BACKGROUND: Few prospective studies of the incidence and outcome of paediatric in-hospital cardiopulmonary arrest have been reported to enable quality assurance comparisons within and between institutions. METHODS: All cardiac and respiratory arrests and their management over a 41-month period in children not subject to palliative treatment or to a ‘do not resuscitate’ order were recorded and analysed using the Utstein template. RESULTS: Cardiac arrest occurred in a total of 111 of 104,780 admissions (1.06/1000) while respiratory arrest alone occurred in 36 (0.34/1000). Return of spontaneous circulation (ROSC) was achieved in 81 patients (73%) in cardiac arrest but only 40 (36%) were discharged from hospital and 38 (34%) survived for 1 year. The 1-year survival from respiratory arrest alone was 97%. Cardiac arrest was four times more common (89 versus 22) and approximately 90 times the incidence in the intensive care unit compared with wards but 1-year survival was similar (34% versus 36%). The initial heart rhythms were hypotensive-bradycardia in 73 (66%) with 38% survival; asystole in 17 (15%) with 12% survival; ventricular fibrillation (VF) or pulseless ventricular tachycardia (VT) in 10 (9%) with 40% survival; pulseless electrical activity (PEA) in 10 (9%) with 30% survival and SVT in 1 with survival. Secondary ventricular fibrillation occurred in 15 children given adrenaline (epinephrine) for treatment of asystole, hypotensive-bradycardia or PEA of whom 11 had received adrenaline in an initial dose of > 15 mcg/kg and 4 had < 15 mcg/kg (P = 0.0013). Eleven of 15 patients (73%) in secondary VF never achieved ROSC. CONCLUSIONS: In-patient paediatric cardiac arrest has a mediocre outcome with a better outlook if the initial rhythm is hypotensive-bradycardia, VF or pulsatile VT. Doses of adrenaline greater than 15 mcg/kg given for non-shockable rhythms may cause secondary VF which has a worse outcome than primary VF.