ALS is the unifying term for this heterogeneous syndrome and time to question the dogma forged at El Escorial

Historically characterised as a neuromuscular disease involving progressive loss of both central and peripheral motor neurons, amyotrophic lateral sclerosis (ALS) now sits among the cerebral neurodegenerative syndromes, involving clinical, pathological and genetic overlap with frontotemporal dementia (FTD). Descriptions of progressive lower motor neuron degeneration and associated muscle wasting, termed progressive muscular atrophy (PMA), were published by anatomists such as Charles Bell as early as 1836. Jean-Martin Charcot’s later contribution was the unique clinicopathological linking of the upper and lower motor neuron degeneration, which he termed ALS, since when PMA has become associated with a lower motor neuron-only variant. Wilhelm Erb and Pierre Marie are associated with identifying the upper motor neuron-only variant of progressive weakness, termed primary lateral sclerosis (PLS). In their JNNP paper, de Vries and colleagues1 demonstrate that so-called PMA and PLS cases show cognitive and behavioural impairment within the spectrum inherent to more typical ALS. In doing this so definitively, with large well-characterised groups, they add to a body of evidence that should call time on the confusing lexicon of phenotypic terms that frequents the penumbra of ALS.

‘What type of ALS do I have?’ This common question in the clinic reflects public awareness of variation in the clinical presentation and speed of progression of the neurodegenerative disorder most famously diagnosed in both Lou Gehrig and Stephen Hawking, despite starkly different natural histories. Neither the site of first symptom nor the spread of regional involvement is random in ALS, but progression of weakness is axiomatic. What varies most obviously is the rate of decline in function, although strikingly constant for the individual patient. ALS, despite all its phenotype variation and molecular complexity, has a final common pathway that is typically unmistakable to the experienced clinician.2 ALS might be increasingly considered the genus for a series of cell pathway-defined species, within a family of neurodegenerative disorders.

The clinical essence of ALS is as a variable combination of upper and lower motor neuron signs. It has been recognised in clinic-based and population-based studies that the rate of progression is significantly slower at the two extremes3 4 with a clinical continuum spanning what may be termed upper and lower motor neuron-predominant forms either side of the equal mixture of ‘classical’ ALS. Cognitive impairment in ALS is also associated with a worse prognosis, yet the natural history of ‘pure’ FTD is typically three times the median survival of ALS. The broader ALS-FTD syndrome can therefore be envisaged as three variably overlapping pathological ‘networks’ (lower motor neuron, upper motor neuron and frontotemporal), with more slowly progressing symptoms associated with the more isolated forms of each.

A smaller study previously identified mild cognitive impairment in cases of so-called PMA.5 Histopathological studies demonstrated central nervous system pathology in similar lower motor neuron cases,6 7 and neuroimaging has permitted in vivo demonstration of a similar signature of cerebral white matter involvement common to classical ALS.8 The authors in the present study used the strict Gordon criteria for their cases of PLS, in which the requirement to be 4 years from symptom onset helps to distinguish it from upper motor neuron-predominant ALS. In this setting PLS represents <3% of the wider syndrome of ALS, but the boundary of PLS with upper motor neuron-predominant ALS is more distinct in comparison to where PMA becomes lower motor neuron-predominant ALS. Patients with PLS have a consistently younger mean age of onset and the condition is often not life-shortening in contrast to ALS. The consistent presence in PLS of neuronal cytoplasmic aggregates of the 43 kDa transactive response DNA-binding protein (TDP-43) is uncertain, but other neuropsychological studies have reported FTD-spectrum cognitive impairment in PLS.9

The diagnostic taxonomy of ALS defined in El Escorial in Spain, and subsequently revised in Airlie in the USA and Awaji-shima in Japan, relies in simple terms on the number of body regions where clinical upper motor neuron signs are found in combination with clinical or electromyographic evidence of lower motor neuron-related muscle denervation. Electromyography (EMG) is only 60% sensitive however,10 with PMA and PLS regarded as outliers in this scheme. The categories of ‘possible’ ALS (one region with combined involvement) and ‘probable’ ALS (two regions) are all too easily interpreted by patients as reflecting diagnostic uncertainty on the part of the clinician, whereas in reality they reflect the insensitivity of both the clinical examination11 and EMG. Those labelled ‘possible’ ALS, as well as PMA and PLS cases, typically find themselves outside the inclusion criteria for therapeutic trials. ‘Possible’ ALS cases may never move out of this category despite succumbing to the same pattern of functional decline as those with ‘definite’ ALS.12

Archaic terms such as PMA and others (eg, Progressive Bulbar Palsy), for which we have lacked consistent definitions, are unhelpful distractors within a syndrome that can be comfortably unified by the single term ALS. Even the synonymous term motor neuron disease has limited recognition outside the UK and is increasingly anachronistic for a condition that consistently involves extramotor pathology and is not a single disease in genetic aetiological terms. The case for PLS being subsumed is less certain given its highly eccentric natural history and pending additional histopathological characterisation.

There is undoubtedly a need for improved prognostic stratification at therapeutic trial enrolment to maximise the chance of detecting benefit. Upper versus lower motor neuron burden, along with the extent of cognitive and behavioural impairment, has a role in this, alongside site of symptom onset. Stratification might be more meaningfully based on the rate of symptom progression and is increasingly likely to involve genotyping. A diagnostic biomarker based on the signature of TDP-43 dysregulation common to 97% of all ALS cases is highly desirable but elusive.13 Nonetheless, this study also makes it timely to reflect hard on whether the diagnostic rubric forged in El Escorial, and its derivatives, is fit for the future.