Gallo et al. 1 suggest that smoking may affect the pathogenesis of amyotrophic lateral sclerosis (ALS). The editorial by Weisskopf and Ascherio 2 maintains that incidence and mortality figures are not always identical and this may complicate the analysis. Indeed, if the incidence of ALS is identical in smokers and nonsmokers, but survival is shorter in smokers, more patients will die in a given period among smokers in a given cohort, but in a steady state the incidence and mortality will be the same.
Smoking is also a major contributor to stroke, supposedly by enhanced atherosclerosis. In Alzheimer's disease (AD) recent data prove that smoking increases the incidence of dementia, probably by inducing a mixed dementia combining a neurodegenerative process and vascular lesions. 3 A vascular contribution, however, is unlikely in ALS.
In Parkinson's disease (PD) smoking seems to be protective. This is surprising as both oxidative stress and vascular factors may be involved in PD pathogenesis. This led to the suggestion of an inverse causation; ie, that PD patients are less likely to smoke, perhaps reflecting personality traits. Paralleling this logic, some premorbid factors or certain biologic processes that "cause" ALS may also trigger smoking behavior years before the disease is expressed. However, if smoking "causes" ALS through an intoxicant, a dose-response relationship might be reflected as an earlier age at disease onset among smokers, as examined by us in PD. 4 Another possibility is that smoking may unmask subclinical ALS. If the preclinical phase of ALS is long, as in AD and PD, then possibly an unknown number of preclinical ALS patients die before they are diagnosed clinically. The nicotinic effects of smoking at the neuromuscular junction could allow earlier expression of the disease, particularly of lower motor neuron manifestations. Thus, if smoking uncovers the disease, it may identify patients early and artificially increase the number of smokers diagnosed with ALS.
This hypothesis can be tested (at least partially) if age at diagnosis and disease duration were provided. Thus, if smoking allows an earlier expression of the disease, the age at diagnosis should be lower among smokers. If, however, smoking accelerates the disease process, then disease duration should be reduced among smokers (although other factors may also affect survival). Thus data on age of onset and disease duration among smokers and nonsmokers would be of interest.