Natural history of amyotrophic lateral sclerosis in a database population Validation of a scoring system and a model for survival prediction

Abstract

Over 1200 patients with motor neuron disease have been carefully diagnosed, followed, and included in a detailed database delineating characteristics of the disease. Of these patients, 831 were identified as exhibiting typical, sporadic amyotrophic lateral sclerosis (ALS). The progression of the disease in these patients has been followed with our scoring system, and the ALS score was verified as a significant covariate of survival. Age at first symptom, delay from first symptom to entering ALS clinic, and rate of change of respiratory function were also identified as significant covariates of survival. These measures, applied to the Cox proportional hazards model, were used to develop a mathematical model for prediction of survival time in ALS, which proved highly accurate for the 80% of patients at intermediate risk. For those patients, a second model was developed which accurately predicts, after an initial period of observation, the time over which ALS patients will decline a set number of points in total ALS score. Such validation permits initial trials for drug therapies in ALS by comparison of relatively small groups of treated patients to this historical control group, based on the model of predicted time to a particular decrement in total ALS score.