Should echocardiography be performed to assess effects of antihypertensive therapy? Test-retest reliability of echocardiography for measurement of left ventricular mass and function

AbstractObjectives. The purpose of this study was to determine the test-retest stability of echocardiography for the measurement of left ventricular mass and function in patients with hypertension.Background. Determination of changes in left ventricular mass may be impaired by study variability. The amount by which variables of mass and left ventricular function must change in an individual patient to exceed temporal variability has not been determined in a multicenter trial.Methods. Ninety-six patients with hypertension had two-demensional targeted, M-mode Doppler echocardiography repeated at 6 ± 8 days by the same technician utilizing the same machine. Left ventricular mass and variables of systolic and diastolic function were measured. Test-retest reliability and the width of the 95% confidence intervals of variable change, as well as the contributions of age, study quality and body size to measurement reliability, were determined.Results. Despite excellent reliability (intraclass coefficient of correlation 0.86), the 95% confidence interval width of a single replicate measurement of left ventricular mass was 59 g, exceeding usual decreases in mass during treatment. Study quality, which was dependent on age and weight, influenced test reliability. Although the confidence interval width for ejection fraction was narrow (5 U), those for peak early (E) and late (A) diastolic velocities were wide, resulting in a confidence interval width for the E/A ratio of 1.5.Conclusions. The temporal variability, particularly in obese or elderly patients, or both, of echocardiography for measurement of left ventricular mass precludes its use to measure changes in mass of the magnitude likely to occur with therapy. Measurement stability is affected by study quality, and age and body weight both influence study quality. Although ejection fraction shows little temporal variability, the large width of the confidence interval of the Doppler E/A ratio impairs its use to serially measure diastolic function