A Novel Innovative Sleeping Mattress Could Possibly Save Lives and the National Economy: Small but Solid Evidence from a Medical Perspective

The RAND Corporation reported that insufficient sleep causes an economic loss of $138 billion (2.92% of GDP) in Japan every year. In this study, we in-vestigated the sleep improvement effect of a novel innovative sleeping mattress called “AiR SI” (Nishikawa Co., Ltd.) to find a simple method to solve the problem of sleep deprivation. We conducted a 2-week randomized, open-label, crossover, self-controlled study in 14 healthy adults in their 30s to 50s (control: 1 week, AiR SI: 1 week), with sleep quality as the primary endpoint as well as salivary components (cortisol, melatonin) and autonomic nervous function (sympathetic nerve, parasympathetic nerve) as secondary endpoints. Trends toward improvement in sleep were suggested for all the endpoints, regardless of differences in the subject background. The results suggested that the use of AiR SI for 1 week not only improved sleep but nor-malized both the endocrine and autonomic functions as well. We conclude that, by using a sleeping mattress with a high sleep improvement effect, the user may easily achieve higher labor productivity and have a lower mortality risk. Eliminating sleep deprivation with a sleeping mattress may lead to a reduction in national economic losses and ultimately produce significant economic effects.


Introduction
According to the RAND Corporation, a nonprofit global policy think tank in the U.S., an individual that sleeps on average less than 6 hours per night has a 2.4 percentage point higher productivity loss in the workplace and a 13% higher mortality risk than an individual sleeping between 7 and 9 hours per night. Insufficient sleep among their populations costs the five OECD countries under consideration (U.S., UK, Germany, Japan, Canada) up to $680 billion of economic output every year. Among these countries, relative to the size of the overall economy, the estimated annual economic loss for Japan is the largest, which is $138 billion (2.92% of its GDP) [1]. In addition, the survey results reported by the Ministry of Health, Labor and Welfare show that almost half of the Japanese people in their 30s to 50s have an average sleeping time (average for the past month) of less than 6 hours, and about 30% of them are not obtaining sufficient rest by sleep, highlighting the fact that many Japanese people have insufficient sleep, especially around their 40s [2].
When calculated simply by labor productivity at the time of reporting by the RAND Corporation, if the sleep time of all people sleeping less than 6 hours was extended by one hour, an economic effect of up to $254 billion and $52 billion could be produced for the five OECD countries and Japan, respectively. Furthermore, the amount of economic loss is expected to increase to up to $775 billion for the five OECD countries and $156 billion (3.30% of GDP) for Japan in 2030 [1]. The economic effect of sleep improvement can be enormous, taking into account the reduced mortality risk as well.
Since insufficient sleep duration has been linked with the leading causes of death including heart disease, cancer, unintentional injuries, stroke, Alzheimer disease, diabetes, influenza, pneumonia, hypertension, and suicide [3], it affects not only the health condition of the individuals but also their work efficiency. Therefore, it is expected that sufficient rest by improving sleep will lower mortality risk and improve social function and labor productivity, eventually leading to curtailment of the economic loss at the national level in the future.
Sleep improvement with a sleeping mattress can be one of the simplest methods to realize a comfortable sleeping environment, but few reports have scientifically verified it. We conducted 2-week randomized, open-label, crossover, self-controlled study for exploratory purposes to examine the effects of a sleeping mattress on sleep quality from a medical perspective, using the "AiR SI (Nishikawa Co., Ltd.)" as the test sample and the individual sleeping mattress of each subject as the control (Figure 1). The AiR SI mattress has a special structure called a three-dimensional cross-slit, and is designed to provide a comfortable sleeping environment that supports the body in a balanced manner in both postures, sideways and supine.
Sleep quality assessment (evaluating "sleep quality" with eight sleep variables) was chosen as the primary endpoint. Since sleep is closely related to stress [4], and endocrine function involved in the stress response is improved by eliminating M. Omura et al. . This study was a 2-week randomized, open-label, crossover, self-controlled study. The subjects were randomly assigned to one of two patterns of changing the sleeping mattress from "control to AiR SI" or "AiR SI to control." Of the 2 weeks of the study period, subjects used the control mattress for 1 week and AiR SI for another week.
sleep deprivation [5], we decided to measure physiological changes that relate to the subjects' stress response as a secondary endpoint. Two neuroendocrine systems are involved in the expression of the stress response, the hypothalamic-pituitary-adrenal cortex system (HPA axis) and the hypothalamic-sympatheticadrenal-medullary system (SAM axis) [6] [7]. Thus, salivary components (cortisol, melatonin) were included as the HPA index, and autonomic nervous (sympathetic, parasympathetic) function as the SAM index.

1) Subjects
This study was conducted in 14 healthy adults in their 30s to 50s from whom consent was obtained, with the approval of the ethics committee of SOUKEN (Table 1). The inclusion criteria were: 1) have not used AiR SI or similar/competitive products within the past 3 months, 2) have not changed their sleeping mattress within the past 3 months, 3) do not smoke, 4) do not drink alcohol regularly, and 5) are not taking medication. Seven men and seven women who met the criteria were randomly assigned to one of two patterns of changing the sleeping mattress after 1 week from "control to AiR SI" or "AiR SI to control." This study was conducted at each subject's home, using their own pillows and comforters. After completion of the study, all the data were tabulated for each endpoint and compared between groups by designating the week using the control mattress as the control group and the week using the AiR SI as the AiR SI group.

2) Inspection Methods a) Experimental design
Since the experimental conditions other than the sleeping mattress could not be standardized, we instructed the subjects to avoid extreme changes in their life rhythm during the study period by waking up, having meals, taking a bath, and going to bed at the same time of day as much as possible in order to eliminate other effects on sleep, the endocrine system, and the autonomic nervous system. The subject wore an actigraph for 24 hours except while bathing for 14 days throughout the study period, without taking it off during sleep [8]. In this study, we used the "FS-760 (ACOS CO., LTD.)", a waist-mounted actigraph, for which a sleep/wake determination coincidence rate of 88.4% with sleep polysomnography (PSG) has been reported [9]. For the sleep/wake analysis based on the activity data, we used "SleepSignAct (KISSEI COMTEC CO., LTD.)", a sleep-wake rhythm research program. In this study, we compared changes in the following  c) Measurement of salivary components For saliva sampling, we used "Saliva Collection Aid (Salimetrics LLC)" to collect saliva [10] [11] [12]. For 14 days during the study period, subjects collected saliva four times a day by passive drool in the predetermined time windows [morning: upon waking up, noon: 11:00-13:00 (before lunch), evening: 17:00-19:00 (before meal), night: before going to bed], recording the saliva sampling time [13]. We instructed the subjects to immediately store the saliva samples in a frozen state, and retrieved them later. For analysis of the collected saliva samples, cortisol and melatonin in saliva were quantified by enzyme immunoassay (ELISA) using the "Salivary Cortisol Enzyme Immunoassay Kit (Salimetrics LLC)" and the "Salivary Melatonin Enzyme Immunoassay Kit (Salimetrics LLC) [14]". Absorbance was measured using the "Thermo Scientific Multiskan FC Absorbance Microplate Reader (Thermo Fisher Scientific K.K.)" with the absorbance filter set at 450 nm. In order to minimize the data variation, data from two wells per subject were quantified, and the mean value of absorbance was calculated. We compared the results between the control group and the AiR SI group, focusing on changes in the salivary concentrations of cortisol in the morning and melatonin at night. d) Analysis of the autonomic nervous system The autonomic nervous function was measured using the ultra-compact wearable heart rate sensor "myBeat WHS-1 (UNION TOOL CO.)". The heart rate periodically fluctuates under the influence of autonomic nervous activity related to breathing (respiratory sinus arrhythmia) and circulation (Mayer wave) [15] [16]. This "spontaneous fluctuation" is heart rate variability (HRV), which is separated into low-frequency (LF) components of 0.04 to 0.15 Hz and highfrequency (HF) components of 0.15 to 0.4 Hz by frequency analysis [17] [18]. As indices in this study, the LF/HF power ratio was used for the sympathetic activity level, and the HF component for the parasympathetic activity level [19] [20] [21]. The heart rate sensor was directly attached to the skin on the left chest using an adhesive electrode. In the event of rash, itching, and so forth at the attachment site, the position of the sensor was to be shifted. Also, the subjects were instructed to remove the sensor when bathing or getting extremely wet, and to replace the electrode seal with a new one when attaching the sensor again. The autonomic nervous function was measured only for 24 hours on the 4th day (Day 4) of the first week and the second week in this study, taking into account the possibility of adverse reactions at the sensor attachment site, the time lag until chronic stress affects the biological rhythm, the washout period for "carry-over effect", and so on. The measured data on autonomic nervous function were exported in CSV file format using "Advanced Viewer Software (WIN Frontier Co., Ltd.)". According to the saliva collection records submitted by the subjects, we selected and extracted continuous data on LF, HF, and LF/HF for several minutes before and after four respective timepoints of saliva collection (morning, noon, evening, night) from the 24-hour continuous data on Day 4 of each week using the control mattress or AiR SI, calculating the mean values at each timepoint by subject to compare changes in the autonomic nervous function between the control group and the AiR SI group.
3) Statistical Analyses A one-way repeated-measures analysis of variance was performed for the data on sleep quality assessment (each of the eight sleep variables), salivary cortisol concentration in the morning, and salivary melatonin concentration at night, whereas a two-way repeated-measures analysis of variance was performed for the autonomic function data on sympathetic nerve (LF/HF) and parasympathetic nerve (HF). In both cases, Bonferroni's method was used for multiple comparisons. For statistical analysis, IBM SPSS Statistics Version 22 was used, and the significance level was set at 5%.

Results
The number of valid subjects was twelve (5 males: 39.8 ± 7.5 years old, 7 females: 47.3 ± 9.8 years old). We adopted data from eight subjects for sleep quality assessment, excluding subjects with significant missing measurements or outliers.
1) Effect on sleep quality Regarding sleep quality assessment, Table 2 shows the activity measurement results of both the control group and the AiR SI group on Days 1, 4, and 7, and Figure 2 illustrates the comparison between the groups for each sleep variable on Day 7.  2) Effect on salivary components Table 3 shows the quantitative results of the salivary component concentrations in the control group and the AiR SI group, and Figure 3 and Figure 4 illustrate the changes in cortisol and melatonin levels by the group. In addition, Figure 5 and Figure 6 show a comparison of changes over time in the morning cortisol and night melatonin levels, respectively, between the groups.
The changes in salivary cortisol concentration over time were similar between the control group and the AiR SI group on Day 1. The cortisol levels were highest when waking up, and then decreased as the day progressed on Days 1, 4, and 7 in both groups (Figure 3(a) and Figure 3(b)). The cortisol levels in the morning, however, were almost unchanged in the control group over Days 1, 4, and 7, which, in contrast, gradually decreased to almost half the level of Day 1 in the AiR SI group on Day 7, with a reduced rate of 49.6%. There was statistical significance between the groups on Day 7 (P = 0.014), as well as within the AiR SI group between Day 1 and Day 7 (P = 0.027) ( Figure 5).
The morning and night salivary melatonin levels were similar between the groups on Day 1, whereas the morning level was lower and the night level was significantly higher in the AiR SI group on Day 7. None of the melatonin levels on Days 1, 4, and 7 showed a specific pattern in the control group (Figure 4(a)).     The night melatonin levels are compared between the groups. The error bars represent standard error. There is no statistical significance between any timepoints (P < 0.05).
On the other hand, they were low from morning through daytime, and increased rapidly from the evening toward the night, showing a specific pattern in the AiR SI group on Days 4 and 7 (Figure 4(b)). The night melatonin levels were almost unchanged over Days 1, 4, and 7 in the control group, whereas they gradually increased from Day 1 and peaked on Day 7 in the AiR SI group. Comparing Days 1 and 7 in the AiR SI group, the melatonin concentration at night increased five-fold (500.8%), but no statistical significance was observed (P = 0.311) ( Figure 6). Table 4 shows the measurement results of autonomic nervous function in the control group and the AiR SI group on Day 4. Figure 7 and Figure 8 illustrate changes in the activity levels of the sympathetic nerve (LF/HF) and the parasympathetic nerve (HF) by group, and Figure 9 and Figure 10 compare the respective activity levels at the four measurement time points (morning, noon, evening, night) between the groups.

3) Effect on the autonomic nervous function
In the control group, the sympathetic activity level was high from night (5.   For the sympathetic activity levels (LH/HF), there was statistical significance between the corresponding timepoints at noon (P = 0.014) and at night (P = 0.042) when comparing the control and AiR SI groups. Furthermore, statistical significance was observed between the morning and noon (P = 0.007) within the AiR SI group (Figure 9). As for the parasympathetic activity level (HF), there was no statistical significance between any of the corresponding timepoints when comparing the groups (P = 0.253 to 0.747) (Figure 10).  The parasympathetic activity levels (HF) are compared between the groups. The error bars represent standard error. There is no statistical significance between any timepoints (P < 0.05).

Discussion
It has been more than a decade since the concept of sleep debt was introduced [22], and nowadays lack of sleep is a serious problem worldwide [1]. We con- level is normally high in the morning and low from daytime to night-time due to a phenomenon known as the cortisol awakening response (CAR) [23] [24], in which the amount of cortisol secretion increases rapidly upon waking up in the morning, and it has been reported that those who are exposed to chronic or long-term stress have a higher rate of this increase [25].
It is highly likely that sleepiness was induced before bedtime in the AiR SI  [29]. The LF/HF value is less than 2.0 in a very resting state [30], and a decrease in "sleep quality" has been reported when it is 5.0 or more [31]. The autonomic rhythm was disturbed in the control group, whereas it showed a pattern along with the normal circadian rhythm in the AiR SI group. The results of this study not only confirm the normalization of autonomic rhythm by the use of AiR SI but also indicate the possibility of improved autonomic balance by suppressing excessive sympathetic activation and activating parasympathetic nerves.
The total sleep time (TST) on Day 7 in the control group and the AiR SI group was 5 hours 57 minutes and 6 hours 8 minutes, respectively, which was not a sufficient sleep time in either group. However, compared to the control group, the AiR SI group showed not only a tendency toward sleep improvement but also normalization in both autonomic and endocrine functions, which seemed to be due to reduction in stress and fatigue. By using the AiR SI for a week, the subjects appeared to be more active during the day, and resting and relaxed from bedtime to the morning. Therefore, it is assumed that use of the AiR SI has led to an improvement in "sleep quality" even though the improvement in total sleep time was slight.
In this study, although there were favorable tendencies in the AiR SI group when comparing the mean values between the groups, statistically significant results were not obtained with respect to sleep quality assessment, salivary melatonin concentration, or parasympathetic activity level. We believe these were mainly due to variation in the subject background. Consequently, in the next phase, we should conduct a longer-term study with a stricter design and further standardization of conditions in a larger number of subjects who have been narrowed down by thoroughly examining their background.
If it is demonstrated that the required length of sleep with sufficient quality can be secured by using the AiR SI, and its use spreads mainly in the population group with insufficient sleep, we can expect higher labor productivity and lower mortality risk in the users, with a possibility that this will ultimately lead to the curtailment of economic losses at the national level.

Conclusion
From the viewpoint of stress reduction and fatigue recovery of the users, the sleeping mattress can be an effective solution for alleviating sleep deprivation, a serious social problem worldwide, and this theme deserves further research in the future.