C-Reactive Protein at Discharge, Diabetes Mellitus and ≥1 Hospitalization During Previous Year Predict Early Readmission in Patients with Acute Exacerbation of Chronic Obstructive Pulmonary Disease

Figures & data

Figure 1. Study flow diagram.

Table 1. General characteristics of enrolled patients

Variables	All patients	Non-Readmitted Patients	Readmitted Patients	p
No (%)	125 (100)	96 (76.8)	29 (23.2)	–
Age, y	69.2 ± 9.8	68.6 ± 10.1	71.0 ± 8.6	0.318
Male,%	93.6	92.7	96.5	0.108
Smoking habit
Current/former,%	28/72	29.2/70.8	24.1/75.9	0.459
FEV1, litres	1.12 (0.88–1.48)	1.12 (0.89–1.62)	1.08 (0.87–1.36)	0.629
FEV1,% predicted	45.9 (34.8–55.2)	46 (36–60)	47 (33.7–52.4)	0.209
FEV1/FVC,%	46.5 ± 12.0	47.3 ± 12.3	44.0 ± 11.0	0.219
COPD-GOLD stage
Mild/moderate/severe/very severe,%	0.9/35.5/35.5/28	1.2/39.5/33.3/25.9	0/23.1/42.3/34.6	0.384
LTOT,%	18.4	17.7	20.7	0.833
MRC dyspnea score ≥ 2,%	85.6	85.4	86.2	0.880
Exacerbations in the preceding year
Patients with ≥ 2 events,%	29.7	24.2	48.1	0.017
Patients with ≥ 4 events,%	10.4	4.1	31	<0.001
Rate (events/patients)
Requiring antibiotics	1.36	1.09	2.27	<0.001
Requiring hospitalizations	1.08	0.83	1.93	<0.001
Chronic co-morbidities
Congestive heart failure	27 (21.6)	21 (21.8)	6 (20.7)	0.988
Chronic renal failure	1 (0.8)	1 (1)	0 (0)	0.433
Diabetes	17 (13.6)	9 (9.3)	8 (27.6)	0.008
NIDDM/IDDM	7/10	4/5	3/5	1.000
Non-cirrhotic liver disease	3 (2.4)	2 (2)	1 (3.4)	0.676
Neurologic disease	8 (6.4)	6 (6.2)	2 (6.9)	0.939
Non-active cancer	12 (9.6)	7 (7.2)	5 (17.2)	0.117
Use of inhaled bronchodilators
SABA	84 (67.2)	61 (63.5)	23 (79.3)	0.177
Salbutamol/terbutaline	79/5	58/3	21/2	0.670
LABA	77 (61.6)	61 (63.5)	16 (55.2)	0.428
Formoterolo/salmeterol	25/52	19/42	6/10	0.656
Anticholinergics	84 (67.2)	64 (66.7)	20 (69)	0.811
Ipatropium/tiotropium	33/51	24/40	9/11	0.649
Use of inhaled corticosteroids	74 (59.2)	56 (58.3)	18 (62.1)	0.890
Budesonide/fluticasone	32/42	26/30	6/12	0.498
Use of theophylline	9 (7.2)	6 (6.2)	3 (10.3)	0.451

Values are reported as mean ± SD, as median (25th -75th percentiles) or as frequency (%).

^†: evaluated at baseline.

Abbreviations: FEV₁ indicates forced expiratory volume in the 1st second; FEV₁/FVC, ratio of FEV₁ to forced vital capacity (FVC); GOLD, global initiative for chronic obstructive lung disease guidelines; LTOT, long-term oxygen therapy; MRC, medical research council dyspnea score; NIDDM and IDDM, non and insulin-dependent diabetes mellitus; SABA, short action ß₂ agonist; LABA, long action ß₂ agonist.

Table 2. Clinical and microbiological variables evaluated at admission and outcomes

Variables	Non-Readmitted Patients	Readmitted Patients	p
Fever,*%	40.6	34.5	0.549
Chills,%	28.1	31	0.770
Cough,%	83.3	75.9	0.260
Characteristics of sputum Absence/mucoid/purulent/rusty,%	28.1/54.2/15.6/2.1	34.5/58.6/6.9/0	0.391
Classification of AECOPD according to Anthonisen criteria,% Type I/Type II/Type III	41.8/29.7/28.6	34.5/31/34.5
pH	7.40 (7.37–7.43)	7.41 (7.36–7.44)	0.374
PCO2, mmHg	47 (41.1–56)	47 (41.3–54.5)	0.547
PaO2/FiO2	257.1 (227.6–280.9)	241.9 (201.4–262.3)	0.046
HCO3, mmol/L	28 (25.3–31.2)	29.3 (27–34.1)	0.082
BE, mmol/L	3 (1–5.9)	4 (2.7–7)	0.087
Confirmed bacterial etiology, n/n total of patients for each group in whom sputum was provided (%)	38/91 (41.7)	10/27 (37.0)	0.661
Gram-positive bacteria
Streptococcus pneumoniae	4 (10.5)	1 (10)	1.000
Streptococcus viridans	6 (15.8)	1 (10)	1.000
Staphylococcus aureus	4 (10.5)	0 (0)	0.572
Gram-negative bacteria
Haemophilus influenzae	9 (23.7)	3 (30)	1.000
Haemophilus parainfluenzae	4 (10.5)	0 (0)	0.572
Pseudomonas aeruginosa	6 (15.8)	3 (30)	0.432
Moraxella catarrhalis	1 (2.6)	0 (0)	1.000
Chlamydia pneumoniae	1 (2.6)	0 (0)	1.000
Enterobacteriaceae	2 (5.3)	1 (10)	0.550
Legionella pneumophila	0 (0)	1 (10)	0.232
Neisseria spp.	1 (2.6)	0 (0)	1.000
LOS, days	7 (5–10)	8 (6-13)	0.368
Use of systemic corticosteroids,%	86.5	100	0.023
Total doses of systemic corticosteroids, mg	300 (216–406)	332 (220–600)	0.214
Use of antibiotics,%	78.1	86.2	0.476
Duration of antibiotics treatment, days	7 (5–9)	8 (6–12)	0.017
Number of classes of antibiotics used,% Only one class/two classes	77.3/22.7	72/28	0.767
Classes of antibiotics used
Penicillins	39 (48.1)	11 (42.3)	0.795
Fluoroquinolones	36 (44.4)	13 (50)	0.479
Macrolides	10 (12.3)	3 (11.5)	1.000
Cephalosporins	15 (18.5)	5 (19.2)	0.780
Carbapenems	1 (1.2)	1 (3.8)	0.412
Use of both systemic corticosteroids and antibiotics,%	80.2	89.7	0.241
Require of NIMV,%	14.5	13.8	0.732
Mental status alteration,%	7.2	13.8	0.290
Admission to ICU,%	3.1	3.4	0.860

Values are reported as median (25th–75th percentiles) or as frequency (%). For microbiological data values are reported as frequency and percentage (%) related to the number of patients with etiological diagnosis in each group.

^‡Including Escherichia coli and Proteus mirabilis.

^§Calculated on patients with use of antibiotics only.

Abbreviations: PaCO₂ indicates partial arterial carbon dioxide pressure; PaO₂/FiO₂, ratio of partial arterial oxygen pressure to the fraction of inspired oxygen; HCO₃, serum bicarbonate; BE, base excess; LOS, length of hospital stay; NIMV, non-invasive mechanical ventilation; ICU, intensive care unit.

Table 3. Serum laboratory data measured at admission

Parameters	n	Non-Readmitted Patients	Readmitted Patients	p
Total leukocytes, counts × 10^9/L	n = 125	10.15 (7.7–12.8)	9.1 (7.5–13.5)	0.670
Hematocrit,%	n = 125	45 (42–49)	45 (41–48.1)	0.529
Glucose, mg/dL	n = 125	118 (104.2–150.2)	132.7 (109–161.5)	0.262
Creatinine, mg/dL	n = 125	0.9 (0.7–1.1)	1.1 (0.8–1.2)	0.081
Sodium, mmol/L	n = 125	136.8 (134.2–139)	137 (135–140)	0.480
Potassium, mEq/L	n = 124	4.4 (4–4.7)	4.5 (4.1–4.7)	0.529
Aspartate transaminase-AST, IU/L	n = 108	22 (18–30)	22 (19–29)	0.847
Alanine transaminase-ALT, IU/L	n = 110	19 (14–27.5)	17 (14–24.8)	0.785
Total bilirubin, mg/dL	n = 103	0.8 (0.6–1.1)	0.9 (0.7–1.3)	0.167
Creatine phosphokinase-CPK, mcg/L	n = 100	72.5 (44.5–110)	70.5 (35–128.9)	0.756
Lactate dehydrogenase-LDH, IU/L	n = 107	351 (274–460)	331 (244–460)	0.702

Values are reported as median (25th-75th percentiles).

Figure 2. Panel of inflammatory response collected at admission, at day 3 and at discharge. White and grey boxplots represent non-readmitted and readmitted AECOPD patients, respectively. CRP indicates C-reactive protein; PCT, procalcitonin; TNF-α tumour necrosis factor-alpha; IL, interleukin. The horizontal bar and box length represent the median and the 25th–75th percentiles, respectively. Circles indicate outliers. * and **: p value <0.05 and 0.001 between groups, respectively.

Table 4. Logistic regression analysis

Risk factors	OR	SE	95% CI	p	OR	SE	95% CI	p
CRP at discharge, ≥7.6 mg/L	4.68	0.52	1.68 to 13.06	0.003	7.41	0.87	1.34 to 40.91	0.022
IL-6 at discharge, ≥19.5 pg/mL	3.27	0.48	1.26 to 8.43	0.014	4.84	0.82	0.95 to 24.51	0.056
Number of co-morbidities	1.38	0.19	0.95 to 2.02	0.086	1.34	0.23	0.84 to 2.14	0.215
Diabetes, presence	3.64	0.54	1.25 to 10.56	0.017	11.03	0.93	1.77 to 68.54	0.010
Hospitalization for AECOPD, ≥1 events	4.70	0.54	1.62 to 13.57	0.004	8.04	0.82	1.61 to 40.17	0.011
PaO2/FiO2	0.99	0.00	0.98 to 1.00	0.075	0.98	0.00	0.97 to 0.99	0.135
Duration of antibiotics treatment, days	1.10	0.03	1.02 to 1.19	0.011	1.08	0.06	0.96 to 1.22	0.186

Dependent variable: readmission to hospital in a period of 30 days.

Hosmer and Lemeshow Tests: X² 10.47 p = 0.863.

^†Cut-offs obtained by receiver operating characteristic (ROC) analysis.

^‡Data evaluated during the preceding year.

Abbreviations: OR indicates odds ratio; SE, standard error; CI, confidence interval; CRP, C-reactive protein; IL, interleukin.

Table 5. Analysis of predicted values, likelihood ratios values, sensitivity and specificity for risk factors

Risk factors	n	True Prevalence	PPV (95% CI)	NPV (95% CI)	Sensitivity	Specificity	LR+	LR-
CRP at discharge ≥7.6 mg/L	n = 125	0.226	0.360 (0.292 to 0.427)	0.892 (0.850 to 0.933)	0.750	0.609	1.91/0.56	0.41/0.11
Presence of diabetes	n = 125	0.233	0.470 (0.348 to 0.591)	0.803 (0.764 to 0.841)	0.275	0.905	2.89/0.87	0.80/0.24
≥1 hospitalization for AECOPD during previous year	n = 125	0.226	0.333 (0.273 to 0.392)	0.903 (0.861 to 0.944)	0.807	0.528	1.70/0.49	0.36/0.10
CRP at discharge ≥7.6 mg/L and Presence of diabetes	n = 56	0.196	0.750 (0.596 to 0.903)	0.895 (0.850 to 0.939)	0.545	0.955	12.11/2.95	0.47/0.11
CRP at discharge ≥7.6 mg/L and ≥1 hospitalization for AECOPD during previous year	n = 42	0.381	0.714 (0.615 to 0.812)	0.952 (0.905 to 0.998)	0.937	0.769	4.05/2.49	0.08/0.05
Presence of diabetes and ≥1 hospitalization for AECOPD during previous year	n = 56	0.160	0.667 (0.509 to 0.824)	0.936 (0.900 to 0.971)	0.667	0.936	10.42/1.98	0.35/0.06
CRP at discharge ≥7.6 mg/L and Presence of diabetes and ≥1 hospitalization for AECOPD during previous year	n = 24	0.208	1.000 (1.000 to 1.000)	1.000 (1.000 to 1.000)	1.000	1.000	∞/∞	0.00/0.00

Abbreviations: PPV and NPV indicate positive and negative predictive values, respectively; LR+ and LR- indicate positive and negative likelihood ratios, respectively; CRP, C-reactive protein.

^†Data reported as conventional and weighted for prevalence.