Comprehensive ECG reference intervals in C57BL/6N substrains provide a generalizable guide for cardiac electrophysiology studies in mice

Abstract

Reference ranges provide a powerful tool for diagnostic decision-making in clinical medicine and are enormously valuable for understanding normality in pre-clinical scientific research that uses in vivo models. As yet, there are no published reference ranges for electrocardiography (ECG) in the laboratory mouse. The first mouse-specific reference ranges for the assessment of electrical conduction are reported herein generated from an ECG dataset of unprecedented scale. International Mouse Phenotyping Consortium data from over 26,000 conscious or anesthetized C57BL/6N wildtype control mice were stratified by sex and age to develop robust ECG reference ranges. Interesting findings include that heart rate and key elements from the ECG waveform (RR-, PR-, ST-, QT-interval, QT corrected, and QRS complex) demonstrate minimal sexual dimorphism. As expected, anesthesia induces a decrease in heart rate and was shown for both inhalation (isoflurane) and injectable (tribromoethanol) anesthesia. In the absence of pharmacological, environmental, or genetic challenges, we did not observe major age-related ECG changes in C57BL/6N-inbred mice as the differences in the reference ranges of 12-week-old compared to 62-week-old mice were negligible. The generalizability of the C57BL/6N substrain reference ranges was demonstrated by comparison with ECG data from a wide range of non-IMPC studies. The close overlap in data from a wide range of mouse strains suggests that the C57BL/6N-based reference ranges can be used as a robust and comprehensive indicator of normality. We report a unique ECG reference resource of fundamental importance for any experimental study of cardiac function in mice.

Fig. 1

Representative averaged waveform of a C57BL/6N mouse as a function of voltage over time, reflecting the most commonly applied ECG annotations with P-, Q-, R-, S-, and T-peak detection. Some differences in the placement of S- and T-annotations between centers participating in the IMPC were observed. Differences in T-peak placement are represented as T_Conscious and T_Anesthetized

Fig. 2

Coefficient of variation (COV) analysis of data split by sex (female and male) and age (EA and LA) identified parameters with excess variability (COV > 30%) that were excluded from further analysis (white bars). Parameters in blue were below the COV threshold of 30% and were retained for further analysis. These were in ascending COV percentage QTc Mitchell, PQ-, QT-, QRS complex, ST-, HR-, RR- and PR-interval

Fig. 3

Histograms presenting the distribution of each selected ECG parameter for male and female mice separately. By visual inspection, no sexual dimorphism was apparent. Panel a: Recorded in the conscious state in EA (Subpanels A–G) and LA mice (Subpanels H–N). Panel b: Recorded under isoflurane anesthesia in EA (Subpanels A–G) and LA mice (Subpanels H–N). Panel c: Recorded under tribromoethanol anesthesia in EA (Subpanels A–G). No LA data are available for tribromoethanol anesthesia

Fig. 3

Histograms presenting the distribution of each selected ECG parameter for male and female mice separately. By visual inspection, no sexual dimorphism was apparent. Panel a: Recorded in the conscious state in EA (Subpanels A–G) and LA mice (Subpanels H–N). Panel b: Recorded under isoflurane anesthesia in EA (Subpanels A–G) and LA mice (Subpanels H–N). Panel c: Recorded under tribromoethanol anesthesia in EA (Subpanels A–G). No LA data are available for tribromoethanol anesthesia

Fig. 4

Comparison of the anesthetic regimes with the conscious state recordings. Distribution of the seven selected ECG parameters presented by histograms, stratified for female and male mice in EA (Subpanels A–G) and LA populations (Subpanels H and N). Color code: Conscious, Isoflurane and Tribromoethanol anesthesia

Fig. 5

Testing age-differences in conscious mice. T-test results when comparing conscious EA versus LA data show high significance for all parameters (p < .001) and negligible to medium Cohen´s d standardized effect sizes. Panel a: Females, Panel b: Males. Panel c. Bootstrap analysis of power estimates for sample sizes ranging from 5 to 100 mice, presented for each of the seven selected ECG parameters. Note: X-axis: bootstrapped sample sizes from 5 to 100; Y-axis: proportion of significant tests with p < .05

Fig. 6

Reference ranges split by anesthetic regimen showing median, and 95% reference ranges (2.5th and 97.5th percentile). Female data are directly above the male data for EA (Subpanels A–G) and LA populations (Subpanels H–N). For the ST-interval in anesthetized mice (Subpanels F and M) data were provided from only one center and for the LA range (Subpanel M) the number of mice was below the recommended number to generate a reliable estimate (Supplementary Table 2). No LA data were available for tribromoethanol anesthesia. Color code: Conscious, Isoflurane and Tribromoethanol anesthesia

Fig. 7

Independent, non-IMPC study on six of the founder strain mice reported in “The Collaborative Cross: A Recombinant Inbred Mouse Population for the Systems Genetic Era” (Threadgill et al. 2011) study, including 129S1/SvlmJ, A/J, C57BL/6 J, NOD/ShiLtJ, NZO/HlLtJ, and PWK/PhJ inbred strains show a close alignment to the reference ranges reported herein for PR-, QRS complex, QT-, RR-, and ST-interval based on multiple C57BL/6N substrains indicating good utility for those reference ranges. Data for HR was not available in this study. Mice were conscious, split by sex and ~ 12 weeks of age, equivalent to the IMPC EA timepoint. Red dotted lines depict the boundaries of the sex-specific reference ranges calculated herein, for each parameter