Complications of a lung biopsy for severe respiratory failure: A systematic review and meta-analysis

Background This systematic review and meta-analysis aimed to evaluate the complications of lung biopsy in patients with acute respiratory failure (ARF), including acute respiratory distress syndrome (ARDS). Methods We searched the MEDLINE and Cochrane Central Register of Controlled Trials. The primary outcomes were biopsy-related death, respiratory failure, cardiac complications, bleeding, and other major complications. We used the McMaster Quality Assessment Scale of Harms (McHarm) to evaluate the risk of bias. A random-effects model was used to calculate the pooled frequencies. Results Thirteen studies (consisting of 574 patients) were included in the meta-analysis. Furthermore, most of the included studies had a high or unclear risk of bias in half of the items in McHarm. All included studies evaluated surgical lung biopsies. The median overall hospital mortality was 53% (range: 17%–90%). The pooled frequencies of biopsy-related death, respiratory failure, cardiac complication, bleeding, and other major complications were 0.00% (95% confidence interval [CI]: 0.00%–0.21%), 1.30% (95% CI: 0.00%–5.69%), 1.03% (95% CI: 0.00%–3.73%), 1.46% (95% CI: 0.16%–3.56%), and 4.26% (95% CI: 0.00%–13.0%), respectively. Conclusions The results of this study will be valuable information in considering the indications of lung biopsy in patients with ARF, including ARDS. Trial registration The protocol was registered with the University Hospital Medical Information Network Clinical Trials Registry (UMIN 000040650).


Introduction
Acute respiratory failure (ARF) and acute respiratory distress syndrome (ARDS) are caused by various underlying conditions, such as pneumonia, sepsis, and trauma [1]. Therefore, in their management, diagnosing the underlying condition and prompting specific treatment is essential because the mortality rate of ARDS remains high (approximately 30%) [2]. Previous studies have suggested lung biopsy is useful for verifying the cause of ARF or ARDS. It has been reported that open lung biopsy (OLB) provides a specific diagnosis in 85% of patients and contributes to management changes in 73% of patients with ARDS [3]. As a specific diagnosis, interstitial lung disease (25%), infectious disease (24%), and neoplastic disease (12%) were more common than diffuse alveolar damage (DAD) (9%), which is commonly observed in ARDS [3]. Therefore, pathological assessment using lung biopsy may play an important role in managing ARF and ARDS.
However, the safety and feasibility of lung biopsy for severe respiratory failure remain unclear. Furthermore, some metaanalyses on lung biopsy in patients with ARDS have shown that 22%e29% of patients undergoing OLB suffer biopsyrelated complications such as pneumothorax, persistent air leaks, bleeding, and infections [3,4]; these studies were not derived from a systematic review using a rigorous approach. Furthermore, while the definition of ARDS was revised from the American-European Consensus Conference (AECC) criteria to the Berlin criteria in 2012 [1], the latest metaanalysis of 14 studies assessing the complications of a lung biopsy for patients with ARDS contained only one study published after the ARDS definition change [4]. Therefore, an updated systematic review and meta-analysis focused on the safety of lung biopsy for severe respiratory failure is required.
As a result, we conducted a systematic review and metaanalysis of the frequency of complications caused by lung biopsy in patients with ARF or ARDS.

Materials and methods
This study is a systematic review and meta-analysis of the frequency of lung biopsy complications in patients with ARDS or ARF. We used the Preferred Reporting Items for Systematic reviews and Meta-analysis (PRISMA) [5] and its extension, which addresses harm (PRISMA harms) [6], to report our findings. The review protocol was pre-registered with the University Hospital Medical Information Network Clinical Trials Registry (UMIN 000040650) and has been previously published [7]. In addition, the need for ethical approval and consent was waived due to the nature of the systematic review.

Population and study eligibility
The study participants were critically ill adult patients (16 years or older) with ARDS or ARF in an ICU or emergency department who required mechanical ventilation and those who underwent lung biopsy. For lung biopsy, we defined the procedures, including surgical lung biopsy (SLB), which consists of OLB and video-assisted thoracoscopic surgery (VATS), transbronchial lung biopsy (TBLB), and cryobiopsy. In OLB, the lung tissue was surgically removed through an incision between the ribs. In VATS, a thoracoscope and forceps were inserted into the chest cavity through minimal incisions in the chest wall, and the lung tissue was collected under the view of the thoracoscope. TBLB and cryobiopsy were performed using flexible bronchoscopy, and the lung tissue was collected by forceps or cryoprobe (for freezing the lung tissue). In addition, we considered the definitions used in primary studies. We included all studies, including randomized controlled trials and observational (cross-sectional, prospective cohort, and retrospective cohort) studies. We excluded case reports, case-control studies, review articles, and studies that did not attempt to evaluate the complications of lung biopsy.

Study selection and data collection
Two authors (HS and YF) independently screened all studies identified in the search strategy by reviewing the titles and abstracts. After the first screening, two independent authors (YF and HI) assessed each study for eligibility. Disagreements r e s p i r a t o r y i n v e s t i g a t i o n 6 1 ( 2 0 2 3 ) 1 2 1 e1 3 2 between the reviewers were resolved through discussion or by a third reviewer (HS). Furthermore, two authors (YF and HI, or HS and YY) independently extracted information from the included studies, and the mismatch was resolved through discussion. We attempted to contact the corresponding authors through email regarding the unknown data.
The following data were extracted using a pre-defined data extraction form: study characteristics (author, year of publication, country, design, sample size, clinical settings, number of studies, and funding source), patient characteristics (inclusion/exclusion criteria, number of dropouts with reason, and patient demographics such as age and sex), type of lung biopsy procedure, and frequency of lung biopsy complications (biopsy-related death, respiratory failure, cardiac complication, bleeding, pneumothorax, infection, and other complications). The primary outcomes were biopsy-related death, respiratory failure, cardiac complications, bleeding, or other major complications, and the secondary outcomes were pneumothorax, infection, or other minor complications.

2.4.
Assessment of study quality and risk of bias  (Table S2) [8]. McHarm is a reliable criterion developed for assessing the internal validity of harms in intervention studies by Santaguida and colleagues using a literature review of harms and the Delphi method. Two investigators (YF and HI, or HS and YY) independently evaluated the risk of bias in each included study. Any disagreements were resolved through discussion or by a third reviewer. Due to the absence of evidence for publication bias in the studies on harm and the lack of reliable methods for its assessment, no statistical evaluation of publication bias was performed. In addition, the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) approach was used to assess the quality of evidence for each complication of lung biopsy [9].

Statistical analysis and data synthesis
The frequency of complication estimates with 95% confidence intervals (CIs) was indicated in paired forest plots to inspect the between-study variance. Because the frequency of adverse events was often expected to be small or zero, we used Freeman-Tukey Double Arcsine transformation to stabilize the variances and then performed a random effects metaanalysis using the DerSimonian-Laird method [10,11]. We excluded studies lacking data for each outcome from the meta-analysis. We performed post hoc sensitivity analyses for the included studies to assess the design bias. All analyses were performed using Review Manager 5.4 (Cochrane Collaboration, London, United Kingdom) and STATA 16 (StataCorp LLC, College Station, TX, USA).

Results
A total of 2214 studies were identified and screened. Among them, 13 studies (consisting of 574 patients) met the eligibility criteria and were included in the quality assessment and meta-analysis (Fig. 1). The characteristics of the included studies are presented in Table 1 and Table S3 [12e24]. No randomized controlled trial was identified. Two studies were prospective cohort studies, and the remaining 11 were retrospective cohort studies. The median number of patients in each study was 36 (range: 14e101). The median age of the patients and the proportion of male sex in each study ranged from 33 to 67 years and 44%e75%, respectively. The median overall hospital mortality was 53% (range: 17% [in 7 days]e 90%). The conditions of the participants were ARF requiring mechanical ventilation (five studies) and acute lung injury or ARDS (eight studies). Five studies and three studies used the AECC and Berlin criteria for diagnosing ARDS, respectively. All studies assessed the complications of SLB in ICU patients, and biopsy procedures were performed in the ICU or operating room. Complications in the included studies are presented in Table 2.

Risk of bias assessment
The summary of the risk of bias in the included studies is shown in Fig. 2. Furthermore, most studies were evaluated as having a high or unclear risk of bias in half of the items in McHarm. For example, in the item regarding the timing and frequency of collection of the harms, 10 of 13 studies were considered to have a high risk of bias because these studies did not clearly specify in the methods section when or how often information on complications was collected. We present the GRADE evidence profile of each complication of lung biopsy in Table 3. For each outcome, we judged the component for the risk of bias as very serious or serious based on the number of studies evaluated as high or unclear risk of bias in the items in McHarm.

Primary outcome
We present the forest plot for the pooled frequency of primary outcomes in Fig. 3. Respiratory failure and cardiac complications had high heterogeneity in the primary outcomes (Table  3). For biopsy-related death, no cases were observed in 11 studies (n ¼ 502), and the pooled frequency was 0.00% (95% CI: 0.00%e0.21%, I 2 ¼ 0.0%) (Fig. 3A).

Secondary outcome
We present a forest plot for the secondary outcomes in Fig. 4. All secondary outcomes showed high heterogeneity (Table 3).

Sensitivity analysis
In the sensitivity analysis, the pooled frequencies in prospective cohort studies tended to be lower than those in retrospective cohort studies; however, the results were similar (Table S4). The results of this study are robust.

Key findings
This systematic review and meta-analysis, including 13 studies, revealed that the median overall hospital mortality was 53%, and no biopsy-related deaths were reported in lung biopsies for ARDS or ARF. Furthermore, less than 5% of other severe complications occurred. While most included studies had a high or unclear risk of bias in half of the items in McHarm, we believe these results may help consider the feasibility of performing lung biopsy for patients with ARF or ARDS.

Strengths
To the best of our knowledge, no previous PRISMA-compliant systematic reviews and meta-analyses have focused on the complications of a lung biopsy for severe respiratory failure.    *Persistent air leak that required surgery. **Persistent air leak that did not require surgery. a Wound infection (n ¼ 2). b Wound infection (n ¼ 1) and empyema (n ¼ 1). c Persistent air leak that did not require surgery (n ¼ 8) and subcutaneous emphysema (n ¼ 3). d Persistent air leak that did not require surgery (n ¼ 9) and bronchopleural fistula (n ¼ 2).  OLB, open lung biopsy; VATS, video-assisted thoracoscopic surgery; CI, confidence interval; GRADE, grading of recommendations, assessment, development, and evaluation. a Persistent air leak that required surgery. b Persistent air leak that did not require surgery, subcutaneous emphysema, and bronchopleural fistula. c We judged this component as very serious because more than two-thirds of the studies were considered to have a high or unclear risk of bias in half of the items in the McMaster Quality Assessment Scale for Harms. d We judged this component as serious because a certain number of studies were considered to have a high or unclear risk of bias in the items in the McMaster Quality Assessment Scale for Harms. e Downgrade due to number of participants was less than optimal information size. We found two previous meta-analyses; one assessed 22 studies and reported the frequency of all complications [3], and another assessed 14 studies and reported the frequency of all complications and persistent air leaks [4]. However, both studies focused on the diagnostic use of lung biopsy rather than safety and did not estimate 95% CIs in the frequency of each complication. In these studies, the frequencies of all complications, mixed from severe to mild complications, were insufficient for clinical decisions. In addition, the risk of bias assessment was not performed in the included studies, and the literature review process was not rigorous.
Furthermore, this study is the first systematic review and meta-analysis following the reporting guidelines (PRISMAharm) and appropriate methodology. We assessed the risk of bias using McHarm and calculated the frequency of each complication with 95% CIs. Therefore, this study may be more valuable than previous studies regarding appropriate methodology.

Clinical implications
Clinicians should determine whether to perform a lung biopsy for ARDS based on a balance between disadvantages and clinical benefits; therefore, the results of this study may be valuable as evidence considering safety and disadvantages. The disadvantages of lung biopsy include complications and costs, such as medical resources. No study that assessed the cost of lung biopsy was included. In this study, serious complications of lung biopsy were less than 5%, with the most common complication being persistent air leak. These results are consistent with those of previous studies [3,4]. According to a recent meta-analysis published in 2015, the complication rate of OLB was 29% (95% CI: 25%e33%), and the most common complication was persistent air leak [4]. These results are similar to those of this study, although we excluded studies lacking data for each outcome from the meta-analysis to avoid underestimation. Therefore, the results of this study are reasonable. Persistent air leaks that required surgery were relatively rare (two of 77 persistent air leaks) and may not be a reason to avoid lung biopsy if properly managed. However, considering the potential risk of indirect lung injury associated with the biopsy procedures and complications is also necessary. For instance, air leaks may limit high positive-end expiratory pressure, and prone positioning is difficult if chest tube placement is required.
In addition to the merits of confirmation or exclusion of diagnosis, some clinical benefits of a lung biopsy for ARDS have been reported. First, lung biopsy for ARDS may be useful in predicting the response to corticosteroid therapy. While it is still controversial, corticosteroids are considered effective against ARDS [25]. It has been reported that OLB could identify a corticosteroid-sensitive pathology associated with lower hospital mortality in patients with ARDS [18]. Furthermore, lung biopsy for ARDS may be able to estimate the prognosis of ARDS. It has been reported that ARDS patients with pathological DAD have poorer survival than those without DAD [26,27].
Consequently, the information about the complications of lung biopsy in patients with severe respiratory failure, including ARDS, may be useful, particularly in patients with unknown etiology and in settings that require justification for using corticosteroids.

Limitations
This study has some limitations. First, this systematic review and meta-analysis did not include all studies that should be included. We did not search for EMBASE and excluded non-English publications because of poor accessibility in Japan. Furthermore, some studies were excluded from our metaanalysis due to a lack of data, although we had tried to contact the authors of the primary studies; therefore, a risk of selection bias may be present. Second, the number of studies and patients included in this study may be too small to assess rare complications, such as biopsy-related death. In addition, most included studies had a high risk of bias, and the quality of the study design in r e s p i r a t o r y i n v e s t i g a t i o n 6 1 ( 2 0 2 3 ) 1 2 1 e1 3 2 each included study was poor. No randomized controlled trial was identified in this study, and only two of the 13 included studies were prospective cohort studies. These limitations may decrease the certainty of the results of this study.
Furthermore, the inclusion criteria for each included study had heterogeneity (i.e., ARDS defined by the Berlin criteria or by the AECC criteria and ARF). Patients' conditions, such as coagulopathy, have not been evaluated appropriately. Therefore, more studies and patients may enable us to create subgroups that assess heterogeneity. In addition, studies differed in their definition of complications. It was difficult to distinguish between biopsy-related deaths and deaths due to underlying diseases, particularly in a retrospective cohort study. While no biopsy-related death was identified in our study, real-world data showed that non-elective SLB for interstitial lung disease had a 16.0% in-hospital mortality [28]. This discrepancy suggests that the results of the meta-analysis may be underestimated, and further studies are needed to clarify the definition of complications.
Finally, we could not determine the complications of TBLB or cryobiopsy for ARDS because of a lack of data. In addition, this study did not evaluate the effect of bronchoalveolar lavage, which is often performed simultaneously with a lung biopsy. Therefore, further studies on the complications of lung biopsies using bronchoscopy for ARDS are required.

Conclusions
We conducted a systematic review and meta-analysis that focused on the complications of a lung biopsy for ARF or ARDS. In this study, no biopsy-related deaths were identified under the median overall hospital mortality of 53%, and the pooled frequency of severe complications was <5%. These results will be valuable information in considering the indications of lung biopsy in patients with ARF, including ARDS.

Ethics approval and consent to participate
Not Applicable.

Consent for publication
Not applicable.

Availability of data and materials
The data used in this meta-analysis were obtained from the articles corresponding to references in our reference list.