a b s t r a c t

Background: traumatic pneumothorax management has evolved to include the use of smaller caliber tube thoracostomy and even observation alone. Data is limited comparing tube thoracostomy to observation for small Traumatic pneumothoraces. We aimed to investigate whether observing patients with a small traumatic pneumothorax on initial chest radiograph is associated with improved outcomes compared to tube thoracostomy.

Methods: We retrospectively reviewed trauma patients at our level 1 trauma center from January 1, 2016 through December 31, 2020. We included those with a pneumothorax size <30 mm as measured from apex to cupola on initial CXR. We excluded patients with injury severity score >= 25, operative requirements, hemothorax, bilateral pneumothoraces, and intensive care unit admission. Patients were grouped by Management strategy (observa- tion vs tube thoracostomy). Our primary outcome was length of stay with secondary outcomes of pulmonary in- fection, failed trial of observation, readmission, and mortality. Results are listed as mean +- standard error of the mean.

Results: Of patients who met criteria, 39 were in the observation group, and 34 were in the tube thoracostomy group. Baseline characteristics were similar between the groups. Average pneumothorax size on CXR was 18 +- 1.0 mm in the observation group and 18 +- 0.84 mm in the tube thoracostomy group (p > 0.99). Average pneumothorax sizes on computed tomography were 25 +- 2.1 and 37 +- 3.9 mm in the observation and tube thoracostomy groups, respectively (p = 0.01). Length of stay in the observation group was significantly shorter than the tube thoracostomy group (3.6 +- 0.33 vs 5.8 +- 0.81 days, p < 0.01). While pneumothorax size on com- puted tomography was associated with tube thoracostomy, only tube thoracostomy correlated with length of stay on multivariable analysis; pneumothorax size on CXR and computed tomography did not. There were no deaths or readmissions in either cohort. One patient in the observation group required tube thoracostomy after 18 h for worsening subcutaneous emphysema, and one patient in the tube thoracostomy group developed an empyema.

Conclusions: Select patients with small traumatic pneumothoraces on initial chest radiograph who were treated with observation experienced an average length of stay over two days shorter than those treated with tube thoracostomy. Outcomes were otherwise similar between the two groups suggesting that an observation-first strategy may be a superior treatment approach for these patients.

(C) 2023 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://

creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

Traumatic pneumothorax management has evolved to include the use of smaller caliber tube thoracostomy (TT), and at times, observation alone [1-7]. While these studies suggest that observation for small

* Corresponding author at: Department of Surgery, UCSF East Bay, 1411 E 31st St., Oakland, CA 94602, USA.

E-mail address: [email protected] (K.C. Banks).

traumatic pneumothorax may be safe, they do not compare outcomes after observation to those after TT among patients with similar sized pneumothoraces [2-4]. Additionally, they include heterogeneous popu- lations of patients with varying injury severity scores (ISS), mechanical ventilation requirements, and operative requirements [2-4]. There has also been a similar trend in the management of spontaneous pneumo- thorax with evidence suggesting that even select large pneumothoraces may be safely observed [8]. Other studies have specifically highlighted the safety and benefits of observation for Occult traumatic

https://doi.org/10.1016/j.ajem.2023.01.017

0735-6757/(C) 2023 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Fig. 1. Example of the measurement of pneumothorax size from apex to cupola on a chest radiograph.

pneumothoraces seen on computed tomography alone [5-7]. However, the optimal management for small traumatic pneumothoraces visible on initial chest radiograph (CXR) remains undetermined.

If successful, management of small traumatic pneumothoraces with observation would have the potential to lead to reduced hospital length of stay while avoiding complications associated with TT. We aimed to compare outcomes of similar groups of patients with a unilat- eral small traumatic pneumothorax on initial CXR. Our hypothesis was that observation is associated with shorter length of stay with otherwise similar rates of treatment failure, pulmonary infection, readmission, and mortality compared to TT for a small traumatic pneumothorax among blunt and penetrating trauma patients with similar injury severity, similar sized pneumothoraces, and lack of mechanical ventilation requirement.

1. Methods

We reviewed trauma patients at our level 1 trauma center from Jan- uary 1, 2016 through December 31, 2020. We used the 2001 American College of Chest Surgeons Guidelines for Spontaneous pneumothorax which define small pneumothoraces as <3 cm from apex to cupola (Fig. 1) [9]. The decision to pursue observation or tube thoracostomy at our institution is dependent on trauma team preference. As a result, our population receives a mix of the two treatment strategies.

We included blunt and penetrating trauma patients with small trau- matic pneumothoraces, and we set strict exclusion criteria so that, aside from treatment strategy, the comparison groups would otherwise be as similar as possible. We excluded patients with ISS >=25, operative requirement due to initial traumatic injury, hemothorax, tension phys- iology, bilateral pneumothoraces, ventilator requirement, and intensive care unit admission. A prospectively collected trauma database was used to identify the patients meeting such criteria, and manual chart re- view was performed to collect imaging measurements and details re- garding patient course. The study was approved by our Institutional Review Board (IRB22-03161EX).

Comparison groups were those managed by observation versus TT (Fig. 2). TT was performed with tubes 24 to 32 French. We collected var- iables on age, sex, race/ethnicity, Size of pneumothorax on CXR (mea- sured from apex to cupola), size of pneumothorax on CT (linear distance from parietal to visceral pleura at deepest pocket), ISS, abbrevi- ated injury scale related to chest injury (AIS Chest), presence of pulmo- nary contusions, Revised Trauma Score, and mechanism of injury. Our primary outcome was length of stay (LOS) with secondary outcomes of pulmonary infection, failED management. readmission, and mortality. A failed trial of observation or the need for repeat TT or surgery after ini- tial TT were considered failed management.

Chi square tests were used for categorical data. t-tests and Mann- Whitney U tests were used for normal and nonnormal distributions of quantitative data, respectively. multivariable regression was performed to assess for associations between variables of interest and patient out- comes. For ordinal variables of ISS and AIS Chest, dichotomous variables with cutoffs of ISS <15 and AIS Chest <3 were used in multivariable analyses. Data is listed as mean +- standard error of the mean (SEM). Statistical analyses were performed with IBM SPSS Statistics Version 27.

1. Results

Of patients who met criteria, 39 were in the observation group, and 34 were in the TT group. Baseline characteristics of age, sex, and race/ ethnicity were not different between the groups (Table 1). Average pneumothorax size on CXR was 18 mm in both groups (+-1.0 mm in ob- servation group, +-0.84 mm in TT group; p > 0.99). Average pneumo- thorax sizes on CT were 25 +- 2.1 and 37 +- 3.9 mm in the observation and TT groups, respectively (p = 0.01). ISS, RTS, and AIS chest were the same in both groups.

There were no deaths or readmissions in either cohort. LOS in the observation group was significantly shorter than in the TT group (3.6 +- 0.33 vs 5.8 +- 0.81 days; p < 0.01) (Table 2). While pneumotho- rax size on CT was associated with TT, TT correlated with LOS on multi- variable analysis; pneumothorax size on CXR and CT did not (Table 3). There was one episode of failed management in each of the two groups (p > 0.99). The patient that failed management in the observa- tion group required TT after 18 h due to worsening subcutaneous

Fig. 2. Treatment strategies for patients with small traumatic pneumothoraces.

Table 1 Baseline Characteristics.
		Type of Management
Characteristic	Total (N = 73)	Observation (N = 39)	Tube Thoracostomy (N = 34)	p-value
Mean Age, years (SEM)	44 (2.1)	48 (3.1)	39 (2.5)	0.06b
Female, N (%)	18 (25)	11 (28)	7 (21)	0.45c
Race/Ethnicity, N (%)
Asian	8 (11)	3 (8)	5 (15)
Black	27 (37)	13 (33)	14 (41)
Hispanic	12 (16)	6 (15)	6 (18)
White	21 (29)	14 (36)	7 (21)
Other	5 (7)	3 (8)	2 (6)	0.60c
Mean Pneumothorax Size on CXR, mm (SEM)	18 (0.66)	18 (1.0)	18 (0.84)	>0.99a
Mean Pneumothorax Size on CT, mm (SEM)?	30 (2.1)	25 (2.1)	37 (3.9)	0.01b
Mean ISS (SEM)	12 (0.59)	12 (0.72)	12 (0.98)	0.57b
Mean AIS Chest (SEM)	2.7 (0.09)	2.7 (0.11)	2.7 (0.15)	0.92b
Mean RTS (SEM)	7.8 (0)	7.8 (0)	7.8 (0)	-
Presence of pulmonary contusions, N (%)	25 (34)	15 (39)	10 (29)	0.42c
Penetrating Trauma, N (%)	17 (23)	8 (21)	9 (27)	0.55c

^a Independent Sample t-test.

^b Mann-Whitney U test.

^c Chi-Square Test.

* 24/34 tube thoracostomy patients underwent computed tomography and 37/38 observed patients underwent computed tomography. SEM, standard error of the mean; CXR, chest radiograph; CT, computed tomography scan; ISS, injury severity score; AIS Chest, Abbreviated Injury Scale related to chest injury; RTS, revised trauma score.

Table 2 Outcomes.
		Type of Management
	Total (N = 73)	Observation (N = 39)	Tube Thoracostomy (N = 34)	p-value
Mean LOS, days (SEM)	4.6 (0.43)	3.6 (0.33)	5.8 (0.81)	<0.01a
Empyema, N (%)	1 (1)	0 (0)	1 (3)	0.47b
Failed Management, N (%)	2 (3)	1 (3)	1 (3)	>0.99b

^a Mann-Whitney U Test.

^b Chi-Square Test; SEM, standard error of the mean; LOS, length of stay.

emphysema on clinical exam. The patient that failed management in the TT group developed an empyema requiring thoracoscopic decortication.

1. Discussion

The management of traumatic pneumothorax is evolving. While several studies have suggested that certain traumatic pneumothoraces can be safely observed, these studies do not compare similar groups of patients with comparably sized pneumothoraces. In this study, we com- pared similarly injured patients with small traumatic pneumothoraces on CXR with the hypothesis that observation would be associated with improved outcomes compared to TT. Ultimately, we found that

observation of small traumatic pneumothoraces was associated with a shorter length of stay compared to TT.

While TT may be effective in the treatment of traumatic pneumothoraces, complication rates have been reported to be as high as 40% [7,10-12]. Our findings support the conclusion that observation is a safe treatment strategy for small traumatic pneumothoraces, so if the complication risk associated with TT can be avoided altogether, ob- servation may be the superior treatment strategy for these patients. While we found similar rates of failed management in each group, failed management in the observation group led to TT and a subsequently un- eventful hospital stay. Alternatively, the case of failed management in the TT group resulted in empyema and ultimately a VATS decortication- a considerably more morbid hospital course.

Table 3

Univariable and multivariable analyses of variable associations with hospital length of stay in days.

	Univariable Analysis			Multivariable Analysis
Variable	Unstandardized B Coefficient	95% Confidence Interval		Unstandardized B Coefficient	95% Confidence Interval
Age (years)	0.00	-0.05-0.05		-	-
Male	0.17	-1.85-2.19		-	-
ISS	3.22	1.35-5.09		1.85	-0.32-4.03
AIS Chest Penetrating Mechanism	2.47 -1.44	0.83-4.11 -3.47-0.60		1.94 -1.18	0.06-3.83 -3.32-0.95
Pneumothorax Size on CXR (cm)	-0.08	-0.23-0.08		-	-
Pneumothorax Size on CT (cm)	0.06	0.00-0.11		0.02	-0.04-0.07
Presence of pulmonary contusions	0.87	-0.96-2.69		-	-
Tube Thoracostomy	2.12	0.45-3.79		2.09	0.19-3.98

Univariable analysis cutoff of p < 0.20 used for inclusion in multivariable model. ISS, injury severity score treated as dichotomous variable with cutoff <15; AIS Chest, abbreviated injury scale related to chest injury treated as dichotomous variable with cutoff <3; CXR, chest radiograph; cm, centimeters; CT, computed tomography.

Perhaps the most interesting finding in our study, however, is that observation was associated with a LOS over two days shorter than tube thoracostomy for patients with small traumatic pneumothoraces. This suggests that in addition to being a safe treatment strategy, obser- vation may be associated with improved outcomes for these patients. Additionally, given the shorter hospitalization time, observation would appear to be a more cost-effective option for both patients and the hos- pital system.

Given that pneumothoraces were compared based on size on CXR, this study may be particularly helpful among institutions with limited access to computed tomography. Although we found size of pneumo- thorax on CT was associated with likelihood of TT, the size of the pneu- mothorax was not associated with LOS. Based on these results, it is possible that small traumatic pneumothoraces as measured on CXR may not require TT regardless of size on CT.

It is also worth noting that all patients were admitted for at least one day. This allowed for a period of observation to determine whether TT would become necessary. While the one patient with failed manage- ment in the observation group underwent TT 18 h after arrival to the emergency department, the optimal duration of observation is yet to be determined.

1. Limitations

In addition to its retrospective nature, our study has a few limita- tions. First, sample sizes are small. This is largely due to the strict exclu- sion criteria applied to allow for comparable groups. We felt that including very severely injured patients, patients with hemothoraces, or patients requiring positive pressure ventilation for any reason would allow for too many confounding variables to draw reliable out- comes comparisons. Another limitation is that patients undergoing ob- servation did not have standardized observation protocols. All of them were admitted for at least one day with a follow up CXR, however, there were not standardized discharge criteria. Finally, we did not in- clude small bore (pigtail) catheters in our study. Further studies would need to be performed to compare observation to placement of these small bore catheters.

1. Conclusions

We found that select patients with small traumatic pneumothorax on initial CXR who were treated with observation had an average LOS over two days shorter than similar patients treated with TT. Though pneumothorax size on CT may have influenced the decision for TT placement in some patients, neither size on CXR nor CT correlated with LOS. An understanding of the Optimal treatment strategy for small traumatic pneumothoraces on CXR may be helpful in practice set- tings where access to CT is limited. Observation and the associated shorter LOS may help avoid significant morbidity and reduce the Healthcare costs related to TT.

Funding

This research did not receive any specific grant from funding agen- cies in the public, commercial, or not-for-profit sectors.

Prior presentation

2022 Pacific Coast Surgical Association Meeting (February 13, 2022).

Declaration of Competing Interest

None.

Acknowledgements

This research did not receive any specific grant from funding agen- cies in the public, commercial, or not-for-profit sectors.

References

1. Inaba K, Lustenberger T, Recinos G, Georgiou C, Velmahos GC, Brown C, et al. Does size matter? A Prospective analysis of 28-32 versus 36-40 French chest tube size in trauma. J Trauma Acute Care Surg. 2012;72:422-7. https://doi.org/10.1097/TA. 0b013e3182452444.
2. Bou Zein Eddine S, Boyle KA, Dodgion CM, Davis CS, Webb TP, Juern JS, et al. Observ- ing pneumothoraces: the 35-millimeter rule is safe for both blunt and penetrating chest trauma. J Trauma Acute Care Surg. 2019;86:557-64. https://doi.org/10.1097/ TA.0000000000002192.
3. Gerhardy BC, Liebenberg P, Simpson G. Conservative management of traumatic pneumothoraces: A retrospective cohort study. Emerg Med Australas. https://doi. org/10.1111/1742-6723.13846.
4. Walker SP, Barratt SL, Thompson J, Maskell NA. Conservative Management in Trau- matic Pneumothoraces. Chest. 2018;153:946-53. https://doi.org/10.1016/j.chest. 2017.10.015.
5. Yadav K, Jalili M, Zehtabchi S. Management of traumatic Occult pneumothorax. Resuscitation. 2010;81:1063-8. https://doi.org/10.1016/j.resuscitation.2010.04.030.
6. Beattie G, Cohan CM, Tang A, Chen JY, Victorino GP. Observational management of penetrating Occult pneumothoraces: outcomes and risk factors for interval tube thoracostomy placement. J Trauma Acute Care Surg. 2022;92:177-84. https://doi. org/10.1097/TA.0000000000003415.
7. Zhang M, Teo LT, Goh MH, Leow J, Go KTS. Occult pneumothorax in blunt trauma: is there a need for tube thoracostomy? Eur J Trauma Emerg Surg Off Publ Eur Trauma Soc. 2016;42:785-90. https://doi.org/10.1007/s00068-016-0645-x.
8. Brown SGA, Ball EL, Perrin K, Asha SE, Braithwaite I, Egerton-Warburton D, et al. Conservative versus interventional treatment for spontaneous pneumothorax. N Engl J Med. 2020;382:405-15. https://doi.org/10.1056/NEJMoa1910775.
9. Baumann MH, Strange C, Heffner JE, Light R, Kirby TJ, Klein J, et al. Management of Spontaneous Pneumothorax. Chest. 2001;119:590-602. https://doi.org/10.1378/ chest.119.2.590.
10. Chan L, Reilly KM, Henderson C, Kahn F, Salluzzo RF. Complication rates of tube thoracostomy. Am J Emerg Med. 1997;15:368-70. https://doi.org/10.1016/S0735- 6757(97)90127-3.
11. Sritharen Y, Hernandez MC, Haddad NN, Kong V, Clarke D, Zielinski MD, et al. Exter- nal validation of a tube Thoracostomy complication classification system. World J Surg. 2018;42:736-41. https://doi.org/10.1007/s00268-017-4260-8.
12. Hernandez MC, El Khatib M, Prokop L, Zielinski MD, Aho JM. Complications in tube thoracostomy: systematic review and Meta-analysis. J Trauma Acute Care Surg. 2018;85:410-6. https://doi.org/10.1097/TA.0000000000001840.