Abstract

Background: Cardiovascular disease (CVD) prevention guidelines typically dichotomise patients by history of CVD, as patients with prior CVD are assumed to be at high CVD risk, whatever their CVD risk profiles.

Objective: To assess the appropriateness of this practice by comparing CVD event rates of patients with and without prior CVD, over and above risk predicted by standard CVD risk factors.

Methods: Between 2002 and 2007 CVD risk assessments were generated using a web-based Framingham risk prediction algorithm in routine primary care. Individual risk profiles were subsequently linked to national hospitalisation and death records. Observed and predicted (Framingham) CVD risk were compared in patients with and without prior CVD.

Results: 35 760 patients were assessed including 10.4% with prior CVD. Of 1216 first CVD events during an average follow-up of 2.05 years, 42% occurred in those with prior CVD. Among those without prior CVD, the predicted Framingham five-year CVD risk was similar to the observed risk extrapolated to five years; in the highest Framingham risk band (>20% five-year risk), observed risk was 25.3%. Among those with prior CVD the observed risk extrapolated to five years rose from 21.7% in the lowest Framingham risk band (<5%) to 49% in the highest (>20%).

Conclusions: Patients with prior CVD have five-year CVD risks approximately 20% higher, in absolute terms than patients without prior CVD, after accounting for standard risk factors. Almost half the CVD events occurred in those with prior CVD. These patients should be the highest priority for intensive preventive management in primary care.