Prognostic value of modified Erasmus Guillaiun Barre' syndrome outcome score (mEGOS) in an Italian cohort of patients: Challenges, insights and limitations.

backgroundClinical course and outcome of Guillain Barre' syndrome (GBS) are highly variable. among countries. The modified Erasmus GBS Outcome Score (mEGOS) is a prediction model proposed to predict risk of walking inability within six months of the disease. Our study aimed to validate the mEGOS in an Italian cohort of patients.

methodsWe used dataset from 230 adults GBS admitted to Neurology Unit of Modena over 22 years. We studied whether mEGOS applied at entry and at week 1 to patients unable to walk could predict the to restoration of independent ambulation at 4 and 24 weeks. Model performance was based on discrimination, using area under receiver operating characteristic curve and on calibration (observed versus predicted probability). Moreover, we assessed the predictive ability of individual variables included and not included in the mEGOS model.

resultsFor validation, eligible patients were 62 at entry and 162 at week 1.The mEGOS discriminative ability at entry was great for prediction at 4 weeks (AUC 0.70), at 6 months (AUC 0.82) and excellent when applied at day 7 (AUC > 0.85).With regard to calibration, the model showed non optimal performance, because it underestimated risk of worse outcome, especially for patients with low predicted risk.

conclusionsThe mEGOS scoring system can discriminate Italian GBS patients with low or high risk of poor outcome, when used in first week of disease. However, it should be considered moderately reliable predictor of independent ambulation at 4 weeks and at 6 months in our patients.