Antonella Zambon1, Andrea Arfè, Giovanni Corrao, Alberto Zanchetti. 1. aDepartment of Statistics and Quantitative Methods, Division of Biostatistics, Epidemiology and Public Health, University of Milano Bicocca bIstituto Auxologico Italiano, University of Milan, Milan, Italy.
Abstract
BACKGROUND AND OBJECTIVE: Guidelines for management of cardiovascular diseases stratify absolute cardiovascular risk into categories with a high-risk threshold defined at a 20% cardiovascular events risk in 10 years, but it is unclear whether only major events or the Framingham-extended definition should be considered. The 2013 ESH-ESC hypertension guidelines, instead, define cardiovascular risk as a risk of cardiovascular death in 10 years, as in the SCORE model, setting the threshold for high risk at the 5% level. It would be therefore convenient to know the quantitative relationship between the risks of the different outcomes adopted by the different guidelines, especially because some outcome definitions include serious nonfatal cardiovascular events relevant in cardiovascular prevention. We have therefore analysed these relationships in trials of antihypertensive therapy as an aid to defining total cardiovascular risk in hypertensive patients. DESIGN AND METHODS: Sixty-one trials were identified, and 51 retained for analysis of the relationship of cardiovascular death to the incidence of all-cause death, major cardiovascular events and inclusive (Framingham) cardiovascular events. The relationship between cardiovascular death rates and each type of event rates was explored by fitting flexible regression models. RESULTS: The included trials provided 15164 cardiovascular deaths and 1674427 patient-years. The relation of each event rate to cardiovascular death rate was best explained by a model considering the logarithm of each event rate as a dependent variable and the logarithm of cardiovascular death rate as a predictor. Mean patients' age and treatment were also predictors, but to a minor extent. The increase of the incidence rates of all types of events was less steep the higher the CV death rate: the rate ratios of all-cause death to cardiovascular death were 2.2, 1.9 and 1.8 at low-moderate (cardiovascular death <5% in 10 years), high (cardiovascular death 5% to <10%) and very high risk (cardiovascular death ≥ 10%), respectively; the rate ratios of major cardiovascular events to cardiovascular death were 3.9, 2.7, 2.3 and the rate ratios of extended cardiovascular events to cardiovascular death were 8.4, 5.6 and 4.6, respectively. Ratios only slightly changed when 12 trials of secondary prevention were excluded. CONCLUSION: Ratios of various event rates to cardiovascular death rate vary with cardiovascular disease severity, cardiovascular mortality representing an increasing proportion of total cardiovascular risk when the former is higher. From the models, a total risk can be estimated in groups of hypertensive patients whose cardiovascular death risk is calculated by the SCORE model.
BACKGROUND AND OBJECTIVE: Guidelines for management of cardiovascular diseases stratify absolute cardiovascular risk into categories with a high-risk threshold defined at a 20% cardiovascular events risk in 10 years, but it is unclear whether only major events or the Framingham-extended definition should be considered. The 2013 ESH-ESC hypertension guidelines, instead, define cardiovascular risk as a risk of cardiovascular death in 10 years, as in the SCORE model, setting the threshold for high risk at the 5% level. It would be therefore convenient to know the quantitative relationship between the risks of the different outcomes adopted by the different guidelines, especially because some outcome definitions include serious nonfatal cardiovascular events relevant in cardiovascular prevention. We have therefore analysed these relationships in trials of antihypertensive therapy as an aid to defining total cardiovascular risk in hypertensivepatients. DESIGN AND METHODS: Sixty-one trials were identified, and 51 retained for analysis of the relationship of cardiovascular death to the incidence of all-cause death, major cardiovascular events and inclusive (Framingham) cardiovascular events. The relationship between cardiovascular death rates and each type of event rates was explored by fitting flexible regression models. RESULTS: The included trials provided 15164 cardiovascular deaths and 1674427 patient-years. The relation of each event rate to cardiovascular death rate was best explained by a model considering the logarithm of each event rate as a dependent variable and the logarithm of cardiovascular death rate as a predictor. Mean patients' age and treatment were also predictors, but to a minor extent. The increase of the incidence rates of all types of events was less steep the higher the CV death rate: the rate ratios of all-cause death to cardiovascular death were 2.2, 1.9 and 1.8 at low-moderate (cardiovascular death <5% in 10 years), high (cardiovascular death 5% to <10%) and very high risk (cardiovascular death ≥ 10%), respectively; the rate ratios of major cardiovascular events to cardiovascular death were 3.9, 2.7, 2.3 and the rate ratios of extended cardiovascular events to cardiovascular death were 8.4, 5.6 and 4.6, respectively. Ratios only slightly changed when 12 trials of secondary prevention were excluded. CONCLUSION: Ratios of various event rates to cardiovascular death rate vary with cardiovascular disease severity, cardiovascular mortality representing an increasing proportion of total cardiovascular risk when the former is higher. From the models, a total risk can be estimated in groups of hypertensivepatients whose cardiovascular death risk is calculated by the SCORE model.