Secondary Logo

Journal Logo

Critical Care and Trauma: Research Report

Radiologic Assessment of Potential Sites for Needle Decompression of a Tension Pneumothorax

Wax, David B. MD; Leibowitz, Andrew B. MD

Author Information
doi: 10.1213/01.ane.0000282827.86345.ff
  • Free

Anesthesiologists must be prepared to rapidly recognize and decompress a tension pneumothorax. The recommended treatment of suspected tension pneumothorax is immediate needle decompression followed by chest tube thoracostomy. There are, however, published case reports of failed needle decompression of pneumothoraces, resulting in diagnostic dilemma and therapeutic delay (1–3). Such failure may be due to inadequate needle length or bore, catheter kinking, or improper needle siting. In addition, cases of life-threatening hemothorax have been reported secondary to vascular injury from attempted needle decompression (4). As an alternative to the long-standing practice of using the second intercostal space at the midclavicular line, needle decompression in the fourth or fifth intercostal spaces at the mid- or anterior axillary lines (MAL, AAL) (the traditional sites of tube thoracostomy) has been proposed (2,4,5). However, the safety and efficacy of using these other sites have not been well described, and their use has not become standard practice. We performed this radiological investigation to determine the optimal needle length and relative safety of each potential decompression site.


With IRB approval, we reviewed thoracic computed tomography (CT) scans of 100 adult patients in our hospital's radiology archive. The scans were selected randomly from a list of patients who had undergone anesthesia for video-assisted thoracoscopy in the prior 12 months, and were therefore likely to have an archived chest CT scan. Scans were viewed using a GE Centricity PACS workstation (GE Healthcare, Fairfield, CT).

Using soft-tissue windowing and the electronic calipers of the PACS workstation, we measured distances between various anatomic sites (Fig. 1). For each subject's scan, we first identified the scan slice at the level of the sternal angle, which typically corresponds to the second intercostal space. In this slice, we measured the horizontal distance from the midline of the sternum to the internal mammary vessels and from the midline of the sternum to the midline of the hemithorax (MHL). At the MHL, we measured the vertical distance from the skin surface to the pleura and from the skin surface to any visible major soft tissue structure (e.g., heart, liver). Next, we identified the scan slice at the level of the xiphoid process, which typically corresponds to the fifth intercostal space. At that level, the perpendicular distance from the skin surface to the pleura and skin surface to any major soft-tissue structure at the MAL was measured. Finally, the perpendicular distance from skin surface to pleura and skin surface to any major soft-tissue structure at the AAL was measured.

Figure 1.
Figure 1.:
Study measurements of thoracic anatomy on computed tomography (CT) imaging. Left image – CT scout film of thorax: MHL = midhemithoracic line (MHL) at level of sternal angle; AAL = anterior axillary line at level of xiphoid process; MAL = midaxillary line at level of xiphoid process. Center image – thoracic CT at level of sternal angle: A = distance from midsternum to MHL; B = distance from midsternum to internal mammary vessels; C = distance from skin surface to pleura at MHL; D = distance from skin surface to major intrathoracic structure at MHL. Right image – thoracic CT at level of xiphoid process: E = distance from skin surface to pleura at AAL; F = distance from skin surface to major intrathoracic structure at AAL; G = distance from skin surface to pleura at MAL; H = distance from skin surface to major intrathoracic structure at MAL.

Patient age, gender, height, and weight were also tabulated, as was the presence, or absence, of radiologic evidence of prior sternotomy. In patients with obesity or macromastia who had soft tissues extending beyond the visible scan area, or in patients with grafted mammary vessels, unobtainable measurements were necessarily omitted from the dataset. All measurements were made bilaterally and the results tabulated for analysis. Descriptive statistics were calculated, and χ2, Student's t-test, and Spearman's rank correlation were used as appropriate, with P < 0.05 considered statistically significant.


The dataset included data from scans of 100 subjects (58 male, 42 female) with a median age of 62 years (range, 23–86), median height of 168 cm (range, 142–191), median weight of 72 kg (range, 47–118), and median body mass index (BMI) of 26 (range, 16–41). There were 13 subjects with evidence of previous sternotomy and eight had grafted mammary vessels. There were seven subjects with incomplete data at the axillary sites because of body habitus.

The median distances from the midline of the sternum at the level of the sternal angle to the MHL and internal mammary vessels were 6.1 and 3.0 cm, respectively, with a 3.1 cm median gap between the two (Table 1). Median differences between the left and right sides and between genders were <1 cm.

Table 1
Table 1:
Horizontal Distance from Midline to Midhemithorax and Internal Mammary Vessels at Level of Sternal Angle

Median depth-to-pleura below the skin surface at the MHL, MAL, and AAL sites was 3.1, 3.5, and 2.6 cm, respectively (Table 2). Median differences between the left and right sides and genders were 1 cm or less. Both increasing weight ([rho] = 0.65, r2 = 0.42 at MHL; [rho] = 0.30, r2 = 0.09 at AAL; [rho] = 0.52, r2 = 0.27 at MAL) and increasing BMI ([rho] = 0.65, r2 = 0.42 at MHL; [rho] = 0.50, r2 = 0.25 at AAL; [rho] = 0.60, r2 = 0.36 at MAL) had a statistically significant correlation (P < 0.01) with increasing depth-to-pleura at all sites. Both male (versus female) gender ([rho] = −0.40, r2 = 0.16) and increasing height ([rho] = −0.20, r2 = 0.04) had a statistically significant correlation (P < 0.05) with decreasing depth-to-pleura at the AAL site, but not at other sites. There was no statistically significant correlation between age and depth-to-pleura at any site.

Table 2
Table 2:
Depth of Pleura Below Skin Surface at Thoracic Sites

Table 3 shows that there was a larger proportion of subjects with major soft-tissue structures within both 5 and 10 cm of the needle entry site and directly adjacent to the chest wall for bilateral MAL and AAL sites compared with the ipsilateral MHL site. Overall, there was less safe distance on the left side compared with the right side, and the safe distance was greatest for the MHL site and least for the AAL site on either side.

Table 3
Table 3:
Proportion of Subjects with Major Soft-Tissue Structures Within Selected Depth Ranges Below Skin Surface at Thoracic Sites


Suspected tension pneumothorax requires rapid intervention. Needle thoracentesis is both diagnostic and therapeutic; it should be performed in a manner to maximize safety and efficacy, regardless of whether the diagnosis is correct or incorrect.

Regarding needle length, one study of 54 adults using ultrasonic measurements in the second intercostal space at the midclavicular line found chest wall thickness of up to 4.4 cm in males and 5.2 cm in females, and recommended a minimum needle length of 4.5 cm to succeed in 96% of cases (6). Another study of 111 patients using CT scans showed chest wall thickness in the second intercostal space at the midclavicular line of over 5 cm in nearly one-quarter of subjects (maximum, 8.2 cm), suggesting that even a 4.5 cm catheter length would often be inadequate (7). We are not aware of any published data regarding the axillary sites. Our data suggest that a needle length of 7 cm would be adequate for nearly all patients at the MHL site, but may be too short for the axillary sites in some patients.

As might be expected, increasing weight and BMI both correlated with increasing depth-to-pleura at all sites. Females might be expected to have more subcutaneous chest wall (i.e., breast) tissue at all sites than males, but this was not found in our sample, perhaps because males tend to weigh more (which correlated with greater depth-to-pleura) than females. A multivariate regression analysis might further describe the complex relationship between height, weight, gender, and depth-to-pleura, but the clinical usefulness of this is limited, since there will always be individual variation, and because clinicians cannot be expected to perform complex calculations during rarely encountered, life-threatening emergencies that demand immediate action, such as tension pneumothorax.

Specialized needle/catheter sets specifically intended for drainage of a pneumothorax are available. There are designs that include features such as needle tip shields and whistle valves (8). In the absence of such equipment, a standard 8.9 cm 17-gauge epidural needle should, according to our and the other cited data, be sufficient to reach pleura in all patients at the MHL site. Such a needle may be used with stepwise advancement of the needle and intermittent removal of the obturator to check for gas escape. Stepwise advancement of the epidural needle allows for variability in depth to pleura among patients, and minimizes the potential danger in advancing the needle deeper than necessary. The obturator prevents clogging of the needle with tissue, the semiblunt tip may be less likely to lacerate vessels, entry into the pleural space can be felt by a “pop,” and the rigid needle resists kinking. These benefits may make an epidural needle a better choice than a standard over-the-needle IV catheter, with its sharp introducer tip, flexible material, and relatively short length, all of which have been cited as reasons for failed decompression attempts. Since hemodynamically significant tension pneumothorax would typically involve a sufficient volume of trapped gas to extend to all studied sites, blind decompression at the MHL site should be successful if needle length, bore, and patency are adequate.

Regarding needle siting, literature from anesthesiology and other specialties suggests a variety of sites and needle sizes for decompression of tension pneumothorax. A 14-gauge, 3–6 cm long or 18-gauge over-the-needle catheter in the second intercostal space at the midclavicular line, a 14-gauge IV catheter at any point in the superior, anterior, or lateral chest wall, or a large-bore needle inserted into the pleural space through the second anterior intercostal space have all been suggested (9–12). One reported study of 25 emergency medicine specialists found that 88% named the second intercostal site as the preferred site, but only 60% correctly marked the site on a volunteer, with the remainder marking the first and third interspace (13). In that study, 95% of the marked sites were located medial to the midclavicular line (with a range of 3 cm). Such errors may be due to a failure of clinicians to appreciate that the clavicle extends laterally to meet the acromion process in the shoulder region rather than stopping at the edge of the hemithorax. For this reason, we chose to evaluate a site in the midhemithorax, which is medial to the midclavicular line and a more likely spot to be used by clinicians.

Locating the appropriate level for needle insertion using easily palpated bony surface landmarks may increase consistency and save time in a crisis, compared with counting intercostal spaces. We chose to evaluate a site at the level of the sternal angle because it is easily palpated on most patients and is typically adjacent to the second intercostal space. Our data show that, in all cases, the internal mammary vessels were located within 4.4 cm (median, 3 cm) of the midline and that the MHL is at least 4.8 cm (median, 6.1 cm) from the midline. This means that the proper MHL site (at a safe distance beyond the vessels) may be quickly determined by inserting the needle 3 fingerbreadths (approximately 5–6 cm for the average clinician's hand) laterally to the midline of the sternal angle. For the axillary sites, we used the level of the xiphoid process at the sternal notch, which is also easily palpated on patients and typically corresponds to the fifth interspace.

Of the three sites studied, the MHL appears to be the safest, since it has the least likelihood of having a vital structure in the path of a needle up to 10 cm in length, especially in patients who have had a prior sternotomy. The MHL site also showed the least frequency of intrathoracic structures immediately adjacent (possibly adherent) to the chest wall. We are unaware of previous reports of safety at the axillary sites. Our data suggest that axillary sites should be used only in cases where needle decompression at the MHL site has failed (as may be possible with a loculated pneumothorax), in prone patients, or for traditional tube thoracostomy, which uses blunt dissection and should therefore be safer at any site compared with blind needle decompression.

In the presence of tension pneumothorax, intrathoracic structures are typically displaced. Although awake patients are instructed to hold their breath with lungs inflated during chest CT scans, this does not accurately simulate tension pneumothorax, and thus the distances to intrathoracic structures and our conclusions regarding safety might not apply. However, since the diagnosis of tension pneumothorax is often uncertain and may be incorrect, the MHL is still the safest site for diagnostic (and possibly therapeutic) needle decompression. Because of limitations in the ability to identify smaller intercostal and pulmonary vessels on CT scan, some potential for vessel injury may not have been identified. However, intercostal and pulmonary vessels are in proximity to all considered sites. Also, our subjects were taken from a pool of patients who may have had a greater prevalence of thoracic abnormalities, and thus may not have been representative of the population at large.

In conclusion, needle decompression of a suspected tension pneumothorax should be attempted in the MHL at the level of the sternal angle using a needle at least 7 cm long inserted perpendicular to the horizontal plane. Half of all patients should have entry into their pleural space accomplished in under 3.1 cm and the remainder within 7 cm. This approach should yield the highest success rate and margin of safety compared to axillary sites.


1. Castle N, Tagg A, Owen R. Bilateral tension pneumothorax. Resuscitation 2005;65:103–5
2. Jenkins C, Sudheer PS. Needle thoracocentesis fails to diagnose a large pneumothorax. Anaesthesia 2000;55:925–6
3. Gilligan P, Hegarty D, Hassan TB. The point of the needle. Occult pneumothorax: a review. Emerg Med J 2003;20:293–6
4. 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;20:383–4
5. Barton ED. Tension pneumothorax. Curr Opin Pulm Med 1999;5:269–74
6. Britten S, Palmer SH, Snow TM. Needle thoracocentesis in tension pneumothorax: insufficient cannula length and potential failure. Injury 1996;27:321–2
7. Givens ML, Ayotte K, Manifold C. Needle thoracostomy: implications of computed tomography chest wall thickness. Acad Emerg Med 2004;11:211–3
8. Roberts K, Steyn R, Bleetman A. New technique for treating spontaneous pneumothorax. Thorax 2004;59:355–6
9. Morgan GE, Mikhail MS, Murray MJ. Clinical anesthesiology. 4th ed. New York: McGraw-Hill, 2006
10. American College of Surgeons. Managing life-threatening thoracic injuries. 1994. Available at (Last accessed August 24, 2007)
11. Tintinalli JE, Kelen GD, Stapczynski JS, Ma OJ, Cline DM. Tintinalli's emergency medicine: a comprehensive study guide. 6th ed. New York: McGraw-Hill, 2004
12. Kasper DL, Braunwald E, Fauci AS, Hauser SL, Longo DL, Jameson JL, Isselbacher KJ. Harrison's principles of internal medicine. 16th ed. New York: McGraw-Hill, 2005
13. Ferrie EP, Collum N, McGovern S. The right place in the right space? Awareness of site for needle thoracocentesis. Emerg Med J 2005;22:788–9
© 2007 International Anesthesia Research Society