I read the report by Katamreddy et al1 with great interest. The authors conducted a median of 8.6 years follow-up study to investigate the effect of day-to-day variation in sleep pattern, all-cause and cardiovascular mortality in 388 patients. The authors used Actiwatch device to retrieve sleep onset, duration, and sleep-wake cycles for 7 consecutive days, and sleep irregularity was assessed by coefficient of variation using mean and standard deviations. By using mean sleep duration, inflammation, diabetes, age, body mass index, renal function, and blood pressure, the adjusted hazard ratio (95% confidence interval) of sleep duration variation in the highest vs the lowest tertile for all-cause mortality was 4.85 (1.52–15.49). In contrast, there was no statistically significant increase in cardiovascular mortality with variation in sleep duration. Irregularity in the onset of sleep was not associated with all-cause mortality or cardiovascular mortality. I have comments regarding statistical procedure and Actiwatch device limitation.
First, a total of 37 patients died, including 10 deaths resulting from cardiovascular causes. The authors used 7 factors for the adjustment and 8 independent variables were used for Cox regression analysis. The number of events per independent variable should be kept 10 or higher to keep validity in prediction model,2,3 and 70 events are needed. I suspect that a wide range of 95% confidence interval may reflect unstable estimation, and the lack of statistical power may partly relate to the statistical results.
Second, the authors used Actiwatch device to measure objective sleep parameters. As an accelerator-based physiological apparatus cannot reflect sleep status by measuring sleep polysomnography,4 patients having a risk of death may present dissociation with actual sleep status. In addition, Yetish et al5 reported ethnic differences in nightly intraindividual sleep variation existed.
Citation: Kawada T. Day-to-day variations in sleep parameters and subsequent risk of mortality. J Clin Sleep Med. 2022;18(9):2331.
DISCLOSURE STATEMENT
The author has seen and approved this manuscript. Work for this study was performed at Nippon Medical School. The author reports no conflicts of interest.
REFERENCES
- 1. Katamreddy A , Uppal D , Ramani G , et al . Day-to-day variation in sleep duration is associated with increased all-cause mortality . J Clin Sleep Med. 2022. ; 18 ( 3 ): 921 – 926 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Concato J , Peduzzi P , Holford TR , Feinstein AR . Importance of events per independent variable in proportional hazards analysis. I. Background, goals, and general strategy . J Clin Epidemiol. 1995. ; 48 ( 12 ): 1495 – 1501 . [DOI] [PubMed] [Google Scholar]
- 3. Peduzzi P , Concato J , Feinstein AR , Holford TR . Importance of events per independent variable in proportional hazards regression analysis. II. Accuracy and precision of regression estimates . J Clin Epidemiol. 1995. ; 48 ( 12 ): 1503 – 1510 . [DOI] [PubMed] [Google Scholar]
- 4. Messman BA , Slavish DC , Dietch JR , Jenkins BN , Ten Brink M , Taylor DJ . Associations between daily affect and sleep vary by sleep assessment type: What can ambulatory EEG add to the picture? Sleep Health. 2021. ; 7 ( 2 ): 219 – 228 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Yetish G , Kaplan H , Gurven M . Sleep variability and nighttime activity among Tsimane forager-horticulturalists . Am J Phys Anthropol. 2018. ; 166 ( 3 ): 590 – 600 . [DOI] [PMC free article] [PubMed] [Google Scholar]