The Suitability of Four Nutritional Screening Tools for Nutritional Screening in Patients with Cirrhosis: a Network Meta-analysis

Background Malnutrition has become one of the adverse complications in patients with cirrhosis, and early nutritional screening and early identification of patients with cirrhosis can effectively improve clinical outcomes, however, the types of nutritional screening tools for patients with cirrhosis are varied and not yet standardized, and further research is needed. Objective To evaluate the applicability of 4 nutritional screening tools for patients with cirrhosis. Methods CNKI, VIP, Wanfang Data, PubMed, Embase, Cochrane Library and Web of Science were searched for diagnostic studies related to nutritional screening tool screens for malnutrition in cirrhotic patients. The search time was limited to December 2023. Two researchers individually read and filtered the literatures, extracted data, and assessed the bias risk of the incorporated researches. RevMan 5.4.1, Meta-DiSc, and StataMP 17.0 were used to perform network meta-analysis. The sensitivity, specificity, positive predictive value and negative predictive value of different nutritional screening tools were ranked by the surface under the cumulative ranking curve (SUCRA) . Results 5 Chinese and 5 English literatures were incorporated, totaling 10, including 1 299 patients; and four nutritional screening tools were included: the Nutritional Risk Screening 2002 (NRS2002), the Royal Free Hospital-Nutritional Prioritization Tool (RFH-NPT), the Malnutrition Universal Screening Tool (MUST), and the Subjective Global Assessment (SGA). The findings of meta-analysis revealed that the combined sensitivity of the four nutritional screening tools was 0.65 (95%CI=0.56-0.73), 0.93 (95%CI=0.89-0.96), 0.77 (95%CI=0.72-0.82), respectively. The combination specificity was 0.87 (95%CI=0.83-0.91), 0.72 (95%CI=0.64-0.79), 0.81 (95%CI=0.68-0.90), and MUST only be studied separately, without the combination sensitivity and combination specificity. The results of network meta-analysis showed the sensitivity and negative predictive value of SGA were lower than that of RFH-NPT (OR=0.03, 95%CI=0-0.55; OR=0.08, 95%CI=0.01-0.81, P<0.05) ; and the sensitivity and negative predictive value of RFH-NPT were higher than that of NRS2002 (OR=44.33, 95%CI=3.94-498.52; OR=17.68, 95%CI=2.13-147.05, P<0.05). The results of the combined subject work characterization curve (SROC) showed that the area under the SROC curve (AUC) for screening for malnutrition in cirrhotic patients was 0.86 for NRS2002, 0.90 for RFH-NPT, and 0.85 for SGA. The SUCRA values of the tools ranked in terms of combined sensitivity from highest to lowest were RFH-NPT (SUCRA=99.5%) >MUST (SUCRA=43.0%) >SGA (SUCRA=39.0%) >NRS2002 (SUCRA=18.5%) ; the SUCRA values of these tools ranked in terms of combined specificity from highest to lowest were: MUST (SUCRA=91.4%) >NRS2002 (SUCRA=49.1%) >SGA (SUCRA=39.8%) >RFH-NPT (SUCRA=19.7%) ; the SUCRA values of these tools ranked in terms of positive predictive value from highest to lowest were MUST (SUCRA=95.2%) >RFH-NPT (SUCRA=37.4%) >NRS2002 (SUCRA=36.1%) >SGA (SUCRA=31.3%) ; the SUCRA values of these tools ranked in terms of negative predictive value from highest to lowest were RFH-NPT (SUCRA=99.1%) >MUST (SUCRA=44.9%) >SGA (SUCRA=39.4%) >NRS2002 (SUCRA=16.7%) . Conclusion The current evidence shows that RFH-NPT and MUST are suitable, but this conclusion still needs to be further confirmed by large samples and multiple high-quality studies.