The incidence of blunt abdominal injury has significantly increased, and the liver is one of the most commonly damaged organs. In this study, we explored and established a nomogram model for patients with liver ruptures undergoing surgical treatment. A retrospective analysis was conducted for 66 adult patients with liver rupture, who were admitted to our hospital from January 2011 to October 2018. These patients were classified into two groups, according to whether the patient had surgery: surgery group (41 cases) and non-surgical group (25 cases). The following data were collected from these two groups of patients: gender, age, injury mechanism, liver damage, laboratory test results, and hospitalization. Multivariate logistic regression analysis was performed to screen the risk factors of patients who require surgical treatment, establish a predictive model based on the selected indicators, and draw the nomogram. Receiver operating characteristic curves and the calibration curve were used to evaluate the predictive value of the model. Compared to the non-surgical group, the body temperature decreased, the heart rate increased, the injury severity score grade increased, the blood urea nitrogen, blood uric acid, creatinine (Cr), arterial partial pressure of oxygen, alkali excess, blood lactic acid and creatine kinase isoenzymes MB (CK-MB) increased, and the HCO- and Glasgow Coma Scale (GCS) coma scores decreased for patients in the surgical group (all, p<0.05). The logistic regression analysis revealed that Cr, arterial partial pressure of oxygen, HCO3-, CK-MB, and the Glasgow coma score were the influencing factors for surgical intervention for liver rupture. The nomo-gram model constructed based on these five indicators had a good degree of discrimination (area under the curve = 0.971, 95% CI: 0.896-0.997) and accuracy. A nomogram model established based on Cr, arterial partial pressure of oxygen, HCO3-, CK-MB, the GCS, and other parameters can accurately predict the surgical treatment of patients with liver rupture.
Read full abstract