Evaluation of the safety and effectiveness of an accelerated treatment protocol for traumatic haemothorax and pneumothorax: a prospective, multicentre observational study in South Africa.

BACKGROUND

Penetrating pneumothorax and haemothorax are common injuries seen in South African trauma centres. Hospitals in the Western Cape Province of South Africa utilize a protocol for the expedited management of patients with chest injuries requiring intercostal drains (ICD) within emergency department observation units, with emphasis on early mobilization and pulmonary rehabilitation. The study objective was to evaluate the safety and effectiveness of the Western Cape ICD protocol for unilateral traumatic pneumothorax, haemothorax, or pneumo-haemothorax.

METHODS

This was a prospective, observational, multi-centre cohort study that enrolled adult (≥ 18 years) patients within four public hospitals (primary, secondary, tertiary) in the Western Cape Province from 2024 to 2025. Patients were eligible for inclusion if they received ICD protocol care at any of the participating hospitals. We performed a descriptive analysis of patient injury profiles, complications, hospital length of stay, in-hospital mortality and 30-day readmissions. The primary outcome was hospital length-of-stay with a hypothesis that > 50% of patients will be discharged ≤ 72 h.

RESULTS

497 patients were enrolled, with 92% male, and median age 29.2. Most (87.7%) were injured by stabbing/cutting mechanisms, and 53.5% had a haemothorax. There were procedural complications among 10.3% of cases (n = 51), including deep placement (n = 25), shallow placement (n = 8), kinked ICD (n = 8), sub-optimal intrapleural placement (too cephalad/caudad; n = 4) and extra-pleural location (n = 4). There were delayed complications among 7.6% of the cohort (n = 38). Common delayed complications included ICD dislodgment (n = 9), premature ICD removal (n = 6), retained haemothorax (n = 6), and infection (n = 15, including 3 empyema, 5 pneumonia, and 9 superficial ICD site infection). Hospital length of stay was a median of 3 days (IQR 2-4). There were no in-hospital deaths, and only 14 patients (2.8%) were readmitted within 30 days.

CONCLUSION

Our cohort had a composite complication rate of 16.7% at the patient-level, combining malpositioning and delayed complications, suggesting that the Western Cape protocol is safe compared to standard of care for the treatment of simple traumatic pneumothorax and haemothorax.