Development of allometric equations to establish normal limits for M-mode echocardiographic measurements in the Italian population. Gruppo Ligure Della Società Italiana di Ecografia Cardiovascolare.

BACKGROUND

Despite general acceptance of the normal limits for M-mode echocardiographic measurements established in the U.S.A. in the late seventies, several studies have shown that those limits of normality have some limitations which restrict their use in clinical practice, and that these limits should be specific for the population that should be studied. To date, however, there is no reference method to calculate normal limits of M-mode echocardiographic measurements in the Italian population.

METHODS

627 healthy volunteers (382 males, 245 females) ranging in age from 4 to 89 years, were prospectively examined, using two-dimensionally directed M-mode echocardiography, in 13 echocardiographic laboratories. End-diastolic and-systolic left ventricular diameters, interventricular septum and posterior wall thickness, end-diastolic aortic dimension and end-systolic left atrial diameter were measured following the recommendations of the American Society of Echocardiography. Age, gender, weight, height, heart rate and blood pressure were used as determinants of the echocardiographic measurements. The specificity of the resulting normal limits of echocardiographic measurements was then tested on an independent population of 145 normal subjects (92 males, 53 females) who were prospectively examined after the development of the model.

RESULTS

Gender, age and weight were significant and independent determinants in nearly all the echocardiographic measurements performed. After logarithmic transformation of all the variables, we developed sex specific regression models which allow calculation of 95% prediction intervals for the cardiac chamber size and wall thickness, using an exponential multivariate regression model. Ninety-one percent of the echocardiographic measurements performed on the control population (range 84.6-94.1%) were correctly identified as normal using our model. Conversely, the specificity of the existing reference models developed in the U.S.A. or in the Netherlands identified, respectively, only 77% (range 53.1-92.4%) and 74% (range 53.8-87.6%) of those measurements as normal (p < 0.0001 vs our model for both). Differences in specificity between our models and the U.S. one also persisted after correcting the U.S. model for the Italian population (82.8% versus 91.3%, respectively; p = 0.003). Conversely, differences in specificity between our models and the Dutch ones were no longer detectable after correcting the Dutch models (which has a statistical approach very similar to ours) for the Italian population (91.3 versus 94.7%, respectively; p = NS).

CONCLUSIONS

Our study provides reliable mathematical model based on exponential growth (allometric) equations to calculate upper and lower normal limits of M-mole echocardiographic measurements obtained from Italian subjects given their gender, height, weight, heart rate, age and blood pressure.