The cumulative effects of historical and physical examination findings on the prognostic value of the electrocardiogram

The electrocardiogram has been shown in epidemiologic studies to be an inde pendent predictor of survival. These studies have adjusted for selected cov ariates simultaneously. This article assesses the value of the electrocardi ogram as a predictor of survival when introduced at progressive stages of t he common clinical encounter. Data collected from 4,518 patients ages of 60 to 96 years with isolated systolic hypertension who were followed-up for u p to 6 years as part of the Systolic Hypertension in the Elderly Project we re analyzed. Survival curves and 3- and 5-year survival races (Cox regressi on methods) of groups with normal and abnormal resting electrocardiograms w ere compared. Blocks of covariates representing demographic information, ri sk factors for cardiovascular disease, clinical history, and physical exami nation findings were added to the survival models sequentially to mimic the sequence of the common clinical encounter, and the independent significanc e of the electrocardiogram as a predictor of survival was assessed at each step. An abnormal electrocardiogram was associated with reduced survival wh en no adjustment for covariates is made. Survival was also significantly (P < .05) different for groups with normal and abnormal electrocardiograms wh en demographic and risk factor variables were included in the statistical m odels, but not after findings from the clinical history and physical examin ation were added. The prognostic value of the electrocardiogram varies with the stage in the clinical encounter in which it is introduced.