In the context of a significant increase in obesity rates, quantifying the relationship between body mass index (BMI) and risk of pulmonary embolism (PE) is an essential component of accurate clinical risk assessment. This observational study is the first to explore this association by clinician-defined cause of the PE. We demonstrate that the association between BMI and PE is driven by patients with otherwise ‘unprovoked’ PE where there is a strong positive correlation with odds ratios equivalent to well-recognised major risk factors such as cancer, pregnancy and surgery. We make a case for the inclusion of BMI in risk-prediction tools.