| Objective(s):This study aimed to determine the independent predictors of albumin infusion in patients with sepsis or septic shock during early fluid therapy for the first 48 intensive care unit(ICU)hours,and construct a nomogram to predict the probability of albumin infusion.Methods:This study retrospectively reviewed patients with sepsis or septic shock admitted to ICU(emergency ICU,surgical ICU and medical ICU)of the Second Affiliated Hospital of Kunming Medical University between August 2020 and August 2022.The following baseline variables were collected: general characteristic information(gender and age),baseline vital signs(systolic blood pressure,diastolic blood pressure,mean arterial blood pressure,heart rate and body temperature),baseline Acute Physiology and Chronic Health Evaluation Ⅱ(APACHE Ⅱ)scores,sequential organ failure assessment(SOFA)scores,septic shock,co-existing conditions(hypertension,diabetes,coronary atherosclerotic heart disease,chronic obstructive pulmonary disease,stroke and cancer),baseline laboratory tests(lactate concentration,white blood cell count,platelet count,procalcitonin level,total protein level,albumin level,total bilirubin level,direct bilirubin level,indirect bilirubin level,urea level and creatinine level),norepinephrine requirement and the maximum norepinephrine infusion rate,invasive ventilation requirement,renal replacement therapy requirement,volume of crystalloids administered during ICU admitted day and the first 48 ICU hours.The patients were randomly grouped into modeling and validation sets.Modeling patients were divided into two groups according to whether or not they had administration of albumin.Independent predictors were identified and a nomogram infographic was developed by the least absolute shrinkage and selection operator(LASSO)regression and multivariate logistic regression using the modeling set.The bootstrap resampling method was used for internal validation in modeling group.The performance of the nomogram was assessed using receiver operating characteristic curve(ROC),area under the curve(AUC),calibration curve and decision curve.Results:1.In total,420 included sepsis or septic patients were randomly split in a 7:3ratio into modeling(n=296)and validation(n=124)sets.Modeling patients were devided into albumin group(n=216)and no-albumin group(n=80).2.In the modeling set,baseline albumin level(OR=0.732,95%CI: 0.655～0.817),baseline direct bilirubin level(OR=1.011,95%CI: 1.001～1.023)and the amount of cumulative crystalloids(m L/h)administration over the first 48 ICU hours(OR=1.023,95%CI: 1.012～1.034)were independent predictors for albumin infusion during early fluid therapy.3.The nomogram infographic was constructed based on the resulting prediction model.Repeating the bootstrap self-sampling and internal validation 1000 times,thesampling results showed a concordance index(C-index)of 0.853.AUCs for predicting albumin infusion were 0.861(95%CI:0.814～0.907)and 0.851(95%CI:0.768～0.934)in modeling group and validation group respectively,indicating an acceptable discrimination of the nomogram.The calibration curve showed adequate agreement between the predictive nomogram and actual observations in the modeling and validation groups,indicting a robust calibration of the nomogram.The decision curve showed that patients would gain a significant positive benefit from the nomogram when the threshold probability of a patient was above 30%,indicating a good clinical applicability of the nomogram.Conclusion(s):1.Baseline albumin level,baseline direct bilirubin level and the volume of cumulative crystalloids(m L/h)administration over the first 48 ICU hours were independent predictors of albumin infusion in patients with sepsis or septic shock during early fluid resuscitation.2.A resulting prediction model based on the independent predictors identified above could perform well with good discrimination,calibration and clinical applicability.3.A nomogram infographic based on the resulting model could be used as an effective and convenient tool to guide albumin administration in ICU patients with sepsis and septic shock during early fluid therapy. |
