Article Text
Abstract
Background: Identified aetiological factors for chronic widespread pain (CWP) are largely related to emotional and behavioural factors, but current management leads to modest improvement in symptoms. Vitamin D deficiency has been suggested as a new modifiable risk factor for CWP.
Objective: To examine the association between vitamin D status (measured by 25-hydroxyvitamin D (25(OH)D)) and CWP in a nationwide population sample of white British adults, accounting for potential mediating and confounding lifestyle factors.
Methods: 9377 participants born 1 week in March 1958, in England, Scotland or Wales and completing a biomedical assessment at age 45; 6824 eligible participants had data on 25(OH)D and completed pain manikins.
Results: Prevalence of CWP varied by 25(OH)D concentration in women but not in men, with the lowest prevalence observed for women with 75–99 nmol/l (14.4% for <25 nmol/l, 14.8% for 25–49 nmol/l, 11.6% for 50–74 nmo/l, 8.2% for 75–99 nmol/l and 9.8% for participants with ⩾100 nmol/l). There was an interaction between 25(OH)D concentration and gender in relation to CWP (interaction, p = 0.006), which was not fully explained by differences in lifestyle or social factors (adjusted interaction, p = 0.03). For women, the association between 25(OH)D concentration and CWP persisted after full adjustment (odds ratio (OR) for <75 nmol/l vs 75–99 nmol/l 1.57, 95% CI 1.09 to 2.26), while no evidence for an association was apparent in men (OR = 1.03, 95% CI 0.75 to 1.43).
Conclusion: Current vitamin D status was associated with CWP in women but not in men. Follow-up studies are needed to evaluate whether higher vitamin D intake might have beneficial effects on the risk of CWP.
Statistics from Altmetric.com
Chronic widespread pain (CWP) affects approximately 10% of the UK population at any time.1 2 It is thought to be a multifactorial condition: to date, the focus has been on psychosocial factors, and although associations with general psychological distress,3 depression4 and somatisation,5 among others, have been consistently observed, the translation into successful management strategies has been modest.6 The aetiology of CWP is not well understood, and the possible influence of endocrine or immunological factors remains to be firmly established.7
One putative biological risk factor for CWP is vitamin D deficiency. Vitamin D is a hormone precursor (obtained either through diet or skin synthesis), which before exerting its metabolic effects undergoes two successive hydroxylations; the first converting vitamin D to 25-hydroxyvitamin D (25(OH)D, which provides an indicator of vitamin D status)8 and the second to the main active hormonal form, 1,25-dihydroxyvitamin D (1,25(OH)2D).
A clinical feature of osteomalacia, a disorder of bone mineralisation caused by severe vitamin D deficiency, is generalised bone pain.9 Up to 93% of patients with persistent pain have been reported to have hypovitaminosis D.10 11 Furthermore, vitamin D supplementation for 3 months resulted in pain resolution in 95% of patients with chronic low back pain in one study,12 and after 6 months, symptom resolution (including, for 52% of patients, chronic back pain) among women with hypovitaminosis D in another study.13
The relationship between vitamin D status and pain in the general population is not well characterised. A greater prevalence of low 25(OH)D concentrations has been reported in participants with chronic pain compared with others among national and immigrant Turkish in Germany14 and British South Asians.15 However, small sample sizes prevented appropriate examination of the role of potential confounding factors in these studies. Whether variations in 25(OH)D concentrations are associated with the prevalence or risk of CWP, and how personal characteristics, lifestyle and social variations affect these associations, has not to our knowledge been previously investigated in general population samples.
The threshold for optimum vitamin D status is known to be considerably higher than that required to prevent osteomalacia.16 Hence, it is important to establish the risk of pain across a broader range of 25(OH)D concentrations. Hypovitaminosis D is a worldwide problem,17 which is particularly prevalent in Britain.18 A recent biomedical survey of the nationwide 1958 British birth cohort (aged 45 years) included measurements of CWP and serum 25(OH)D in nearly 7000 people. The 1958 cohort is a predominantly white Caucasian population (98%), which has been particularly well characterised for factors that may confound any association between vitamin D status and pain. Given the sparse data from white population-based samples, we examined the association between serum 25(OH)D concentrations, lifestyle variations and the prevalence of CWP.
PATIENTS AND METHODS
The data are from the 45-year biomedical survey of the 1958 British birth cohort. All cohort members were born in England Scotland and Wales, during 1 week in March 1958, with follow-up at intervals from birth to 45 years.19–21 At age 45 a target of 11 971 cohort members were invited to a clinical examination undertaken in their home by a trained nurse. A total of 9377 participants were seen between September 2002 and March 2004, of whom 7591 provided a blood sample from which 25(OH)D was measured. Participants from non-white ethnic groups (n = 159), or who were taking neuroleptic or anti-epileptic drugs (known to affect vitamin D metabolism) (n = 37) or who were pregnant (n = 1) were excluded. Of the remainder, 570 did not provide information on musculoskeletal pain, leaving 6824 for the current analyses. Ethical approval was obtained from the South East Multi-Centre Research Ethics Committee (ref 01/1/44).
Measures
25(OH)D was measured from non-fasting venous blood samples using an ELISA (IDS OCTEIA Elisa; IDS, Bolton, UK) and an analyser (BEP 2000; Dade-Behring, Marburg, Germany) with sensitivity of 5.0 nmol/l, linearity ⩽155 nmol/l, and intra-assay coefficient of variation 5.5–7.2%.22 25(OH)D concentrations were standardised according to the mean of the values found by the Vitamin D External Quality Assessment Scheme (DEQAS).
Pain was assessed by a questionnaire sent in advance of the nurse interview, using blank body manikins on which respondents shaded areas of pain lasting for 1 day or longer in the past month; participants were then asked whether they had been aware of the pain for at least 3 months. CWP was defined according to the American College of Rheumatology criteria for fibromyalgia—namely, contralateral body quadrant pain and axial skeleton pain lasting for at least 3 months.23
Covariates
Social class was determined by the Registrar General’s classification,24 based on current or most recent occupation at age 42 (or age 33 if missing) and categorised as I and II (professional and managerial), III (skilled) non-manual, III (skilled) manual and IV and V (partly skilled and unskilled).
Cigarette smoking was recorded at age 42 with reports of never smoking verified using data from ages 23 and 33. At age 45, alcohol consumption was measured using AUDIT (Alcohol Use Disorders Identification Test), within which typical frequency of drinking and standard drinks per session were recorded.25 Drinks per week were calculated and consumption classified into non-drinker, light (<7 drinks/week), moderate (7–14 drinks/week), heavy (14–21 drinks/week) and very heavy (>21 drinks/week).
Information on usual time spent outdoors in daylight hours in the previous month (<1 h/day, 1–2 h/day, ⩾3 h/day) and time per day spent watching TV or using a PC, or both (<7 h/week, 7–20 h/week, ⩾21 h/week).
Dietary supplements in the past month including cod liver/fish oil or other supplements containing vitamin D were recorded. Consumption of oily fish such as salmon, trout, mackerel, sardines and fresh tuna was classified into weekly, less than weekly and never.
Recreation metabolic equivalent (MET) hours per week at 45 years were derived from reported frequencies and usual durations for up to 37 activities, and published MET scores.26 For analyses presented here, recreational activity was divided into quartiles, calculated for men and women separately. An additional category was created for implausibly high values (participants with weekly recreation hours of 3SD above the gender mean). Participants engaging in vigorous activity were those who recorded an activity with a MET score >6.27
Height and weight at age 45 were measured with the participant lightly clothed and unshod. Waist circumference was measured by the nurse midway between the costal margin and iliac crest.
Geographic location, based on region of residence at age 45, was grouped as follows: South (South West, South East, Greater London), Middle (East Anglia, Midlands, Wales), North (North, North West, Yorkshire and the Humber) and Scotland.
Statistical analysis
Social and lifestyle factors were examined for their association with mean 25(OH)D concentrations and CWP in order to identify relevant confounding factors. Differences in CWP prevalence were examined using non-parametric tests for trend; likelihood ratio tests (LRTs) from linear regression models were used to test for trend in mean 25(OH)D. Serum 25(OH)D varies by season, hence mean values are presented weighted by the number of subjects in each season (“standardised by season”). The distribution of 25(OH)D was skewed. To achieve a normal distribution we used a logarithmic transformation for calculations of geometric means and for use of 25(OH)D as an outcome in regression models.
Associations between 25(OH)D and CWP were examined using logistic regression: the odds ratios (ORs) for CWP were calculated for (a) each category of 25(OH)D concentration (reference category 75–100 nmol/l) and (b) per unit increase of 25(OH)D (ie, modelled as a continuous variable). To evaluate whether the association of 25(OH)D and CWP was linear, we added a quadratic (squared) term of 25(OH)D to the model: deviation from linearity was not found to be statistically significant. However, given the suggestion of non-linearity at the extremes of 25(OH)D concentrations, both categorical and continuous analyses are presented.
Models were performed in three stages: (a) adjusted for month of measurement; (b) adjusted in addition for social and lifestyle factors (adult social class, smoking status, alcohol consumption, TV/PC time, recreation MET hours and vigorous activity) and (c) adjusted further for social and lifestyle factors plus measures of obesity (body mass index (BMI), waist circumference). All data are presented separately for men and women, given a significant gender interaction in logistic regression analyses of CWP (tested with interaction between gender and continuous 25(OH)D, p<0.05, adjusted for month of measurement). We further adjusted the gender interaction model for all indicators in supplementary table 1 (available online), including supplement use, oily fish intake and time spent outdoors, in addition to indicators in model (c) above. This was done to evaluate whether variations in vitamin D intake or lifestyle might explain the gender difference in the association of 25(OH)D and CWP.
Information on the social and lifestyle factors which were included in the final models were missing for 497 participants. We used multiple imputation to estimate information on missing covariates.28 Results were not affected by treatment of missing information, hence final models are presented for the sample with complete data (n = 6336). All analyses were conducted using Stata 9.2.
RESULTS
The prevalence of CWP was 11.4% and 12.5% for men and women, respectively. Women with CWP had mean 25(OH)D concentrations of 46.6 nmol/l compared with 52.2 nmol/l for those who did not report CWP; equivalent concentrations for men were 53.1 nmol/l and 53.4 nmol/l, respectively.
Prevalence of CWP did not vary by season or region of residence, while the association with BMI was U-shaped with a higher prevalence of CWP in underweight and severely obese subjects, these groups having the lowest 25(OH)D concentrations (table 1). Some variation in the average 25(OH)D concentrations was observed by season, region and all social and lifestyle factors included in supplementary table 1 (available online). The prevalence of CWP was higher in smokers than in non-smokers, and in alcohol abstainers than in consumers (supplementary table 1). Social gradients in CWP were stronger than those observed for vitamin D. From the direct indicators of vitamin D intake, vitamin D supplementation was not associated with CWP, while an association with time spent outdoors was observed for men but not for women (LRT interaction p = 0.01). An association between physical activity (as indicated either by vigorous activity or recreational MET hours) and CWP varied by gender (LRT interaction p⩽0.01 for all comparisons), with a stronger association in men than in women.
There was an interaction between 25(OH)D concentration and gender in relation to CWP prevalence (LRT interaction p = 0.006), and this was not explained by gender differences in associated social and lifestyle factors (LRT interaction after adjustment for indicators in table 1 and supplementary table 1, p = 0.03). Table 2 shows the odds ratios for CWP prevalence for each category of 25(OH)D concentration and per 10 nmol/l increase of 25(OH)D in unadjusted and adjusted models.
For women, there was a significant association between 25(OH)D concentration and CWP (p<0.001). The lowest prevalence of CWP was observed for women with 25(OH)D 75–99 nmol/l. The association between 25(OH)D and CWP in women was somewhat attenuated by adjustment for social and lifestyle factors (social class, smoking, alcohol, physical activity and TV/PC time), with further attenuation after adjustment for adiposity (BMI and waist circumference; table 2). After full adjustment, there was an association between 25(OH)D and CWP with the main contrast being between women with 25(OH)D <75 nmol/l compared with 75–99 nmol/l (OR after full adjustment 1.57, 95% CI 1.09 to 2.26). In contrast, there was no evidence of an association between 25(OH)D and CWP in men across the range of 25(OH)D concentrations, nor were men with <25 nmol/l 25(OH)D at increased risk (table 2). Additional adjustment for time spent outdoors did not alter any of the observed associations between 25(OH)D and CWP (data not shown).
DISCUSSION
Our findings from a Caucasian, nationwide population sample of British adults provide little support for current vitamin D status as a major determinant for CWP. A significant association was observed between 25(OH)D and CWP prevalence for women. However, there was no indication for a dose-dependent effect and 25(OH)D was not associated with CWP in men. In our study the lowest prevalence of CWP was observed for women with 25(OH)D 75–99 nmol/l. This is intriguing given that 25(OH)D >75 nmol/l has been previously suggested as the cut-off point for optimal bone health.29 30 To our knowledge, this study is the largest population-based examination of the association between vitamin D status and CWP to date, and the first to consider related variations in lifestyle factors or to focus on white ethnic groups.
Methodological considerations
A strength of this study is its basis in a large population sample, well characterised with respect to potential confounders, covering a range of concurrent social, lifestyle and demographic factors. CWP was defined according to American College of Rheumatology criteria.
We acknowledge some limitations in relation to these findings. Although 25(OH)D is accepted as the best measure of vitamin D status during the past 3–4 weeks,31 it may not represent the participants’ status during the time that CWP developed, which might have been years earlier. Furthermore, given the cross-sectional design, it is not possible to establish whether suboptimal vitamin D status results in an increased risk of CWP, or whether changes in the behaviour of subjects with CWP result in reduced 25(OH)D concentrations (eg, functional limitation associated with CWP may affect the proportion of time spent indoors). However, the association between 25(OH)D and CWP in women persisted even after adjustment for several factors that might potentially mediate a decrease in 25(OH)D concentrations among those with CWP.
Analyses were restricted to Caucasian participants, hence our results may not apply to non-white ethnic groups. Immigrants to northern countries are typically reported to have a higher prevalence of CWP and lower concentrations of 25(OH)D than the native population,15 hence it is possible that hypovitaminosis D may play a greater role in CWP in subjects from non-white ethnic groups. As previously described, some sample attrition occurred during the follow-up to age 45 years and although the sample remains broadly representative of the surviving cohort, there is some under-representation of specific minority groups.32
Comparisons with other studies
Evidence for an association between vitamin D status and pain has been obtained from case studies reporting low 25(OH)D concentrations in affected subjects.10–13 33 Pain is a symptom of osteomalacia caused by severe vitamin D deficiency.34 In various case studies of vitamin D-deficient subjects, pain has been shown to be relieved by vitamin D supplementation.13 35 An earlier study suggested a markedly greater prevalence of low 25(OH)D (<10 nmol/l, OR = 3.5, 95% CI 0.4 to 31) in British South Asian women with CWP than in those not reporting CWP.15 Compared with unaffected subjects, lower 25(OH)D concentrations have been reported in those with prolonged bone pain in a mixed sample of German and Turkish populations.14 In this previous study gender was a predictor of bone pain; however, the authors did not investigate gender differences in the association between 25(OH)D and CWP. The lack of a stronger association between 25(OH)D and CWP in our general population sample may suggest that pain is primarily indicative of a severe vitamin D deficiency rather than a more gradual response to varying concentrations.
The prevalence of CWP was similar to the reported UK population prevalence of 11–13%2; we observed a slight female preponderance, though not as great as the 1.5 excess reported elsewhere.1 Both cigarette smoking and obesity have been identified as risk factors for a number of chronic pains, including CWP.36–38 Interestingly, in an earlier study of low back pain in the cohort, those with high BMI at 23 years and those with BMI <15th centile were at increased risk of low back pain.36 It is likely that for some of the lifestyle factors (eg, alcohol consumption, physical activity), associations may be explained by reverse causation, with those experiencing CWP changing their behaviour owing to the illness.
Explanations
It is unclear why an association between 25(OH)D and CWP in our study was seen in women but not in men. Given that women in the 1958 cohort are still mostly premenopausal (45 years), it can be speculated that the known influences of hormonal vitamin D on the regulation of oestrogen activity may at least partly contribute to this gender difference.39
We examined variations in lifestyle as an explanation for the observed gender interaction, but statistical evidence for a differing association of 25(OH)D with CWP in men and women persisted even after full adjustment for social and lifestyle characteristics. We found some indication that men and women with CWP behave differently in relation to factors that are associated with vitamin D status—notably, physical activity and time spent outdoors. Differences in the way in which men and women with CWP cope with, or manage, their pain have previously been reported. Women with pain are more likely to seek healthcare than men,40 41 and to exhibit behaviours reflecting the experience of pain.42 Furthermore, there is evidence to suggest that CWP is a manifestation of the process of somatisation,5 with strong associations seen between mood and CWP. Information on somatisation is not available from the cohort. However, adjustment for related factors such as symptoms of depression or anxiety did not affect the associations of 25(OH)D and CWP (data not presented). Nevertheless, given the observational nature of these data, we cannot exclude the possibility that our finding of an association between vitamin D status and CWP in women (or the lack of any association in men) is confounded by unmeasured factors.
Osteomalacia, a disease of extreme vitamin D deficiency, is associated with isolated or generalised bone pain.9 In addition to its action on bone, there is increasing evidence of non-calcaemic roles for 1,25(OH)2D, the hormonally active form of vitamin D.43 1,25(OH)2D is involved in the regulation of immune responses and this is believed to explain the associations with autoimmune diseases such as type 1 diabetes.44 Intriguingly, one hypothesis for the cause of fibromyalgia (of which CWP is a major constituent) is the activation of inflammatory cytokine responses after infection; cytokines having been shown to modulate pain perception both peripherally and centrally.45
CONCLUSIONS
These results do not support current vitamin D status as a key determinant for CWP; some evidence for an association was observed for women, while in men CWP was unrelated to 25(OH)D concentration. Follow-up studies are needed to evaluate whether higher vitamin D intake might have beneficial effects on CWP risk.
Acknowledgments
We thank Dr Ian Gibb and Steve Turner (Royal Victoria Infirmary, Newcastle-upon-Tyne) for carrying out the laboratory assays. Data were provided by Centre for Longitudinal Studies, Institute of Education (original data producers).
REFERENCES
Footnotes
Competing interests: None.
Funding: Data collection at age 44-46 years and support for KA and DJB was provided by the UK Medical Research Council (grant G0000934), 25(OH)D assays were funded by the BUPA Foundation. EH is funded by the Department of Health (UK) Public Health Career Scientist Award. This work was undertaken at Great Ormond Street Hospital/University College London, Institute of Child Health, which received a proportion of funding from the Department of Health’s National Institute of Health Research (“Biomedical Research Centres” funding). The Medical Research Council provides funds for the MRC Centre of Epidemiology for Child Health.
Ethics approval: Ethical approval was obtained from the South East Multi-Centre Research Ethics Committee (ref 01/1/44).
▸ An additional table is published online only at http://ard.bmj.com/content/vol68/issue6