Table 1.

Predictor distribution per dataset.

Source	n	Gender (male/female)	Onset (spinal/bulbar)	Age (years)	Symptom duration (months)	Baseline weight (kg)	Baseline ALSFRS (score)	Baseline ALSFRS decline rate (score/month)
PRO-ACT	3,971	2,485/1,486	3,117/854	$56.2\pm 11.3$ (18:81)	$20.8\pm 12.7$ (0.5:140.4)	$74.8\pm 15.8$ (30:148.6)	$30.1\pm 5.7$ (7:40)	$-0.61\pm 0.51$ ($-$6.09:0)
Trophos	431	277/154	346/85	$56.7\pm 11.1$ (26:79)	$16.4\pm 8.0$ (5:38)	$71.5\pm 12.7$ (41:130)	$32.5\pm 4.1$ (16:40)	$-0.55\pm 0.39$ ($-$2.67:0)
Exonhit	172	118/54	129/43	$55.6\pm 12.0$ (26.3:77.9)	$24.7\pm 11.9$ (5:58)	$70.1\pm 13.8$ (45:112)	$27.5\pm 6.4$ (10:39)	$-0.60\pm 0.40$ ($-$3.14:$-$0.05)
Real world	646	345/301	458/188	$62.2\pm 12.1$ (26.3:92.2)	$22.1\pm 21.6$ (0:228.5)	$70.4\pm 13.2$ (40:140)	$28.6\pm 7.4$ (3.5:40)	$-0.78\pm 0.65$ ($-$4.16:0)
Overall	5,220	3,225/1,995	4,050/1,170	$57.0\pm 11.6$ (18:92.2)	$20.7\pm 13.9$ (0:228.5)	$73.8\pm 15.3$ (30:148.6)	$30\pm 5.9$ (3.5:40)	$-0.63\pm 0.52$ ($-$6.09:0)

Numerical predictors are described using mean ± standard deviation (range).