Abstract
Context
Seasonal variation in cold and light exposure may influence metabolic health.
Objective
We assessed the associations of bright sunlight and outdoor temperature with measures of glucose and lipid metabolism in two populations of middle-aged European subjects.
Design
Cross-sectional study.
Setting
Two population-based European cohorts.
Participants
Middle-aged nondiabetic subjects from the Oxford Biobank (OBB; N = 4327; mean age, 41.4 years) and the Netherlands Epidemiology of Obesity (NEO) study (N = 5899; mean age, 55.6 years).
Interventions
Data on outdoor bright sunlight and temperature collected from local weather stations.
Main Outcome Measures
Insulin resistance and fasting lipid levels. Multivariable regression analyses were adjusted for age, sex, percentage body fat, season, and either outdoor temperature or bright sunlight.
Results
In the OBB cohort, increased bright sunlight exposure was associated with lower fasting insulin [−1.27% (95% CI, −2.09 to −0.47%) per extra hour of bright sunlight], lower homeostatic model assessment for insulin resistance (−1.36%; 95% CI, −2.23 to −0.50), lower homeostatic model assessment for β-cell function (−0.80%; 95% CI, −1.31 to −0.30), and lower triglyceride (−1.28%; 95% CI, −2.07 to −0.50) levels. In the NEO cohort generally unidirectional but weaker associations were observed. No associations between outdoor temperature and measures of glucose or lipid metabolism were detected following adjustment for bright sunlight.
Conclusions
Bright sunlight, but not outdoor temperature, might be associated with increased insulin sensitivity and lower triglyceride levels.
Outdoor temperature and sunlight were studied in relationship to cardiometabolic traits, and it was found that specifically outdoor sunlight showed beneficial associations with insulin resistance.
Several studies have been conducted to date to identify modifiable (e.g., lifestyle) and nonmodifiable (e.g., genetic variation) risk factors for developing cardiometabolic disease. Recent preclinical and clinical studies have linked environmental temperature to cardiometabolic health, with thermogenesis in muscle and brown adipose tissue as potential mediators (1–7). For example, Blauw et al. (7) reported an increased incidence of type 2 diabetes (T2D) with higher mean annual outdoor temperature both across US states and worldwide. Similarly, Speakman and Heidari-Bakavoli (8) calculated that ambient temperature explained 12.4% of the variation in the prevalence of T2D across the mainland United States after accounting for obesity, poverty, and race. These findings may be attributable to lifestyle adaptations to changes in outdoor temperature, but it has also been hypothesized that metabolism can rapidly adapt to changes in environmental temperature. For example, seasonal fluctuations in measures of insulin sensitivity were reported in the Rotterdam study cohort, which could only partly be attributed to seasonal changes in lifestyle (9, 10). Furthermore, using a hyperinsulinemic-euglycemic clamp, researchers in another study (11) found that insulin sensitivity was lowest during winter in addition to demonstrating a positive effect of outdoor temperature on insulin sensitivity irrespective of season. More directly, Hanssen et al. (5) showed that short-term cold acclimation improves insulin sensitivity in patients with T2D, an effect that was associated with markedly increased basal skeletal muscle GLUT4 translocation.
As outdoor temperature is associated with outdoor bright sunlight, the interpretation of these studies is complicated. Increased exposure to bright sunlight can increase nocturnal melatonin levels (12), which in turn inhibit pancreatic insulin secretion (13). Vitamin D, which is primarily produced in the skin in response to sunlight exposure, has also been linked with the development of insulin resistance (14). Additionally, brown adipose tissue exhibits a daily rhythm in glucose and lipid uptake (15, 16), which at least in rodents is modulated by the duration of daily light exposure (16). In this study, we aimed to examine the associations of outdoor temperature and bright sunlight with measures of body composition, insulin sensitivity, and fasting lipid concentrations in a combined cohort of >10,000 middle-aged healthy European subjects enrolled in the Oxford Biobank (OBB) and the Netherlands Epidemiology of Obesity (NEO) study.
Methods
Study settings
The OBB is a population-based cohort of randomly selected healthy participants aged 30 to 50 years from Oxfordshire, UK. Individuals with a history of myocardial infarction, diabetes mellitus, heart failure, untreated malignancy, other ongoing systemic diseases, or ongoing pregnancy were not eligible for study inclusion. Enrolment of participants started in 1999 and as of May 2015 the OBB cohort comprised 7185 individuals (4054 women and 3131 men). More detailed descriptions of the study recruitment criteria and population characteristics are reported elsewhere (17).
The NEO study is a population-based prospective cohort study of men and women aged between 45 and 65 years with an oversampling of individuals with a body mass index (BMI) ≥27 kg/m2 living in the greater area of Leiden (in the west of the Netherlands). Additionally, all inhabitants aged between 45 and 65 years from one municipality (Leiderdorp) were invited in the study irrespective of their BMI to allow for a reference distribution of BMI. Between September 2008 and September 2012, 6671 individuals were included in the study. Detailed information about the study design and data collection has been described previously (18).
In both cohorts, participants were invited for a detailed baseline assessment, conducted after an overnight fast, which included blood sampling and anthropometry. Both studies were approved by local ethics committees, and written informed consent was obtained from all study participants.
Study population
In the OBB, we excluded individuals with missing data on mean outdoor temperature and bright sunlight on the day of the study visit (N = 406) and individuals with missing data on body composition (N = 2452). From the NEO study, we excluded individuals with both treated and diagnosed diabetes, as well as subjects with a fasting glucose concentration >7.0 mmol/L (N = 749), to have a uniform population, in terms of glycemic status, as that of the OBB. Additionally, we excluded participants who were not fasted (self-reported; N = 23).
Data collection on outdoor temperature and bright sunlight
Data on mean outdoor temperature and hours of bright sunlight (defined as global radiation >120 W/m2) were collected from local weather stations. Based on these data, we calculated the mean outdoor temperature and bright sunlight duration during a 7- and 30-day period before the date of blood sampling. For the OBB, data were obtained from the Radcliffe Meteorological Station (School of Geography and the Environment, University of Oxford, Oxford, UK) whereas for the NEO study we obtained meteorological data from the Royal Dutch Meteorological Institute, which has the closest proximity to the city of Leiden.
Laboratory assays
In the OBB, fasting glucose, total cholesterol, and triglyceride concentrations were measured in plasma using Instrumentation Laboratory IL Test™ kits on ILab 600/650 clinical chemistry analyzers (Werfen, Warrington, UK). High-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol levels were measured in plasma using the Randox direct clearance method adapted for use on the ILab 600/650 analyzers (Randox Laboratories, Crumlin, UK). Fasting insulin levels were measured with a Millipore human insulin–specific radioimmunoassay (Millipore UK, Watford, UK).
In the NEO study, fasting serum glucose, total cholesterol, and triglyceride concentrations were determined by enzymatic colorimetric methods [Roche Modular Analytics P800; Roche Diagnostics, Mannheim, Germany; coefficient of variation (CV) <5%]. HDL cholesterol concentration was measured with the homogeneous HDL cholesterol method (third generation; Roche Modular Analytics P800; Roche Diagnostics; CV <5%). Serum insulin concentration was measured with an immunometric method (Siemens Immulite 2500; Siemens Healthcare Diagnostics, Breda, Netherlands; CV <5%). LDL cholesterol concentration was estimated using the Friedewald equation (19).
In both cohorts, the homeostatic model assessment for insulin resistance (HOMA-IR) was calculated as [fasting insulin (μU/mL)] × [fasting glucose (mmol/L)]/22.5, and the homeostatic model assessment for β-cell function (HOMA-B) was calculated as [20 × fasting insulin (μU/mL)]/[glucose (mmol/L) − 3.5] (20, 21).
Measurements
In the OBB, body fat percentage was determined by dual-energy X-ray absorptiometry using a GE Lunar iDXA. All data were analyzed with Encore software (version 11.0; GE Medical Systems, Madison, WI). For the NEO study, body weight and percentage body fat were measured by a Tanita bioimpedance balance (TBF-310; Tanita International Division, Yiewsley, Middlesex, UK). Height and weight were measured by research nurses at the OBB and NEO study centers. BMI was calculated by dividing the weight in kilograms by the height in meters squared. Season was derived from the date of the blood sampling (winter, December through February; spring, March through May; summer, June through August; autumn, September through November). Use of lipid-lowering medication was determined from the medication inventory.
Statistical analyses
In the NEO study, individuals with a BMI ≥27 kg/m2 were oversampled. To correctly represent associations for the general population (22), adjustments for the oversampling of participants with a BMI ≥27 kg/m2 were made. This was done by weighting all participants toward the BMI distribution of participants from the Leiderdorp municipality (23), whose BMI distribution was similar to the BMI distribution of the general Dutch population (18). All results in the NEO study were based on weighted analyses. Consequently, the results apply to a population-based study without oversampling of individuals with a BMI ≥27 kg/m2. As a result of the weighting procedure, the numbers of participants per group in the NEO study are presented as percentages.
All analyses in the OBB and the NEO study were performed using Stata version 12.1 (StataCorp, College Station, TX). Characteristics of the OBB and NEO study populations separately were expressed as (weighted) mean (with SD), (weighted) median (interquartile range), or (weighted) proportion (percentage).
As most of the outcome variables were not normally distributed, we log-transformed all outcome variables to be able to better compare the effect sizes of the different study outcomes. Associations of mean bright sunlight and temperature with measures of insulin resistance and dyslipidemia were examined using multivariable linear regression analyses separately for the OBB and NEO study populations. Estimates retrieved from the analyses in the OBB and NEO study populations were subsequently meta-analyzed using inverse-variance weighting using the rmeta statistical package in R. β estimates from the separate results and results from the meta-analyses were back-transformed and were expressed as the percentage difference in outcome variable per unit increase in outdoor temperature or bright sunlight (95% CI). Model 1 was adjusted for age, sex, and percentage body fat. Model 2 was additionally adjusted for season. Model 3 was additionally adjusted either for the mean temperature or mean hours of bright sunlight. In sensitivity analyses on blood lipid levels, we further adjusted for use of lipid-lowering medication. We additionally adjusted all associations for BMI rather than for percentage body fat. Furthermore, we repeated the analyses using a mean temperature/bright sunlight exposure during 30 days as exposure. For presentation purposes, we analyzed the data per 5°C increase in temperature and per hour increase in sunlight exposure.
Results
Characteristics of the study populations
The total study population (N = 10,226) was composed of 4327 individuals from the OBB and 5899 individuals from the NEO study (Table 1). Participants from the OBB were younger (mean age, 41.4 vs 55.6 years) and had higher mean percentage body fat (42.1% vs 31.5%) compared with participants from the NEO study. Median insulin concentrations (11.5 vs 7.5 mIU/L), median HOMA-IR values (2.6 vs 1.8), and median HOMA-B values (106 vs 87) were higher in the OBB than in the NEO cohort, whereas use of lipid-lowering medication was higher in the NEO cohort (8.2% vs 0.7%). Mean outdoor temperature was similar in the two geographic regions, although the median number of hours of bright sunlight duration was lower in Oxfordshire than in the Leiden area during the study period (3.8 vs 5.0 hours).
Table 1.
Characteristics of the Study Populations
OBB (N = 4327) | NEO Study (N = 5899) | |
---|---|---|
Age, y, mean (SD) | 41.4 (5.9) | 55.6 (6.0) |
Men, % | 43.2 | 42.8 |
BMI, kg/m2, mean (SD) | 25.9 (4.7) | 26.1 (4.3) |
Percentage body fat, mean (SD) | 42.1 (12.0) | 31.5 (8.6) |
Glucose, mmol/L, mean (SD) | 5.2 (0.5) | 5.3 (0.5) |
Insulin, mIU/L, median (IQR) | 11.5 (8.7, 15.2) | 7.5 (5.1, 11.3) |
HOMA-IR, median (IQR) | 2.6 (1.9, 3.6) | 1.8 (1.2, 2.7) |
HOMA-B, median (IQR) | 106 (89, 126) | 87 (61, 126) |
LDL cholesterol, mmol/L, mean (SD) | 3.2 (1.2) | 3.6 (1.0) |
HDL cholesterol, mmol/L, mean (SD) | 1.4 (0.4) | 1.6 (0.5) |
Triglycerides, mmol/L, median (IQR) | 0.9 (0.7, 1.3) | 1.0 (0.7; 1.5) |
Use of lipid-lowering medication, % | 0.7 | 8.2 |
Winter, % | 24.5 | 24.0 |
Spring, % | 25.0 | 26.5 |
Summer, % | 23.1 | 24.3 |
Autumn, % | 27.4 | 25.3 |
Outdoor temperature (7 d), °C, mean (SD) | 10.6 (5.2) | 10.5 (5.6) |
Outdoor bright sunlight, h (7 d), median (IQR) | 3.8 (2.4, 5.9) | 5.0 (2.7, 7.0) |
Results of the NEO study population are weighted toward the body mass distribution of the general population.
Abbreviation: IQR, interquartile range.
Temperature and bright sunlight during a representative 3-month period for Oxfordshire and Leiden are presented in an online repository (24). Whereas bright sunlight duration fluctuated markedly on a day-to-day basis, variation in temperature was less pronounced. As expected, a higher number of hours of bright sunlight was associated with a higher outdoor temperature [1.30°C per 1 extra hour of bright sunlight (95% CI, 1.25 to 1.35) in Oxfordshire and 0.96 per 1 extra hour of bright sunlight (95% CI, 0.91 to 1.01) in Leiden]. Independent of season, outdoor bright sunlight was associated with a lower BMI in the OBB cohort (−0.09 kg/m2 per 1 hour; 95% CI, −0.16 to −0.02) and a higher BMI in the NEO cohort (0.06 kg/m2 per 1 hour; 95% CI, 0.00 to 0.12). No association between duration of bright sunlight and percentage total body fat was observed in either cohort. Independent of season, a higher outdoor temperature was associated with a lower BMI in both the OBB cohort (−0.19 kg/m2 per 5°C; 95% CI, −0.43 to 0.05) and the NEO cohort (−0.44 kg/m2 per 5°C; 95% CI, −0.62 to −0.25). Directionally similar associations were observed for percentage total body fat in both populations.
Outdoor temperature and glucose and lipid metabolism
Cohort-specific results for associations between outdoor temperature and measures of glucose and lipid metabolism are presented in Table 2. After adjustment for age, sex, and percentage body fat, higher mean outdoor temperature during the 7 days prior to the center visit was associated with lower HOMA-IR (−1.42% per 5°C increase in mean outdoor temperature; 95% CI, −2.86 to −0.01) in the OBB. This association became stronger after additional adjustment for season (−3.11%; 95% CI, −5.74 to −0.54) and was primarily driven by lower glucose levels at higher mean outdoor temperatures (−0.51%; 95% CI, −0.92 to −0.09). However, both associations attenuated toward the null after additional adjustment for mean bright sunlight. In the NEO cohort, no associations between mean outdoor temperature and glucose or lipid metabolism were detected in any of the models studied.
Table 2.
Association Between Outdoor Temperature and Measures of Glucose and Lipid Metabolism in the OBB and NEO Study Populations
Model 1 | Model 2 | Model 3 | |
---|---|---|---|
Glucose metabolism | |||
Glucose, %, OBB | −0.06 (−0.29 to 0.17) | −0.51 (−0.92 to −0.09) | −0.38 (−0.85 to 0.09) |
Glucose, %, NEO study | 0.22 (−0.09 to 0.52) | 0.19 (−0.36 to 0.73) | 0.23 (−0.33 to 0.78) |
Insulin, %, OBB | −1.10 (−2.44 to 0.22) | −2.22 (−4.67 to 0.17) | −0.79 (−3.50 to 1.84) |
Insulin, %, NEO study | −0.37 (−2.21 to 1.44) | −0.94 (−3.85 to 1.95) | −0.69 (−3.70 to 2.41) |
HOMA-IR, %, OBB | −1.42 (−2.86 to −0.01) | −3.11 (−5.74 to −0.54) | −1.47 (−4.38 to 1.36) |
HOMA-IR, %, NEO study | −0.15 (−2.08 to 1.82) | −0.75 (−3.75 to 2.34) | −0.46 (−3.81 to 2.76) |
HOMA-B, %, OBB | −0.65 (−1.48 to 0.18) | −0.53 (−2.04 to 0.96) | 0.19 (−1.49 to 1.83) |
HOMA-B, %, NEO study | −0.85 (−2.56 to 0.82) | −1.47 (−4.55 to 1.52) | −1.25 (−4.49 to 1.88) |
Lipid metabolism | |||
LDL cholesterol, %, OBB | −0.20 (−0.91 to 0.50) | −0.33 (−1.61 to 0.93) | −0.08 (−1.49 to 1.32) |
LDL cholesterol, %, NEO study | 0.00 (−0.86 to 0.84) | 1.26 (−0.26 to 2.76) | 1.46 (−0.14 to 3.04) |
HDL cholesterol, %, OBB | −0.19 (−0.92 to 0.53) | −0.59 (−1.91 to 0.71) | −0.56 (−2.04 to 0.90) |
HDL cholesterol, %, NEO study | −0.76 (−1.55 to 0.03) | −0.81 (−2.20 to 0.56) | −0.83 (−2.28 to 0.60) |
Triglycerides, %, OBB | 0.08 (−1.21 to 1.35) | −2.02 (−4.40 to 0.31) | −0.31 (−2.93 to 2.25) |
Triglycerides, %, NEO study | −0.37 (−1.97 to 1.21) | −0.25 (−3.17 to 2.59) | 0.38 (−2.58 to 3.27) |
Presented associations are from the OBB (N = 4327) and NEO study (N = 5899) populations. Results from the NEO study are weighted toward BMI distribution of the general population. Associations are presented as the percentage difference in the outcome measure per 5°C increased outdoor temperature during the 7 d before the center visit (95% CI). In model 1, analyses were adjusted for age, sex, and percentage body fat. In model 2, analyses were adjusted for age, sex, percentage body fat, and season. In model 3, analyses were adjusted for age, sex, percentage body fat, season, and outdoor bright sunlight.
In the meta-analysis (24), only the association between mean outdoor temperature and HOMA-IR persisted after adjustment for age, sex, percentage body fat, and season (−2.15%; 95% CI, −4.16 to −0.18). Additional adjustment for mean daily hours of outdoor bright sunlight attenuated this association (−1.04%; 95% CI, −3.22 to 1.10). Results were not materially different when we adjusted associations for BMI rather than percentage body fat. When taking into account the mean outdoor temperature during the 30 days prior to the center visit in the basic model (24), the association with HOMA-IR became stronger (−1.22%; 95% CI, −2.44 to −0.01). No associations between mean outdoor temperature and glucose or lipid levels were detected, although, as highlighted in the preceding paragraph, we observed heterogeneity between the two cohorts.
Bright sunlight and glucose and lipid metabolism
Cohort-specific results on the associations between bright sunlight hours and measures of glucose and lipid metabolism are presented in Table 3. In the OBB, after adjustment for age, sex, and percentage body fat, longer mean bright sunlight duration was associated with lower HOMA-IR (−1.05%; 95% CI, −1.65 to −0.45) and lower HOMA-B (−0.54%; 95% CI, −0.89 to −0.19). Both of these associations were driven by lower fasting insulin levels (−0.92% per hour increase in bright sunlight; 95% CI, −1.48 to −0.36) in the presence of prolonged sunshine. Longer mean bright sunlight hours were additionally associated with lower LDL cholesterol (−0.36%; 95% CI, −0.66 to −0.07) and lower triglyceride levels (−0.57%; 95% CI, −1.12 to −0.03). With the exception of LDL cholesterol, all associations persisted and indeed became stronger after adjustment for season and mean outdoor temperature. No association between outdoor bright sunlight hours and glucose concentration was detected. In the NEO cohort, associations between bright sunlight hours and glucose and lipid traits were in the same direction as those observed in the OBB but with smaller effect sizes.
Table 3.
Association Between Outdoor Bright Sunlight and Measures of Glucose and Lipid Metabolism in the OBB and NEO Study Populations
Model 1 | Model 2 | Model 3 | |
---|---|---|---|
Glucose metabolism | |||
Glucose, %, OBB | −0.05 (−0.15 to 0.05) | −0.08 (−0.21 to 0.04) | −0.03 (−0.17 to 0.11) |
Glucose, %, NEO study | −0.02 (−0.14 to 0.10) | −0.04 (−0.20 to 0.13) | −0.05 (−0.22 to 0.12) |
Insulin, %, OBB | −0.92 (−1.48 to −0.36) | −1.37 (−2.09 to −0.65) | −1.27 (−2.09 to −0.47) |
Insulin, %, NEO study | 0.21 (−0.47 to 0.88) | −0.40 (−1.37 to 0.55) | −0.36 (−1.38 to 0.65) |
HOMA-IR, %, OBB | −1.05 (−1.65 to −0.45) | −1.54 (−2.32 to −0.77) | −1.36 (−2.23 to −0.50) |
HOMA-IR, %, NEO study | 0.19 (−0.53 to 0.90) | −0.44 (−1.47 to 0.58) | −0.41 (−1.50 to 0.67) |
HOMA-B, %, OBB | −0.54 (−0.89 to −0.19) | −0.78 (−1.23 to −0.33) | −0.80 (−1.31 to −0.30) |
HOMA-B, %, NEO study | 0.27 (−0.40 to 0.93) | −0.39 (−1.30 to 0.52) | −0.31 (−1.26 to 0.64) |
Lipid metabolism | |||
LDL cholesterol, %, OBB | −0.36 (−0.66 to −0.07) | −0.19 (−0.57 to 0.19) | −0.16 (−0.58 to 0.27) |
LDL cholesterol, %, NEO study | −0.09 (−0.44 to 0.25) | −0.19 (−0.64 to 0.26) | −0.28 (−0.76 to 0.19) |
HDL cholesterol, %, OBB | −0.25 (−0.56 to 0.06) | 0.00 (−0.39 to 0.39) | −0.01 (−0.45 to 0.43) |
HDL cholesterol, %, NEO study | −0.10 (−0.41 to 0.21) | −0.03 (−0.43 to 0.37) | 0.02 (−0.39 to 0.44) |
Triglycerides, %, OBB | −0.57 (−1.12 to −0.03) | −1.36 (−2.07 to −0.66) | −1.28 (−2.07 to −0.50) |
Triglycerides, %, NEO study | −0.97 (−1.62 to −0.32) | −0.89 (−1.75 to −0.03) | −0.91 (−1.80 to −0.03) |
Presented associations are from the OBB (N = 4327) and the NEO study (N = 5899) populations. Results from the NEO study are weighted toward BMI distribution of the general population. Associations are presented as the percentage difference in the outcome measure per hour increase in bright sunlight during the 7 d before the center visit (95% CI). In model 1, analyses were adjusted for age, sex, and percentage body fat. In model 2, analyses were adjusted for age, sex, percentage body fat, and season. In model 3, analyses were adjusted for age, sex, percentage body fat, season, and outdoor temperature.
In the meta-analysis, results did not materially differ from the data presented for the OBB cohort (24). This was also true after we adjusted our data for BMI rather than percentage body fat (data not shown) or when we additionally adjusted the associations between bright sunlight duration and lipid levels for use of lipid-lowering medication. Specifically after meta-analyses, the results become somewhat weaker considering a 30-day period to calculate the mean hours of bright sunlight (24).
Discussion
In the current study, we examined the associations of outdoor temperature and bright sunlight with measures of glucose and lipid metabolism. Using two independent study populations with a combined sample size of >10,000 nondiabetic subjects of European descent, we detected little evidence for associations between mean outdoor temperature and glucose or lipid traits. In contrast, we found that bright sunlight was associated with a “healthier” metabolic profile conditional on season and mean outdoor temperature. These associations between bright sunlight duration and measures of glucose and lipid metabolism were generally stronger in the OBB than in the NEO study. This may have been due to differences in body composition and/or metabolic profile between the two cohorts. Nevertheless, to our knowledge, our study is the first to suggest that regular exposure to bright sunlight could improve glucose and lipid metabolism, thereby decreasing the lifetime risk of developing cardiometabolic disease.
Previous studies have suggested a link between lower outdoor temperature and reduced T2D risk, with thermogenesis in muscle and brown adipose tissue as potential mediators (7, 8). When taking into account the mean outdoor temperature in the 30 days prior to the center visit we found that, in nondiabetic subjects, a higher mean outdoor temperature was associated with lower HOMA-IR. We suspect that this might be primarily related to increased physical activity and/or altered habitual food intake when the outdoor temperature starts to increase, as evidenced by the association of higher outdoor temperature with lower BMI in both cohorts, although we were not able to test this hypothesis. At least from the perspective in which cold acclimatization promotes the thermogenic activity of brown adipose tissue (25–27), this was an unexpected result. We can think of several reasons for the lack of expected associations. First, our study populations were comprised primarily of middle-aged overweight subjects. Both age and BMI are known to be inversely associated with the amount and activity of brown adipose tissue as measured by the uptake of [18F]fluorodeoxyglucose (25). We note, however, that older and heavier subjects may benefit most from strategies that promote brown fat activation and are therefore an important population to study. Second, the impact of mean outdoor temperature on cardiometabolic outcomes could have been diluted due to controlled indoor climate and clothing habits. To further dissect the roles of outdoor temperature and bright sunlight on systemic metabolism while simultaneously accounting for physical activity, larger studies with more dense phenotyping are required.
To the best of our knowledge, this is the first human population study to show a positive association between hours of ambient bright sunlight and metabolic health. We acknowledge that we have no data available as to whether the participating individuals in our study were actually exposed to more sunlight during days with more sunshine. Nonetheless, such effects would dilute the exposure variables, and hence the true effect of bright sunlight on measures of glucose and lipid metabolism may actually be larger than observed herein. One candidate mechanism via which bright sunlight exposure may improve insulin sensitivity and lipid metabolism is through enhanced vitamin D generation (14). However, this is unlikely due to the lack of a causal association between vitamin D and T2D (28). A more conceivable biological mechanism, in addition to light-associated changes in diet and/or lifestyle, is the involvement of melatonin signaling, with nocturnal melatonin concentrations increasing with prolonged exposure to bright light (12). In turn, increased nocturnal melatonin levels may lead to reduced insulin secretion, as has been demonstrated in rodent studies (13). Interestingly, a role for melatonin in glucose metabolism has also been demonstrated by genetic association studies; common genetic variation in MTNR1B, which encodes one of the melatonin receptors, was associated with fasting glucose levels and T2D risk (29), and rare partial and total loss-of-function MTNR1B variants were associated with increased T2D risk (30). Furthermore, clinical trials have revealed that treatment with melatonin improves blood pressure and lipid metabolism in subjects with the metabolic syndrome (31). Based on these findings and our own data, we postulate that bright sunlight exposure might be of benefit to individuals with an unhealthy cardiometabolic risk profile through activation of melatonin signaling. Indeed, an ongoing clinical trial is currently assessing the role of light therapy in improving insulin sensitivity in patients with depression and T2D (32). Further results from such studies are necessary to determine the clinical relevance of our findings and to test our hypothesis that altered melatonin signaling mediates the beneficial effects of bright sunlight exposure on systemic metabolism.
The current study has a number of strengths and limitations. The study made use of data collected from two independent cohorts with a combined sample size of >10,000 participants. Another strength of our study was that both the OBB and NEO study participants were residing in relatively small geographic areas, and hence we were able to accurately couple data from local weather stations with anthropometric and biochemical data collected during the center visits. Limitations include that effect sizes were small, and although generally unidirectional in the two study populations, associations were generally stronger in the OBB than in the NEO study. This is likely due to differences in population characteristics as discussed above (see also Table 1). Owing to this inherent heterogeneity in cohort characteristics, the results of the meta-analyses should be interpreted with caution. Given that all outcome variables, especially those related to glucose metabolism, are highly interrelated we did not correct our data for multiple testing. Notwithstanding, associations between bright sunlight duration and metabolic measures within the OBB were robust and most would have survived multiple testing correction. Finally, we were unable to examine the potential contribution of lifestyle factors such as habitual food intake, sleep patterns, and physical activity to the observed associations. However, none of these factors can affect our exposures, that is, outdoor temperature and bright sunlight duration. Hence, instead of acting as confounders, these factors are more likely to act as mediators and could potentially explain the biological mechanisms behind our observations.
In summary, in this large study we have found evidence of an association between bright sunlight, but not outdoor temperature, and a healthier metabolic profile. Although all associations had a similar direction of effect in both of our study cohorts, positive findings were predominantly observed in the OBB population, and only to a lesser extent in the NEO cohort. Additionally, causality as well as the potential direct effects of sunlight on a healthier metabolic profile remain to be investigated.
Acknowledgments
We express our gratitude to all individuals who participate in the OBB and NEO study. We are also grateful to all participating general practitioners for inviting eligible participants and all research nurses for collecting the data.
Financial Support: F.K. is supported by British Heart Foundation Programme Grant RG/17/1/32663. C.C. is a British Heart Foundation Intermediate Clinical Research Fellow (FS/16/45/32359 11). D.v.H. is supported by the European Commission–funded project HUMAN (Health-2013-INNOVATION-1-602757). S.K. is supported by Dutch Heart Foundation Grant 2017T016 and the European Foundation for the Study of Diabetes (EFSD Albert Renold Travel Fellowship and EFSD Rising Star Fellowship program). We acknowledge the support of the Netherlands Cardiovascular Research Initiative, an initiative with the support of the Dutch Heart Foundation (CVON2014-02 ENERGISE). The OBB is supported by the National Institute for Health Research Oxford Biomedical Research Center Obesity and Lifestyle theme. The NEO study is supported by the participating departments, the division, and the board of directors of the Leiden University Medical Center, and by the Leiden University research profile area “Vascular and Regenerative Medicine.”
Disclosure Summary: The authors have nothing to disclose.
Glossary
Abbreviations:
- BMI
body mass index
- CV
coefficient of variation
- HDL
high-density lipoprotein
- HOMA-B
homeostatic model assessment for β-cell function
- HOMA-IR
homeostatic model assessment for insulin resistance
- LDL
low-density lipoprotein
- NEO
Netherlands Epidemiology of Obesity
- OBB
Oxford Biobank
- T2D
type 2 diabetes
References
- 1. Bartelt A, John C, Schaltenberg N, Berbée JF, Worthmann A, Cherradi ML, Schlein C, Piepenburg J, Boon MR, Rinninger F, Heine M, Toedter K, Niemeier A, Nilsson SK, Fischer M, Wijers SL, van Marken Lichtenbelt W, Scheja L, Rensen PC, Heeren J. Thermogenic adipocytes promote HDL turnover and reverse cholesterol transport. Nat Commun. 2017;8:15010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Berbée JF, Boon MR, Khedoe PP, Bartelt A, Schlein C, Worthmann A, Kooijman S, Hoeke G, Mol IM, John C, Jung C, Vazirpanah N, Brouwers LP, Gordts PL, Esko JD, Hiemstra PS, Havekes LM, Scheja L, Heeren J, Rensen PC. Brown fat activation reduces hypercholesterolaemia and protects from atherosclerosis development. Nat Commun. 2015;6(1):6356. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Bartelt A, Bruns OT, Reimer R, Hohenberg H, Ittrich H, Peldschus K, Kaul MG, Tromsdorf UI, Weller H, Waurisch C, Eychmüller A, Gordts PL, Rinninger F, Bruegelmann K, Freund B, Nielsen P, Merkel M, Heeren J. Brown adipose tissue activity controls triglyceride clearance. Nat Med. 2011;17(2):200–205. [DOI] [PubMed] [Google Scholar]
- 4. Bakker LE, Boon MR, van der Linden RA, Arias-Bouda LP, van Klinken JB, Smit F, Verberne HJ, Jukema JW, Tamsma JT, Havekes LM, van Marken Lichtenbelt WD, Jazet IM, Rensen PC. Brown adipose tissue volume in healthy lean south Asian adults compared with white Caucasians: a prospective, case-controlled observational study. Lancet Diabetes Endocrinol. 2014;2(3):210–217. [DOI] [PubMed] [Google Scholar]
- 5. Hanssen MJ, Hoeks J, Brans B, van der Lans AA, Schaart G, van den Driessche JJ, Jörgensen JA, Boekschoten MV, Hesselink MK, Havekes B, Kersten S, Mottaghy FM, van Marken Lichtenbelt WD, Schrauwen P. Short-term cold acclimation improves insulin sensitivity in patients with type 2 diabetes mellitus. Nat Med. 2015;21(8):863–865. [DOI] [PubMed] [Google Scholar]
- 6. van der Lans AA, Hoeks J, Brans B, Vijgen GH, Visser MG, Vosselman MJ, Hansen J, Jörgensen JA, Wu J, Mottaghy FM, Schrauwen P, van Marken Lichtenbelt WD. Cold acclimation recruits human brown fat and increases nonshivering thermogenesis. J Clin Invest. 2013;123(8):3395–3403. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Blauw LL, Aziz NA, Tannemaat MR, Blauw CA, de Craen AJ, Pijl H, Rensen PC. Diabetes incidence and glucose intolerance prevalence increase with higher outdoor temperature. BMJ Open Diabetes Res Care. 2017;5(1):e000317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Speakman JR, Heidari-Bakavoli S. Type 2 diabetes, but not obesity, prevalence is positively associated with ambient temperature. Sci Rep. 2016;6(1):30409. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Cepeda M, Koolhaas CM, van Rooij FJA, Tiemeier H, Guxens M, Franco OH, Schoufour JD. Seasonality of physical activity, sedentary behavior, and sleep in a middle-aged and elderly population: the Rotterdam study. Maturitas. 2018;110:41–50. [DOI] [PubMed] [Google Scholar]
- 10. Cepeda M, Muka T, Ikram MA, Franco OH, Schoufour JD. Seasonality of insulin resistance, glucose, and insulin among middle-aged and elderly population: the Rotterdam study. J Clin Endocrinol Metab. 2018;103(3):946–955. [DOI] [PubMed] [Google Scholar]
- 11. Berglund L, Berne C, Svärdsudd K, Garmo H, Melhus H, Zethelius B. Seasonal variations of insulin sensitivity from a euglycemic insulin clamp in elderly men. Ups J Med Sci. 2012;117(1):35–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Park SJ, Tokura H. Bright light exposure during the daytime affects circadian rhythms of urinary melatonin and salivary immunoglobulin A. Chronobiol Int. 1999;16(3):359–371. [DOI] [PubMed] [Google Scholar]
- 13. Peschke E, Bähr I, Mühlbauer E. Melatonin and pancreatic islets: interrelationships between melatonin, insulin and glucagon. Int J Mol Sci. 2013;14(4):6981–7015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Pilz S, Kienreich K, Rutters F, de Jongh R, van Ballegooijen AJ, Grübler M, Tomaschitz A, Dekker JM. Role of vitamin D in the development of insulin resistance and type 2 diabetes. Curr Diab Rep. 2013;13(2):261–270. [DOI] [PubMed] [Google Scholar]
- 15. Lee P, Bova R, Schofield L, Bryant W, Dieckmann W, Slattery A, Govendir MA, Emmett L, Greenfield JR. Brown adipose tissue exhibits a glucose-responsive thermogenic biorhythm in humans. Cell Metab. 2016;23(4):602–609. [DOI] [PubMed] [Google Scholar]
- 16. van den Berg R, Kooijman S, Noordam R, Ramkisoensing A, Abreu-Vieira G, Tambyrajah LL, Dijk W, Ruppert P, Mol IM, Kramar B, Caputo R, Puig LS, de Ruiter EM, Kroon J, Hoekstra M, van der Sluis RJ, Meijer OC, Willems van Dijk K, van Kerkhof LW, Christodoulides C, Karpe F, Gerhart-Hines Z, Kersten S, Meijer JH, Coomans CP, van Heemst D, Biermasz NR, Rensen PC. A diurnal rhythm in brown adipose tissue causes rapid clearance and combustion of plasma lipids at wakening. Cell Reports. 2018;22(13):3521–3533. [DOI] [PubMed] [Google Scholar]
- 17. Karpe F, Vasan SK, Humphreys SM, Miller J, Cheeseman J, Louise Dennis A, Neville MJ. Cohort profile: the Oxford Biobank. Int J Epidemiol. 2018;47(1):21–21g. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. de Mutsert R, den Heijer M, Rabelink TJ, Smit JW, Romijn JA, Jukema JW, de Roos A, Cobbaert CM, Kloppenburg M, le Cessie S, Middeldorp S, Rosendaal FR. The Netherlands Epidemiology of Obesity (NEO) study: study design and data collection. Eur J Epidemiol. 2013;28(6):513–523. [DOI] [PubMed] [Google Scholar]
- 19. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem. 1972;18(6):499–502. [PubMed] [Google Scholar]
- 20. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and β-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28(7):412–419. [DOI] [PubMed] [Google Scholar]
- 21. Wallace TM, Levy JC, Matthews DR. Use and abuse of HOMA modeling. Diabetes Care. 2004;27(6):1487–1495. [DOI] [PubMed] [Google Scholar]
- 22. Korn EL, Graubard BI. Epidemiologic studies utilizing surveys: accounting for the sampling design. Am J Public Health. 1991;81(9):1166–1173. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Lumley T. Analysis of complex survey samples. J Statistical Software. 2004;9(8):1–19. [Google Scholar]
- 24. Noordam R, Ramkisoensing A, Loh NY, Neville MJ, Rosendaal FR, Willems van Dijk K, van Heemst D, Karpe F, Christodoulides C, Kooijman S. Data from: Associations of outdoor temperature, bright sunlight, and cardiometabolic traits in two European population-based cohorts. figshare 2019. Accessed 26 January 2019. https://figshare.com/s/af69f38a950879d2709c.
- 25. van Marken Lichtenbelt WD, Vanhommerig JW, Smulders NM, Drossaerts JM, Kemerink GJ, Bouvy ND, Schrauwen P, Teule GJ. Cold-activated brown adipose tissue in healthy men. N Engl J Med. 2009;360(15):1500–1508. [DOI] [PubMed] [Google Scholar]
- 26. Chen KY, Brychta RJ, Linderman JD, Smith S, Courville A, Dieckmann W, Herscovitch P, Millo CM, Remaley A, Lee P, Celi FS. Brown fat activation mediates cold-induced thermogenesis in adult humans in response to a mild decrease in ambient temperature. J Clin Endocrinol Metab. 2013;98(7):E1218–E1223. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Au-Yong IT, Thorn N, Ganatra R, Perkins AC, Symonds ME. Brown adipose tissue and seasonal variation in humans. Diabetes. 2009;58(11):2583–2587. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Ye Z, Sharp SJ, Burgess S, Scott RA, Imamura F, Langenberg C, Wareham NJ, Forouhi NG; InterAct Consortium. Association between circulating 25-hydroxyvitamin D and incident type 2 diabetes: a mendelian randomisation study. Lancet Diabetes Endocrinol. 2015;3(1):35–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Bouatia-Naji N, Bonnefond A, Cavalcanti-Proença C, Sparsø T, Holmkvist J, Marchand M, Delplanque J, Lobbens S, Rocheleau G, Durand E, De Graeve F, Chèvre JC, Borch-Johnsen K, Hartikainen AL, Ruokonen A, Tichet J, Marre M, Weill J, Heude B, Tauber M, Lemaire K, Schuit F, Elliott P, Jørgensen T, Charpentier G, Hadjadj S, Cauchi S, Vaxillaire M, Sladek R, Visvikis-Siest S, Balkau B, Lévy-Marchal C, Pattou F, Meyre D, Blakemore AI, Jarvelin MR, Walley AJ, Hansen T, Dina C, Pedersen O, Froguel P. A variant near MTNR1B is associated with increased fasting plasma glucose levels and type 2 diabetes risk. Nat Genet. 2009;41(1):89–94. [DOI] [PubMed] [Google Scholar]
- 30. Bonnefond A, Clément N, Fawcett K, Yengo L, Vaillant E, Guillaume JL, Dechaume A, Payne F, Roussel R, Czernichow S, Hercberg S, Hadjadj S, Balkau B, Marre M, Lantieri O, Langenberg C, Bouatia-Naji N, Charpentier G, Vaxillaire M, Rocheleau G, Wareham NJ, Sladek R, McCarthy MI, Dina C, Barroso I, Jockers R, Froguel P; Meta-Analysis of Glucose and Insulin-Related Traits Consortium (MAGIC). Rare MTNR1B variants impairing melatonin receptor 1B function contribute to type 2 diabetes. Nat Genet. 2012;44(3):297–301. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Koziróg M, Poliwczak AR, Duchnowicz P, Koter-Michalak M, Sikora J, Broncel M. Melatonin treatment improves blood pressure, lipid profile, and parameters of oxidative stress in patients with metabolic syndrome. J Pineal Res. 2011;50(3):261–266. [DOI] [PubMed] [Google Scholar]
- 32. Brouwer A, van Raalte DH, Diamant M, Rutters F, van Someren EJ, Snoek FJ, Beekman AT, Bremmer MA. Light therapy for better mood and insulin sensitivity in patients with major depression and type 2 diabetes: a randomised, double-blind, parallel-arm trial. BMC Psychiatry. 2015;15(1):169. [DOI] [PMC free article] [PubMed] [Google Scholar]