Conservative versus Interventional Treatment for Spontaneous Pneumothorax

Conservative versus Interventional Treatment for Spontaneous Pneumothorax

Bottom line

Conservative management of stable unilateral, spontaneous primary pneumothorax, is not inferior to Seldinger drain, with lower risk of serious adverse events.

Majority of patient with unilateral primary spontaneous pneumothorax is safe for conservative management and discharge at 4 hrs, after a repeat CXR

High fragility index due to lost to follow-up means data is not statistically robust

Study Question

Whether conservative management is an acceptable alternative to interventional management for uncomplicated, moderate-to-large primary spontaneous pneumothorax is unknown.

Study Design

  • Open labeled, prospective, randomized, noninferiority trial
  • -9% as noninferiority margin (90% vs estimated 99% resolution in intervention group)


  • Multicenter. 39 metropolitan and rural hospitals from Australia and New Zealand
  • 316 patients (154 to intervention arm vs 162 to conservative arm)
  • Patient 14 to 50 yo

Inclusion criteria

  • Unilateral pneumothorax. Must be primary and spontaneous
  • 32% or more using Collins method

Exclusion criteria

  • Previous ipsilateral primary spontaneous pneumothorax
  • Secondary pneumothorax (acute trauma, asthma on preventive medications or symptoms in preceding 2 years)
  • coexisting hemothorax
  • bilateral pneumothorax
  • “tension pneumothorax” (SBP <90 mmHg, MAP <65 mmHg, shock index HR/ SBP >1)
  • pregnancy at time of enrolment
  • social circumstances (inadequate support after discharge to re-attend hospital if required or unlikely to present for study follow-up)
  • Planned air travel within the following 12 weeks


All patient received standard care

  • analgesia (panadol, ibuprofen, oral or IV opioids), oxygen if SpO2 <92% on room air

Intervention arm

    • small bore (<12 Fr) Seldinger-style chest tube, connected to underwater seal (no active suction).
    • Repeat CXR 1 hr later
    • If lung re-expanded, underwater drain not bubbling, drain is clamped
    • Repeat CXR 4 hr later
    • Drain is removed If pneumothorax did not recur, and patient discharged
    • Patient is admitted with same Seldinger drain open to underwater seal, if:
      • no radiographic resolution of pneumothorax, or
      • pneumothorax recurred under observation
      • These patient is subsequently managed at discretion of admitting doctor

Conservative-management arm

    • Pt is observed for minimum of 4 hrs. before repeat CXR
    • Discharged with instruction and analgesia, if:
    • not needing oxygen supplement
    • walking comfortably
    • Intervention (drainage) is allowed if:
    • significant symptoms persist despite analgesia
    • Chest pain or dyspnea prevented mobilization
    • Pt unwilling to continue with conservative treatment
    • Pt became unstable (SBP <90 mmHg, HR became greater or equal to SBP, RR >30, SpO2 <90% on room air), or
    • Repeat CXR showed enlarging pneumothorax with physiological instability

Follow up

  • Follow up at assessment at 24, 72 hrs, and at 2 weeks, 4 weeks and 8 weeks visits.
  • Repeat CXR at follow-up visits
  • Questionnaire regarding symptoms, analgesia use, patient satisfaction
  • Telephone call at 6 months and 12 months after randomization for pneumothorax recurrence


  • from July 2011 to March 2017
  • 316 patients underwent randomization (WA sites are Armadale/ Freemantle/ FSH/ SJOG Midland/ RPH/ SCGH)
  • 154 (intervention group) vs 162 (conservative-management group)
  • Baseline character  (table 1) similar except
  • more male in conservative arm (88% vs 84%)
  • Intervention arm longer symptoms duration (44 hr vs 34 hr), larger pneumothorax (68% vs 64%), more tobacco smoker and more pack years
  • These data suggest some problem with randomization in my opinion due to design(* open label study)

Primary outcome is resolution of pneumothorax at follow-up

  • In intervention arm 129/131 (98.5%) within 8 weeks vs 118/125 (94.4%) in conservative-management arm (risk difference -4.1%, 95% CI -8.6 to 0.5, p=-0.02 for non inferiority). This is within the estimated -9% point.
  • At 63 days, intervention arm 134/136 (98.5%) vs conservative arm 129/136 (94.9%) (risk difference -3.7%, CI -7.9 to 0.6)
  • However, if missing patients at 56 days were assumed to be non resolution (treatment failure) – 129/138 (93.5%) vs 118/143 (82.5%) (risk difference -11%, 95% CI -18.4 to -3.5).

Secondary outcome (table 2)

  • Radiographic resolution was 16 days in intervention group vs 30 days in conservative group (HR 0.49, 95% CI 0.39 to 0.63)
  • By 8 weeks, interventional group has 124/126 (98.4%) had resolution vs 123/130 (94.6%) in conservative group (risk difference -3.8%, 95% CI -8.3 to 0.7)
  • Pt needing 1 or more procedures was 145/154 in interventional group (94.2%) vs 25/162 (15.4%) in conservative group
  • Adverse event rate was 41/154 (26.6%) interventional arm vs 13/162 (8%) in conservative arm. Serious adverse event 19/154 (12.3%) in intervention arm vs 6/162 (3.7%) in conservative arm (relative risk 3.30, 95% CI 1.85 to 5.95)
  • Median length of hospital stay was 3.8 days in intervention arm vs 0.2 days in conservative arm.


  • Study design may favor intervention group as treating clinician most likely to report resolution of pneumothorax in this group.
  • Still no data to tell us if patient can be sent home from ED without 4 hr admission and repeat CXR, if they are stable!, and who these patients are.