Gordon et al. 2006 — The Natural History of Primary Lateral Sclerosis
Before 2006, PLS had no large prospective natural history study. Clinicians knew PLS was slower than ALS and that it primarily affected upper motor neurons, but the actual rate of functional decline — measured, quantified, across a real patient cohort — had never been established. Gordon and colleagues changed that. Their analysis of 15 years of NEALS registry data gave the field its first durable clinical benchmark.
What they did
The Northeast ALS Consortium (NEALS) had maintained a prospective registry of motor neuron disease patients across multiple academic centers in the United States. Gordon and colleagues analyzed the PLS patients within that registry — approximately 250 individuals — followed over a 15-year period. Data collection included clinical features at onset, functional status measured by the ALS Functional Rating Scale — Revised (ALSFRS-R), and survival.
This was not a randomized trial and not an interventional study. It was an observational cohort — the appropriate design for understanding what a disease does when left to its natural course. The strength of the study was its size (unprecedented for PLS at the time), its duration (15 years of follow-up), and its use of the NEALS infrastructure, which had already been validated for ALS research and could be applied directly to PLS.
The ALSFRS-R was not designed for PLS specifically — it includes items for respiratory function and swallowing that are rarely affected in PLS. Nevertheless, it provided a standardized, validated functional scale that allowed comparison with ALS data and permitted tracking of functional decline over time. This methodological choice would later motivate the development of the PLSFRS (PLS Functional Rating Scale), which is more specific to PLS.
What they found
The average annual rate of functional decline in PLS was approximately 1.6 ALSFRS-R points per year. For context, ALS typically declines at 4–6 ALSFRS-R points per year. This means PLS patients in this cohort declined at roughly one-quarter to one-third the rate of ALS patients — a clinically meaningful difference that validated the longstanding clinical impression that PLS was a substantially slower disease.
Symptoms began in the lower limbs in the majority of cases — consistent with the 56–84% lower limb onset figures cited in later reviews. Corticobulbar onset (affecting speech and swallowing from the start) was less common but not rare. Survival was markedly longer than ALS. Patients lived for years to decades after diagnosis, and the study confirmed that pure PLS — without lower motor neuron involvement — carried a fundamentally different prognosis.
The cohort also demonstrated that a small but significant proportion of patients initially meeting PLS criteria eventually developed lower motor neuron signs and received an ALS diagnosis. This conversion rate — later quantified at approximately 23% within the first four years — was a key finding motivating the diagnostic threshold that would eventually be formalized in the 2020 consensus criteria.
Why it matters
Gordon 2006 is the paper that every subsequent PLS natural history study either replicates, updates, or explicitly builds upon. The 1.6 ALSFRS-R points per year figure appears in the 2025 Scirocco review, in the PNHS design documents, in the 2020 consensus criteria background, and in virtually every PLS clinical study published in the past 18 years. It is the field's reference standard.
Beyond the specific numbers, this study established several things that were not previously agreed upon: that PLS is distinctly slower than ALS at a population level; that it can be reliably tracked using standardized functional scales; that the NEALS Consortium infrastructure was capable of supporting PLS-specific research; and that a large prospective PLS cohort was feasible to assemble and maintain.
For patients and families, this study is the scientific basis for the clinical statement that PLS has a fundamentally different prognosis than ALS. When a neurologist says "PLS is much slower than ALS," the 1.6 points per year figure from Gordon 2006 is part of what underlies that statement.
The study also highlighted the absence of disease-modifying therapy. At 15 years of follow-up, decline was measurable but slow — slow enough that any future clinical trial would face the challenge of detecting a treatment effect against a baseline rate of change that was already modest. This problem of trial design in a slowly progressive disease remains the central challenge for PLS research as of 2025.
Limitations
Several limitations are worth noting. The ALSFRS-R was designed for ALS, not PLS — it includes subscales for swallowing, respiratory function, and handwriting that are rarely impaired in early PLS, which compresses the measurable range and may underestimate disease-specific functional impact. This was one motivation for later development of the PLSFRS. The cohort was drawn from NEALS academic centers in the United States, introducing selection bias — patients at specialty centers may differ systematically from the broader PLS population. And 15 years of registry data, while substantial, included patients diagnosed before modern neuroimaging and EMG standards, which may have included some cases that would today be reclassified.
How this connects
The 1.6 ALSFRS-R points per year figure established here is updated by Scirocco 2025, which synthesizes two decades of subsequent evidence and cites consistent findings across independent cohorts. The concern about ALS conversion motivated the diagnostic thresholds now formalized in the 2020 consensus criteria. The PNHS Mayo Clinic study represents the methodological successor — a prospective, multicenter study with biospecimen collection that Gordon 2006 could not include. For what the progression rate means in practice, see the Prognosis page. For the research infrastructure that produced this study, see the Natural History hub.
Citation
Gordon PH, Cheng B, Katz IB, Pinto M, Hays AP, Mitsumoto H, Rowland LP. The natural history of primary lateral sclerosis. Neurology. 2006;66(5):647–653. doi:10.1212/01.wnl.0000201254.89710.20