Early respiratory decline around diagnosis and short-term post-landmark outcomes in amyotrophic lateral sclerosis: a 6-month landmark cohort study

In amyotrophic lateral sclerosis (ALS), respiratory decisions rely on serial trends rather than a single value. We evaluated whether early respiratory decline around diagnosis provides prognostic information in a real-world landmark framework.

This single-center retrospective cohort screened 94 consecutive patients diagnosed between April 2019 and December 2025. A 6-month landmark was used. Early decline was estimated from %FVC values between − 30 and + 180 days around diagnosis. The primary model included age and early %FVC decline; robustness analyses included time-varying Cox, piecewise Cox, RMST, included-vs-excluded comparison, death-only analysis, and slope-quality filtering.

Of 94 screened patients, 62 met baseline eligibility, 56 had calculable early slope, and 45 entered the landmark cohort; 28 post-landmark composite events occurred. In the Cox model, faster early %FVC decline was associated with higher hazard of death or invasive mechanical ventilation via tracheostomy (HR 1.33 per 1%/month faster decline, 95% CI 1.14–1.55, p < 0.001). PH diagnostics suggested non-proportionality (%FVC p = 0.031; NIV p = 0.034 in the expanded model), so this HR was interpreted as an average follow-up association and complemented by PH-robust analyses. The signal was stronger early than late, remained consistent in a death-only analysis, and favored the slower-decline group by RMST at 24 and 36 months.

In this selected measurement-capable landmark cohort, early respiratory decline provided a clinically meaningful short-to-medium term prognostic signal for post-landmark adverse outcomes. External validation is required before broader generalization beyond measurement-capable landmark populations.