A Nomogram for Predicting Severe Exacerbations in Stable COPD Patients

Figures & data

Table 1 Baseline Characteristic of Study Population

	Total COPD Cohort (N=1711)	Severe Exacerbations During Follow-up		P-value
		No (N=1188)	Yes (N=523)
Demographics
Age (years)	65.2 ± 8.0	65.6 ± 7.9	64.4 ± 8.2	0.005
Male	984 (57.5%)	713 (60.0%)	271 (51.8%)	0.002
BMI (kg/m^2)	27.3 ± 5.3	27.4 ± 5.2	27.0 ± 5.5	0.176
Clinical Data
GOLD Stage				<0.001
1	367 (21.4%)	315 (26.5%)	52 (9.9%)
2	764 (44.7%)	571 (48.1%)	193 (36.9%)
3	409 (23.9%)	224 (18.9%)	185 (35.4%)
4	171 (10.0%)	78 (6.6%)	93 (17.8%)
Current smoker	572 (34.0%)	388 (33.2%)	184 (35.9%)	0.286
Smoking pack-years, median (IQR)	46.0 (35.0–62.0)	45.0 (35.0–60.0)	48.0 (36.2–63.8)	0.235
Severe exacerbations history	259 (15.4%)	99 (8.5%)	160 (31.1%)	<0.001
Comorbidities
History of coronary artery disease	179 (10.6%)	121(10.3%)	58(11.3%)	0.548
History of congestive heart failure	53 (3.1%)	28 (2.4%)	25 (4.9%)	0.007
History of GERD	534 (31.6%)	355 (30.2%)	179 (34.9%)	0.054
History of diabetes	230 (13.6%)	145 (12.3%)	85 (16.5%)	0.020
History of asthma	381 (22.8%)	220 (19.0%)	161 (31.3%)	<0.001
History of sleep apnea	321 (19.9%)	212 (18.9%)	109 (22.3%)	0.115
Physiology
Post-bronchodilator FEV1 (% predicted)	60.8 ± 23.0	65.4 ± 22.3	50.5 ± 21.1	<0.001
CAT	15.5 ± 8.0	14.0 ± 7.8	18.7 ± 7.7	<0.001
mMRC score	1.3 ± 1.0	1.1 ± 1.0	1.6 ± 1.0	<0.001
Inflammation Biomarker
White blood cell, median (IQR), (10^9/L)	7.0 (5.9–8.4)	6.9 (5.8–8.2)	7.2 (6.0–8.7)	<0.001
Treatment
Currently using oral corticosteroids	60 (3.6%)	25 (2.1%)	35 (6.8%)	<0.001
Inhaled bronchodilators	1120 (66.2%)	691 (58.7%)	429 (83.5%)	<0.001
Inhaled corticosteroids	793 (46.8%)	465 (39.4%)	328 (63.7%)	<0.001

Abbreviations: GERD, gastroesophageal reflux disease; CAT, COPD assessment test; mMRC, modified Medical Research Council.

Figure 1 Study flow chart. SPIROMICS: Subpopulations and Intermediate Outcome Measures in COPD Study.

Table 2 Identified Predictors of Severe Exacerbations from Cox Proportional Hazards Model, N=1438

	HR	95%CI	P-value
BMI, <21 kg/m^2	1.49	(1.11–2.00)	0.0076
Severe exacerbations history	1.90	(1.54–2.35)	<0.0001
Post-Bronchodilator FEV1 (% predicted)
≥80
50–79	1.68	(1.20–2.35)	0.0024
30–49	2.53	(1.79–3.58)	<0.0001
<30	3.33	(2.26–4.92)	<0.0001
Comorbidity Index
0
1	1.29	(1.04–1.59)	0.0194
2	1.89	(1.45–2.47)	<0.0001
≥3	3.53	(2.09–5.97)	<0.0001
CAT	1.05	(1.03–1.06)	<0.0001
White blood cell, ≥10×10^9/L	1.35	(1.03–1.77)	0.0318

Figure 2 A nomogram to predict the risk of severe exacerbations in stable COPD patients. To use the nomogram, draw a vertical line to identify corresponding points of each variable according to their actual status. Then sum up the points of all variables and find the position on the total points axis. With the same line mentioned above, you can determine the risk of severe exacerbations of 3- and 5-year at the lower line of the nomogram. Comorbidity index is identified through counting the number of comorbid diseases one had among congestive heart failure, GERD, diabetes, and asthma.

Table 3 C-Index of Nomogram, ADO, BODE and DOSE Risk Scores

Prediction Models	C-Index	95%CI
Nomogram	0.74	0.71–0.76
ADO	0.69	0.66–0.71
BODE	0.69	0.66–0.72
DOSE	0.70	0.68–0.73

Figure 3 Calibration plot showing predicted probability of severe exacerbations against the observed proportion of outcomes. The dashed line is the ideal calibration line. Vertical lines in grouped observation represents 95%CI. All groups of predicted probabilities fitting close to the ideal calibration line show perfect calibration. (A) 3-year incidence; (B) 5-year incidence.

Figure 4 Decision curves.