Due to the lack of a comprehensive evaluation of the prognosis of small bowel obstruction (SBO), recent clinical strategies have remained subjective and controversial. The recognition of pretreatment risk factors and tailored treatment could improve SBO outcomes. A series of posttreatment laboratory tests were integrated into a two-step clustering (TSC) analysis. The TSC outcome was determined according to different predictor importance (PI). A risk score (RS) system for the TSC outcome model was constructed by multivariable analysis. Receiver operating characteristic (ROC) curves and the area under the curve (AUC) were calculated to assess prediction accuracy. Of the 355 patients, 66 (18.6%) were sorted into the better prognosis group (BPG), 149 (42.0%) were sorted into the poor prognosis group (PPG), and 140 (39.4%) were sorted into the severe prognosis group (SPG) by TSC analysis. For the TSC outcome, four variables with higher PI were identified, namely, Ca (PI = 1), albumin (PI = 0.62), WBC count (PI = 0.5) and NE% (PI = 0.45). Compared with the SPG, the BPG presented better outcomes after surgery events. The TSC outcome model was efficient in distinguishing the duration of bowel function recovery and hospital stay by Kaplan‒Meier curves. Via multivariate analysis, a RS consisting of four risk factors, namely, constipation duration (OR = 1.002), APTT (OR = 0.923), PT (OR = 1.449) and PCT (OR = 1.540), was identified. The AUC of the RS on the TSC outcome model was 0.719 (95% CI, 0.635-0.804). A novel TSC outcome model and RS system was constructed to comprehensively reflect the tailored treatment, surgical events and posttreatment recovery for SBO patients.
Read full abstract