Our study demonstrates two important findings. First, to our knowledge, it is the first report to reveal significant relationships between insufficient sleep and high blood pressure, asthma, and arthritis. In addition, significant relationships between insufficient sleep and obesity, diabetes, CHD, and stroke are consistent with recent reports [11, 12, 31]. Second, these significant associations were still present after adjustment for age, sex, race/ethnicity, education, FMD, and obesity. FMD and obesity were moderate mediators but did not fully explain the relationships of insufficient sleep to these chronic diseases. Therefore, insufficient sleep appears to be an important correlate of the leading chronic diseases.
In the BRFSS insufficient sleep is a newly developed sleep indicator and is not accompanied by reports of hours of sleep per night in the full study population. As an ad hoc analysis, we analyzed data from 74,944 respondents who had information on both insufficient sleep and sleep duration in 12 states in 2009. Our results indicate that the prevalence of reporting ≥14 days of perceived insufficient sleep was greater among those reporting <7 hours sleep (51.0%) in contrast to those reporting 7–9 hours (12.1%) or ≥10 hours sleep (16.1%) in a 24-hour period. Furthermore, the percentage of those reporting an optimal sleep (7–9 hours) also differed by number of days of perceived insufficient sleep (78.2% of 0 day, 67.4% of 1–13 days, 32.4% of 14–29 days, and 19.7% of all 30 days). These results are consistent with the findings from a previous study in Finland  and suggest that insufficient sleep is strongly related to sleep duration but they definitely are not redundant because they share only partially same information. Additionally, insufficient sleep may partly reflect some sleep disorders such as obstructive sleep apnea, which is also associated with chronic diseases [33–36] but we could not assess that relationship due to lack of data. Considerable data suggest that chronic sleep loss can result in insulin resistance and changes in appetite-regulating hormones, such as increased ghrelin and decreased leptin. The latter conditions further develop into chronic metabolic impairments such as obesity, and may subsequently lead to the development of diabetes, hypertension, heart disease, stroke, and even mortality [37–39].
Experimental evidence suggested that sleep loss was associated with exacerbation of arthritis through the increase of the sensitivity of pain among persons with rheumatoid arthritis [40, 41]. However, further study is needed to understand the directionality of the association between sleep loss and arthritis. In addition, a few studies indicated that poor sleep may affect lung function or lower the quality of life to exacerbate symptoms of asthma among youth [42, 43] or worsen the bronchoconstriction among persons with nocturnal asthma . The evidence may shed light on the mechanism of sleep loss associated with asthma.
Depression and anxiety often co-exist with sleep loss and many chronic diseases [45, 46]. Consistent with previous data from prospective studies, our study demonstrated a highly significant relationship between insufficient sleep and FMD, an indicator of psychological distress, and between FMD and chronic disease [10, 47]. Furthermore, our results suggested that FMD partially mediated the relationships between frequent insufficient sleep and chronic diseases although prospective studies are needed to clarify the pathway.
We assessed obesity as a potential mediator in the relationship between insufficient sleep and chronic disease because sleep loss is associated with increased risk for obesity and obesity is a robust risk factor for diabetes, hypertension, and cardiovascular diseases [3, 6, 11]. Our findings that obesity might partially mediate the relationships between frequent insufficient sleep and diabetes and high blood pressure were consistent with the results of studies assessing sleep duration [3, 4].
Strengths of this study include the large sample size in this population-based study (N = 375,653). Furthermore, this study represents the adult population in each of the 50 states. However, our study is subject to several limitations. First, due to the cross-sectional nature of the study, it is not possible to confirm whether there is a causal relationship between insufficient sleep and chronic disease. Nevertheless, such a causal link is biologically plausible because sleep loss may lead to metabolic abnormalities and weight gain through the regulation of hypothalamic-pituitary-adrenal axis [47, 48]. In addition, it is also very difficult to determine whether FMD and chronic conditions result in insufficient sleep or whether the direction of effect is reversed. Second, perceived insufficient sleep is a subjective measure of sleep health and has not been validated with polysomnography or other objective measures of sleep loss. Therefore, our results may have bias due to the misclassifications of exposure variable. Third, a significantly higher percentage of short sleep duration (average ≤6 hours per 24–hour period) was recently reported among workers with regular night shifts than among those with regular daytime shifts  supporting the role of circadian rhythm disruption in short sleeper and the possible association with chronic diseases . Therefore, shift work status may affect our results. However, we are unable to assess the effect of shift work status or circadian rhythm disruption on the relationship of chronic disease with insufficient sleep due to lack of data in BRFSS. Fourth, although we excluded those respondents who had missing values from the analyses, this is likely to bias our results toward the null if those non-respondents were included in our analyses . The results from a repeated analysis also confirm this assumption after including those non-respondents (data not shown). Additionally, self-reports may lead to underreporting of chronic diseases and obesity [50, 51], which could result in a much stronger relationship of insufficient sleep with the outcomes in this study. Finally, sample selection bias due to the low response rate is also possible because persons with severely impaired physical or mental health might not complete the BRFSS. Institutionalized persons and persons residing in households without landline telephones were also not included in the survey. However, the effect of potential systematic bias might be limited as a significant relationship between insufficient sleep and chronic disease revealed in our study was consistent with the results from prior research measuring sleep duration [2–6].