Article Text

Download PDFPDF

Original article
Working patterns and vitamin D status in mid-life: a cross-sectional study of the 1958 British birth cohort
  1. Maria Ward1,2,
  2. Diane J Berry1,
  3. Chris Power1,
  4. Elina Hyppönen1
  1. 1Centre for Paediatric Epidemiology and Biostatistics and MRC Centre of Epidemiology for Child Health, UCL Institute of Child Health, London, UK
  2. 2School of Medicine, Dentistry and Biomedical Sciences, Queen's University Belfast, Northern Ireland, UK
  1. Correspondence to Dr Elina Hyppönen, MRC Centre for the Epidemiology of Child Health, UCL Institute of Child Health, 30 Guilford Street, London WC1N 1EH, UK; e.hypponen{at}


Objectives To assess the association between working patterns and vitamin D status in men and women and to determine the potential influence of related lifestyle and socioeconomic factors.

Methods The authors used data from the 1958 British birth cohort (aged 45 years) and 6154 participants, who were in full-time work, were included in current analyses. Vitamin D status was measured by circulating concentrations of 25-hydroxyvitamin D (25(OH)D). Information on working patterns and lifestyle factors was obtained using a structured questionnaire administered at 45 years.

Results Manual social class was strongly associated with vitamin D-related lifestyle factors, with those in manual classes not only spending more time outdoors, but also spending more time watching TV/using PC, consuming less supplements and oily fish. Associations between working patterns and vitamin D-related lifestyles were less clear: night work was not strongly associated with lifestyles in either gender, while working hours were associated with time spent outside, PC/TV leisure time and use of supplements in men but not in women. In men, working patterns were not associated with lower 25(OH)D concentrations. In women, 25(OH)D concentrations were 8% lower (95% CI 15% to 2%) in night workers compared with others, while women working less than 35 h/week had 5% higher concentrations of 25(OH)D (95% CI 1% to 8%) compared with those working 35–40 h/week after adjustment for season, social class and body mass index (BMI).

Conclusions Women working nights and longer hours may be vulnerable to deficits in vitamin D status and associated health hazards.

  • Vitamin D
  • occupational status
  • health behaviour
  • epidemiology
  • public health
  • nutrition
  • preventive medicine
  • shift work

Statistics from

What this paper adds

  • Work patterns (long hours and night work) were potential influences on vitamin D status: this was the first study to assess these relationships.

  • Night work and long work hours were associated with adverse vitamin D status in British women in mid-adulthood.

  • Socioeconomic status was not associated with vitamin D status, but the gender-patterned variations require further investigation.


Vitamin D deficiency is common with up to 1 billion people estimated to be affected globally.1 Prevalence of deficiency is particularly high in Western populations,2 where the problem is believed to be aggravated by working patterns including long hours of work indoors.3 Although initially classified as ‘a vitamin’, vitamin D is more correctly a pro-hormone, much of which is derived from synthesis by the skin upon exposure to UVB radiation from sunlight. It is also found naturally in certain foods such as oily fish.4 Vitamin D obtained from these sources is converted in the liver to 25-hydroxyvitamin D (25(OH)D), which is the major circulating metabolite and the best available indicator for individual vitamin D status. Renal conversion of 25(OH)D to the active metabolite 1,25-dihydroxyvitamin D (1,25(OH)2D) is tightly regulated by homeostatic mechanisms.1 There has been extensive interest in the health effects of vitamin D during the past decade. In addition to the long appreciated key role in bone health, evidence now suggests the importance of vitamin D for a wide-range of diseases such as cancer, autoimmune disorders, cardiovascular and infectious diseases.1

As a substantial proportion of time in adult life is spent working indoors, this has a major influence on sunlight exposure of individuals. It is therefore expected that patterns and type of work would affect an individual's level of sun exposure and their subsequent vitamin D status. Despite extensive literature on lifestyle and vitamin D deficiency, remarkably little is known about the role of working conditions on vitamin D status. Studies examining the association of social class and vitamin D status have generally not found large differences.5 6 There is some interesting anecdotal evidence from case reports showing vitamin D deficiency in night workers. For example, there is a report of osteomalacia in a night nurse who worked for 20 years between the hours of 15:00 and 23:00 and avoided direct sunlight exposure by sleeping during daylight hours.7 In 1984, Kiriyama et al8 described a case of recurrent osteomalacia in a 31-year-old bartender who rested during the daytime. The essential importance of sunlight exposure to ensure sufficient vitamin D status was also demonstrated by a study of submariners who despite having a ‘good diet’ showed a marked deterioration of vitamin D status after an absence of sunlight for 2 months.9

The aim of this study was to determine whether vitamin D status (measured by 25(OH)D) is associated with working patterns, specifically night work and duration of the working week. We used data from a large cross-sectional study of middle-aged Britons (namely the 1958 British birth cohort)10 with analyses carried out among 6154 participants in paid employment at age 45 years. Due to the extensive information available,5 we were able to examine whether working patterns were associated with lifestyle factors related to vitamin D intake (such as vitamin D supplementation or oily fish consumption) and to quantify how these factors contributed to any differences in vitamin D status observed.

Participants and methods

Participants are from the 1958 British birth cohort, which consists of a large sample of individuals born in England, Scotland or Wales in 1 week in March and enrolled in the Perinatal Mortality Survey.11 Individuals have been surveyed throughout the life course. At age 45 years a target sample (n=11971) was identified from those who had not died, emigrated or registered permanent refusal: between September 2002 and April 2004, 9377 (78%) individuals completed a questionnaire of whom 8302 provided a blood sample, and 7591 had valid 25(OH)D measures. For the current analyses, we excluded the small minority of participants from non-Caucasian ethnic groups (n=158, 2.1%) and one participant who was pregnant at the time of the study. The sample was further restricted to those who were in paid work at 45 years leaving 6154 participants for the analyses (figure 1). Ethical approval for the 45-year survey was obtained from the South East Multi-Centre Research Ethics Committee.

Figure 1

A schematic representation of numbers of participants in the 45-year survey of the 1958 British birth cohort and the selection of participants for the current study.


Vitamin D status

Measurement of serum 25(OH)D concentrations was conducted using automated application of an ELISA (IDS OCTEIA Elisa; IDS, Bolton, UK) and an analyser (BEP 2000; Dade Behring, Milton Keynes, UK) with sensitivity of 5.0 nmol/l, linearity of 155 nmol/l and intra-assay coefficient of variation of 5.5%–7.2%.12 The concentrations were standardised according to the mean of the values from the Vitamin D External Quality Assurance Survey to ease the interstudy comparison.12 For this paper, <50 nmol/l was chosen as the cut-off value to indicate vitamin D insufficiency.13 14

Work patterns

At age 45 years, participants recorded in a self-completed booklet the number of hours spent in paid work per average week in their main job, including work brought home. Based on the European Union Working Time Directive,15 the categories used in this study were as follows: <35 h/week, 35–40 h/week, 41–48 h/week and >48 h/week corresponding to part-time, full-time, lower overtime and extended working. Participants were classified as regular night workers if they reported that they normally worked during the night (between 24:00 and 06:00) in a self-completed questionnaire given along with their saliva sample collection.

Lifestyle factors

Time spent outside, PC/TV leisure time, supplement use and oily fish consumption were also reported in a questionnaire completed by the participants at age 45 years. For the purpose of statistical analyses, each lifestyle factor was re-coded into binary categories. Usual time spent outside during daylight hours in the previous month was coded as ‘<3 h spent outside/day’ and ‘>3 h spent outside/day’. Oily fish consumption, reported as the frequency of eating oily fish such as salmon, trout, mackerel or sardines, was coded as ‘weekly’ (incorporating from 2–4 times/day to 1–2 days/week) and ‘less than weekly’ (less than once/week, occasionally or never). A question regarding ‘the use of cod liver or fish oil or any other vitamin D-containing supplement’ was used to determine if vitamin D supplementation was used. The indicator for time spent using a PC or watching TV (‘PC/TV leisure time’) combined the reported average leisure time spent using a computer or viewing TV (or video) during the last year (coded as ‘<3 h/day’ vs ‘≥3 h/day’). Usual frequency and duration of various physical activities were reported and used to calculate recreational metabolic equivalent (MET) hours/week.16 Recreational MET hours were categorised into gender-specific quartiles, with an additional category created for implausibly high values (participants with weekly recreation hours of 3SD above the gender mean).17

Other covariates

Obesity was defined as body mass index (BMI)>30. To calculate BMI (weight in kg/height in m2), weight and height were measured using scales and a stadiometer with participants wearing light clothes and no shoes. Data missing at 45 years were supplemented with data from 42 years. Socioeconomic position at birth and 42 years was determined by the social class based on occupational class (Registrar General's classification)18 and grouped into four categories: professional/managerial (I and II), non-manual (III), manual (III) and unskilled manual (IV and V). Those in a manual occupation were coded as working in categories manual III, IV or V.

Data analysis

Statistical analyses were conducted using the software application Stata version 11. We included the comparison of participants who were in paid work versus those not in paid work in the initial descriptive analyses only, while all subsequent analyses evaluating associations with working patterns were restricted to cohort members who were all in paid work. The unadjusted relationship between work patterns (paid work, night work, hours worked per week and manual work) and dichotomised lifestyle factors (time spent outside, time spent using/watching PC/TV, vitamin D supplements, oily fish consumption and obesity) was assessed using two-way tables and tested with Pearson's χ2 test of independence. Associations between work pattern exposures and dichotomised lifestyle outcomes were evaluated with logistic regression adjusting for adulthood social class.

The distribution of 25(OH)D was slightly skewed and was natural log transformed to achieve a normal distribution. Accordingly, geometric means are presented for 25(OH)D. Associations between night work and hours worked per week with 25(OH)D were investigated using multiple linear regression, adjusting for (1) season of blood sample collection, (2) season, social class at birth and 42 years, (3) season, social class at birth and 42 years and BMI, (4) season, social class at birth and 42 years, BMI and smoking, alcohol consumption, recreational MET hours, PC/TV leisure time, time spent outdoors, oily fish consumption and vitamin D supplementation. The relation of work pattern exposure variables with vitamin D insufficiency was investigated by logistic regression. Missing data on covariates were imputed 20 times using multiple imputation as implemented in Stata version 11.19 Proportion with missing information in lifestyle factors varied between 0.1% (obesity) and 3.3% (PC/TV leisure time). Similar associations were seen in the models regressed on the imputed datasets compared with the complete dataset. Sensitivity analyses were carried out also by imputing information on night work (missing for 891 men and 767 women), with similar results obtained from imputed and complete case models. As night work was a key exposure variable in this study, and the proportion of missing information was fairly large, night work in table 1 is presented for participants with available information while missing data on lifestyle and social covariates has been imputed.

Table 1

Associations of working patterns and manual social class with obesity and selected lifestyles in the 1958 British birth cohort

All analyses were conducted stratified by gender. The rationale for conducting analysis on men and women separately is based on literature describing occupational gender segregation: men typically work different hours in different occupations from women even within the same broad socioeconomic groups.20 Interactions between gender and work patterns with lifestyle factors were tested using logistic regression.


Average 25(OH)D concentrations were higher in participants who were in paid work (52.9 nmol/l) compared with others (48.3 nmol/l, p<0.001) (table 2). There were differences also in time spent outside and time spent watching TV/using a PC between those in paid work compared with others, however, paid work was not associated with the use of vitamin D supplementation or the frequency of oily fish consumption. The associations between working patterns and lifestyle factors in participants who were at paid work are shown in table 1—9.2% of men and 7.7% of women worked nights regularly (p=0.07). Men tended to have longer work hours with 58.6% of men and 18.2% of women working more than 41 h/week (p<0.001). More men worked in a manual occupation compared with women (43.8% and 28.2%, respectively, p<0.001).

Table 2

25(OH)D concentrations and related lifestyle indicators by working status in the 1958 British birth cohort

Among those in paid work, both men and women who did manual work were more likely to spend longer periods outside than those who did non-manual work (table 1). However, manual workers were also more likely to spend more time using PC/watching TV, to be obese, to take less vitamin D supplements and to consume oily fish less frequently. Night work was not associated with PC/TV leisure time, supplement use or oily fish consumption in men or women. There was a suggestion of an association between night work and time spent outdoors for both genders; however, for men it was explained by adjustment for social class. In men, but not in women, work hours were significantly associated with time spent outside, PC/TV leisure time and supplement use. In both genders, night work was associated with obesity but in opposing directions: women who worked nights tended to be more commonly obese compared with others while the opposite was found for men (interaction p=0.027). Obesity was more common among those working longer (>48) hours in both genders, and associations were not explained by social class differences.

There was no difference in serum 25(OH)D concentrations between participants in manual occupations compared with others (p=0.45 for men and 0.33 for women). There was an association between night work and 25(OH)D concentrations in women; concentrations were 8% lower (95% CI 15% to 2%) in night workers compared with others after adjustment for season, social class and BMI. No association was seen in men, although there was a suggestion of higher mean 25(OH)D in men who worked nights (table 3). Associations between work hours and 25(OH)D concentrations were in opposing directions for men and women: concentrations increased as work hours increased for men while a decrease was seen for women. Further adjustment for lifestyle factors including smoking, alcohol consumption, physical activity, PC/TV leisure time, time spent outdoors, oily fish consumption and vitamin D supplementation did not attenuate associations between hours worked and 25(OH)D concentrations in men or women (table 3). Overall, working patterns in women appear to be more strongly associated with adverse 25(OH)D status than in men. Mutual adjustment for night work and working hours did not affect the observed associations (data not presented).

Table 3

The association of night work and the duration of the working week with the prevalence of vitamin D insufficiency and log 25(OH)D concentrations in British men and women aged 45 years*

Only one association was observed for prevalence of vitamin D insufficiency (25(OH)D<50 nmo/l), with a 1.46 times greater risk (95% CI 1.09 to 1.96, adjusted for season, social factors and BMI) in women working >48 h compared with those working <35 h.


Working patterns determine much of how we use our time during waking hours and influence our lifestyles and health.21 Our findings from the 1958 British birth cohort suggest that night work and a long working week may have an adverse influence on vitamin D status in women but not in men. Some lifestyle factors related to vitamin D status varied by working patterns (most notably duration of the working week) and, paradoxically, associations were observed mainly for men and less so in women. The gender differences observed between work patterns and vitamin D status in this study are intriguing and merit further research in order to establish if there are longer-term health consequences. Interestingly, a recent case–control study demonstrated an inverse association between occupational UV exposure and risk of renal cell carcinoma in men but not in women.4 Vitamin D is suspected to be a contributing factor for the potential UV exposure and cancer risk association. Other studies have shown differences between men and women which are yet to be explained. For example, in Nordic men, occupational variation in the incidence of colon cancer was found with the lowest incidence in forestry workers, farmer and gardeners. No such variation in incidence of colon cancer was found in women.22

It is worth noting that working in a manual occupation had no association with vitamin D status and yet it was associated with several lifestyle indicators related to vitamin D status in both genders. This can be explained by the fact that the adverse effects of the lifestyle factors on vitamin D status (obesity, less oily fish and supplement use and more time using PC/TV) were counteracted by the higher proportion of time spent outside among those involved in manual work. The net result of these opposing trends is a null association between manual occupation and vitamin D status. Differences in lifestyle may also contribute to the gender differences observed in our study. Research has shown that women are still responsible for most of the housework and family responsibilities and as a result have less leisure time than their male partners.23 Our results add strength to this argument by confirming that more women spent less time outdoors and had less PC/TV leisure time than men.23

Implications of findings

Research on working patterns and health has thus far focused on areas such as shift work, night work and their impact on mental and cardiovascular health.24 To our knowledge, this is the first study to directly assess the relationship between work patterns and vitamin D status. Vitamin D status is currently attaining importance as a public health issue and a systematic review in 200825 stated that sufficient sun exposure to prevent vitamin D deficiency would probably avoid a larger burden of disease than the skin cancer burden due to excessive UVB exposure. In the global context, another consideration regarding our findings is the future implication in developing countries striving for economic growth through labour-intensive employment, particularly in relation to the well-being of women.26

The segregation of work according to gender is well documented whereby men and women are aggregated within different types of work (horizontal segregation) and in different roles within the same type of work (vertical segregation). Emslie et al20 found that, on average, women tend to aggregate in the lower end of the status scale within each category of work thus resulting in less autonomy and greater time pressure. The majority of women worked in female-dominated occupations such as nursing. It is also argued that women have often only gained entry to an occupation because it became less appealing to men due to perceived ‘female’ duties or poorer conditions of work. Working patterns were associated with vitamin D status in women, although no strong associations were seen with related lifestyle factors. Our study included many of the key lifestyle indicators believed to influence vitamin D status (notably time spent outdoors, supplement use and oily fish consumption), suggesting that further studies on factors underlying the association between working patterns and vitamin D status are warranted. Potential associations between working patterns and vitamin D status may have important implications in occupational medicine, especially as many societies currently appear to be tending towards both genders working nights and long hours.

It is noteworthy that those who work nights and long hours tend to have reduced sleep duration.27 28 Studies in sleep medicine postulate that the balance between circadian rhythm and homeostatic processes, which normally allow the body adequate sleep time, are disrupted by night work thus leading to shorter sleep duration.29 30 Low vitamin D status in night workers could aggravate any potential adverse influences of short sleep duration, as lack of sleep and vitamin D deficiency have both been implicated with negative health outcomes such as cardiovascular disease and depression.1 27

Strengths and limitations

The main strength of this study lies in the large nationwide sample, and in the detailed information on lifestyle and other factors. Other studies have been conducted on this population which help support this study, for example, the standardised measurement of 25(OH)D12 and the description of working patterns and the association with cardiovascular disease24 and cortisol secretion.31 The population is largely representative of the general white population; however, no information is available from other ethnic groups, which reduces the generalisability of our findings to modern British populations.10 However, it could be argued that given the very high prevalence of vitamin D deficiency in ethnic immigrant populations32 33 and possible differences in occupational distributions,34 vitamin D deficiency associations may differ. Furthermore, our study did not address the possible importance of unpaid work on 25(OH)D concentrations. Hence, caution will be required when making inferences based on the current findings outside the group of white Caucasian individuals in paid work in Britain.

Information on working patterns and lifestyle indicators available in our study were obtained by self-report and available indicators were relatively crude. Night work was ascertained from a question on whether participants normally worked nights which is not an exact definition. However, for the purpose of evaluating our hypothesis on the possible influence of night work on serum 25(OH)D concentrations this type of question was useful, as it was likely to identify individuals who worked nights often enough to have a potential influence on their lifestyle and consequently, vitamin D status. A shortcoming with our study was that information on night work was missing from a fairly large number of participants. However, associations between night work and 25(OH)D concentrations were similar whether obtained from multiple imputation or complete case analyses. Despite having a fairly extensive range of relevant lifestyle indicators (for a study of this kind), it is possible that data limitations contribute to the fact that the observed association between working patterns and serum 25(OH)D concentrations were not fully explained by time spent outdoors or other available indicators on lifestyle. Further research is needed in this area in order to establish causal patterns between vitamin D and its proposed health consequences and to allow advocacy for this public health issue.

In conclusion, this study found that night work and long working hours adversely affected vitamin D status independently of socioeconomic status in women but not in men. As vitamin D has recently been associated with a wide range of health benefits, working patterns may have significant health implications especially in women. This association warrants further investigation and preventive measures ought to be considered in the future.



  • Funding MW is funded by the Child Health Research Appeal Trust (CHRAT) studentship and EH by the Department of Health (UK) Public Health Career Scientist Award. Data collection at age 44–46 years and statistical analyses were funded by the UK Medical Research Council (grants G0000934, G0601653), 25(OH)D assays by The BUPA Foundation. This work was undertaken at GOSH/UCL Institute of Child Health which received a proportion of funding from the Department of Health's NIHR Biomedical Research Centres funding scheme. MRC provided support for the MRC Centre of Epidemiology for Child Health.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval This study was conducted with the approval of the South East Multi-Centre Research Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.