were expressed as frequencies and percentages. Coagulation parameters and D-dimer were treated as continuous predictors in receiver operating characteristic analyses.

Statistical methods: data were analysed using SPSS version 20.0. Comparisons between survivors and non-survivors were performed using the chi-square or Fisher's exact test for categorical variables and the Mann-Whitney U test for continuous variables. Receiver operating characteristic curve analysis was used to evaluate the predictive performance of coagulation parameters and D-dimer for mortality, DIC, and MODS, with calculation of the area under the curve and 95% confidence intervals. Optimal cut-off values were determined based on the balance between sensitivity and specificity. A two-sided p-value <0.05 was considered statistically significant.

Ethical consideration statement: this observational non-interventional study did not alter standard clinical care and was approved by the Ethics Committee of Nghe An Obstetrics and Pediatrics Hospital. All patient data were anonymised and handled in accordance with institutional regulations to ensure confidentiality and protection of pediatric participants' rights.

Results     

Demographic characteristics: among 59 pediatric patients with septic shock, the median age was 16 months, with no significant age difference between survivors and non-survivors. The sex distribution was similar between groups, with males accounting for approximately two-thirds of cases. Comorbidities were present in 27.1% of patients and did not differ significantly by survival status. A higher proportion of non-survivors had more than two organ failures compared with survivors (85.7% vs. 64.5%, p = 0.05). Median vasoactive-inotropic score at 24 hours and PELOD-2 score were significantly higher in non-survivors than in survivors. Overall mortality was 47.5% (Table 1).

Descriptive statistics: non-survivors had significantly lower platelet counts and prothrombin time percentages, and significantly higher INR and activated partial thromboplastin time compared with survivors. Fibrinogen levels and D-dimer concentrations tended to differ between groups but did not reach statistical significance. Disseminated intravascular coagulation was markedly more frequent among non-survivors than survivors (71.4% vs. 29.0%, p = 0.020) (Table 2).

Diagnostic values analysis: ROC analysis demonstrated that prothrombin time and INR showed strong discriminatory ability for mortality, with AUCs of 0.899 and 0.813, respectively, and statistically significant p-values. Platelet count and fibrinogen showed moderate discrimination but did not reach statistical significance. APTT and D-dimer showed modest predictive performance, with lower AUCs and limited specificity (Table 3). All evaluated coagulation parameters demonstrated statistically significant predictive value for disseminated intravascular coagulation. Platelet count showed excellent discrimination with an AUC of 0.943, followed by INR and prothrombin time with high AUC values. Fibrinogen and D-dimer also demonstrated good predictive performance, with balanced sensitivity and specificity at the identified cut-off points (Table 4). For the prediction of multiple organ dysfunction involving more than two organs, prothrombin time, INR, and APTT demonstrated strong discriminatory ability, with AUCs exceeding 0.80 and statistically significant p-values. Platelet count and fibrinogen showed moderate predictive performance, while D-dimer demonstrated lower discrimination with reduced sensitivity compared with other parameters (Table 5).

Discussion     

This study demonstrates that abnormalities in routine coagulation parameters are closely associated with disease severity and adverse outcomes in pediatric septic shock. Coagulation indices reflecting impaired clot formation and activation of fibrinolysis showed clear prognostic relevance for mortality, disseminated intravascular coagulation, and multiple organ dysfunction, alongside established clinical severity markers. Collectively, the findings indicate that basic, widely available coagulation tests provide meaningful prognostic information beyond general demographic characteristics.

The present Receiver operating characteristic (ROC) analyses indicate that coagulation abnormalities provide clinically meaningful discrimination for mortality in pediatric septic shock, with prothrombin time and INR emerging as the strongest individual predictors. The high AUC for PT (0.899) and INR (0.813), together with favorable sensitivity-specificity trade-offs at pragmatic cut-offs, suggests that impairment of the extrinsic coagulation pathway and global clotting function closely tracks fatal disease trajectories. This aligns with prior pediatric shock cohorts in which prolonged PT/INR and thrombocytopenia were consistently associated with higher mortality, reflecting consumptive coagulopathy and endothelial injury in severe sepsis [11,12]. By contrast, platelet count and fibrinogen showed moderate discrimination for mortality, while D-dimer demonstrated limited specificity, a pattern also reported in earlier studies where fibrinolysis markers were sensitive to inflammation but less specific for death when considered in isolation [8,13].

For disseminated intravascular coagulation, all evaluated parameters demonstrated strong and statistically significant predictive performance, with platelet count showing excellent discrimination (AUC 0.943), followed by INR and PT. These findings are concordant with pediatric DIC literature in which conventional coagulation tests are markedly deranged in DIC-positive septic shock and form the core of ISTH-based scoring systems [12]. Prior studies have shown that higher ISTH DIC scores are strongly associated with mortality in children, with reported death rates of 50% at higher score thresholds and a 35% increase in death odds per one-point score increment [11]. Similar associations have been observed across emergency and intensive care settings, including higher vasopressor requirements and worse long-term survival among children with elevated DIC scores [14], as well as increased PICU mortality in pediatric hemato-oncology septic shock [13]. The good performance of fibrinogen and D-dimer for DIC prediction in the present study further supports their role as markers of advanced consumptive and fibrinolytic activation rather than early mortality per se.

Regarding multiple organ dysfunction, PT, INR, and APTT demonstrated strong discriminatory ability (AUCs >0.80), whereas platelet count and fibrinogen showed moderate performance, and D-dimer had lower sensitivity. This gradient mirrors the pathophysiological progression of sepsis, in which sustained coagulation pathway derangements accompany worsening organ failure. Previous work has shown that prolonged clotting times and persistent aPTT elevation parallel the development of organ dysfunction and interact with established severity indices such as VIS, PELOD-2, and pSOFA [11,15,16]. In low- and middle-income country settings, where advanced biomarkers are often unavailable, the present findings reinforce evidence that basic coagulation tests, used alongside clinical severity scores, offer a pragmatic, data-supported framework for early identification of children at high risk of DIC, multiple organ dysfunction, and death in septic shock [8,13,14].

From a clinical perspective, the findings support the use of basic coagulation parameters and D-dimer as accessible, low-cost tools for early risk stratification in pediatric septic shock, particularly in resource-limited settings. Incorporating these indices into routine assessment may aid clinicians in identifying high-risk patients, anticipating complications such as DIC or multiple organ dysfunction, and prioritising intensive monitoring and supportive interventions. Their widespread availability enhances their potential utility in daily practice without the need for specialised testing.

Several limitations should be acknowledged. The single-centre design and relatively small sample size may limit the generalizability of the findings. The observational nature of the study precludes causal inference, and unmeasured confounders may have influenced the observed associations. Additionally, coagulation parameters were assessed at limited time points, preventing evaluation of dynamic changes over the disease course, which may further refine prognostic accuracy. The relatively small sample size may also have contributed to the overestimation of ROC-derived performance measures, particularly AUC values. Furthermore, confidence intervals for sensitivity and specificity and internal validation methods such as bootstrapping were not included in the present analysis. Future studies with larger multicenter cohorts and internal validation approaches are needed to confirm the robustness and reproducibility of these predictive findings.

Conclusion     

In conclusion, this study provides evidence that routine coagulation indices and D-dimer are closely associated with mortality, disseminated intravascular coagulation, and multiple organ dysfunction in pediatric septic shock. These readily available markers complement established clinical severity measures and offer valuable prognostic insight. Integrating coagulation assessment into the early evaluation of pediatric septic shock may enhance risk stratification and support timely clinical decision-making, particularly in settings where advanced biomarkers are not readily accessible.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Bui Thi Huong: study conception and design, patient recruitment, data collection, and manuscript drafting. Nguyen Hung Manh: clinical data acquisition and database management. Nguyen Thi Quynh Trang: data collection and preliminary analysis. Le Thi Kim Dung: methodological support and data validation. Hoang Thi Hue: statistical analysis and interpretation. Chu Thi Ha: literature review and manuscript editing. Nguyen Thi Xuan Huong: methodological consultation and critical revision. Ngo Anh Vinh: study supervision, conceptual guidance, and critical revision of the manuscript. All authors read and approved the final manuscript.

Acknowledgments     

The authors thank the medical and nursing staff of the Pediatric Intensive Care Unit, Nghe, an Obstetrics and Pediatrics Hospital, for their support in patient care and assistance with data collection.

Tables     

Table 1: baseline characteristics of the study population (n = 59)

Table 2: comparison of basic coagulation parameters and D-dimer between survivors and non-survivors

Table 3: predictive value of coagulation parameters for mortality in patients with septic shock (ROC analysis, N = 59)

Table 4: predictive performance of basic coagulation parameters and D-dimer for disseminated intravascular coagulation (DIC) (ROC analysis, N = 59)

Table 5: predictive value of basic coagulation parameters and D-dimer for multiple organ dysfunction (>2 organs)

References