a b s t r a c t

Introduction: The purposes of this study were to measure the chest wall thicknesses (CWTs) at second intercostal space (ICS) mid-clavicular line (MCL) and fifth ICS MAL directly, and compare the actual success rates of needle thoracostomies (NTs) by inserting a 5-cm-long syringe needle. Predictive values of weight, body mass index (BMI) and CWT were also analyzed.

Materials and Methods: This study included 199 measurements of 50 adult fresh cadavers from both hemithoraces. Five-centimeter-long syringe needles were inserted and secured. Penetration into the pleural cavity was assessed, and CWTs at 4 locations were measured. Achieved power of this study for the primary aim of CWT comparison from 2^nd and 5^th ICSs was .94.

Results: Overall mean CWTs at 2^nd ICS MCL and 5^th ICS MAL were measured as 2.46 +- 0.78 and 2.89 +- 1.09, re- spectively, and 5^th ICS MAL was found to be statistically thicker (P = .002). The success rate of NT at 2^nd ICS MCL was 87% (95% CI, 80-94), and that at 5^th ICS MAL was 78% (95% CI, 70-86; P = .3570). Only 6 (17.1%) of 35 failed NTs had a CWT greater than 5-cm. Needle thoracostomy has failed in 29 (14.9%) of 194 locations, despite a CWT less than 5-cm. Below a weight of 72 kg, BMI of 23 kg/m², or CWT of 2.4 cm, all NTs were successful.

Discussion and Conclusions: In this report, we present the largest cadaver-based cohort to date to the best of our knowledge, and we observed a statistically nonsignificant 9% more NT success rate at 2^nd ICS at a power of 88% and statistically significant more success rate in males at 5^th ICS was (47.7%). We also observed thinner CWTs and higher success rates than previous imaging-based studies. A BMI of 23 kg/m² or less and weight of 72 kg or less seem to accurately rule-out NT failure in cadavers, and they seem to be better predictors at the bedside.

(C) 2016

Introduction

Needle thoracostomy (NT) is a lifesaving procedure for converting a Tension pneumothorax into a simple one [1,2]. Advanced Trauma Life Support (ninth edition) guideline suggests, “inserting a large-caliber needle into the second ICS in the MCL of the affected hemithorax” for

? Meetings: This manuscript was presented as an oral abstract presentation at the 10th Turkish Emergency Medicine Congress; on October 10, 2015; at Girne, Cyprus.

* Corresponding author at: Marmara Universitesi Pendik Egitim ve Arastirma Hastanesi, Acil Tip Anabilim Dali, Fevzi Cakmak Mah, Muhsin Yazicioglu Cad, No: 10 Ust Kaynarca, Pendik, Istanbul, Turkey. Tel.: +90 216 625 45 45-70 20.

E-mail addresses: [email protected] (C. Ozen), [email protected], [email protected] (H. Akoglu), [email protected] (R.O. Ozdemirel), [email protected] (E. Omeroglu), [email protected] (C.U. Ozpolat), [email protected] (O. Onur), [email protected] (Y. Buyuk), [email protected] (A. Denizbasi).

¹ Present Address: Department of Emergency Medicine, Umraniye Education and Re- search Hospital, Istanbul, Turkey.

the emergency management of tension pneumothorax [3]. Needle thoracostomy success rates vary by location, and its use has also been questioned in the literature [4-6]. Various needle lengths and locations were proposed and fifth intercostal space (ICS) emerged as the most popular alternative. Because it is unethical and nearly impossible to perform a randomized controlled trial comparing the success and complication rates at second ICS mid-clavicular line (MCL) to fifth ICS MAL in TPX patients, secondary ways were sought, and imaging-based chest wall thickness (CWT) measurement has emerged [1,5,7-18]. In those imaging-based studies, CWTs of the locations were compared with the length of typical needles (4.5, 5, and 6 cm) to estimate the success rate [9,19]. Some researchers also evaluated the effect of obesity and body mass index (BMI) and found that BMI is directly but negatively correlated with CWT [1,5,9,12,13].

The CWT measurement technique and assumptions of the imaging- based approach have some drawbacks. In all of these trials, measurement is indirect. Techniques are variable; an actual needle is not inserted; elasticity and compressibility of the tissues are ignored; and a CWT

http://dx.doi.org/10.1016/j.ajem.2016.08.038

0735-6757/(C) 2016

thinner than a specific length (4.5, 5, and 6 cm) is assumed to mean a successful NT. To the best of our knowledge, in only one trial, cadavers were used instead of imaging; CWT was measured directly, and actual needles were inserted to observe the success rate [10]. Furthermore, to the best of our knowledge, all studies to indicate a success rate according to BMI are CT-based studies, and they report estimated success rates according to needle length-CWT comparison.

The purposes of this study were to measure the CWT at 2 most popular locations directly on fresh cadavers, and compare both CWTs and success rates at each location after inserting 5-cm-long needles. Second, we aimed to determine the cutoff values for weight, CWT, and BMI to rule out NT failure. Study was approved by the Ethical Board of the Council of Forensic Medicine of Ministry of Justice (Approval: 21589509/633).

Methods

Study design and setting

This study was completed at the Mortuary Department of Council of Forensic Medicine of Ministry of Justice in Istanbul, Turkey, one of the biggest autopsy laboratories in the world with more than 5000 autop- sies for each year. This investigation was a nonblinded, comparative, ca- daver study, which was conducted between September 2014 and January 2015 with 50 adult fresh cadavers.

Selection of participants

All consecutive cadavers with an intact torso that were approved by The Council authority whom the needle insertion would not change the legal outcomes of the autopsy report were recruited. Cadavers with prominent evidence of neoplastic, congenital, and developmental malformations, with significant tissue loss due to trauma or with any kind of needle marks on the chest were excluded.

We kept the cadavers in refrigerated storage at +4?C until 1 hour be- fore the procedure, and then allowed them to warm to room tempera- ture. Then, cadavers were placed on the autopsy tables which have weighting scale properties in the supine position with arms at 90?. Heights were measured by the special tapes on the tables.

Interventions

Ordinary 6.5-cm syringe tips with 5-cm-long needles were inserted to the second ICS MCL and fifth ICS MAL on each hemithorax just above the ribs as specified by the Advanced Trauma Life Support guidelines and secured using standard medical adhesive tapes.

Methods and measurements

Then, the CWTs of 4 locations on each cadaver were measured (second and fifth ICS on the right and left hemithorax), and penetration to the pleural cavity was assessed after midline thoracotomy. If the needles were dislodged, we used pleural puncture as evidence of penetration.

Two emergency medicine and 2 forensic medicine physicians (FMPs) completed all measurements. Before the study has begun, FMPs participated in a hands-on training about the needle insertion technique and measurement methods. intraclass correlation coefficient was calculated for the first 40 measurements of the initial 10 cadavers as

.91. Remaining 159 measurements in 40 cadavers were made by FMPs. Body mass index was calculated using a standard formula (weight (kg)/[height (m)]²) and grouped according to World Health

Organization classification [20].

Analysis

In this study, the maximum type I error was 0.05, and the level of sig- nificance was accepted as P b .05. Continuous variables, frequencies, their differences, odds ratios (ORs), and Relative risks (RRs) were all re- ported with 95% confidence intervals (CIs). Correlations were calculated as defined by Pearson and their effect sizes were named according to Dancey and Reidy [21]. threshold values of criterion values at the fixed sensitivity of 100% were selected to rule-out NT failure. CWTs were transformed to categories according to commonly reported needle lengths of 4.5, 5, and 6 cm, and achieved success rates within those thresholds were also reported. BMI, CWT, and sex were selected for the logistic regression model because of their clinical importance.

We planned an equivalence study of the CWT at second and fifth ICS with an ? value of .05, ? value of .10, and equivalence limit to .5 cm and estimated the SD of CWT as 1.0 cm. If there was no difference between both sites, 87 measurements per ICS would be required to be 90% sure that the limits of a 2-sided 90% CI will exclude a difference in means of more than .5 cm. We, therefore, planned to recruit 50 cadavers for 100 measurements for each location (on both hemithorax). The achieved power of this study for the comparison of CWT between second and fifth ICS was .88 (difference, .42 cm; 95% CI, .16-.69; Gpower v3.1.9.2, Franz Faul, Universitat Kiel, Germany). All other analyses were per- formed using MedCalc v15.8 (MedCalc Software bvba, Oostende, Belgium).

Results

All cadavers were adults according to growth plate evaluations, but age could not be determined in 31 (62%) of the cadavers. Mean age of the remaining 19 (38%) cadavers was 45.37 +- 18.29 years (95% CI,

36.55-54.18). Forty-one (82%) of the cadavers were male and 9 (18%) were female. Mean weight and height were 78.83 +- 22.84 kg (95% CI, 71.62-86.04) and 174.49 +- 7.04 cm (95% CI, 172.26-176.71) for male,

73.78 +- 18.31 kg (95% CI, 59.70-87.85) and 159.33 +- 6.69 cm (95% CI,

154.19-164.48) for female cadavers, respectively. Calculated mean BMI was 25.91 +- 7.41 kg/m² (95% CI, 23.57-28.45) for males and

29.01 +- 6.87 kg/m² (95% CI, 23.74-34.29) for females, which pointed to a slightly overweight cadaver cohort (BMIN 25 kg/m²).

We observed no difference between the success rates of second and fifth ICSs in males (89.0% [n = 73/82] vs 86.6% [n = 71/82]; P = .8120), but success rate was higher at second ICS in females (77.8% [n = 14/18] vs 38.9% [n = 7/18]; P = .0409). Fifth ICS was statistically significantly thicker in both sex compared with second ICS, and thicker in females compared with males on both locations (Figs. 1 and 2). Conversely,

Fig. 1. Boxplot distribution graphics of the CWTs at second ICS according to sex.

mean CWTs of the second ICS were similar in both sexes (Fig. 1). Being female and obese were statistically significant risk factors for NT failure with RRs of 3.42 and 12.67, respectively. All those main outcomes are presented in Table 1 in detail.

We compared the proportion of cadavers in each World Health Organization BMI class with the reference values for the geographic region and found that our sample was a good representative of the regional population [22]. In normal BMI class, second and fifth ICS CWTs were comparable regardless of sex. In male cadavers, 96.5% of the failed NTs were observed in preobese or heavier cadavers. In fe- males, all of the failed NTs were observed in obese cadavers.

Chest wall thickness, BMI, and weight were the most accurate variables for predicting an NT failure with areas under curve (AUCs) of 0.92, 0.89, and 0.81, respectively (Table 2). Pairwise comparison of Receiver operating curves, as described by DeLong et al [23], revealed that CWT (P = .0023) and BMI (P = .0074) were statistically more precise than weight, but indifferent from each other (P = .28).

No failed NTs were observed in cadavers with a weight less than 72 kg, BMI less than 23 kg/m², and CWT less than 2.4 cm (Table 2). In this study, no NT was successful if CWT was 4.3 cm or thicker. On the contrary, nearly half (20/34; 58.8%) of the failed NTs were observed at locations thinner than 4.3 cm. Of the 34 failed NTs, 14 (41.2%), 10

(29.4%), and 5 (14.7%) of them had thicker CWTs than 4.3, 4.5, and 5 cm, respectively. On the contrary, NT has failed in 20 (58.8%) of 185, 26 (13.6%) of 191, and 29 (14.9%) of 194 sites despite adequate length, at the same order. If we have used 4.3, 4.5, or 5 cm as thresholds for success, we would estimate only 14 (41.2%), 10 (29.4%), and 5 (14.7%) of the 34 actually failed NTs as failed, respectively.

A logistic regression model was tested to ascertain the effects of BMI (obese or not), CWT (cm), and sex (male or female), and found to be statistically significant (?² = 102.198, P b .0001). The model explained 67% (Nagelkerke R²) of the variance of the success rate and correctly classified 91.0% of cases. Analysis showed that females had 4.46 times higher odds for an NT failure (95% CI, 1.24-15.99; P = .0219). Also, increasing CWT was strongly (OR, 9.82; 95% CI, 4.21-22.88; P b .0001), and increasing BMI was slightly (OR, 1.10; 95% CI 1.00-1.12; P = .0485) associated with a reduced likelihood of NT success.

3

Discussion

First, overall success rates of NT at second ICS MCL and fifth ICS MAL were 87% and 78%, respectively, without a statistically significant differ- ence. Mean CWT of failed NTs was significantly higher at fifth ICS in all cadavers, Males and females. Second, all NTs were successful in cadavers with a BMI of 23 kg/m² or less. Of the failed NTs, 96.5% of males were at least preobese, and 68.2% of males and 100% of females were at least obese.

Fig. 2. Boxplot distribution graphics of the CWTs at fifth ICS according to sex.

The NT success rate that we have achieved at second ICS MCL (87%) was considerably higher compared with the previous but smaller cadaver-based study (n = 20, 23/40 [57.5%]) which is the lowest reported rates to date [5,10]. This rate is lower compared with other published studies (66.4%-94.3%) except 3 trials [8,11,13,15,17]. A recent meta-analysis, which also includes our prior CT-based trial [17], con- cluded that an optimally placed 5-cm-long catheter would have enough length to decompress 72% of the cases [24]. Therefore, we observed a higher (but not the highest) success rate compared with previous cadaver and imaging-based studies.

We think that this difference was mainly originated from the mea- surement bias in imaging-based studies, which overestimates the CWT because of the following reasons: (1) CWT might be measured perpendicular to CT gantry because it is not practical to measure CWT perpendicular to skin; (2) needle thoracostomy is performed from the anatomically thinnest point; however, CT is not sensitive enough to dis- criminate this particular point; (3) CT-based CWT measurements ignore the elasticity and compressibility of the tissues; and (4) all imaging- based studies assume that if the CWT is thinner than a specific needle length (4.3, 4.5, 5, and 6 cm, according to previous studies), intervention would be successful. However, if we have used 4.3, 4.5, or 5 cm as thresholds for success, we would correctly estimate only 14 (41.2%), 10 (29.4%), and 5 (14.7%) of the 34 actually failed NTs. Our results clearly show that estimating success just because the CWT is thinner than a particular length is a false assumption.

To the best of our knowledge, this is the first study to report the threshold values of weight, BMI, and CWT to rule-out an NT failure (Table 2). According to our results, physicians may be confident that a failure is unlikely in a male patient with normal stature and weight if the procedure was done correctly. However, preobese and obese patients carry the risk of NT failure despite adequate CWT.

Limitations

This study has some limitations: first, postmortem weight may not represent the actual weight. Second, despite the selection of only fresh cadavers, water content, stability, and tension of peripheral pleura do change postmortem and may have resisted against needle penetration. Third, cadavers of traumatic deaths with prominent thoracal tissue loss were excluded; however, those were the exact population NT may have been used. Fourth, this study was conducted in a department which serves to a population of 17 million. Therefore, the busy schedule and enormous number of daily examinations may have limited the consec- utive recruitment of all eligible cadavers. To test the effect of this sam- pling problem, we compared the BMI distribution of our sample with regional metrics and found that our cadaver sample is compatible with the population. Fifth, because CWTs were measured after thoracot- omy, the original tension and elasticity of the skin, muscles, and fascias may have been lost, and thus, the measurements may not have been ac- curate. On the contrary, while measuring the CWT with hand-held scissor-type calipers, pressure exerted by researchers may have reduced the measured CWT. Therefore, we do not know the magnitude and di- rection of measurement bias if there is any. Sixth, during thoracotomy, the vibration, and movement of the chest wall might cause the misplacement of some needles. To avoid this, we extensively searched the pleura for any puncture marks (wounds) when needles were dislodged. Last, the nonblinded nature of the study may have affected the reported success rate in cadavers with dislodged needles. Still, we think that these limitations are acceptable as compared with probable measurement bias lies in the technique used in imaging- based studies.

Conclusion

In summary, cadaver-based research creates a different perspective to determine the success rate at various locations. Both imaging-based

Table 1

Primary outcomes of the study according to cadaver and intervention variables

Success rate CWT

Subgroups n

			n (%)	RR, 95% CI, P	Difference (%), 95% CI, Pa		CWT (cm), mean +- SD, 95% CI	Difference (%), 95% CI, Pb
Overall		200	165 (82.5)	-			2.68 +- 0.96, 2.54-2.81	-
Location	Second	100	87 (87.0)	1.69, 0.90-3.17, .1001	9.0, -3.4-21.5, .1864		2.47 +- 0.77, 2.31-2.62	0.42, 0.16-0.69, .0018c
	Fifth	100	78 (78.0)				2.89 +- 1.09, 2.67-3.11d

Sex

3.42, 1.94-6.01, .0001 29.50, 7.71-53.23, .0017 2.62 +- 0.90, 2.48-2.76

2.95 +- 1.18, 2.54-3.35

b e

Male	164	144 (87.8)
Female	36	21 (58.3)

0.33, -0.10 to 0.76, .1250^c

Side	Right	100	82 (82.0)	0.94, 0.52-1.73, .8524	1.0, -11.4 to 13.4, .9714	2.66 +- 0.94, 2.47-2.85	0.03, -0.24 to 0.30, .8015
	Left	100	83 (83.0)			2.69 +- 0.99, 2.50-2.89d
BMI	Preobese or	152	145 (95.4)	12.67, 5.91-27.14, b.0001	53.70,	2.40 +- 0.76, 2.27-2.52	1.19, 0.87-1.51, b.0001c
grouped	leaner				29.41-75.15, b.0001
	Overweight	48	20 (41.7)			3.59 +- 1.00, 3.29-3.88f
	(obese)
cMI	Underweight	12	12 (100.0)	- -		1.38 +- 0.29, 1.19-1.56	-
Class	Normal	76	75 (98.7)	- -		2.13 +- 0.54, 2.00-2.25	-
	Overweight	64	58 (90.6)	- -		2.91 +- 0.68, 2.74-3.08	-
	(preobese)
	Overweight	48	20 (41.7)	- -		3.59 +- 1.00, 3.29-3.88f	-
	(obese)

Statistically significant P values are shown in bold.

^a ?² Test with Yates correction.

^b n= 199.

^c Welch test was used because variances were unequal. All percentages are row percentages.

^d n = 99.

^e n = 35.

^f n = 47, n of success rate is 1 less than n of chest wall thickness measurements because we were unable to measure an unsuccessful intervention at the fifth ICS of an obese female cadaver.

and cadaver-based studies are limited by unique risk of measurement bias; however, it is not feasible and ethical to conduct a randomized- control trial in patients with tension pneumothorax. In this report, we present the largest cadaver-based cohort to date to the best of our knowledge, and we observed a statistically nonsignificant 9% more NT success rate at second ICS at a power of 88% and statistically significant more success rate in males at fifth ICS was (47.7%). We also observed thinner CWTs and higher success rates than previous imaging-based studies. A BMI of 23 kg/cm² or less and weight of 72 kg or less seems to accurately rule out NT failure, and they seem to be promising predictors at the bedside.

Funding

No funding or grants were received.

Conflicts of interest

The authors declare that they have no conflicts of interest.

Author contributions

Initials of the contributing authors were listed in brackets after the relevant parts of the research: Literature search (CO, HA, AD, OO), study design (CO, HA, AD, OO), legislative applications (CO, HA), data collection

(CO, ROO, EO, CUO), supervision and quality control (HA, AD, OO, YB), statistical advice (AD), statistical data analysis (CO, HA), data interpreta- tion (CO, HA, AD, OO), drafting the manuscript (CO, HA). All authors were involved in the writing and critical revision of the manuscript and approved the final version. CO and HA take responsibility for the paper as a whole.

References

1. Powers WF, Clancy TV, Adams A, West TC, Kotwall CA, Hope WW. Proper catheter selection for needle thoracostomy: a height and weight-based criteria. Injury 2014;45:107-11. http://dx.doi.org/10.1016/j.injury.2013.08.026.
2. Cameron P, Jelinek G, Kelly A-M, Brown AFT, Little M. Pneumothorax. In: Cameron P, Jelinek G, Kelly A-M, Brown AFT, Little M, editors. Textbook of adult emergency medicine. 1st ed. Elsevier Health Sciences; 2014. p. 1036.
3. The American College of Surgeons COT. Thoracic trauma. Proper catheter selection for needle thoracostomy: a height and weight-based criteria. 9 ed. The American College of Surgeons; 2012.
4. Cullinane DC, Morris JA, Bass JG, Rutherford EJ. Needle thoracostomy may not be in- dicated in the trauma patient. Injury 2001;32:749-52.
5. Inaba K, Ives C, McClure K, Branco BC, Eckstein M, Shatz D, et al. Radiologic evalua- tion of alternative sites for needle decompression of tension pneumothorax. Arch Surg 2012;147:813-8. http://dx.doi.org/10.1001/archsurg.2012.751.
6. Martin M, Satterly S, Inaba K, Blair K. Does needle thoracostomy provide adequate and effective decompression of tension pneumothorax? J Trauma Acute Care Surg 2012;73:1412-7. http://dx.doi.org/10.1097/TA. 0b013e31825ac511.
7. Sanchez LD, Straszewski S, Saghir A, Khan A, Horn E, Fischer C, et al. Anterior versus lateral needle decompression of tension pneumothorax: comparison by computed

   Table 2

   AUCs of receiver operating curves for weight, BMI, and CWT and calculated threshold values to rule out NT failure

   Variable	Subgroup	n	Accuracy (AUC), 95% CI	Accuracy (AUC), P	Threshold at 100% sensitivitya	Specificity at 100% sensitivity, 95% CI
   Weight (kg)	All	200	0.81, 0.75-0.86	b.0001	<=72	48.5, 40.6-56.4
   	Female	36	0.77, 0.60-0.89	.0008	<=(56-81)a	57.1, 34.0-78.2
   	Male	164	0.84, 0.77-0.89	b.0001	<=72	47.2, 38.9-55.7
   BMI (kg/cm2)	All	200	0.89, 0.83-0.93	b.0001	<=23	41.2, 33.6-49.1
   	Female	36	0.77, 0.60-0.89	.0017	<= (24-30)a	57.1, 34.0-78.2
   	Male	164	0.87, 0.80-0.91	b.0001	<=23	41.7, 33.5-50.2
   CWT (cm)	All	199b	0.92, 0.87-0.95	b.0001	<=2.4	59.4, 51.5-67.0
   	Female	35b	0.96, 0.83-1.00	b.0001	<=2.4	66.7, 43.0-85.4
   	Male	164	0.90, 0.84-0.94	b.0001	<=2.4	58.3, 49.8-66.5

   ^a No female cadavers were between 56 and 81 kg and had a BMI between 24 and 30 kg/cm². Hence, we estimate that the population cutoffs are within reported limits.

   ^b We were unable to measure an unsuccessful intervention at the fifth ICS of an obese female cadaver.

   tomography chest wall measurement. Acad Emerg Med 2011;18:1022-6. http://dx. doi.org/10.1111/j.1553-2712.2011.01159.x.

   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;64:111-4. http://dx.doi.org/10.1097/01.ta. 0000239241.59283.03.

8. McLean AR, Richards ME, Crandall CS, Marinaro JL. Ultrasound determination of chest wall thickness: implications for needle thoracostomy. Am J Emerg Med 2011;29:1173-7. http://dx.doi.org/10.1016/j.ajem.2010.06.030.
9. Inaba K, Branco BC, Eckstein M, Shatz DV, Martin MJ, Green DJ, et al. Optimal posi- tioning for emergent needle thoracostomy: a cadaver-based study. J Trauma Acute Care Surg 2011;71:1099-103. http://dx.doi.org/10.1097/TA.0b013e31822d9618.
10. Givens ML, Ayotte K, Manifold C. Needle thoracostomy: implications of computed tomography chest wall thickness. Acad Emerg Med 2004;11:211-3.
11. Wax DB, Leibowitz AB. Radiologic assessment of potential sites for needle decom- pression of a tension pneumothorax. Anesth Analg 2007;105:1385-8. http://dx. doi.org/10.1213/01.ane.0000282827.86345.ff.
12. Schroeder E, Valdez C, Krauthamer A, Khati N, Rasmus J, Amdur R, et al. Average chest wall thickness at two anatomic locations in trauma patients. Injury 2013;44: 1183-5. http://dx.doi.org/10.1016/j.injury.2013.03.027.
13. Stevens RL, Rochester AA, Busko J, Blackwell T, Schwartz D, Argenta A, et al. Needle thoracostomy for tension pneumothorax: failure predicted by chest computed tomography. Prehosp Emerg Care 2009;13:14-7. http://dx.doi.org/10.1080/ 10903120802471998.
14. Lamblin A, Turc J, Bylicki O, Loheas D, Martinez J-Y, Derkenne C, et al. Measure of chest wall thickness in French soldiers: which technique to use for needle decompression of tension pneumothorax at the front? Mil Med 2014;179:783-6. http://dx.doi.org/10.7205/MILMED-D-13-00512.
15. Yamagiwa T, Morita S, Yamamoto R, Seki T, Sugimoto K, Inokuchi S. Determination of the appropriate catheter length for needle thoracostomy by using computed tomography scans of trauma patients in Japan. Injury 2012;43:42-5. http://dx.doi. org/10.1016/j.injury.2010.11.022.
16. Akoglu H, Akoglu EU, Evman S, Akoglu T, Altinok AD, Guneysel O, et al. Determination of the appropriate catheter length and place for needle thoracostomy by using computed tomography scans of pneumothorax patients. Injury 2013;44:1177-82. http://dx.doi.org/10.1016/ j.injury.2012.10.005.
17. Chang SJ, Ross SW, Kiefer DJ, Anderson WE, Rogers AT, Sing RF, et al. Evaluation of 8.0-cm needle at the fourth anterior axillary line for needle chest decompression of tension pneumothorax. J Trauma Acute Care Surg 2014;76:1029-34. http://dx. doi.org/10.1097/TA.0000000000000158.
18. Britten S, Palmer SH, Snow TM. needle thoracocentesis in tension pneumothorax: insufficient cannula length and potential failure. Injury 1996;27:321-2.
19. World Health Organization. body mass index classification. AppsWhoInt. http:// apps.who.int/bmi/; 2015. [accessed 09.04.2015 2015].
20. Dancey CP, Reidy J. Statistics without maths for psychology. Pearson education; 2007.
21. Iseri A, Arslan N. Obesity in adults in Turkey: age and regional effects. Eur J Pub Health 2009;19:91-4. http://dx.doi.org/10.1093/eurpub/ckn107.
22. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837-45.
23. Clemency BM, Tanski CT, Rosenberg M, May PR, Consiglio JD, Lindstrom HA. Sufficient catheter length for pneumothorax needle decompression: a meta- analysis. Prehosp Disaster Med 2015;30:1-5. http://dx.doi.org/10.1017/ S1049023X15004653.