Background
An epidemic of Coronavirus Disease 2019(COVID-19) struck Wuhan, China and rapidly spread to the entire country and around the globe [
1‐
3]. Till August 31, 2020, A cumulative total of nearly 25 million cases and 800,000 deaths globally were reported since the start of the outbreak according to the World Health Organization [
4] and the National Health Commission of China reported a total of 85,058 confirmed COVID-19 cases, 4634 deaths, and 80,208 cured cases in China [
5,
6]. COVID-19 was caused by the severe adult respiratory syndrome coronavirus 2 (SARS-CoV-2). The diagnosis of COVID-19 was established on the basis of contact history, clinical features, imaging findings and results of RT-PCR tests [
7]. Given the wide clinical spectrum of COVID-19, understanding the factors that can predict disease severity were essential since this would help frontline clinical staff to stratify patients with increased confidence [
8]. Pleural effusion (PE), lung cavitation, lymphadenopathy and calcification were rarely seen in COVID-19 patients [
9‐
11]. Previous studies demonstrated that PE exerted a significant influence on the final outcome of patients suffering from acute lung injury or acute respiratory distress syndrome [
12]. Recent study found that severe/critical patients showed more lymph node enlargement, pericardial effusion, and pleural effusion, which suggesting these extrapulmonary lesions may indicate the occurrence of severe inflammation, However, the sample size of that research was relatively small [
13]. Additionally, The possibility of PE prediction for progression to critical condition of COVID-19 patients was not yet analyzed.
In the present study, we preliminarily characterized the imaging findings of 476 COVID-19 patients. Then, we compared COVID-19 patients with and without PE in terms of their clinical futures and outcomes. Finally, a predictive model based on PE and other clinical features was created to identify COVID-19 patients who may progress to critical condition.
Discussion
COVID-19 was diagnosed based on the patients’ contact history, clinical presentations, imaging findings and laboratory results [
7,
14,
16]. Chest CT plays an important role in the initial diagnosis of COVID-19. Typical chest CT findings in patients with COVID-19 principally included multiple bilateral patchy ground-glass opacities in lobules with peripheral distribution [
7]. Pleural effusion is very commom with the pathological accumulation of fluid in the pleural space. There are many causes of pleural effusion, including viral pleuritis, congestive heart failure or cancer [
17]. Patients with a non-malignant pleural effusion have a one-year mortality in the range of 25 to 57% [
18]. A recent study found that pleural effusion occurred in 10.3% COVID-19 patients and those refractory patients had a higher incidence of pleural effusion than general COVID-19 patients, suggesting a more obviously inflammatory response in the lung [
19]. However, no clinical studies with larger sample size have especially focused on COVID-19 complicated with PE and the implication of PE is underestimated in clinical practice. To our knowledge, this was the first clinical study to examine the imaging features and clinical characteristics of COVID-19 with PE. Our study demonstrated that PE was an uncommon imaging sign and its presence signified unfavorable clinical outcomes.
Typical CT findings of COVID-19 included peripherally distributed multifocal ground-glass opacities plus patchy consolidations, with a potential to involve posterior parts or lower lobes [
7]. In the present study, 261 subjects with COVID-19 had PE, including outpatients and inpatients. The incidence rate was 7.09% (261/3679), which was different from that of Severe Acute Respiratory Syndrome. Previous study revealed that Severe Acute Respiratory Syndrome patients radiologically presented more frequently with “ground-glass” changes without PE [
20]. Meanwhile, CT findings in 153 inpatients with PE showed that PE was bilateral in most patients (65.36%). COVID-19 patients with PE presented with different types of pulmonary lesions: ground glass opacities, pulmonary consolidation, and liner opacities. Some patients were complicated with pleural thickening (20.26%), pericardial effusion (7.84%) and pulmonary emphysema (5.23%). All aforementioned imaging findings suggested that COVID-19 patients with PE had more involved pulmonary changes.
Common symptoms of COVID-19 included fever, cough, myalgia and fatigue [
16]. Compared to patients without PE, COVID-19 patients with PE exhibited more specific symptoms, such as high fever, worse cough and breath shortness. Previous researches demonstrated that evidently decreased lymphocytes, increased platelets, CRP, LDH and D-dimer in COVID-19 patients might indicate that inflammation was severe and disease might deteriorate [
2,
21‐
23]. Our study showed that changes of these indicators were more conspicuous in PE group than in none-PE group. Meanwhile, the partial pressure of oxygen and oxygen saturation were significantly lower in PE group than in none-PE group. PE might substantially inhibit the respiratory function and lower the partial pressure of oxygen and oxygen saturation, eventually exacerbating acute respiratory distress syndrome in patients with severe or critical COVID-19.
Severe COVID-19 patients tended to rapidly progress to acute respiratory failure, acute respiratory distress syndrome, metabolic acidosis, coagulopathy, and septic shock. Early identification of risk factors for severe COVID-19 could lead to prompt supportive care and early admission to the intensive care unit [
24]. Hasley et al [
25] reported that the presence of bilateral PE was an independent predictor for short-term mortality in patients with community-acquired pneumonia. A study examined patients with MERS-CoV and found that the presence of PE and higher chest radiographic scores were indicative of poor prognosis and higher short-term mortality [
26]. In our series, no COVID-19 patients with PE were of general type. Moreover, the median length of hospital stay was longer in PE group than in none-PE group (
P < 0.0001). The mortality rate was significantly higher in patients with PE than in their counterparts without PE (5.54% & 0.93%,
P = 0.001). COVID-19 patients with PE might had poor prognosis, suggesting a more obviously inflammatory response in the lung. Early diagnosis and timely and proper treatment for those patients might have satisfactory effect.
Meanwhile, the independent risk factors for critical illness were also screened. The results of the LASSO regression analysis showed that LDH, D-dimer, PE and TBIL were significant risk factors associated with critical COVID-19. Then, we used a model containing LDH, D-dimer, PE and TBIL, in the form of nomogram to predict the progression to critical condition in COVID-19 patients. The risk predicted by the nomogram was virtually consistent with the actual outcomes, indicating combination of the four indicators promised to work better in the prediction of progression to critical condition. In line with our findings, the results by Mo PZ et al. [
17] showed that 85 refractory patients had higher levels of maximum temperature among fever cases, higher incidence of breath shortness and anorexia, severer disease assessment on admission, high levels of neutrophils, aspartate aminotransferase (AST), LDH and CRP, lower levels of platelets and albumin, and higher incidences of bilateral pneumonia and PE (45.2%). Also, the Chinese COVID-19 diagnosis and treatment plan (trial version 8) recommended dropping lymphocytes, rising inflammatory factors (e.g., IL-6, CRP), increasing LDH and rapidly progressive pulmonary changes are the predictive factors for severe and critical COVID-19. Now, we used predictive model based on PE and other clinical features to identify COVID-19 patients with critical condition and analyzed PE were significantly related to incidence of critical COVID-19. Therefore, we suggested strongly that PE might be included as a complementary risk factor for the identification of severe and critical COVID-19.
This study had some limitations. First, the study was a single-setting study without external validation cohort. Second, due to a small amount of effusion, thoracentesis could not be performed in COVID-19 patients with PE. Unfortunately, the laboratory findings of pleural effusion were not available. Biochemical analyses of PE caused by COVID-19 needed further study. Third, some patients might have self-medication before hospital admission, which could affect the results of CT images. However, this study focused on the clinical features of COVID-19 complicated with pleural effusion and we hope that our results could help clinicians better evaluate and manage COVID-19 patients with PE.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.