Etiology of Emergency Visit and In-Hospital Outcomes of Patients with COPD

Backgrounds Patients with COPD often visit the emergency department (ED) due to exacerbation of respiratory symptoms (dyspnea, cough, and sputum production). Because manifestations of acute exacerbation of COPD (AECOPD) are nonspecific, differential diagnosis is critical in this acute setting. The causes for emergency visiting and the in-hospital outcomes are varied in patients with COPD. This study aimed to investigate the distributions of etiologies and the in-hospital outcomes of patients with COPD who presented to the ED because of exacerbation of respiratory symptoms. Methods This was a retrospective study on COPD patients who had visited the ED and been hospitalized in a tertiary hospital because of worsening respiratory symptoms including cough, sputum production, and dyspnea from January 2017 to April 2020. Demographics, clinical manifestations, and laboratory studies in the ED were collected as the baseline data. The primary diagnosis at discharge or death was recorded. The hospitalization settings (general wards and ICU), the in-hospital outcomes, and associated factors were analyzed. Results During the study period, 392 patients with COPD (male 302 (77.0%)), with a median age of 78 years, visited the ED and hospitalized in this hospital. The first 3 causes for emergency visit were AECOPD (n = 314, 80.1%), acute coronary artery syndrome with or without congestive heart failure (n = 24, 6.1%), and pulmonary embolism (n = 13, 3.3%). For patients with AECOPD (n = 314), 51.6% (n = 162) was admitted to ICU, and 6.4% (n = 20) died. Multivariate logistic analysis showed that age, atrial fibrillation, NT-pro BNP ≥300 pg/ml, and blood pH <7.3 were independent risk factors for ICU admission. Age, comorbid malignancy, NT-pro BNP ≥1800 pg/ml, and pneumonia on CT scan were independent risk factors for hospital mortality in patients with AECOPD. Conclusion In COPD patients visiting the ED because of worsening respiratory symptoms, nearly 20% were due to non-AECOPD causes. For those with AECOPD, age, atrial fibrillation, NT-pro BNP ≥300 pg/ml, and blood pH <7.3 were independent risk factors for ICU admission, while advanced age, underlying malignancy, elevated NT-pro BNP, and pneumonia on CT scan were risk factors for hospital mortality.


Introduction
Chronic obstructive pulmonary disease (COPD) is a globally prevalent disease [1]. In China, the prevalence of spirometrydiagnosed COPD in adults ≥20 years old was 8.6% in a population-based study [2]. Acute exacerbation, defined as worsening of respiratory symptoms needing change of usual care, is an important event in the natural course of the disease [3]. AECOPD can be classified as mild, moderate, and severe, and severe exacerbation often needs emergency visit and/or hospitalization [4]. However, the diagnosis of AECOPD is based solely on clinical manifestations, which are nonspecific, and therefore differential diagnosis is critical for patients presenting to the emergency department (ED). e hallmark symptom of AECOPD is aggravated dyspnea, which is also a manifestation of other critical diseases, including pulmonary and extra-pulmonary diseases, such as heart failure (HF), pulmonary embolism (PE), cardiac infarction, and pneumothorax. A large observational study in Europe and the Asian-Pacific region for patients presenting to EDs with dyspnea as the main complaint showed that the most common causes were lower respiratory tract infection, HF, AECOPD, and asthma [5]. However, the causes for ED visiting in COPD patients with exacerbated respiratory symptoms (dyspnea, cough, and/or sputum) are rarely studied, though the data are important for differential diagnosis and management in this acute setting. erefore, we conducted a single-center retrospective cohort study on the causes and outcomes of COPD patients presenting to our ED and hospitalized for exacerbation of respiratory symptoms.

Study Subjects.
is was a single-center, retrospective observational study performed in Peking University ird Hospital. Patients with COPD presenting to the ED and hospitalized from January 2017 to April 2020 were consecutively enrolled. e inclusion criteria were (1) age ≥40 years; (2) diagnosis of COPD verified by medical records; (3) presenting with exacerbated respiratory symptoms including dyspnea, cough, and/or sputum production; and (4) hospitalized within 72 h after ED evaluation and management. Patients who died at the ED were excluded.

Data Collection.
All hospitalized patients with a diagnosis of COPD were retrieved and their medical records were reviewed, and if they had any ED visit during the study period, they were enrolled in the study. For patients with repeated visits, only the first visit was included for analysis. Baseline data and diagnostic studies at the ED and during hospitalization were collected, including demographics, comorbidities, laboratory tests, chest CT, and echocardiography. e symptoms (cough, sputum production, and dyspnea) were derived from the emergency medical records. Only laboratory results from the first ED visit were included in this study. Based on the emergency diagnosis and the discharge diagnosis, we ascertained the etiology for ED visiting. For patients with AECOPD, underlying diseases were recorded as comorbidities, which served as variates in the analysis. e outcomes were admission to the ward or ICU, length of stay, and survival.

Chest CT Review.
Chest CT scans performed within 72 h after presentation to the ED were collected and analyzed. A radiologist and a pulmonary physician reviewed the CT scans. Signs consistent with pneumonia were identified [6,7].

Statistical Analysis.
Statistical analysis was performed using Stata 16.0 software. Comparison between categorical variables was made using the chi-square test. e independent-samples t-test (for normal distribution parameters) and Mann-Whitney U-test (for abnormal distribution parameters) were adopted for comparisons of continuous data between two groups. Poisson regression and logistic regression were used to analyze the risk factors for ICU admission and in-hospital mortality for patients with AECOPD. P values <0.05 were considered as statistically significant.

Demographics and Baseline
Data. From January 2017 to April 2020, 392 eligible COPD patients, with a median age of 78 years, presented to our ED with the main complaint of worsening respiratory symptoms including cough, sputum production, and dyspnea. e demographics and the baseline data are presented in Table 1.

Distribution of Etiologies for ED Presentation.
e primary diagnosis of the causes for ED visiting of the COPD patients complaining of exacerbated respiratory symptoms are outlined in Figure 1 and Figure 2. Most of the patients (n � 314, 80.1%) presented to the ED and then hospitalized because of AECOPD. For patients with consistent manifestations and infiltrates on chest CT, we classified them as those with pneumonic AECOPD [6,7] and hence included in the AECOPD group. It is notable that in our cohort, non-AECOPD accounted for 19.9% of the primary causes for ED presentation, including acute coronary syndrome and/or HF, PE, and pneumothorax.

In-Hospital Outcomes of the Non-AECOPD Patients.
Of the 78 patients (19.9%) with non-AECOPD causes for ED visiting, 24 (24/78, 30.8%) were due to acute coronary syndrome and/or HF (8 with reduced ejection fraction), and 2 cases died in hospital.
irteen patients (13/78, 16.7%) were found to suffer from PE by computed tomography pulmonary angiogram (CTPA) and/or ventilation-perfusion scan. Ten patients (10/78, 12.8%) had pneumothorax, in which 1 was caused by lung puncture for diagnosis of lung lesions and 1 patient died. Eight patients were found to have lung cancer (8/78, 10.3%), among whom two had pleural effusion and 1 had obstructive pneumonia. Six patients (6/ 78, 7.7%) were confirmed to have fungal pulmonary infection, and 1 died. Other causes and the in-hospital outcomes are shown in Table 2. were receiving long-term oxygen therapy or NIPPV. As in the whole population, hypertension, arrhythmia, and cerebrovascular disease (CVD) were the 3 most common comorbidities (see Table 3).

In-Hospital Outcomes of the AECOPD Patients.
Of the AECOPD patients presenting to the ED, more than half (51.6%, 162/314) were admitted to ICU, and others (152, 48.4%) were admitted to the general ward. Among patients in the ward, 5 died; and among those in ICU, 15 died, resulting in a total in-hospital mortality of 6.4% (20/314). e median length of stay in ICU was 15 days (interquartile range, , and the length of stay in hospital of all the patients was 13 days (IQR 10-19).

Risk Factors for ICU Admission and In-Hospital Mortality in AECOPD Patients Presenting to the ED.
Comparison between patients admitted to ICU and the general ward is shown in Table 3. Patients admitted to the ICU were older and had higher levels of D-dimer, PCT, and NT-pro BNP, and more patients had ≥2 comorbidities, arrhythmia or HF. Because the latest heart failure guidelines recommend NT-pro BNP <300 pg/ml as a threshold to exclude heart failure in emergency situations [8], we used this cut-off value for analysis. e results showed that age (odds ratio (OR) 1.04, 95% confidence interval (CI) 1.005-1.069, P value 0.021), atrial fibrillation (OR 2.99, 95% CI 1.16-7.71, P value 0.023), NT-pro BNP ≥300 pg/ml (OR 2.58, 95% CI 1.38-4.80, P value 0.003), and pH <7.3 (OR 8.73, 95% CI 2.35-32.4, P value 0.001) were independent risk factors for ICU admission.

Discussion
AECOPD is defined as an acute worsening of respiratory symptoms resulting in additional therapy. As comorbidities that may worsen respiratory symptoms are common in COPD patients, clinical assessments to rule out alternative Emergency Medicine International 3 diagnoses should be performed before confirmation of a COPD exacerbation [10]. In clinical practice, because patients with COPD often visit the ED due to exacerbation of respiratory symptoms, differential diagnosis is critical in this acute setting. However, detailed, practical studies of the causes for ED visiting in the COPD population are scarce.
Here, we described the distributions of etiologies and the inhospital outcomes of patients with COPD who presented to the ED because of exacerbation of respiratory symptoms. We found that the top five causes for ED visiting in our cohort were AECOPD, acute coronary syndrome and/or heart failure, PE, pneumothorax, and lung cancer, confirmed after careful examinations in the ED and after hospitalization. We also found that advanced age, malignancy, atrial fibrillation, elevated NT-pro BNP, pH <7.3, and pneumonia on CT scans increased the risk of ICU admission and/or in-hospital mortality in patients with AECOPD. A large population-based study in Canada showed that, for the 27,705 admissions with a diagnosis of COPD at the EDs, the primary diagnoses at discharge were mostly related to COPD per se (n � 20, 848, 75.2%), followed by pulmonary infections (n � 2916, 10.5%). Cardiac conditions including heart failure, ischemic heart disease, or chest pain represented 8.3% of the admissions (n � 2311, 8.3%). Other serious conditions complicating COPD (for example, pneumothorax, PE, or stroke) were reported infrequently (n � 129, 0.5%) [11]. However, this study was based on provincial administrative databases, which lacked detailed clinical information at the EDs. In a recent study on COPD patients admitted to the hospital due to acute worsening of respiratory symptoms, PE was detected in 5.9% of patients using a predefined diagnostic algorithm [12]. e pooled prevalence of PE in "unexplained" AECOPD was 16.1% (95% CI, 8.3%-25.8%) in a total of 880 patients [13].
Current guidelines recommend that patients suspected with AECOPD should be differentiated from diagnosis of pneumonia, pneumothorax, PE, cardiac events, and arrhythmia [10]. As COPD patients with exacerbated respiratory symptoms often present to the EDs, our results    Smoking and aging are common risk factors for COPD and other chronic diseases, which are associated with multiple comorbidities in COPD [14]. e prevalence of HF in COPD patients ranges from 7.1% to 31.3%, and the prevalence of coronary heart diseases (including myocardial infarction, angina, and ischemic heart disease) ranges from 4.7% to 60% [15]. Cardiac complications were highly associated with an increased risk of death [16] in COPD patients. In this study, our findings were consistent with others in terms of the risk factors for ICU admission and inhospital mortality in patients with AECOPD, including older age [17], cardiovascular complications [18] (arrhythmia, HF), malignancy [16], pneumonia [19][20][21], anemia [22], and high levels of NT-pro BNP [23].
BNP has moderate accuracy in detecting HF in the EDs [24], but interestingly, NT-pro BNP was a strong and independent predictor of in-hospital mortality in AECOPD patients [25]. Buchan et al. found that elevated BNP or NTpro BNP levels in AECOPD were associated with increased mortality from cardiovascular diseases [26]. An earlier study showed that AECOPD was definitely associated with acute left-heart dysfunction in 31.1% and possibly with left ventricular dysfunction in 13.5% of the patients [27]. Furthermore, COPD is associated with a high incidence of pulmonary hypertension, which is linked with exercise limitation and a worse prognosis [27]. In AECOPD patients without underlying left ventricular dysfunction, log-transformed NT-pro BNP levels were positively associated with echocardiographically estimated right ventricular systolic pressure [28]. It should be noted that the NT-pro BNP cutoff values for assessing ICU admission and in-hospital mortality were different in our study because the diagnostic threshold of NT-pro BNP was varied. A level <300 pg/ml is often used to rule out HF [8]. When it comes to assessing inhospital mortality, the diagnosis of HF is more important. erefore, we used NT-pro BNP ≥1800 pg/ml for analysis of in-hospital mortality, as the median age of our cohort was older than 75 years. In addition, in this study, we used two regression models for analysis of the risk factors because the number of patients who died in hospital was relatively small. Poisson regression and logistic regression were used at the same time to make the regression result more robust, and the two regression results were similar.
Our study had several limitations. As a retrospective study, some examination results and lab data were not available for all our patients. In this hospital, patients requiring noninvasive or invasive mechanical ventilation were all admitted to ICU, and therefore the ICU admission rate was high. As a single-center study, the sample size was relatively small, and the results may not be applicable to other ED practice.

Conclusion
Our study found that, for COPD patients who presented to the EDs because of exacerbated respiratory symptoms, non-AECOPD causes accounted for about 20%, most commonly acute coronary syndrome and/or heart failure, and PE. For patients with AECOPD, age, atrial fibrillation, NT-pro BNP ≥300 pg/ml, and blood pH <7.3 were independent risk factors for ICU admission, while advanced age, underlying malignancy, NT-pro BNP ≥1800 pg/ml, and pneumonia on CT scan increased the risk of in-hospital mortality.

OR:
Odds ratios CI: Confidence intervals SD: Standard deviation