Construction and evaluation of a prognostic prediction model based on the mEGOS score for patients with Guillain-Barré syndrome

Abstract

BackgroundGuillain-Barré syndrome (GBS) is an immune-mediated acute peripheral neuropathy in which up to 20% patients remain unable to walk independently after 6 months of onset. This study aimed to develop a clinical prognostic model based on the modified Erasmus GBS Outcome Score (mEGOS) for predicting the prognosis of GBS patients at 6 months of onset.MethodsThe clinical data of 201 GBS patients were retrospectively analyzed. According to the GBS disability score (GBS-DS) at 6 months of onset, patients were divided into a good prognosis group (GBS-DS <3 points) and a poor prognosis group (GBS-DS≥3 points). Univariate and multivariate analysis was used to screen out independent risk factors for poor prognosis, and a prediction model was accordingly constructed for GBS prognosis.ResultsThe mEGOS score, serum albumin (ALB) and fasting plasma glucose (FPG) were independent risk factors for poor prognosis in patients with GBS, and the above risk factors were used to construct a prognostic model of mEGOS-I and a nomogram. The receiver operating characteristic (ROC) curve showed that the area under curve (AUC) of mEGOS-I at admission and at 7 days of admission to predict poor prognosis at 6 months of GBS onset was 0.891 and 0.916, respectively, with sensitivities of 82.7% and 82.6% and specificities of 86.5% and 86.6%, respectively. Decision curve analysis showed that the nomogram had a very high clinical benefit.ConclusionTo our knowledge, this is the first report of the construction of a prognostic prediction model based on the mEGOS score, ALB, and FPG that can accurately and stably predict the prognosis of GBS patients at 6 months of onset.