Chronobiology International, Early Online: 1–7, (2013) ! Informa Healthcare USA, Inc. ISSN: 0742-0528 print / 1525-6073 online DOI: 10.3109/07420528.2013.864299
ORIGINAL ARTICLE
Independent associations of exposure to evening light and nocturnal urinary melatonin excretion with diabetes in the elderly Kenji Obayashi,1 Keigo Saeki,1 Junko Iwamoto,2 Yoshito Ikada,3 and Norio Kurumatani1 1
Department of Community Health and Epidemiology, Nara Medical University School of Medicine, Nara, Japan, Department of Nursing, Tenri Health Care University, Nara, Japan, and 3Department of Surgery, Nara Medical University School of Medicine, Nara, Japan
Chronobiol Int Downloaded from informahealthcare.com by 123.227.89.20 on 12/11/13 For personal use only.
2
Circadian misalignment between internal and environmental rhythms dysregulates glucose homeostasis because of disruption of the biological clock, and increases risk of diabetes. Although exposure to evening light and decreased melatonin secretion are both associated with the circadian misalignment, it remains unclear whether they are associated with diabetes. In this cross-sectional study on 513 elderly individuals (mean age, 72.7 years), we measured ambulatory light intensity during the 4 h prior to bedtime at 1-min intervals during two consecutive days and overnight urinary 6-sulfatoxymelatonin excretion (UME) along with glucose metabolism. The median average intensity of evening light exposure and UME were 25.4 lux (interquartile range 17.5–37.6) and 6.6 mg (interquartile range 3.9–9.7), respectively. Both log-transformed average intensity of evening light exposure and log-transformed UME were significantly associated with diabetes in a multivariate logistic regression model adjusted for covariates, including gender, body mass index, duration in bed, and night-time light exposure [adjusted odds ratio (OR), 1.72; 95% confidence interval (CI), 1.12–2.64; p ¼ 0.01; and adjusted OR, 0.66; 95% CI, 0.44–0.97; p ¼ 0.04; respectively]. An increase in evening light exposure from 17.5 to 37.6 lux (25–75th percentiles) was associated with a 51.2% (95% CI, 8.2–111.4%) increase in prevalent diabetes, and an increase in UME from 3.9 to 9.7 mg (25–75th percentiles) was associated with a 32.0% (95% CI, 1.9–52.8%) decrease in prevalent diabetes. In conclusion, this study in elderly individuals demonstrated that evening light exposure in home settings and UME were significantly and independently associated with diabetes. Keywords: Circadian rhythm, diabetes, elderly, evening light, melatonin
INTRODUCTION
Wyse et al., 2011). Physiologically, exposure to LAN is the most important environmental cue for disruption of the SCN function, leading to circadian misalignment (Zeitzer et al., 2000). Melatonin is hypothesized to be a major contributor to the association between exposure to LAN and circadian misalignment; however, epidemiological studies have reported that exposure to LAN in home settings is not significantly associated with a decrease in nocturnal total melatonin levels (Davis et al., 2001; Levallois et al., 2001; Obayashi et al., 2012). Therefore, exposure to LAN in home settings may cause circadian misalignment and diabetes through a mechanism independent of changes in nocturnal total melatonin levels. According to the circadian phase-response curve to light (Khalsa et al., 2003), exposure to LAN before bedtime (evening light exposure) delays the subsequent
Circadian regulation of glucose homeostasis is controlled by the suprachiasmatic nucleus (SCN) of the hypothalamus, an essential component of the master biological clock (Asher & Schibler, 2011; Marcheva et al., 2010; Turek et al., 2005). Chronic misalignment between internal and environmental rhythms is typically found in night-shift workers who are exposed to increased light at night (LAN) and decreased melatonin secretion (Davis et al., 2012; Dumont et al., 2001). Epidemiologic data have demonstrated a higher risk of diabetes in night-shift workers (Mikuni et al., 1983; Morikawa et al., 2005; Pan et al., 2011; Suwazono et al., 2006, 2008). In modern society, increased exposure to LAN is observed not only in night-shift workers, but also in humans with normal circadian lifestyles because of the use of artificial lighting (Navara & Nelson, 2007;
Submitted April 15, 2013, Returned for revision October 7, 2013, Accepted November 6, 2013
Correspondence: Kenji Obayashi, MD, PhD, Department of Community Health and Epidemiology, Nara Medical University School of Medicine, 840 Shijocho, Kashiharashi, Nara 634-8521, Japan. Tel: þ81-744-22-3051. Fax: þ81-744-25-7657. E-mail: [email protected]
1
Chronobiol Int Downloaded from informahealthcare.com by 123.227.89.20 on 12/11/13 For personal use only.
2
K. Obayashi et al.
circadian phase more than light exposure in the other period of time. In previous experimental studies, evening light exposure is associated with alterations in the timing and amount of melatonin secretion, expression of clock genes, and timing of food intake (Cajochen et al., 2006; Gooley et al., 2011; Shuboni & Yan, 2010). Therefore, evening light exposure may be more closely associated with risk of diabetes than light exposure in the other period of time, although it remains unclear whether evening light exposure in home settings is associated with diabetes. Melatonin is a pineal gland hormone secreted predominantly at night. Melatonin secretion is involved in regulation of circadian internal rhythm, with decreased melatonin secretion being associated with circadian misalignment (Brzezinski, 1997). Previous experimental studies have shown that melatonin changes the plasma levels of insulin and leptin (Nishida et al., 2003; Wolden-Hanson et al., 2000) and that signaling by melatonin is associated with a risk of diabetes (Lyssenko et al., 2009). However, it remains unclear whether melatonin secretion is associated with diabetes in humans. In this cross-sectional study on 513 elderly individuals, we examined the associations of evening light exposure in home settings and melatonin secretion with prevalent diabetes. We measured ambulatory light intensity using a wrist light meter during the 4 h prior to bedtime as an index of evening light exposure. Urinary 6-sulfatoxymelatonin excretion (UME), the major metabolite of melatonin, was used as an index of melatonin secretion because there is evidence that it correlates closely with the secretion levels (Baskett et al., 1998).
MATERIALS AND METHODS Participants Between September 2010 and April 2012, 537 community-based elderly subjects were recruited with the cooperation of local resident associations and elderly resident clubs, and voluntarily enrolled in a study titled ‘‘Housing Environments and Health Investigation among Japanese Older People in Nara, Kansai Region: a prospective community-based cohort (HEIJO-KYO) study’’. Of these, 513 home-dwelling participants met the inclusion criteria of age 60 years and complete records of UME and evening light measurement. We selected this aging group with most retired their jobs to minimize the effect of current night-shift work on the associations of evening light exposure and UME with glucose metabolism. All the participants provided written informed consent, and the study protocol was performed in accordance with the ethics committee of Nara Medical University and the ethical standards of the Journal (Portaluppi et al., 2010).
Study protocol The protocols for measuring UME and light exposure were described in our previously reported study (Obayashi et al., 2012). In brief, we visited the participants’ homes, collected overnight fasting venous blood samples, and gathered demographic and medical information using a standardized questionnaire. In addition, we initiated 48-h measurements of light exposure and instructed the participants to collect their urine the following night and to maintain a standardized sleep diary logging the time they went to bed and the period of time they spent in bed. Measurements of light exposure Ambulatory evening and daytime light exposure were measured at 1-min intervals using a wrist light meter (Actiwatch 2; Respironics Inc., Pittsburgh, PA) that was worn on the non-dominant wrist. Values 51 lux during the out-of-bed period were considered to be missing data because of the clothing covering the sensor and were not included in the analyses (Scheuermaier et al., 2010). If the amount of time for which data was missing exceeded half of each period, the parameters were treated as blank data. Night-time light exposure was measured at 1-min intervals using a light meter (LX-28SD; Sato Shouji Inc., Kanagawa, Japan) with the sensor fixed at 60 cm above the floor, near the head of the bed and facing the ceiling. Three parameters of light exposure were defined as follows: evening light (ELavg), the average light intensity during the 4 h prior to bedtime, daytime light (DLavg), the average light intensity during the out-of-bed period and night-time light (NLavg), the average light intensity during the in-bed period. Urinary 6-sulfatoxymelatonin excretion The urine collection protocol involved discarding the last void at bedtime and collecting each subsequent void until the first morning void. The samples were stored in a dark bottle at room temperature, the total volume measured, and then stored at 20 C until assay. Urinary 6-sulfatoxymelatonin concentration was measured at a commercial laboratory (SRL, Inc., Tokyo, Japan) using a highly sensitive enzyme-linked immunosorbent assay kit (RE54031; IBL International, Hamburg, Germany). UME was calculated as follows: UME (mg) ¼ 6-sulfatoxymelatonin concentration (mg/mL) total overnight urine volume (mL). UME data were considered missing if the urine was not collected according to the protocol. The reproducibility of UME in the initial 188 participants was assessed by an additional collection of urine samples approximately four months later. The intraclass correlation coefficient between the two UME levels was 0.66 (95% CI, 0.57–0.73). Chronobiology International
Evening light, melatonin and diabetes
Chronobiol Int Downloaded from informahealthcare.com by 123.227.89.20 on 12/11/13 For personal use only.
Physical activity and sleep measurements Physical activity and sleep measurements were recorded at 1-min intervals for 48 h using an actigraph (Actiwatch 2). Four actigraphic parameters were determined using Actiware version 5.5 (Respironics Inc.) with a default wake threshold value (Philips Respironics Actiware Tutorials, 2013) used to identify periods of wake or sleep: (1) daytime physical activity, the average valid physical activity counts per minute during the outof-bed period; (2) total sleep time, the sleep time during the in-bed period; (3) sleep efficiency, the total sleep time divided by the duration in bed and (4) sleep-onset latency, the interval from bedtime to the first minute scored as sleep. Other measurements Body mass index (BMI) was calculated as weight (kg)/ height (m2). Current smoking status, habitual alcohol consumption and socioeconomic status, such as current household income and past education levels, were evaluated using a questionnaire. Venous blood samples were analyzed at a commercial laboratory (SRL Co. Inc., Tokyo, Japan) using a standard clinical chemistry analyzer to determine the concentrations of glycated hemoglobin (HbA1c) and fasting plasma glucose (FPG). Diabetes mellitus was diagnosed based on medical history, current diabetes treatment or if FPG was 7.0 mmol/L and HbA1c level was 6.5% of the National Glycohemoglobin Standardization Program value. The day length in Nara (latitude: 34 N) from sunrise to sunset on the measurement days was retrieved from the National Astronomical Observatory of Japan website (National Astronomical Observatory of Japan, 2013). Statistical analyses Variables with a normal distribution were expressed as the mean (SD), while variables with an asymmetric
distribution were reported as medians and interquartile ranges (IQRs). Means and medians were compared between the diabetes and non-diabetes groups using the unpaired t-test and Mann–Whitney U-test, respectively. The 2 test was used for comparison of categorical data. Univariate comparisons between the diabetes and non-diabetes groups included variables of age, gender, BMI, current smoking status, habitual alcohol consumption, current household income, past education levels, duration in bed, time to bed, day length, light exposure (evening, daytime and night-time light), UME and actigraphic parameters (daytime physical activity and sleep quality). The average of two consecutive days was used for further analysis of data on light exposure, actigraphy, sleep diary and day length. Data of evening and daytime light intensity, UME and sleep-onset latency were naturally log-transformed for analysis. For BMI and night-time light intensity which were not normally distributed even after natural log transformation, we used categorical analysis of tertile groups. Odds ratios (ORs) for diabetes were simultaneously adjusted for variables that were associated marginally to significantly with diabetes (p50.25) in the univariate comparisons (Tables 1 and 2). The statistical analyses were performed using SPSS version 19.0 for Windows (IBM SPSS Inc., Chicago, IL), with a two-sided p value of 50.05 being considered statistically significant.
RESULTS The mean age of the participants was 72.7 (SD, 6.5) years. Compared with the non-diabetes group, the diabetes group had significantly higher BMI (p50.01), more males (p ¼ 0.20) and longer duration in bed (p ¼ 0.21; Table 1). The median ELavg and UME were 25.4 lux (IQR, 17.5– 37.6) and 6.6 mg (IQR, 3.9–9.7), respectively. Compared
TABLE 1. Comparisons of demographic and clinical parameters between diabetes and non-diabetes. Variables
Diabetes (n ¼ 69)
Demographic parameters Age and mean, years Gender (male), number BMI and median, kg/m Current smoker, number Alcohol consumption (30 g/day), number Household income (4 billion JPY/year), number Past education (13 years), number
72.4 37 224.1 5 10 29 17
(6.1) (53.6) (21.5–26.1) (7.2) (14.5) (47.5) (24.6)
Clinical parameters FPG and mean, mg/dL HbA1c and mean, % HbA1c and mean, mmol/mol Duration in bed and mean, min Time to bed and mean, clock time Day length (511 h), number
129.9 6.6 49 507.2 22:24 34
(53.8) (1.0) (10.9) (80.0) (1:08) (49.3)
Non-diabetes (n ¼ 444) 72.8 201 22.7 19 53 192 109 95.6 5.3 34 494.8 22:31 217
p Value
(6.5) (45.3) (20.8–24.3) (4.3) (11.9) (43.2) (24.5)
0.66 0.20 50.01 0.28 0.55 0.88 0.99
(9.7) (0.4) (4.4) (75.8) (1:07) (48.9)
50.01 50.01 50.01 0.21 0.37 0.95
Data are expressed as means (SD), number (%), or median (interquartile range). BMI, body mass index; JPY, Japanese yen; FPG, fasting plasma glucose. t Test after natural log transformation. !
Informa Healthcare USA, Inc.
3
4
K. Obayashi et al. TABLE 2. Comparisons of light exposure, melatonin and actigraphic parameters between diabetes and non-diabetes. Diabetes (n ¼ 69)
Non-diabetes (n ¼ 444)
p Value
30.7 (20.1–38.8) 332.4 (192.4–753.3) 1.4 (0.2–5.9)
25.1 (16.9–37.2) 346.7 (174.0–755.6) 0.8 (0.1–3.3)
50.01a 0.76a 0.19
Variables Light exposure parameters Evening (ELavg) and median, lux Daytime (DLavg) and median, lux Night-time (NLavg) and median, lux Melatonin parameter UME and median, mg Actigraphic parameters Daytime physical activity and mean, counts/min Total sleep time and mean, min Sleep efficiency and mean, % Sleep-onset latency and median, min
6.2 (3.3–9.0) 286.2 443.7 83.7 21.0
0.17a
6.7 (4.0–9.8)
(88.4) (84.9) (8.8) (9.3–41.8)
297.6 437.3 84.2 19.5
(107.3) (78.5) (8.3) (10.5–37.0)
0.40 0.53 0.59 0.99a
Chronobiol Int Downloaded from informahealthcare.com by 123.227.89.20 on 12/11/13 For personal use only.
Data are expressed as mean (SD) or median (interquartile range). DLavg, average intensity of exposure to daytime light; ELavg, average intensity of exposure to evening light; NLavg, average intensity of exposure to night-time light; UME, urinary 6-sulfatoxymelatonin excretion. a t Test after natural log transformation. TABLE 3. Univariate and multivariate logistic regression analysis for the association between variables and diabetes. Variables Log-transformed ELavg (per log lux) Log-transformed UME (per log mg) Gender (male) BMI (per tertile increment) Duration in bed (per min) NLavg (per tertile increment)
Crude OR for diabetes
95% CI
1.77 0.77 1.40 1.65 1.002 1.17
1.17 0.53 0.84 1.19 0.999 0.85
2.67 1.12 2.33 2.28 1.005 1.61
p
Adjusted OR for diabetesa
95% CI
50.01 0.17 0.20 50.01 0.21 0.33
1.72 0.66 1.24 1.62 1.002 1.07
1.12 0.44 0.71 1.15 0.998 0.77
p Value 2.64 0.97 2.16 2.28 1.005 1.49
0.01 0.04 0.44 50.01 0.28 0.69
OR, odds ratio; CI, confidence interval; ELavg, average intensity of exposure to evening light; UME, urinary 6-sulfatoxymelatoninexcretion; BMI, body mass index; NLavg, average intensity of exposure to night-time light. a Adjusted for all covariates shown.
with the non-diabetes group, the diabetes group showed significantly higher ELavg (p50.01) and marginally higher NLavg (p ¼ 0.19) and lower UME (p ¼ 0.17; Table 2). The day-to-day correlation of ELavg was moderately high (Spearman’s rank correlation coefficient, 0.61). None of the actigraphic parameters differed marginally between the two groups. The multivariate model simultaneously adjusted for gender, BMI, duration in bed and NLavg (Table 3) showed that log-transformed ELavg and logtransformed UME were both significantly associated with diabetes (adjusted OR, 1.72; 95% CI, 1.12–2.64; p ¼ 0.01; and 0.66; 95% CI, 0.44–0.97; p ¼ 0.04; respectively). The dose-response slopes (Figure 1) were drawn by substituting the proportion of males, the mean categorical numbers for BMI and NLavg, and the means for duration in bed and log-transformed ELavg or log-transformed UME in the regression formula, are shown in Table 3. The slopes indicated that an increase in evening light exposure from 17.5 to 37.6 lux (25–75th percentiles) was associated with a 51.2% (95% CI, 8.2–111.4%) increase in prevalent diabetes (Figure 1A), and that an increase in UME from 3.9 to 9.7 mg (25–75th percentiles) was associated with a 32.0% (95% CI, 1.9–52.8%) decrease in prevalent diabetes (Figure 1B).
DISCUSSION This study demonstrated that evening light exposure in home settings and UME are significantly and independently associated with diabetes in elderly individuals, as evidenced by the fact that both ELavg and UME were significantly associated with diabetes in a multivariate model simultaneously adjusted for gender, BMI, duration in bed and NLavg (adjusted OR, 1.72; 95% CI, 1.12– 2.64; p ¼ 0.01; and adjusted OR, 0.66; 95% CI, 0.44–0.97; p ¼ 0.04; respectively). The magnitude of the influences of evening light exposure in home settings and UME on diabetes are shown by the dose-response slopes, which indicate that an increase in evening light exposure from 17.5 to 37.6 lux (25–75th percentiles) was associated with 51.2% (95% CI, 8.2–111.4%) increase in prevalent diabetes, and that an increase in UME from 3.9 to 9.7 mg (25–75th percentiles) was associated with 32.0% (95% CI, 1.9–52.8%) decrease in prevalent diabetes. In human physiology, these dose-response effects of evening light and endogenous melatonin, even at low levels, were reported in previous studies (Dijk & Cajochen, 1997; Zeitzer et al., 2000). These results are consistent with those of previous studies that demonstrated an association between night-shift workers and diabetes. Although epidemiologic data have demonstrated a higher risk of diabetes in Chronobiology International
Evening light, melatonin and diabetes (A)
5
(B) 1.3
2.5 1.2
1.1
1.0 Adjusted Odds Ratio for Diabetes
Adjusted Odds Ratio for Diabetes
2.0
1.5
.9
.8
.7
.6
1.0
Chronobiol Int Downloaded from informahealthcare.com by 123.227.89.20 on 12/11/13 For personal use only.
.5
.5
10
15
20
.4
37.6 lux (75th percentile)
17.5 lux (25th percentile) 25
30
35
40
45
Evening Light Exposure (lux)
.3
3.9 µg (25th percentile) 3
4
5
9.7 µg (75th percentile) 6
7
8
9
10
11
Urinary 6-sulfatoxymelatonin Excretion (µg)
FIGURE 1. The dose-response slopes for evening light exposure (A) and UME (B) in relation to the adjusted OR for diabetes in elderly individuals. The 95% CIs are indicated by dotted lines. Reference values of evening light exposure and UME indicate the 25th and 75th percentiles.
night-shift workers (Mikuni et al., 1983; Morikawa et al., 2005; Pan et al., 2011; Suwazono et al., 2006, 2008), these data were collected using a questionnaire, and the light intensity and endogenous melatonin levels experienced by the workers were not measured. To the best of our knowledge, this is the first study to report significant and independent associations of quantified evening light exposure in home settings and nocturnal melatonin levels with diabetes in the general elderly population. Our finding that ELavg showed a stronger association with diabetes than that shown by NLavg is consistent with the circadian phase-response curve to light, although both light exposures at either of these times may cause circadian misalignment (Cajochen et al., 2006; Gooley et al., 2011; Shuboni & Yan, 2010). The phase-response curve indicates that LAN from evening to midnight delays the circadian phase, whereas LAN from midnight to early morning advances it (Khalsa et al., 2003). Exposure to LAN during the in-bed period (NLavg) causes both circadian phase delay and advance, and in our previous and current study (Obayashi et al., 2013), we found no significant association between exposure to LAN during the in-bed period and diabetes. A significant association between UME and diabetes was observed in the multivariate model, although the association was marginal (p ¼ 0.17) in the univariate model. The covariates included in the multivariate model were reported to be associated with nocturnal melatonin levels. Nocturnal melatonin levels are higher in females than in males (Cain et al., 2010), and are !
Informa Healthcare USA, Inc.
known to be associated with the duration of darkness (Davis et al., 2001). There is also evidence that females with a higher BMI have lower nocturnal melatonin levels (Schernhammer et al., 2006) and that evening and nighttime light exposure suppresses nocturnal melatonin secretion (Gooley et al., 2011; Zeitzer et al., 2000). Although the effects of evening light exposure and melatonin secretion on diabetes are not fully understood, possible mechanisms are suggested by data regarding hormonal levels, circadian rhythm, sleep quality and sympathetic nerve activity. Evening light exposure delays the subsequent circadian phase, as demonstrated by the phase-response curve (Khalsa et al., 2003), and is associated with alterations in the timing and amount of melatonin secretion, expression of clock genes and timing of food intake (Cajochen et al., 2006; Gooley et al., 2011; Shuboni & Yan, 2010). Pinealectomy increases nocturnal plasma levels of insulin, which is reversed by administration of melatonin (Nishida et al., 2003; Wolden-Hanson et al., 2000). Melatonin secretion is involved in the regulation of circadian internal rhythm, with decreased melatonin secretion being associated with circadian misalignment, resulting in metabolic consequences (Asher & Schibler, 2011; Brzezinski, 1997). The circadian misalignment induced by evening light exposure or decreased melatonin secretion is also related to impaired sleep quality, as inadequate sleep is associated with decreased leptin levels, increased ghrelin levels, and incident diabetes (Gangwisch, 2009; Nilsson et al., 2004; Scheer et al., 2009; Spiegel et al., 2004). Furthermore, circadian
Chronobiol Int Downloaded from informahealthcare.com by 123.227.89.20 on 12/11/13 For personal use only.
6
K. Obayashi et al.
misalignment and sleep deprivation cause increased urinary and plasma catecholamine levels (Irwin et al., 1999; Scheer et al., 2009). Night-shift work diminishes the normal decrease in urinary norepinephrine and epinephrine levels during the non-work period (Yamasaki et al., 1998), and is also associated with increased low-frequency power and an increased ratio of low to high frequency in heart rate variability, indicating higher sympathetic nerve activity (Chung et al., 2009). The risk of diabetes may be reduced by decreasing the intensity of evening light exposure and/or increasing nocturnal melatonin levels. Causality of the associations of evening light exposure and UME with diabetes cannot be ascertained in our study as it is cross-sectional in nature. The effects of evening light exposure and nocturnal melatonin levels on diabetes are currently under investigation, with this association being supported by previous studies (Cajochen et al., 2006; Chung et al., 2009; Fonken et al., 2010; Gangwisch, 2009; Gooley et al., 2011; Irwin et al., 1999; Nilsson et al., 2004; Nishida et al., 2003; Scheer et al., 2009; Shuboni & Yan, 2010; Spiegel et al., 2004; Wolden-Hanson et al., 2000; Yamasaki et al., 1998). Nocturnal melatonin levels are influenced by daytime, evening and night-time light exposure (Gooley et al., 2011; Obayashi et al., 2012; Zeitzer et al., 2000), and evening exposure to moderately high-intensity artificial light is common in modern society (Navara & Nelson, 2007; Wyse et al., 2011). Further research that includes a longitudinal design and the definitions of evening light exposure and nocturnal melatonin levels used in the present study is required to better understand the associations of light exposure and nocturnal melatonin levels with diabetes. The strengths of the present study include quantification of evening light exposure and large sample of UME measurements. The present study had two limitations in addition to cross-sectional analysis. A limitation was that evening light exposure and UME were only measured for two consecutive days and a single night, respectively. However, we showed the reproducibility of evening light exposure and UME to be moderately high. Therefore, the values of evening light exposure and UME may be an accurate measure of the central tendency. Another limitation was non-random sampling because participants were recruited with the cooperation of local resident associations and elderly resident clubs, possibly leading to selection bias. However, the generalizability of our findings may be acceptable given that some basic data (e.g. BMI and estimated glomerular filtration rate) were consistent with those of the National Health and Nutrition Survey in Japan in 2010 (The National Health and Nutrition Survey Japan, 2010). Furthermore, the transmission rates of lenses are reduced by aging, and they would be potential intermediator underlying the association between evening light exposure and metabolic abnormality (Turner et al., 2010). However, considering younger population with
higher transmission rates of lenses, the possible effect of evening light exposure on glucose metabolism may be stronger in younger population than the aging population in this study. In conclusion, this study demonstrated that evening light exposure in home settings and UME are significantly and independently associated with diabetes in the general elderly population.
ACKNOWLEDGEMENTS We are indebted to all the participants of this study. We would also like to thank Sachiko Uemura and Naomi Takenaka for their valuable support during data collection.
DECLARATION OF INTEREST This work was supported by Grants from the Department of Indoor Environmental Medicine, Nara Medical University; Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology; Mitsui Sumitomo Insurance Welfare Foundation; Meiji Yasuda Life Foundation of Health and Welfare; Osaka Gas Group Welfare Foundation; Japan Diabetes Foundation; and the Japan Science and Technology Agency. All the authors report no conflicts of interest.
REFERENCES Asher G, Schibler U. (2011). Crosstalk between components of circadian and metabolic cycles in mammals. Cell Metab. 13: 125–37. Baskett JJ, Cockrem JF, Antunovich TA. (1998). Sulphatoxymelatonin excretion in older people: Relationship to plasma melatonin and renal function. J Pineal Res. 24:58–61. Brzezinski A. (1997). Melatonin in humans. N Engl J Med. 336: 186–95. Cain SW, Dennison CF, Zeitzer JM, et al. (2010). Sex differences in phase angle of entrainment and melatonin amplitude in humans. J Biol Rhythms. 25:288–96. Cajochen C, Jud C, Mu¨nch M, et al. (2006). Evening exposure to blue light stimulates the expression of the clock gene PER2 in humans. Eur J Neurosci. 23:1082–6. Chung MH, Kuo TB, Hsu N, et al. (2009). Sleep and autonomic nervous system changes – Enhanced cardiac sympathetic modulations during sleep in permanent night shift nurses. Scand J Work Environ Health. 35:180–7. Davis S, Kaune WT, Mirick DK, et al. (2001). Residential magnetic fields, light-at-night, and nocturnal urinary 6-sulfatoxymelatonin concentration in women. Am J Epidemiol. 154:591–600. Davis S, Mirick DK, Chen C, et al. (2012). Night shift work and hormone levels in women. Cancer Epidemiol Biomarkers Prev. 21:609–18. Dijk DJ, Cajochen C. (1997). Melatonin and the circadian regulation of sleep initiation, consolidation, structure, and the sleep EEG. J Biol Rhythms. 12:627–35. Dumont M, Benhaberou-Brun D, Paquet J. (2001). Profile of 24-h light exposure and circadian phase of melatonin secretion in night workers. J Biol Rhythms. 16:502–11. Chronobiology International
Chronobiol Int Downloaded from informahealthcare.com by 123.227.89.20 on 12/11/13 For personal use only.
Evening light, melatonin and diabetes Fonken LK, Workman JL, Walton JC, et al. (2010). Light at night increases body mass by shifting the time of food intake. Proc Natl Acad Sci USA. 107:18664–9. Gooley JJ, Chamberlain K, Smith KA, et al. (2011). Exposure to room light before bedtime suppresses melatonin onset and shortens melatonin duration in humans. J Clin Endocrinol Metab. 96:463–72. Gangwisch JE. (2009). Epidemiological evidence for the links between sleep, circadian rhythms and metabolism. Obes Rev. 10:37–45. Irwin M, Thompson J, Miller C, et al. (1999). Effects of sleep and sleep deprivation on catecholamine and interleukin-2 levels in humans: Clinical implications. J Clin Endocrinol Metab. 84: 1979–85. Khalsa SB, Jewett ME, Cajochen C, et al. (2003). A phase response curve to single bright light pulses in human subjects. J Physiol. 549:945–52. Levallois P, Dumont M, Touitou Y, et al. (2001). Effects of electric and magnetic fields from high-power lines on female urinary excretion of 6-sulfatoxymelatonin. Am J Epidemiol. 154: 601–9. Lyssenko V, Nagorny CL, Erdos MR, et al. (2009). Common variant in MTNR1B associated with increased risk of type 2 diabetes and impaired early insulin secretion. Nat Genet. 41:82–8. Marcheva B, Ramsey KM, Buhr ED, et al. (2010). Disruption of the clock components CLOCK and BMAL1 leads to hypoinsulinaemia and diabetes. Nature. 466:627–31. Mikuni E, Ohoshi T, Hayashi K, et al. (1983). Glucose intolerance in an employed population. Tohoku J Exp Med. 141 Suppl: 251–6. Morikawa Y, Nakagawa H, Miura K, et al. (2005). Shift work and the risk of diabetes mellitus among Japanese male factory workers. Scand J Work Environ Health. 31:179–83. National Astronomical Observatory of Japan. (2013). Available at: http://eco.mtk.nao.ac.jp/koyomi/index.html.en (accessed on April 7, 2013). Navara KJ, Nelson RJ. (2007). The dark side of light at night: physiological, epidemiological, and ecological consequences. J Pineal Res. 43:215–24. Nilsson PM, Ro¨o¨st M, Engstro¨m G, et al. (2004). Incidence of diabetes in middle-aged men is related to sleep disturbances. Diabetes Care. 27:2464–9. Nishida S, Sato R, Murai I, et al. (2003). Effect of pinealectomy on plasma levels of insulin and leptin and on hepatic lipids in type 2 diabetic rats. J Pineal Res. 35:251–6. Obayashi K, Saeki K, Iwamoto J, et al. (2012). Positive effect of daylight exposure on nocturnal urinary melatonin excretion in the elderly: A cross-sectional analysis of the HEIJO-KYO study. J Clin Endocrinol Metab. 97:4166–73. Obayashi K, Saeki K, Iwamoto J, et al. (2013). Exposure to light at night, nocturnal urinary melatonin excretion, and obesity/ dyslipidemia in the elderly: A cross-sectional analysis of the HEIJO-KYO study. J Clin Endocrinol Metab. 98:337–44.
!
Informa Healthcare USA, Inc.
7
Pan A, Schernhammer ES, Sun Q, et al. (2011). Rotating night shift work and risk of type 2 diabetes: Two prospective cohort studies in women. PLoS Med. 8:e1001141. Philips Respironics Actiware Tutorials. (2013). Available at: http:// learnactiware.com/tutorials/ (accessed on April 7, 2013). Portaluppi F, Smolensky MH, Touitou Y. (2010). Ethics and methods for biological rhythm research on animals and human beings. Chronobiol Int. 27:1911–29. Scheer FA, Hilton MF, Mantzoros CS, et al. (2009). Adverse metabolic and cardiovascular consequences of circadian misalignment. Proc Natl Acad Sci USA. 106:4453–8. Schernhammer ES, Kroenke CH, Dowsett M, et al. (2006). Urinary 6-sulfatoxymelatonin levels and their correlations with lifestyle factors and steroid hormone levels. J Pineal Res. 40:116–24. Scheuermaier K, Laffan AM, Duffy JF. (2010). Light exposure patterns in healthy older and young adults. J Biol Rhythms. 25: 113–22. Shuboni D, Yan L. (2010). Nighttime dim light exposure alters the responses of the circadian system. Neuroscience. 10:1172–8. Spiegel K, Tasali E, Penev P, et al. (2004). Sleep curtailment in healthy young men is associated with decreased leptin levels, elevated ghrelin levels, and increased hunger and appetite. Ann Intern Med. 141:846–50. Suwazono Y, Dochi M, Sakata K, et al. (2008). A longitudinal study on the effect of shift work on weight gain in male Japanese workers. Obesity. 16:1887–93. Suwazono Y, Sakata K, Okubo Y, et al. (2006). Long-term longitudinal study on the relationship between alternating shift work and the onset of diabetes mellitus in male Japanese workers. J Occup Environ Med. 48:455–61. The National Health and Nutrition Survey Japan. (2010). Available at: http://www.mhlw.go.jp/bunya/kenkou/eiyou/h22-houkoku.html (Japanese; accessed on April 8, 2013). Turek FW, Joshu C, Kohsaka A, et al. (2005). Obesity and metabolic syndrome in circadian Clock mutant mice. Science. 308:1043–5. Turner PL, Van Someren EJ, Mainster MA. (2010). The role of environmental light in sleep and health: Effects of ocular aging and cataract surgery. Sleep Med Rev. 14:269–80. Wyse CA, Selman C, Page MM, et al. (2011). Circadian desynchrony and metabolic dysfunction; did light pollution make us fat? Med Hypotheses. 77:1139–44. Wolden-Hanson T, Mitton DR, McCants RL, et al. (2000). Daily melatonin administration to middle-aged male rats suppresses body weight, intraabdominal adiposity, and plasma leptin and insulin independent of food intake and total body fat. Endocrinology. 141:487–97. Yamasaki F, Schwartz JE, Gerber LM, et al. (1998). Impact of shift work and race/ethnicity on the diurnal rhythm of blood pressure and catecholamines. Hypertension. 32:417–23. Zeitzer JM, Dijk D, Kronauer RE, et al. (2000). Sensitivity of the human circadian pacemaker to nocturnal light: Melatonin phase resetting and suppression. J Physiol. 526: 695–702.
ORIGINAL ARTICLE
Independent associations of exposure to evening light and nocturnal urinary melatonin excretion with diabetes in the elderly Kenji Obayashi,1 Keigo Saeki,1 Junko Iwamoto,2 Yoshito Ikada,3 and Norio Kurumatani1 1
Department of Community Health and Epidemiology, Nara Medical University School of Medicine, Nara, Japan, Department of Nursing, Tenri Health Care University, Nara, Japan, and 3Department of Surgery, Nara Medical University School of Medicine, Nara, Japan
Chronobiol Int Downloaded from informahealthcare.com by 123.227.89.20 on 12/11/13 For personal use only.
2
Circadian misalignment between internal and environmental rhythms dysregulates glucose homeostasis because of disruption of the biological clock, and increases risk of diabetes. Although exposure to evening light and decreased melatonin secretion are both associated with the circadian misalignment, it remains unclear whether they are associated with diabetes. In this cross-sectional study on 513 elderly individuals (mean age, 72.7 years), we measured ambulatory light intensity during the 4 h prior to bedtime at 1-min intervals during two consecutive days and overnight urinary 6-sulfatoxymelatonin excretion (UME) along with glucose metabolism. The median average intensity of evening light exposure and UME were 25.4 lux (interquartile range 17.5–37.6) and 6.6 mg (interquartile range 3.9–9.7), respectively. Both log-transformed average intensity of evening light exposure and log-transformed UME were significantly associated with diabetes in a multivariate logistic regression model adjusted for covariates, including gender, body mass index, duration in bed, and night-time light exposure [adjusted odds ratio (OR), 1.72; 95% confidence interval (CI), 1.12–2.64; p ¼ 0.01; and adjusted OR, 0.66; 95% CI, 0.44–0.97; p ¼ 0.04; respectively]. An increase in evening light exposure from 17.5 to 37.6 lux (25–75th percentiles) was associated with a 51.2% (95% CI, 8.2–111.4%) increase in prevalent diabetes, and an increase in UME from 3.9 to 9.7 mg (25–75th percentiles) was associated with a 32.0% (95% CI, 1.9–52.8%) decrease in prevalent diabetes. In conclusion, this study in elderly individuals demonstrated that evening light exposure in home settings and UME were significantly and independently associated with diabetes. Keywords: Circadian rhythm, diabetes, elderly, evening light, melatonin
INTRODUCTION
Wyse et al., 2011). Physiologically, exposure to LAN is the most important environmental cue for disruption of the SCN function, leading to circadian misalignment (Zeitzer et al., 2000). Melatonin is hypothesized to be a major contributor to the association between exposure to LAN and circadian misalignment; however, epidemiological studies have reported that exposure to LAN in home settings is not significantly associated with a decrease in nocturnal total melatonin levels (Davis et al., 2001; Levallois et al., 2001; Obayashi et al., 2012). Therefore, exposure to LAN in home settings may cause circadian misalignment and diabetes through a mechanism independent of changes in nocturnal total melatonin levels. According to the circadian phase-response curve to light (Khalsa et al., 2003), exposure to LAN before bedtime (evening light exposure) delays the subsequent
Circadian regulation of glucose homeostasis is controlled by the suprachiasmatic nucleus (SCN) of the hypothalamus, an essential component of the master biological clock (Asher & Schibler, 2011; Marcheva et al., 2010; Turek et al., 2005). Chronic misalignment between internal and environmental rhythms is typically found in night-shift workers who are exposed to increased light at night (LAN) and decreased melatonin secretion (Davis et al., 2012; Dumont et al., 2001). Epidemiologic data have demonstrated a higher risk of diabetes in night-shift workers (Mikuni et al., 1983; Morikawa et al., 2005; Pan et al., 2011; Suwazono et al., 2006, 2008). In modern society, increased exposure to LAN is observed not only in night-shift workers, but also in humans with normal circadian lifestyles because of the use of artificial lighting (Navara & Nelson, 2007;
Submitted April 15, 2013, Returned for revision October 7, 2013, Accepted November 6, 2013
Correspondence: Kenji Obayashi, MD, PhD, Department of Community Health and Epidemiology, Nara Medical University School of Medicine, 840 Shijocho, Kashiharashi, Nara 634-8521, Japan. Tel: þ81-744-22-3051. Fax: þ81-744-25-7657. E-mail: [email protected]
1
Chronobiol Int Downloaded from informahealthcare.com by 123.227.89.20 on 12/11/13 For personal use only.
2
K. Obayashi et al.
circadian phase more than light exposure in the other period of time. In previous experimental studies, evening light exposure is associated with alterations in the timing and amount of melatonin secretion, expression of clock genes, and timing of food intake (Cajochen et al., 2006; Gooley et al., 2011; Shuboni & Yan, 2010). Therefore, evening light exposure may be more closely associated with risk of diabetes than light exposure in the other period of time, although it remains unclear whether evening light exposure in home settings is associated with diabetes. Melatonin is a pineal gland hormone secreted predominantly at night. Melatonin secretion is involved in regulation of circadian internal rhythm, with decreased melatonin secretion being associated with circadian misalignment (Brzezinski, 1997). Previous experimental studies have shown that melatonin changes the plasma levels of insulin and leptin (Nishida et al., 2003; Wolden-Hanson et al., 2000) and that signaling by melatonin is associated with a risk of diabetes (Lyssenko et al., 2009). However, it remains unclear whether melatonin secretion is associated with diabetes in humans. In this cross-sectional study on 513 elderly individuals, we examined the associations of evening light exposure in home settings and melatonin secretion with prevalent diabetes. We measured ambulatory light intensity using a wrist light meter during the 4 h prior to bedtime as an index of evening light exposure. Urinary 6-sulfatoxymelatonin excretion (UME), the major metabolite of melatonin, was used as an index of melatonin secretion because there is evidence that it correlates closely with the secretion levels (Baskett et al., 1998).
MATERIALS AND METHODS Participants Between September 2010 and April 2012, 537 community-based elderly subjects were recruited with the cooperation of local resident associations and elderly resident clubs, and voluntarily enrolled in a study titled ‘‘Housing Environments and Health Investigation among Japanese Older People in Nara, Kansai Region: a prospective community-based cohort (HEIJO-KYO) study’’. Of these, 513 home-dwelling participants met the inclusion criteria of age 60 years and complete records of UME and evening light measurement. We selected this aging group with most retired their jobs to minimize the effect of current night-shift work on the associations of evening light exposure and UME with glucose metabolism. All the participants provided written informed consent, and the study protocol was performed in accordance with the ethics committee of Nara Medical University and the ethical standards of the Journal (Portaluppi et al., 2010).
Study protocol The protocols for measuring UME and light exposure were described in our previously reported study (Obayashi et al., 2012). In brief, we visited the participants’ homes, collected overnight fasting venous blood samples, and gathered demographic and medical information using a standardized questionnaire. In addition, we initiated 48-h measurements of light exposure and instructed the participants to collect their urine the following night and to maintain a standardized sleep diary logging the time they went to bed and the period of time they spent in bed. Measurements of light exposure Ambulatory evening and daytime light exposure were measured at 1-min intervals using a wrist light meter (Actiwatch 2; Respironics Inc., Pittsburgh, PA) that was worn on the non-dominant wrist. Values 51 lux during the out-of-bed period were considered to be missing data because of the clothing covering the sensor and were not included in the analyses (Scheuermaier et al., 2010). If the amount of time for which data was missing exceeded half of each period, the parameters were treated as blank data. Night-time light exposure was measured at 1-min intervals using a light meter (LX-28SD; Sato Shouji Inc., Kanagawa, Japan) with the sensor fixed at 60 cm above the floor, near the head of the bed and facing the ceiling. Three parameters of light exposure were defined as follows: evening light (ELavg), the average light intensity during the 4 h prior to bedtime, daytime light (DLavg), the average light intensity during the out-of-bed period and night-time light (NLavg), the average light intensity during the in-bed period. Urinary 6-sulfatoxymelatonin excretion The urine collection protocol involved discarding the last void at bedtime and collecting each subsequent void until the first morning void. The samples were stored in a dark bottle at room temperature, the total volume measured, and then stored at 20 C until assay. Urinary 6-sulfatoxymelatonin concentration was measured at a commercial laboratory (SRL, Inc., Tokyo, Japan) using a highly sensitive enzyme-linked immunosorbent assay kit (RE54031; IBL International, Hamburg, Germany). UME was calculated as follows: UME (mg) ¼ 6-sulfatoxymelatonin concentration (mg/mL) total overnight urine volume (mL). UME data were considered missing if the urine was not collected according to the protocol. The reproducibility of UME in the initial 188 participants was assessed by an additional collection of urine samples approximately four months later. The intraclass correlation coefficient between the two UME levels was 0.66 (95% CI, 0.57–0.73). Chronobiology International
Evening light, melatonin and diabetes
Chronobiol Int Downloaded from informahealthcare.com by 123.227.89.20 on 12/11/13 For personal use only.
Physical activity and sleep measurements Physical activity and sleep measurements were recorded at 1-min intervals for 48 h using an actigraph (Actiwatch 2). Four actigraphic parameters were determined using Actiware version 5.5 (Respironics Inc.) with a default wake threshold value (Philips Respironics Actiware Tutorials, 2013) used to identify periods of wake or sleep: (1) daytime physical activity, the average valid physical activity counts per minute during the outof-bed period; (2) total sleep time, the sleep time during the in-bed period; (3) sleep efficiency, the total sleep time divided by the duration in bed and (4) sleep-onset latency, the interval from bedtime to the first minute scored as sleep. Other measurements Body mass index (BMI) was calculated as weight (kg)/ height (m2). Current smoking status, habitual alcohol consumption and socioeconomic status, such as current household income and past education levels, were evaluated using a questionnaire. Venous blood samples were analyzed at a commercial laboratory (SRL Co. Inc., Tokyo, Japan) using a standard clinical chemistry analyzer to determine the concentrations of glycated hemoglobin (HbA1c) and fasting plasma glucose (FPG). Diabetes mellitus was diagnosed based on medical history, current diabetes treatment or if FPG was 7.0 mmol/L and HbA1c level was 6.5% of the National Glycohemoglobin Standardization Program value. The day length in Nara (latitude: 34 N) from sunrise to sunset on the measurement days was retrieved from the National Astronomical Observatory of Japan website (National Astronomical Observatory of Japan, 2013). Statistical analyses Variables with a normal distribution were expressed as the mean (SD), while variables with an asymmetric
distribution were reported as medians and interquartile ranges (IQRs). Means and medians were compared between the diabetes and non-diabetes groups using the unpaired t-test and Mann–Whitney U-test, respectively. The 2 test was used for comparison of categorical data. Univariate comparisons between the diabetes and non-diabetes groups included variables of age, gender, BMI, current smoking status, habitual alcohol consumption, current household income, past education levels, duration in bed, time to bed, day length, light exposure (evening, daytime and night-time light), UME and actigraphic parameters (daytime physical activity and sleep quality). The average of two consecutive days was used for further analysis of data on light exposure, actigraphy, sleep diary and day length. Data of evening and daytime light intensity, UME and sleep-onset latency were naturally log-transformed for analysis. For BMI and night-time light intensity which were not normally distributed even after natural log transformation, we used categorical analysis of tertile groups. Odds ratios (ORs) for diabetes were simultaneously adjusted for variables that were associated marginally to significantly with diabetes (p50.25) in the univariate comparisons (Tables 1 and 2). The statistical analyses were performed using SPSS version 19.0 for Windows (IBM SPSS Inc., Chicago, IL), with a two-sided p value of 50.05 being considered statistically significant.
RESULTS The mean age of the participants was 72.7 (SD, 6.5) years. Compared with the non-diabetes group, the diabetes group had significantly higher BMI (p50.01), more males (p ¼ 0.20) and longer duration in bed (p ¼ 0.21; Table 1). The median ELavg and UME were 25.4 lux (IQR, 17.5– 37.6) and 6.6 mg (IQR, 3.9–9.7), respectively. Compared
TABLE 1. Comparisons of demographic and clinical parameters between diabetes and non-diabetes. Variables
Diabetes (n ¼ 69)
Demographic parameters Age and mean, years Gender (male), number BMI and median, kg/m Current smoker, number Alcohol consumption (30 g/day), number Household income (4 billion JPY/year), number Past education (13 years), number
72.4 37 224.1 5 10 29 17
(6.1) (53.6) (21.5–26.1) (7.2) (14.5) (47.5) (24.6)
Clinical parameters FPG and mean, mg/dL HbA1c and mean, % HbA1c and mean, mmol/mol Duration in bed and mean, min Time to bed and mean, clock time Day length (511 h), number
129.9 6.6 49 507.2 22:24 34
(53.8) (1.0) (10.9) (80.0) (1:08) (49.3)
Non-diabetes (n ¼ 444) 72.8 201 22.7 19 53 192 109 95.6 5.3 34 494.8 22:31 217
p Value
(6.5) (45.3) (20.8–24.3) (4.3) (11.9) (43.2) (24.5)
0.66 0.20 50.01 0.28 0.55 0.88 0.99
(9.7) (0.4) (4.4) (75.8) (1:07) (48.9)
50.01 50.01 50.01 0.21 0.37 0.95
Data are expressed as means (SD), number (%), or median (interquartile range). BMI, body mass index; JPY, Japanese yen; FPG, fasting plasma glucose. t Test after natural log transformation. !
Informa Healthcare USA, Inc.
3
4
K. Obayashi et al. TABLE 2. Comparisons of light exposure, melatonin and actigraphic parameters between diabetes and non-diabetes. Diabetes (n ¼ 69)
Non-diabetes (n ¼ 444)
p Value
30.7 (20.1–38.8) 332.4 (192.4–753.3) 1.4 (0.2–5.9)
25.1 (16.9–37.2) 346.7 (174.0–755.6) 0.8 (0.1–3.3)
50.01a 0.76a 0.19
Variables Light exposure parameters Evening (ELavg) and median, lux Daytime (DLavg) and median, lux Night-time (NLavg) and median, lux Melatonin parameter UME and median, mg Actigraphic parameters Daytime physical activity and mean, counts/min Total sleep time and mean, min Sleep efficiency and mean, % Sleep-onset latency and median, min
6.2 (3.3–9.0) 286.2 443.7 83.7 21.0
0.17a
6.7 (4.0–9.8)
(88.4) (84.9) (8.8) (9.3–41.8)
297.6 437.3 84.2 19.5
(107.3) (78.5) (8.3) (10.5–37.0)
0.40 0.53 0.59 0.99a
Chronobiol Int Downloaded from informahealthcare.com by 123.227.89.20 on 12/11/13 For personal use only.
Data are expressed as mean (SD) or median (interquartile range). DLavg, average intensity of exposure to daytime light; ELavg, average intensity of exposure to evening light; NLavg, average intensity of exposure to night-time light; UME, urinary 6-sulfatoxymelatonin excretion. a t Test after natural log transformation. TABLE 3. Univariate and multivariate logistic regression analysis for the association between variables and diabetes. Variables Log-transformed ELavg (per log lux) Log-transformed UME (per log mg) Gender (male) BMI (per tertile increment) Duration in bed (per min) NLavg (per tertile increment)
Crude OR for diabetes
95% CI
1.77 0.77 1.40 1.65 1.002 1.17
1.17 0.53 0.84 1.19 0.999 0.85
2.67 1.12 2.33 2.28 1.005 1.61
p
Adjusted OR for diabetesa
95% CI
50.01 0.17 0.20 50.01 0.21 0.33
1.72 0.66 1.24 1.62 1.002 1.07
1.12 0.44 0.71 1.15 0.998 0.77
p Value 2.64 0.97 2.16 2.28 1.005 1.49
0.01 0.04 0.44 50.01 0.28 0.69
OR, odds ratio; CI, confidence interval; ELavg, average intensity of exposure to evening light; UME, urinary 6-sulfatoxymelatoninexcretion; BMI, body mass index; NLavg, average intensity of exposure to night-time light. a Adjusted for all covariates shown.
with the non-diabetes group, the diabetes group showed significantly higher ELavg (p50.01) and marginally higher NLavg (p ¼ 0.19) and lower UME (p ¼ 0.17; Table 2). The day-to-day correlation of ELavg was moderately high (Spearman’s rank correlation coefficient, 0.61). None of the actigraphic parameters differed marginally between the two groups. The multivariate model simultaneously adjusted for gender, BMI, duration in bed and NLavg (Table 3) showed that log-transformed ELavg and logtransformed UME were both significantly associated with diabetes (adjusted OR, 1.72; 95% CI, 1.12–2.64; p ¼ 0.01; and 0.66; 95% CI, 0.44–0.97; p ¼ 0.04; respectively). The dose-response slopes (Figure 1) were drawn by substituting the proportion of males, the mean categorical numbers for BMI and NLavg, and the means for duration in bed and log-transformed ELavg or log-transformed UME in the regression formula, are shown in Table 3. The slopes indicated that an increase in evening light exposure from 17.5 to 37.6 lux (25–75th percentiles) was associated with a 51.2% (95% CI, 8.2–111.4%) increase in prevalent diabetes (Figure 1A), and that an increase in UME from 3.9 to 9.7 mg (25–75th percentiles) was associated with a 32.0% (95% CI, 1.9–52.8%) decrease in prevalent diabetes (Figure 1B).
DISCUSSION This study demonstrated that evening light exposure in home settings and UME are significantly and independently associated with diabetes in elderly individuals, as evidenced by the fact that both ELavg and UME were significantly associated with diabetes in a multivariate model simultaneously adjusted for gender, BMI, duration in bed and NLavg (adjusted OR, 1.72; 95% CI, 1.12– 2.64; p ¼ 0.01; and adjusted OR, 0.66; 95% CI, 0.44–0.97; p ¼ 0.04; respectively). The magnitude of the influences of evening light exposure in home settings and UME on diabetes are shown by the dose-response slopes, which indicate that an increase in evening light exposure from 17.5 to 37.6 lux (25–75th percentiles) was associated with 51.2% (95% CI, 8.2–111.4%) increase in prevalent diabetes, and that an increase in UME from 3.9 to 9.7 mg (25–75th percentiles) was associated with 32.0% (95% CI, 1.9–52.8%) decrease in prevalent diabetes. In human physiology, these dose-response effects of evening light and endogenous melatonin, even at low levels, were reported in previous studies (Dijk & Cajochen, 1997; Zeitzer et al., 2000). These results are consistent with those of previous studies that demonstrated an association between night-shift workers and diabetes. Although epidemiologic data have demonstrated a higher risk of diabetes in Chronobiology International
Evening light, melatonin and diabetes (A)
5
(B) 1.3
2.5 1.2
1.1
1.0 Adjusted Odds Ratio for Diabetes
Adjusted Odds Ratio for Diabetes
2.0
1.5
.9
.8
.7
.6
1.0
Chronobiol Int Downloaded from informahealthcare.com by 123.227.89.20 on 12/11/13 For personal use only.
.5
.5
10
15
20
.4
37.6 lux (75th percentile)
17.5 lux (25th percentile) 25
30
35
40
45
Evening Light Exposure (lux)
.3
3.9 µg (25th percentile) 3
4
5
9.7 µg (75th percentile) 6
7
8
9
10
11
Urinary 6-sulfatoxymelatonin Excretion (µg)
FIGURE 1. The dose-response slopes for evening light exposure (A) and UME (B) in relation to the adjusted OR for diabetes in elderly individuals. The 95% CIs are indicated by dotted lines. Reference values of evening light exposure and UME indicate the 25th and 75th percentiles.
night-shift workers (Mikuni et al., 1983; Morikawa et al., 2005; Pan et al., 2011; Suwazono et al., 2006, 2008), these data were collected using a questionnaire, and the light intensity and endogenous melatonin levels experienced by the workers were not measured. To the best of our knowledge, this is the first study to report significant and independent associations of quantified evening light exposure in home settings and nocturnal melatonin levels with diabetes in the general elderly population. Our finding that ELavg showed a stronger association with diabetes than that shown by NLavg is consistent with the circadian phase-response curve to light, although both light exposures at either of these times may cause circadian misalignment (Cajochen et al., 2006; Gooley et al., 2011; Shuboni & Yan, 2010). The phase-response curve indicates that LAN from evening to midnight delays the circadian phase, whereas LAN from midnight to early morning advances it (Khalsa et al., 2003). Exposure to LAN during the in-bed period (NLavg) causes both circadian phase delay and advance, and in our previous and current study (Obayashi et al., 2013), we found no significant association between exposure to LAN during the in-bed period and diabetes. A significant association between UME and diabetes was observed in the multivariate model, although the association was marginal (p ¼ 0.17) in the univariate model. The covariates included in the multivariate model were reported to be associated with nocturnal melatonin levels. Nocturnal melatonin levels are higher in females than in males (Cain et al., 2010), and are !
Informa Healthcare USA, Inc.
known to be associated with the duration of darkness (Davis et al., 2001). There is also evidence that females with a higher BMI have lower nocturnal melatonin levels (Schernhammer et al., 2006) and that evening and nighttime light exposure suppresses nocturnal melatonin secretion (Gooley et al., 2011; Zeitzer et al., 2000). Although the effects of evening light exposure and melatonin secretion on diabetes are not fully understood, possible mechanisms are suggested by data regarding hormonal levels, circadian rhythm, sleep quality and sympathetic nerve activity. Evening light exposure delays the subsequent circadian phase, as demonstrated by the phase-response curve (Khalsa et al., 2003), and is associated with alterations in the timing and amount of melatonin secretion, expression of clock genes and timing of food intake (Cajochen et al., 2006; Gooley et al., 2011; Shuboni & Yan, 2010). Pinealectomy increases nocturnal plasma levels of insulin, which is reversed by administration of melatonin (Nishida et al., 2003; Wolden-Hanson et al., 2000). Melatonin secretion is involved in the regulation of circadian internal rhythm, with decreased melatonin secretion being associated with circadian misalignment, resulting in metabolic consequences (Asher & Schibler, 2011; Brzezinski, 1997). The circadian misalignment induced by evening light exposure or decreased melatonin secretion is also related to impaired sleep quality, as inadequate sleep is associated with decreased leptin levels, increased ghrelin levels, and incident diabetes (Gangwisch, 2009; Nilsson et al., 2004; Scheer et al., 2009; Spiegel et al., 2004). Furthermore, circadian
Chronobiol Int Downloaded from informahealthcare.com by 123.227.89.20 on 12/11/13 For personal use only.
6
K. Obayashi et al.
misalignment and sleep deprivation cause increased urinary and plasma catecholamine levels (Irwin et al., 1999; Scheer et al., 2009). Night-shift work diminishes the normal decrease in urinary norepinephrine and epinephrine levels during the non-work period (Yamasaki et al., 1998), and is also associated with increased low-frequency power and an increased ratio of low to high frequency in heart rate variability, indicating higher sympathetic nerve activity (Chung et al., 2009). The risk of diabetes may be reduced by decreasing the intensity of evening light exposure and/or increasing nocturnal melatonin levels. Causality of the associations of evening light exposure and UME with diabetes cannot be ascertained in our study as it is cross-sectional in nature. The effects of evening light exposure and nocturnal melatonin levels on diabetes are currently under investigation, with this association being supported by previous studies (Cajochen et al., 2006; Chung et al., 2009; Fonken et al., 2010; Gangwisch, 2009; Gooley et al., 2011; Irwin et al., 1999; Nilsson et al., 2004; Nishida et al., 2003; Scheer et al., 2009; Shuboni & Yan, 2010; Spiegel et al., 2004; Wolden-Hanson et al., 2000; Yamasaki et al., 1998). Nocturnal melatonin levels are influenced by daytime, evening and night-time light exposure (Gooley et al., 2011; Obayashi et al., 2012; Zeitzer et al., 2000), and evening exposure to moderately high-intensity artificial light is common in modern society (Navara & Nelson, 2007; Wyse et al., 2011). Further research that includes a longitudinal design and the definitions of evening light exposure and nocturnal melatonin levels used in the present study is required to better understand the associations of light exposure and nocturnal melatonin levels with diabetes. The strengths of the present study include quantification of evening light exposure and large sample of UME measurements. The present study had two limitations in addition to cross-sectional analysis. A limitation was that evening light exposure and UME were only measured for two consecutive days and a single night, respectively. However, we showed the reproducibility of evening light exposure and UME to be moderately high. Therefore, the values of evening light exposure and UME may be an accurate measure of the central tendency. Another limitation was non-random sampling because participants were recruited with the cooperation of local resident associations and elderly resident clubs, possibly leading to selection bias. However, the generalizability of our findings may be acceptable given that some basic data (e.g. BMI and estimated glomerular filtration rate) were consistent with those of the National Health and Nutrition Survey in Japan in 2010 (The National Health and Nutrition Survey Japan, 2010). Furthermore, the transmission rates of lenses are reduced by aging, and they would be potential intermediator underlying the association between evening light exposure and metabolic abnormality (Turner et al., 2010). However, considering younger population with
higher transmission rates of lenses, the possible effect of evening light exposure on glucose metabolism may be stronger in younger population than the aging population in this study. In conclusion, this study demonstrated that evening light exposure in home settings and UME are significantly and independently associated with diabetes in the general elderly population.
ACKNOWLEDGEMENTS We are indebted to all the participants of this study. We would also like to thank Sachiko Uemura and Naomi Takenaka for their valuable support during data collection.
DECLARATION OF INTEREST This work was supported by Grants from the Department of Indoor Environmental Medicine, Nara Medical University; Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology; Mitsui Sumitomo Insurance Welfare Foundation; Meiji Yasuda Life Foundation of Health and Welfare; Osaka Gas Group Welfare Foundation; Japan Diabetes Foundation; and the Japan Science and Technology Agency. All the authors report no conflicts of interest.
REFERENCES Asher G, Schibler U. (2011). Crosstalk between components of circadian and metabolic cycles in mammals. Cell Metab. 13: 125–37. Baskett JJ, Cockrem JF, Antunovich TA. (1998). Sulphatoxymelatonin excretion in older people: Relationship to plasma melatonin and renal function. J Pineal Res. 24:58–61. Brzezinski A. (1997). Melatonin in humans. N Engl J Med. 336: 186–95. Cain SW, Dennison CF, Zeitzer JM, et al. (2010). Sex differences in phase angle of entrainment and melatonin amplitude in humans. J Biol Rhythms. 25:288–96. Cajochen C, Jud C, Mu¨nch M, et al. (2006). Evening exposure to blue light stimulates the expression of the clock gene PER2 in humans. Eur J Neurosci. 23:1082–6. Chung MH, Kuo TB, Hsu N, et al. (2009). Sleep and autonomic nervous system changes – Enhanced cardiac sympathetic modulations during sleep in permanent night shift nurses. Scand J Work Environ Health. 35:180–7. Davis S, Kaune WT, Mirick DK, et al. (2001). Residential magnetic fields, light-at-night, and nocturnal urinary 6-sulfatoxymelatonin concentration in women. Am J Epidemiol. 154:591–600. Davis S, Mirick DK, Chen C, et al. (2012). Night shift work and hormone levels in women. Cancer Epidemiol Biomarkers Prev. 21:609–18. Dijk DJ, Cajochen C. (1997). Melatonin and the circadian regulation of sleep initiation, consolidation, structure, and the sleep EEG. J Biol Rhythms. 12:627–35. Dumont M, Benhaberou-Brun D, Paquet J. (2001). Profile of 24-h light exposure and circadian phase of melatonin secretion in night workers. J Biol Rhythms. 16:502–11. Chronobiology International
Chronobiol Int Downloaded from informahealthcare.com by 123.227.89.20 on 12/11/13 For personal use only.
Evening light, melatonin and diabetes Fonken LK, Workman JL, Walton JC, et al. (2010). Light at night increases body mass by shifting the time of food intake. Proc Natl Acad Sci USA. 107:18664–9. Gooley JJ, Chamberlain K, Smith KA, et al. (2011). Exposure to room light before bedtime suppresses melatonin onset and shortens melatonin duration in humans. J Clin Endocrinol Metab. 96:463–72. Gangwisch JE. (2009). Epidemiological evidence for the links between sleep, circadian rhythms and metabolism. Obes Rev. 10:37–45. Irwin M, Thompson J, Miller C, et al. (1999). Effects of sleep and sleep deprivation on catecholamine and interleukin-2 levels in humans: Clinical implications. J Clin Endocrinol Metab. 84: 1979–85. Khalsa SB, Jewett ME, Cajochen C, et al. (2003). A phase response curve to single bright light pulses in human subjects. J Physiol. 549:945–52. Levallois P, Dumont M, Touitou Y, et al. (2001). Effects of electric and magnetic fields from high-power lines on female urinary excretion of 6-sulfatoxymelatonin. Am J Epidemiol. 154: 601–9. Lyssenko V, Nagorny CL, Erdos MR, et al. (2009). Common variant in MTNR1B associated with increased risk of type 2 diabetes and impaired early insulin secretion. Nat Genet. 41:82–8. Marcheva B, Ramsey KM, Buhr ED, et al. (2010). Disruption of the clock components CLOCK and BMAL1 leads to hypoinsulinaemia and diabetes. Nature. 466:627–31. Mikuni E, Ohoshi T, Hayashi K, et al. (1983). Glucose intolerance in an employed population. Tohoku J Exp Med. 141 Suppl: 251–6. Morikawa Y, Nakagawa H, Miura K, et al. (2005). Shift work and the risk of diabetes mellitus among Japanese male factory workers. Scand J Work Environ Health. 31:179–83. National Astronomical Observatory of Japan. (2013). Available at: http://eco.mtk.nao.ac.jp/koyomi/index.html.en (accessed on April 7, 2013). Navara KJ, Nelson RJ. (2007). The dark side of light at night: physiological, epidemiological, and ecological consequences. J Pineal Res. 43:215–24. Nilsson PM, Ro¨o¨st M, Engstro¨m G, et al. (2004). Incidence of diabetes in middle-aged men is related to sleep disturbances. Diabetes Care. 27:2464–9. Nishida S, Sato R, Murai I, et al. (2003). Effect of pinealectomy on plasma levels of insulin and leptin and on hepatic lipids in type 2 diabetic rats. J Pineal Res. 35:251–6. Obayashi K, Saeki K, Iwamoto J, et al. (2012). Positive effect of daylight exposure on nocturnal urinary melatonin excretion in the elderly: A cross-sectional analysis of the HEIJO-KYO study. J Clin Endocrinol Metab. 97:4166–73. Obayashi K, Saeki K, Iwamoto J, et al. (2013). Exposure to light at night, nocturnal urinary melatonin excretion, and obesity/ dyslipidemia in the elderly: A cross-sectional analysis of the HEIJO-KYO study. J Clin Endocrinol Metab. 98:337–44.
!
Informa Healthcare USA, Inc.
7
Pan A, Schernhammer ES, Sun Q, et al. (2011). Rotating night shift work and risk of type 2 diabetes: Two prospective cohort studies in women. PLoS Med. 8:e1001141. Philips Respironics Actiware Tutorials. (2013). Available at: http:// learnactiware.com/tutorials/ (accessed on April 7, 2013). Portaluppi F, Smolensky MH, Touitou Y. (2010). Ethics and methods for biological rhythm research on animals and human beings. Chronobiol Int. 27:1911–29. Scheer FA, Hilton MF, Mantzoros CS, et al. (2009). Adverse metabolic and cardiovascular consequences of circadian misalignment. Proc Natl Acad Sci USA. 106:4453–8. Schernhammer ES, Kroenke CH, Dowsett M, et al. (2006). Urinary 6-sulfatoxymelatonin levels and their correlations with lifestyle factors and steroid hormone levels. J Pineal Res. 40:116–24. Scheuermaier K, Laffan AM, Duffy JF. (2010). Light exposure patterns in healthy older and young adults. J Biol Rhythms. 25: 113–22. Shuboni D, Yan L. (2010). Nighttime dim light exposure alters the responses of the circadian system. Neuroscience. 10:1172–8. Spiegel K, Tasali E, Penev P, et al. (2004). Sleep curtailment in healthy young men is associated with decreased leptin levels, elevated ghrelin levels, and increased hunger and appetite. Ann Intern Med. 141:846–50. Suwazono Y, Dochi M, Sakata K, et al. (2008). A longitudinal study on the effect of shift work on weight gain in male Japanese workers. Obesity. 16:1887–93. Suwazono Y, Sakata K, Okubo Y, et al. (2006). Long-term longitudinal study on the relationship between alternating shift work and the onset of diabetes mellitus in male Japanese workers. J Occup Environ Med. 48:455–61. The National Health and Nutrition Survey Japan. (2010). Available at: http://www.mhlw.go.jp/bunya/kenkou/eiyou/h22-houkoku.html (Japanese; accessed on April 8, 2013). Turek FW, Joshu C, Kohsaka A, et al. (2005). Obesity and metabolic syndrome in circadian Clock mutant mice. Science. 308:1043–5. Turner PL, Van Someren EJ, Mainster MA. (2010). The role of environmental light in sleep and health: Effects of ocular aging and cataract surgery. Sleep Med Rev. 14:269–80. Wyse CA, Selman C, Page MM, et al. (2011). Circadian desynchrony and metabolic dysfunction; did light pollution make us fat? Med Hypotheses. 77:1139–44. Wolden-Hanson T, Mitton DR, McCants RL, et al. (2000). Daily melatonin administration to middle-aged male rats suppresses body weight, intraabdominal adiposity, and plasma leptin and insulin independent of food intake and total body fat. Endocrinology. 141:487–97. Yamasaki F, Schwartz JE, Gerber LM, et al. (1998). Impact of shift work and race/ethnicity on the diurnal rhythm of blood pressure and catecholamines. Hypertension. 32:417–23. Zeitzer JM, Dijk D, Kronauer RE, et al. (2000). Sensitivity of the human circadian pacemaker to nocturnal light: Melatonin phase resetting and suppression. J Physiol. 526: 695–702.
