From the Authors:

We thank Ravaglia and colleagues for their interest in MIST-3 (Third Multicenter Intrapleural Sepsis Trial) and their letter to the Journal about our recent article (1). In modern practice, the role of empyema staging in predicting failure of optimal image-guided drain placement remains unclear, given that most patients have mixed-stage disease and stage 3 empyema being a predictor of risk for conversion rather than a direct contraindication to video-assisted thoracoscopic surgery (2, 3). The point here is that, regardless of stage, an appropriately trained surgeon proceeding to video-assisted thoracoscopic surgery has the prerequisite training and expertise to proceed to decortication and convert to open thoracotomy as and when required in the safest environment with appropriate anesthetic support, including single-lung ventilation.

This is at odds with medical thoracoscopy (MT), which, by definition, is performed by physicians with no formal surgical training, in most cases with the patient under local anesthetic with awake sedation, as in the series referenced. MT has been proposed as “effective and safe” on the basis of a retrospective series equivalent to three patients per hospital per year, which questions the generalizability and applicability. These are the two largest multicenter series in the literature; however, the selection bias is a major limitation in the MT evidence base (4). To date, the only randomized controlled trial (RCT) addressing this question was underpowered (n = 32) and was methodologically flawed in its primary outcome (5).

Brutsche and colleagues reported that MT was “primarily successful in 91% of cases”; yet, half of these patients required postintervention fibrinolytics “for 3–5 days,” which, by virtue of this need for additional intervention, equates to a treatment failure rate of 49% (6). In the more recent cohort described by Ravaglia and colleagues (n = 131), treatment success, similar to the RCT by Kheir and colleagues (5), was reported at 75%. This is an improvement on the 66% success rate of chest tube and antibiotics (7) but still falls significantly short compared with surgery and combination intrapleural enzyme therapy (IET), both with success rates in excess of 85–90% (2). Moreover, MT has a considerable pooled complication rate relating to post-procedural surgical emphysema and persistent air leak that has been reported as high as 26.7% (4).

The authors suggested that stratification by RAPID (renal [urea], age, fluid purulence, infection source, dietary [albumin]) score (7), the only prospectively and externally validated prognostic tool in the literature, was inadequate because it did not consider ultrasound. The evidence linking sonographic parameters to clinical outcomes is limited to small retrospective case series (8).

We applaud Ravaglia and colleagues for their MT expertise and access, but their practice is based on what we perceive to be a very limited evidence base. A recent consensus statement suggested that physicians only occasionally consider MT in multiloculated pleural infection in elderly and frail patients considered to be at high surgical risk (2). We strongly encourage the authors to proceed with a well-designed, adequately powered, multicenter RCT to address this unmet need for a higher level of evidence in the literature.