TABLE 1.

Predicted versus observed impact of anti-inflammatory treatments according to baseline biomarkers

	Included trial, control versus active intervention				Weighted mean reduction control versus active
	Novel START, salbutamol versus any low-dose ICS #,¶	CAPTAIN, FF100 versus FF200 #	QUEST, placebo versus dupi200	DREAM, placebo versus any mepo
Type-2 low: blood Eos <0.15×109 per L and FENO <25 ppb	n=18 versus 60	n=194 versus 211	n=106 versus 139	n=23 versus 63	n=341 versus 473
Observed
Control arm	0.05	0.27	0.56	1.98
Active arm	0.03	0.22	0.58	1.71
Reduction	41%	20%	−4%	14%	14%
Predicted
Control arm	0.07	0.19	0.53	0.74
Active arm	0.07	0.19	0.53	0.74
Reduction	0%	0%	0%	0%	0%
Type-2 high: blood Eos ≥0.15×109 per L or FENO ≥25 ppb	n=201 versus 377	n=903 versus 909	n=514 versus 484	n=145 versus 392	n=1763 versus 2162
Observed
Control arm	0.05	0.40	1.07	2.46
Active arm	0.03	0.26	0.41	1.19
Reduction	35%	35%	62%	52%	41%
Predicted
Control arm	0.13	0.32	0.93	1.28
Active arm	0.07	0.19	0.53	0.74
Reduction	42%	42%	42%	42%	42%
Type-2 very high: blood Eos ≥0.30×109 per L and FENO ≥50 ppb	n=51 versus 54	n=67 versus 71	NA	NA	n=118 versus 125
Observed
Control arm	0.13	0.62	NA	NA
Active arm	0.02	0.25	NA	NA
Reduction	81%	60%	NA	NA	69%
Predicted
Control arm	0.26	0.65	1.88	2.60
Active arm	0.07	0.19	0.82	1.14
Reduction	72%	72%	72%	72%	72%

Data are presented as annual severe asthma attack rate unless otherwise stated. Data from [4–6, 8], applied to the prototype scale reported in [2]. ICS: inhaled corticosteroids; FFx: fluticasone furoate x μg·day⁻¹; dupi200: dupilumab 200 mg over 2 weeks; mepo: mepolizumab; Eos: eosinophils; F_ENO: exhaled nitric oxide fraction; NA: not available. ^#: data of patients with a baseline F_ENO <20 ppb were regrouped into the <25-ppb group, as the difference of 5 ppb in F_ENO is not clinically relevant [13]; ^¶: only the percentage of patients with one or more severe attack(s) in the 52 weeks of follow-up was reported so a rate was imputed as −log₁₀(1−%incidence).