Association Between Vital Signs and Urine Output for the Prediction of Sequential Organ Failure Assessment (SOFA)

One of the most common causes of death for hospitalized patients is sepsis. Acute kidney injury (AKI) brought on by sepsis is associated with worse outcomes. To avoid a worse prognosis, assessing and quickly identifying the function of the static kidney while determining AKI is essential. However, compared to analyzing vital signs that can be obtained at the bedside, urine output takes much longer to obtain. This investigation aims to: (i) identify the relationship between urine output and vital signs in determining SOFA between sepsis AKI and sepsis non-AKI; and (ii) assess the performance of predictive models of vital signs and urine output on both cohorts. This is a retrospective study on 30 sepsis AKI and 30 sepsis non-AKI patients admitted to the Intensive Care Units of Hospital U niversiti Sains Malaysia (HUSM). Data collected include age, gender, heart rate, respiratory rate, diastolic blood pressure, systolic blood pressure, mean arterial pressure, Glasgow Coma Scale, temperature, and urine output. The data included patients older than 18 years old and admitted to the ICU for more than 24 hours. Statistical analysis was performed using Matlab software. The results indicated that sepsis AKI and sepsis non-AKI were significantly different for urine output (p-value < 0.05). The performance of the multiple regression model on sepsis AKI using vital signs and urine output was better compared to the model with only vital signs as a predictor, with 3.53 on the root mean square error (RMSE) and 0.376 for the coefficient of determination (R
²
). However, the model of sepsis non-AKI presented similar performance for both sets of these predictors. Vital signs require slightly less time than vital signs with urine output to identify the Sequential Organ Failure Assessment (SOFA) and prevent delays in sepsis treatment, even though vital signs and urine output can improve the prognosis.