OBJECTIVES: The purpose of this study was to examine clinical and echocardiographic predictors of outcome in a cohort of patients with Marfan's syndrome. BACKGROUND: Serial echocardiographic measurements of aortic root dimensions are an important clinical method for monitoring patients with Marfan's syndrome. However, there are few data on the prognostic importance of echocardiographic variables for risk stratification and timing of aortic root replacement. METHODS: In 89 consecutive patients with Marfan's syndrome (age range 1-54 years) clinical and serial echocardiographic data (n = 62) were evaluated as potential predictors of outcome (mean (range) follow up 4 (< 1-16) years). Aortic sinus diameter and an aortic ratio normalised for age and body surface area were examined using Kaplan-Meier life table and Cox regression analysis, with the end point defined as death or surgery for ascending aortic dissection and events defined as an end point or surgery for ascending aortic aneurysm. RESULTS: Overall actuarial survival at two and five years was 96% and 92% and event free survival was 85% and 76%, respectively. There were five deaths due to aortic dissection, four patients survived surgery for ascending dissection, and nine underwent root replacement with a composite graft for ascending aneurysm. Those with aortic events were older (35 (12) v 25 (13) years, P = 0.007) and had greater initial aortic root dimensions (47 (14) v 33 (8) mm, P < 0.0001) and ratios (1.6 (0.5) v 1.3 (0.2), P < 0.0001). In the 62 patients with serial echocardiographic follow up, the rate of aortic root dilatation was more rapid in those with events (15 (17) v 0 (3)%/year, P < 0.0001). Utilising a Cox proportional hazards model, the groups with an initial aortic ratio > or = 1.3 or an annual change in aortic ratio > or = 5% had a relative risk of an aortic complication of 2.7 and 4.1, respectively (95% confidence limits 1.5 to 4.8 and 1.8 to 9.3). Only one of 31 patients with an initial aortic ratio of < 1.3 and a rate of change of < 5% had an event (five year event free survival 97%). CONCLUSIONS: A low risk subgroup of patients with Marfan's syndrome can be identified as those with an aortic ratio < 1.3 and an annual change in aortic ratio of < 5%. These findings are helpful in optimising echocardiographic monitoring and risk stratification of patients with Marfan's syndrome.