Contributions
Abstract: EP1396
Type: Poster Sessions
Abstract Category: Clinical aspects of MS - Clinical assessment tools
Background: In multiple sclerosis (MS), severity and symptoms of relapses are highly variable and the residual disability is difficult to predict.
We aimed to develop a clinical-based model for predicting the risk of disease progression at six months post-relapse in MS patients.
Methods: For learning and internal validation, we used data of the COPOUSEP (« Corticothérapie Orale dans les Poussées de Sclérose en Plaques ») study. It consisted in 186 patients, aged 18 to 55, respecting the 2005 McDonald criteria and having an EDSS ≤ 5 at inclusion. The predictive outcome was defined as an increase of ≥1 point of EDSS six months after relapse. We studied the following prognostic factors: gender, age, disease duration and disease modifying therapy at relapse time, EDSS before and during the relapse, and symptoms of the relapse (motor, sensory dysfunction, visual or cerebellar impairment, cranial nerve dysfunction, bladder/bowels disorders, or cognitive impairment). We used a logistic regression with LASSO penalization to construct the model and Bootstrap Cross-validation for internal validation. The model was externally validated by using a cohort constituted by 174 patients followed in Bordeaux University Hospital.
Results: Patients experienced mostly sensory disturbances (74%) and motor dysfunction (39%). Six factors were retained to construct the SCOPOUSEP model (“SCOring des POUssées de Sclérose en Plaques”): age > 40 yo, an increase of EDSS at relapse time higher than 1.5 points, disease duration, EDSS higher to or equal than 1 point before relapse, proprioceptive ataxia and subjective sensory disorders. The model presented acceptable discriminative capacities for both the internal (AUC=0.82, CI95% from 0.72 to 0.91) and external (AUC=0.71, CI95% from 0.61 to 0.79) validations.
Conclusion: We developed and validated a simple predictive model to help neurologists to evaluate the severity of a relapse in daily practice as defined by the risk of progression 6 months after the relapse.
Disclosure: All : Nothing to disclose
Abstract: EP1396
Type: Poster Sessions
Abstract Category: Clinical aspects of MS - Clinical assessment tools
Background: In multiple sclerosis (MS), severity and symptoms of relapses are highly variable and the residual disability is difficult to predict.
We aimed to develop a clinical-based model for predicting the risk of disease progression at six months post-relapse in MS patients.
Methods: For learning and internal validation, we used data of the COPOUSEP (« Corticothérapie Orale dans les Poussées de Sclérose en Plaques ») study. It consisted in 186 patients, aged 18 to 55, respecting the 2005 McDonald criteria and having an EDSS ≤ 5 at inclusion. The predictive outcome was defined as an increase of ≥1 point of EDSS six months after relapse. We studied the following prognostic factors: gender, age, disease duration and disease modifying therapy at relapse time, EDSS before and during the relapse, and symptoms of the relapse (motor, sensory dysfunction, visual or cerebellar impairment, cranial nerve dysfunction, bladder/bowels disorders, or cognitive impairment). We used a logistic regression with LASSO penalization to construct the model and Bootstrap Cross-validation for internal validation. The model was externally validated by using a cohort constituted by 174 patients followed in Bordeaux University Hospital.
Results: Patients experienced mostly sensory disturbances (74%) and motor dysfunction (39%). Six factors were retained to construct the SCOPOUSEP model (“SCOring des POUssées de Sclérose en Plaques”): age > 40 yo, an increase of EDSS at relapse time higher than 1.5 points, disease duration, EDSS higher to or equal than 1 point before relapse, proprioceptive ataxia and subjective sensory disorders. The model presented acceptable discriminative capacities for both the internal (AUC=0.82, CI95% from 0.72 to 0.91) and external (AUC=0.71, CI95% from 0.61 to 0.79) validations.
Conclusion: We developed and validated a simple predictive model to help neurologists to evaluate the severity of a relapse in daily practice as defined by the risk of progression 6 months after the relapse.
Disclosure: All : Nothing to disclose