TABLE 4

Sensitivity analysis in EpiHealth

Protein	Abbreviation	UniProtNo	EpiHealth^#		EpiHealth^¶
			FEV₁ % predicted		FEV₁ % predicted
			β-coefficient (95% CI)	Adjusted p-value	β-coefficient (95% CI)	Adjusted p-value
Leptin	LEP	P41159	−4.80 (−5.87– −3.74)	4.9×10⁻¹⁶	−3.44 (−4.46– −2.41)	1.6×10⁻⁸
Interleukin-6	IL-6	P05231	−1.86 (−2.54– −1.18)	5.8×10⁻⁶	−1.78 (−2.54– −1.03)	0.0003
Paraoxonase 3	PON3	Q15166	1.87 (1.12–2.61)	4.5×10⁻⁵	0.91 (0.69–1.75)	0.27*
Interleukin-1 receptor antagonist	IL-1RA	P18510	−1.54 (−2.32– −0.77)	0.003	−1.16 (−2.07–0.46)	0.17*
Fatty acid-binding protein 4	FABP4	P15090	−2.61 (−3.51– −1.72)	1.4×10⁻⁶	−2.3 (−3.27– −1.33)	0.0003
Agouti-related protein	AGRP	O00253	1.01 (0.35–1.67)	0.04	0.7 (−0.07–1.46)	0.36*
Growth hormone	GH	P01241	1.15 (0.39–1.92)	0.04	1.21 (0.33–2.08)	0.11*
Plasminogen activator inhibitor	KIM1	Q96D42	−1.30 (−2.00– −0.60)	0.007	−1.37 (−2.20– −0.55)	0.025
C-C motif chemokine 16	CCL16	O15467	−1.23 (−1.91– −0.54)	0.01	−0.90 (−1.66– −0.14)	0.22*
Metalloproteinase inhibitor 4	TIMP4	Q99727	−1.20 (−1.89– −0.50)	0.01	−1.36 (−2.15– −0.57)	0.023
Retinoic acid receptor responder 2	RARRES2	Q99969	−1.96 (−2.68– −1.24)	5.8×10⁻⁶	−1.68 (−2.46–0.9)	0.001
Receptor for advanced glycosylation and products	RAGE	Q15109	1.02 (0.36–1.68)	0.04	0.92 (0.17–1.67)	0.2*
Trefoil factor 3	TFF3	Q07654	−1.08 (−1.77– −0.39)	0.04	−1.30 (−2.08– −0.53)	0.025
Matrix metalloproteinase 7	MMP7	P09237	−1.15 (−1.85– −0.45)	0.02	−0.81 (−1.65–0.02)	0.31*
Plasminogen activator inhibitor	PAI	P05121	−1.00 (−1.66– −0.34)	0.04	−0.76 (−1.50– −0.01)	0.31*
Adrenomedullin	ADM	P35318	−1.92 (−2.69– −1.15)	4.5×10⁻⁵	−1.65 (−2.54– −0.76)	0.01
Growth/differentiation factor 15	GDF15	Q99988	−1.36 (−2.15– −0.57)	0.01	−1.58 (−2.54–0.63)	0.03
Tissue-type plasminogen activator	tPA	P00750	−1.56 (−2.28– −0.58)	0.0006	−1.52 (−2.32– −0.73)	0.009
Adhesion G protein-coupled receptor G2	ADGRG2	Q8IZP9	1.64 (0.95–2.32)	4.5×10⁻⁵	1.34 (0.56–2.11)	0.02

Cross-sectional associations between plasma proteins and FEV₁% predicted in EpiHealth before and after exclusion of individuals with known CVD, diabetes or obesity. Multivariable linear regression analysis for each protein. Data are presented as regression coefficients for proteins increased by at least 1 standard deviation. Adjusted for age, sex, cohort, body mass index, smoking, pack-years and physical activity. For the sensitivity analysis this multivariable linear regression model was adjusted for hypertension, hyperlipidaemia and waist circumference (normal versus high) instead of BMI. FEV₁: forced expiratory volume in 1 s; CI: confidence interval; CVD: cardiovascular disease. ^#: n=2229; ^¶: n=1685; *: non-significant.