Background Kawasaki disease (KD) is the commonest cause of paediatric acquired heart disease in the Western world. The Kobayashi score has been shown not to be a reliable indication of disease severity for the UK population and such scoring is needed to predict which children will benefit from adjunct therapy at presentation.

Methods The recent survey (BPSU) ran from February 2013 to February 2015 covering the UK and Ireland, including questions relating to date of symptom appearance. We determined the link between specific symptoms, in those where the date of appearance of each symptom was recorded and cardiac disease. “Coronary artery abnormality” (CAA) was any abnormality of coronary arteries whereas “Cardiac involvement” also included pericardial effusion, valve regurgitation or myocarditis.

Results 291 children fulfilled the inclusion criteria for this substudy. CAA rate was 20(±5.5)% for all complete KD (n = 208), but for those with all 5 symptoms at diagnosis, it was just 12.5 (±6.57)%. For those with just 4 symptoms (n = 144), the CAA rate was 23.6(±6.94)%. If the mucositis was absent, CAA rate was 37.5% (n = 8). If mucosa or extremity changes were present (n = 50), there were significantly fewer CAA (12.5% p < 0.05, Chi squared). At diagnosis, if lymphadenopathy, conjunctival involvement or rash were present, then CAA rate was higher (28%). Chi square for trend suggests significantly less cardiac involvement the greater the number of key symptoms present at diagnosis.

For incomplete KD (n = 55), the cardiac involvement rate was 5.45(±6.0)%. Atypical KD with CAA (n = 28), was seen predominantly in those <1 year and 12% had additional cardiac involvement. Most had mucositis if there were no CAA or cardiac involvement. The least likely symptoms were non-purulent conjunctivitis or rash (p < 0.05, chi squared).

Conclusions In this largest worldwide population series of Kawasaki disease, the rate of cardiac involvement remains extremely high. Those with mucosal involvement are least likely to have cardiac disease. Those without mucosal involvement or extremity changes have a much higher rate of cardiac involvement. This suggests that there may be significant underdiagnosis of Kawasaki disease in the UK, especially if mucosal or extremity changes are not present.