A novel method optimizing the normalization of cardiac parameters in small animal models: the importance of dimensional indexing.

For indexing cardiac measures in small animal models, tibia length (TL) is a recommended surrogate for body weight (BW) that aims to avoid biases because of disease-induced BW changes. However, we question if indexing by TL is mathematically correct. This study aimed to investigate the relation between TL and BW, heart weight, ventricular weights, and left ventricular diameter to optimize the current common practice of indexing cardiac parameters in small animal models. In 29 healthy Wistar rats (age 5-34 wk) and 116 healthy Black 6 mice (age 3-17 wk), BW appeared to scale nonlinearly to TL1 but linearly to TL3. Formulas for indexing cardiac weights were derived. To illustrate the effects of indexing, cardiac weights between the 50% with highest BW and the 50% with lowest BW were compared. The nonindexed cardiac weights differed significantly between groups, as could be expected (P < 0.001). However, after indexing by TL1, indexed cardiac weights remained significantly different between groups (P < 0.001). With the derived formulas for indexing, indexed cardiac weights were similar between groups. In healthy rats and mice, BW and heart weights scale linearly to TL3. This indicates that not TL1 but TL3 is the optimal surrogate for BW. New formulas for indexing heart weight and isolated ventricular weights are provided, and we propose a concept in which cardiac parameters should not all be indexed to the same measure but one-dimensional measures to BW1/3 or TL1, two-dimensional measures to BW2/3 or TL2, and three-dimensional measures to BW or TL3. NEW & NOTEWORTHY In healthy rats and mice, body weight (BW) scales linearly to tibia length (TL) to the power of three (TL3). This indicates that for indexing cardiac parameters, not TL1 but TL3 is the optimal surrogate for BW. New formulas for indexing heart weight and isolated ventricular weights are provided, and we propose a concept of dimensionally consistent indexing. This concept is proposed to be widely applied in small animal experiments.