Metabolic changes of different high-resolution computed tomography phenotypes of COPD after budesonide–formoterol treatment

Figures & data

Table 1 Demographics and clinical characteristics of COPD patients in the two phenotypes

Variables	Phenotype E (n=22)	Phenotype M (n=28)	P-value
Gender, male	20 (90.91%)	25 (89.29%)	0.849
Age, years	73.64±6.31	70.18±8.51	0.118
BMI (kg/m^2)	21.21±2.25	19.65±3.39	0.069
Current smoker	6 (27.27%)	8 (28.57%)	0.919
Smoking, pack/year	43.82±24.35	44.28±26.03	0.949
FEV^1/FVC% in pretreatment	48.53±9.35	50.92±11.26	0.427
FEV^1/FVC% in posttreatment	63.62±13.94a	74.43±14.73a,b	0.011
Change in FEV^1/FVC% in pre–posttreatment	14.93±3.24	25.56±4.12b	<0.001
FEV1% predicted in pretreatment	47.46±10.78	46.12±14.21	0.715
FEV1% predicted in posttreatment	62.49±11.32a	73.15±15.22a,b	0.009
Change in FEV1% predicted in pre–posttreatment	15.17±3.55	26.40±5.13b	<0.001
SGRQ total score in pretreatment	39.23±7.98	58.06±10.42b	<0.001
SGRQ total score in posttreatment	30.69±6.59a	38.38±6.51a,b	0.001
Change in SGRQ total score in pre–posttreatment	8.79±2.19	19.65±4.73b	<0.001

Notes: Data are presented as mean ± SD.

a P<0.05 compared with pretreatment.

b P<0.05 compared with phenotype E. Phenotype E, emphysema without bronchial wall thickening; Phenotype M, emphysema with bronchial wall thickening.

Abbreviations: BMI, body mass index; FEV₁, forced expiratory volume in 1 second; FVC, forced vital capacity; SGRQ, St George’s Respiratory Questionnaire.

Figure 1 (A) The spectrum shows a wide range of metabolites. Representative ¹H-NMR spectrum of plasma samples obtained from (B) phenotype E pretreatment, (C) phenotype E posttreatment, (D) phenotype M pretreatment and (E) phenotype M posttreatment.

Abbreviation: ¹H-NMR, ¹H nuclear magnetic resonance spectroscopy.

Figure 2 PCA (A–D) and PLS-DA (E–H) score plots demonstrated good separation between different groups.

Note: Ellipses display 95% confidence regions.
Abbreviations: EH, phenotype E posttreatment; EQ, phenotype E pretreatment; MH, phenotype M posttreatment; MQ, phenotype M pretreatment; PCA, principal component analysis; PLS-DA, partial least squares discriminant analysis.

Table 2 Summary of the parameters for assessing the quality of the PLS-DA model

Model	No	R^2X	R^2Y	Q^2
MH and MQ	100	0.029	0.624	0.338
EH and EQ	100	0.107	0.920	0.273
EQ and MQ	100	0.203	0.750	0.547
EH and MH	100	0.102	0.828	0.0743

Notes: No represents the number of principal component obtained by cross-validation; R²X, variation explanation in X; R²Y, variation explanation in the response to Y; Q², predict variation capability of compound.

Abbreviations: EH, phenotype E posttreatment; EQ, phenotype E pretreatment; MH, phenotype M posttreatment; MQ, phenotype M pretreatment; PLS-DA, partial least squares discriminant analysis.

Table 3 Quantitative comparison and the relative integrals of metabolites in the EQ, MQ, EH and MH groups

Compound	ppm	VIP	EQ (n=22)	MQ (n=28)	EH (n=22)	MH (n=28)
ADP	8.28	1.5864	(2.34±1.04) ×10^−4,a	(1.74±1.17) ×10^−4	(5.19±3.73) ×10^−4	(5.71±2.71) ×10^−4,b
Guanosine	7.98	1.2857	(3.44±1.03) ×10^−4,a	(3.28±0.92) ×10^−4	(3.68±1.45) ×10^−3	(3.22±1.35) ×10^−3,b
Tyrosine	7.16	1.6328	(1.52±1.13) ×10^−4,a	(1.28±1.19) ×10^−4	(2.62±1.20) ×10^−4	(2.52±1.26) ×10^−4,b
Uridine	5.88	1.5898	(1.14±0.93) ×10^−2	(1.59±1.19) ×10^−2	(1.05±0.78) ×10^−2	(0.95±0.58) ×10^−2,b
Maltose	5.22	1.0970	(1.20±1.24) ×10^−1	(1.45±1.03) ×10^−1	(1.52±1.41) ×10^−1	(1.15±0.94) ×10^−1
Sucrose	4.26	1.6069	(2.78±1.10) ×10^−1	(2.35±0.84) ×10^−1	(3.22±1.15) ×10^−1	(2.58±0.62) ×10^−1
l-Threonine	4.25	2.1840	(2.73±0.88) ×10^−1	(2.20±0.97) ×10^−1	(3.21±1.25) ×10^−1	(2.71±0.77) ×10^−1,b
d-Glucose	3.70	1.5992	6.25±1.39	5.18±0.61c	5.97±1.75d	4.94±0.99
Glycine	3.57	2.5918	(7.70±1.58) ×10^−1,a	(6.14±1.26) ×10^−1,c	(8.79±1.91) ×10^−1,d	(7.33±1.70) ×10^−1,b
Proline	3.33	1.9525	(1.99±0.54) ×10^−1,a	(1.61±0.99) ×10^−1	(4.80±0.77) ×10^−1,d	(5.34±0.98) ×10^−1,b
Betaine	3.24	1.4522	2.12±0.47	1.93±0.28	2.06±0.39	2.02±0.40
Choline	3.18	2.2438	(2.62±0.89) ×10^−1,a	(2.14±0.84) ×10^−1	(3.38±0.33) ×10^−1	(2.91±1.59) ×10^−1,b
Malonate	3.12	1.6003	(6.36±2.02) ×10^−3,a	(7.49±2.19) ×10^−3	(8.63±2.92) ×10^−3	(8.06±2.23) ×10^−3
l-Lysine	3.07	1.3124	(3.34±1.74) ×10^−1	(3.07±0.92) ×10^−1	(3.41±1.25) ×10^−1	(3.53±1.44) ×10^−1
Creatine	3.05	1.3495	(2.54±0.63) ×10^−1	(1.90±0.28) ×10^−1	(2.64±0.91) ×10^−1	(2.02±0.49) ×10^−1
Asparagine	2.86	1.1751	(2.51±1.12) ×10^−2	(2.11±1.78) ×10^−2	(2.91±1.61) ×10^−2	(2.69±1.25) ×10^−2
Aspartate	2.82	1.1520	(5.73±1.12) ×10^−2	(5.16±1.53) ×10^−2	(6.23±1.64) ×10^−2	(5.91±1.44) ×10^−2
Succinate	2.38	1.3575	(2.08±0.71) ×10^−1	(2.67±0.68) ×10^−1,c	(1.97±0.50) ×10^−1	(2.25±0.88) ×10^−1
Pyruvic acid	2.35	1.6409	8.77±2.76a	1.29±0.31c	7.37±1.37d	1.34±0.42
Acetone	2.22	1.5726	(5.20±2.06) ×10^−1,a	(4.89±0.79) ×10^−1	(4.07±1.48) ×10^−1	(3.79±0.57) ×10^−1,b
Ornithine	1.82	1.6560	(9.44±2.30) ×10^−3	(8.49±2.70) ×10^−3	(7.99±2.76) ×10^−3	(7.51±2.40) ×10^−3
l-Alanine	1.46	1.0100	(1.38±0.75) ×10^−1,a	(1.03±0.34) ×10^−1	(2.24±1.10) ×10^−1	(1.94±0.38) ×10^−1,b
Lactate	1.30	1.0361	9.76±4.29a	12.58±4.60c	7.29±2.57	6.38±2.82b
Isopropyl alcohol	1.16	2.4802	1.67±0.43	1.39±0.68	1.42±0.42	1.18±0.51
l-Valine	1.01	1.4451	(2.27±0.94) ×10^−1,a	(2.02±0.93) ×10^−1	(3.42±0.93) ×10^−1,d	(3.91±0.79) ×10^−1,b
Leucine	0.94	1.5185	1.04±0.12a	(9.78±1.55) ×10^−1	1.15±0.18d	1.25±0.13b

Notes: A total of 26 characteristic resonances were observed in plasma of the four groups (VIP value >1.0).

a P<0.05 compared with EH.

b P<0.05 compared with MH.

c P<0.05 compared with MQ;

d P<0.05 compared with MH. ppm, chemical shift.

Abbreviations: ADP, adenosine diphosphate; EQ, phenotype E pretreatment; MQ, phenotype M pretreatment; EH, phenotype E posttreatment; MH, phenotype M posttreatment; VIP, variable importance.

Figure 3 Results of the MSEA for COPD patients.

Notes: The horizontal bar graph shows the most affected metabolic pathways. The most significant differences include protein biosynthesis, amino acid metabolism, gluconeogenesis and glycolysis.
Abbreviation: MSEA, metabolite set enrichment analysis.