Study design
A prospective analysis was performed on COPD patients who reported to a single outpatient clinic in Poland with a COPD exacerbation and whose treatment decision was guided by the Prospector method between 01/OCT/2015 and 04/JAN/2019. The analysis was performed in a Private Health Care Facility outpatient clinic in Suchanino, Gdansk, Poland. Informed consent was obtained from all patients to participate in the study. All methods were performed in accordance with the
relevant guidelines and regulations
For comparison, a retrospective analysis was performed on patients who reported to the same clinic at the same time period but whose decision was guided by the treating physician’s best knowledge. Approval for this study was received from the Bioethics Committee of District Medical Chamber in Gdansk (nr KB 14/2/15). The authors confirm that all research was performed in accordance with relevant guidelines/regulations.
Settings and participants
Inclusion criteria for both groups were: previous COPD diagnosis (ICD-10 code J44 and J44 /45, J44/J20), age > 40 years, COPD exacerbation, post 400 µg salbutamol FEV1/FVC<0.7 and/or the lower limit of normal (LLN) adjusted for age, and availability of medical records [2]. Patients in the Prospector group (PG) (n=40, 12 lost to follow-up) were treated by the first study author using the Prospector calculator. The control group (CG) (n=37) were treated by physicians other than the first author, according to their best knowledge, where Anthonisen's criteria were used [16]. Patients were tracked for exacerbations, hospitalizations, failure to recover, and death. The characteristics of both groups are presented in Table 1.
Table 1. Patient characteristics of the Prospector and control group.
|
Prospector group
n=28
|
Control Group
n=37
|
sex
|
|
|
female
|
10 (32%)
|
17 (46%)
|
male
|
18 (64%)
|
20 (54%)
|
Age (mean)
|
66,6 years
|
70,7 years
|
COPD severity
|
|
|
Mild +moderate (FEV1>50)
|
13 (46%)
|
31 (84%)
|
Severe +very severe (FEV1<50)
|
14 (54%)
|
6 (16%)
|
Concomitant disease ICD-10
|
I50 2[7%],I25 4[14%], E11 6[21%], I10 18[64%], I48 3[11%]
|
I50 3[8%],I25 5[14%]E11 10[27%], I10[24]64%, I48 8[22%]
|
Number of chronic diseases other than COPD
|
0-1
|
9[32%],
|
7[19%]
|
2
|
7[25%]
|
8[22%],
|
3+
|
12[43%]
|
22[59%]
|
The Prospector method
Prospector was created based on available data from the publications included in the article and the GOLD guidelines - regarding the diagnosis of COPD exacerbation and clinical experience in the management of patients with COPD exacerbation. In the publications, the authors validated the variables in terms of their usefulness in diagnosing bacterial exacerbation of COPD, but in none of these studies did they combine them. The simple design of the calculator was necessitated by clinical experience and taking into account the above-mentioned variables, which had already been validated in previous studies for their usefulness in diagnosing bacterial exacerbation of COPD. Simple scoring also forced a cut off point, arbitrarily set at 10 - antibiotic administration, < 10, i.e. 9 or less against antibiotic administration. Stat analysis showed that a more precise cut-off point could be 8, which is covered in the text.
Prospector is a novel clinical calculator developed by the study authors where information on patient symptoms, exacerbation, and medical history of COPD is translated into a score (Table 2). The score is compared against a cutoff point. A cutoff point was set arbitrarily as ≥10 for the recommendation of antibiotics treatment and ≤9 for recommendation against antibiotics treatment. Upon scoring ≥10, the patient was prescribed an antibiotic on visit 1. Upon scoring ≤9, the decision on antibiotic therapy was deferred to visit 2. In cases of suspected pneumonia, the investigator ordered a chest X-ray. Patients with ≤9 points were collected C-reactive protein (CRP) and complete blood count (CBC) tests to recognize markers of bacterial infection, which were set based on the literature on CRP ≥40, neutrophils ≥7500/µl with leukocytes ≥10000/µl and phenotyping based on eosinophils ≥200/µl as a candidate for OCS treatment [3, 18, 19]. A follow-up visit (visit 2) was recommended 2 to 3 days later to assess eosinophil count and verify intervention efficacy.
Table2. Scoring system visit 1.
Manifestation
|
Scoring
|
Reference
|
1. Cough greater than usual
|
1 point
|
[16]
|
2. Increased amount of sputum
|
1 point
|
[4]
[16]
|
3. Increased sputum purulence (change of sputum color to purulent-green, brown)
|
2 points
|
[20]
[21]
[16]
|
4. Increased dyspnea
|
1 point
|
[22]
[14]
[16]
|
5. Feeling of illness, symptoms of infection
|
1 point
|
[23]
|
6. Exacerbations in the last 12 months: 2 treated on an outpatient basis or 1 or more hospitalizations due to exacerbation of COPD
|
1 point
|
[18]
[4]
[24]
|
7. Airflow limitation severity (COPD grade)
|
very severe, 3 points; severe, 2 points; medium, 1 point; mild, 0 points
|
[3]
|
8. Concomitant diseases: cardiovascular, rheumatological, renal failure, chronic liver disease, cancer, metabolic and endocrine
|
1 point
|
[25]
|
9. Temperature >38⁰C
|
1 point
|
[23]
|
10. Saturation (SaO2) <95
|
1 point
|
[26]
|
11. Auscultatory changes other than ordinary daily wheezing and crackles
|
4 points
|
[23]
|
Visit 1
During the first visit, exacerbation history (past 12 months), medical examination, and history related to concomitant diseases were collected from the study group. Based on these data, patients were assessed using the Prospector method.
For enrollment into the Prospector group, patients needed to meet 2 criteria from points 1-5 (Table 2). To manage their exacerbation, patients were given bronchodilators for nebulization (0.5 mg fenoterol and 0.25 mg/ml ipratropium bromide or 0.25 mg/ml ipratropium bromide) and OCS (30 mg prednisone or equivalent if in the source documentation, the historical value of eosinophils was ≥200/µl). If eosinophil blood count (EBC) was unknown, patients were ordered CRP and CBC tests.
Verification (visit 2)
During the second visit, laboratory results were evaluated, the history of the last 2-3 days was collected, the patient was examined, and the Prospector method verified intervention efficacy. Patients scoring ≥10 points (Table 3) did not pass the verification and antibiotics were introduced. If the patient had antibiotics introduced at visit 1 and the verification was still negative, treatment was adjusted at the physician’s discretion. OCS were continued for 5-7 days if the new EBC was ≥200/µl; if not, OCS were discontinued.
Table 3. Scoring system on visit 2.
Manifestation
|
Scoring
|
Reference
|
1. Leukocytosis 10000/µl with neutrophils 7500/µl
|
4 points
|
[19]
|
2. CRP ≥40
|
4 points
|
[18]
|
3. Chest X-rays confirm pneumonia
|
4 points
|
[19]
|
4. Deterioration during treatment
|
5 points
|
[27]
|
5. Improvement during treatment
|
-5 points
|
[28]
|
6. Auscultatory changes
|
increased, 2 points; same as on visit 1, 1 point; no change, 0 points
|
[28]
|
In the control group, antibiotic therapies were implemented at the physician's discretion, often using Anthonisen's criteria. A second visit, if applicable, was set up at the physician's discretion.
Data collection
Patient demographics, exacerbation history in the previous 12 months, hospitalizations, symptoms of exacerbations, concomitant diseases, airflow limitation severity (COPD grade: 1,2,3,4), temperature, saturation, and physical examination findings were collected according to available electronic data retrieved from Serum system version 4.5.8.26 Kamsoft S.A. (Poland) and from the Prospector database. All laboratory tests were retrieved from a local authorized laboratory, Bruss (Poland).
Statistical analysis:
The statistical analyses have been performed using the statistical suite TIBCO Software Inc. (2017) and Statistica data analysis software system, version 13 (http://statistica.io). The quantitative variables were characterized by the arithmetic mean of standard deviation or median or max/min (range) and a 95% confidence interval. The qualitative variables were presented with the use of count and percentage. The statistical significance of differences between the two groups was processed with the student T-test or Mann-Whitney u-test. Chi-squared tests for independence were used for qualitative variables. In all the calculations, the statistical significance level of p=0.05 has been used. ROC (receiver operating characteristic curve) and AUC (area under curve) was implemented to verify accuracy of arbitrary set cut-off point. Value of P < 0.05was considered statistically significant.