ABSTRACT
Objective:To develop and analyze the predictive value of a combined predictive model based on the lower limb skin mottling score (SMS) for the outcome of sepsis patients admitted to ICU.Methods:Using a multicenter prospective, observational study, 458 patients admitted to six ICUs in 4 grade A tertiary hospitals in Fujian Province from March 1, 2019 to December 31, 2019 with first diagnosis of sepsis were selected as study subjects. Demographic data, clinical biochemical indexes, lower limb SMS, Sequential Organ Failure Assessment (SOFA), and Acute Physiology and Chronic Health EvaluationⅡ(APACHEⅡ) were collected. Binary Logistic regression was used to analyze the outcome of patients with sepsis after admission to ICU, a joint prediction model was constructed, ROC curves were drawn, and a comparison of the predictive efficacy of the lower limb SMS and the joint prediction model was performed.Results:Logistic multifactorial regression analysis showed that lower limb SMS, SOFA score, APACHEⅡscore, lactate, mean arterial pressure were the influencing factors of outcome of patients with sepsis after admission to ICU (all P<0.05). The area under the curve of lower limb SMS predicted ICU outcome risk was 0.718 (95% CI 0.675-0.959, P<0.001) with the sensitivity of 51.96% and specificity of 85.11%. The area under the curve of predictive ability of the combined model based on the lower limb SMS for ICU outcome of sepsis patients was 0.835 (95% CI 0.762-0.878, P<0.01) with the sensitivity of 81.40% and specificity of 73.00%, the difference was significant ( P<0.05). Conclusions:The combined prediction model based on lower limb SMS can more accurately predict the prognosis of sepsis patients after admission to ICU, which can be a new nursing assessment tool for nurses to evaluate the prognosis of sepsis patients and help nurses to make more accurate clinical care decisions and reduce the morbidity and mortality rate.