Objectives: A pilot study to assess the feasibility of a randomization trial protocol comparing low (2.5mg) and standard (10mg) doses of intrapleural tissue plasminogen activator (tPA) with deoxyribonucleases (DNase) in unresolved pleural infection.

Hypothesis: The proposed randomized protocol comparing low dose and standard doses of intrapleural tPA (with DNase) therapy will be feasible and acceptable.

Design and subjects: A single-center, two-arm, double-blinded, randomized controlled feasibility study which includes subjects with unresolved pleural infection eligible for intrapleural tPA/DNase injection, with follow-ups till 3 months after hospital discharge.

Interventions: Recruited subjects will be randomized in 1:1 ratio to receive a maximum of 6 doses of intrapleural tPA (with DNase) starting at either 2.5mg or 10mg. A clinical decision is allowed at or after the third dose of tPA to continue with the assigned regimen (blinded) or convert to open-label use of 10mg doses of tPA to complete the course based on the clinical response.

Main outcome measures: The main outcome is the feasibility of the trial protocol, based on the percentage of eligible patients enrolled, retained to discharge, and completing 3 months of follow-up. Other important outcomes include survival at 3 months follow-up and without the need for surgical intervention, the need for additional pleural interventions, the number of decisions to convert to open-labelled use of 10mg intrapleural tPA, clinical and radiographic response after the treatment course, safety profiles, especially bleeding complications and the number and reason for protocol violation.

Data analysis and expected results: Feasibility outcomes will be reported as descriptive data. Comparison of outcomes between the two treatment groups will be analyzed on an intention-to-treat basis. Safety outcomes will be reported descriptively for each group. The reported estimates of recruitment rates, adherence, follow-up completeness, and variability and event rates for key clinical and bleeding outcomes will be used, to inform the design and sample size considerations future studies incorporating the current study design

Pneumonia is one of the commonest diseases that require hospitalization and it ranks second as the cause of death in Hong Kong. Parapneumonic effusion (PPE) is frequently associated with pneumonia and leads to a worse prognosis. Complicated PPE (CPPE) and empyema, collectively known as pleural infection, can increase mortality and consumption of healthcare resources. Complete evacuation of pleural fluid can be challenging even with adequate large-bore chest drain placement, due to the viscous nature of the fluid and extensive septations partitioning the infected fluid into multiple locules in the pleural space. Antibiotic treatment and drainage of infected pleural effusion can only cure 70% of pleural infections. Surgical decortication is associated with unavoidable bleeding, anaesthetic risk and a high rate of chronic post-thoracotomy pain. The advent of intrapleural fibrinolysis therapy by a combination of tissue plasminogen activator (tPA) and deoxyribonuclease (DNase) can greatly improve drainage of pleural fluid and avoid surgery in more than 90% of patients.

The optimal dose of intrapleural tPA (to be used with DNase) remains undefined. The original "standard" dose (10mg) used in the landmark MIST-2 randomized controlled trial (RCT) was chosen empirically and had not been subjected to conventional dose-escalation assessment or long-term pharmacovigilance follow-up. Tissue plasminogen activator accounts for the majority of the cost of the combination tPA/DNase therapy and has potential bleeding risks.

Rates of bleeding following intrapleural tPA/DNase therapy vary from <5% to as high as 17% in published studies. Clinically significant bleeding warrants additional intervention including packed cell transfusion, radiological interventions and even surgery. This risk is therefore one of the driving forces urging the need to find a safer and lower effective dose. It is likely that the risk of pleural bleed is dose-dependent. A study using 20mg of intrapleural tPA was associated with a doubling of serious bleeding complications (28%; versus 10mg of tPA). Studies of de-escalation of tPA doses appear to show a reducing trend in bleeding rates with lower starting doses of intrapleural tPA: 4.9% and 2.9% when using 5mg and 2.5mg of tPA respectively.

The multi-center dose de-escalation ADAPT and ADAPT-2 studies have confirmed that a starting dose of 5mg and 2.5mg intrapleural tPA were safe and effective. They employed a pragmatic approach to begin therapy with a lower tPA dose (2.5mg or 5mg) and allow escalation to the conventional dose (10mg) if attending clinicians are concerned with a lack of clinical response. In the ADAPT-2 study, a starting dose of tPA at 2.5mg remained efficacious, with only 2 patients (2.9%) requiring surgery and a low rate of bleeding complications (2.9%), comparable with original RCT using 10mg of tPA. The lower dose of tPA, with potentially lower bleeding risk, provides an attractive alternative, especially in patients who require therapeutic anticoagulation and those with acquired coagulopathy due to overwhelming sepsis.

There is no high-quality, RCT data comparing the therapeutic outcomes and bleeding rates between low (2.5mg) and standard doses (10mg) of intrapleural tPA in pleural infection. The lack of this important evidence is also reflected by heterogeneous practice in using different doses, ranging from 1 to 10mg, of intrapleural tPA from a recent international survey among experts in pleural medicine. The respondents would consider using a lower starting dose of tPA (with the possibility of escalation if clinically needed) if a median of 80% of patients could be successfully treated at that dose. A multicenter RCT involving low and standard doses of intrapleural tPA is required to compare their therapeutic efficacy and safety profiles. Such studies are challenging to design. Clinically relevant and patient-oriented outcomes are clearly preferable over conventional radiographic clearance. There are no set definitions of clinical resolution of pleural infection due to many reasons. These patients are complex; separating the pleural infection from concurrent/underlying source of (usually lung) infection is difficult as inflammatory parameters and blood markers are confounded. Patients typically have multiple comorbidities; clinical "improvement" or "deterioration" has to be considered taking into account multiple parameters. Decisions to stop/step down antibiotics or to refer to surgical decortication, to insert more or to remove chest tubes, and when to discharge home, all need to be individualized and cannot be governed by set rules. Clinical practice varies across institutions and among different countries. For these reasons, the study dose is preferably blinded, given the element of subjective decision on treatment success or failure. Therefore, a special and pragmatic trial design is necessary to allow the clinicians to determine if the randomized and blinded dose is providing the expected clinical response. Allowing the clinicians to revert to an open-label use of "standard" 10mg of tPA at any point after the initial 24 hours of treatment would alleviate any concern from physicians and patients on delay in receiving effective treatment doses.

Given the high success rate and relatively low bleeding complication rate, a full-scale non-inferiority study will require a very large cohort. Before that should be undertaken, a pilot feasibility study is necessary to assess this pragmatic trial design, including its acceptability to patients and clinicians (in the form of enrolment rates), the practical running of the protocol (e.g. compliance) and completion rates. The likelihood of clinicians reverting to open-labelled 'standard' 10mg dosing will also provide useful data to guide future trial design.

This RCT should also incorporate the flexibility of tPA dose escalation if an ineffective initial clinical response occurs in the low-dose tPA group with pre-set criteria. This would alleviate any concern from participants and enrolling clinicians that patients may be disadvantaged in receiving lower doses of treatment.

This pilot feasibility RCT will test the feasibility of a study design mimicking real-world use of intrapleural tPA/DNase for unresolved pleural infection, with two different starting doses of intrapleural tPA, while preserving the clinical flexibility of dose escalation during the course of treatment.

Inclusion Criteria:

• Clinical features suggesting uncontrolled pleural infection with incomplete drainage of pleural effusion, at least 1 day after insertion of pleural drain (French size 12 or above) and administration of antibiotics.
• Intend to administer intrapleural fibrinolytic
• Written informed consent obtained

Exclusion Criteria:

• Previously received intrapleural tPA to the ipsilateral pleural space for the current episode of pleural infection.
• Known sensitivity to tPA or DNase.
• A coincidental stroke, major haemorrhage or major trauma.
• Frank bleeding or evidence of puncture to the intercostal artery during chest drain insertion.
• Ongoing frank bleeding from the ipsilateral pleural space.
• Has had puncture of a non-compressible vessel in the previous 14 days.
• Has had major surgery in the previous 14 days.
• Has had unprovoked gastrointestinal bleeding or intracranial haemorrhage in the last 3 months.
• Active use of anticoagulation (except prophylaxis for deep vein thrombosis) or dual-antiplatelet agents.
• Active use of any systemic fibrinolytic therapy or any airway DNase therapy.
• On long-term macrolide antibiotics (as they may interact with DNase).
• Uncorrectable bleeding diathesis or baseline INR > 1.5.
• Has had a previous pneumonectomy (either on the same or contralateral side).
• Presence of active bronchopleural fistula.
• Age less than 18 years old.
• Patients who are pregnant or lactating (females of childbearing potential must have a negative pregnancy test before randomisation).
• Expected survival less than three months from a different pathology to this empyema (e.g. metastatic lung carcinoma).
• Use of agents under research or not registered in the 30 days prior to the study.
• Inability to give informed consent.

The research team, clinicians, nursing staff, and patients will be blinded to the treatment arm. A copy of the randomisation code will be held only by a research team member performing randomisation. Access to this code for a patient before the completion of the trial, will only be permitted in exceptional circumstances, such as being necessary to determine the best acute medical care for a patient, and only after discussion with the Principal Investigator (PI) or his deputy. Pharmacy staff will receive unblinded randomisation information in order to prepare the study drug of assigned doses. The labelling of the blinded drugs will be under the supervision of a dedicated pharmacy. Each course of trial treatments will be pre-prepared in a trial numbered pack and dispensed to the ward at randomisation.