Left ventricular ejection fraction (EF) is a valuable prognostic index in patients with congestive heart failure (CHF). Although EF can be readily measured, many clinicians use roentgenographic heart size as a clue to differentiate systolic from diastolic dysfunction, even in the absence of solid supportive data. To test the hypothesis that the cardiothoracic ratio (CTR) measured from the chest roentgenogram can be used to estimate left ventricular EF in individuals with CHF. To answer this question, the database of the Digitalis Investigation Group trial was used. The CTR, determined using the Danzer method, and quantitative EF, measured locally using angiographic, radionuclide, or 2-dimensional echocardiographic techniques, were compared in 7476 patients with clinical CHF (New York Heart Association functional classes I-IV) due to acquired left-sided cardiac disease of ischemic, hypertensive, idiopathic, and alcohol-related causes. Mean (+/-SD) CTR for the cohort was 0.53+/-.07. Mean (+/-SD) EF was 31.7%+/-12.2%. A weak, negative correlation between CTR and EF was observed (r=-0.176). Similar findings were obtained when the results were stratified by cause of CHF, presence of clinically defined right ventricular dysfunction, and method of EF measurement. Categorical analysis failed to yield a CTR cutoff point that facilitated useful segregation of individuals with an EF greater than 35% or 35% and below; greater than 40% or 40% and below; and greater than 45% or 45% and below in any patient group. Although a weak, negative correlation exists between CTR and EF, this relationship does not allow for accurate determination of systolic function in individual patients with CHF. Considering the morbidity and mortality associated with CHF, and the clinical implications of systolic function in this syndrome, direct measurement of EF is recommended.