Utilization and Outcomes of Observation for Spontaneous Pneumothorax at an Integrated Health System

Introduction: Though limited, recent evidence supports observation rather than intervention for spontaneous pneumothorax management. We sought to compare the utilization and outcomes between observation and intervention for patients with primary and secondary spontaneous pneumothoraces.

Methods: A retrospective cohort study of all adults presenting to Kaiser Permanente Northern California emergency rooms with spontaneous pneumothorax from 2016 to 2020 was performed. Those with prior pneumothoraces, tension physiology, bilateral pneumothoraces, effusions, and prior thoracic procedures or surgery on the affected side were excluded. Groups included observation versus intervention. Baseline clinicodemographic variables and outcomes were compared. Treatment was considered successful if further interventions were not required for pneumothorax resolution. Wilcoxon rank-sum tests, chi-square tests, Fischer exact tests, and multivariable logistic regression models were performed.

Results: Of the 386 patients with primary spontaneous pneumothorax, age, race/ethnicity, body mass index, smoking status, and the Charlson comorbidity index were not different between treatment groups. Of 86 patients with secondary spontaneous pneumothorax, age, gender, and smoking status were not different between treatment groups. Among patients with primary pneumothoraces, 83 underwent observation while 303 underwent intervention. The success rate was 92.8% for observation and 60.4% for intervention (P < 0.0001). Among patients with secondary pneumothoraces, 15 underwent observation while 71 underwent intervention, with a successful rate of 73.3% for observation and 32.4% for intervention (P = 0.003).

Conclusions: Given the high success rates for observation of both small and moderate primary and secondary pneumothoraces, observation should be considered for clinically stable patients. Observation may be the superior choice for decreasing morbidity and healthcare costs.