Amyotrophic lateral sclerosis established as a multistep process across phenotypes.

Background and purpose: Given the accepted multistep process of disease causation in amyotrophic lateral sclerosis (ALS), the present study was undertaken to determine the number of steps required for disease onset across each of the ALS phenotypes.

Methods: Clinical and demographic data were prospectively accumulated using the Australian Motor Neurone Disease Registry (2005-2016), and age-specific incidence rates were calculated. Poisson regression was utilized to assess the relationship between log age-specific incidence and log age of onset, with McFadden's R² used to assess the goodness of fit of the model.

Results: In total, 2647 ALS patients were included, with mean disease-onset age being 62.2 ± 12.1 years. A linear relationship between log incidence and log age was established across ALS phenotypes, with variable slope estimates: bulbar 5.1 (95% confidence interval [CI] 4.6-5.6); cervical 2.7 (95% CI 2.3-3.0); lumbar 3.5 (95% CI 3.2-3.9); flail arm 4.7 (95% CI 3.9-5.5); flail leg 3.6 (95% CI 2.6-4.5); primary lateral sclerosis 2.7 (95% CI 1.8-3.7). Slope estimates were significantly higher in the bulbar compared to the cervical, lumbar and primary lateral sclerosis phenotypes. McFadden's R² values were >0.4 for all phenotypes indicating excellent model fit.

Discussion: A multistep process has been established across all ALS phenotypes with variable slope estimates, suggesting that the number of steps to develop disease is different across clinical presentations. Identification of mechanisms underlying slope estimate variability could exert pathophysiological significance.