- Email: [email protected]
Routine Chest Radiography for the Evaluation of Pneumothorax Following Bronchoscopy
Original Investigation
Routine Chest Radiography for the Evaluation of Pneumothorax Following Bronchoscopy Christopher P Centonze, MD, Matthew S Davenport, MD, Eric S White, MD, MS, Ella A Kazerooni, MD, MS
Rationale and objectives: To determine the utility of routine postbronchoscopy chest radiography to detect pneumothorax. Materials and Methods: This retrospective quality improvement cohort study was approved by the Institutional Review Board. All outpatients (n = 1443) who underwent protocol-driven postbronchoscopy chest radiography in one health system from January 2010 to July 2017 were identified by electronic medical record query. The prevalence of pneumothorax (with 95% confidence intervals [CI]) and clinical outcome were determined following coded review of chest radiography reports and review of the electronic medical record. The effect of smoking and lung disease on risk of pneumothorax was determined with Chi Square tests. Results: Of 1443 subjects undergoing interventional bronchoscopy, 6% (93/1443) were current smokers, 35% (505/1442) were former smokers, and 35% (540/1443) had known lung disease. Pneumothorax prevalence was 3.4% (49/1443; 95% CI: 2.6% 4.5%) following any intervention and 4.1% (42/1032; 95% CI: 3.9% 5.5%) following transbronchial intervention. In those without known pre-existing pneumothorax or a confirmed false positive diagnosis, the real overall pneumothorax rate was 2.9% (42/1443; 95% CI: 2.1% 3.9%). The risk of pneumothorax did not differ based on smoking history (p = 0.99) or history of lung disease (p = 0.19). Of 49 subjects with pneumothorax, 13 were symptomatic, and 10 had a change in management including chest tube placement (N = 2), inpatient admission (N = 3), and/or observation (N = 7). No pneumothorax-related intervention was performed in asymptomatic patients. Conclusion: Pneumothorax following interventional outpatient bronchoscopy is uncommon, usually asymptomatic, and often clinically insignificant. Asymptomatic postbronchoscopy patients are very low risk and may not need routine imaging. Key Words: Quality; Chest radiography; Bronchoscopy; Pneumothorax. © 2019 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.
INTRODUCTION
F
lexible bronchoscopy is a common procedure that can be used for diagnostic or therapeutic purposes, with low prevalence of morbidity and mortality (1,2). The overall complication rate ranges from 0.8% 5%,
Acad Radiol 2019; 26:585 590 From the Department of Radiology, Michigan Medicine, Ann Arbor, Michigan (C.P.C., M.S.D., E.A.K.); Michigan Radiology Quality Collaborative, Ann Arbor, Michigan (C.P.C., M.S.D., E.A.K.); Department of Radiology, University of Michigan Health System, 1500 E. Medical Center Dr. B2-A209P, Ann Arbor MI, 48108 (M.S.D.); Division of Pulmonary and Critical Care Medicine, Michigan Medicine, Ann Arbor, Michigan (E.S.W.). Received October 29, 2018; revised November 14, 2018; accepted November 17, 2018. Disclosures: (Resident) Chris Centonze No relevant financial disclosures. Matt Davenport Royalties from Wolters Kluwer. Eric White No relevant financial disclosures. (AUR member, senior author) Ella Kazerooni No relevant financial disclosures. Funding: No extramural funding solicited or used for this work. Address correspondence to: M. S. D. e-mail: [email protected] © 2019 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.acra.2018.11.025
with major complications including pulmonary hemorrhage, pneumothorax, and respiratory failure (3,4). Pneumothorax is the most common major complication. The greatest risk for pneumothorax occurs following interventions during bronchoscopy, such as transbronchial biopsy, with reported prevalence of 0%-5% (5). Our historical practice has been to obtain routine postprocedure chest radiographs on all patients following interventional fiberoptic bronchoscopy to exclude pneumothorax prior to outpatient discharge; however, there are no clear data to support this practice. According to the 2013 British Thoracic Society Guideline for Diagnostic Flexible Bronchoscopy in Adults, “a chest radiograph should be obtained if a patient is symptomatic or there is a clinical suspicion of possible pneumothorax after TBLB” (transbronchial lung biopsy) and “patients should be advised of the potential for delayed complications following TBLB and provided with written information regarding likely symptoms and action required.” Both of these were given a grade D recommendation due to evidence being available from nonanalytic studies, such as case reports or case series, or from expert opinion (6). The previous American 585
CENTONZE ET AL
Thoracic Society guideline from 1987 does not mention the use of chest radiography (7). Chest radiography is a common low-risk radiology procedure (0.06 0.25 mSv for 2 views) (8) and is historically the first-line test to evaluate for suspected pneumothorax, although some data suggest thoracic ultrasound may have superior sensitivity and negative predictive value in the setting of trauma (9). However, routine chest radiography can increase length of stay, increase radiation exposure, increase cost, create incidental findings that lead to additional tests, and produce false positive results that lead to unnecessary interventions (10 12). To evaluate the utility of routine radiography following outpatient bronchoscopy, we reviewed six years of data from our health system to determine the diagnostic yield and clinical outcomes following this practice.
Academic Radiology, Vol 26, No 5, May 2019
interventional pulmonologist and depends on patient symptoms, size of the pneumothorax, and the patient's living situation. Options for management include: discharge without intervention, discharge with follow-up chest radiography, repeat chest radiography, placement of chest tube, admission to observation, and admission to inpatient status. Risk Factors for Pneumothorax
Risk factors for pneumothorax including smoking history (smoking status and, if applicable, pack-year history) and pulmonary diseases were collected by electronic medical record query of coded fields, ICD-9 codes, and ICD-10 codes. Clinical Outcomes
MATERIALS AND METHODS Study Population
The study population included all adult outpatients who underwent 1- or 2-view chest radiography within four hours following interventional bronchoscopy between 1/1/2010 and 7/1/2017. There were no exclusion criteria. All bronchoscopy and related chest radiography procedures were identified using electronic medical record query of the Current Procedural Terminology (CPT) billing codes. Bronchoscopy
Following informed consent, the patient typically receives local anesthesia of the nasal passages and oropharynx prior to administration of intravenous conscious sedation. After insertion of the flexible bronchoscope, a visual inspection of the airway is performed. Once inspection is completed, selected diagnostic or therapeutic procedures are performed based on clinical indication. When appropriate, the procedure is guided by fluoroscopy, ultrasound, or virtual bronchoscopy (utilizing a previously obtained computed tomography [CT] of the chest). These imaging modalities are used independently or in conjunction with the other imaging modalities. In other cases, no imaging assistance is utilized. At the completion of the procedure, the patient is brought to postprocedure recovery for monitoring. Postprocedure Chest Radiography
There is a standard process for evaluating patients following bronchoscopy. Once the patient successfully recovers from conscious sedation, he or she is transported from recovery to radiology where either 1- (N = 278) or 2-view (N = 1165) chest radiography is performed. Upon completion, typically the interventional pulmonology fellow and attending review the radiograph for pneumothorax. Occasionally the pulmonologists await formal interpretation by a cardiothoracic radiologist. If the chest radiograph is negative for pneumothorax and the patient is clinically stable, the patient is discharged. If the chest radiograph is positive for pneumothorax, action is determined by the 586
All final postprocedure chest radiography reports were manually reviewed by a single member of the study team (C.C.) who was blinded to the interventional technique. Reports were coded as positive or negative for pneumothorax based on the text in the report. If listed, the size of the pneumothorax (if any) was recorded. Pneumothorax interventions were determined by electronic medical record query or CPT code analysis: admission to observation or inpatient status within 24 hours of the procedure, open-guided thoracostomy placement (32551), image-guided thoracostomy placement (75989). Manual electronic medical record review was performed for all patients diagnosed with pneumothorax to determine the specific bronchoscopy intervention(s) (e.g., brushing or washing, bronchoalveolar lavage, endobronchial biopsy, transbronchial biopsy, and dilatation), postprocedure symptoms (e.g., dyspnea, cough, and chest pain), postprocedure change in vital signs (e.g., tachycardia, and hypoxia), use of follow-up chest radiography, placement of a thoracostomy tube, and admission to observation, or inpatient status. Data Analysis
The overall population and subgroups (pneumothorax and no pneumothorax) were summarized with descriptive statistics, including counts, percentages, means, and ranges. The prevalence of pneumothorax and pneumothorax intervention were calculated, and 95% confidence intervals (CI) were derived. An unpaired t-test was used to evaluate differences in age between the two subgroups. Chi-square test or Fisher's exact test was performed to determine if there were significant differences in categorical variables between the two subgroups. The risk of pneumothorax with and without transbronchial intervention was compared with Fisher's exact test. In patients with pneumothorax, characteristics of the bronchoscopy procedure and imaging findings were summarized with descriptive statistics. Institutional Review Board approval was obtained and informed consent was waived for this single-center Health Insurance Portability and Accountability Act (HIPAA)-compliant retrospective study (Table 1).
Academic Radiology, Vol 26, No 5, May 2019
TABLE 1. List of International Classification of Diseases (ICD) Codes Used in the Data Query to Establish a History of “Lung Disease” Pulmonary Disease Unspecified Chronic Obstructive Pulmonary Disease Other Emphysema Simple Chronic Bronchitis Mucopurulent chronic bronchitis Other Chronic Bronchitis Unspecified Chronic Bronchitis Emphysema Cystic Fibrosis with Pulmonary Fibrosis Unspecified Pulmonary Fibrosis Idiopathic Pulmonary Fibrosis Rheumatoid Lung Unspecified Interstitial Pulmonary Disease Respiratory Bronchiolitis Interstitial Lung Disease Unspecified Chronic Obstructive Pulmonary Disease Unspecified Emphysema Mucopurulent Chronic Bronchitis Mixed Simple and Mucopurulent Chronic Bronchitis Unspecified Chronic Bronchitis Cystic Fibrosis with Pulmonary Manifestations Unspecified Pulmonary Fibrosis Other Specified Interstitial Pulmonary Diseases Idiopathic Pulmonary Fibrosis Rheumatoid Lung Disease with Rheumatoid Arthritis of Unspecified Site Unspecified Interstitial Pulmonary Disease Respiratory Bronchiolitis Interstitial Lung Disease Lymphoid Interstitial Pneumonia Idiopathic Nonspecific Interstitial Pneumonitis Cryptogenic Organizing Pneumonia Desquamative Interstitial Pneumonia Lymphangioleiomyomatosis Adult Pulmonary Langerhans Cell Histiocytosis
ICD-9 and ICD-10 Codes 491.2, 493.2, 496 492.8 491 491.1 491.8 491.9 492 277.02 515 516.31 714.81 516.9 516.34 J44.9 J43.9 J41.1 J41.8 J42 E84.0 J84.10 J84.89 J84.112 M05.10
J84.9 J84.115 J84.2 J84.113 J84.116 J84.117 J84.81 J84.82
ROUTINE CHEST RADIOGRAPHY AFTER BRONCHOSCOPY
(N = 540) had lung disease, 6% (N = 93) were current smokers, and 35% (N = 505) were former smokers [Table 2]. Patients with and without pneumothorax had similar predefined risk factors (all p >0.05) [Table 2]. The most common intervention was transbronchial biopsy (65.7% [N = 948]) [Table 3]. The majority of patients (60.8% [N = 878]) underwent at least two interventions during the same procedure. Forty-nine patients (3.4%; 95% CI: 2.5% 4.5%) had a pneumothorax identified on post-procedure chest radiography. Of these, 42 (4.1%; 95% CI: 3.9% 5.5%) had undergone transbronchial intervention (36 transbronchial biopsy (-ies) and/or 16 transbronchial needle aspiration (s). Numbers total greater than 42 because some patients underwent more than one intervention [Table 4]. Transbronchial intervention was associated with a higher risk of pneumothorax compared to those undergoing bronchoscopy without transbronchial intervention (p = 0.03). Of the 49 patients with a postprocedure pneumothorax, 53% of procedures used fluoroscopic image guidance, 45% used ultrasound guidance, and 37% used virtual bronchoscopy. Mean pneumothorax size (reported for 45 of 49 pneumothoraces) was 17 mm (range 1 150 mm) [Table 4]. In those without a reported size (N = 4), two were described as “small” or “tiny”, one was described as decreased from preprocedure imaging, and one had no qualitative descriptor. Twenty-six of the 49 patients diagnosed with pneumothorax had at least one follow-up chest radiograph (mean: 2.3 radiographs) and in one, the pneumothorax was determined to be likely artifactual. Thirteen (27%) patients with pneumothorax had postprocedure symptoms of dyspnea, chest pain, or cough. Five of 13 symptomatic patients had hypoxia and tachycardia, seven had stable vital signs, and two had no vital signs recorded. Conversely, thirty-two (65%) patients diagnosed with pneumothorax were asymptomatic. Six of the 49 diagnosed with pneumothorax had a pre-existing pneumothorax that either improved or did not change on postprocedure radiography. Excluding the 6 subjects with a pre-existing pneumothorax and the 1 subject with a confirmed false positive resulted in a real postprocedure pneumothorax rate of 2.9% (42/1443; 95% CI: 2.1% 3.9%). Seven patients with pneumothorax were placed in observation (four with symptoms, three without symptoms, or vital sign abnormalities); none of these seven required interventions and all were discharged following repeat radiography. Three patients with pneumothorax were admitted to the hospital; all three had symptoms of pneumothorax and two (0.1% of entire cohort) required chest tube placement. No chest tubes were placed in asymptomatic patients.
RESULTS
DISCUSSION
The study cohort (Fig 1) consisted of 1443 adults (mean age: 58 years [range: 18-92 years], 683 females [mean age: 58 years, range: 18 92 years], 760 males [mean age: 58 years, range: 19 90 years]) who had chest radiography within four hours of outpatient interventional bronchoscopy. Of the subjects 35%
Pneumothorax following interventional bronchoscopy is uncommon (overall: 2.9% [95% CI: 2.1% 3.9%]). When it occurs, it is usually asymptomatic and clinically insignificant. No chest tubes were placed in asymptomatic patients, and only two chest tubes were placed (both in symptomatic 587
CENTONZE ET AL
Academic Radiology, Vol 26, No 5, May 2019
Figure 1. Study population flowchart.
TABLE 2. Study Population Details Characteristic N (%) Mean Age (years § SD) Female Male Current Smoker Former Smoker Lung Disease
Total Population
No Pneumothorax
Pneumothorax
1443 58.2 § 13.9 47% (683) 53% (760) 6% (93) 35% (505) 35% (540)
1394 (97%) 58.1 § 13.8 47% (658) 53% (736) 6% (90) 35% (488) 38% (526)
49 (3%) 61.9 § 15.3 51% (25) 49% (24) 6% (3) 35% (17) 29% (14)
patients) after 1443 procedures (0.1%). If postprocedure radiography had been restricted to symptomatic patients (6), a large fraction of radiographs could have been avoided. Based on these data, the following approach to postbronchoscopy imaging might be considered: If the patient is asymptomatic,
p 0.09 0.61 0.99 0.99 0.19
no X-ray is indicated unless there are unusual circumstances (e.g., impending air travel). If the patient is symptomatic, an X-ray should be performed. The prevalence of postbronchoscopy pneumothorax we observed was similar to that reported in other studies
TABLE 3. Prevalence of Pneumothorax Following Interventional Bronchoscopy. Procedure Types are not Mutually Exclusive Procedure (not Mutually Exclusive) Interventional Bronchoscopy Bronchoalveolar Lavage, Aspiration, or Brushing Endobronchial or Bronchial Biopsy Transbronchial Intervention Transbronchial Biopsy, Single, or Multiple Lobes Transbronchial Aspiration, Single, or Multiple Lobes Placement or Revision of Tracheobronchial Stent(S) Tracheobronchial Dilatation or Fracture Reduction Thermoplasty, Single, or Multiple Lobes
588
N
No Pneumothorax
Pneumothorax
1443 (100%) 995 (69%) 160 (11%) 1032 (72%) 948 (66%) 263 (18%) 30 (2%) 105 (7%) 0 (0%)
1394 (97%) 968 (97%) 157 (98%) 990 (96%) 912 (96%) 247 (94%) 30 (100%) 103 (98%) 0 (0%)
49 (3%) 27 (3%) 3 (2%) 42 (4%) 36 (4%) 16 (6%) 0 (0%) 2 (2%) 0 (0%)
Academic Radiology, Vol 26, No 5, May 2019
ROUTINE CHEST RADIOGRAPHY AFTER BRONCHOSCOPY
TABLE 4. Pneumothorax Management (N = 49) Stratified by Related Symptoms (Cough, Dyspnea, and Chest Pain)
All pneumothoraces (N = 49) Symptomatic (N = 13) Asymptomatic (N = 36)
Mean Size [mm] (Range)
Follow-up Radiography
Chest Tube Placement
Inpatient Admission
Unexpected Observation
17 (1 150) 31 (4 150) 11 (1 47)
26 (53%) 10 (77%) 16 (44%)
2 (4%) 2 (15%) 0 (0%)
7 (14%) 4 (31%) 3 (8%)
3 (6%) 3 (23%) 0 (0%)
(3 5). In 1989, Milam et al (13) performed a single center retrospective cohort study of 207 patients undergoing flexible bronchoscopy, including 98 with transbronchial biopsy and 14 with transbronchial needle aspiration. In that cohort, one hundred thirty had an immediate postprocedure chest radiograph, with only one pneumothorax diagnosed (patient with pleuritic chest pain). The investigators suggested that postprocedure radiography should be restricted to specific high-risk groups, which they defined as comatose, mentally obtunded, or confused patients; patients on positive pressure ventilation (due to the risk of tension pneumothorax) or with respiratory compromise (who could not tolerate a small pneumothorax); and outpatients in whom delayed radiography could be used to detect a delayed pneumothorax, if patients became symptomatic. In 1990, Frazier et al. (14) conducted a retrospective study of 305 patients who underwent routine chest radiography after fluoroscopically guided transbronchial biopsy, and found only two pneumothoraces. Both patients had chest pain and a fluoroscopic abnormality. Based on these results, the authors concluded that routine chest radiography was not necessary in the absence of symptoms or a fluoroscopic abnormality. In 2003, Wu et al. (15) evaluated 454 consecutive inpatients and outpatients undergoing flexible bronchoscopy at a community-based hospital without pulmonary medicine or critical care fellows. They found one (0.2%) pneumothorax on routine postprocedure chest radiography in a patient who was clinically suspected of having a pneumothorax. The authors concluded that routine chest radiography after flexible bronchoscopy may not be cost effective or medically necessary in all settings. In 2006 (16) and 2016 (2), Izbicki et al. performed a pair of prospective studies to evaluate the utility of routine postprocedure chest radiography following transbronchial biopsy. The first study (16) included 350 consecutive patients who underwent fluoroscopically guided flexible bronchoscopy with transbronchial biopsy and had routine chest radiography performed two hours postprocedure. Ten (2.9%) patients had radiographically detected pneumothorax, including seven symptomatic patients with clinical suspicion for pneumothorax. The other three were asymptomatic, discharged, and had spontaneous resolution of their pneumothorax on follow-up imaging. The authors concluded that routine postprocedure chest radiography was not necessary in asymptomatic patients without other risk factors (e.g., mental status changes, positive
pressure ventilation). To test this conclusion, they (2) performed a second prospective study with 201 consecutive patients undergoing fluoroscopically guided flexible bronchoscopy with transbronchial biopsy. The patients were evaluated for symptoms at two hours in post procedure observation, and at 24 hours and 48 hours by telephone interview. A chest radiograph was only obtained, if pneumothorax was clinically suspected. Eight patients (4%) were diagnosed with pneumothorax. Six were symptomatic during the two-hour observation period including two requiring chest tubes. The remaining two pneumothoraces were detected outside the two-hour observation window (one presented to the emergency room at 26 hours, one had pneumothorax detected incidentally nine days later when imaged for another reason). The authors concluded routine chest radiography after transbronchial biopsy is not necessary in asymptomatic patients who maintain a stable oxygenation level. Our series is approximately five-times larger than the previously largest study and the results support prior work, but also have some limitations. It was a singlecenter retrospective review of outpatient bronchoscopy performed in consecutive patients in a dedicated bronchoscopy unit that conducts a high volume of procedures with support of pulmonary and critical care medicine fellows. It may not apply to sites with a smaller volume of procedures and different levels of experience. Postprocedure radiography was defined as a chest radiographic examination obtained within four hours of procedure completion. It is possible that some procedures were not captured within the defined cutoff time, although this is unlikely. Data was collected by query of the institutional electronic medical record, followed by manual review of all radiology reports and pneumothorax care patterns. There is a possibility that some procedures may have been missed, and some data may have been absent from the medical record. However, we consider it unlikely that clinically important pneumothoraces would not have been recorded. In conclusion, pneumothorax following interventional outpatient bronchoscopy is uncommon and when it occurs it usually is asymptomatic and clinically insignificant. Asymptomatic postbronchoscopy patients are very low risk and may not need routine imaging. In select circumstances (e.g., known impending air travel) it might be helpful to confirm the absence of a radiographically detected pneumothorax, but in most patients it is likely unneeded. 589
CENTONZE ET AL
REFERENCES 1. 2. 3. 4.
5.
6.
7. 8.
590
Casal R, Ost D, Eapen G. Flexible bronchoscopy. Clin Chest Med 2013; 34:341–352. Izbicki G, Ayal R, Nissim A, et al. Avoiding routine chest radiography after transbronchial biopsy is safe. Respiration 2016; 92:176–181. Pue C, Pacht P. Complications of fiberoptic bronchoscopy at a university hospital. Chest 1995; 107:430–432. Smith L, Singer J, Hayes M, et al. An analysis of potential risk factors for early complications from fiberoptic bronchoscopy in lung transplant recipients. Trans Int 2011; 25:172–178. Steinfort D, Khor P, Manser R, et al. Radial probe endobronchial ultrasound for the diagnosis of peripheral lung cancer: Systematic review and meta-analysis. Eur Resp J 2010; 37:902–910. Rand I, Blaikley J, Booton R, et al. British thoracic society guideline for diagnostic flexible bronchoscopy in adults: Accredited by NICE. Thorax 2013; 68:I1–I44. Diederich S, Lenzen H. Radiation exposure associated with imaging of the chest. Cancer 2000; 89:2457–2460. Sokolowski JW, Burgher LW, Jones FL, et al. Position paper on guidelines for fiberoptic bronchoscopy in adults. Am Rev Resp Dis 1987; 136:1066.
Academic Radiology, Vol 26, No 5, May 2019
9. Nagarsheth K, Kurek S. Ultrasound detection of pneumothorax compared with chest X-ray and computed tomography scan. Am Surg 2011; 77:480–484. 10. Brivet FG, Jacobs FM, Maitre S, et al. A clinically based chest X-ray strategy reduces health care costs without adverse consequences. Intensive Care Med 2006; 32:S136. ndez-Aguado I. Incidental findings in 11. Lumbreras B, Donat L, Herna imaging diagnostic tests: A systematic review. Brit J Radiol 2010; 83:276–289. 12. Ding A, Eisenberg JD, Pandharipande PV. The economic burden of incidentally detected findings. Radiol Clin North Am 2011; 49:257–265. 13. Milam M, Evins A, Sahn S. Immediate chest roentgenography following fiberoptic bronchoscopy. Chest 1989; 96:477–479. 14. Frazier W, Pope T, Findley L. Pneumothorax following transbronchial biopsy. Chest 1990; 97:539–540. 15. Sun S, Zabaneh R, Carrey Z. Incidence of pneumothorax after fiberoptic bronchoscopy (FOB) in community-based hospital: are routine postprocedure chest roentgenograms necessary? Chest 2003; 145S. 16. Izbicki G, Shitrit D, Yarmolovsky A, et al. Is routine chest radiography after transbronchial biopsy necessary? Chest 2006; 129:1561–1564.
Routine Chest Radiography for the Evaluation of Pneumothorax Following Bronchoscopy Christopher P Centonze, MD, Matthew S Davenport, MD, Eric S White, MD, MS, Ella A Kazerooni, MD, MS
Rationale and objectives: To determine the utility of routine postbronchoscopy chest radiography to detect pneumothorax. Materials and Methods: This retrospective quality improvement cohort study was approved by the Institutional Review Board. All outpatients (n = 1443) who underwent protocol-driven postbronchoscopy chest radiography in one health system from January 2010 to July 2017 were identified by electronic medical record query. The prevalence of pneumothorax (with 95% confidence intervals [CI]) and clinical outcome were determined following coded review of chest radiography reports and review of the electronic medical record. The effect of smoking and lung disease on risk of pneumothorax was determined with Chi Square tests. Results: Of 1443 subjects undergoing interventional bronchoscopy, 6% (93/1443) were current smokers, 35% (505/1442) were former smokers, and 35% (540/1443) had known lung disease. Pneumothorax prevalence was 3.4% (49/1443; 95% CI: 2.6% 4.5%) following any intervention and 4.1% (42/1032; 95% CI: 3.9% 5.5%) following transbronchial intervention. In those without known pre-existing pneumothorax or a confirmed false positive diagnosis, the real overall pneumothorax rate was 2.9% (42/1443; 95% CI: 2.1% 3.9%). The risk of pneumothorax did not differ based on smoking history (p = 0.99) or history of lung disease (p = 0.19). Of 49 subjects with pneumothorax, 13 were symptomatic, and 10 had a change in management including chest tube placement (N = 2), inpatient admission (N = 3), and/or observation (N = 7). No pneumothorax-related intervention was performed in asymptomatic patients. Conclusion: Pneumothorax following interventional outpatient bronchoscopy is uncommon, usually asymptomatic, and often clinically insignificant. Asymptomatic postbronchoscopy patients are very low risk and may not need routine imaging. Key Words: Quality; Chest radiography; Bronchoscopy; Pneumothorax. © 2019 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.
INTRODUCTION
F
lexible bronchoscopy is a common procedure that can be used for diagnostic or therapeutic purposes, with low prevalence of morbidity and mortality (1,2). The overall complication rate ranges from 0.8% 5%,
Acad Radiol 2019; 26:585 590 From the Department of Radiology, Michigan Medicine, Ann Arbor, Michigan (C.P.C., M.S.D., E.A.K.); Michigan Radiology Quality Collaborative, Ann Arbor, Michigan (C.P.C., M.S.D., E.A.K.); Department of Radiology, University of Michigan Health System, 1500 E. Medical Center Dr. B2-A209P, Ann Arbor MI, 48108 (M.S.D.); Division of Pulmonary and Critical Care Medicine, Michigan Medicine, Ann Arbor, Michigan (E.S.W.). Received October 29, 2018; revised November 14, 2018; accepted November 17, 2018. Disclosures: (Resident) Chris Centonze No relevant financial disclosures. Matt Davenport Royalties from Wolters Kluwer. Eric White No relevant financial disclosures. (AUR member, senior author) Ella Kazerooni No relevant financial disclosures. Funding: No extramural funding solicited or used for this work. Address correspondence to: M. S. D. e-mail: [email protected] © 2019 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.acra.2018.11.025
with major complications including pulmonary hemorrhage, pneumothorax, and respiratory failure (3,4). Pneumothorax is the most common major complication. The greatest risk for pneumothorax occurs following interventions during bronchoscopy, such as transbronchial biopsy, with reported prevalence of 0%-5% (5). Our historical practice has been to obtain routine postprocedure chest radiographs on all patients following interventional fiberoptic bronchoscopy to exclude pneumothorax prior to outpatient discharge; however, there are no clear data to support this practice. According to the 2013 British Thoracic Society Guideline for Diagnostic Flexible Bronchoscopy in Adults, “a chest radiograph should be obtained if a patient is symptomatic or there is a clinical suspicion of possible pneumothorax after TBLB” (transbronchial lung biopsy) and “patients should be advised of the potential for delayed complications following TBLB and provided with written information regarding likely symptoms and action required.” Both of these were given a grade D recommendation due to evidence being available from nonanalytic studies, such as case reports or case series, or from expert opinion (6). The previous American 585
CENTONZE ET AL
Thoracic Society guideline from 1987 does not mention the use of chest radiography (7). Chest radiography is a common low-risk radiology procedure (0.06 0.25 mSv for 2 views) (8) and is historically the first-line test to evaluate for suspected pneumothorax, although some data suggest thoracic ultrasound may have superior sensitivity and negative predictive value in the setting of trauma (9). However, routine chest radiography can increase length of stay, increase radiation exposure, increase cost, create incidental findings that lead to additional tests, and produce false positive results that lead to unnecessary interventions (10 12). To evaluate the utility of routine radiography following outpatient bronchoscopy, we reviewed six years of data from our health system to determine the diagnostic yield and clinical outcomes following this practice.
Academic Radiology, Vol 26, No 5, May 2019
interventional pulmonologist and depends on patient symptoms, size of the pneumothorax, and the patient's living situation. Options for management include: discharge without intervention, discharge with follow-up chest radiography, repeat chest radiography, placement of chest tube, admission to observation, and admission to inpatient status. Risk Factors for Pneumothorax
Risk factors for pneumothorax including smoking history (smoking status and, if applicable, pack-year history) and pulmonary diseases were collected by electronic medical record query of coded fields, ICD-9 codes, and ICD-10 codes. Clinical Outcomes
MATERIALS AND METHODS Study Population
The study population included all adult outpatients who underwent 1- or 2-view chest radiography within four hours following interventional bronchoscopy between 1/1/2010 and 7/1/2017. There were no exclusion criteria. All bronchoscopy and related chest radiography procedures were identified using electronic medical record query of the Current Procedural Terminology (CPT) billing codes. Bronchoscopy
Following informed consent, the patient typically receives local anesthesia of the nasal passages and oropharynx prior to administration of intravenous conscious sedation. After insertion of the flexible bronchoscope, a visual inspection of the airway is performed. Once inspection is completed, selected diagnostic or therapeutic procedures are performed based on clinical indication. When appropriate, the procedure is guided by fluoroscopy, ultrasound, or virtual bronchoscopy (utilizing a previously obtained computed tomography [CT] of the chest). These imaging modalities are used independently or in conjunction with the other imaging modalities. In other cases, no imaging assistance is utilized. At the completion of the procedure, the patient is brought to postprocedure recovery for monitoring. Postprocedure Chest Radiography
There is a standard process for evaluating patients following bronchoscopy. Once the patient successfully recovers from conscious sedation, he or she is transported from recovery to radiology where either 1- (N = 278) or 2-view (N = 1165) chest radiography is performed. Upon completion, typically the interventional pulmonology fellow and attending review the radiograph for pneumothorax. Occasionally the pulmonologists await formal interpretation by a cardiothoracic radiologist. If the chest radiograph is negative for pneumothorax and the patient is clinically stable, the patient is discharged. If the chest radiograph is positive for pneumothorax, action is determined by the 586
All final postprocedure chest radiography reports were manually reviewed by a single member of the study team (C.C.) who was blinded to the interventional technique. Reports were coded as positive or negative for pneumothorax based on the text in the report. If listed, the size of the pneumothorax (if any) was recorded. Pneumothorax interventions were determined by electronic medical record query or CPT code analysis: admission to observation or inpatient status within 24 hours of the procedure, open-guided thoracostomy placement (32551), image-guided thoracostomy placement (75989). Manual electronic medical record review was performed for all patients diagnosed with pneumothorax to determine the specific bronchoscopy intervention(s) (e.g., brushing or washing, bronchoalveolar lavage, endobronchial biopsy, transbronchial biopsy, and dilatation), postprocedure symptoms (e.g., dyspnea, cough, and chest pain), postprocedure change in vital signs (e.g., tachycardia, and hypoxia), use of follow-up chest radiography, placement of a thoracostomy tube, and admission to observation, or inpatient status. Data Analysis
The overall population and subgroups (pneumothorax and no pneumothorax) were summarized with descriptive statistics, including counts, percentages, means, and ranges. The prevalence of pneumothorax and pneumothorax intervention were calculated, and 95% confidence intervals (CI) were derived. An unpaired t-test was used to evaluate differences in age between the two subgroups. Chi-square test or Fisher's exact test was performed to determine if there were significant differences in categorical variables between the two subgroups. The risk of pneumothorax with and without transbronchial intervention was compared with Fisher's exact test. In patients with pneumothorax, characteristics of the bronchoscopy procedure and imaging findings were summarized with descriptive statistics. Institutional Review Board approval was obtained and informed consent was waived for this single-center Health Insurance Portability and Accountability Act (HIPAA)-compliant retrospective study (Table 1).
Academic Radiology, Vol 26, No 5, May 2019
TABLE 1. List of International Classification of Diseases (ICD) Codes Used in the Data Query to Establish a History of “Lung Disease” Pulmonary Disease Unspecified Chronic Obstructive Pulmonary Disease Other Emphysema Simple Chronic Bronchitis Mucopurulent chronic bronchitis Other Chronic Bronchitis Unspecified Chronic Bronchitis Emphysema Cystic Fibrosis with Pulmonary Fibrosis Unspecified Pulmonary Fibrosis Idiopathic Pulmonary Fibrosis Rheumatoid Lung Unspecified Interstitial Pulmonary Disease Respiratory Bronchiolitis Interstitial Lung Disease Unspecified Chronic Obstructive Pulmonary Disease Unspecified Emphysema Mucopurulent Chronic Bronchitis Mixed Simple and Mucopurulent Chronic Bronchitis Unspecified Chronic Bronchitis Cystic Fibrosis with Pulmonary Manifestations Unspecified Pulmonary Fibrosis Other Specified Interstitial Pulmonary Diseases Idiopathic Pulmonary Fibrosis Rheumatoid Lung Disease with Rheumatoid Arthritis of Unspecified Site Unspecified Interstitial Pulmonary Disease Respiratory Bronchiolitis Interstitial Lung Disease Lymphoid Interstitial Pneumonia Idiopathic Nonspecific Interstitial Pneumonitis Cryptogenic Organizing Pneumonia Desquamative Interstitial Pneumonia Lymphangioleiomyomatosis Adult Pulmonary Langerhans Cell Histiocytosis
ICD-9 and ICD-10 Codes 491.2, 493.2, 496 492.8 491 491.1 491.8 491.9 492 277.02 515 516.31 714.81 516.9 516.34 J44.9 J43.9 J41.1 J41.8 J42 E84.0 J84.10 J84.89 J84.112 M05.10
J84.9 J84.115 J84.2 J84.113 J84.116 J84.117 J84.81 J84.82
ROUTINE CHEST RADIOGRAPHY AFTER BRONCHOSCOPY
(N = 540) had lung disease, 6% (N = 93) were current smokers, and 35% (N = 505) were former smokers [Table 2]. Patients with and without pneumothorax had similar predefined risk factors (all p >0.05) [Table 2]. The most common intervention was transbronchial biopsy (65.7% [N = 948]) [Table 3]. The majority of patients (60.8% [N = 878]) underwent at least two interventions during the same procedure. Forty-nine patients (3.4%; 95% CI: 2.5% 4.5%) had a pneumothorax identified on post-procedure chest radiography. Of these, 42 (4.1%; 95% CI: 3.9% 5.5%) had undergone transbronchial intervention (36 transbronchial biopsy (-ies) and/or 16 transbronchial needle aspiration (s). Numbers total greater than 42 because some patients underwent more than one intervention [Table 4]. Transbronchial intervention was associated with a higher risk of pneumothorax compared to those undergoing bronchoscopy without transbronchial intervention (p = 0.03). Of the 49 patients with a postprocedure pneumothorax, 53% of procedures used fluoroscopic image guidance, 45% used ultrasound guidance, and 37% used virtual bronchoscopy. Mean pneumothorax size (reported for 45 of 49 pneumothoraces) was 17 mm (range 1 150 mm) [Table 4]. In those without a reported size (N = 4), two were described as “small” or “tiny”, one was described as decreased from preprocedure imaging, and one had no qualitative descriptor. Twenty-six of the 49 patients diagnosed with pneumothorax had at least one follow-up chest radiograph (mean: 2.3 radiographs) and in one, the pneumothorax was determined to be likely artifactual. Thirteen (27%) patients with pneumothorax had postprocedure symptoms of dyspnea, chest pain, or cough. Five of 13 symptomatic patients had hypoxia and tachycardia, seven had stable vital signs, and two had no vital signs recorded. Conversely, thirty-two (65%) patients diagnosed with pneumothorax were asymptomatic. Six of the 49 diagnosed with pneumothorax had a pre-existing pneumothorax that either improved or did not change on postprocedure radiography. Excluding the 6 subjects with a pre-existing pneumothorax and the 1 subject with a confirmed false positive resulted in a real postprocedure pneumothorax rate of 2.9% (42/1443; 95% CI: 2.1% 3.9%). Seven patients with pneumothorax were placed in observation (four with symptoms, three without symptoms, or vital sign abnormalities); none of these seven required interventions and all were discharged following repeat radiography. Three patients with pneumothorax were admitted to the hospital; all three had symptoms of pneumothorax and two (0.1% of entire cohort) required chest tube placement. No chest tubes were placed in asymptomatic patients.
RESULTS
DISCUSSION
The study cohort (Fig 1) consisted of 1443 adults (mean age: 58 years [range: 18-92 years], 683 females [mean age: 58 years, range: 18 92 years], 760 males [mean age: 58 years, range: 19 90 years]) who had chest radiography within four hours of outpatient interventional bronchoscopy. Of the subjects 35%
Pneumothorax following interventional bronchoscopy is uncommon (overall: 2.9% [95% CI: 2.1% 3.9%]). When it occurs, it is usually asymptomatic and clinically insignificant. No chest tubes were placed in asymptomatic patients, and only two chest tubes were placed (both in symptomatic 587
CENTONZE ET AL
Academic Radiology, Vol 26, No 5, May 2019
Figure 1. Study population flowchart.
TABLE 2. Study Population Details Characteristic N (%) Mean Age (years § SD) Female Male Current Smoker Former Smoker Lung Disease
Total Population
No Pneumothorax
Pneumothorax
1443 58.2 § 13.9 47% (683) 53% (760) 6% (93) 35% (505) 35% (540)
1394 (97%) 58.1 § 13.8 47% (658) 53% (736) 6% (90) 35% (488) 38% (526)
49 (3%) 61.9 § 15.3 51% (25) 49% (24) 6% (3) 35% (17) 29% (14)
patients) after 1443 procedures (0.1%). If postprocedure radiography had been restricted to symptomatic patients (6), a large fraction of radiographs could have been avoided. Based on these data, the following approach to postbronchoscopy imaging might be considered: If the patient is asymptomatic,
p 0.09 0.61 0.99 0.99 0.19
no X-ray is indicated unless there are unusual circumstances (e.g., impending air travel). If the patient is symptomatic, an X-ray should be performed. The prevalence of postbronchoscopy pneumothorax we observed was similar to that reported in other studies
TABLE 3. Prevalence of Pneumothorax Following Interventional Bronchoscopy. Procedure Types are not Mutually Exclusive Procedure (not Mutually Exclusive) Interventional Bronchoscopy Bronchoalveolar Lavage, Aspiration, or Brushing Endobronchial or Bronchial Biopsy Transbronchial Intervention Transbronchial Biopsy, Single, or Multiple Lobes Transbronchial Aspiration, Single, or Multiple Lobes Placement or Revision of Tracheobronchial Stent(S) Tracheobronchial Dilatation or Fracture Reduction Thermoplasty, Single, or Multiple Lobes
588
N
No Pneumothorax
Pneumothorax
1443 (100%) 995 (69%) 160 (11%) 1032 (72%) 948 (66%) 263 (18%) 30 (2%) 105 (7%) 0 (0%)
1394 (97%) 968 (97%) 157 (98%) 990 (96%) 912 (96%) 247 (94%) 30 (100%) 103 (98%) 0 (0%)
49 (3%) 27 (3%) 3 (2%) 42 (4%) 36 (4%) 16 (6%) 0 (0%) 2 (2%) 0 (0%)
Academic Radiology, Vol 26, No 5, May 2019
ROUTINE CHEST RADIOGRAPHY AFTER BRONCHOSCOPY
TABLE 4. Pneumothorax Management (N = 49) Stratified by Related Symptoms (Cough, Dyspnea, and Chest Pain)
All pneumothoraces (N = 49) Symptomatic (N = 13) Asymptomatic (N = 36)
Mean Size [mm] (Range)
Follow-up Radiography
Chest Tube Placement
Inpatient Admission
Unexpected Observation
17 (1 150) 31 (4 150) 11 (1 47)
26 (53%) 10 (77%) 16 (44%)
2 (4%) 2 (15%) 0 (0%)
7 (14%) 4 (31%) 3 (8%)
3 (6%) 3 (23%) 0 (0%)
(3 5). In 1989, Milam et al (13) performed a single center retrospective cohort study of 207 patients undergoing flexible bronchoscopy, including 98 with transbronchial biopsy and 14 with transbronchial needle aspiration. In that cohort, one hundred thirty had an immediate postprocedure chest radiograph, with only one pneumothorax diagnosed (patient with pleuritic chest pain). The investigators suggested that postprocedure radiography should be restricted to specific high-risk groups, which they defined as comatose, mentally obtunded, or confused patients; patients on positive pressure ventilation (due to the risk of tension pneumothorax) or with respiratory compromise (who could not tolerate a small pneumothorax); and outpatients in whom delayed radiography could be used to detect a delayed pneumothorax, if patients became symptomatic. In 1990, Frazier et al. (14) conducted a retrospective study of 305 patients who underwent routine chest radiography after fluoroscopically guided transbronchial biopsy, and found only two pneumothoraces. Both patients had chest pain and a fluoroscopic abnormality. Based on these results, the authors concluded that routine chest radiography was not necessary in the absence of symptoms or a fluoroscopic abnormality. In 2003, Wu et al. (15) evaluated 454 consecutive inpatients and outpatients undergoing flexible bronchoscopy at a community-based hospital without pulmonary medicine or critical care fellows. They found one (0.2%) pneumothorax on routine postprocedure chest radiography in a patient who was clinically suspected of having a pneumothorax. The authors concluded that routine chest radiography after flexible bronchoscopy may not be cost effective or medically necessary in all settings. In 2006 (16) and 2016 (2), Izbicki et al. performed a pair of prospective studies to evaluate the utility of routine postprocedure chest radiography following transbronchial biopsy. The first study (16) included 350 consecutive patients who underwent fluoroscopically guided flexible bronchoscopy with transbronchial biopsy and had routine chest radiography performed two hours postprocedure. Ten (2.9%) patients had radiographically detected pneumothorax, including seven symptomatic patients with clinical suspicion for pneumothorax. The other three were asymptomatic, discharged, and had spontaneous resolution of their pneumothorax on follow-up imaging. The authors concluded that routine postprocedure chest radiography was not necessary in asymptomatic patients without other risk factors (e.g., mental status changes, positive
pressure ventilation). To test this conclusion, they (2) performed a second prospective study with 201 consecutive patients undergoing fluoroscopically guided flexible bronchoscopy with transbronchial biopsy. The patients were evaluated for symptoms at two hours in post procedure observation, and at 24 hours and 48 hours by telephone interview. A chest radiograph was only obtained, if pneumothorax was clinically suspected. Eight patients (4%) were diagnosed with pneumothorax. Six were symptomatic during the two-hour observation period including two requiring chest tubes. The remaining two pneumothoraces were detected outside the two-hour observation window (one presented to the emergency room at 26 hours, one had pneumothorax detected incidentally nine days later when imaged for another reason). The authors concluded routine chest radiography after transbronchial biopsy is not necessary in asymptomatic patients who maintain a stable oxygenation level. Our series is approximately five-times larger than the previously largest study and the results support prior work, but also have some limitations. It was a singlecenter retrospective review of outpatient bronchoscopy performed in consecutive patients in a dedicated bronchoscopy unit that conducts a high volume of procedures with support of pulmonary and critical care medicine fellows. It may not apply to sites with a smaller volume of procedures and different levels of experience. Postprocedure radiography was defined as a chest radiographic examination obtained within four hours of procedure completion. It is possible that some procedures were not captured within the defined cutoff time, although this is unlikely. Data was collected by query of the institutional electronic medical record, followed by manual review of all radiology reports and pneumothorax care patterns. There is a possibility that some procedures may have been missed, and some data may have been absent from the medical record. However, we consider it unlikely that clinically important pneumothoraces would not have been recorded. In conclusion, pneumothorax following interventional outpatient bronchoscopy is uncommon and when it occurs it usually is asymptomatic and clinically insignificant. Asymptomatic postbronchoscopy patients are very low risk and may not need routine imaging. In select circumstances (e.g., known impending air travel) it might be helpful to confirm the absence of a radiographically detected pneumothorax, but in most patients it is likely unneeded. 589
CENTONZE ET AL
REFERENCES 1. 2. 3. 4.
5.
6.
7. 8.
590
Casal R, Ost D, Eapen G. Flexible bronchoscopy. Clin Chest Med 2013; 34:341–352. Izbicki G, Ayal R, Nissim A, et al. Avoiding routine chest radiography after transbronchial biopsy is safe. Respiration 2016; 92:176–181. Pue C, Pacht P. Complications of fiberoptic bronchoscopy at a university hospital. Chest 1995; 107:430–432. Smith L, Singer J, Hayes M, et al. An analysis of potential risk factors for early complications from fiberoptic bronchoscopy in lung transplant recipients. Trans Int 2011; 25:172–178. Steinfort D, Khor P, Manser R, et al. Radial probe endobronchial ultrasound for the diagnosis of peripheral lung cancer: Systematic review and meta-analysis. Eur Resp J 2010; 37:902–910. Rand I, Blaikley J, Booton R, et al. British thoracic society guideline for diagnostic flexible bronchoscopy in adults: Accredited by NICE. Thorax 2013; 68:I1–I44. Diederich S, Lenzen H. Radiation exposure associated with imaging of the chest. Cancer 2000; 89:2457–2460. Sokolowski JW, Burgher LW, Jones FL, et al. Position paper on guidelines for fiberoptic bronchoscopy in adults. Am Rev Resp Dis 1987; 136:1066.
Academic Radiology, Vol 26, No 5, May 2019
9. Nagarsheth K, Kurek S. Ultrasound detection of pneumothorax compared with chest X-ray and computed tomography scan. Am Surg 2011; 77:480–484. 10. Brivet FG, Jacobs FM, Maitre S, et al. A clinically based chest X-ray strategy reduces health care costs without adverse consequences. Intensive Care Med 2006; 32:S136. ndez-Aguado I. Incidental findings in 11. Lumbreras B, Donat L, Herna imaging diagnostic tests: A systematic review. Brit J Radiol 2010; 83:276–289. 12. Ding A, Eisenberg JD, Pandharipande PV. The economic burden of incidentally detected findings. Radiol Clin North Am 2011; 49:257–265. 13. Milam M, Evins A, Sahn S. Immediate chest roentgenography following fiberoptic bronchoscopy. Chest 1989; 96:477–479. 14. Frazier W, Pope T, Findley L. Pneumothorax following transbronchial biopsy. Chest 1990; 97:539–540. 15. Sun S, Zabaneh R, Carrey Z. Incidence of pneumothorax after fiberoptic bronchoscopy (FOB) in community-based hospital: are routine postprocedure chest roentgenograms necessary? Chest 2003; 145S. 16. Izbicki G, Shitrit D, Yarmolovsky A, et al. Is routine chest radiography after transbronchial biopsy necessary? Chest 2006; 129:1561–1564.