Venous Thromboembolism in Critical Illness and Trauma: Pediatric Perspectives

Derivation and Validation of a Machine Learning Algorithm for Predicting Venous Thromboembolism in Injured Children

BACKGROUND

Venous thromboembolism (VTE) causes significant morbidity in pediatric trauma patients. We applied machine learning algorithms to the Trauma Quality Improvement Program (TQIP) database to develop and validate a risk prediction model for VTE in injured children.

METHODS

Patients ≤18 years were identified from TQIP (2017-2019, n = 383,814). Those administered VTE prophylaxis ≤24 h and missing the outcome (VTE) were removed (n = 347,576). Feature selection identified 15 predictors: intubation, need for supplemental oxygen, spinal injury, pelvic fractures, multiple long bone fractures, major surgery (neurosurgery, thoracic, orthopedic, vascular), age, transfusion requirement, intracranial pressure monitor or external ventricular drain placement, and low Glasgow Coma Scale score. Data was split into training (n = 251,409) and testing (n = 118,175) subsets. Machine learning algorithms were trained, tested, and compared.

RESULTS

Low-risk prediction: For the testing subset, all models outperformed the baseline rate of VTE (0.15%) with a predicted rate of 0.01-0.02% (p < 2.2e^-16). 88.4-89.4% of patients were classified as low risk by the models.

HIGH-RISK PREDICTION

All models outperformed baseline with a predicted rate of VTE ranging from 1.13 to 1.32% (p < 2.2e^-16). The performance of the 3 models was not significantly different.

CONCLUSION

We developed a predictive model that differentiates injured children for development of VTE with high discrimination and can guide prophylaxis use.

LEVEL OF EVIDENCE

Prognostic, Level II.

TYPE OF STUDY

Retrospective, Cross-sectional.

Incidence of Deep Vein Thrombosis in Cerebral Palsy Following an Orthopaedic Surgical Event

PURPOSE

The purpose of this study was to identify the incidence of venous thromboembolism (VTE) and characterize the demographics, comorbidities, and risk factors for patients with cerebral palsy (CP) having orthopaedic surgery.

METHODS

All patients diagnosed with CP who underwent an orthopaedic surgical procedure at one institution between 2008 and 2017 were identified. Diagnosis codes and associated patient events were recovered from the electronic medical record. Each VTE event was reviewed to ascertain an actual VTE episode related to a surgical event.

RESULTS

The review included 2583 orthopaedic surgical events in 1371 patients. Of the initial 88 cases identified, 28 cases had a deep thrombosis documented. Six cases of VTE occurred within 3 months following the surgical event. Three of these cases had thigh thrombosis, and 2 patients had upper arm thrombosis, and 1 patient had a superior vena cava thrombosis. On further workup, 5 of these 6 patients were identified as having a congenital hypercoagulable condition.

CONCLUSIONS

VTE is a relatively rare occurrence after orthopaedic surgery in pediatric patients with CP, but when it occurs, a full hematologic workup for a congenital hypercoagulable condition is indicated. Based on the low incidence of thigh thrombosis, routine pharmacological or intermittent mechanical calf compression is not recommended. A careful clinical and family history should be performed to identify patients with possible genetic hypercoagulable conditions who would merit prophylaxis.

LEVEL OF EVIDENCE

Level IV.

Incidence of venous thromboembolic events not related to vascular catheters in a prospective cohort of critically ill children

The risk for venous thromboembolism (VTE) is considered to be low in the general paediatric intensive care unit (PICU) population, and pharmacological thromboprophylaxis is not routinely used. PICU patients considered at high-risk of VTE could possibly benefit from pharmacological thromboprophylaxis, but the incidence of VTE in this group of patients is unclear. This was an observational, prospective study at a tertiary multi-disciplinary paediatric hospital. We used comprehensive ultrasonography screening for VTE in critically ill children with multiple risk factors for VTE. Patients admitted to PICU ≥ 72 h and with ≥ two risk factors for VTE were included. Patients receiving pharmacological thromboprophylaxis during their entire PICU stay were excluded. The primary outcome of the study was VTEs not related to the use of a CVC. Ultrasonography screening of the great veins was performed at PICU discharge. Seventy patients with median (interquartile range) 3 (2-4) risk factors for VTE were evaluated. Median age was 0.3 years (0.03-4.3) and median PICU length of stay 9 days (5-17). Regarding the primary outcome, no symptomatic VTEs occurred and no asymptomatic VTEs were found on ultrasonography screening, resulting in an incidence of VTEs not related to a vascular catheter of 0% (95% CI: 0-5.1%).

CONCLUSION

Our results indicate that VTEs not related to a vascular catheter are a rare event even in a selected group of severely ill small children considered to be at high risk of VTE.

WHAT IS KNOWN

• Children in the PICU often have several risk factors for venous thromboembolism (VTE). • The incidence of VTE in PICU patients is highly uncertain, and there are no evidence-based guidelines regarding VTE prophylaxis.

WHAT IS NEW

• This study found an incidence of VTEs not related to a vascular catheter of 0% (95% CI: 0-5.1%). • This indicates that such VTE events are rare even in PICU patients with multiple risk factors for VTE.

Implementation of automatic data extraction from an enterprise database warehouse (EDW) for validating pediatric VTE decision rule: a prospective observational study in a critical care population

Multiple clinical risk prediction tools for hospital acquired venous thromboembolism (HA-VTE) have been developed. The objectives of this study were to develop and assess the feasibility of data extraction from Electronic Medical Records (EMR) from an enterprise database warehouse (EDW) and to test the validity of a previously developed Pediatric Clot Decision Rule (PCDR). This single-center prospective observational cohort study was conducted between March 2016 and March 2017 and included eligible patients admitted to the intensive care units. Risk score was calculated using the PCDR tool. Sensitivity, specificity, positive and negative predicted value (PPV and NPV) were calculated based on a cut-point of 3. A total of 2822 children were eligible for analysis and 5.1% (95% CI 4.2-6.2) children had a PCDR score of 3. Children with PCDR score of ≥ 3 had a 3 times higher odd of developing VTE compared to those with scores < 3 (OR 3.1; 95% CI 1.93-4.80; p < 0.001). The model performance showed that at the cutoff point of ≥ 3, both the specificity and sensitivity of the PCDR in predicting VTE was 69% and NPV of 98%. We successfully demonstrated using our EDW to populate a research database using an automatic data import. A PCDR score of ≥ 3 was associated with VTE. Collaboration through large registries will be useful in informing practices and guidelines for rare disorders such as pediatric VTE.

Venous thromboembolism prophylaxis after pediatric trauma

In recent years, there has been an increased focus on developing and validating venous thromboprophylaxis guidelines in the pediatric trauma population. We review the current literature regarding the incidence of and risk factors for venous thromboembolism (VTE) and the use of prophylaxis in the pediatric trauma population. Risk factors such as age, injury severity, central venous catheters, mental status, injury type, surgery, and comorbidities can lead to a higher incidence of VTE. Risk stratification tools have been developed to determine whether mechanical and/or pharmacologic prophylaxis should be implemented depending on the degree of VTE risk. When VTE risk is high, pharmacologic prophylaxis, such as with low molecular weight heparin, is often initiated. However, the timing and duration of VTE prophylaxis is dependent on patient factors including ambulatory status and contraindications such as bleeding. In addition, the utility of screening ultrasound for VTE surveillance has been evaluated and though they are not widely recommended, no formal guidelines exist. While more research has been done in recent years to assess the most appropriate type, timing, and duration of VTE prophylaxis, further studies are warranted to create optimal guidelines for decreasing the risk of VTE after pediatric trauma.

Risk Assessment and Outcome of Venous Thromboembolism in Pediatric Population in an Academic Care Center of a Low-Middle Income Country

Venous thromboembolism (VTE) is a recognized complication of hospital stay in young patients in many developed countries, but such an information is largely unavailable from a low middle-income country (LMIC). This study aimed at identifying the frequency, risk factors, treatment options and outcome of deep venous thrombosis/pulmonary embolism (DVT/PE) in pediatric population in a tertiary care center from a LMIC. International classification of disease, ninth revision (ICD-9) was used to identify VTE in patients aged 0-18 years during January 2011 to September 2019. In-house computerized system was used to collect data for demographics, clinical and laboratory details. SPSS version 19 was used to analyzed data. The study was approved by Institutional ethical review committee (3872-Pat-ERC-15). During the study period, 134617 pediatric patients were hospitalized, DVT/PE was observed in 77 unique patients (47 males and 30 females) with a median (IQR) age of 14 (5-16) years equivalent to 5.9 VTE events /10,000 hospital admissions. Malignancy, community acquired infections and autoimmune diseases were the predominant risk factors (75%) in adolescent age-group while surgery for congenital heart anomalies was the primary reason (71%) in infants. Overall, lower extremity thrombosis was the most frequent (51%) followed by pulmonary embolism (25%). and upper extremity thrombosis (24%). Enoxaparin and unfractionated heparin were mainly used to treat VTE and all-cause mortality was 13% in the cohort studied. We observed substantial VTE events in pediatric patients during their hospital stay in a tertiary care center of a low-middle income country.

Risks of venous thrombosis and bleeding in critically ill adolescents after trauma or major surgery

BACKGROUND

The risks of venous thromboembolism (VTE) and bleeding in critically ill adolescents based on interventions received and anatomic site of trauma or major surgery may identify a cohort eligible for enrollment in a trial of pharmacologic prophylaxis.

METHODS

This retrospective cohort study using the Virtual Pediatric Systems database included adolescents admitted to pediatric intensive care units after trauma or major surgery between 2013 and 2017. Mixed effects logistic regression was used to determine the adjusted risks of VTE and bleeding with central venous catheterization (CVC), mechanical ventilation (MV) and anatomic site of trauma or major surgery. The adjusted risks were used to identify the cohort eligible for enrollment.

MEASUREMENTS AND MAIN RESULTS

VTE developed in 212 (0.8%) of 27,647 adolescents. The adjusted risk of VTE was >2% with CVC and 2 or more of MV and trauma or major surgery to the brain or abdomen. Excluding those with bleeds present on admission or at high risk of bleeding, 375 (1.4%) adolescents would be eligible for enrollment.

CONCLUSIONS

VTE is generally uncommon in adolescents after trauma or major surgery. The small proportion of adolescents who are at high risk of VTE and at low risk of bleeding impacts the feasibility of a trial.

LEVEL OF EVIDENCE

Prognostic Study Level II.

Association of perioperative red blood cell transfusion with postoperative venous thromboembolism in pediatric patients: A propensity score matched analysis

OBJECTIVE

To examine the association between perioperative red blood cell (RBC) transfusion and postoperative venous thromboembolism (VTE) in pediatric surgical patients.

METHODS

Retrospective cohort study using the National Surgical Quality Improvement Project Pediatric, a validated registry of 118 United States children's hospitals. Patients under 19 years of age undergoing a surgical procedure between 2012 and 2017 were included, with the main exposure being RBC transfusion in the perioperative period (48 hours prior to operation to 72 hours after operation). The primary 30-day outcome of interest was a postoperative VTE requiring therapy. Risk-adjusted odds ratios (aOR) were calculated using multiple logistic regression. Subgroup analyses were performed across multiple surgical specialties. Sensitivity analyses were performed after (a) imputation for missing variables and (b) propensity score matching.

RESULTS

During the study years, 482 867 pediatric patients (56.7% male; median age, 6 years [interquartile range, 1-12 years]) underwent an operation. Of these, 30 879 (6.4%) received at least one perioperative RBC transfusion. Postoperative VTE requiring therapy occurred in 618 patients (0.13%). After adjustment for multiple risk factors, perioperative RBC transfusion was associated with an increased risk of VTE (aOR 2.4; 95% CI, 1.9-3.0). The increased VTE risk persisted after imputation of missing demographic and clinical data as well as after 1:1 propensity score matching (29 811 matched pairs, aOR 2.2; 95% CI, 1.7-2.8).

CONCLUSIONS

Perioperative RBC transfusion is associated with an increased, albeit still very low, risk of postoperative VTE in pediatric patients. Patients receiving blood in the perioperative period may benefit from additional monitoring or VTE prophylaxis.

Multicenter Review of Current Practices Associated With Venous Thromboembolism Prophylaxis in Pediatric Patients After Trauma

OBJECTIVES

Frequency of venous thromboembolism in pediatric trauma patients admitted to PICUs is not insignificant, ranging up to 6%. Risk factors have been identified in this population. However, there is little consensus of actual venous thromboembolism prophylaxis practice. We examined factors associated with venous thromboembolism prophylaxis in PICUs.

DESIGN

A retrospective study evaluating associations with mechanical venous thromboembolism prophylaxis, pharmacologic venous thromboembolism prophylaxis, or dual therapy (DUAL) prophylaxis compared with no venous thromboembolism prophylaxis. Multivariable logistic regression explored the relationship between prophylaxis type and selected covariates with stepwise selection method to identify the independent predictors of venous thromboembolism prophylaxis utilization.

SETTING

Five level I/II pediatric trauma centers in the United States.

PATIENTS

Children less than 18 years from January 1, 2013, to December 31, 2013, admitted to the PICU after a trauma, identified through combined trauma registry and Virtual Pediatric Systems database.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Six hundred ninety-two patients were included in the database, with 55 excluded for missing data. Of the remaining 637 patients, 538 (84.5%) had no venous thromboembolism prophylaxis by 48 hours, 77 (12.1%) had only mechanical venous thromboembolism prophylaxis, 11 (1.7%) had DUAL, and 11 (1.7%) had pharmacologic venous thromboembolism prophylaxis alone. Multivariable analysis showed increased age, and orthopedic procedure was associated with all forms of prophylaxis. Orthopedic procedures were associated with higher utilization of dual prophylaxis use (odds ratio, 5.2; 95% CI, 1.2-21.8), pharmacologic venous thromboembolism prophylaxis (odds ratio, 8.5; 95% CI, 2.3-31.7), and mechanical venous thromboembolism prophylaxis (odds ratio, 2.2; 95% CI, 1.1-4.2) alone. Brain/spinal cord procedures (odds ratio, 3.7; 95% CI, 1.9-7.3) and abdominal procedures (odds ratio, 6.6; 95% CI, 2.5-17.1) were associated with mechanical venous thromboembolism prophylaxis. Head injury was associated with a decreased use of any prophylaxis (odds ratio, 0.5; 95% CI, 0.3-0.9). Patient comorbidities were associated with decreased use of mechanical venous thromboembolism prophylaxis (odds ratio, 0.5; 95% CI, 0.3-1.0).

CONCLUSIONS

Pharmacologic venous thromboembolism prophylaxis is not common in critically ill children after trauma. Patient age, orthopedic and vascular procedures, and higher injury severity are associated with pharmacologic venous thromboembolism prophylaxis.