Intrapleural use of urokinase and DNase in pleural infections managed with repeated thoracentesis: A comparative cohort study

Introduction Evacuation of infected fluid in pleural infections is essential. To date, the use of an intrapleural fibrinolytic agent such as urokinase and DNase has not yet been assessed in infections managed by repeated therapeutic thoracentesis (RTT). Methods We performed a retrospective comparative study of two successive cohorts of consecutive patients with pleural infections from 2001 to 2018. Between 2001 and 2010, patients had RTT with intrapleural urokinase (RTT-U). After 2011, patients received intrapleural urokinase and DNase with RTT (RTT-UD). Data were collected through a standardized questionnaire. Results One hundred and thirty-three patients were included: 93 were men and the mean age was 59 years (standard deviation 17.2). Eighty-one patients were treated with a combination of intrapleural urokinase and DNase, and 52 were treated with intrapleural urokinase only. In the RTT-UD, RTT failure occurred in 14 patients (17%) compared to 10 (19%) in the RTT-U group (P = 0.82). There was no difference between the two groups in intensive care unit admission, surgical referrals or in-hospital mortality. RTT-UD was associated with faster time to apyrexia (aOR = 0.51, 95%CI [0.37–0.72]), a reduced length of hospital stay (aOR = 0.61, 95%CI [0.52–0.73]) and a higher volume of total pleural fluid retrieved (aOR = 1.38, 95%CI [1.02–1.88]). Complications were rare with only one hemothorax in the RTT-UD group and no pneumothorax requiring drainage in either group. Conclusion Compared to urokinase only, intrapleural use of urokinase and DNase in RTT was associated with quicker defervescence, shorter hospital stay and increased volumes of pleural fluid drained. Randomized controlled trials evaluating urokinase and DNase with RTT technique would be required to confirm these results.


Introduction
Pleural infections are severe infectious diseases, with a mortality rate ranging from 10 to 20% [1,2]. Incidence of pleural infections has increased in the last few decades and reached 15 cases per 100 000 per year [3][4][5]. According to guidelines, the two pillars of treatment are effective antibiotics and drainage of the infected liquid [6][7][8]. Hippocrates already underlined that 'if an empyema does not rupture, death will occur'. To date, chest tube insertion is the most widespread technique of drainage and is considered as the gold standard treatment in pleural infections. Nevertheless, another strategy is possible, namely repeated therapeutic thoracentesis (RTT). The theoretical advantages of RTT are: i) targeting spaces where loculated collections are the most voluminous; ii) allowing early ambulation; iii) allowing early pleural physiotherapy; iv) avoiding clogging of chest tube. Several studies reported RTT as an effective alternative to chest tube. We previously described a cohort of pleural infections managed by RTT as first-line treatment, with a clinical success rate beyond 80% and low rate of complications related to RTT [9]. In a comparative study, thoracenteses were not inferior to chest tube, with reduced duration of hospitalization and complications rates [10].
Furthermore, releasing fibrinolytic agents into pleural cavity has shown benefits in improving drainage effectiveness in a large randomized controlled trial [11]. In this study, Rahman et al. demonstrated that intrapleural use of tissue plasminogen activator (tPA) and deoxyribonuclease (DNase) reduces surgical referral at 3 months and length of hospital stay. In a previous trial, administration of a single fibrinolytic agent-tPA-did not improve any outcome [1]. Intrapleural fibrinolytic therapy has only been assessed once agents have been released through the chest tube and retrieved after one hour. In RTT, fibrinolytic agents and DNase released into the pleural space have a prolonged exposure. Though urokinase or tPa have very short half-lives (30-45 minutes), DNase has prolonged half-life (3-4 hours) increasing its activity on loculations [12,13]. We aimed to compare outcomes in patients with pleural infections managed by RTT with urokinase and DNase (RTT-UD), versus RTT with urokinase (RTT-U).

Study population
Rennes University Hospital is a 1500-bed tertiary-care hospital, which serves as a referral center in Western France (population catchment area, 1,000,000 inhabitants). We performed a retrospective study of all consecutive patients managed with RTT as first-line treatment for pleural infections in the Rennes University Hospital, France, during years 2001-2018. Non-complicated parapneumonic effusions were excluded from the study. To define pleural infections, we included all adults patients (age � 18 years) with at least one clinical symptom of current lung infection and at least one of the following characteristics for pleural fluid: frank pus (empyema), microorganisms observed after Gram staining, pH <7.2, glucose level <2.2 mmol/L, intrapleural loculations or abundant parapneumonic effusion. We divided our global cohort of patients with pleural infections treated by RTT into two consecutive groups: i) the first from 2001 to 2010, when patients were managed by RTT with adjunction of urokinase (RTT-U) ii) the second from 2011 to 2018, when most patients were managed by RTT with intrapleural use of urokinase and DNase (RTT-UD), following changes in our local protocol.

Data collection
Medical files of all patients treated by RTT as first-line regimen were reviewed. Data were collected with a standardized questionnaire including demographic, clinical, biological and imaging data. Outcomes were in-hospital mortality, one-year mortality, RTT failure (defined by death related to pleural infection, surgical referral or the need to switch to chest tube drainage), ICU admission, time to apyrexia, length of hospital stay and volume of pleural fluid drained. Apyrexia was defined as an axillary temperature < 38˚C during at least 72 hours.

RTT protocol
The RTT protocol was standardized and was performed at the patient's bedside under local anesthesia (lidocain 1%) using 8-French laparoscopic trocar ('Boutin Trocart' -Novatech1, La Ciotat, France). Intrapleural fibrinolysis protocol was standardized, based on urokinase (Eumedica1, Biarritz, France) 100 000 UI and DNase (Pulmozyme, Roche1) 5000 UI diluted in 50 mL of saline, then instilled by gravity in the pleural space using the pleural trocar at the end of thoracentesis. Patients at risk of hemorrhage or with a bronchopleural fistula were excluded, as these are contraindication for fibrinolysis. Each thoracentesis attempted to evacuate a maximum volume of infected pleural effusion. Thoracentesis were repeated as many times as required, at the discretion of the physician, until sepsis was under control and pleural effusion had significantly decreased. Complications, such as blank thoracentesis, vasovagal reaction or pleural hemorrhage, were collected.

Statistical analysis
We first described the characteristics of all patients enrolled in the two cohorts. Results were expressed as median and interquartile range (IQR) or mean and standard deviation (SD) for quantitative variables, and relative frequencies for qualitative variables. Missing data for each variable were excluded from the denominator. The Student t-test or Mann-Whitney test, and Fisher's exact test were used, as appropriate. We then performed logistic and linear regression analyses to evaluate the 'use of urokinase and DNase'. A logarithmic transformation was applied for a non-normally distributed response variable, namely 'total volume of pleural fluid retrieved', 'time to apyrexia after initial pleural drainage' or 'hospital length of stay'. A propensity score for treatment allocation was calculated using an adjusted logistic regression model with potential confounders: age, sex, underlying comorbidities (such as COPD, chronic heart failure, diabetes mellitus, malignancy, immunosuppression), severity at admission (based on respiratory failure and sepsis criteria), number of thoracentesis performed, abundance of pleural effusion at admission, positive pleural fluid direct examination, exact RAPID-score (which is a validated prognosis score in pleural infections) and ultrasonography-guided thoracentesis. Thereafter, the overlap weighting method based on propensity score was applied to a logistic regression for the following outcomes: in-hospital mortality, one-year mortality, RTT failure, ICU admission; and linear regression for the following outcomes: length of hospital stay, time to apyrexia, total volume of pleural fluid drained. Covariate balance after overlap weighting was evaluated by checking standardized differences (S1 Fig). Two-tailed p-values were reported, with p<0.05 considered as statistically significant. Statistical analysis was carried out using R-Studio 2015, Integrated Development for R (R-Studio, Boston, MA, USA).

Ethics
The study was approved by the Ethics Committee of Rennes University Hospital (approval number 12.52). Written informed consent was obtained from all study surviving patients.

Results
Of the 443 patients with suspected pleural infection, 133 fulfilled criteria for complicated parapneumonic effusion (CPPE) or pleural empyema (PE) (Fig 1). Data on patients from the first group were previously published [9].
The number of thoracentesis per patient in the two groups were equivalent (4 [3][4][5][6] vs. 4 [3][4][5]; P = 0.49) ( Table 2). Ultrasonography-guided procedure was more frequent in the recent cohort (54% vs. 84%; P = 0.007). Complications of RTT such as iatrogenic pneumothorax (n = 9/133, 6.8%) and vasovagal reaction (4 out of a total of 632 thoracenteses) were low in both groups but almost one third of patients (n = 46/133, 34.6%) had at least one blank thoracentesis during the management of the pleural infection. Antibiotics used did not differ between the two groups, with a large proportion of patients treated by amoxicillin/clavulanic acid combination (58% in RTT-U group vs. 60% in RTT-UD group). Amoxicillin alone was prescribed in cases of pleural infection due to S. pneumoniae or oral streptococci (15% in RTT-U group vs. 26% in RTT-UD group). Patients in the RTT-U group received longer antibiotic courses compared to RTT-UD (Table 2).
Time to apyrexia was significantly shorter in patients treated with the combination of urokinase and DNase (median fever duration: 10 vs. 7 days, P = 0.003). Fever duration after the first thoracentesis was also reduced in patients treated by RTT-UD (P<0.001). In-hospital mortality did not differ between the two groups (P = 0.25). Intensive care unit (ICU) admission rates were similar (15% vs. 15, P = 1).
Estimates of weighted odds-ratios (wORs) obtained with the overlap weighting method based on the propensity score are reported in Table 3

Discussion
Intrapleural fibrinolytic and DNase agents in pleural infections have been assessed only on patients drained with chest tube. To our knowledge, this study is the first to evaluate intrapleural fibrinolytic agent-urokinase-and DNase in pleural infections managed by RTT. Our work underlines many important messages: i) Time to apyrexia, length of hospital stay and the total volume of pleural fluid retrieved were optimized with a combination of intrapleural urokinase and DNase, ii) Mortality rates, RTT failures and surgical referral were low and did not differ between the two groups, iii) RTT is safe, with a low rate of iatrogenic complications, iv) in patients treated with RTT-UD who had non-purulent pleural fluid, no failure was reported.
In our center, we opted for urokinase instead of streptokinase or alteplase because it seems that urokinase is associated with a lower incidence of bleeding or allergic reaction [12,[14][15][16]. Moreover, two randomized placebo-controlled trials found that urokinase decreased duration of hospitalization, with faster defervescence and improved pleural drainage [17][18][19]. In addition, a meta-analysis supported that urokinase use could reduce surgical intervention (OR = 0.33; 95%CI: 0.14-0.78) [20]. Adding DNase is based on clear-cut benefits demonstrated in the MIST-2 trial [21,22]. It has also been proven that this treatment strategy is highly cost-effective [23]. Moreover, a combination including DNase was more effective in experimental studies, with a reduction of pleural fluid viscosity and an increased volume of pleural fluid drained [24,25]. These are the reasons why we used this combination with RTT. In line with previous studies, we found that the total volume of pleural fluid drained was higher with intrapleural urokinase and DNase, even though pleural effusion seemed to be less abundant and more loculated in the RTT-UD group. Of note, loculations were more commonly detected in the more recent group, probably explained by development of pleural ultrasonography in our department. On the other hand, we have no explanation for the fact that abundance of pleural effusion on chest imaging was lower in the RTT-UD group. Our work supports an enhanced effective drainage with the use of urokinase and DNase in pleural infections managed by RTT. In addition, time to apyrexia has been assessed as an outcome parameter in infectious diseases, where inappropriate therapies were associated with longer duration of fever [26,27]. Moreover, several studies identified prolonged fever as a factor of poor outcome [28][29][30]. In this study, defervescence was lower in the RTT-UD group, arguing for its determinant role in controlling pleural infections. In a landmark randomized controlled trial, Rahman et al. showed a reduction in hospital stay and a decrease in surgical referral with the use of t-PA and DNase in patients treated with chest tube. Herein, we did not find differences in mortality or surgical referral, but our study may have been underpowered due to limited sample size. The absence of difference in mortality rates comes as no surprise since no evidence of reduction in overall mortality has ever been proven [15]. In a meta-analysis, Janda and Swiston. brought to light that there was no difference in mortality rates with intrapleural fibrinolytic therapy [31]. In this work, we reported mortality and surgical referral rates concordant with other studies in literature, mostly using chest tube drainage strategy [32].
RTT had previously been identified as an effective alternative to chest tube in complicated parapneumonic effusion [9]. An experimental study focused on thoracentesis in the treatment of pleural empyema in rabbits revealed that thoracentesis was at least as effective as chest tube in terms of mortality [33]. In a recent review of literature, only few and old studies of limited sample size reported low rates of success with this strategy of drainage [34]. In other studies, clinical success rates ranged from 81 to 100% [9,10,35,36]. To our knowledge, our study is the largest to date which assessed RTT in pleural infection management. We reported clinical success in 82% of cases, and only three patients required chest tube as second-line treatment in the RTT-UD group. Additionally, thoracentesis with adjunctive fibrinolytic and DNase could be a strategy avoiding more invasive procedures [37,38]. Considering the half-life of DNase of 3-4 hours, a greater exposition time may improve effects on loculations [13]. In our study, even if fibrinolytic agents (i.e urokinase) and DNase remained in the pleural space until the next thoracentesis, adverse effects were rare with only one hemothorax and no pneumothorax requiring drainage.
Ferguson et al. found that pleural empyema was associated with failure of RTT [39]. However, in the RAPID-score, non-purulent pleural fluid was a predictor of poor outcome. We reported no failure in patients with non-purulent pleural fluid treated by RTT-UD. It suggests that RTT-UD could be considered as an effective alternative treatment strategy of pleural infections with non-purulent fluid. In case of empyema, switching to chest tube or surgery should be considered if necessary.
Our work has limitations. First, as a single-center study, our findings may differ from others. Second, as data were collected retrospectively over a large period of time (2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018), this study has potential biases. Nonetheless, heterogeneity in practice between the two groups was limited to adjunction of intrapleural DNase after 2011 and increased use of pleural ultrasonography. In addition, there were few missing data in the final database. Third, this study focused solely on pleural infections managed by RTT as first-line treatment. Hence, our population did not enroll most patients with severe presentation, such as patients admitted straight to ICUwhere chest tube is first-line treatment-or patients who required early surgical referral. Fourth, the study was not designed to compare outcomes between patients receiving intrapleural fibrinolytic agent plus DNase through chest tube and by RTT. Lastly, this study did not assess the impact of the main partner for treatment success: antimicrobial treatment. However, our practices regarding antibiotic use for pleural infections did not significantly change during the study period.
In conclusion, intrapleural use of urokinase and DNase was associated with better outcomes than urokinase only in the management of pleural infections managed with RTT. Pleural infections are hard-to-treat infections, but RTT failures were infrequent, confirming the efficacy of RTT-UD as an alternative strategy to chest tube drainage, especially in complicated parapneumonic effusions.