The ACVIM consensus statement definition of left ventricular enlargement in myxomatous mitral valve disease does not...

To determine how frequently the current criteria for left ventricular enlargement in dogs misclassify healthy dogs as having left ventricular enlargement; to examine the effect of breed on diastolic left ventricular normalized dimensions (LVIDDN); to propose appropriate scaling exponents and reference limits for dogs.

Echocardiographic data from 1,124 healthy adult dogs, including 454 dogs weighing <20 kg.

We calculated power regression parameters (allometric scaling), including exponents and proportionality constants, for various subsets of the dogs (all dogs, dogs < 20 kg, generic dogs, and individual breeds with >10 observations) and derived upper reference limits for LVIDDN. We determined the proportions of dogs that would be identified as having left ventricular enlargement with each regression model compared to previously published reference limits or guidelines. We then identified breeds failing to conform to generic dog models.

The American College of Veterinary Internal Medicine-recommended scaling exponent (0.294) and criterion for identifying left ventricular enlargement (1.7) identified >10% of apparently healthy dogs as having left ventricular enlargement, with specific breeds being misclassified up to 50% of the time. However, with a scaling exponent of 0.33, a constant of 1.7 represented a normal left ventricular size in 97.5% of healthy dogs in both generic and non-conforming breeds.

Left ventricular internal dimension in diastole normalized to bodyweight is breed-dependent. A constant of 1.7 with a scaling exponent of 0.294 does not always represent ventricular enlargement; a scaling exponent of 0.33, with breed-specific reference limits for breeds that fail to conform to allometric models of generic dogs, reduces the misclassification of healthy dogs as having left ventricular enlargement.