Normal values of left atrial volume in pediatric age group using a validated allometric model

Circ Cardiovasc Imaging. 2012 Nov;5(6):791-6. doi: 10.1161/CIRCIMAGING.112.974428. Epub 2012 Oct 16.


Background: Left atrial volume (LAV) increase is an indicator of diastolic dysfunction and a surrogate marker of significant left to right shunts. Normalization of LAV is currently performed by indexing to body surface area(1) (BSA(1)). The indexed LAV thus derived does not account for the nonlinear relationship of physiologic variables to BSA and has not been tested for independence to body size. Our objective was to identify a valid allometric model for indexing LAV and use it to develop Z-scores in children.

Methods and results: LAV was measured in 300 normal subjects by echocardiography using the biplane area length method. LAV/BSA(1) had a residual relationship to BSA (r=0.52, P<0.0001). The allometric exponent (AE) derived for the entire cohort (1.27) using the least squares regression analysis also failed to eliminate the residual relationship to BSA (r=-0.15, P=0.01). Dividing the cohort in two groups with a BSA cut-off of 1 m(2) provided the best-fit allometric model. The AE for each group was 1.48 and 1.08 for BSA≤1 m(2) and >1 m(2), respectively, and was validated against an independent sample. The mean indexed LAV±SD for BSA≤1 m(2) and >1 m(2) is 31.5±5.5 mL and 26.0±4.2 mL, respectively, and was used to derive Z-scores.

Conclusions: This study demonstrates the fallacy of using "per-BSA(1) standards" for normalization of LAV in pediatrics. LAV/BSA(1.48) for children with BSA≤1 m(2) and LAV/BSA(1.08) for those with BSA>1 m(2) is accurate and can be used to derive Z-scores.

Publication types

  • Comparative Study

MeSH terms

  • Adolescent
  • Atrial Function*
  • Cardiac Volume*
  • Child
  • Child, Preschool
  • Echocardiography / methods*
  • Female
  • Heart Atria / diagnostic imaging*
  • Humans
  • Infant
  • Infant, Newborn
  • Male
  • Models, Theoretical*
  • ROC Curve
  • Reference Values
  • Reproducibility of Results