Diagnostic and prognostic implications of bacteremia in patients with complicated pleural infection

Study population

We conducted a retrospective study on patients who had received pleural drainage due to complicated parapneumonic effusion or empyema at Kyungpook National University Hospital, a tertiary referral hospital in South Korea, which included consecutive cases between January 2011 and May 2021. Patients who were enrolled had either a complicated parapneumonic effusion, which was defined as a pleural fluid pH of less than 7.2, glucose less than 40 mg/dL, lactate dehydrogenase (LDH) greater than 1,000 U/L, or positive gram stain or culture or an empyema, defined by the presence of frank pus, based on the Light’s classification [22]. Those who did not undergo pleural drainage were excluded. Only the first occurrence of pleural infection was considered for patients with recurrent episodes.

Data collection and definitions

Our institution’s standard protocol for patients suspected with pleural infection is to perform pleural fluid smear and cultures including pairs of aerobic and anaerobic blood culture bottles to detect aerobic and anaerobic microbial pathogens, mycobacteria, and fungi as well as two sets of blood cultures [11]. Pleural infection due to a nonbacterial pathogen such as fungus or mycobacterium was excluded in this study. Bacteremia was defined when there was the identification of the same organism in the blood culture and other cultures obtained from the pleural fluid/sputum or the isolation of the organism from two or more blood cultures drawn at different times. Except for methicillin-resistant Staphylococcus aureus, multidrug resistance (MDR) was defined as the lack of susceptibility to at least one agent in three or more categories of antimicrobial agents in drug sensitivity testing [23]. The standard initial treatment regimen involved the administration of empirical antibiotics and percutaneous catheter drainage, with or without antifibrinolytic therapy (urokinase or tissue plasminogen activator) given at the physician’s discretion. Details on data regarding demographics, comorbidities, infection acquisition site (community vs. hospital), blood and pleural fluid testing results, causative pathogens, and clinical outcome were collected from electronic medical record. RAPID score (Renal [urea], age, fluid purulence, infection source, and dietary [albumin]) was also calculated [24]. The data on in-hospital and post-discharge mortality, as well as the last-follow-up date, were collected.

A retrospective evaluation of radiological findings was conducted by a thoracic radiologist and a pulmonary specialist, regarding the amount of pleural effusion and loculation status of the effusion as seen on chest radiographs, and the presence of cavitary lesions observed on chest computed tomography scans. The clinical, laboratory, and radiologic data and outcome were compared between the two groups with and without bacteremia. The study protocols were approved by the Institutional Review Board of Kyungpook National University Hospital. Informed consent was waved due to the retrospective nature of the study.

Statistical analysis

The statistical analyses were conducted utilizing IBM SPSS Statistics Version 22.0 for Windows (IBM Corp., Armonk, NY, USA). For the comparison of continuous variables, which were expressed as median and interquartile range (IQR), either the t-test or the Mann-Whitney U test was used, whereas the chi-square or Fisher’s exact test was utilized for comparing categorical variables, which were presented as number and percentage. Univariable and multivariable logistic regression analyses were performed to identify the factors associated with bacteremia in patients with complicated pleural infection, and the resulting odds ratio (OR) and 95 % confidence intervals (CI) were reported. Additionally, the Cox-proportional hazard regression analysis was employed to assess the risk factors associated with overall survival. Variables with p-values<0.05 were considered statistically significant.

Comparisons of clinical, laboratory, and radiologic findings between complicated pleural infection patients with and without bacteremia

In this study, 341 patients with complicated pleural infections who received pleural drainage were included. Among them, 25 (7 %) had a positive blood culture. The median age for patients with bacteremia and those without bacteremia was 68 and 65 years, respectively (Table 1). In both groups, approximately 80 % of the patients were male. In total, 23 % had comorbid diabetes mellitus. Malignancy was observed to be more prevalent in patients with bacteremia than in those without (16 vs. 6 %, p=0.087). There was no significant difference in the occurrence of pleural infection due to extrapulmonary infection between the two groups. On the other hand, there was a tendency for hospital-acquired infections to occur more frequently in patients with bacteremia (24 vs. 13 %, p=0.133).

Patients with bacteremia had significantly lower levels of serum albumin and higher levels of blood urea nitrogen (BUN) and serum LDH on presentation compared to those without bacteremia (albumin: 2.7 [2.4–3.2] vs. 3.1 [2.6–3.5] g/dL, p=0.005; BUN: 23.5 [15.5–44.7] vs. 15.4 [11.2–22.7] mg/dL, p<0.001; LDH: 273 [191–406] vs. 193 [162–237] U/L, p=0.006). The RAPID score was significantly higher in patients with bacteremia than in those without bacteremia (5 [4–5] vs. 3 [2–4], p<0.001).

The analysis of pleural fluid revealed that patients with bacteremia had elevated levels of LDH and reduced levels of protein. In addition, patients with bacteremia had a significantly higher rate of positive pleural fluid culture results (68 vs. 32 %, p<0.001), with a two-fold increase in occurrence compared to those without bacteremia. Furthermore, patients with bacteremia exhibited a higher occurrence of cavitary lesions and loculated pleural effusion on radiological evaluation, compared to those without bacteremia (20 vs. 6 % for cavitary lesions, p=0.027; and 40 vs. 23 % for loculated effusion, p=0.047).

Multivariable analysis showed that cavitary lesion (OR=3.88 [95 % CI=1.20–12.52], p=0.023), RAPID score≥5 (OR=2.70 [95 % CI=1.12–6.51], p=0.027), and pleural fluid culture positivity (OR=4.40 [95 % CI=1.75–11.03], p=0.002) were the independent factors associated with bacteremia in patients with complicated pleural infection.

Identification of microbial etiologies from the blood and pleural fluid cultures in patients with complicated pleural infection

Among 341 patients with pleural infection, 127 (37 %) had the causative pathogens identified through blood (n=25, 7 %) or pleural fluid culture (n=119, 35 %) (Table 2). Eight among 25 patients with bacteremia showed no growth in pleural fluid culture, accounting for 2 % of the total study population and 6 % of the population with identified pathogens. The causative pathogens from the blood culture included Staphylococcus species (36 %; S. aureus [n=8] and Staphylococcus hemolyticus [n=1]), Streptococcus species (32 %; Streptococcus pneumoniae [n=3], Streptococcus pyogenes [n=2], Streptococcus sanguinis [n=1], Streptococcus parasanguinis [n=1], and Streptococcus salivarius [n=1]), Klebsiella pneumoniae (20 %, n=5), and others (12 %; Escherichia coli [n=2], and Enterococcus fascium [n=1]). Among 17 patients with complicated pleural infection with both blood and pleural fluid culture-positive results, 14 yielded consistent organisms across blood and pleural fluid cultures, while three had different organisms in blood and pleural fluid cultures (Supplementary Table 1). Polymicrobial infections and MDR-pathogens were identified in nine (7 %) and 40 (32 %) out of 127 patients with a positive blood or pleural fluid culture, respectively. The proportion of MDR pathogens was higher in patients with bacteremia than in those without bacteremia (52 vs. 27 %, p=0.014). Drug susceptibility testing results for 14 isolates from both the blood and pleural fluid simultaneously were consistent between blood and pleural fluid.

On the other hand, among the 102 patients who tested positive for pleural fluid culture but did not have bacteremia, the primary causative pathogens were identified as viridans streptococci species (n=54, 53 %), K. pneumoniae (n=24, 24 %), and Staphylococcus species (n=10, 10 %).

Comparisons of clinical outcomes between complicated pleural infection patients with and without bacteremia

Approximately 50 % of patients received intrapleural fibrinolytic therapy, which was similar between patients with and without bacteremia (Table 3). The length of pleural catheter drainage was significantly longer in patients with bacteremia, with a median of 14 days (7–23) compared to 8 days (5–13) in those without bacteremia (p=0.003). However, there were no significant differences in the number of patients who underwent decortication operation or experienced in-hospital mortality between the two groups (decortication operation: 8 vs. 3 %, p=0.217; in-hospital mortality: 3 vs. 4 %, p=1.0). The in-hospital mortality rates for the groups with positive blood cultures alone, positive pleural fluid cultures alone, positive cultures in both blood and pleural fluid, and negative cultures in both were 0 , 6.9, 7.7, and 2.3 %, respectively.

In patients with complicated pleural infection, several factors were found to be associated with overall mortality according to univariable Cox proportional hazard regression analysis. These factors included old age, advanced liver disease, hospital acquired infection, serum albumin levels, pleural fluid culture positivity, and RAPID score. However, in multivariable analysis, only the RAPID score was significantly associated with overall mortality (hazard ratio [HR]=1.96 [95 % CI=1.35–2.84], p<0.001) (Table 4). Pleural fluid culture positivity showed an increased tendency for overall mortality (HR=2.25 [95 % CI=0.86–5.88], p=0.098).