Risks, Benefits, and the Optimal Time to Resume Deep Vein Thrombosis Prophylaxis in Patients with Intracranial Hemorrhage

Application of artificial neural network in daily prediction of bleeding in ICU patients treated with anti-thrombotic therapy

OBJECTIVES

Anti-thrombotic therapy is the basis of thrombosis prevention and treatment. Bleeding is the main adverse event of anti-thrombosis. Existing laboratory indicators cannot accurately reflect the real-time coagulation function. It is necessary to develop tools to dynamically evaluate the risk and benefits of anti-thrombosis to prescribe accurate anti-thrombotic therapy.

METHODS

The prediction model,daily prediction of bleeding risk in ICU patients treated with anti-thrombotic therapy, was built using deep learning algorithm recurrent neural networks, and the model results and performance were compared with clinicians.

RESULTS

There was no significant statistical discrepancy in the baseline. ROC curves of the four models in the validation and test set were drawn, respectively. One-layer GRU of the validation set had a larger AUC (0.9462; 95%CI, 0.9147-0.9778). Analysis was conducted in the test set, and the ROC curve showed the superiority of two layers LSTM over one-layer GRU, while the former AUC was 0.8391(95%CI, 0.7786-0.8997). One-layer GRU in the test set possessed a better specificity (sensitivity 0.5942; specificity 0.9300). The Fleiss' k of junior clinicians, senior clinicians, and machine learning classifiers is 0.0984, 0.4562, and 0.8012, respectively.

CONCLUSIONS

Recurrent neural networks were first applied for daily prediction of bleeding risk in ICU patients treated with anti-thrombotic therapy. Deep learning classifiers are more reliable and consistent than human classifiers. The machine learning classifier suggested strong reliability. The deep learning algorithm significantly outperformed human classifiers in prediction time.

Leukocyte as an Independent Predictor of Lower-Extremity Deep Venous Thrombosis in Elderly Patients With Primary Intracerebral Hemorrhage

Objective

Due to the interaction of leukocytes with platelets and coagulation factors, they may in turn play a role in hemostasis or the formation of thrombi. This study aimed to investigate the association of leukocytosis on admission with an increased risk of acute lower-extremity deep venous thrombosis (LEDVT) in elderly patients with primary intracerebral hemorrhage (ICH).

Methods

This was a single-center, retrospective observational study of consecutive patients observed with spontaneous ICH aged 60 years or above at Lanzhou University Second Hospital from January 2017 to September 2021. Clinical data and demographic information were collected and analyzed. Univariate and multivariate analyses were conducted to identify independent risk factors of acute LEDVT. One-to-one matching was implemented to balance important patient characteristics by the groups' propensity score matching (PSM) analysis.

Results

A total of 371 elderly patients with primary ICH fulfilled requirements for inclusion and exclusion, of whom 33 (8.89%) experienced LEDVT. Leukocyte counts were statistically higher in the LEDVT group compared to the non-LEDVT group [12.89 (8.80–14.61) × 10⁹ cells/L vs. 8.31 (6.60–10.75) × 10⁹ cells /L, p < 0.001]. Multivariate logistic regression models adjusted for several potential confounding factors were performed, and leukocytes were consistently a significant independent predictor of LEDVT. The optimal cut-off value of leukocyte counts calculated from the receiver operating characteristic (ROC) curve to predict LEDVT was 10.22 × 10⁹ cells /L (area under the curve:0.714, 95%CI 0.665–0.759; the sensitivity was 72.73%; the specificity was 71.01%) in elderly patients with primary ICH. After one-to-one PSM, compared to the matched non-LEDVT group, the matched LEDVT group had significantly higher leukocyte counts [11.98 (8.40–13.94) × 10⁹ cells/L vs. 6.12 (4.68–12.00) × 10⁹ cells/L, p = 0.003]. After PSM, the ROC curve was plotted for leukocytes as a predictor of LEDVT, with an AUC of 0.722 (95%CI 0.593–0.828, p = 0.001; the sensitivity was 87.10%, and the specificity was 61.29%). Elevated leukocytes remained independently significant as predictors of LEDVT in elderly patients with primary ICH.

Conclusion

Leukocyte at admission is an independent risk factor of LEDVT in elderly patients with primary ICH.