Amyotrophic lateral sclerosis (ALS): an animated, cited guide

This page needs JavaScript for its animations. The outline and key answers are below, and every claim in the full version carries a numbered citation.

Not medical advice. ALS Companion is an educational synthesis of published literature. It does not replace a clinician.

What is in it

What ALS is Who gets it, the numbers, and the ways it starts.
Upper and lower motor neurons Two neurons, two sets of signs, animated.
TDP-43 and what goes wrong A protein leaves the nucleus and clumps.
Genes: C9orf72, SOD1 and more Familial ALS, repeat expansions and genetic testing.
Diagnosis and criteria The Gold Coast, revised El Escorial and Awaji criteria, as a calculator.
The differential Conditions that can look like ALS, and how they differ.
ALSFRS-R and the course The 12-item scale, the progression rate and the stages.
Riluzole, edaravone and others What each drug targets and what the trials found.
Tofersen An antisense drug that lowers SOD1, for SOD1-ALS.
Breathing and NIV Weak breathing muscles, the tests, and when ventilation is offered.
Swallowing and feeding tubes Why weight matters and when a gastrostomy is discussed.
Living with ALS Communication, cognition, symptoms, planning and end of life.
Nursing care Assessment, giving the medicines, teaching and documentation.
Toward a cure Every route under study, with the evidence for each.
History and people From Charcot to tofersen.
Flashcards, quiz and glossary Revision for students.
Sources Every reference behind the guide.

Part 2 · Diagnosis & course

ALSFRS-R and the course

The ALSFRS-R adds up 12 everyday abilities into one score from 48 down to 0. Press Play to follow three illustrative courses past the milestones of King's staging, then score the 12 items yourself in the calculator below.

Timeline

Domain order

The lines are arithmetic, 48 minus rate × months, using the lower quartile, median and upper quartile of ΔFS at diagnosis in a French cohort: 0.5, 1.0 and 2.0 points a month. ²¹⁷ For comparison, PRO-ACT trial participants fell by 1.02 a month on average ²⁹ and a UK and Australian clinic cohort by 0.58. ⁶⁶ They are illustrations, not observed courses. Stage bands and event markers are mean months from onset in people who had died in a London clinic cohort: means, not medians, and other cohorts differ. ⁴⁵¹ The order in which the four domains fall is schematic.

ALSFRS-R calculator

This calculator shows how the published scale, the ΔFS formula and published cut-offs classify a set of item scores. It does not diagnose ALS and it is not a prognosis for any person. The ALSFRS-R is copyrighted and is scored with its own wording; the descriptions here are short paraphrases of each item's topic and levels, written for this guide. Each item is scored 0 to 4, so the total runs from 0 to 48, where 48 is normal function. ^66,79,94

Load an example

Months since symptom onset

Gastrostomy in place: item 5 is then scored as 5b, handling the tube and fastenings, instead of 5a, cutting food. ⁹

Prognosis: the ENCALS model

ENCALS is a published model that sorts people into five risk groups. The figures below are medians for groups of people in the development data, as published. They describe populations, not what will happen to one person, and the model's own inputs come from a clinical assessment.

Data came from 14 ALS centres in nine European countries, for people seen between 1992 and 2016. The endpoint combined three events, counted from symptom onset: non-invasive ventilation for more than 23 hours a day, tracheostomy, or death. Of 16 candidate predictors, eight entered the model. ⁵³⁴

Tested in external data, the model separated longer from shorter times fairly well (c statistic 0.78, 95% CI 0.77-0.80) and its predictions matched observed times (calibration slope 1.01). ⁵³⁴

Predictor	Univariable HR	Reading the hazard ratio
Bulbar onset	1.71	Above 1: endpoint sooner
Age at onset (per year)	1.03	Above 1: endpoint sooner
Definite against probable or possible ALS	1.47	Above 1: endpoint sooner
Diagnostic delay	0.52	Below 1: a longer delay goes with a later endpoint
FVC (forced vital capacity)	0.99	Below 1: higher FVC goes with a later endpoint
Progression rate (ΔFRS)	6.33	Above 1: faster decline, endpoint sooner
Frontotemporal dementia	1.34	Above 1: endpoint sooner
C9orf72 repeat expansion	1.45	Above 1: endpoint sooner

Hazard ratios as published. ⁵³⁴

Bars: predicted medians of 17.7, 25.3, 32.2, 43.7 and 91.0 months, as given in a 2022 review. ¹⁸⁰ Marks: observed medians of 16.5, 25.2, 32.8 and 44.6 months in the first four groups. ⁵³⁴

In plain wordsA model like this groups people who share similar features. Even inside one group, some people reach the endpoint much sooner and some much later than the group's middle value.

What the numbers mean, and their limits

How much change matters

In a UK and Australian clinic cohort, the minimal important difference over about 3 months was 3.8 points on the usual summed scale (2 on an interval version); other methods gave 2.02 to 5.43. ⁶⁶

90% of 65 surveyed clinicians and researchers rated a 20% slowing of decline as at least somewhat meaningful, and 93% rated a 50% slowing as very meaningful. ⁹¹ In a survey of 522 people with ALS, the median minimum slowing expected from a drug was 20%. ³³⁰

One total, or four domains?

In an Irish register cohort, bulbar and motor subscore slopes fitted prognostic models better than the total slope, and the respiratory subscore added nothing once the other two were known. The authors argue for reporting domains. ⁴⁵⁵

The scale is analysed as if each point were equal, although its items are categories; an interval (Rasch) transformation has been proposed. ⁶⁶

Trials and clinics differ

Pooled trial data average about 1 point a month (PRO-ACT 1.02, ±2.3). ²⁹ A clinic cohort averaged 0.58, and fast progressors were more likely to drop out of follow-up. ⁶⁶ Survival from onset ranged from 4 months to 11.9 years in one German cohort. ²⁷⁰

King's and MiToS measure different things

King's counts regions involved and then the need for gastrostomy or NIV; MiToS counts domains in which independence is lost. In LiCALS trial data they agreed only fairly (weighted kappa 0.21): King's gives more detail early, MiToS late, and the authors propose using both. ^111,168,451

Losing at least one MiToS domain, or an ALSFRS-R fall of 6 to 9 points, over 6 months was the best cut-off for predicting death, tracheotomy or more than 23 hours a day of NIV at 12 to 18 months. ⁴⁹⁴

Exam hookALSFRS-R: 12 items × 0-4 = 48. ΔFS = (48 − score) / months since onset; a higher ΔFS at diagnosis predicts shorter survival. King's 1-3 count regions, 4A is gastrostomy, 4B is NIV. ^265,280,451

In plain wordsA 48-point score tracks everyday abilities. How quickly it falls says more about the future than any one score, and the fall is rarely a perfectly straight line.