The impact of scalar variable and process on athlete-control comparisons of cardiac dimensions. Med. Sci. Sports Exerc., Vol. 30, No. 6, pp. 824-830, 1998.
This study compared linear left ventricular dimensions and mass (LVM), before and after normalizing for body dimensions via allometric and ratio-standard scaling.
Height (HT; m), body mass (BM; kg), body surface area (BSA; m2), and fat-free mass (FFM; kg) were measured in elite male weight lifters (N = 11) and age-matched controls (N = 45). Septum (ST), posterior wall (PWT), and internal dimension in diastole (LVIDd) were measured from M-mode echocardiographic traces and used to calculate LVM. Via multivariate allometric scaling, common group power function exponents were identified for all cardiac dimensions related to all body size scalars. t-tests were used to compare group differences in absolute and scaled data.
BM, FFM, and BSA, as well as absolute LVM (262 ± 54 vs 206 ± 39) and ST (11 ± 1 vs 9 ± 1), were greater in the athletes (P < 0.05). All exponents conformed to dimensionality theory within 95% confidence limits. Fat-free mass presented the highest multiple R value and the least residual sum of squares of any scalar variable. If FFM was used to scale, no difference in LVM remained (P > 0.05).
Data suggest that any group effect on cardiac dimensions is substantially altered by the scaling procedure. The choice of the most appropriate variable and process for partitioning out any effect of body dimensions on cardiac dimensions in similar studies requires attention.
Department of Exercise and Sport Science, the Manchester Metropolitan University, Crewe and Alsager Faculty, Hassall Road, Alsager, Cheshire, ST7 2HL, UNITED KINGDOM
Submitted for publication February 1997.
Accepted for publication September 1997.
Address for correspondence: Keith P. George, Department of Exercise and Sport Science, the Manchester Metropolitan University, Crewe and Alsager Faculty, Hassall Road, Isager, Cheshire, ST7 2HL, United Kingdom. E-mail: firstname.lastname@example.org.