Article Text

Management of paediatric spontaneous pneumothorax: a multicentre retrospective case series
  1. Paul D Robinson1,2,
  2. Carol Blackburn3,
  3. Franz E Babl3,
  4. Lalith Gamage4,
  5. Jacquie Schutz4,
  6. Rebecca Nogajski5,
  7. Stuart Dalziel6,7,
  8. Colin B Donald6,
  9. Dino Druda8,
  10. David Krieser8,
  11. Jocelyn Neutze9,
  12. Jason Acworth10,
  13. Mark Lee11,
  14. Peter K Ngo1
  15. on behalf of the Paediatric Emergency Departments International Collaborative (PREDICT) research network
    1. 1Department of Respiratory Medicine, Children's Hospital at Westmead, Sydney, New South Wales, Australia
    2. 2Discipline of Paediatrics and Child Health, University of Sydney, Sydney, Australia
    3. 3Royal Children's Hospital, Murdoch Children's Research Institute and University of Melbourne, Melbourne, Australia
    4. 4Emergency Department, Children, Youth and Women's Health Service, Adelaide, Australia
    5. 5Emergency Department, Sydney Children's Hospital, Sydney, Australia
    6. 6Starship Hospital, Auckland, New Zealand
    7. 7Liggins Institute, University of Auckland, New Zealand
    8. 8Emergency Medicine Department, Sunshine Hospital, St Albans, Victoria, Australia
    9. 9Emergency Medicine Department, Kidzfirst Middlemore Hospital, Otahuhu, New Zealand
    10. 10Emergency Medicine Department, Royal Children's Hospital, Brisbane, Australia
    11. 11Emergency Medicine Department, John Hunter Hospital, Newcastle, New South Wales, Australia
    1. Correspondence to Dr Paul D Robinson, The Children's Hospital at Westmead, Locked Bag 4001, Westmead, NSW 2145, Australia; paul.robinson1{at}


    Objectives Paediatric guidelines are lacking for management of spontaneous pneumothorax. Adult patient-focused guidelines (British Thoracic Society 2003 and 2010) introduced aspiration as first-line intervention for primary spontaneous pneumothorax (PSP) and small secondary spontaneous pneumothoraces (SSP). Paediatric practice is unclear, and evidence for aspiration success rates is urgently required to develop paediatric-specific recommendations.

    Methods Retrospective analysis of PSP and SSP management at nine paediatric emergency departments across Australia and New Zealand (2003–2010) to compare PSP and SSP management.

    Results 219 episodes of spontaneous pneumothorax occurred in 162 children (median age 15 years, 71% male); 155 PSP episodes in 120 children and 64 SSP episodes in 42 children. Intervention in PSP vs SSP episodes occurred in 55% (95% CI 47% to 62%) vs 70% (60% to 79%), p<0.05. An intercostal chest catheter (ICC) was used in 104/219 (47%) episodes. Aspiration was used in more PSP than in SSP episodes with interventions (27% (18% to 37%) vs 9% (3% to 21%), p<0.05). Aspiration success was 52% (33% to 70%) overall and not significantly different between PSP and SSP. Aspiration success was greater in small vs large pneumothoraces (80% (48% to 95%) vs 33% (14% to 61%), p=0.01). Small-bore ICCs were used in 40% of ICCs and usage increased during the study.

    Conclusions In this descriptive study of pneumothorax management, PSP and SSP management did not differ and ICC insertion was the continuing preferred intervention. Overall success of aspiration was lower than reported results for adults, although success was greater for small than for large pneumothoraces. Paediatric prospective studies are urgently required to determine optimal paediatric interventional management strategies.

    • Paediatric Practice
    • Respiratory
    • Accident & Emergency

    Statistics from

    What is already known on this topic?

    • No paediatric-specific guidelines for the management of spontaneous pneumothoraces in children are available.

    • In adults, aspiration is increasingly used in the management of pneumothorax.

    • Optimal management of pneumothorax in children is unclear—in particular, the role of aspiration.

    What this study adds?

    • Clinicians tend not to differentiate between primary and secondary pneumothorax.

    • The management of pneumothorax in children is still largely based on intercostal catheters, and aspiration was successful in only half the cases in which it was attempted.

    • This study highlights the need for a large observational study to clarify optimal management of pneumothorax in children.



    Despite recent studies of primary spontaneous pneumothorax (PSP) in adults,1–3 paediatric data remain scarce and limited to small single-centre case series or studies in adult patients also containing adolescents.4 Paediatric-specific management has not been separately examined in any international guideline,5–7 and published paediatric studies fail to distinguish PSP from secondary spontaneous pneumothorax (SSP, defined as those with an existing diagnosis of underlying lung disease).8–12


    The distinction between PSP and SSP is important because their management differs. In SSP the presence of underlying lung disease leads to more severe symptoms, and breathlessness is disproportionate to the size of the pneumothorax; hence, the threshold for intervention is lower. The 2003 British Thoracic Society (BTS) international guidelines6 recommend aspiration as first-line intervention for PSP and small SSP, as opposed to intercostal chest catheter (ICC) insertion. These recommendations remain in the 2010 update.7 Success rates of aspiration in children have not yet been assessed, and adult and adolescent studies dealing with this topic typically contain few children.4

    Goals of this investigation

    The Paediatric Research in Emergency Departments (EDs) International Collaborative (PREDICT) is a network of Australian and New Zealand paediatric EDs.13 The primary aim of our study was to describe the management of PSP and SSP in paediatric patients presenting to EDs. The null hypotheses tested were that there would be no difference in PSP and SSP management in paediatric subjects, despite the BTS recommendations,6 ,7 and that aspiration success in children would not differ from that reported in adults.


    Study design and setting

    We carried out a multicentre retrospective study of PSP and SSP presentations over 7 years (June 2003–June 2010) at nine EDs in the PREDICT network (annual paediatric census 19 000–69 000 patients).13 Two centres developed protocols, based on the BTS guidelines,6 ,7 for management of spontaneous pneumothorax during the study period. The study was approved by the ethics committee at each of the sites taking part in the study.

    Selection of participants

    Presentations were identified using ICD-9 and ICD-10. Exclusion criteria included malignancy, traumatic or tension pneumothorax, neonates (post-gestational age <1 month), age >18 years in mixed adult and paediatric EDs (but all subjects presenting to paediatric EDs were recruited). Spontaneous pneumothoraces coinciding with minor trauma and not considered to be the direct cause of the pneumothorax were not excluded.


    Primary outcome was the management methods used (conservative, ICC, aspiration, later surgery) for PSP and SSP. Secondary outcomes were aspiration success, size of ICC and pneumothorax recurrence.

    Study process

    A data collection instrument was piloted by all sites after training the site researchers with standardised teaching materials.14 Site researchers were not blinded to the protocol and outcomes examined in this study. Data were collected on demographics, signs and symptoms at presentation, activity at symptom onset, past medical history (including details of previous pneumothorax and surgical management) and management occurring in the ED, inpatient ward and during follow-up (before the next episode, if recurrent). Chest X-ray (CXR) findings and CT scans were assessed for pneumothorax size and treatment response.

    Data from each subject were examined by a site investigator. To aid quality control, a second investigator reviewed a portion of the data and all X-ray findings. Inter-rater reliability was not assessed.


    SSP was defined as pneumothorax in a patient with a pre-existing diagnosis/condition with lung involvement or implications. Conservative management was defined as presentations where no interventions occurred. Pneumothorax size was classified according to the BTS guidelines.7 Small was defined as <2 cm gap between lung and chest wall (at the level of the hilum) in the presentation CXR; if measurement was not possible, then a description in the medical notes of ‘small’ (or similar terminology), or <50% of the hemithorax (if percentage was documented), was used. Large was defined as a pneumothorax that was not small. Aspiration was defined as successful if a decrease in pneumothorax size was seen on CXR examination, or recorded in the medical record and an ICC was not subsequently placed. ICC size was defined as small- (<16 French (F)), medium- (16–24 F) or large-bore (>24 F).15 Pneumothorax recurrence was defined as an episode occurring during the study, in a subject who had a history of previous pneumothorax recorded in the medical notes or past medical history.


    Data from each site were reviewed centrally at the Children's Hospital, Westmead. A power calculation to detect differences in management or intervention success was not possible because of insufficient published evidence in children, but the study was designed to maximise recruitment during the 7 years since the BTS guideline publication.7 Data were entered into a Microsoft Access database and analysed using SPSS (V., Chicago, SPSS Inc, USA). Proportions are presented as percentages with 95% CIs calculated using a modified Wald method.16 Comparisons between proportions used χ2 tests. Parametric and non-parametric data were compared using t tests and Mann–Whitney tests, respectively. Distribution was determined by visual inspection of histograms for each variable.


    Characteristics of study subjects

    A total of 219 presentations of spontaneous pneumothorax (162 children with 120 PSP and 42 SSP) were identified across the nine sites, with two sites each contributing 20% of the cases. There were 155 episodes (71%) of PSP and 64 episodes (29%) of SSP (figure 1, table 1). Of all episodes, 134 (61%) were the first known episode in that child. Overall, the median (range) age at initial presentation was 15.2 years (1 month to 19 years). Only 21 patients (13.0%) presented aged <13 years, with three patients presenting aged <1 year. There was no difference in the age at first presentation between PSP and SSP subjects. The 42 children presenting with SSP had the following diagnoses: 25 (60%) had asthma, 6 (14%) had cystic fibrosis; Marfan's syndrome, bronchiolitis obliterans, congenital cystic adenomatoid malformation each had two cases; congenital lobar emphysema, ataxia telangiectasia, previous methicillin-resistant Staphylococcus aureus pneumonia, or previous right lung hydatid cyst removal each had one case; and, one case had no additional information.

    Table 1

    Summary of episodes of spontaneous pneumothorax episodes, categorised as either primary or secondary spontaneous pneumothorax (PSP and SSP, respectively)

    Figure 1

    Overview of cases presenting during the study period. SP, spontaneous pneumothorax. *By the end of the study period.

    Clinical findings in the whole cohort

    At presentation, the commonest symptom was chest pain (187/216, 87%) followed by shortness of breath (93/216, 43%). Cough was rare (5%), and only two cases (1%) were asymptomatic. Most commonly, symptom onset occurred at rest (142/187, 76%). Symptom onset during activity (45/187, 24%) was most commonly sport-related (62%), followed by walking (16%), lifting (4%) or minor trauma (4%). Median (range) duration of symptoms before presentations was 1.0 (0–30) days, and did not differ between PSP and SSP cases. The weight z score was −0.1 (−9.1 to 2.7). Height and body mass index data were unavailable.

    In 185/219 (84%) episodes, the size of the pneumothorax could be classified: 144 by CXR examination and 41 from the clinical notes. The pneumothorax was left sided in 52% of presentations. The pneumothorax was small in 115/185 (62%) of presentations.

    Management of PSP

    Conservative management was used in 42% of PSP episodes (table 1). Of these, 30/65 (46%) were discharged home from the ED. Active intervention occurred in 55% of episodes (table 2). ICC was the most common initial choice (79%, 104/131). Aspiration was performed in the other episodes. Aspiration success rate overall was 52% (95% CI 33% to 70%).

    Table 2

    Summary of management of spontaneous pneumothorax episodes, categorised as either primary or secondary spontaneous pneumothorax (PSP and SSP, respectively)

    Overall, the majority of patients with PSP were admitted to the ward from the ED (81%) and had a mean (SD) length of stay of 5.6 (4.4) days. CT scans were performed in 35% of PSP episodes (either as an inpatient or as an outpatient), and 38% had a surgical procedure following that episode. Of those undergoing surgery, the commonest procedure was apical pleurectomy (27/59 where documented, 46%), followed by chemical pleurodesis (37%) and stapling (27%). The main approach used was video-assisted thoracoscopic surgery (36/59, 61%).

    Comparison of SSP management with PSP

    In comparing management of SSP and PSP episodes, the following findings are noteworthy. In subjects with SSP, conservative management was less common, and a trend towards more oxygen use at presentation was seen (p=0.055, table 1). Intervention was more frequent in SSP cases with greater ICC use, as both first-line intervention and overall for the entire SSP cohort (table 2). Aspiration was more likely to be attempted in subjects with PSP (although not statistically significant, p=0.19), with comparable success rates in subjects with SSP.

    The proportion discharged home from the ED was similar in patients with SSP and PSP. In those admitted to the ward, the length of stay was longer for patients with SSP (∼12 days). Subsequent CT scans and surgery were performed in a similar proportion for PSP and SSP cohorts. Of those patients with SSP undergoing surgery, the commonest procedure performed was apical pleurectomy (43%), followed by stapling (37%) and chemical pleurodesis (23%). The main approach used, when documented, was video-assisted thoracoscopic surgery (14/19, 74%).

    Management and procedure operators

    Table 3 summarises which service directed management and undertook procedures. The team directing management could be determined in 91% and 98% of PSP and SSP cases with interventions, respectively. ED staff supervised most cases and performed most of the procedures.

    Table 3

    Summary of supervising team and team performing the interventions

    Success rate of aspiration

    Overall, aspiration was performed as the first-line intervention for 27 episodes (20% of those receiving an intervention, and 12% of all spontaneous pneumothorax episodes) (table 2). Aspiration was successful in 14/27 treatments (52%, 95% CI 34% to 69%). In those where aspiration was unsuccessful, in one case repeat aspiration was also unsuccessful, and 11 had an ICC inserted. One other child was discharged home from ED with a small pneumothorax unchanged in size from presentation (before aspiration). For PSP cases, the duration of symptoms at presentation did not differ between successful and unsuccessful aspiration.

    Use of aspiration in PSP management as a total number or percentage of interventions over the course of the study did not increase (see online supplementary figure S1). Across all network sites, aspiration as first-line therapy for PSP varied from 0% to 48%. Aspiration was rarely performed in young children, being used in only 1/23 (4%) aged <13 years. CXR examination showed a small pneumothorax in 86/130 (66%) cases (figure 2). Aspiration was used in only four SSP episodes despite 30/64 (47%) subjects with SSP being classified as having a small SSP. Aspiration success was 2/4, and in the two SSP subjects with a large SSP was unsuccessful on both occasions. For PSP alone, there was no significant difference in the success rate of aspiration in small versus large sized pneumothoraces (6/8 vs 4/10, p=0.13). Combining PSP and SSP subjects, the difference in aspiration success rates between small and large pneumothoraces was significant (8/10 vs 4/12, p=0.014).

    Figure 2

    Summary of management of primary spontaneous pneumothorax episodes where an objective assessment of pneumothorax size was available. ICC, intercostal chest catheter.

    Size of ICC used

    Information on ICC size was available in 76/117 (65%) cases with an ICC. Small-bore ICC were used in 30/76 (39%) cases overall (PSP vs SSP: 15/28 vs 15/48, p=0.09). The proportion of small-bore ICC used increased from year to year during the study, from 0% in 2003 to 75% in 2010. The remaining drains were all medium-bore apart from one large-bore ICC inserted into a 15-year-old presenting with a first PSP episode. There was no difference in length of stay (median (range)) between PSP children receiving small- and medium/large-bore ICC (excluding those who had surgery before discharge): 6 (4–19) vs 6 (2–21) days.

    Pneumothorax recurrence

    The overall recurrence rate was 95% CI (24% to 39%) (figure 1), and identical for PSP and SSP over very similar follow-up periods (mean (SD) 3.9 (2.1) vs 3.9 (2.2) years). For PSP, there was one recurrence in 65% (24/37) of children, two in 24% (9/37), three in 5% (2/37), and the two remaining children had four and five recurrences. The median interval (range) between presentations was 62 (7–930) days. Recurrence occurred on the same side in 30/37 (81%) cases. Of those in whom previous surgery had been documented (12/37, 32%), recurrence occurred on the same side as previous surgery in nine (75%). For subjects with SSP, the majority again had a single recurrence during the study: one recurrence in 54% (7/13) children, two in 23% (3/13), whilst the three remaining children had three, four and eight recurrences. Recurrence occurred on the same side in 92%, and of those who had had previous surgery (5/13), on the same side as in the previous surgery in 40%. Across the entire cohort, CT scans had been performed in 76/219 (34.7%). Blebs/bullae were documented on CT in 46% of these scans (table 2).


    This multicentre study of 219 episodes of paediatric spontaneous pneumothorax in the Australia and New Zealand PREDICT network13 has three important observations. First, management is different from management in adults according to current guidelines.7 For PSP, ICC insertion continues to be our first-line intervention. Second, in comparison with PSP, episodes of SSP have a greater rate of active intervention, ICC insertion and longer duration of admission. Third, aspiration use was low, did not differ between PSP and SSP and did not increase during the study. Aspiration success rates were higher in small, than in large, pneumothoraces.

    Data from a US hospital administrative database (∼7000 children) has highlighted the epidemiology of pneumothorax in children, such as its increasing incidence and the high intervention rate,11 but these data are unable to inform interventions or compliance with guidelines. In our study we examined the interventions. Over the 7 years since the BTS guidelines,7 ICC remains the most common intervention used, with aspiration chosen for only 27% of PSP and 9% of SSP first-line interventions. Higher ICC rates for SSP were not surprising given the reported lower success rates of aspiration in SSP.6 We found aspiration use was higher in paediatric PSP (although not statistically significant), that management differed from BTS guidelines and that management was similar for PSP and SSP. Our study also shows that respiratory physicians are not commonly involved in decision-making in these patients, yet current guidelines originate from their professional organisation.

    We found that aspiration was successful in 52% of episodes, which is at the lower end of that quoted in adults (50–83%),17 ,18 in whom aspiration is considered less invasive, more cost-effective and associated with fewer complications.19 Choice of intervention was not influenced by PSP size. Consistent with most,20–23 but not all,24 ,25 literature in adult cases, we demonstrated a higher aspiration success rate in small spontaneous pneumothoraces. If this observation is confirmed in future studies, this new finding has implications for when aspiration should be used in treatment.

    The demographic findings in our population were similar to those previously reported; a predominance of male adolescents.4 ,11 ,26 The majority (55%) of children with PSP managed conservatively were admitted to hospital from the ED. This practice differs from recent adult practice, where outpatient management is increasing.1 Recurrence rates in our study mirrored adult reports (∼30%),27 and were far lower than those reported in small paediatric case series (up to 61%).9 ,28 The relationship between blebs/bullae and subsequent recurrence risk remains controversial. The prevalence of blebs/bullae on CT was lower in our series than in adult literature (up to 88%).26 ,29 ,30

    Our study has a number of limitations. (1) It is retrospective. (2) In one-third of cases there was no diagnostic CXR available, although 84% of episodes could be classified as small or large, with information from the medical notes. (3) The definition of pneumothorax size reflects recommendations in adults and may not be appropriate for children. The focus of this paper was whether BTS guidelines6 ,7 were being followed, and the sizing definitions used are beyond the scope of this manuscript. However, the lack of adherence was present across the paediatric age range, including adolescents where size definitions may be more applicable. (4) Our study was multicentre and spanned 7 years, but few young children presented with pneumothorax. This low rate limits the generalisability of the findings to young children. (5) The data abstractors were not blinded to the study's purpose, but we did follow guidelines for chart review.14 (6) Aspiration success was based on avoidance of subsequent ICC and any improvement in pneumothorax size.19 These criteria can lead to an overestimation of success. (7) Only two EDs had a management protocol during the study: Royal Children's Hospital, Melbourne (from 2004) and Children's Hospital at Westmead, Sydney (from 2009), accounting collectively for 17/23 (74%) of all aspiration cases.

    In conclusion, our study indicates that in contemporary paediatric practice, management of spontaneous pneumothorax does not follow the BTS guidelines.6 ,7 This finding was evident both for the adolescent cohort alone and across the wider paediatric age range. In particular, ICC remains the preferred intervention, with relatively low success rates of aspiration when used. Based on these data, appropriately designed prospective paediatric multicentre studies are urgently required to clarify the utility of aspiration in children presenting with spontaneous pneumothorax and guide future paediatric-specific guideline recommendations.


    The authors acknowledge the assistance of Gilad Chayen MD (Children's Hospital at Westmead) in data collection and Vi Chong (Royal Children's Hospital Melbourne) in the initial assessment of internal quality control for this project.


    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

      Files in this Data Supplement:


    • Collaborators The PREDICT network hospitals and investigators in Australia: Children's Hospital at Westmead (Paul Robinson, Peter Ngo), Royal Children's Hospital Melbourne (Franz Babl, Carol Blackburn), Children's and Women's Hospital Adelaide (Lalith Gamage, Jacquie Schutz), Sydney Children's Hospital (Rebecca Nogajski, Liz Cotterill), Sunshine Hospital Melbourne (Dino Druda, David Krieser), Royal Children's Hospital Brisbane (Jason Acworth), and John Hunter Children's Hospital, Newcastle (Mark Lee); and in New Zealand, Starship Children's Hospital (Stuart Dalziel, Colin Donald) and Kidzfirst Children's Hospital in Auckland (Jocelyn Neutze, Alan Drage).

    • Contributors All the authors were involved in the study design, data collection and writing of the manuscript. PDR and PKN collated the data centrally from the centres involved in this study and performed the data analysis.

    • Competing interests None.

    • Ethics approval The study was approved by the ethics committee at each of the sites involved in the study.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Linked Articles