Subphenotypes of Mild-to-Moderate COPD by Factor and Cluster Analysis of Pulmonary Function, CT Imaging and Breathomics in a Population-Based Survey

Figures & data

Figure 1  Flowchart of subjects in the study. Flowchart of the study. *Due to logistic reasons. COPD criteria are based on the GOLD guidelines: COPD is defined as the presence of postbronchodilator FEV1/FVC ratio < 0.70.

Figure 2  Dendrogram. Hierarchical cluster analysis by Ward's method resulted in a dendogram identifying 4 clusters.

Table 1.  Demographics and clinical characteristics of the total cohort and COPD clusters identified by factor and cluster analysis

	Total COPD cohort	Cluster 1	Cluster 2	Cluster 3	Cluster 4	p-value*
Number of subjects	157	35	48	60	14
Sex (% female)	24	31	25	23	0	0.046*
Age (yr)	63.2 (5.7)	63.1 (5.7)	63.7 (6.2)	63.3 (5.7)	62.3 (4.2)	0.877
BMI (kg/m2)	26.8 (3.9)	26.5 (3.8)	26.7 (4.5)	26.6 (3.4)	27.9 (3.5)	0.705
Post-BD FEV1/FVC	0.61 (0.08)	0.62 (0.07)	0.56 (0.09)	0.64 (0.05)	0.60 (0.06)	<0.001*
Pre-BD FEV1 (%pred)	88 (18)	85 (13)	75 (15)	102 (13)	83 (13)	<0.001*
Post-BD FEV1 (%pred)	92 (17)	90 (13)	77 (14)	106 (10)	86 (12)	<0.001*
GOLD stage 1/2/3 (%)	70/27/3	63/37/0	46/44/10	95/5/0	64/36/0
Reversibility in FEV1 (%)	3.4 (6.4)	4.4 (4.5)	2.0 (6.6)	4.3 (7.2)	2.6 (5.6)	0.219
KCO (%)	79 (26)	80 (21)	75 (19)	76 (19)	75 (14)	0.642
Pack-years	45 (22)	38 (11)	45 (12)	36 (12)	96 (27)	<0.001*
Current smokers,%	55	66	50	53	50	0.156

*p-value from analysis of variance or Chi-square test between the clusters.

Numeric data is expressed as mean (SD). Post-BD: post-bronchodilator values after 400 ug salbutamol. Variables listed in table 3 were used for cluster formation.

Table 2.  Description of the total cohort and COPD clusters identified by factor and cluster analysis

	Total COPD cohort	Cluster 1	Cluster 2	Cluster 3	Cluster 4	p-value*
Number of subjects	157	35	48	60	14
Sputum production%	28	29	38	23	14	0.111
Chronic bronchitis%	22	14	31	20	14	0.062
Dyspnea at exercise%	53	54	58	43	71	0.060
Dyspnea at rest%	5	3	6	2	21	0.003*
Chronic unproductive cough%	11	14	13	7	14	0.301
2nd hand smoke%	24	20	29	17	14	0.632
Cancer, any form,%	5	6	4	2	7	0.258
Inhaled corticosteroid use%	7	6	10	7	0	0.324
Long-acting bronchodilator use%	9	14	13	5	0	0.163
Hypercholesterolaemia%	32	31	23	42	29	0.041*
Asthma%	10	14	8	5	14	0.216
Diabetes%	13	3	13	15	21	0.121
Cardiovascular disease%	20	17	21	22	21	0.676
Hypertension%	24	17	33	22	14	0.101
CCQ, median (IQR)	6 (4–11)	6 (3–10)	8 (5-12)	6 (3–11)	12 (5–16)	0.025*
Perc15 (HU), mean (SD)	–943 (22)	–918 (19)	-953 (13)	–950 (15)	–934 (18)	<0.001*
VI-HU910 (%), mean (SD)	42 (15)	23 (8)	49 (8)	46 (11)	35 (13)	<0.001*
VI-HU950 (%), mean (SD)	14 (9)	5 (2)	18 (7)	16 (7)	10 (6)	<0.001*
eNose PC1, mean (SD)	0.00 (1.00)	–0.05 (0.98)	-0.05 (0.98)	0.06 (1.07)	0.03 (0.90)	0.931
eNose PC2, mean (SD)	0.00 (1.00)	–0.07 (0.95)	-0.04 (1.09)	0.07 (0.99)	–0.01 (0.92)	0.903
eNose PC3, mean (SD)	0.00 (1.00)	–0.18 (0.83)	0.18 (1.08)	–0.05 (1.02)	0.08 (1.03)	0.410
eNose PC4, mean (SD)	0.00 (1.00)	–0.14 (0.96)	0.33 (0.93)	–0.18 (1.09)	–0.04 (0.73)	0.047*

CCQ: clinical COPD questionnaire. Perc15: 15^th percentile point. VI-HU910: Voxel index at a threshold of -910 Hounsfield Units. VI-HU950: Voxel index at a threshold of -950 Hounsfield Units. PCx: principal component x of exhaled breath electronic nose. Chronic bronchitis is defined as the presence of cough and sputum production for at least 3 months in each of 2 consecutive years. Variables listed in table 3 were used for cluster formation.

Table 3.  Loadings of variables per factor

Factor	Variable	Loading
1	Postbronchodilator FEV1%pred	0.934
2	CT emphysema score (perc15)	–0.955
3	Chronic bronchitis	–0.932
4	Use of long-acting bronchodilators	0.885
5	Dyspnea at rest	0.757
6	Cardiovascular disease	0.772
7	Body Mass Index	0.787
8	Diabetes	0.801
9	Reversibility in FEV1	–0.858
10	eNose PC1	0.715
11	eNose PC4	0.751
12	Pack-years	–0.605