Background: Pneumothorax is a rare complication of acupuncture and the risk factors are unclear. Objective: This study analysed the incidence of post-acupuncture pneumothorax requiring hospitalisation in a one-million-sample cohort derived from Taiwan’s National Health Insurance Research Database. Methods: We tracked this cohort between 1997 and 2012 and recorded all medical insurance information. Subjects were categorised according to gender, insurance amount, comorbidities, residential area, and number of acupuncture treatments. Pneumothorax risk was evaluated according to different demographic and medical variables by logistic regression analysis using an adjusted odds ratio (aOR) with a 95% confidence interval (95% CI). Results: Overall, 411 734 patients undergoing 5 407 378 acupuncture treatments were identified with data collected over the first 7 days after acupuncture. The incidence rates of iatrogenic pneumothorax were 0.87 per 1 000 000 acupuncture treatments overall and 1.75 per 1 000 000 acupuncture treatments in “at-risk” anatomical areas. Multivariate logistic regression demonstrated that a history of thoracic surgery (aOR 7.85, 95% CI 3.49 to 9.25), chronic bronchitis (aOR 2.61, 95% CI 1.03 to 6.87), emphysema (aOR 4.87, 95% CI 1.03 to 7.96), pneumonia (aOR 2.09, 95% CI 1.44 to 2.72), tuberculosis (aOR 3.65, 95% CI 1.39 to 9.56), and lung cancer (aOR 3.85, 95% CI 1.53 to 9.73) may increase the post-acupuncture risk of iatrogenic pneumothorax. Men had a higher risk of pneumothorax than women (aOR 3.41, 95% CI 1.36 to 8.57). The number of treatments was not associated with risk of pneumothorax. Conclusions: Patients with a history of lung disease including chronic bronchitis, emphysema, tuberculosis, lung cancer and pneumonia, and a history of thoracic surgery, might have an increased post-acupuncture risk of pneumothorax. This information may possibly help physicians avoid post-acupuncture pneumothorax.