Proposed limits for ordinal echocardiographic estimates of left atrial enlargement in dogs

To propose echocardiographic left-atrial-to-aortic ratio (LA:Ao) limits for defining ordinal categories of left atrial (LA) enlargement in dogs.

Right parasternal short-axis echocardiographic images from 33 dogs with various degrees of LA enlargement. Right parasternal short-axis and long-axis echocardiographic measurements from 238 healthy dogs. Images were duplicated and randomized. Duplicate images contained a LA:Ao estimate. Participants categorized the LA in each image into one of four categories: normal, mildly, moderately or severely enlarged. Distributions of categorization were compared between cardiologists and non-cardiologists. Intra-observer intra-study and inter-study agreement were examined. Effect of measurement was evaluated on agreement between participants. A parametric estimate of LA enlargement was calculated for both short-axis and long-axis views.

Cardiologists and non-cardiologists provided similar distributions of LA size estimates, and showed similarly high intra-observer agreement (kappa = 0.84). Having a measurement provided with the image increased agreement for categorizing LA as normal or mildly enlarged (P<0.001). Parametric and consensus-based approaches provided similar limits for categorizing left atrial size in the right parasternal short-axis view - normal = LA:Ao < 1.6, mildly enlarged = 1.6 < LA:Ao < 1.9, moderately enlarged = 1.9 < LA:Ao < 2.3, severely enlarged = LA:Ao ≥ 2.3. A parametric approach for the right parasternal long-axis view provided the following: normal = LA:Ao < 2.1, mildly enlarged = 2.1 < LA:Ao < 2.5, moderately enlarged = 2.5 < LA:Ao < 2.7, severely enlarged = LA:Ao ≥ 2.7.

Participants mostly classified LA sizes into four ordinal categories that corresponded to the aforementioned limits. Clinicians estimating LA size in early diastole can use these limits to increase inter-observer agreement when identifying LA enlargement.