Appropriate nutritional management requires comprehensive decisions based on multiple nutritional and clinical indicators. In this study, we retrospectively evaluated the usefulness of blood biomarkers as prognostic factors in NST rounds to lead to appropriate nutritional management of undernourished patients. We targeted 176 patients who received NST rounds at our hospital, adjusted for age and gender by propensity score matching. The dependent variable was the outcome at discharge as assessed by the attending physician (favorable group, unfavorable group), and the explanatory variables were blood biomarkers (ALB, ChE, CRP, Hb, RBP, TC, Tf, TLC, TTR, WBC) tested before NST rounds and statistically analyzed. Multivariate logistic regression analysis showed significant differences in Hb, RBP, TLC, WBC, and AUC ranged from [0.650 (95%CI: 0.570–0.731)] to [0.722 (95%CI: 0.647–0.797)] in comparison between before (basal model) and after (enhanced model) adding the four items mentioned above. The IDI value was [0.076 (95%CI: 0.036–0.116), p < 0.001]. In the decision tree analysis, TLC was an important factor in the prognosis of undernourished patients. Combining the tests of immunocompetence and nutritional assessment protein with short half-lives improves prediction accuracy in undernourished patients. We expect future visual prognosis prediction by IDI and decision tree analysis using patient information on detailed nutritional assessment. We found that measured TLC, RBP, Hb, and WBC values before NST rounds were important prognostic factors.
View full abstract