Journal of Photochemistry and Photobiology B: Biology 143 (2015) 74–81

Contents lists available at ScienceDirect

Journal of Photochemistry and Photobiology B: Biology journal homepage: www.elsevier.com/locate/jphotobiol

Weekend personal ultraviolet radiation exposure in four cities in Australia: Influence of temperature, humidity and ambient ultraviolet radiation Fan Xiang a,⇑, Simone Harrison b,c, Madeleine Nowak b,d, Michael Kimlin c, Ingrid Van der Mei e, Rachel E. Neale f, Craig Sinclair g, Robyn M. Lucas a,h, the AusD Study Investigator Group a

National Centre for Epidemiology and Population Health, The Australian National University, Canberra, Australia School of Public Health, Tropical Medicine and Rehabilitation Sciences, Australian Institute of Tropical Health and Medicine, James Cook University, Townsville, Australia c School of Public Health, Institute of Health and Biomedical Innovation, Queensland University of Technology, Brisbane, Australia d School of Medicine and Dentistry, James Cook University, Townsville, Australia e Menzies Research Institute Tasmania, Hobart, Australia f Queensland Berghofer Institute of Medical Research, Brisbane, Australia g Cancer Council Victoria, Melbourne, Australia h Telethon Kids Institute, University of Western Australia, Perth, Australia b

a r t i c l e

i n f o

Article history: Received 13 June 2014 Received in revised form 22 October 2014 Accepted 27 December 2014 Available online 9 January 2015

a b s t r a c t Purpose: To examine the effects of meteorological factors on weekend sun exposure behaviours and personal received dose of ultraviolet radiation (UVR) in Australian adults. Methods: Australian adults (n = 1002) living in Townsville (19°S, 146°E), Brisbane (27°S, 153°E), Canberra (35°S, 149°E) and Hobart (43°S, 147°E) were recruited between 2009 and 2010. Data on sun exposure behaviours were collected by daily sun exposure dairies; personal UVR exposure was measured with a polysulphone dosimeter. Meteorological data were obtained from the Australian Bureau of Meteorology; ambient UVR levels were estimated using the Ozone Monitoring Instrument data. Results: Higher daily maximum temperatures were associated with reduced likelihood of wearing a longsleeved shirt or wearing long trousers in Canberra and Hobart, and higher clothing-adjusted UVR dose in Canberra. Higher daily humidity was associated with less time spent outdoors in Canberra. Higher ambient UVR level was related to a greater clothing-adjusted personal UVR dose in Hobart and a greater likelihood of using sunscreen in Townsville. Conclusion: The current findings enhance our understanding of the impact of weather conditions on the population’s sun exposure behaviours. This information will allow us to refine current predictive models for UVR-related diseases, and guide future health service and health promotion needs. Ó 2015 Elsevier B.V. All rights reserved.

1. Introduction Exposure to solar ultraviolet radiation (UVR) has both adverse and beneficial effects on human health [17]. Exposure to UVR is the most important risk factor for melanoma and non-melanoma skin cancers (NMSC) [2], and eye disorders such as cataract and pterygium [30], and UVR-induced immunosuppression can cause reactivation of latent virus infections [20]. The best-known beneficial effect of UVR exposure is initiation of vitamin D synthesis in

⇑ Corresponding author at: National Centre for Epidemiology and Population Health, The Australian National University, ACT 2601, Australia. Tel.: +61 2 6125 2312; fax: +61 2 6125 5614. E-mail address: [email protected] (F. Xiang). http://dx.doi.org/10.1016/j.jphotobiol.2014.12.029 1011-1344/Ó 2015 Elsevier B.V. All rights reserved.

the skin. Adequate vitamin D is essential for bone health and may decrease the risk or improve outcomes of a wide range of diseases [12]. The balance of risks and benefits of UVR exposure and its contribution to the total disease burden at a population-level depends on demographic factors such as age and sex (e.g. NMSCs are more common in the elderly with higher incidence in men than women), as well as on the population distributions of different skin types and of levels of UVR exposure. The latter is a function of ambient UVR and sun exposure behaviour. In populations of similar ethnic origin, the incidence of NMSC increases as ambient UVR levels increase [29]. For any individual, their own risk of UVR-induced disease also depends on their sun exposure behaviour. Globally, outdoor workers receive about 10% of available ambient UVR and

F. Xiang et al. / Journal of Photochemistry and Photobiology B: Biology 143 (2015) 74–81

indoor workers and children receive about 3% [11]. Within a population, this dose can vary widely from one tenth to ten times the mean value [9]. Changes in weather conditions, such as temperature and humidity, alter people’s sun exposure behaviour [10] but the level and direction of the effect may vary across population subgroups and locations. For example, on hotter days, people living in already hot climates may prefer to spend time indoors, whereas those in cool climates may spend more time outdoors. Whether sun exposure is more affected by changes in relative temperature – hotter or colder – or absolute temperature, is not clear. In a previous study of self-reported sun exposure behaviour on the preceding weekend, Australian adults were more likely to wear sunscreen, sunglasses and a hat on warmer than on cooler days. Of the participants who spent more than 15 min outside, the reported time outdoors was greater when it was warmer, and there were more sunburn episodes [10]. However, on hotter days (>28 °C), some participants did not spend more than 15 min outside, that is, during peak UVR times they stayed indoors. Understanding how people’s sun exposure behaviour may change in relation to meteorological factors is crucial in the context of managing current behaviour and in predicting UVR-related health risks under future climate change scenarios. Here, we examine the effects of meteorological factors (temperature, humidity, and ambient UVR) on weekend sun exposure behaviours (amount of time spent outdoors and use of sun protection such as clothing and sunscreen) and personal received dose of UVR (adjusted for clothing) in Australian adults living in four cities with different thermal and UVR climates. 2. Methods This analysis uses data from the multi-centre AusD Study [4], the primary aim of which was to examine determinants of vitamin D status in adults (aged 18 + years) resident in four eastern Australian cities, each separated by approximately 8° of latitude: Townsville (19°S, 146°E), Brisbane (27°S, 153°E), Canberra (35°S, 149°E) and Hobart (43°S, 147°E). The latitude range of the study regions provided wide variation in ambient UVR levels and other meteorological factors, and standardised data collection across the four regions ensured comparable measurement of sun exposure behaviours. The methods of the AusD Study have been described elsewhere [4]. In brief, 1002 adults aged 18–75 years were recruited from the Australian Electoral Roll between May 2009 and December 2010, approximately equally distributed by sex, age group and study region. Data were collected by self-administered questionnaire, face-to-face interviews (including physical examination and collection of blood) and daily sun-exposure diaries. Polysulphone UVR dosimeters were used to monitor personal UVR exposure [22]. The AusD Study was approved by the Human Research Ethics Committees of the participating institutions (Australian National University #2008/451; Queensland University of Technology #0600000224; James Cook University #H3124; University of Tasmania #H0010277). All participants signed written informed consent. 2.1. Demographic information Questionnaires were self-completed at baseline and captured information about date of birth, education level (using the question ‘‘What is the highest technical, professional or academic qualification that you have completed?’’ with 8 options), type of employment (using the question ‘‘Which of the following best describes your current employment status?’’ with 9 options), and parents’ ethnicity

75

(using the question ‘‘What is your mother’s/father’s ethnic origin (that is, the place where the most of their ancestors came from)?’’ with 20 possible options) [4]. 2.2. Assessment of personal sun exposure and duration of time spent outdoors Participants recorded the amount of time spent outdoors during each hour from 5 am to 7 pm on 10 consecutive days, using the following categories: 0 min, 0 mm) were excluded. Daily time spent outdoors was calculated as the sum of time spent outdoors from 6 am to 6 pm Australian Eastern Standard Time (AEST), using the midpoint value of each of the five time categories (i.e., 0, 7.5, 22.5, 37.5, 52.6 min) [15]. We used the clothing on each of the body areas – upper body, lower body and head – as a separate outcome variable. We focused on clothing worn during the middle of the day (11 am to 2 pm AEST) as this is the time of peak UVR, most likely to affect both vitamin D synthesis and UV-induced skin damage. Upper body clothing was coded into three categories: no clothing on the upper body, bikini, swimsuit, crop top, or singlet; short-sleeved top; longsleeved top. Lower body clothing was also coded into three categories: no lower body clothing, briefs, shorts or short skirt; kneelength shorts or skirt; long trousers/jeans, long skirt. Head covering was coded into two categories: no head cover or a hat without a brim; any sunhat including cap, legionnaire’s cap, bucket hat or wide-brimmed hat. For regression analysis, upper body clothing

76

F. Xiang et al. / Journal of Photochemistry and Photobiology B: Biology 143 (2015) 74–81

was considered as long-sleeved top, yes/no; lower body clothing as long trousers/skirt, yes/no. Where clothing changed during the 11 am to 2 pm period, the most frequent category of clothing (clothing worn most frequently) was assigned. The polysulphone dosimeters provided the dose of UVR received at the wrist from the time of waking in the morning until the time of going to bed at night on each day. The percentage of the body surface area exposed was calculated as 100% minus the percentage of skin covered by clothing when outdoors (as previously described [15]) and multiplied by the UVR dose from the dosimeter to estimate the clothing-adjusted personal UVR dose. For descriptive analyses, the median and interquartile range (IQR) or mean and standard deviation (SD) were presented depending on the distribution of the variable. We used scatter plots with fitted lines plotted using a locally weighted regression smoothing function (Stata function ‘‘lowess’’) to show variation in time outdoors and clothing adjusted UVR exposure in relation to climatic factors. For regression analysis with continuous outcomes (such as time spent outdoors and adjusted personal UVR dose) we used random effects multilevel linear regression modelling [7]. We modelled each outcome variable (amount of time spent outdoors, clothing [separately upper body, lower body, head] and sunscreen use, and personal received dose of UVR) in relation to each climatic factor (temperature, humidity, ambient UVR), adjusting for those demographic and other factors that fulfilled the criteria for being a potential confounder factor (i.e., an independent risk factor for the outcome, associated with the exposure, and not an intermediate between the exposure and outcome [27]). Robust standard errors were obtained to take account of minor levels of heteroscedasticity and lack of normality of residuals. Random effects multilevel logistic regression was used where the outcome was dichotomous (using sunscreen, and wearing a long-sleeved top, long leg cover, or sunhat). In order to account for the uneven distribution of data collection by season across the four study regions, all multivariable regression analyses were stratified by study region. This was preferable to adjustment for season/month of interview in combined analyses because of the high correlation between season/month and temperature/humidity/ambient UVR. Statistical analysis was performed using Stata 12.0 (Stata Statistical Software, Release 12.0, StataCorp, College Station, TX). P < 0.05 was considered to be statistically significant. 3. Results 3.1. General characteristics The general characteristics of the study participants have been described elsewhere [4]. In summary, 80% of all participants were born in Australia and 54% were female. Approximately 50% reported Australian parental ethnicity, and 40% reported European parental ethnicity, while less than 10% reported non-Australian and non-European parental ethnicity. Almost half of the participants were working full-time (48%) and over two-thirds of the working participants worked mainly indoors (69%). Compared with the other three study regions, Canberra participants were more likely to have been born overseas, to hold a tertiary qualification and to work predominantly indoors (all P < 0.001). In addition, participants from Canberra were more likely to have been interviewed during winter (Canberra: 39%, Townsville: 25%, Brisbane 35%, Hobart 28%, P = 0.001) and less likely to have been interviewed during summer (Canberra: 6%, Townsville: 11%, Brisbane 13%, Hobart 12%, P = 0.001). Overall, participants spent a median time outdoors of 105 (interquartile range = 161) minutes outdoors and 26% people wore sunscreen on a weekend day. Prevailing weather conditions during the study at each location are included in Supplementary Table 1.

3.2. Descriptive analysis Completed weekend sun diary data were available for 953 (95%) AusD Study participants. Overall, men spent more time outdoors per day (127 [203] min vs. 105 [143] min, P < 0.001), received more ambient UVR (155 [292] J/m2 vs. 111 [175], P < 0.001), had higher clothing-adjusted UVR dose (36 [83] J/m2 vs. 23 [48] J/m2) and were less likely to use sunscreen (14% vs. 37%, P < 0.001), than women. During the middle of the day (11 am to 2 pm AEST) women were more likely to wear a long-sleeved top (44% vs. 30%, P < 0.001) and long trousers/skirt (57% vs. 47%, P < 0.001) and were less likely to wear a sunhat (10% vs. 23%, P < 0.001) than men. The relationship between time spent outdoors and the meteorological factors is shown in Fig. 1. Hobart and Canberra had broader ranges of daily maximum temperature than Townsville and Brisbane. For Hobart, the highest latitude location, time spent outdoors on weekends increased with higher daily maximum temperature; this trend was not clearly apparent for the other three regions (Fig. 1A). However, when the analysis was restricted to days where the temperature was over 30 °C participants living in Brisbane spent slightly less time outside on days with higher temperatures. With increasing daily relative humidity, participants living in Brisbane spent more time outside while those living in Canberra spent less time outside; there was no marked difference in time spent outdoors in relation to relative humidity in Townsville and Hobart (Fig. 1B). The relationship between time spent outdoors and ambient UVR (Fig. 1C) was similar to that for temperature (Fig. 1A), most likely due to the high correlation between daily maximum temperature and daily average ambient UVR (Pearson’s correlation coefficient: 0.75). Overall, participants received 4% (median: 4%; mean: 8%) of the ambient UVR. After adjustment for clothing cover this decreased to 1% (median: 1%; mean: 2%). The relationship between clothingadjusted personal UVR and higher daily maximum temperature and relative humidity was similar to that for time spent outdoors (Fig. 2A and B). Clothing-adjusted personal UVR dose did not vary significantly across the range of daily average ambient UVR (Fig. 2C). The use of sunscreen and style of clothing worn in relation to meteorological factors differed across the four study regions (Supplementary Figs. 1–4). At the highest quartile of daily maximum temperature (>28.2 °C), more than half participants living in Hobart reported using sunscreen, while one third of participants in Brisbane and around 20% participants in Townsville and Canberra used sunscreen. As daily maximum temperature or daily ambient UVR increased, the percentage of participants wearing a long-sleeved top or long trousers/skirt decreased in all four regions; however as daily relative humidity increased, this percentage increased only in Canberra. The percentage of participants wearing a sunhat was low across all four regions (11–15%) and did not appear to be associated with differences in weather conditions. 3.3. Regression analysis We explored the association between sun exposure behaviours and meteorological factors using multivariable regression analysis, stratifying by study region (Table 1). There were marked differences in the patterns of association between the various sun exposure behaviours and meteorological factors, across the four regions. With increasing daily maximum temperature, participants living in Canberra and Hobart were less likely to wear a longsleeved top or long trousers; Canberra participants had higher clothing-adjusted UVR dose (regression coefficient: 3.52, 95% CI: 0.24, 6.71). Daily relative humidity was inversely associated with time spent outdoors only in Canberra, where humidity is typically

F. Xiang et al. / Journal of Photochemistry and Photobiology B: Biology 143 (2015) 74–81

B

C

Time spent outdoors (minutes per day)

A

77

Daily maximum temperature (°C)

Daily relative humidity (%)

Daily average ambient UVR (kJ/m2)

Red lines are the lowess smoothing curves produced with Stata function “lowess”. Fig. 1. Time spent outdoors according to temperature/humidity/ambient UVR in adults in four Australian cities. Red lines are the lowess smoothing curves produced with Stata function ‘‘lowess’’. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

relatively low. Higher levels of ambient UVR were associated with a greater clothing-adjusted personal UVR dose only in Hobart. The use of sunhats was not significantly associated with any of the meteorological factors in any study region. We repeated the analyses using daily maximum apparent temperature (perceived outdoor temperature, derived from the combination of temperature, wind speed and relative humidity) [21], instead of daily maximum temperature. The results were essentially the same (data not shown). Due to the effect of regional acclimatisation, people’s sun exposure behaviour may be influenced by the variation of weather conditions, rather than the absolute weather. We further examined the influence of weather deviation on each of the sun exposure behaviours (Table 2). The results were similar to those from the primary analysis (Table 1), with no evidence to support an acclimatisation effect. 4. Discussion In this multicentre study, we found that daily maximum temperature (either absolute or apparent), relative humidity and ambient UVR had significant, although modest, impacts on weekend sun exposure behaviours and received UVR doses of adults living in four eastern Australian cities. The patterns of the associations between sun exposure behaviours and meteorological factors differed across the four study regions. Daily maximum temperature and ambient UVR had more influence on sun exposure behaviours than did relative humidity. Higher daily temperature was associated with a lower likelihood of wearing a long-sleeved top or wear-

ing long trousers in Canberra and Hobart; higher ambient UVR with greater clothing-adjusted personal UVR in Hobart. In more humid conditions, Canberra participants tended to spend more time indoors. There was a similar effect of absolute and relative weather conditions on people’s sun exposure behaviours. The key strength of this study was the population-based recruitment over a range of latitudes and over more than 12 months, which allowed for assessment of effects of different meteorological environments. Good quality meteorological data were derived from official meteorological databases. While participants’ sun exposure behaviours were self-reported, the use of a daily sun diary is likely to have minimised recall bias [26] and there was objective measurement of the received dose of UVR. The study does, however, also have some limitations. This was a cross-sectional study, with each participant providing data on sun exposure behaviours on up to two consecutive weekend days. Longitudinal data would better elucidate changes in an individual’s behaviour according to these meteorological factors. AusD Study participants were more likely to be older, Australia-born women, and there was a bias towards indoor workers compared to the population-based Australian National Health Survey (2007–2008) [4]. The distribution of interviews across the seasons was not consistent across study regions, with Canberra participants more likely to be interviewed during winter and less likely to be interviewed in summer than participants from other regions. Rather than adjusting for study region – which is a strong determinant of differences in meteorological factors – we accounted for this by stratifying all of the regression analyses by study region. In the AusD Study, participants wore a polysulphone dosimeter at the wrist.

78

F. Xiang et al. / Journal of Photochemistry and Photobiology B: Biology 143 (2015) 74–81

A

C

Clothing-adjusted personal UVR (J/m2)

B

Daily maximum temperature (°C)

Daily average ambient UVR (kJ/m2)

Daily relative humidity (%)

Red lines are the lowess smoothing curves produced with Stata function “lowess”. Fig. 2. Clothing-adjusted personal UVR exposure according to temperature/humidity/ambient UVR in adults in four Australian cities. Red lines are the lowess smoothing curves produced with Stata function ‘‘lowess’’. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 1 Association between adults’ sun exposure behaviours and meteorological factors in four Australian cities. Townsville (19.3°S) Adjusted coefficienta (95% CI)

Brisbane (27.5°S) Adjusted coefficienta (95% CI)

Canberra (35.3°S) Adjusted coefficienta (95% CI)

Hobart (42.8°S) Adjusted coefficienta (95% CI)

Time spent outdoors (minutes per day) Maximum temperature (°C) 1.31 ( 8.76, 6.13) Relative humidity (1%) 0.45 ( 1.57, 0.67) Ambient UVR (kJ/m2) 0.45 ( 11.53, 10.63)

3.18 ( 2.74, 9.11) 0.89 ( 0.73, 2.51) 6.47 ( 18.87, 5.93)

0.44 ( 2.41, 3.30) 1.33* ( 2.54, 0.12) 9.48 ( 22.43, 3.46)

2.28 ( 3.50, 8.06) 0.87 ( 1.54, 3.28) 11.39 ( 0.25, 23.04)

Adjusted personal UVR (J/m2) Maximum temperature (°C) Relative humidity (1%) Ambient UVR (kJ/m2)

2.77 ( 10.27, 4.72) 0.68 ( 1.99, 0.63) 5.65 ( 3.5 5, 14.85)

0.81 ( 3.69, 5.31) 0.18 ( 1.35, 0.99) 1.29 ( 8.61, 11.19)

3.52* (0.34, 6.71) 1.53 ( 3.07, 0.01) 2.75 ( 15.34, 9.83)

2.15 ( 2.36, 6.65) 0.12 ( 1.07, 1.31) 13.59* (1.54, 25.62)

Adjusted odds ratioa (95% CI)

Adjusted odds ratioa (95% CI)

Adjusted odds ratioa (95% CI)

Adjusted odds ratioa (95% CI)

0.54 (0.27, 1.07) 1.03 (0.94, 1.13) 3.13* (1.09, 9.01)

1.01 (0.78, 1.31) 0.98 (0.92, 1.04) 1.16 (0.66, 2.04)

0.90 (0.77, 1.15) 0.95 (0.89, 1.01) 1.41 (0.73, 2.72)

0.98 (0.79, 1.23) 1.03 (0.94, 1.12) 1.86 (1.00, 3.36)

Wore long-sleeved top (yes/no)b Maximum temperature (°C) 0.90 (0.55, 1.45) Relative humidity (1%) 0.94 (0.86, 1.01) Ambient UVR (kJ/m2) 0.63 (0.28, 1.39)

0.79 (0.62, 1.01) 0.99 (0.94, 1.04) 0.58 (0.32, 1.04)

0.44* (0.22, 0.86) 0.89 (0.77, 1.02) 0.47 (0.12, 1.78)

0.73* (0.60, 0.88) 0.97 (0.91, 1.03) 0.73 (0.52, 1.02)

Wore long trousers/skirt (yes/no)b Maximum temperature (°C) 1.40 (0.90, 2.16) Relative humidity (1%) 1.01 (0.96, 1.07) 2 Ambient UVR (kJ/m ) 0.59 (0.33, 1.00)

0.78 (0.57, 1.06) 1.04 (0.97, 1.11) 0.32* (0.13, 0.78)

0.67* (0.49, 0.92) 1.01 (0.92, 1.09) 0.84 (0.39, 1.79)

0.81* (0.69, 0.95) 0.95 (0.89, 1.00) 0.69* (0.51, 0.95)

Wore sunhat (yes/no)b Maximum temperature (°C) Relative humidity (1%) Ambient UVR (kJ/m2)

1.11 (0.88, 1.38) 1.01 (0.96, 1.06) 0.89 (0.56, 1.42)

1.08 (0.87, 1.32) 0.99 (0.91, 1.07) 1.13 (0.48, 2.68)

1.04 (0.92, 1.16) 1.01 (0.96, 1.06) 1.26 (0.98, 1.62)

Used sunscreen (yes/no)b Maximum temperature (°C) Relative humidity (1%) Ambient UVR (kJ/m2)

*

0.92 (0.67, 1.27) 0.98 (0.93, 1.03) 1.59 (0.07, 2.59)

P < 0.05. Adjusted coefficients and odds ratios were from multivariable regression models including daily maximum temperature, relative humidity, ambient UVR, sex and age group as covariates. b Analysis was restricted to participants who spent more than 15 min outdoors from 11 am to 2 pm Australian Eastern Standard Time (AEST). a

Table 2 Association between adults’ sun exposure behaviours and deviation of meteorological factors from the averagec, in four Australian cities. Townsville (19.3°S) Adjusted coefficienta (95% CI)

b c

Hobart (42.8°S) Adjusted coefficienta (95% CI)

7.15 ( 16.20, 1.90) 0.91 ( 2.07, 0.24) 1.80 ( 15.43, 11.82)

2.22 ( 3.99, 8.44) 0.70 ( 0.96, 2.35) 0.76 ( 18.07, 16.54)

0.10 ( 2.99, 3.19) 1.25* ( 2.47, 0.03) 9.46 ( 23.54, 4.62)

4.65 ( 1.46, 10.76) 0.05 ( 2.11, 2.20) 24.00 (8.97, 39.02)

Clothing-adjusted UVR dose (J/m2) Deviation of maximum temperature (°C) Deviation of relative humidity (1%) Deviation of ambient UVR (kJ/m2)

1.47 ( 6.99, 9.93) 0.02 ( 1.36, 1.33) 3.23 ( 8.01, 14.47)

1.10 ( 3.89, 6.09) 0.03 ( 1.38, 1.44) 3.07 ( 9.71, 15.86)

6.09* (1.89, 10.30) 1.16 ( 2.64, 0.33) 3.15 ( 15.92, 9.62)

3.30 ( 1.35, 7.96) 1.03 ( 2.28, 0.23) 21.74* (3.53, 39.95)

Adjusted odds ratioa (95% CI)

Adjusted odds ratioa (95% CI)

Adjusted odds ratioa (95% CI)

Used sunscreen (Yes/No) Deviation of maximum temperature (°C) Deviation of relative humidity (1%) Deviation of ambient UVR (kJ/m2)

0.46 (0.20, 1.06) 1.03 (0.94, 1.14) 5.05* (1.05, 24.22)

1.06 (0.80, 1.40) 0.99 (0.93, 1.05) 0.89 (0.45, 1.74)

0.95 (0.79, 1.14) 0.96 (0.90, 1.02) 0.84 (0.38, 1.85)

1.05 (0.85, 1.32) 0.99 (0.90, 1.08) 4.60 (0.61, 13.15)

Wore long-sleeved top (Yes/No)b Deviation of maximum temperature (°C) Deviation of relative humidity (1%) Deviation of ambient UVR (kJ/m2)

1.37 (0.76, 2.48) 0.97 (0.89, 1.06) 0.68 (0.25, 1.83)

0.69* (0.52, 0.91) 1.03 (0.98, 1.08) 1.09 (0.61, 1.97)

0.35* (0.22, 0.55) 0.85* (0.74, 0.98) 2.03 (0.43, 9.67)

0.80* (0.67, 0.97) 1.02 (0.97, 1.08) 0.74 (0.48, 1.16)

Wore long trousers/skirt (Yes/No)b Deviation of maximum temperature (°C) Deviation of relative humidity (1%) Deviation of ambient UVR (kJ/m2)

1.47 (0.94, 2.30) 1.02 (0.96, 1.07) 0.58 (0.29, 1.14)

0.64* (0.43, 0.93) 1.11* (1.02, 1.21) 1.35 (0.58, 3.16)

0.63* (0.45, 0.88) 0.97 (0.89, 1.06) 1.06 (0.41, 2.73)

0.84 (0.71, 1.01) 0.98 (0.92, 1.04) 0.86 (0.54, 1.35)

Wore sunhat (Yes/No)b Deviation of maximum temperature (°C) Deviation of relative humidity (1%) Deviation of ambient UVR (kJ/m2)

0.95 (0.66, 1.36) 0.99 (0.94, 1.04) 1.95 (0.98, 3.87)

1.18 (0.93, 1.49) 1.02 (0.98, 1.07) 0.69 (0.40, 1.19)

1.10 (0.87, 1.37) 1.00 (0.92, 1.09) 0.56 (0.19, 1.63)

1.07 (0.94, 1.22) 0.98 (0.93, 1.03) 1.54* (1.03, 2.30)

P < 0.05. Adjusted coefficients and odds ratios were from multivariable regression models including daily deviation of maximum temperature, deviation of relative humidity, deviation of ambient UVR, sex and age group as covariates. Analysis was restricted to participants who spent more than 15 min outdoors from 11 am to 2 pm Australian Eastern Standard Time (AEST). E.g., Deviation of maximum temperature: absolute maximum temperature on the day of data collection minus average daily maximum temperature of the season for 2009 and 2010 for the study region.

F. Xiang et al. / Journal of Photochemistry and Photobiology B: Biology 143 (2015) 74–81

a

Canberra (35.3°S) Adjusted coefficienta (95% CI)

Time spent outdoors (minutes per day) Deviation of maximum temperature (°C) Deviation of relative humidity (1%) Deviation of ambient UVR (kJ/m2)

Adjusted odds ratioa (95% CI)

*

Brisbane (27.5°S) Adjusted coefficienta (95% CI)

79

80

F. Xiang et al. / Journal of Photochemistry and Photobiology B: Biology 143 (2015) 74–81

Wearing the dosimeter at this site may increase the risk of Type-II noncompliance (e.g., covering badges with clothing, exposing them to water or removing them during outdoor activities) [22]. The orientation of the wrist is also different from that of the rest of the body, although previous research indicates that it receives comparable UVR exposure to other parts of the body [23]. Sun exposure behaviour is a key modifier of an individual’s received UVR dose and is thus an important determinant of the risk of UVR-related health outcomes [8]. At the population level, over 80% of the variability in the incidence of the NMSCs is accounted for by age, sex, calendar year and average ambient UVR [29]. However, within a population, sun exposure behaviour varies significantly by age, sex, geographic location, and weather conditions [10,8], thus modifying the risks of UVR-related diseases. There have been very few studies that have examined the effects of temperature and humidity on sun exposure behaviour. In the 2003–4 national sun survey of Australian adolescents and adults, information about time spent outdoors and other sun exposure behaviours during peak UVR hours (10 am to 2 pm AEST) on the previous summer weekend was collected by telephone interview [10]. Temperature (at 3 pm) was the ‘‘strongest and most consistent determinant of adults’ sun-protective behaviours’’. When the temperature was greater than 22 °C (temperature < 22 °C as the reference category), adults spent more time outdoors, and were more likely to wear hats and sunscreen, but were much less likely to wear clothing that covered the arms or legs. Other meteorological factors including cloud cover and wind speed were not associated with people’s sun exposure behaviour. Our study extends this earlier work, by including data across a full year, with more detailed sun exposure behaviour provided by a daily diary, and objective measurement of UVR exposure. Furthermore, the current analysis has a greater range of temperatures, from 9 °C to 37 °C, and was able to model temperature as a continuous variable. In previous analyses there was no adjustment or stratification according to the participant’s location, so that any effect of temperature on behaviour may have been confounded by the effect of ‘‘region’’, which could include acclimatisation effects, strength of sun protection programs, clothing norms and other factors. In our stratified analysis, there was no uniform behavioural change in relation to meteorological factors across the four study regions. AusD participants living in Canberra received a higher clothing-adjusted UVR dose when the daily maximum temperature was higher. This was mainly due to more time in the sun or a greater body surface area exposed. As a result, higher temperatures may increase the risk of adverse effects of sun exposure (e.g., skin cancer). The greater amount of time spent outdoors in association with higher daily maximum temperature in Hobart participants was similar to the pattern previously described by Dobbinson and colleagues (Fig. 2A), but was not statistically significant after adjustment for age and sex. Daily maximum temperature did not appear to influence the use of sunhats or sunscreen in any region. We found that higher daily relative humidity was associated with less time spent outdoors on weekends only in Canberra. Canberra is the only inland city at an altitude of 580 m, whereas the other study regions are all coastal cities at the sea-level. Therefore, Canberra is usually relatively dry (annual relative humidity: Townsville: 58%; Brisbane: 52%; Canberra: 37%; Hobart: 58%). This may explain some of the distinctive patterns of changes in behaviours in response to temperature and humidity in Canberra; for example, when people are not acclimatised to high humidity and avoid it by being indoors. Time spent outdoors is commonly used as a surrogate for personal sun exposure [28]. However we found only a moderate correlation between time spent outdoors on weekends and UVR

exposure as measured by the polysulphone badge worn at the wrist (Spearman’s correlation coefficient: 0.61). Furthermore, the correlation between time spent outdoors and clothing-adjusted UVR dose was even lower (Spearman’s correlation coefficient: 0.51). In Townsville and Brisbane, clothing-adjusted personal UVR dose was not associated with any of the meteorological factors, but increased in line with daily maximum temperature in Canberra and higher ambient UVR in Hobart. Townsville and Brisbane are lower latitude regions where it is relatively warm all year round, possibly accounting for the lack of variation in sun exposure behaviours across the year and in association with meteorological factors. In Townsville, participants tended to already be wearing a minimal amount of clothing, with little opportunity to wear less during warmer temperatures. In regions of higher latitude (Canberra and Hobart), there is much greater seasonal variation in the local climate (especially temperature), that is mirrored by larger changes in sun exposure behaviours such as time spent outdoors and the types of clothing worn. Overall, only one quarter of AusD Study participants applied sunscreen on weekends when they were outdoors (Spring: 28%, Summer: 26%, Autumn: 21%, Winter: 25%), with no significant variation across the study regions. This low percentage of sunscreen use is consistent with the findings from national surveys in the United States [5]. Participants in Townsville were more likely to report using sunscreen when the ambient UVR was higher (AOR = 3.13 [95% CI, 0.34, 6.71] per increase of 1 kJ/m2), but this relationship was not apparent in the other locations. Unlike temperature or humidity, ambient UVR cannot be perceived directly, but the UV Index (UVI) is published in Australian newspapers, on the internet and with the television weather forecast, and provides a visual method of communicating the ambient UVR for the day. A recent systematic review reported that the UVI generally had a very limited influence on people’s sunscreen use [14]. Our results, showing greater sunscreen use with higher ambient UVR levels, show that Australians living in a high ambient UVR environment (Townsville) were more likely to use sun protection according to the ambient UVR than people in lower ambient UVR environments. Notably, though, we did not ask people if they checked the daily UVI to guide their sun protection behaviour. Modelling studies indicate that temperature, humidity and ambient UVR will increase in tropical regions in the future [13]. Non-melanoma skin cancers are the most common cancers in many countries, with high incidence particularly in fairskinned populations living in high ambient UVR locations [1]. Understanding the effect of climatic factors on a population’s sun exposure behaviours allows modelling of future risks that takes account not only of changes in ambient UVR [25,24], but also changes in behaviour that will affect the received personal UVR dose and the amount of skin that is sun-exposed. These factors are critical inputs to risk calculations and will be important to guide future health service and health promotion needs.

Funding/support This work was supported by the National Health and Medical Research Council (NHMRC Project Grant 497220). F.X. is supported by a NHMRC Centre of Research Excellence in Sun and Health Postdoctoral Fellowship. S.H. and M.N. received salary support from Queensland Health. M.K. is supported by a Cancer Council Queensland Senior Research Fellowship. I.v.d.M. is supported by Australian Research Council Future Fellowships. R.E.N. is supported by a NHMRC Research Fellowship. R.L. is supported by a NHMRC Career Development Fellowship.

F. Xiang et al. / Journal of Photochemistry and Photobiology B: Biology 143 (2015) 74–81

Conflict of interest None declared. Acknowledgement We are indebted to Mr Ivan Hanigan for his assistance with derivation of the ambient UVR according to study region. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jphotobiol.2014. 12.029. References [1] M. Almahroos, A.K. Kurban, Ultraviolet carcinogenesis in nonmelanoma skin cancer. Part I: incidence rates in relation to geographic locations and in migrant populations, SKINmed 3 (2004) 29–36. [2] B.K. Armstrong, A. Kricker, D.R. English, Sun exposure and skin cancer, Australas. J. Dermatol. 38 (Suppl 1) (1997) S1–S6. [3] BOM website, Australian Bureau of Meteorology. (accessed 01.08.13). [4] A.M. Brodie, R.M. Lucas, S.L. Harrison, I.A.F. van der Mei, B. Armstrong, A. Kricker, R.S. Mason, A.J. McMichael, M. Nowak, D.C. Whiteman, M.G. Kimlin, The AusD study: a population-based study of the determinants of serum 25hydroxyvitamin D concentration across a broad latitude range, Am. J. Epidemiol. 177 (2013) 894–903. [5] D.B. Buller, V. Cokkinides, H.I. Hall, A.M. Hartman, M. Saraiya, E. Miller, L. Paddock, K. Glanz, Prevalence of sunburn, sun protection, and indoor tanning behaviours among Americans: review from national surveys and case studies of 3 states, J. Am. Acad. Dermatol. 65 (2011) S114–S123. [6] CIE, Erythemal Reference Action Spectrum and Standard Erythema Dose CIE 007/E:1998, 1998. [7] A.M. De Livera, S. Zaloumis, J.A. Simpson, Models for the analysis of repeated continuous outcome measures in clinical trials, Respirology 19 (2014) 155– 161. [8] B. Diffey, Climate change, ozone depletion and the impact on ultraviolet exposure of human skin, Phys. Med. Biol. 49 (2004) R1–R11. [9] B.L. Diffey, H.P. Gies, The confounding influence of sun exposure in melanoma, Lancet 351 (1998) 1101–1102. [10] S. Dobbinson, M. Wakefield, D. Hill, A. Girgis, J.F. Aitken, K. Beckmann, A.I. Reeder, N. Herd, A. Fairthorne, K.A. Bowles, Prevalence and determinants of Australian adolescents’ and adults’ weekend sun protection and sunburn, summer 2003–2004, J. Am. Acad. Dermatol. 59 (2008) 602–614. [11] D.E. Godar, UV doses worldwide, Photochem. Photobiol. 81 (2005) 736–749.

81

[12] M.F. Holick, Vitamin D deficiency, N. Engl. J. Med. 357 (2007) 266–281. [13] IPCC, Climate Change 2013: The Physical Science Basis, 2013. (accessed 01.08.13). [14] N. Italia, E.A. Rehfuess, Is the global solar UV index an effective instrument for promoting sun protection? A systematic review (vol. 7, pg. 200, 2012), Health Educ. Res. 27 (2012) 1129–1131. [15] M.G. Kimlin, R.M. Lucas, S.L. Harrison, I. van der Mei, B.K. Armstrong, D.C. Whiteman, A. Kricker, M. Nowak, A.M. Brodie, J.D. Sun, The contributions of solar ultraviolet radiation exposure and other determinants to serum 25hydroxyvitamin D concentrations in Australian adults: the AusD study, Am. J. Epidemiol. 179 (2014) 864–874. [16] NASA website, NASA OMI Data, National Aeronautics and Space Administration. (accessed 02.06.12). [17] M. Norval, R. Lucas, A.P. Cullen, F. De Gruijl, J. Longstreth, Y. Takizawa, J.C. Van der Leun, The human health effects of ozone depletion and interactions with climate change, Photochem. Photobiol. Sci. 10 (2011) 199–225. [18] D.L. O’Riordan, K. Glanz, P. Gies, T. Elliott, A pilot study of the validity of selfreported ultraviolet radiation exposure and sun protection practices among lifeguards, parents and children, Photochem. Photobiol. 84 (2008) 774–778. [19] A.V. Parisi, L.R. Meldrum, M.G. Kimlin, J.C.F. Wong, J. Aitken, J.S. Mainstone, Evaluation of differences in ultraviolet exposure during weekend and weekday activities, Phys. Med. Biol. 45 (2000) 2253–2262. [20] T. Schwarz, The dark and the sunny sides of UVR-induced immunosuppression: photoimmunology revisited, J. Invest. Dermatol. 130 (2010) 49–54. [21] R.G. Steadman, Norms of apparent temperature in Australia, Aust. Met. Mag. 43 (1994) 1–16. [22] J.D. Sun, R.M. Lucas, S.L. Harrison, I. van der Mei, D.C. Whiteman, R. Mason, M. Nowak, A.M. Brodie, M.G. Kimlin, Measuring exposure to solar ultraviolet radiation using a dosimetric technique: understanding participant compliance issues, Photochem. Photobiol. 90 (2014) 919–924. [23] E. Thieden, M.S. Agren, H.C. Wulf, The wrist is a reliable body site for personal dosimetry of ultraviolet radiation, Photodermatol. Photo 16 (2000) 57–61. [24] J.C. van der Leun, R.D. Piacentini, F.R. de Gruijl, Climate change and human skin cancer, Photochem. Photobiol. Sci. 7 (2008) 730–733. [25] A. van Dijk, H. Slaper, P.N. den Outer, O. Morgenstern, P. Braesicke, J.A. Pyle, H. Garny, A. Stenke, M. Dameris, A. Kazantzidis, K. Tourpali, A.F. Bais, Skin cancer risks avoided by the montreal protocol – worldwide modelling integrating coupled climate-chemistry models with a risk model for UV, Photochem. Photobiol. 89 (2013) 234–246. [26] L.M. Verbrugge, Health diaries, Med. Care 18 (1980) 73–95. [27] P. Webb, C. Bain, Essential Epidemiology: An Introduction for Students and Health Professionals, Cambridge University Press, 2010. [28] D.C. Whiteman, M. Stickley, P. Watt, M.C. Hughes, M.B. Davis, A.C. Green, Anatomic site, sun exposure, and risk of cutaneous melanoma, J. Clin. Oncol. 24 (2006) 3172–3177. [29] F. Xiang, R. Lucas, S. Hales, R. Neale, Incidence of nonmelanoma skin cancer in relation to ambient UV radiation in white populations, 1978–2012: empirical relationships, JAMA Dermatol. 150 (2014) 1063–1071. [30] J. Yam, A. Kwok, Ultraviolet light and ocular diseases, Int. Ophthalmol. 34 (2014) 383–400.

Weekend personal ultraviolet radiation exposure in four cities in Australia: influence of temperature, humidity and ambient ultraviolet radiation.

To examine the effects of meteorological factors on weekend sun exposure behaviours and personal received dose of ultraviolet radiation (UVR) in Austr...
1MB Sizes 0 Downloads 7 Views