Clinical characteristics and related factors of fungal empyema thoracis

Object Data on the characteristics and related factors of fungal empyema thoracis (FET) are limited. Our aim is to investigate the clinical characteristics and related factors of FET. Methods We conducted a retrospective study of patients with positive culture from pleural effusion who were admitted to the First Aliated Hospital of Soochow University between January 2007 and January 2018. The clinical and laboratory characteristics of all study participants were collected. Logistic regression models were used to assess the related factors of FET. Results There were 30 patients diagnosed with FET. The median age of patients with FET was 62.7 years old and 24 (80.0%) were male. The most frequent pathogens of FET were Candida albicans (55.9%). Diabetes mellitus (23.3% vs. 7.9%, P = 0.019) and upper gastrointestinal tract perforation or rupture (20.0% vs. 2.1%, P < 0.001) were associated with increased risk of FET. The adjusted odds ratios (95% condent interval) were 3.11 (1.02–9.56) for diabetes mellitus and 12.71 (3.47–46.55) for upper gastrointestinal tract perforation or rupture. There were 10 deaths (33.3%) among FET patients after one-year of follow-up.


Introduction
Empyema thoracis (ET) is de ned as the accumulation of pus within pleural cavity [1]. It had been a lifethreatening disease until thoracic drainage and antibiotic were recommended, although Hippocrates put forward the rst hypothesis and experimental treatment of empyema thoracis in 500 BC [2]. It is reported that more than 18,000 patients were diagnosed as ET in the USA each year and mortality rate was up to 15%-75% in the history [3][4][5].
Fungal empyema thoracis (FET) is a rare ET subtype caused by fungus, and Candida species are the major kinds of pathogens [6]. A retrospective study conducted in Australia indicated that FET was commonly seen in postoperative immunosuppressed patients with the incidence of only 7.0% in the ET patients [7]. However, a high mortality was observed in patients with FET, for example, one study in American showed a 6-week mortality of 34%, and another two Taiwan studies reported the mortality was up to 61.9%-73.0% [8][9][10]. Pneumonia, surgery and gastrointestinal tract rupture were reported to be the most common causes of FET [8]. Despite the high mortality rate of FET, there has been no large-scale study of FET describing the characteristics and related factors of FET in mainland China.
Therefore, we conducted this retrospective study to investigate the clinical characteristics, related factors and outcome of FET in a Chinese population with positive effusion culture, in order to provide references for clinicians to select a reasonable treatment strategy.

Patients
The diagnosis of empyema thoracis was suspected by the nding of a pleural effusion on chest imaging and/or ultrasound examination and con rmed by diagnostic pleural aspiration from which organisms were grown. All patients admitted at the First A liated Hospital of Soochow University from January 2007 to January 2018 with positive culture from pleural effusion were included except tuberculous empyema. This study was approved by the Ethics committee of the First A liated Hospital of Soochow University and written informed consent was waived as a retrospective study.

Data collection
The demographic data (sex, age), co-morbidity (hypertension, diabetes, chronic lung disease, cirrhosis, chronic renal disease, malignancy), pathogen and outcome were collected by means of hospital database. Infection and nutrition related laboratory examination, such as white blood cell (WBC), neutrophil count, lymphocyte count, hemoglobin and albumin were recorded.

De nitions
Charlson's Comorbidity Index (CCI) included 19 different medical conditions with variable scores based on severity: 1 point for chronic obstructive pulmonary disease (COPD), dementia, connective tissue disease (CTD), congestive heart insu ciency, cerebral or peripheral vascular disease, cardiac infarction, peptic ulcer and mild diabetes or hepatopathy; 2 points for hemiplegia, moderate/severe renal failure, end-stage diabetes, any neoplasm, leukemia or lymphoma; 3 points for moderate/severe hepatopathy and 6 points for acquired immune de ciency syndrome (AIDS), and metastatic parenchymal tumor [11]. Age-adjusted Charlson's Comorbidity Index (ACCI) was calculated by one point that added to the CCI for each decade from the beginning of 50 years old [11]. Possible factors that may lead to empyema thoracis were classi ed into 4 groups: pneumonia, surgery (including thoracic and upper abdominal operation), gastrointestinal tract injury (including esophageal leak, gastroduodenal perforation) and thoracic trauma. Hospital acquired empyema was regarded as patients who developed empyema after 48 hours of hospitalization except those complicated to previous infection on admission (such as communityacquired pneumonia) [12].

Statistical analysis
Continuous variables were expressed as median (Interquartile range, IQR) and were compared using the Manne Whitney U-test. Categorical variables were expressed with proportions (%) and were compared using the chi-square test or Fisher's exact test if necessary. Logistic regression models were applied for multivariate analysis to determine the in uence of the variables on FET. Odds ratio (OR) and corresponding 95% con dence interval (CI) were calculated. Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) for Windows (Version 22.0; SPSS, Inc., Chicago, IL, USA).

Patient characteristics
Of the 202 cases with culture-positive pleural effusion from January 2007 to January 2018, 8 cases were excluded (4 were mediastinal infections and the others were contaminated samples). The remaining 194 cases were then classi ed into FET group (n = 30) and non-FET group (n = 164) according their pathogen from pleural effusion ( Fig. 1).
The possible primary causes of FET thoracis seen in this series were as follows: pneumonic, 11 cases (36.7%); surgery, 13 cases (43.3%); Upper gastrointestinal tract perforations or rupture, 6 cases (20.0%). All surgeries were chest approach, including 6 esophagectomies, 3 gastrectomies and 4 pulmonary lobectomies. Upper gastrointestinal tract perforation or rupture included 3 spontaneous esophagus ruptures and 3 gastroduodenal perforations. There were ten deaths and one-year mortality was 33.3% (Table 1).   (Table 1). Patients whose ACCI were over All variants signi cantly associated with higher FET risk found in univariate analysis were included in multivariate model. The ndings revealed that diabetes mellitus (P = 0.047), and upper gastrointestinal tract perforation or rupture (P < 0.001) were associated increased risk of FET (  Discussion FET remains as a rare but fatal disease and leads to a puzzling condition for clinicians. In our study, the proportion of FET case was 15.5% (30/194), which is signi cantly higher than prior reports of 7.0% conducted in Australia [7]. The growth in the rate of FET is likely related to the increasing application of broad-spectrum antibiotics, transthoracic surgery and the number of critical or immunocompromised patients, as well as better fungal detection techniques recently [8]. In addition, the one-year mortality was up to 33.3% (10/30) in our study, the rate was similar to the study in American, which showed the 6-week all-cause mortality of patients with FET due to Candida or Aspergillus spp. was 34%. However, Ko  Similar to previous studies, Candida was the most common fungal pathogens in our cohort (30/34, 91.2% vs. 47/73, 63.5%) [8]. Candida albicans remains the major causative agent while non-Candida albicans species such as Candida glabrata, Candida tropicalis and Candida parapsilosis were not rare as FET pathogens (11/33, 33.3%). Candida tropicalis was one of most commonly-detected fungi in human gut mycobiome and was detected in 67.0% of samples based on 17 gut mycobiome studies [13]. Notably, 4 of 5(80.0%) Candida tropicalis empyema thoracis in our study had gastrointestinal tract surgery or injury and it is believable that Candida tropicalis is commonly associated with digestive diseases. Apart from Candida, one cryptococcosis empyema thoracis case in our study was comorbid with malignancy B cell lymphoma. Cryptococcosis infection has been reported in immunocompromised patients such as HIV infection, liver cirrhosis or Bruton's agammaglobulinemia and most cryptococcosis empyema thoracis has similarly been described in immunode ciency patients [14].
A number of factors had been proved to predispose patients to fungal infection. Impaired T-cell function followed by high-dose glucocorticoid therapy, chemotherapy, or AIDS, as well as depressed neutrophil number or function were considered to increase the risk of fungal infection [8]. But researches about factors of predispose patients to FET were limited. Masayuki N et al. studied 97 FET patients complicated with malignant tumors and indicated that previous surgical operation was a risk factor for candida infection and the presence of uncommon mold species suggested as the contamination of pleural effusion specimens [15]. FET had been reported as a complication of operation or esophageal pleural stula [16,17]. Takashi I et al. found that all of candida empyema (5/5) were secondary to esophageal or gastric stula and proposed a hypothesis that Candida spices in the pleural effusion can be an important clue for suspecting gastrointestinal tract perforation [18]. From 194 cases of culture-positive empyema thoracis, our study identi ed 10 cases due to upper gastrointestinal tract perforations, rupture or stulas including spontaneous esophagus ruptures (6 cases) and gastroduodenal perforations (4 cases), 6 of whom were diagnosed as FET and had con rmed statistical difference compared to non-FET. Those patients with gastroduodenal perforation had variable causes (portal hypertension with cirrhosis, gastric ulcer, iatrogenic perforation or duodenal trauma) and it was hard to assess the risk difference in gastroduodenal perforations with different causes duo to limited simples in this study. Candida species is a kind of normal commensals of humans throughout the entire gastrointestinal tract and can be the pathogen causing empyema when they break out the gastrointestinal tract barrier and enter the pleural cavity through esophagus rupture directly or gastroduodenal perforation [13]. Thus, the damage of gastrointestinal tract barrier had potential risk for development of fungal infections.
Diabetes mellitus has become one of the leading chronic disease burden worldwide and the overall morbidity was estimated to be 11.6% in the China [19]. Patients with diabetes mellitus were proven to have increased respiratory infectious risk due to inadequate clearance or the disturbance of normal pulmonary immune function [20]. It has been reported that diabetic patients conferred a 1.71-fold increased risk of empyema thoracis without any comorbidity [21]. A series of retrospective studies showed that the hazard of developing empyema was higher for patient with diabetes mellitus than those with chronic obstructive lung disease or chronic liver disease and cirrhosis [20,22]. Gosiewski T et al found the quantity of candida in the feces of patients with diabetes was signi cantly higher compared to non-diabetic controls [23]. However, the association between diabetes and FET has not yet been fully evaluated. Our study showed that patients with diabetes have an increased risk of FET, suggesting that diabetes mellitus could be an important risk factor for FET.

Conclusion
Our study rstly provided fundamental clinical characteristics and related factors about FET in mainland China. We demonstrated that upper gastrointestinal tract perforation or rupture and diabetes mellitus (I) Conception and design: EJZ, HWQ, CJ; (II) Collection and assembly of data: EJZ, CQH, XFQ, LRZ, JL; (III) Data analysis and interpretation: EJZ, HWQ, CKZ, YHZ; (IV) Manuscript writing: All authors; (V) All authors read and approved the nal manuscript.

Figure 1
Flow chart representing case lter