The optimal measurement period of actigraphy for circadian rhythm in relation to adiposity: A retrospective case-control study

Background: This study focused on the relationship between adiposity and Rest-Activity Rhythms (RAR), utilizing both parametric cosine-based models and non-parametric algorithms. The emphasis was on the impact of varying measurement periods (7-28 days) on this relationship. Methods: We retrieved actigraphy data from two datasets, encompassing a diverse cohort recruited from an obesity outpatient clinic and a workplace health promotion program. Participants were required to wear a research-grade wrist actigraphy device continuously for a minimum of four weeks. The final dataset included 115 individuals (mean age 40.7 +/- 9.5 years, 51 % female). We employed both parametric and non-parametric methods to quantify RAR using six standard variables. Additionally, the study evaluated the correlations between three key adiposity indices Body Mass Index (BMI), Visceral Adipose Tissue (VAT) area, and Body Fat Percentage (BF%) and circadian rhythm indicators, controlling for factors like physical activity, age, and gender. Results: The obesity group displayed a significantly lower relative amplitude (RA) as per non-parametric algorithm findings, with a decreased amplitude noted in the parametric algorithm analysis, in comparison to the overweight and control groups. The relationship between circadian rhythm indicators and adiposity metrics over 7- to 28-day periods was examined. A notable negative correlation was observed between RA and both BMI and VAT, while correlation coefficients between adiposity indicators and non-parametric circadian parameters increased with extended durations of actigraphy data. Specifically, RA over a 28-day period was significantly correlated with BF%, a trend not seen in the 7-day measurement (p = 0.094) in multivariate linear regression. The strength of the correlation between BF% and 28-day RA was more pronounced than that in the 7-day period (p = 0.044). However, replacing RA with amplitude as per parametric cosinor fitting yielded no significant correlations for any of the measurement periods. Conclusion: The study concludes that a 28-day measurement period more effectively captures the link between disrupted circadian rhythms and adiposity. Non-parametric algorithms, in particular, were more effective in characterizing disrupted circadian rhythms, especially when extending the measurement period beyond the standard 7 days.