RESUMO
INTRODUCTION: Determining the risk of recurrence of primary spontaneous pneumothorax is challenging. The objective of this study was to develop a risk assessment model to predict the probability of recurrence in patients with spontaneous pneumothorax. METHODS: A retrospective study was performed of all episodes of pneumothorax diagnosed in the last 12 years in a hospital, in patients not initially submitted to surgery. Logistic regression was used to estimate the probability of recurrence. Based on a set of variables, a predictive model was built with its corresponding ROC curve to determine its discrimination power and diagnostic precision. RESULTS: Of the 253 patients included, 128 (50.6%) experienced recurrence (37% within the first year). Recurrence was detected within 110 days in 25% of patients. The median of time to recurrence for the whole population was 1120 days. The presence of blebs/bullae was found to be a risk factor of recurrence (OR: 5.34; 95% CI: 2.81-10.23; p=0.000), whereas chest drainage exerted protective effect (OR: 0.19; 95% CI: 0.08-0.40; p=0.000). The variables included in the regression model constructed were hemoglobin and leukocyte count in blood, treatment received, and presence of blebs/bullae, with a fair discriminative power to predict recurrence [AUC=0.778 (95% CI: 0.721-0.835)]. CONCLUSION: The overall recurrence rate was high and was associated with the presence of blebs/bullae, failure to perform an active intervention (chest drainage) and low levels of hemoglobin and leukocytes in blood. Recurrence rarely occurs later than three years after the first episode. Once validated, this precision model could be useful to guide therapeutic decisions.