There has been an increased interest in reducing iatrogenic adverse effects from procedures during hospitalization. One such adverse effect is an iatrogenic pneumothorax from either a thoracentesis or thoracostomy tube placement. Studies show the iatrogenic pneumothorax rate in performing these procedures is between 6% and 9% in most patients, and up to 38% in mechanically ventilated patients (1–4). Other adverse effects include pain, infection, hemorrhage, injury to the intercostal neurovascular bundle, and tube malposition (1, 3, 5–7). Risk factors for adverse effects include both small (< 250 mL) and large effusions (> 1.5 L), mechanical ventilation, coagulopathy, blind needle insertion without ultrasound evaluation prior to the procedure, and provider inexperience (8–10). To our knowledge, ours is the first study of ultrasound guidance methodology to prevent these adverse effects in the ICU.
Ultrasound is quickly becoming the imaging modality of choice for both diagnosis of pleural effusion and to guide thoracentesis. Lung ultrasound has higher diagnostic accuracy than chest radiography in detecting pleural effusions. Furthermore, ultrasound evaluation of pleural effusions prior to thoracentesis for marking of the chest entry site has decreased pneumothorax rates from 9.3% to 4.0% in a few studies (3, 9). One of these studies in the outpatient setting showed that ultrasound-guided thoracentesis decreased the rate of iatrogenic pneumothoraces, but in conjunction with extra simulation training and limiting the number of physicians able to perform these procedures (9). Other studies in the ICU setting showed that ultrasound marking reduced the rates of hemothorax, but they did not evaluate pneumothoraces or real-time ultrasound guidance (11, 12).
The primary aim of our study was to assess whether or not using real-time ultrasound guidance while performing pleural procedures in the ICU reduced the rate of iatrogenic pneumothoraces. Secondary aims included reducing other adverse effects, such as bleeding and malpositioned thoracostomy tubes, and whether any effect seen would apply in the subgroup of mechanically ventilated patients.
MATERIALS AND METHODS
We conducted a retrospective cohort study at a single tertiary referral center, comparing real-time ultrasound guidance and ultrasound marking for performing thoracentesis or thoracostomy tube placement. The Mayo Clinic Institutional Review Board approved the study protocol 17-003073.
Between January 1, 2014, and March 31, 2017, we identified all patients (≥ 18 yr old) admitted to the ICU undergoing either thoracentesis or thoracostomy tube placement by an intensivist. The ICU was composed of medical, surgical, and transplantation (heart, lung, liver, and kidney) units. Informed consent was not necessary because this study was performed using retrospective data. Patients were excluded if there was inadequate documentation in the medical record to determine what, if any, type of ultrasound was used for the procedure. A mandatory chest radiograph following the procedure was required in our institution, allowing diagnosis of complications including all pneumothoraces. All patients in this study with a suspected pneumothorax following the procedure had a thoracostomy tube placed with subsequent chest radiography showing complete lung expansion allowing for the verification of true pneumothorax versus unexpandable lung. Indications for a diagnostic thoracentesis versus thoracostomy tube placement were determined by intensivist discretion. Therapeutic drainage of a pleural effusion was dictated by the estimation of pleural effusion size either by ultrasound and/or chest radiograph in combination with the presence of acute respiratory failure or failed weaning from mechanical ventilation.
Ultrasound-guided procedures began with maintaining the patient in the semirecumbent position followed by identification of the diaphragm and pleural effusion with the phased array probe. This probe allowed for a more thorough evaluation of the effusion, including estimation of the volume formula by Balik et al (13): Volume (mL) = 20 × pleural separation (mm). Once the effusion was adequately visualized and deemed appropriate for drainage, noting the distance to the pleural fluid and lung parenchyma, a linear probe was used for enhanced spatial resolution, allowing real-time guidance of the needle insertion through the thoracic wall planes (Video 1, Supplemental Digital Content 1, http://links.lww.com/CCM/E512; and Video 2, Supplemental Digital Content 2, http://links.lww.com/CCM/E513; legend, Supplemental Digital Content 7, http://links.lww.com/CCM/E518) (14). The procedure at this point became sterile in regard to the intensivist, patient, and ultrasound probe. The linear probe was placed craniocaudally, out-of-plane, with the indicator in the cranial position and local anesthetic administered. The needle was inserted 0.5–1 cm away from the probe with an angle of about 20–30° toward the probe, carefully controlling the needle as the intensivist visualized the needle passing through the soft tissue, superiorly to the rib, and entering the pleural space (Video 3, Supplemental Digital Content 3, http://links.lww.com/CCM/E514; and Video 4, Supplemental Digital Content 4, http://links.lww.com/CCM/E515; legend, Supplemental Digital Content 7, http://links.lww.com/CCM/E518). At this point, either fluid was removed by the insertion needle or a guidewire was inserted through the needle for placement of a thoracostomy tube, depending on whether the procedure was diagnostic or therapeutic in nature. Insertion of the guidewire into the pleural space required ultrasound visualization of the correct location—guidewire over the diaphragm and freely moving in the pleural effusion. Immediately following the procedure, the ultrasound was used to look for a post-procedure pneumothorax. Ultrasound-marked procedures were performed by a single intensivist per procedure, who initially marked the pleural effusion with the phased array probe, then immediately performed the pleural procedure. For both types of procedures, there was no patient position change once the procedure began. Furthermore, post-procedural malposition of thoracostomy tubes was ascertained by chest radiograph review by blinded radiologists who were not involved in the procedure. All 19 intensivists at this institution were previously trained on both procedure methods, and no intensivist exclusively performed one or the other procedure.
The primary outcome of our study was the frequency of iatrogenic pneumothoraces when real-time ultrasound was used for the procedure. Secondary outcomes were all complications and bleeding rates from the procedure, along with malposition of the thoracostomy tube.
JMP Version 13.0.0 (SAS Institute, Cary, NC) was used for statistical analysis. For analysis, coagulopathy was converted to a categorical variable as defined as platelets less than 50 × 109/L, international normalized ratio greater than 1.5, and activated partial thromboplastin time greater than 50 seconds. The 19 different providers were used a categorical variable in the analysis. Continuous data were analyzed using a nonparametric Wilcoxon rank-sum test. Pearson chi-square or Fisher exact test, depending on the size of analyzed variables, was used for categorical data. We performed a multiple logistic regression analysis on all variables that had a p value of less than or equal to 0.10 on the univariate analysis, which resulted in models for pneumothorax rate, overall complications, and malposition of the thoracostomy tube. Providers as a variable did not meet this criteria, but because of the potential that different providers have a large impact on complication rates, they were included in a separate multivariate analysis. The analysis for pneumothorax rate was performed on patients without a baseline pneumothorax. Adverse effect rates were analyzed by exact binomial 95% CIs. p value of less than 0.05 was considered significant.
Baseline and Procedural Characteristics
A total of 394 patients were identified as having either a thoracentesis or thoracostomy tube placed by an intensivist in the specified time period. Thirty-two patients with no adverse effects from their procedures were excluded from the study based upon inadequate documentation, leaving 362 patients meeting inclusion criteria. Patient characteristics of the real-time ultrasound guidance (n = 159) and ultrasound marking (n = 203) groups were similar at baseline (Table 1). All patients with a baseline pneumothorax had resolution of their pneumothorax post-procedure by either ultrasound or chest imaging. Although coagulopathy was a variable that could have potentially caused adverse effects, this variable was also similar between the two groups at baseline (Table 2). There were 19 different intensivists who performed the procedure with a mean time in practice of 12.3 ± 7.4 years.
The rate of pneumothoraces was significantly different between the two procedures (0.70% [95% CI, 0.12–3.9%] vs 5.01% [95% CI, 2.70–9.38%]; p = 0.03), with an odds ratio of 0.14 (95% CI, 0.02–0.88) when using ultrasound guidance (Table 3). There was a significant difference in the rate of adverse effects between real-time ultrasound-guided (0.63% [95% CI, 0.11–3.4%]) and ultrasound-marked (6.89% [95% CI, 4.15–11.24%]; p ≤ 0.01) procedures. Mechanically ventilated patients had a significant difference in overall adverse effect rates between real-time ultrasound-guided (0%) and ultrasound-marked (5.40% [95% CI, 1.50–17.70%]; p = 0.01). All pneumothoraces occurred from procedures done for therapeutic and diagnostic purposes. Despite there being a difference between what type of ICU the patient was located in and whether ultrasound guidance was used or not, there was no significance with any complications.
Ultrasound guidance was the only variable analyzed that was significant for reducing the pneumothorax rate and overall complications (p = 0.02 and < 0.01, respectively) (Supplemental Table 1, Supplemental Digital Content 5, http://links.lww.com/CCM/E516; and Supplemental Table 2, Supplemental Digital Content 6, http://links.lww.com/CCM/E517). There was incomplete data (n = 130, 36.1% of total cohort) for the amount of fluid removed as not all procedures were performed for therapeutic purposes, but there was no significance between volume of fluid removed and pneumothorax rate or overall complication rate (p = 0.90 and 0.93, respectively). When adjusting for potential confounders, ultrasound guidance remained significant in reducing pneumothoraces and overall complications (Table 4). Overall complications were only effected by ultrasound guidance and not by provider (Table 4).
When separated by type of procedure, there remained a significant difference in the rate of adverse effects for thoracenteses between real-time ultrasound guidance (0%) and ultrasound marking (12% [95% CI, 5.62–23.80%]; p = 0.02), with pneumothorax occurring in 0 patients (0%) with real-time ultrasound guidance and five patients (10% [95% CI, 4.34–21.36%]; p = 0.02) with ultrasound marking (Fig. 1). There was no significant difference in the adverse effect rate for thoracostomy tube placement between the two groups (0.89% [95% CI, 0.16–4.88%] vs 5.23% [95% CI, 2.67–9.98%]; p = 0.08).
The only adverse effect of real-time ultrasound was pneumothorax in one patient. Fourteen patients developed adverse effects after ultrasound marking: nine patients developed pneumothorax, requiring thoracostomy tube placement; two (0.99% [95% CI, 0.27–3.52%]) had immediate bleeding, controlled with absorbable hemostat fabric in one and digital pressure in the other; and three (1.48% [95% CI, 0.50–4.25%]) had a misplaced thoracostomy tube, two in the lung parenchyma and one kinked within the soft tissue. There was no mortality associated with any complications.
Our single-center retrospective study compared ultrasound guidance with standard ultrasound usage during thoracenteses and thoracostomy tube placement in the ICU. We found a notable difference between adverse effect rates, especially the rate of iatrogenic pneumothoraces. The difference was only statistically significant with thoracentesis. This decrease in adverse effect rate remained after adjusting for potential confounders.
The use of ultrasound aided pleural procedures has been around for nearly 20 years being first described as a safer way to perform thoracenteses in mechanically ventilated patients (15). Further studies continue to show that using ultrasound during these procedures is safer than not, by showing a decrease in iatrogenic pneumothoraces while increasing the yield of a thoracentesis (16–18). These studies evaluated using ultrasound to identify the procedure site and not evaluating ultrasound methodology. The decrease in adverse effect rate with ultrasound guidance is likely due to the ability to visualize the needle or thoracostomy tube entering the pleural fluid while avoiding the lung and other organs and tissues. The operator is able to change direction or stop insertion if the needle or tube is about to enter an unintended site. In addition, real-time ultrasound guidance ensures an appropriate fluid window is present at the time of insertion. Patient movement occurring between site marking and needle or tube insertion likely increases procedural risk in all patients, especially in patients with atypical positioning. Real-time guidance may be beneficial in small- to moderate-sized pleural effusions, where theoretically, there is a higher risk of adverse effects, but further studies are necessary to show this benefit.
Ultrasound-guided pleural procedures have been used in bundles to decrease the rate of iatrogenic pneumothoraces in the outpatient setting, but these were employed in conjunction with simulation-based education, with difficulty determining which intervention had the greatest impact (8). Interventional radiologists currently use ultrasound guidance as the standard at large referral hospitals in patients stable enough to go to the interventional suite (19). These studies showed that ultrasound guidance was useful for decreasing adverse effect rates in pleural procedures performed by physicians trained in bedside ultrasound; however, critically ill patients in the ICU were not included in the studies.
Prior studies showed that critically ill patients on mechanical ventilation had an iatrogenic pneumothorax rate of 0.6–38.0%, which is an increased risk (4, 9, 10, 20, 21). The wide range is most likely due to the heterogeneity of patient, provider, technique, and methodology conducted in these studies. Smaller studies of fewer than 100 patients displayed higher adverse effect rates (4, 21), whereas larger studies of more than 1,000 patients reported fewer adverse effects (10, 20). Although patients on mechanical ventilation seem to have a higher risk of iatrogenic pneumothorax, our study showed no difference in adverse effect rates with mechanically ventilated patients with the use of ultrasound guidance.
Pneumothorax ex vacuo is typically identified as ipsilateral volume loss and basilar pneumothorax along with evidence of an unexpandable lung. This condition has been reported to occur following a thoracentesis using an ultrasound-marked method by experienced physicians because it is not a result of direct lung puncture (22). There were no cases of pneumothorax ex vacuo, as evidenced by post-procedure imaging of all the patients who developed a pneumothorax following the procedure. Also, all the patients who developed a pneumothorax had a chest tube placed resulting in complete resolution of the pneumothorax.
Our study had limitations, including that it was a retrospective analysis and we could not assure the reason why some patients received the real-time ultrasound guidance method over the other. The most reasonable explanation to this particular limitation is that some providers were more interested in developing technical skills for real-time ultrasound guidance, in such a rapidly evolving field during the last 5 years. Second, our findings are only limited to a single tertiary referral center with physicians experienced in the use of ultrasound-guided procedures. The intensivists who performed these procedures at our institution used a combination of real-time ultrasound guidance and ultrasound marking, so the experience level with ultrasound guidance between intensivists is unknown, and intensivists more confident in their skills may have used real-time ultrasound guidance for more technically difficult thoracenteses. However, our faculty group mixture regarding prior experience and diverse medical background reflects most other critical care practices staffing. Third, baseline faculty training with ultrasound guidance varied as they had different critical care subspecialty backgrounds and there was not consistent simulation-based education, despite each intensivist having one initial simulation-based training session on the ultrasound-guided technique. Current positive impact of simulation-based learning could be a possible solution for such a heterogeneous group seeking proficiency in thoracentesis and thoracostomy tube placement (23). Although our study was retrospective in nature, the ultrasound-marked group acted as the control group. Because of the retrospective nature of our study, sample sizes were not equal, but patient characteristics were not noticeably different between the two groups. The adverse effect rate was also low in both groups, which resulted in lower robustness. Last, we were not able to measure a counterbalance to using real-time ultrasound guidance including additional costs associated with potential increased procedural time, sterile probe covers, and operator training.
Among critically ill adults undergoing thoracentesis or thoracostomy tube placement in the ICU, real-time ultrasound guidance was associated with decreased adverse effect rates, especially pneumothoraces. Future prospective studies are necessary to determine the relative cost-benefit ratio and to determine the best method for developing the skills for real-time guidance. Furthermore, this seems timely as the U.S. Centers for Medicare & Medicaid Services has a hospital-acquired condition reduction program that includes iatrogenic pneumothorax as a patient safety indicator 90.
We thank editing, proofreading, and reference verification were provided by Scientific Publications, Mayo Clinic, Jacksonville, FL.
1. Daniels CE, Ryu JH. Improving the safety of thoracentesis
. Curr Opin Pulm Med 2011; 17:232–236
2. de Latorre FJ, Tomasa A, Klamburg J, et al. Incidence of pneumothorax
and pneumomediastinum in patients with aspiration pneumonia requiring ventilatory support. Chest 1977; 72:141–144
3. Maslove DM, Chen BT, Wang H, et al. The diagnosis and management of pleural effusions in the ICU. J Intensive Care Med 2013; 28:24–36
4. McCartney JP, Adams JW 2nd, Hazard PB. Safety of thoracentesis
in mechanically ventilated patients. Chest 1993; 103:1920–1921
5. Alrahbi R, Easton R, Bendinelli C, et al. Intercostal catheter insertion: Are we really doing well? ANZ J Surg 2012; 82:392–394
6. Diacon AH, Brutsche MH, Solèr M. Accuracy of pleural puncture sites: A prospective comparison of clinical examination with ultrasound
. Chest 2003; 123:436–441
7. Havelock T, Teoh R, Laws D, et al. Pleural procedures and thoracic ultrasound
: British Thoracic Society Pleural Disease Guideline 2010. Thorax 2010; 65(Suppl 2):ii61–ii76
8. Duncan DR, Morgenthaler TI, Ryu JH, et al. Reducing iatrogenic risk in thoracentesis
: Establishing best practice via experiential training in a zero-risk environment. Chest 2009; 135:1315–1320
9. Gordon CE, Feller-Kopman D, Balk EM, et al. Pneumothorax
: A systematic review and meta-analysis. Arch Intern Med 2010; 170:332–339
10. Mayo PH, Goltz HR, Tafreshi M, et al. Safety of ultrasound
in patients receiving mechanical ventilation
. Chest 2004; 125:1059–1062
11. Tu CY, Hsu WH, Hsia TC, et al. The changing pathogens of complicated parapneumonic effusions or empyemas in a medical intensive care unit
. Intensive Care Med 2006; 32:570–576
12. Tu CY, Hsu WH, Hsia TC, et al. Pleural effusions in febrile medical ICU patients: Chest ultrasound
study. Chest 2004; 126:1274–1280
13. Balik M, Plasil P, Waldauf P, et al. Ultrasound
estimation of volume of pleural fluid in mechanically ventilated patients. Intensive Care Med 2006; 32:318
14. Millington SJ, Koenig S. Better with ultrasound
: Pleural procedures in critically ill patients. Chest 2018; 153:224–232
15. Lichtenstein D, Hulot JS, Rabiller A, et al. Feasibility and safety of ultrasound
in mechanically ventilated patients. Intensive Care Med 1999; 25:955–958
16. Barnes TW, Morgenthaler TI, Olson EJ, et al. Sonographically guided thoracentesis
and rate of pneumothorax
. J Clin Ultrasound
17. Grogan DR, Irwin RS, Channick R, et al. Complications associated with thoracentesis
. A prospective, randomized study comparing three different methods. Arch Intern Med 1990; 150:873–877
18. Perazzo A, Gatto P, Barlascini C, et al. Can ultrasound
guidance reduce the risk of pneumothorax
? J Bras Pneumol 2014; 40:6–12
19. Jones PW, Moyers JP, Rogers JT, et al. Ultrasound
: Is it a safer method? Chest 2003; 123:418–423
20. Ault MJ, Rosen BT, Scher J, et al. Thoracentesis
outcomes: A 12-year experience. Thorax 2015; 70:127–132
21. Britten S, Palmer SH, Snow TM. Needle thoracocentesis in tension pneumothorax
: Insufficient cannula length and potential failure. Injury 1996; 27:321–322
22. Heidecker J, Huggins JT, Sahn SA, et al. Pathophysiology of pneumothorax
. Chest 2006; 130:1173–1184
23. Barsuk JH, Cohen ER, Williams MV, et al. Simulation-based mastery learning for thoracentesis
skills improves patient outcomes: A RANDOMIZED Trial. Acad Med 2018; 93:729–735
chest tubes; intensive care unit; mechanical ventilation; pneumothorax; thoracentesis; ultrasound
Supplemental Digital Content
Copyright © 2019 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.