The predictive power of postoperative complications and the agreement between PG-SGA SF and GLIM in identifying malnourishment in patients with pancreatic diseases before pancreatic surgery

Résumé fourni par la source

Background Patients with pancreatic diseases impose several nutritional challenges and identifying malnourished patients is important since malnutrition may increase the risk of postoperative complications. Aims To assess the agreement between Patient Generated-Subjective Global Assessment Short Form (PG-SGA SF) and Global Leadership Initiative on Malnutrition (GLIM) criteria in identifying malnutrition in patients with pancreatic diseases scheduled for elective surgery. To investigate whether there is an association between malnutrition and postoperative complications. Methods A secondary analysis of a prospective observational cross-sectional study in patients with pancreatic diseases undergoing pancreatic surgery from September 2017 to January 2018 at Karolinska University Hospital (Stockholm, Sweden). Nutritional screening was performed by PG-SGA SF and malnutrition diagnosis was defined by GLIM with and without prior screening. Results A total of 34 patients were included (16 men, 18 women), median age 70 (range 38–83) years, median BMI 24.4 (range 19.3–40.4) kg/m 2 . Risk of malnutrition was detected in 14 patients (41 %) by PG-SGA SF. Malnutrition was detected in 9 patients (26 %) by GLIM with prior screening (GLIM + screening) and in 17 patients (50 %) by GLIM without prior screening (GLIM - screening). The agreement between PG-SGA SF and GLIM - screening was 62 %, the sensitivity 64 %, the specificity 60 %, and the Cohen's kappa was 0.235, P =0.08. GLIM - screening had a positive predictive value of 53 % and negative predictive value of 71 %. A total of 14 patients had postoperative complications. The GLIM - screening showed a RR of 1.40 (CI: 0.58–3.37) of postoperative complications when being malnourished. The GLIM + screening showed a RR of 1.66 (CI: 0.49–5.61) of postoperative complications when being malnourished. Conclusions This study showed that PG-SGA SF and GLIM had a slight to fair agreement in detecting malnutrition in patients undergoing pancreatic resection. No firm conclusions in predicting postoperative complications could be found.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.