Interstitial lung disease (ILD) represents a heterogeneous group of diseases with different prognostic and therapeutic implications. The evaluation of patients with ILD is best achieved by a multidisciplinary approach combining clinical, radiologic, and pathologic features. In many circumstances a histological confirmation is required to establish the exact diagnosis.1–5 Several techniques have evolved over the years and are employed in many centres. Conventional transbronchial lung biopsy (TLB) is not currently recommended in fibrotic lung disorders due to its low diagnostic yield6; besides, the procedure presents risks.7

Surgical lung biopsy (SLB) provides adequate lung specimens for histopathological analysis and is recommended as reference standard by current guidelines in ILD when a non-invasive diagnosis cannot confidently be made.6,8 However, it requires hospitalization and chest tube drainage and may be associated with significant morbidity (2–20%), including prolonged air leak and acute exacerbation as well as mortality (2–17%).9–12 Moreover, many subjects are excluded because of increased risks of perioperative morbidity and mortality by a combination of advanced age, comorbidity, severe respiratory failure, and pulmonary hypertension.

Cryotherapy has been used in bronchoscopy since the sixties, mainly in the management of endobronchial lesions in patients with bronchial obstruction.13 More recently, transbronchial cryobiopsy (TBC) has been used for the assessment of ILD. The major advantage of TBC is that large pieces of tissue can be extracted with a higher percentage of alveolar tissue,14–18 fewer crush artefacts and less atelectasis.16–18 The greater volume of alveolar tissue correlates with a better diagnostic yield.15–19 Despite its less invasive nature, some associated complications have been described. The potential for bleeding and the pneumothorax rates up to 33%, often requiring inpatient admission and chest tube drainage, are the most commonly described.19–23

We here investigate the safety of TBC in subjects with ILD and possible factors associated with its complications, in order to improve candidate selection for this procedure.

Obtaining tissue from lung parenchyma is often crucial for diagnosis and treatment in many ILD. Over the last few years TBC has been implemented as a mean of obtaining histological samples in the diagnostic setting of ILD; it has been used as a quite new modality and carried in some institutions. Few studies have explored its safety issues. We have here analyzed the TBC complications and its associated risk factors.

Among the adverse events, pneumothorax represented the most important complication of TBC with an incidence of 22.4% (22 of 90 patients; chest tube drainage was required in 18 cases (20%). The incidence of pneumothorax described in previous studies varies. Some studies reported no episodes of pneumothorax14,26–30 while others reported pneumothorax rates of between 1.4% and 12% (comparable to regular forceps biopsies or even lower).15,31–34 In a retrospective analysis that included 297 ILD patients undergoing cryobiopsy pneumothorax was the most common complication occurring in 60 patients (20.2%), 46 cases (15.5%) requiring drainage.21 Another study showed the occurrence of pneumothorax in 33% (19/58) patients undergoing TBC, with the majority (15/19) requiring chest tube drainage and hospitalization (average 6 days; range 1–9 days).23 In these studies the procedure was taken under spontaneous breathing and the probe was cooled for approximately 5–6s.

Hagmeyer et al. observed more pneumothorax events in patients submitted to jet ventilation (43%; 3/7) when compared to patients sedated and under spontaneous breathing over a bronchoscopy tube (12%; 3/25).22 In our study, all patients underwent general anaesthesia with jet ventilation; it could be a potential reason for the higher pneumothorax rate when compared to other studies that maintained spontaneous ventilation during the procedure. Also, the approximated time that the probe was cooled in our study was higher than the time used in other studies with lower advents frequencies. Furthermore, all our patients were intubated and that could also have played a role in increasing the pneumothorax risk. In a study by Polleti and Hetzel the authors analyzed the experience of previous studies with patients intubated and not intubated. They verified that in the ‘intubated group’ the mean size of samples was higher and there were more complications than in the ‘non-intubated group’.35

In a recent interesting study by Bango-Álvarez et al., TBC was performed under spontaneous breathing, under moderate sedation and without fluoroscopy; the cryoprobe was moved forward distally until it was noted that it could not advance further and then retracted 1–2cm. Using this technique they showed a low occurrence of pneumothorax (4.7%) and lack of pleura in all of the biopsies; which questions the need for fluoroscopy.34 On the other hand, we used fluoroscopy to guide the probe and defined a safe position as having its tip perpendicular and about 1cm from the chest wall. Our study population was limited to ILD patients; in the majority of these patients it is necessary to take peripheral subpleural biopsies in order to obtain samples that adequately represent the histopathologic processes, possibly explaining the high pneumothorax incidence.

We found that the risk of pneumothorax increases when fragments of pleura are present in the biopsy, similar to previous reports.19 In the study by Ravaglia et al. the degree of fibrosis of the lower lung zones was significantly increased in patients with pneumothorax compared to those without pneumothorax; furthermore, the authors also found a significant correlation between the incidence of pneumothorax and the histological UIP pattern.21 In our study we found opposite findings; only one patient with pneumothorax presented UIP/possible UIP pattern and only two patients presented fibrotic histology (comparing to other possible patterns), although with only a marginally statistical significance. This could be explained by the presumable additional caution in the execution of the technique because of the major risk of events in those patients.

It also must be noted that only a proportion of the patients would require chest tube drainage with TBC, compared with 100% patients who undergo surgical lung biopsy. In our sample, the pneumothoraces requiring drainage rapidly resolved; the median drainage time was 1 day, which was lower than the drainage time described in the previously referred studies. This result is important as chest tubes inserted for a prolonged time could be associated with an increased rate of infection.36

Although there is the potential for a large defect in the lung parenchyma following cryobiopsy, the occurrence of bleeding was low in our group (14.4%) and all cases were controlled only by local measures. In the majority of studies on this topic there were no episodes of severe bleeding needing blood transfusion or surgery and some studies did not report haemorrhage events15,30; other studies showed the occurrence of bleeding in up to 56.4%.16,17,22,29,31,33 The systematic review in the study by Ravaglia et al. reported no severe bleeding even in studies not employing the Fogarty balloon and moderate bleeding was documented in 7 out of 12 studies; however, the definition of ‘moderate’ bleeding was vastly divergent between the studies.21

The low incidence of significant bleeding in our cohort could be explained partly due to the preventive use of a bronchial blocker balloon, which is inflated immediately while the bronchoscope is being withdrawn from the airway to remove the biopsy specimen from the cryoprobe. Moreover, the haemostatic effects of cooling by cryo could probably have contributed to diminishing the occurrence of that event. However, the preventive use of a bronchial blocker balloon in our study is considered a ‘bias’ in the interpretation of the results as it does not allow us to quantify the true incidence of bleeding complications.

We found a positive association between the risk of bleeding and BMI that could be explained by the higher restriction and collapse of the peripheral lung areas in heavy patients that make it difficult for the probe to advance more peripherally. Thus the biopsies are done in more central areas increasing the bleeding risk.

Exacerbations of the ILD after the cryobiopsy are rare as well as associated deaths19,21 which occur less than in surgical lung biopsy.9–11 In our cohort, we did not have any acute exacerbation.

One of the major limitations of our analysis was the sample size. Although the collection of the data was prospective some data were missing. We also were not able to evaluate the influence of the greater experience of the bronchoscopist on the risks of complications.

In conclusion, pneumothorax was the most frequent complication, occurring in about a quarter of cases, and the presence of pleura in the biopsy was associated with its increased risk. Increased BMI seemed to be a risk factor for increased bleeding. The present analysis highlights the relative safety of the TBC procedure comparing to other possible approaches. Despite the obvious tendency to an increase in the use of this technique in different centres, there is still a lack of standardization of the procedure, as well as lack of defined indications and contra-indications, which would be of key importance not only to increase the diagnostic yield but also to reduce the potential complications of this technique. Another important issue is the grading of complications like bleeding, since its definition also differs among centres. More studies about TBC complications are needed in order to improve candidate selection for this procedure.

The authors have no conflicts of interest to declare.