Challenging conventional wisdom: A fresh look at the litter effect in rats.

内容摘要

This retrospective study challenges the assumption that physiological variance among neonates is greater between litters than within litters of multiparous species like rats. Conventional wisdom holds that litters - not individual pups - should be the unit of replication due to shared perinatal environment and genetics. We provide empirical evidence contradicting this view. We measured within- versus between-litter variance in body weight of outbred Sprague-Dawley (SD) neonatal rats across three cohorts: (1) pups from 24 mothers weighed on postnatal day 4 (P4); (2) P4 pups from 10 untreated and seven nicotine-exposed litters; and (3) breathing frequency in pups aged P3-P4 from six litters. We used Welch's ANOVA and the intraclass correlation coefficient (ICC) as primary outcome measures. Importantly, ICC quantifies raw variance distribution, unlike ANOVA's mean-square calculation. Despite significant 'litter effects' detected by conventional ANOVA for body weight, ICC analysis revealed that ∼40-75% of total variance in cohort 1 originated within litters. In cohort 2, variance was distributed more equally, but between-litter variance (43%) remained less than within-litter variance (57%). Remarkably, for breathing frequency, nearly 100% of total variance originated within litters. These findings align with the genetics and developmental biology of outbred SD rats: each fetus is genetically unique and possesses its own placenta with distinct blood flow, hormone levels and nutrient delivery. We therefore argue against uniform policies mandating how many offspring per litter to use experimentally. Instead, researchers should base this decision on their specific research question, outcome variables, rat genetics, and careful consideration of the raw variance both within and between litters. KEY POINTS: Because rat littermates share a genetic background and perinatal environment, conventional wisdom assumes variance is greater between litters than within litters, which has led statisticians and reviewers to conclude that the litter, not each animal within the litter, is the true unit of replication. Our retrospective analysis of outbred Sprague-Dawley rat pups revealed the opposite: within-litter variance significantly exceeded between-litter variance for body weight (40-75% within) and breathing frequency (nearly 100% within). Even in a cohort that included pups from nicotine-exposed litters, within-litter variance (57%) exceeded between-litter variance (43%). Each outbred fetus is genetically unique with its own placenta, umbilical cord, and uterine position - making substantial within-litter variation expected, not surprising. These findings challenge conventional wisdom about littermate homogeneity and necessitate reconsideration of experimental design practices in developmental research in outbred rat strains.