Introduction

Pneumothorax (PTX) is common following penetrating thoracic trauma and is usually treated by the insertion of an intercostal chest drain (ICD) [1, 2]. There is, however, evidence to suggest that a small, uncomplicated PTX in an asymptomatic patient may be treated conservatively by active clinical observation [3]. Most of the literature advocating this approach came from our parent institution at King Edward VIII Hospital (KEH) in Durban over three decades ago [4, 5], and we have continued to follow these guidelines in our practice. Since the time of these original reports, there has been a general paucity of literature on this selective conservative approach. The objective of this study was to review the contemporary outcome of small, uncomplicated PTX in a high volume trauma service in South Africa.

Materials and methods

A retrospective study was undertaken at the Pietermaritzburg Metropolitan Trauma Service (PMTS), Pietermaritzburg, South Africa. A retrospective review was performed on our prospectively maintained regional trauma registry over a period of 2 years from January 2012 to December 2013. Ethical approval for this study and to maintain our registry was granted by the Biomedical Research Ethics Committee (BREC) of the University of KwaZulu Natal (reference number: BE 207/09). The PMTS provides definitive trauma care to the city of Pietermaritzburg, the capital of KwaZulu Natal (KZN) Province, as well as the surrounding rural areas. It also serves as the trauma referral service for 19 other district hospitals within the province, and covers a total catchment population of over 3 million people.

All patients who sustained stab wounds to the chest and who did not have an immediate clinical indication for an ICD such as severe respiratory distress, tension pneumothorax, or an indication for operative intervention were eligible for inclusion. Data collected included: basic demographics, side of injury, single or multiple stabs, and type of weapon. All patients who present to our trauma unit are assessed clinically. If no immediate indication for an ICD is present, an erect inspiratory chest radiograph (CXR) is obtained. All patients who were asymptomatic, with CXR evidence of a small PTX were included in the study. We define a small PTX as less than 2 cm in size, measured from the apex of lung to the highest point of the cupula. Figure 1 shows an example of a small PTX (arrow) in a typical patient following stab injury to the right thorax.

Fig. 1
figure 1

A small PTX

Full size image

All patients who met these criteria were admitted to our designated trauma ward for active clinical observation. All patients were actively observed for evidence of clinical deterioration, such as respiratory distress, tachypnoea, and desaturation. We routinely performed a follow-up CXR at 12 h in these patients to determine the need for intervention. Those who showed progression of PTX on CXR, or those who had clinical signs of deterioration were treated immediately with an ICD inserted in the fifth intercostal space under local anaesthesia. The remaining patients who remained asymptomatic with no CXR sign of progression were discharged from the hospital at 24 h, with advice to return if they became symptomatic. All the original CXR images were digitalised for processing and reviewed by the primary author (VYK). The size of each pneumothorax was measured. These measurements were used to classify the patients into four separate groups for purposes of analysis: Group A: <0.5 cm, Group B: ≥0.5 to <1 cm, Group C: ≥1 to <1.5 cm and Group D: ≥1.5 to <2 cm. All relevant data were extracted onto a Microsoft EXCEL© spread sheet for processing and detailed analysis.

Results

Demographics

During the 2-year study period, a total of 125 patients were eligible for inclusion, of which 92 % (115/125) were male and 8 % (10/125) were female, with a median age of 21 (19–24) years.

Pattern of injury

Ninety-seven per cent (121/125) of all stab injuries were inflicted by knives and the remaining 3 % (4/125) by screwdrivers. Sixty-one per cent of injuries were on the left side, and 39 % were on the right. Eighty-two per cent (102/125) of all patients sustained a single stab injury, and the remaining 18 % (23/125) had multiple stab wounds to the thorax.

Size of pneumothoraces

Thirty-nine per cent (49/125) had a PTX <0.5 cm (Group A), 26 % (32/125) were ≥0.5 to <1 cm (Group B), 19 % (24/125) were ≥1 to <1.5 cm (Group C) and 15 % (20/125) were ≥1.5 to <2 cm (Group D). Table 1 summarises the proportion in each subgroup.

Table 1 Size of PTXs
Full size table

Outcomes

Of the 125 patients in this study, 3 % (4/125) eventually required an ICD, while 97 % (121/125) were managed successfully by conservative treatment. Table 2 summarises the outcome of all 125 patients according to the four groups they were categorised into. None of the patients in Group A or B required an ICD. One patient in Group C eventually required an ICD due to radiographic evidence of progression of the PTX. Three patients from Group D eventually required an ICD. Two had progression evident on CXR. The remaining patient developed severe respiratory distress and had an ICD inserted without a further CXR. Subsequent post-insertion CXR showed partial re-expansion of the lung. All but one of the patients who required an ICD had sustained multiple stab injuries to the chest. All four patients who eventually required an ICD had them in situ for 24 h and recovered without any adverse outcomes. All subsequent CXRs showed complete resolution of the original PTX. All other patients who were managed successfully with clinical observation alone were discharged home well. The mean length of hospital stay was 1.2 days. There was no morbidity or mortality as a direct result of our selective conservative approach in this study.

Table 2 Outcome of all 125 patients managed selectively
Full size table

Discussion

Traumatic PTXs are found in up to 40 % of all thoracic trauma patients, and an ICD is usually the standard management for these patients [1–3]. ICD insertion, however, is a potentially dangerous procedure which is associated with an overall complication rate of up to 30 % [6, 7]. In light of the high volume of penetrating trauma in our environment a selective approach is an attractive one. [8]. Selective conservatism for penetrating thoracic trauma was initially described at our parent institution, KEH in Durban over three decades ago [4, 5]. This practice continues to form the basis of our trauma protocol in the PMTS. Hegarty [4] from KEH published a seminal paper in 1976 on the selective conservative management of penetrating thoracic trauma and found that a significant proportion (14 %) of patients with stab and GSW injuries to the chest could be safely observed without the need for ICD insertion. A follow up study by Muckart in 1985 reviewed 85 patients with PTX and, or haemothorax (HTX) who were managed conservatively. In the subgroup of 40 patients with small-uncomplicated traumatic PTXs, only one eventually required an ICD, with no associated adverse outcomes [5]. Several authors have gone on to describe the successful application of a selective approach to these injuries, but none have focused exclusively on stab injuries of the chest [9–11]. Table 3 shows a comparison between different studies to date [4, 5, 9–11]. Our study showed that less than 3 % of all patients who were treated conservatively ultimately required an ICD. This is consistent with the findings from Muckart et al. and Hegarty. We believe that the relatively higher success rate in our study was related to our exclusive focus on stab wounds. Although selected patients with GSWs could probably also be managed conservatively, GSWs are more likely to inflict a greater degree of lung injury [12], and our experience would suggest that these patients seldom have PTX alone as the sole pathology with haemothorax and, or hemopneumothorax being relatively more common.

Table 3 Literature to date (1979–2014)
Full size table

Although much has been written regarding the management of PTX in general, there is often a great deal of confusion surrounding the terminology, which makes it difficult to extrapolate results from one environment to another. PTX simply denotes the presence of air within the pleural cavity and may be traumatic or non-traumatic in origin [13]. In the non-trauma setting, most studies tend to focus on spontaneous PTX, which is further divided into primary or secondary [13]. Although there is sufficient evidence to advocate a conservative approach in these patients [13], the underlying pathophysiology is likely to differ from traumatic PTX [14]. Occult PTX, on the other hand, specifically describes the entity where CXR is considered normal, but computed tomography (CT) scan of the chest reveals evidence of a PTX [15]. Several studies on the management traumatic PTX often include patients with occult PTX as part of their cohort and, this substantially weakens their generalisability [16, 17].

The exact definition of a ‘small pneumothorax’ has been subject to much debate [13]. Different studies report the use of either 1.5 or 2 cm as the cut-off, but many do not specify how the measurements were obtained [9–11]. The British Thoracic Society (BTS) guidelines suggest 2 cm as a cut-off for the definition of small pneumothorax [13]. The BTS guideline stipulates that the presence of a visible rim of >2 cm is to be measured from the lung margin to the chest wall at the level of the hilum. The American College of Chest Physicians (ACCP), however, recommends that measurements be taken from the lung apex to the cupula [18]. We continue to utilise the ACCP definition. Our study differs from previous published studies, in that we had clear categorisation of the actual size of the PTX. The widespread availability of digital processing of radiological images allows for easy manipulation and accurate quantification of the size of a PTX. This was not possible previously.

From our study, it would appear that small-uncomplicated traumatic PTX,s up to 1 cm in size do not require any treatment, other than active clinical observation. The risk factors for failure of conservative treatment of small-uncomplicated PTX are a PTX ≥1.5 to <2 cm in size and the presence of multiple thoracic stab wounds. Of the four patients who required delayed insertion of an ICD, three had a PTX ≥1.5 to <2 cm, which represents 15 % (3/20) of this cohort and all but one had sustained multiple stab injuries to the thorax. We therefore continue to advocate active clinical observation of all small-uncomplicated PTX from stab wounds provided it can be done safely. A period of clinical observation of 12–24 h appears to be more than sufficient.

Conclusions

The majority of asymptomatic patients with a small PTX following a stab injury can be managed conservatively. We advocate that such patients be actively observed by experienced staff for signs of clinical deterioration. Patients with PTX close to 2 cm in size and patients with multiple stabs injuries are the subgroup most likely to complicate and require ICD insertion.