Journal of Preventive Medicine

Keywords

Sleep duration; Diabetes; Middle-aged adults; General health; Mental health; Chronic conditions

Introduction

Sleep duration, whether too short or too long, can negatively impact health. This has become an issues for healthcare providers because the number of disordered sleepers has increased over the last 30 years [1]. Upto one-third of people in the U.S. report less than 6 hours of sleep per night and another third report 9 or more hours, both leading to significant loss of productivity and health [2-4]. Indeed, both short and long sleep durations are linked to chronic health conditions including hypertension, depression, cardiovascular diseases, and metabolic syndrome [1,3,4-7].

Research shows that sleep issues are related to diabetes status. Type 2 diabetes affects over 387 million people worldwide. In the U.S., it is the seventh leading cause of death and currently impacts over 30 million citizens [7-10]. Economically, diabetes costs the U.S. healthcare system $245 billion annually, and this cost causes higher medical expenditures for diabetics than for those without the disease [7,11]. The risk of developing diabetes has been associated with both overly short and long sleep durations [1,2,5-7,10,12-14]. Several systematic review articles show evidence that men and women who slept less than five hours per night or nine or more hours had a higher risk of developing type 2 diabetes [1,2,5,6,10]. Furthermore, in men and women diagnosed with type 2 diabetes, there is a U-shaped curve between hemoglobin A1c levels and sleep duration, indicating that both short and long sleep duration may produce poor diabetic outcomes [1,5,6,14].

While ample evidence may exist for the effect of sleep on diabetes, there is currently insufficient research on the effect of diabetes on sleep issues. Whether or not diabetics tend to have abnormal sleep durations remains uncertain [5]. If abnormal sleep durations are more common in diabetics, it should be treated in addition to their diabetes to improve control of their blood glucose level [5,12,14]. Therefore, our study examines whether sleep duration differs by diabetes status in middle-aged males and females.

Methods

Design

This cross-sectional analysis used 2016 data from the Behavior Risk Factor Surveillance System (BRFSS) conducted across the U.S. by the Centers for Disease Control and Prevention (CDC) [15]. Established in 1984, BRFSS is a system of health-related telephone surveys using random digit dialing techniques to collect data about U.S. residents regarding their health-related risk behaviors, chronic health conditions, and use of preventive services. BRFSS collects data in all 50 states and completes more than 400,000 adult interviews annually. The CDC compiles the data and allows researchers access to deidentified data to conduct secondary data analyses. This study was given exempt status by the Institutional Review Board of the University of North Texas Health Science Center.

Sample

The samples included males and females ages 40-65 years old in Georgia (N=2352), Louisiana (N=2377), New Mexico (N=2832), and Oklahoma (N=2908). These states were chosen because they had larger proportions of diabetic residents in the target age range compared to the other states based on BRFSS prevalence data [16].

Data

The outcome, sleep duration, was originally measured in BRFSS as the average number of hours of sleep per night. We dichotomized this number to reflect two categories of sleep duration: “moderate” (averaging 6 to 8 hours of sleep per night) vs “non-moderate” (averaging more than or less than 6 to 8 hours per night). The factor of interest, diabetes status, was measured as “yes” or “no” to having a diagnosis of diabetes.

Control variables included number of health conditions, general health status, mental health status, tobacco use, physical activity, income level, education level, employment status, gender, ethnicity/race, and age category. The number of health conditions was determined as the number of “yes” responses to diagnoses for the following: heart attack, CHD, stroke, skin cancer, other cancer, COPD, arthritis, depression, kidney disease, and asthma. This number was then categorized as “0 health conditions”, “1 health condition” or “2 or more health conditions” other than diabetes.

Participant’s general health status was categorized as “good or better” or “poor or fair”. Mental health was dichotomized as “yes” or “no” to 30 days of good mental health in the past 30 days. Tobacco use was measured as “current smoker” or “nonsmoker”. Physical activity was dichotomized as “yes” or “no” to “are having performed physical activity or exercise in the past 30 days”. Income level was measured as “$50,000 or more” and “less than $50,000”. Education level was dichotomized as “yes” or “no” to “having graduated college/technical school”. Employment was categorized as “yes” or “no” to being “employed”. Because the majority of participants reported white race, ethnicity/race was dichotomized as “white, non-Hispanic” or “other”. Age was categorized as “40-49”, “50-59” and “60-65.” Descriptive statistics and categories for all variables across the four states are shown in Table 1.

Table 1: Participant characteristics by state.

Analysis

Frequency distributions by state were used to describe the sample and to identify any problems with variable distributions. Multiple logistic regression analysis by state was used to assess the relationship between diabetes status and sleep duration while controlling for factors related to health status, health behaviors, socioeconomic status, and demographic factors. Models were originally run by gender but because the results were very similar, in the spirit of parsimony, we decided to run one model for both genders. We chose to analyze data separately by state to assess patterns in relations between variables across similar samples. As such, a similar result in three or four out of four states was considered reliable evidence for a relationship. Observations with missing data for any of the variables were excluded from the multivariable analysis. All analyses were conducted in STATA 15 (Copyright 1985-2017 StataCorp LLC).

Results

Descriptive statistics

Table 1 lists participant characteristics for middle-aged males and females in Georgia, Louisiana, New Mexico and Oklahoma. Across states, the majority of participants reported a moderate amount of sleep each night (6-8 hours, 76%-81%) and less than one-fifth reported a diagnosis of diabetes (16%-18%). For other health indicators, most reported good or better general health (76%-78%); the majority had one or more health conditions other than diabetes (55%-60%); and the majority reported good mental health in past 30 days (62%-69%). For health behaviors, most did not smoke tobacco (79%-84%) and the majority performed physical activity in the past 30 days (68%-79%).

For socioeconomic status, almost half reported earning $50,000 or more (43%-51%); the majority reported not graduating college or technical school (62%-69%); and the majority reported being employed (57%-61%). For demographic factors, the majority of participants were female (57%-62%) and white, non-Hispanic (51%-78%) with roughly one-third in each category (ages 40-49: 25%-25%; ages 50-59: 41%-43%; ages 60-65: 29%-36%).

Adjusted Statistics

As shown in Table 2, the results of multiple logistic regression analysis for middle-aged males and females in Georgia, Louisiana, New Mexico and Oklahoma indicated that after controlling for all other variables in the model, sleep duration was not significantly related to diabetes status across states. However, compared to those who reported no health conditions, participants in all four states who reported 2 or more health conditions (not including diabetes) were about 1.5 to 2 times less likely to report moderate sleep duration. In contrast, those who reported 30 days of good mental health in all four states were about 1.5 times more likely to report moderate sleep duration compared to those who reported less than 30 days of good mental health, and those who performed physical activity in the past 30 days in three of four states were about 1.5 times more likely to report moderate sleep compared to those who did not perform physical activity in the past 30 days.

Table 2: Adjusted results by state.

Discussion

The purpose of this study was to assess whether sleep duration differed by diabetes status in middle-aged males and females in the U.S. general population. The results of adjusted analyses indicated that sleep duration was not significantly related to diabetes status across states, which differs from previous studies that found a significant relationship between sleep duration and diabetes status [1,3-7,10,12-14]. This discrepancy could be due to measurement differences, as we measured moderate sleep duration as 6 to 8 hours whereas other studies measured it from as low as 5 hours to as high as 10 hours [1]. Discrepancies could also be attributed to differences between the target populations. In a review of 40 prospective cohort studies [3], only six of those studies examined males and females in America. Additionally, the majority of prior research examined the effects of sleep duration on diabetes risk [1,3,6,7,13], whereas this study assessed whether sleep duration differed by diabetes status.

However, the results of this study did indicate that sleep duration is related to number of health conditions. Males and females 40-65 years old with two or more health conditions were about 2 times less likely to report moderate sleep duration. This is consistent with prior research that found abnormal sleep duration to be significantly related to chronic health conditions including cardiovascular disease, coronary heart disease, dyslipidemia, depression, hypertension, and obesity [2,3,7,12,13]. However, this study extends findings by showing that the number of chronic health conditions is also related to sleep duration and not just to each condition.

The results of this study also indicate that moderate sleep duration was consistently and moderately related to mental health and physical activity in males and females ages 40-65. Across all four states, participants who reported good mental health were about 1.5 times more likely to report moderate sleep, a moderate relationship also established in prior studies [17-19]. In addition, results in three of four states indicated that those who reported performing physical activity were about 1.5 times more likely to report moderate sleep, also shown in previous research [17,20] thus, sleep duration may be related more to the severity and management of multiple health issues, including diabetes.

Limitations

BRFSS provides large general population samples to assess variable relations across states; however, the data does not detail the severity or management of diabetes or other health conditions, including information on medications, many of which could induce somnolence or insomnia and affect sleep duration. Also, sleep duration in our study was measured subjectively. Because self-reported sleep duration has been shown previously to be higher than measured sleep duration, some participants may have qualified themselves under moderate sleep duration while actually having short sleep duration [21].

Conclusion

The results of this population-based study may generalize to middle-aged males and females in a primary care setting. Primary care providers may expect up to one-fourth of their middle-aged male and female patients to report non-moderate sleep duration or to have a diabetes diagnosis. However, the results of this study indicated that diabetes status and sleep duration were not related. Thus, primary care providers should screen for diabetes and sleep issues when symptoms present, but treat them separately and refer to specialists as needed. In addition, primary care providers may expect up to one-third of their middle-aged patients to report two or more health conditions, mental health issues, or lack of physical activity. Because the results of this study indicate that sleep duration is moderately related to these, providers should screen for sleep duration, health conditions, mental health, and physical activity if symptoms of any are present. Providers should educate patients about managing their symptoms and chronic conditions, coordinate treatment plans for comorbid conditions, and make referrals to appropriate specialties as needed.

Disclaimers

No author has any conflict of interest.

Source of support

There was no financial support for the conduct or findings of this study.

References

1. Larcher A, Benhamou PY, Pépin JL, Borel AL (2015) Sleep habits and diabetes. Diabetes Metab 41: 263-271.
2. Itani O, Jike M, Watanabe N, Kaneita Y (2017) Short sleep duration and health outcomes: A systematic review, meta-analysis, and meta-regression. Sleep Med 32: 246-256.
3. Liu TZ, Xu C, Rota M, Cai H, Zhang C, et al. (2017) Sleep duration and risk of all-cause mortality: A flexible, non-linear, meta regression of 40 prospective cohort studies. Sleep Med Reviews 32: 28-36.
4. Jike M, Itani O, Watanabe N, Buysse DJ, Kaneita Y, et al. (2018) Long sleep duration and health outcomes: A systematic review, meta-analysis and meta-regression. Sleep Med Reviews 39: 25-36.
5. Arora T, Taheri S (2015) Sleep optimization and diabetes control: A review of the literature. Diabetes Therapy 6: 425-468.
6. Lee SW, Ng KY, Chin WK (2017) The impact of sleep amount and sleep quality on glycemic control in type 2 diabetes: A systematic review and meta-analysis. Sleep Med Reviews 31: 91-101.
7. Anothaisintawee T, Reutrakul S, Van Cauter EV, Thakkinstian A (2016) Sleep disturbances compared to traditional risk factors for diabetes development: Systematic review and meta-analysis. Sleep Med Reviews 30: 11-24.
8. https://www.cdc.gov/diabetes/pdfs/data/statistics/national-diabetes-statistics-report.pdf
9. https://www.cdc.gov/diabetes/basics/type2.html
10. Zhu B, Hershberger PE, Kapella MC, Fritschi C (2017) The relationship between sleep disturbance and glycaemic control in adults with type 2 diabetes: An integrative review. J Clin Nursing 26: 4053-4064.
11. https://www.cdc.gov/diabetes/basics/quick-facts.html
12. Holliday EG, Magee CA, Kritharides L, Banks E, Attia J, et al. (2013) Short sleep duration is associated with risk of future diabetes but not cardiovascular disease: A prospective study and meta-analysis. PLoS One 25: 8.
13. Spiegal K, Knutson K, Leproult R, Tasali E, Cauter E, et al. (2005) Sleep loss: A novel factor for insulin resistance and type 2 diabetes. J Appl Physiol 99: 2008-2019.
14. Tsuneki H, Sasaoka T, Sakurai T (2016) Sleep control, GPCRs, and glucose metabolism. Trends Endocrinol Metab 227: 633-642.
15. https://www.cdc.gov/brfss/about/index.htm
16. https://www.cdc.gov/brfss/brfssprevalence
17. Strine TW, Chapman DP (2005) Associations of frequent sleep insufficiency with health-related quality of life and health behaviors. Sleep Med 6: 23-27.
18. Lippman S, Gardener H, Rundek T, Seixas A, Elkind MS, et al. (2017) Short sleep is associated with more depressive symptoms in a multi-ethnic cohort of older adults. Sleep Med 40: 58-62.
19. Zhai L, Zhang H, Zhang D (2015) Sleep duration and depression among adults: A meta-analysis of prospective studies. Depress Anxiety 32: 664-670.
20. Gerber M, Lindwall M, Börjesson M, Hadzibajramovic E, Jonsdottir IH (2017) Low leisure-time physical activity, but not shift-work, contributes to the development of sleep complaints in Swedish health care workers. Ment Health Phys Act 13: 22-29.
21. Lauderdale DS, Knutson KL, Yan LL, Liu K, Rathouz PJ, et al. (2008) Self-reported and measured sleep duration: How similar are they? Epidemiol 19: 838-845.