Anterior Versus Lateral Needle Decompression of Tension Pneumothorax: Comparison by Computed Tomography Chest Wall Measurement

Transcription

1 ORIGINAL RESEARCH CONTRIBUTION Anterior Versus Lateral Needle Decompression of Tension Pneumothorax: Comparison by Computed Tomography Chest Wall Measurement Leon D. Sanchez, MD, MPH, Shannon Straszewski, MD, Amina Saghir, MBBS, Atif Khan, MBBS, Erin Horn, MD, Christopher Fischer, MD, Faisal Khosa, MD, and Marc A. Camacho, MD Abstract Objectives: Recent research describes failed needle decompression in the anterior position. It has been hypothesized that a lateral approach may be more successful. The aim of this study was to identify the optimal site for needle decompression. Methods: A retrospective study was conducted of emergency department (ED) patients who underwent computed tomography (CT) of the chest as part of their evaluation for blunt trauma. A convenience sample of 159 patients was formed by reviewing consecutive scans of eligible patients. Six measurements from the skin surface to the pleural surface were made for each patient: anterior second intercostal space, lateral fourth intercostal space, and lateral fifth intercostal space on the left and right sides. Results: The distance from skin to pleura at the anterior second intercostal space averaged 46.3 mm on the right and 45.2 mm on the left. The distance at the midaxillary line in the fourth intercostal space was 63.7 mm on the right and 62.1 mm on the left. In the fifth intercostal space the distance was 53.8 mm on the right and 52.9 mm on the left. The distance of the anterior approach was statistically less when compared to both intercostal spaces (p < 0.01). Conclusions: With commonly available angiocatheters, the lateral approach is less likely to be successful than the anterior approach. The anterior approach may fail in many patients as well. Longer angiocatheters may increase the chances of decompression, but would also carry a higher risk of damage to surrounding vital structures. ACADEMIC EMERGENCY MEDICINE 2011; 18: ª 2011 by the Society for Academic Emergency Medicine Atension pneumothorax requires immediate attention, often with needle decompression. Advanced Trauma Life Support (ATLS) guidelines specify the use of a standard 50-mm 14-gauge catheter in the second intercostal space at the midclavicular line for needle decompression. Alternate sites have been proposed, such as the midaxillary line of the fourth or From the Department of Emergency Medicine (LDS, SS, EH, CF) and the Department of Radiology (AS, AK, FK, MAC), Beth Israel Deaconess Medical Center, Boston, MA. Received October 24, 2010; revisions received February 2, April 1, and May 3, 2011; accepted May 4, Presented at the American Society of Emergency Radiology annual meeting, Seattle, WA, August The authors have no relevant financial information or potential conflicts of interest to disclose. Supervising Editor: Scott Wilber, MD. Address for correspondence and reprints: Leon D. Sanchez, MD, MPH; lsanche1@bidmc.harvard.edu. fifth intercostal space, these being the usual sites of chest tube insertion. 1,2 Recent research describes failed needle decompression in the anterior position due to inadequate catheter length. 3,4 It has been hypothesized that a lateral approach may be more successful as there may be a shorter distance from the chest wall to the pleural space. 2,5 The aim of this study was to identify the optimal site for needle decompression using computed tomography (CT) images to assess chest wall thickness at both the anterior and the lateral sites. METHODS Study Design This study was a retrospective chart review approved by the institutional review board. Study Setting and Population The study site was Beth Israel Deaconess, an academic medical center that serves as a tertiary referral center ISSN ª 2011 by the Society for Academic Emergency Medicine 1022 PII ISSN doi: /j x

2 ACADEMIC EMERGENCY MEDICINE October 2011, Vol. 18, No and Level I trauma center in Boston, Massachusetts. The emergency department (ED) has approximately 55,000 visits per year. Study Protocol Emergency department records were searched for patients who underwent CT of the chest as part of the initial evaluation for a blunt trauma presentation from June 2008 through February Chest scanning was performed on a General Electric Volume Zoom CT scanner (GE Healthcare, Worldwide, Little Chalfont, UK), which is a 64-detector scanner with minimal detector width of mm. Initial coronal and sagittal whole-body scout topograms were acquired, principally for setting the protocol and establishing the start and length of scanning. Raw axial slices were reconstructed in 5-mm-thick axial, coronal, and sagittal multiplanar reformatted images. Eligible scans were reviewed and measurements were made on the GE Centricity Picture Archive Communication System (PACS; GE Healthcare Integrated IT Solutions, Barrington, IL). Outcome Measures A total of six measurements from the skin surface to the pleural surface were made for each eligible patient, three on the left and three on the right: anterior second intercostal space (A2R, A2L), lateral fourth intercostal space (L4R, L4L), and lateral fifth intercostal space (L5R, L5L). In the anterior chest wall, measurements were made on each of the right and left sides for each patient. Utilizing the coronal scout topogram, a line was drawn along each clavicle (example along the right shown in Figure 1A). This line was bisected and a true vertical dropped into the hemithorax estimating the midclavicular line, mimicking the clinical determination. The corresponding sagittal slice was then identified via cross reference. On the sagittal slice, measurements were made on each side along a line drawn perpendicular to the skin surface at the respective right or left second intercostal space (A2R, A2L; Figures 1A and 1B). In the lateral chest wall, measurements from the skin surface to the pleura were made at the fourth intercostal space and again at the fifth intercostal space on each side. All measurements were made on the respective axial slice identified by visual inspection, counting ribs and intercostal spaces while scrolling through data sets. Measurements were made in a line parallel to the CT gantry table from the skin surface to the pleura (Figure 1C). A total of four lateral measurements were made on each patient: L4R, L4L, L5R, and L5R. A measurement along the same course was made at each lateral site from the skin surface to the closest underlying vital structure, which was defined as any large mediastinal structure. An attending emergency radiologist with 6 years experience defined the measurements, made a portion of the initial measurements, and trained others to make the remaining measurements. The subsequent measurements were made by an emergency radiology clinical fellow (with 2+ years experience), and a radiology research assistant (supervised by the fellow and attending). (A) (B) (C) Figure 1. (A) Coronal Scout topogram from chest CT. Line drawn along obliquely oriented clavicle (due to arms up positioning) was bisected and a true vertical line dropped through the respective hemithorax to approximate the midclavicular line. This process was repeated for the contralateral side for each patient. (B) Sagittal reconstructed image from chest CT corresponding via cross reference to the vertical line determined in A. On this slice, a measurement along a line (A2R) perpendicular to and originating from the skin surface to pleural surface through the second intercostal space was made. This process was repeated for the contralateral side in each patient. (C) Axial (transverse) slice from chest CT. Axial (transverse) slice from chest CT selected by visual inspection at level where a true horizontal line (parallel to the floor ) would enter thorax through fourth intercostal space. A measurement was then made along this line (L4R). This process was then repeated for the contralateral side and both measurements were again repeated at the fifth intercostal space.

3 1024 Sanchez et al. ANTERIOR VS. LATERAL NEEDLE DECOMPRESSION OF TENSION PNEUMOTHORAX We defined a failure as a patient whose skin to pleura distance exceeded 50 mm because the standard 50-mm catheter would therefore be inadequate to penetrate the pleura. This is the needle size recommended in ATLS guidelines and commonly stocked in EDs. Data Analysis An initial power calculation indicated 198 paired measurements would be needed to demonstrate a mean difference of 0.5 cm (10% of the length of the catheter), at an alpha of 0.05 and 80% power. The 5-mm difference was chosen because it was felt to be a clinically significant difference and it is the smallest difference in catheter lengths commercially available. An initial sample of 198 consecutive patients was formed by reviewing CT scans of eligible patients during the study period. Scans that did not include the skin surface at measurement levels, or where the clavicles were not completely seen, were excluded, as the measurements could not be made. Scans with extensive subcutaneous emphysema or soft tissue injury were also excluded, since anatomic landmarks could be disrupted and measurements may be less reliable. After the initial 198 scans were obtained, patients were also excluded if they were scanned with their arms positioned at their sides, as it is not standard CT scan protocol. Measurements were entered into a spreadsheet created and edited within Microsoft Excel (Microsoft Corp., Redmond, WA). Data were then imported into JMP 9.0 (SAS Institute, Cary, NC) for analysis. When comparing means between two groups (i.e., males vs females), unpaired t-tests were used. When comparing means of values obtained from the same patient (i.e., comparing different approaches in the same subject), paired t-tests were used, since measurements at each location on either side of the same subject are not independent. When comparing means of values obtained from different locations (i.e., comparing anterior vs. lateral approaches among subjects), unpaired t-tests were used with the Bonferroni correction for multiple comparisons with p < considered significant. Fisher s exact test was used for comparison of proportions. The Shapiro-Wilk test for normality with p > 0.05 was used when comparing distributions. RESULTS A total of 225 consecutive scans were reviewed. Of these, 66 scans were excluded (27 because of incomplete visualization of the clavicles or subcutaneous emphysema, 39 because of arm positioning). A total of 159 scans were analyzed. The majority of the scans were on male patients (102 of 159, 64%). The median age of the patients was 41 years for males (interquartile range [IQR] = 24 to 56 years) and 40 years for females (IQR = 25 to 66 years). Table 1A summarizes the mean skin-to-pleura distances. The distances at each location were normally distributed. There was no significant difference in mean difference between the left and right side for any of the measurements within each subject (p > 0.05 with paired t-tests). The distance on the anterior approach was significantly shorter when compared to both the fourth and the fifth intercostal spaces on each side (p < for all comparisons using unpaired t-tests with the Bonferroni correction; Table 1B). Overall, the failure rate at all sites was 54.2% (95% confidence interval [CI] = 51.0% to 57.3%). Failure rates at each site are shown in Table 2. The difference of the failure rates at the three sites were statistically significant (p < 0.05). There was no significant difference between failure rates on either side at each location (p > 0.05 for each comparison). The mean distance to a vital structure via the lateral approach is shown in Table 3. Table 1B Mean Distance (mm) From Skin to Pleura at the Anterior Second, Lateral Fourth, and Lateral Fifth ICS on Each Side Anterior second ICS Lateral fourth ICS Lateral fifth ICS Left Mean Distance 95% CI Mean Distance Right 95% CI p < for all comparisons (anterior second vs. lateral fourth, anterior second vs lateral fifth, and lateral fourth vs lateral fifth) as unpaired t-tests with the Bonferroni correction for multiple comparisons. Table 1A Mean Distance (mm) from Skin to Pleura at the Anterior Second, Lateral Fourth, and Lateral Fifth ICS Mean Distance (Left) Mean Distance (Right) Overall Male Female Overall Male Female Mean Difference Overall p-value (Paired t-test) Anterior second ICS ) Lateral fourth ICS ) Lateral fifth ICS ) Table depicts overall mean for both males and females combined, as well as separated by sex.

4 ACADEMIC EMERGENCY MEDICINE October 2011, Vol. 18, No Table 2 Percentage of Patients for Which a 50-mm Angiocatheter Would Fail to Enter Pleura at the Anterior Second, Lateral Fourth, and Lateral Fifth ICS Based on CT-guided Measurement of Skin-to- Pleura Chest Wall Thickness DISCUSSION Overall Failure Rate (95% CI) Failure Rate Left Right p-value Anterior second 33.6* ( ) ICS Lateral fourth ICS 73.6* ( ) Lateral fifth ICS 55.3* ( ) p-values for comparison of proportions by Fisher s exact test comparing left vs right. *p-value comparing overall failure rate at 2nd ICS to 4th and 5th ICS <0.05. Table 3 Mean Distance to Closest Vital Structure Left Right Lateral fourth ICS (±22.9) (±22.8) Lateral fifth ICS (±20.9) (±22.8) Data are reported as mean (±SD). Mean Distance to Closest Vital Structure (mm) The focus of this study was to identify the optimum site for needle decompression using CT images to determine the distance from skin to pleura in trauma patients. Our patient population appears to be fairly representative of the overall blunt trauma population, with a 2.1:1 male-to-female predominance, similar to the male:female ratio of 2.2 reported by the National Vital Statistics Report of death rates secondary to unintentional trauma, 6 and the 2.6:1 male:female ratio reported in another study of needle thoracostomy. 7 Our results with regards to the failure rate secondary to inadequate catheter length are similar to those reported by Givens et al. 7 At the anterior second intercostal space, our study population would have had an approximately 30% to 36% failure rate, slightly higher than the 22% failure rate reported by Givens and colleagues. 7 The rates of failure are much greater at the lateral sites, ranging up to two-thirds of the patients, which suggest that for our study population, the anterior approach is the better site when considering needle decompression in an emergency setting (Table 2). Although a study by Wax and Leibowitz 8 did not report failure rates, all reported median depth-to-pleura distances were less than 50 mm, suggesting potentially successful decompression at the studied sites: midhemithoracic line (roughly equivalent to our anterior site), midaxillary line (same as our lateral sites), and anterior axillary line. The patient population in this study used CT scans from radiology records on patients undergoing anesthesia for video-assisted thorascopy, which may be different than the patient population in the ED who are undergoing CT scans for trauma. Another study by McLean et al. 9 suggests that a 5-cm catheter would be sufficient to penetrate the pleura at any site. These data were obtained using ultrasound measurements. The pressure placed on the ultrasound probe can displace tissue and could account for the shorter overall measurements. The patient population was composed of recruited medical staff and medical students, which may not be representative of trauma patients in an ED. When considering sex, our study showed that, at every site, women had shorter distances from skin to pleura when compared to men. Contradictory to our results, the data by Givens et al. 7 showed that on average, women had thicker chest walls. This study was conducted in a military trauma center. It is possible that the military population is different enough to account for our divergent findings. Neither study included body mass index in the data set, so we are unable to compare the populations more closely. Longer catheters have been suggested to increase the likelihood of successful needle decompression. 7,8 This is a matter of debate secondary to the argument that the use of longer catheters may lead to more injuries. The study by Rawlins et al. 2 suggested that the lateral approach has the advantage of avoiding the anterior vascular structures such as the internal mammary artery and the subclavian vessels. The lateral approach is still at risk of damaging vital structures. A 70-mm catheter would be needed to assure a 95% success rate of reaching the pleura in our patient population. The heart, diaphragm, liver, or a major vessel could be punctured with this length of needle, as would be the case for 5% of our patient population. We concur with Wax and Leibowitz 8 that the anterior approach is not only the most successful site, but also the safest in regard to distance of vital structures. LIMITATIONS Patients with significant subcutaneous emphysema and tissue destruction were excluded from the study, and while this would likely distort the tissue measurements, it is also likely excludes the study populations in which needle decompression would be useful. A tension pneumothorax often displaces intrathoracic structures, and therefore, the distances to these structures may be under- or overestimated in our data. This was a singlesite study, and therefore our study population may not be generalizable to other institutions. For the comparison of the distances among all of the locations on each side (Table 1B), the distances were compared with multiple t-tests with the Bonferroni correction for multiple comparisons. Since the distances are not independent (there are multiple measurements from each individual subject), we believe that this approach is the most conservative way to maintain the overall familywise error rate of <0.05 when accounting for multiple comparisons. Although our final sample was less than our a

5 1026 Sanchez et al. ANTERIOR VS. LATERAL NEEDLE DECOMPRESSION OF TENSION PNEUMOTHORAX priori determined sample size because of exclusions of some subjects due to the imaging techniques used, we believe that the robustness of the results still supports our conclusions. CONCLUSIONS Comparison of the chest wall thickness within individuals suggests that with commonly available angiocatheters, the lateral approach is less likely to be successful than the anterior approach. The anterior approach may fail in many patients as well. Longer angiocatheters may increase the chances of decompression but would also carry a higher risk of damage to surrounding vital structures. References 1. American College of Surgeons. Student Course Manual. ATLS Advanced Trauma Life Support Program for Doctors. 7th ed. Chicago, IL: American College of Surgeons, Rawlins R, Brown KM, Carr CS, Cameron CR. Life threatening haemorrhage after anterior needle aspiration of pneumothoraces. A role for lateral needle aspiration in emergency decompression of spontaneous pneumothorax. Emerg Med J. 2003; 20: Zengerink I, Brink PR, Laupland KB, Raber EL, Zygun D, Kortbeek JB. Needle thoracostomy in the treatment of a tension pneumothorax in trauma patients: what size needle? J Trauma. 2008; 64: Stevens RL, Rochester AA, Busko J, et al. Needle thoracostomy for tension pneumothorax: failure predicted by chest computed tomography. Prehosp Emerg Care. 2009; 13: Ball CG, Wyrzykowski AD, Kirpatrick AW, et al. Thoracic needle decompression for tension pneumothorax: clinical correlation with catheter length. Can J Surg. 2010; 53: Heron M, Hoyert DL, Murphy SL, Xu J, Kochanek KD, Tejada-Vera B. Deaths: final data for National Vital Statistics Report. 2009; 57: Givens ML, Ayotte K, Manifold C. Needle thoracostomy: implications of computed tomography chest wall thickness. Acad Emerg Med. 2004; 11: Wax DB, Leibowitz AB. Radiologic assessment of potential sites for needle decompression of a tension pneumothorax. Anesth Analg. 2007; 105: McLean AR, Richards ME, Crandall CS, Marinaro JL. Ultrasound determination of chest wall thickness: implications for needle thoracostomy. Am J Emerg Med [Epub ahead of print].