Abstract
Background
Pneumothorax is a common and important complication of transbronchial lung cryobiopsy (TBLC) in patients with interstitial lung disease (ILD), however little is known about what predisposes patients to pneumothorax during TBLC.
Methods
We reviewed the pre-TBLC ILD work-up of 76 patients who underwent TBLC in our tertiary ILD referral center from 2017 to 2023. We assessed factors including pulmonary function tests, bronchoalveolar lavage (BAL), blood tests, and imaging and to see if any of them caused an increased risk of iatrogenic pneumothorax.
Results
19 patients suffered a pneumothorax (25%). Neutrophilia on BAL (37.9% pneumothorax when present vs 12.8% when not present, p = 0.046) and the presence of ground glass opacification on high resolution CT-Thorax (38.2% vs 12.5%, p = 0.014) increased the risk of pneumothorax during TBLC. Other factors such as the presence of pleura in the biopsy sample (47.1% vs 17.2%, p = 0.021), as well as taking 3 or more biopsies (35.7% vs 11.8%, p = 0.019) were also significant predictors of pneumothorax.
Conclusions
Markers of acute inflammation as well as increased number of biopsies taken in a procedure lead to increased risk of pneumothorax. This may lead to changes in planning, patient selection and timing of TBLC procedures in ILD patients.
Transbronchial lung cryobiopsy (TBLC) is accepted as a reasonable alternative to surgical lung biopsy (SLB), and is shown in literature to be a safer, better tolerated and less expensive means of obtaining larger biopsy samples than traditional transbronchial biopsy (TBB) in patients with undifferentiated interstitial lung disease (ILD). 1 Despite advances in patient safety and tolerability, TBLC still comes with significant risk of adverse events, most notably pneumothorax and bleeding. Meta-analysis shows a procedure-related pneumothorax rate of 0%–26%. 2 On extensive literature review however, minimal data was found on factors that may predispose patients to having an iatrogenic TBLC-related pneumothorax.
We reviewed 7 years (2017 to 2023) of TBLC data from our centre, a large tertiary university teaching hospital, to see if we could identify any risk factors for pneumothorax in patients undergoing the procedure. We are a specialist ILD service receiving referrals from a catchment area of over 600,000 people. Following a thorough work-up of clinical and antigen exposure history, ILD-screening blood panel, high-resolution CT-Thorax (HRCT), pulmonary function tests (PFT) and bronchoscopy with bronchoalveolar lavage (BAL), each case is discussed at a multi-disciplinary team meeting involving consultant respiratory physicians, consultant histopathologists and radiologists with a specialist interest in ILD thoracic imaging. A decision is then made to proceed to TBLC in undiagnosed clinically appropriate cases.
Here, we discuss the outcomes of 76 patients who underwent TBLC for undifferentiated ILD in our centre, by looking at the investigations from their ILD work-up to see if there are any factors which may have predicted an increased risk of iatrogenic pneumothorax. A small minority of patients didn’t have full reliable reports of investigations performed in external centres-hence the slight difference in numbers of patients included for each variable. The reason for reviewing this data is to improve patient safety and outcomes in a procedure that comes with a high complication rate alongside high diagnostic yield. 3
TBLC is performed in our endoscopy day-case unit under conscious sedation with intravenous midazolam and fentanyl. Endobronchial balloon tamponade is used to control bleeding events. The procedures are not fluoroscopy-guided, due to lack of availability locally. This is a safe and efficacious means of performing this procedure.4,5 Every patient has a routine post-TBLC chest x-ray to exclude pneumothorax regardless of symptoms. There was an average of 2.7 samples taken per procedure, totalling 205 across 76 patients. The average size of the largest biopsy sample per procedure was 8.7 mm. 19 patients suffered pneumothorax from the study group of 76 (25%), however only 7 of these patients (9.2%) required chest drain insertion for clinically significant pneumothorax-as determined by the performing senior respiratory physician. There was no 30-days mortalities in our cohort.
Statistical analysis was performed by Fisher’s exact test for individual factors for univariate analysis, and a multiple logistic regression of significant factors was performed for multivariate analysis. The Wilcoxon test was used when we were looking at sets in which we were using continuous data-age and sample size. Given our sample size, only those variables that were significant on univariate analysis were included in the multivariate analysis, to avoid over-fitting. All factors that were significant on univariate analysis remained significant and independent of each other following multivariate analysis.
Pneumothorax events by presence or absence of listed factors (not all patients had complete BAL, PFT or HRCT data so not all factors total 76 cases). Univariate analysis by Fisher’s Exact Test and Wilcoxon test depending on whether we were using categorical or continuous data, and multivariate analysis of significant findings by Multiple Logistic Regression. Significance determined as p-value <0.05. Normal BAL values listed in italics, macrophages should be ≥ 85% and make up the remainder of the cell count. The values in bold denote factors that achieved statistical significance.
BAL = Bronchoalveolar lavage, HRCT = High-resolution computed tomography of thorax, GGO = Ground-glass opacity, IPF = Idiopathic pulmonary fibrosis, HP = Hypersensitivity pneumonitis, PFT = Pulmonary function tests, TLCO = Transfer factor for carbon monoxide, FEV1 = Forced expiratory volume in one second, mm = millimetres.
Regarding imaging, the presence of ground-glass opacities on HRCT was also a significant indicator. Other imaging features did not meet significance levels-including fibrosis, bronchiectasis/traction bronchiectasis and emphysema-likely due to lack of volume of cases with these features in the cohort. In assessing our cases, the presence of these features refers to the overall thoracic imaging and not specifically the area biopsied. Diagnosis of the two most common diseases in our group, idiopathic pulmonary fibrosis and hypersensitivity pneumonitis, also had no significant bearing on pneumothorax rate.
Peripheral blood eosinophilia did not confer an increased risk of pneumothorax. Unsurprisingly, patients with pleural tissue in their biopsy had a significantly higher rate of pneumothorax.
In terms of the procedure itself, taking 3 or more biopsy samples increased the risk of pneumothorax also. This could have important implications for patient safety and procedural planning. The same did not apply to the size of the samples themselves.
To summarise, it appears as if markers of acute inflammation in the lung- BAL neutrophilia, as well as the presence of ground-glass opacity on imaging-were significant predictors for pneumothorax in patients undergoing TBLC. It is unclear whether this observation would also apply to TBB and SLB, where there would be a larger cohort of documented cases. While not being an ILD or TBLC study, a 2021 publication showed that pneumothorax was more likely in transbronchial biopsy when patients had ground-glass lesions rather than solid lesions seen on their preceding CT. 6
A Finnish study of patients with TBLC performed under general anaesthetic with fluoroscopy guidance found that traction bronchiectasis was a risk factor for clinically significant complications, i.e., pooled pneumothorax and bleeding but not pneumothorax alone. 7 A study looking at endobronchial ultrasound-guided transbronchial biopsies showed significance for emphysema in causing pneumothorax, but we did not have enough emphysema cases in our study for meaningful statistical analysis. 8 Only three of our cohort were on steroids pre-procedure so we cannot gather any meaningful conclusions on their role either.
While we do have the HRCT report and BAL results prior to TBLC, we do not know whether pleura will be present in the sample until after the procedure. However, we can decide how many samples are taken in the procedure, which our data shows is an important consideration for avoiding pneumothorax in this procedure. This could inform our thinking on how many biopsies we take during the TBLC.
Finding methods of risk stratifying our TBLC candidates is important given they have underlying ILD which already puts them at increased risk of pneumothorax. 9 In practice, having BAL data prior to TBLC in all cases could help optimise the timing of the procedure. In centres with access to fluoroscopy, HRCT imaging could help target the safest and most appropriate site for TBLC. Limitations in our study include the relatively small sample size, and that it is single-centre data and retrospective. However, this data is an interesting beginning to a discussion about TBLC-related complications, as it is still a relatively new procedure in the ILD field and we don’t have enough comprehensive studies yet. In our literature review there was also a paucity of data pertaining to the risk of pneumothorax attributable to inflammatory markers following TBB and SLB. It is not yet clear whether our findings will apply to all modes of lung biopsy, or be specific to TBLC. Further prospective data is needed to see if these findings can help better select or better time TBLC procedures for improved patient outcomes. It is possible that we could defer procedures to a safer time if we knew a patient had increased active inflammation in the lung that would increase the risk of significant complications.
Footnotes
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.