The main cause of death of cystic fibrosis (CF) patients is respiratory disease due to secondary infections, haemoptysis, pneumothorax, and respiratory failure [1]. In a study of 129 French CF patients who died between 2007 and 2010 without receiving lung transplantation, 8.5% of deaths were due to haemoptysis [2]. Factors associated with an increased risk of haemoptysis in CF patients include older age, advanced lung disease (forced expiratory volume in 1 s (FEV 1) <70% predicted), airway colonisation by Pseudomonas aeruginosa [3], CF-related diabetes, portal hypertension, and liver cirrhosis [4]. To our knowledge, there have been no studies of the risk factors for the development of massive haemoptysis (MH) in CF patients who have previously experienced an episode of mild-to-moderate haemoptysis. We retrospectively analysed data from all adult CF patients seen at the Lille CF Centre between 2008 and 2018 who had a history of at least one mild-to-moderate haemoptysis episode. All individuals gave written informed consent. Approval for the use of these data was provided by the Institutional Review Board of the French Learned Society for Pulmonology (CEPRO 2015-004). The patients were followed up from the time of mild-to-moderate haemoptysis (or from the first episode for those with >1 mild-to-moderate haemoptysis) until they experienced an MH event ("with MH" group) or until the end of follow-up ("without MH" group). Clinicopathological data were collected from medical records. Mild-to-moderate haemoptysis was defined as an episode resulting in 5-240 mL externalised blood. MH was defined as: 1) >240 mL externalised blood in 24 h, 2) >100 mL externalised blood over ⩾2 days [5], or 3) any externalised blood loss causing abnormal gas exchange/airway obstruction or haemodynamic instability [6]. Haemoptysis was treated according to the CF Foundation Consensus Conference Report [7], and patients with MH received concomitant therapy with intravenous (i.v.) antibiotics. Long-term drug therapy was defined as medication prescribed for >3 months. The cumulative incidence of MH was estimated using the Kaplan-Meier method by censoring patients lost to follow-up at the last available visit. The median follow-up time from first mild-to-moderate haemoptysis episode was calculated using the reverse Kaplan-Meier method. Associations between baseline characteristics and MH occurrence were analysed using univariable Cox proportional hazard regression models. The log-linearity assumption for continuous factors was checked by examining Martingale residual plots and using restricted cubic spline functions, and the proportional hazard assumption for each potential predictive factor was checked by examining the Schoenfeld residuals. Mild-to-moderate haemoptysis recurrence rates were calculated as the number of episodes per 100 person-months from the time of first mild-to-moderate haemoptysis episode to MH or, for patients without MH, to the last annual review. All statistical tests were performed at a two-tailed α level of 0.05. A total ...