Models to predict expansive intracranial hematomas occurrence for adult traumatic brain injury patients presenting at Accident and Emergency Department at Mulago National referral Hospital in Uganda
![Thumbnail](/xmlui/bitstream/handle/123456789/3730/Models%20to%20predict%20expansive%20intracranial%20hematomas%20occurrence%20for%20adult%20traumatic%20brain%20injury%20patients%20presenting%20at%20Accident%20and%20Emergency%20Department%20at%20Mulago%20National%20Referral%20Hospital%20in%20Uganda.pdf.jpg?sequence=4&isAllowed=y)
View/ Open
Date
2024Author
Kamabu, Larrey Kasereka
Oboth, Ronald
Bbosa, Godfrey S.
Fuller, Anthony T.
Deng, Daniel
Lekuya, Hervé Monka
Ssenyondwa, John Baptist
Sekabunga, Juliet Nalwanga
Kataka, Louange Maha
Obiga, Doomwin Oscar Deogratius
Kaddumukasa, Martin N.
Kiryabwire, Joel
Galukande, Moses
Sajatovic, Martha
Kaddumukasa, Mark
Kitya, David
Haglund, Michael M.
Metadata
Show full item recordAbstract
Background: A timely and accurate surgical decision regarding expansive intracranial hematomas (EIH) is crucial in clinical practice and patient outcomes, particularly in resource-limited settings. Although several predictive models exist, their utility in forecasting EIH remains underexplored, often neglecting the evolving nature of intracranial hemorrhage.
Aim: Study determined the models that can be used to predict the EIH occurrence for patients with TBI in Uganda. Methods: A cross sectional study was conducted to determine characteristics of patients with EIH which was then used to identify applicable models for predicting EIH occurrence among TBI patients. Adult TBI patients with intracranial hematoma undergoing surgical evacuation between June 16, 2021, and December 17, 2022, were included. Participants were categorized based on EIH presence or absence, determined by hematoma volume changes. Logistic regression analyzed factors influencing EIH, including demographics, neurological assessment, hematological parameters, and neuroimaging.
Results: Of the total 324 enrolled patients with intracranial hematomas, 59.3% (n=192) developed EIH, resulting in a proportion of 0.59 (95% CI: 0.54 to 0.65). The final model incorporated age, systolic and diastolic blood pressure, subdural hematoma (SDH), diffuse axonal injury (DAI), skull fracture, and an interaction term between skull fracture and SDH. Each unit increase in systolic blood pressure raised EIH odds by 1.045, while diastolic blood pressure increase lowered odds to 0.942. SDH increased odds by 6.286, and DAI by 4.024. However, in cases of skull fracture, SDH reduced odds to 0.0676. The model's five-fold cross-validated average area under the receiver operating curve (AUC) was 0.722, with 64.5% accuracy.
Conclusion: EIH is common among TBI patients in Uganda with a prevalence of 59.3%. When systolic blood pressure and diastolic are raised by 1 unit from the baseline, having SDH, DAI and skull fracture, the bigger odds of having EIH it becomes. These new models can inform policy and future interventions to predict earlier EIH occurrence and build off the effective treatment modalities for such patients.
Collections
- Research articles [61]