Introduction Bronchial artery embolisation (BAE) is recommended as the therapy of choice for massive haemoptysis in cystic fibrosis (CF) but there are no randomised controlled trials of BAE in this setting. Outcomes from BAE are uncertain and the efficacy of BAE in sub-massive haemoptysis is unclear. We performed a single-centre observational study to investigate the role of BAE in CF-related haemoptysis.

Methods All patients with CF undergoing BAE from March 2011 to January 2015 were identified at the time of the procedure. Patient records were reviewed following hospital discharge or death. Severity of haemoptysis was classified as: massive (>240 ml/24 h or >100 ml/day for ≥2 days), severe (>20 ml/24 h) or mild (<20 ml/24 h). Data were collected on adjuvant therapies, time to recurrence, complications and survival.

Results Twenty-seven patients underwent 49 BAE procedures. Median age was 30 years (range 18–72) and 16 (59%) were male. Mean baseline FEV­₁%-predicted was 51.0% (SD 19.3). BAE was indicated for massive haemoptysis in 18 episodes (37%), severe in 27 (55%) and mild in 4 (8%). Adjuvant therapies included tranexamic acid in 48 (98%), intravenous antibiotics in 47 (96%), intravenous vitamin K in 31 (63%), ethamsylate in 8 (16%), terlipressin in 7 (14%) and propranolol in 5 (10%) of episodes.

Twenty-nine (59%) BAEs were complicated by adverse events including chest pain (29%), headache (12%), paraesthesia (10%), groin pain (6%), limb weakness (4%) and limb ischaemia (2%). Eight patients (30%) required ≥2 BAEs during the study (range 2–7). Median time to first repeat BAE was 213 (range 18–682) days. Overall, haemoptysis recurred after 31/49 (63%) procedures with no significant difference between massive and sub-massive haemoptysis (61.1% vs 64.5%).

Five patients (18.5%) died during the study and this group had a median FEV₁%-predicted of 32% (range 28–82%). Mortality was 3.7% at 30 days following first BAE and 11.1% at six months. Four out of 8 (50%) patients requiring repeat BAE died compared with 1/19 (5%) who needed a single BAE only (p = 0.006).

Conclusion BAE may be life-saving but is associated with considerable morbidity in CF. Need for repeat BAE is associated with increased mortality.