Transbronchial lung cryobiopsy may be of value for nonresolving acute respiratory distress syndrome: case series and systematic literature review

Background Identification of pathologic features is helpful for the management of nonresolving acute respiratory distress syndrome (ARDS). Transbronchial lung cryobiopsy (TBLC) is a novel biopsy technique that may have comparable utility to surgical biopsy. The aim of this study was to assess the value of TBLC in patients with nonresolving ARDS. Methods All patients with nonresolving ARDS who underwent TBLC from January 2019 to August 2019 in a tertiary medical ICU were included. In addition, a literature search of TBLC for ARDS was performed by searching PubMed, EMBASE, ATS/ERS/APSR meeting abstracts, ClinicalTrials.gov, and Google Scholar. Data on complications, histologic diagnosis, management changes, and outcomes were analysed. Results Five patients (three women and two men) underwent TBLC. None of the patients developed pneumothorax, although two patients developed massive bleeding, which was controlled by continuous occlusion using bronchial blockers. There were no procedure-related deaths. Diffuse alveolar damage (DAD) and alternative histologic patterns were found in two and three patients, respectively, resulting in management changes in all cases. The literature search yielded four studies, which together with the present study comprised data from 25 cases in which TBLC was used in nonresolving ARDS. The summary diagnostic yield was 92% (23/25). Only 44% (11/25) of cases were proven to be DAD. TBLC contributed to management changes in 80% of patients (20/25). Procedure-related complications consisted of pneumothorax (16%, 4/25), significant bleeding (12%, 3/25), and persistent air leaks (8%, 2/25). There were no procedure-related deaths. The follow-up survival rate was 61.9% (13/21). Conclusions The complications of TBLC in selected patients with nonresolving ARDS may be acceptable. The procedure may have a high diagnostic yield and can lead to a re-evaluation of the diagnosis as well as changes in patient management. Further investigations with larger sample sizes are required.


Background
Acute respiratory distress syndrome (ARDS) is common in critically ill patients. The majority of studies report a high rate of morbidity and a mortality rate in the range of 35-60% [1,2]. The Berlin definition of ARDS takes into account clinical and radiological criteria in the diagnosis and classification without considering pathologic findings [3]. Diffuse alveolar damage (DAD) is a typical pathological finding in ARDS [4], but it is not seen in all ARDS patients nor is it specific to the disease. Studies have shown that only 43-45% of ARDS patients have DAD [5][6][7][8]. Identifying the pathologic pattern in nonresolving ARDS can facilitate treatment decision-making and improve outcomes [6,8].
When the inciting event that results in ARDS is unclear or when an alternative diagnosis is under consideration, surgical lung biopsy (SLB) is one of the most common procedures to obtain a histologic diagnosis. Several studies have shown a positive impact of the histologic results of SLB in patients with nonresolving ARDS [6,8]. However, SLB is performed in only 4-7% of patients with ARDS [6,9] due to the high surgery risk and poor conditions of most patients who are candidates for the procedure. Moreover, the complication rate of SLB is approximately 30%, with complications including persistent air leak and bleeding [9]. The use of less invasive procedures, such as transbronchial lung biopsy (TBLB), in clinical practice has been limited due to difficulties in obtaining adequate tissue to provide confident pathologic results (the diagnostic yield is < 35%) [9,10]. Transbronchial lung cryobiopsy (TBLC) is an alternative technique that has been widely used for the diagnosis of diffuse parenchymal lung disease (DPLD) [11]. The sample sizes are significantly larger than those obtained with TBLB, and the diagnostic yield approaches that of SLB with lower complication rates [12,13]. These observations suggest that TBLC may be an alternative procedure to SLB in patients with nonresolving ARDS, with potential benefits including adequate tissue for identifying pathologies and less invasiveness. However, to the best of our knowledge, the use of TBLC in nonresolving ARDS has been examined in only a few case reports. Therefore, in this study we assessed the use of TBLC in nonresolving ARDS based on 25 cases from our institute and a literature review.

Patients
TBLC was conducted in the event of persistent respiratory failure after ongoing lung infection was ruled out by previous bronchoalveolar lavage (BAL) or when an alternative diagnosis was suspected based on the patient's history, and clinical and radiologic presentation. The hospital records of all ARDS patients who underwent TBLC from January 2019 to August 2019 in our 26-bed tertiary medical ICU were reviewed. The inclusion criteria were as follows: age ≥ 18 years at the time of TBLC, findings consistent with the Berlin ARDS definition, and ARDS characterized as mild, moderate, or severe, as described in the Berlin definition, at the time of diagnosis and biopsy.

TBLC procedure
Five patients with ARDS underwent TBLC (Fig. 1). In four patients, the procedure was performed at the ICU bedside through an endotracheal tube (ETT), with the patient under deep sedation and supported by pressure control ventilation (F i O 2 100%, PEEP 0 cmH 2 O) or combined with extracorporeal membrane oxygenation (ECMO). The fifth patient received TBLC in a hybrid cone beam computed tomography (CBCT) operating room (OR). The procedure was conducted using rigid bronchoscopy with the patient under general anesthesia and ventilated by a high-frequency jet respirator (F i O 2 100%, respiratory rate 60 bpm, tidal volume 500 mL). In the four bedside procedures, bronchoscopy was first introduced through the nasal route to the lower airway after releasing the balloon on the ETT. Next, a long guidewire was inserted into the target lobe through the bronchoscopy working channel. Bronchoscopy was then pulled out alone and then reinserted through the ETT. Then, a bronchial blocker with a guide channel was introduced along the guidewire to the target lobe. Radial probe endobronchial ultrasound (RP-EBUS) (EU-ME1, Olympus, Tokyo, Japan) was used to identify the proper biopsy site (Fig. 1a). Under the guidance of preprocedurale CT images, the bronchoscope was advanced into the potential target bronchi as far as possible and then retracted 1-2 cm. When the surrounding RP-EBUS showed a heterogeneous echo without vascular presentation, the depth of the probe was marked. A 2.4 mm cryoprobe (ERBE, Solingen, Germany) was inserted to the same position (Fig. 1b, Fig. 1 C1). Cryobiopsy was performed (freeze time: 4 s) following probe positioning using carbon dioxide as the cryogen (Fig. 1d). After each biopsy, the bronchial balloon blocker (CRE balloon, Boston Scientific Microvasive, Natick, MA, USA) was immediately filled (0.5-1 atm) to stop the bleeding (Fig. 1e). Two to five biopsies were performed in each patient, and the sizes of the obtained samples were measured. For patients who underwent TBLC in the hybrid OR, flexible bronchoscopy and bronchial blockers were inserted through a rigid bronchoscope. Prior to TBLC, CBCT images (Artis Zee III ceiling, Siemens AG, Munich, Germany) were acquired to determine the exact position of the cryoprobe (Fig. 1 C2), which was placed under RP-EBUS guidance.
In patients with ECMO, the plan was to stop treatment with unfractionated heparin (UFH) 4 h before TBLC and then monitor the activated clotting time of whole blood (ACT). TBLC was performed when the ACT had decreased to within an acceptable range. After the procedure, the bronchial blocker was not removed until UFH had been reinitiated (Fig. 1f) and the ACT had recovered to the level previously achieved during ECMO with no active bleeding.

Outcomes
Patients routinely underwent a postprocedure chest Xray to screen for pneumothorax. Bleeding severity was graded on a scale of 4: no bleeding, mild bleeding (requiring suction for clearance but no other endoscopic procedures), moderate bleeding (requiring endoscopic procedures such as bronchial occlusion-collapse and/or instillation of ice-cold saline) and severe bleeding (causing haemodynamic or respiratory instability, requiring tamponade or other surgical interventions or transfusions) [14]. Other complications, if any, were recorded.
Data on the pathological diagnosis, treatment changes after TBLC and survival outcomes were obtained from the patients' medical records.

Literature search strategy
A literature search of PubMed, EMBASE, ATS meeting abstracts, ERS meeting abstracts, APSR meeting abstracts, ClinicalTrials.gov, and Google Scholar ending on November 25, 2019, was conducted using the following terms: "acute respiratory distress syndrome" or "ARDS," and "cryobiopsy" or "cryoprobe," without restrictions on language or publication year. The retrieved papers were read in their entirety to assess their appropriateness for this study of the value of TBLC in ARDS. Data on the characteristics of the TBLC procedure, related complications, pathological diagnosis, treatment changes and patient survival outcomes were extracted, and summary proportions were calculated based on individual cases.

Patient characteristics
Five patients (three women and two men) with nonresolving ARDS underwent TBLC at our hospital. Their average age was 53 years (range: 31-68 years). The characteristics of these patients are summarized in Table 1. Four of them had underlying disease, including multiple sclerosis, nephrotic syndrome, impaired glucose tolerance and advanced lung adenocarcinoma. Four of the patients were diagnosed with severe ARDS and one with moderate ARDS according to the Berlin definition. All four patients with severe ARDS received mechanical ventilation (MV) support, including two patients who received MV in combination with veno-venous ECMO. UFH was used in patients with ECMO but was stopped 4 h before TBLC in one patient. In the other patient (patient 1), because flow-through ECMO decreased significantly after UFH was stopped, heparin use was continued until 30 min before the biopsy procedure. Respiratory support in patients with moderate ARDS consisted of high-flow nasal cannula oxygen therapy (HFNC). All TBLC procedures were successfully performed using a 2.4 mm cryoprobe (freeze time: 4 s). A mean of 3.2 samples (range: 2-4 samples) were obtained from one (1 patient) or two (4 patients) lung segments. The samples had a mean size of 27.1 mm 2 (surface area) and were deemed satisfactory for use in histopathology and tissue culture (Fig. 2). The patient with moderate ARDS successfully recovered from general anaesthesia and was placed on HFNC after TBLC. None of the patients experienced significant oxygenation state changes before or after the procedure. No patient developed pneumothorax but two patients suffered massive bleeding, which ceased after 2 and 3 h, respectively, in response to tamponades using bronchial blockers. There was no incident of haemodynamic instability. Three other patients had mild bleeding. There were no procedure-related deaths. The bronchial balloon blockers in patients with ECMO were removed after confirming no active bleeding following the reinitiation of UFH.

Outcomes
Histopathological diagnosis in the five patients included a fibrotic phase of DAD with underlying infection (n = 1), proliferative phase of DAD with cytomegalovirus inclusions (n = 1), foreign body granulomas (n = 1), fibrotic non-specific interstitial pneumonia (n = 1) and cryptogenic organizing pneumonia (COP) (n = 1). One patient had an Acinetobacter baumanii infection, confirmed by tissue culture. Management changes (such as those in antibiotic and steroid usage (Table 1)) were made in all patients. Three patients improved and subsequently underwent rehabilitation treatments. Two patients died: one due to pulmonary embolism and the other because of septic shock.

Discussion
We assessed the safety and value of TBLC for ARDS in five patients with nonresolving ARDS at our hospital and then in a larger study that also included all patients from previous studies.
TBLC is most often used in patients with DPLD. A recent study [13] indicated that the diagnostic yield of TBLC is comparable to that of SLB, with a high level of diagnostic agreement between the two procedures. Furthermore, the feasibility and diagnostic value of TBLC has been demonstrated in patients after lung transplantation [19] and in those with peripheral pulmonary lesions [20]. However, the safety and potential value of TBLC in ARDS has not been established. A literature search of this topic was conducted and resulted in the identification of only three independent studies with a total of 20 patients [15][16][17]. Our combined analyses suggested that complications of TBLC in ARDS are acceptable and similar to those reported for TBLC used in DPLD during the stable stage [12]. The overall complication rate of SLB is approximately 30%, and most of these cases (71%) are persistent air leaks [21]; this rate of complications is significantly higher than that for TBLC (8%). Persistent air leak will increase the difficulty of management in ARDS patients with MV. And it may lead to increased duration of MV, hospital length of stay and hospital mortality [22]. Another advantage of TBLC over SLB is that patient and ventilation conditions may be easier to manage, and the requirement for surgeon skill may be lower. Finally, the cost of TBLC is much lower than that of SLB [23]. In China, the overall cost of TBLC is approximately 10,000 yuan compared to approximately 50,000 yuan for SLB.
However, ARDS is defined only according to clinical and radiological criteria without consideration of the pathological findings, which are known to be heterogeneous. Identifying specific pathological patterns is an important determinant of treatment. Gerard et al. [6] reported that 57% (29/51) of patients who underwent SLB had pathologies other than DAD. Moreover, 37% (19/51) of the patients had a steroid-sensitive pathologic disease pattern (OP, acute interstitial pneumonia, acute exacerbation of usual interstitial pneumonia, eosinophilic pneumonia, pneumocystis pneumonia, alveolar haemorrhage and amiodarone toxicity); these patients had significantly better outcomes than patients with steroid-resistant pathologies (in-hospital mortality rate: 37% vs. 65%; 180-day mortality rate: 37% vs. 75%). In our study, based on the combined data of reported cases, the summary diagnostic yield of TBLC was 92, and 56% of patients had a diagnosis other than DAD. Changes in management occurred after multidisciplinary discussions, based on the clinical condition, laboratory data, BAL results of the patient, and the pathological pattern of TBLC in our institute. TBLC led to management changes in 80% of patients after including reported ECMO is an important life support method for patients with severe ARDS. Almost all patients receive UFH to prevent thrombus formation and maintain normal ECMO operation, although the risk for procedurerelated bleeding is higher. The use of TBLC in ARDS patients on ECMO has not previously been described. However, our study included two patients with severe ARDS who were on ECMO support in whom TBLC was successfully performed. One suffered from procedurerelated massive bleeding, which was successfully stopped by 3 h of continuous occlusion using bronchial blockers. Thus, in patients with ECMO, TBLC should be performed after controlling for the bleeding risk and thrombogenesis, and bronchial blockers should be placed prophylactically. The rate of significant bleeding was 8.7% (2/23) in patients without ECMO.
The safety of TBLC depends to a large extent on the location of the cryoprobe [24]. A distance of < 1 cm to the pleura is associated with a significantly higher risk for pneumothorax, and a biopsy obtained too proximal to the middle third of the lung increases the risk for severe bleeding. RP-EBUS can identify diffuse lung lesions and their surrounding vessels, and encouraging results have been obtained using RP-EBUS-guided TBLC in patients with ILD [25,26]. Our study is the first to describe the use of RP-EBUS-guided TBLC in ARDS. Four patients with severe ARDS underwent ICU-bedside TBLC guided only by RP-EBUS. The patient with moderate ARDS underwent TBLC in a hybrid OR with the combined guidance of CBCT and RP-EBUS. None of the patients developed pneumothorax, but two had massive bleeding. Thus, the value of RP-EBUS guidance requires further investigations in larger populations.
By including all patients from current and previous studies, our study became the largest case series to preliminarily assess the safety and value of TBLC in ARDS. However, it had several limitations. First, it was a retrospective study, although the results support the need for a prospective controlled study with a larger population. Second, the procedure protocols were not similar between studies. Further investigations should be performed based on a standard TBLC protocol for patients with ARDS. Third, the results were limited by the small sample size, and strong support for the routine use of TBLC in patients with nonresolving ARDS is still lacking. TBLC should be performed after multidisciplinary discussion with consideration of the clinical condition, laboratory data of the patient, and other noninvasive methods, such as endotracheal aspirates which can be used to investigate lower airway inflammation [27]. The cost/benefit ratio should also be considered.

Conclusions
The complications of TBLC in selected patients with nonresolving ARDS may be acceptable. The procedure may have a high diagnostic yield and can lead to a reevaluation of the diagnosis as well as changes in patient management. Further investigations with a prospective design and a larger number of patients are required to confirm the value of TBLC for nonresolving ARDS.