ARTHRITIS & RHEUMATISM

Vol. 64, No. 12, December 2012, pp 39263935

DOI 10.1002/art.37687
2012, American College of Rheumatology

Vitamin D, Race, and Experimental Pain Sensitivity in

Older Adults With Knee Osteoarthritis

T. L. Glover,1 B. R. Goodin,1 A. L. Horgas,1 L. L. Kindler,2 C. D. King,1 K. T. Sibille,1

C. A. Peloquin,1 J. L. Riley III,1 R. Staud,1 L. A. Bradley,3 and R. B. Fillingim4

Objective. Low circulating serum levels of 25- underwent quantitative sensory testing, including mea-
hydroxyvitamin D (referred to hereafter as vitamin D) sures of sensitivity to heat-induced and mechanically
have been correlated with many health conditions, in- induced pain.
cluding chronic pain. Recent clinical practice guidelines Results. Blacks had significantly lower levels of
define vitamin D levels <20 ng/ml as deficient and levels vitamin D compared to whites, demonstrated greater
of 2129 ng/ml as insufficient. Vitamin D insufficiency, clinical pain, and showed greater sensitivity to heat-
including the most severe levels of deficiency, is more induced and mechanically induced pain. Low levels of
prevalent in black Americans. Ethnic and race group vitamin D predicted increased experimental pain sensi-
differences have been reported in both clinical and tivity, but did not predict self-reported clinical pain.
experimental pain, with black Americans reporting in- Group differences in vitamin D levels significantly pre-
creased pain. The purpose of this study was to examine dicted group differences in heat pain and pressure pain
whether variations in vitamin D levels contribute to race thresholds at the index knee and ipsilateral forearm.
differences in knee osteoarthritis pain. Conclusion. These data demonstrate that race
Methods. The sample consisted of 94 participants differences in experimental pain are mediated by differ-
(74% women), including 45 blacks and 49 whites with ences in the vitamin D level. Vitamin D deficiency may
symptomatic knee osteoarthritis. Their average age was be a risk factor for increased knee osteoarthritis pain in
55.8 years (range 4571 years). Participants completed black Americans.
a questionnaire on knee osteoarthritis symptoms and
The last decade has witnessed a dramatic in-
Supported by the NIH (National Institute on Aging grant crease in research related to the nonskeletal effects of
AG-033906) and the University of Florida Clinical and Translational vitamin D. In its known role as a vitamin, this micronu-
Science Institute, which is supported by NIH grant RR-029890. Ms
Glover is a John A. Hartford Foundation Building Academic Geriatric
trient aids calcium absorption. However, recent research
Nursing Capacity Scholar and a Mayday Fund grantee (grant AAN on vitamin D has focused on its hormonal actions.
11-116). Recommended daily intake of vitamin D has historically
1
T. L. Glover, MSN, ARNP, FNP-BC, B. R. Goodin, PhD,
A. L. Horgas, PhD, RN, C. D. King, PhD, K. T. Sibille, PhD, C. A.
been aimed at primary prevention of osteomalacia and
Peloquin, PharmD, FCCP, J. L. Riley III, PhD, R. Staud, MD: osteoporosis. In its biologically active form, vitamin D is
University of Florida, Gainesville; 2L. L. Kindler, PhD, RN, CNS: a secosteroid involved in regulating cell differentiation,
University of Portland, Portland, Oregon; 3L. A. Bradley, PhD: Uni-
versity of Alabama at Birmingham; 4R. B. Fillingim, PhD: University proliferation, angiogenesis, and apoptosis (1). Vitamin
of Florida and North Florida/South Georgia Veterans Health System, D is synthesized through the skin with adequate expo-
Rehabilitation Outcomes Research Center, Gainesville, Florida. sure to ultraviolet B (UVB) light. Following synthesis
Dr. Staud has received consulting fees, speaking fees, and/or
honoraria from Pfizer (less than $10,000). Dr. Fillingim has received and conversion, vitamin D is hydroxylated in the liver,
consulting fees, speaking fees, and/or honoraria from Cytogel Pharma, then in the kidneys. Vitamin D receptors can be found
Curatio, Algynomics, Codman, and Medscape (less than $10,000 each) on nearly all nucleated cells (2). Vitamin D deficiency
and owns stock or stock options in Algynomics.
Address correspondence to T. L. Glover, MSN, ARNP, has been correlated with diabetes, cancer, and decreased
FNP-BC, University of Florida, College of Dentistry, 1395 Center immunity (1).
Drive, Dental Tower, Room D2-13, PO Box 100404, Gainesville, FL Less than sufficient circulating serum levels of
32610. E-mail: tglover@ufl.edu.
Submitted for publication January 31, 2012; accepted in 25-hydroxyvitamin D (referred to hereafter as vitamin
revised form August 23, 2012. D) have been noted across populations in epidemiologic
3926
VITAMIN D, RACE, AND PAIN 3927

studies. Seasonal variation in the angle of the suns Research suggests that the brain reorganizes in the
UVB rays decreases opportunities for vitamin D syn- presence of chronic pain, which may reflect fundamental
thesis in northern latitudes and in the winter months. changes in how the brain processes pain-related infor-
Other factors contributing to inadequate vitamin D mation (22). Despite advances in the basic science
levels include more time spent indoors, increased use of understanding of pain pathways and processing, there
sunscreen, and the prevalence of obesity (3). As a remain vast individual differences in response to the
fat-soluble nutrient, vitamin D is sequestered in fat cells, clinical treatment of pain. Furthermore, pain remains
decreasing its availability for hormonal actions in the undertreated, especially for older adults and nonwhite
bloodstream. The bodys ability to synthesize vitamin D populations (2327). Previous research in our laboratory
in the skin lessens with aging (3); thus, older adults are has found ethnic differences in quantitative sensory
at greater risk of inadequate vitamin D levels. Addition- testing results, with blacks reporting increased pain
ally, vitamin D synthesis requires longer periods of sun sensitivity (28,29).
exposure for those with dark skin pigmentation. Thus, The triage theory, proposed by Ames (30), hy-
low levels of vitamin D are prevalent in black Ameri- pothesizes that long-term micronutrient deficiencies
cans, including the most severe levels of deficiency. trigger chronic inflammation. In turn, chronic inflamma-
Estimates indicate that 70% of white Americans and tion leads to chronic health conditions, many of which
95% of black Americans have insufficient levels of are characterized by pain as a disabling symptom. Re-
vitamin D (4). Greater vitamin D deficiency in black cent research by Lee et al (31) supports the hypothesis
Americans may, in part, explain their increased inci- that the etiology of osteoarthritis includes a systemic
dence of chronic health conditions and provide a key to inflammatory component. Heaney (32) theorizes that
reducing health disparities (35).
long-latency chronic diseases are related to insufficient
The optimal serum concentration of vitamin D
micronutrients over extended periods. The US nutri-
is currently debated but is believed to be between 30
tional recommendations for micronutrients are based on
and 60 ng/ml. Recent clinical practice guidelines define
preventing short-latency diseases and not on optimizing
vitamin D levels 20 ng/ml as deficient and levels of
the preventive health effects of micronutrient therapy.
2129 ng/ml as insufficient (6). It has been suggested that
These theories may help to explain relationships be-
different normative values of vitamin D may be war-
tween vitamin D level and chronic pain. The purpose of
ranted for black and white Americans based on the
this study was to examine whether variations in vitamin
inverse relationship between vitamin D and parathyroid
hormone levels (7). However, variation in the vitamin D levels contribute to race differences in symptomatic
Dparathyroid hormone relationship by race and age is knee osteoarthritis pain. We hypothesized that low levels
not fully understood, and clinical trials on vitamin D of vitamin D would contribute to self-reported and
supplementation are needed (710). In 2011, the Insti- experimental knee pain and that vitamin D level would
tute of Medicine report, Dietary Reference Intakes for mediate the relationship between race (referred to as
Calcium and Vitamin D (11), increased the recom- group differences) and knee pain.
mended daily intake of vitamin D from 400 to 600 IU per
day. Many experts believe that this dose is still too low PATIENTS AND METHODS
(12). The Institute of Medicine report recommended
Study design. This cross-sectional study design exam-
interventional research on vitamin D supplementation ined the relationship between low levels of vitamin D and
to ascertain its effect on disease, aging, and racial health symptomatic knee osteoarthritis pain among older adults.
disparities and therefore raised the maximum safe daily The project is a substudy of an ongoing study examining
dose to 4,000 IU. Although vitamin D has been corre- ethnic differences in knee osteoarthritis pain. The measures
lated with many health conditions, few studies have and procedures described are limited to those involved in the
current study. Participants were recruited at the University of
considered the relationship of vitamin D to chronic pain Florida between January 2010 and August 2011.
(1317). Participants. Participants were recruited via posted
Chronic pain is a disease. The 2011 Institute of fliers, radio and print media advertisements, clinic recruitment,
Medicine report, Relieving Pain in America (18), estimates and word-of-mouth referral. The inclusion criteria were as
annual spending on pain to be between $560 billion and follows: 1) ages 4585 years; 2) unilateral or bilateral symp-
tomatic knee osteoarthritis based on American College of
$635 billion. Many older adults contend with pain from Rheumatology criteria (33), regardless of radiographic evi-
osteoarthritisthe most common joint condition and dence of osteoarthritis; and 3) availability to complete the
the leading cause of disability in older adults (1921). 2-session protocol. Participants were excluded if they met any
3928 GLOVER ET AL

of the following criteria: 1) prosthetic knee replacement or tion and/or habituation of cutaneous receptors. The results of
nonarthroscopic surgery to the affected knee; 2) uncontrolled the 3 individual trials were averaged together to create overall
hypertension (150/95), heart failure, or history of acute heat pain threshold and heat pain tolerance scores for the
myocardial infarction; 3) peripheral neuropathy; 4) systemic index knee. Similarly, 3 trials for threshold and 3 for tolerance
rheumatic disorders including rheumatoid arthritis, systemic were delivered to the ipsilateral forearm. The position of the
lupus erythematosus, and fibromyalgia; 5) daily opioid use; thermode was moved among 3 sites an inch above the ventral
6) cognitive impairment (Mini-Mental State Examination wrist and an inch below the antecubital space. The results of
[MMSE] [34] score 22); 7) excessive anxiety regarding the 3 individual trials were averaged together to create overall
protocol procedures (intravenous [IV] catheter insertion, heat pain threshold and heat pain tolerance scores for the
quantitative sensory testing procedures); and 8) hospitalization ipsilateral forearm.
within the preceding year for psychiatric illness. All procedures Mechanical testing procedures. To determine sensitivity
were reviewed and approved by the University of Florida at the site of clinical pain, 6 total trials of pressure pain
Institutional Review Board. Participants provided informed threshold were assessed at the medial (3 trials) and lateral
consent and were compensated for their participation. (3 trials) joint lines of the index knee. Additionally, 3 pres-
Procedures. Within 1 week prior to the health assess- sure pain threshold trials were assessed at the dorsal ipsi-
ment session, participants completed study questionnaires. At lateral forearm. A handheld Medoc digital pressure algometer
the health assessment session, the following measures were (Ramat Yishai) was used for the mechanical procedures. An
obtained: anthropometric data, vital signs, health history, application rate of 30 kPa per second was used. To assess
current medications, MMSE score, and a bilateral joint exam- pressure pain threshold, the examiner applied a constant rate
ination of the hand and knee joints. Using the American of pressure, and the participant was instructed to press a
College of Rheumatology criteria (33) for symptomatic knee button when the sensation first becomes painful, at which
osteoarthritis, a participants most symptomatic/painful knee time the device recorded the pressure in kPa. The average of
was designated the index knee. Within 4 weeks of the health the 3 trials was computed separately for the medial and lateral
assessment session, participants were scheduled for the quan- knee and subsequently combined to create an overall pressure
titative sensory testing session. Additional questionnaires were pain sensitivity score for the index knee. Likewise, the average
completed prior to and during this session. Quantitative sen- results of the 3 trials for the ipsilateral forearm were calculated
sory testing procedures included vital signs, IV insertion for to create an overall pressure pain sensitivity score. The overall
blood collection, and assessment of thermally induced and pain index scores for the index knee and ipsilateral forearm
mechanically induced pain. were included in statistical analysis.
Self-report measures. Center for Epidemiologic Studies Vitamin D assay. Serum was collected at the beginning
Depression Scale (CES-D). The CES-D is a 20-item self-report of the quantitative sensory testing session. Following collec-
tool that measures symptoms of depression including de- tion, plasma was stored in a freezer at 80C. The analyte was
pressed mood, guilt/worthlessness, helplessness/hopelessness, subjected to analysis within 6 months of collection. Vitamin D
psychomotor retardation, loss of appetite, and sleep distur- analysis was performed by high-performance liquid chroma-
bance (35). The total score of the CES-D (range 060) was tography (total 25-hydroxyvitamin D 25[OH]D2 plus
used in the current study as an estimate of the degree of 25[OH]D3). Results were shared with participants, and if their
participants depressive symptoms. The validity and internal vitamin D level was 30 ng/ml, they were encouraged to
consistency of the CES-D in the general population have been discuss this result with their primary care provider.
reported to be acceptable (36). Statistical analysis. Statistical analysis was performed
Western Ontario and McMaster Universities Osteoarthri- using SPSS software, version 19.0 (IBM). Bivariate relation-
tis Index (WOMAC). The WOMAC (37) is frequently used in ships among continuously measured variables were assessed
research to assess symptoms of knee and hip osteoarthritis. using Pearsons correlations, while sex differences were as-
The subscales measure pain, stiffness, and physical function. sessed using analysis of variance. Group differences by race
The total WOMAC score was used in analysis (range 096). were adjusted for covariates and assessed using analysis of
High construct validity and testretest reliability has been found covariance (ANCOVA). The relationships between vitamin D
in paper and computerized versions of the WOMAC (38). level and pain were examined using multiple regressions. P
Quantitative sensory testing. Thermal testing proce- values less than 0.05 were considered significant.
dures. All thermal stimuli were delivered using a computer- To test whether vitamin D level significantly medi-
controlled Medoc Pathway Thermal Sensory Analyzer (Ramat ated (Figure 1) the association between race and clinical and
Yishai). Heat pain thresholds and heat pain tolerances were experimental pain measures, we conducted a bootstrap analy-
assessed using an ascending method of limits. From a baseline sis. Bootstrapping, as put forth by Hayes and Preacher (39,40),
of 32C, thermode temperature increased at a rate of 0.5C per is a nonparametric resampling procedure that has been shown
second until the participant responded by pressing a button on to be a viable alternative to the Baron and Kenny (41)
a handheld device. For heat pain threshold, participants were approach to testing intervening variable effects. Percentile
instructed to press the button when the sensation first be- bootstrap confidence intervals (CIs) were used to minimize
comes painful, and for tolerance, the instruction was to press Type I error rate (42). A percentile bootstrap 95% CI was
the button when they no longer feel able to tolerate the pain. calculated, using the SPSS macro for simple mediation (40) to
Three trials for threshold and 3 for tolerance were delivered to determine the significance of the mediator. In the current
the index knee. The position of the thermode was moved study, path c represents the total effect of the independent
among 3 sites (the medial joint line, the patella, and the tibial variable (race) on the dependent variable (clinical and exper-
tuberosity distal to the joint) between trials to avoid sensitiza- imental pain measures). Path a denotes the effect of race on
VITAMIN D, RACE, AND PAIN 3929

Women demonstrated lower heat pain tolerance

at the forearm (F[1,92] 4.76, P 0.03) and lower
pressure pain thresholds at the knee (F[1,92] 12.92,
P 0.001) and forearm (F[1,92] 13.75, P 0.001)
compared to men. There was a greater proportion of
black current smokers compared to white current smok-
ers (2 4.64, P 0.05). However, clinical and exper-
imental pain, as well as vitamin D level, did not differ
significantly according to smoking status (all P 0.05 for
current smokers versus nonsmokers). Participant age,
sex, body mass index (BMI), and CES-D score corre-
lated with pain response and were included as statistical
controls in all subsequent data analyses.
Group differences in clinical and experimental
pain. Using the total WOMAC score, ANCOVA re-
vealed a significant group difference (F[1,88] 5.67,
Figure 1. Mediation model. The effect of race on vitamin D level is
represented by a, the direct effect of vitamin D level on pain measures
is represented by b, and the total effect of race on pain measures is
represented by c.

vitamin D level (mediator) and path b is the effect of vitamin

D level on clinical and experimental pain measures. The
bootstrapped mediation analysis indicates whether the total
effect (path c) consists of a significant direct effect (path c) of
race on clinical and experimental pain measures and a signif-
icant indirect effect (path a b) of race on clinical and
experimental pain measures through the mediatorvitamin D
level.

RESULTS
Sample characteristics and examination of co-
variates. A total of 107 participants with symptomatic
knee osteoarthritis were recruited. Vitamin D data were
available for 94 participants at the time of analyses.
Figure 2 shows a flow diagram for participant matricu-
lation through the study.
Pearsons correlations for key study variables are
shown in Table 1. Vitamin D level correlated signifi-
cantly and inversely with the total WOMAC score,
suggesting that lower vitamin D levels are associated
with greater knee osteoarthritis pain and dysfunction.
For experimental pain measures, vitamin D level corre-
lated significantly with heat pain threshold and heat pain
tolerance at the index knee and at the ipsilateral fore-
arm. Finally, vitamin D level correlated significantly with
pressure pain thresholds at the knee and at the ipsilat-
eral forearm. Age and CES-D score did not correlate
significantly with vitamin D level; however, the CES-D Figure 2. Flow diagram of the study participants. QST quantitative
score correlated positively with WOMAC score and heat sensory testing; HAS health assessment session; AA African
pain threshold at the knee. American.
3930 GLOVER ET AL

Table 1. Pearsons correlations of variables among the 94 study participants*

Knee Knee Forearm Forearm Knee Forearm
Age, Vitamin D CES-D WOMAC heat pain heat pain heat pain heat pain pressure pressure
years BMI level score score threshold tolerance threshold tolerance pain pain
Age, years
BMI 0.02
Vitamin D level 0.02 0.33
CES-D score 0.18 0.01 0.07
WOMAC score 0.03 0.50 0.27 0.28
Knee heat pain threshold 0.02 0.02 0.27 0.23 0.12
Knee heat pain tolerance 0.11 0.14 0.27 0.07 0.22 0.61
Forearm heat pain threshold 0.07 0.14 0.35 0.18 0.15 0.58 0.59
Forearm heat pain tolerance 0.16 0.21 0.30 0.08 0.20 0.47 0.88 0.64
Knee pressure pain 0.07 0.38 0.39 0.05 0.41 0.37 0.36 0.37 0.35
Forearm pressure pain 0.19 0.16 0.34 0.06 0.22 0.36 0.35 0.41 0.37 0.67

* BMI body mass index; CES-D Center for Epidemiologic Studies Depression Scale; WOMAC Western Ontario and McMaster Universities
Osteoarthritis Index.
P 0.01.
P 0.05.

P 0.02). As shown in Table 2, the mean SD pants demonstrated significantly lower mean pressure
WOMAC score for black participants was 41.5 21.66, pain threshold at the index knee (F[1,88] 10.13, P
compared to a mean SD WOMAC score of 29.4 0.002) and at the ipsilateral forearm (F[1,88] 3.96,
18.76 for white participants, indicating that blacks re- P 0.05) compared to whites.
ported more knee pain, stiffness, and limitations in Group differences in vitamin D level. Adjusted
physical function. Additionally, black participants dem- for covariates, ANCOVA revealed that black partici-
onstrated greater sensitivity to experimental pain mea- pants were characterized by significantly lower mean
sures compared to whites. For example, blacks demon- vitamin D levels than their white counterparts (F[1,88]
strated significantly lower mean heat pain threshold 16.03, P 0.001). The mean SD vitamin D level
(F[1,88] 5.36, P 0.02) and heat pain tolerance in black participants was 19.9 8.6 ng/ml, which based
(F[1,88] 23.14, P 0.001) at the index knee. Similar on the most recent guidelines (6) indicates vitamin D
results were found for the ipsilateral forearm, such that deficiency, compared to a mean SD vitamin D level of
blacks demonstrated significantly lower heat pain 28.2 8.6 ng/ml in whites, which indicates vitamin D
threshold (F[1,88] 9.32, P 0.003) and heat pain insufficiency. Low levels of vitamin D were observed
tolerance (F[1,88] 17.81, P 0.001) compared to across the sample but were found disproportionately in
whites. For mechanically induced pain, black partici- black participants; 38 of 45 black participants (84%) had

Table 2. Descriptive data for key study variables across the groups*
Blacks Whites
(n 45) (n49) P
No. of women/men 31/14 39/10
Age, years 54.6 5.4 (4568) 56.9 7.7 (4571)
Vitamin D level, ng/ml 19.9 8.6 (840) 28.2 8.6 (748) 0.001
BMI, kg/m2 32.4 8.0 (1850) 29.0 6.7 (1949) 0.05
CES-D score (range 060) 16.2 6.5 (030) 16.7 5.2 (038)
WOMAC score (range 096) 41.5 21.7 (494) 29.4 18.8 (486) 0.05
Knee heat pain threshold, C 41.3 3.0 (3646) 42.7 2.9 (3648)
Knee heat pain tolerance, C 44.6 3.3 (3549) 47.0 2.1 (3851) 0.001
Forearm heat pain threshold, C 40.9 3.3 (3447) 42.8 2.7 (3547) 0.05
Forearm heat pain tolerance, C 45.2 2.7 (3750) 47.0 2.1 (3951) 0.001
Knee pressure pain threshold, kPa 263.6 149.0 (53588) 362.4 164.4 (81598) 0.05
Forearm pressure pain threshold, kPa 233.3 153.2 (51748) 275.9 148.9 (77677) 0.05

* Except where indicated otherwise, values are the mean SD (range). BMI body mass index;
CES-D Center for Epidemiologic Studies Depression Scale; WOMAC Western Ontario and
McMaster Universities Osteoarthritis Index.
VITAMIN D, RACE, AND PAIN 3931

vitamin D levels 30 ng/ml compared to 25 of 49 white the results of the analyses that examined whether vita-
participants (51%) (2 11.86, P 0.001). min D level mediated group differences in heat pain
Vitamin D level and pain. Adjusted multiple thresholds at the index knee (model 1) and ipsilateral
regression analyses, controlling for the effects of age, forearm (model 2) after adjusting for covariates. The
sex, BMI, CES-D score, and group, revealed that vita- vitamin D level was found to significantly mediate the
min D level was not significantly associated with the total group difference in heat pain threshold at the index knee
WOMAC score ( coefficient 0.06, P 0.56). How- (percentile bootstrap 95% CI 0.011.11) and at the
ever, low levels of vitamin D predicted increased pain ipsilateral forearm (percentile bootstrap 95% CI 0.09
sensitivity and lower experimental pain thresholds. Di- 1.22) using the percentile bootstrap 95% CI, with 5,000
minished vitamin D levels were significantly associated resamples. These results indicate that indirect effects
with lower heat pain thresholds at the index knee through vitamin D level are significantly different from
( coefficient 0.23, P 0.05) and ipsilateral forearm zero. Thus, group differences in heat pain thresholds are
( coefficient 0.25, P 0.03). Additionally, lower vita- mediated by group differences in vitamin D levels. More
min D levels were significantly associated with greater specifically, blacks possessed lower vitamin D levels than
sensitivity to pressure pain (i.e., lower threshold) at the whites and, in turn, lower vitamin D levels significantly
index knee ( coefficient 0.23, P 0.02) and the predicted lower heat pain thresholds at the index knee
ipsilateral forearm ( coefficient 0.31, P 0.01). How- and ipsilateral forearm.
ever, the vitamin D level was not significantly related to Table 4 displays the results of the analyses that
heat pain tolerance at the index knee ( coefficient 0.10, examined whether the vitamin D level also mediated
P 0.37) or the ipsilateral forearm ( coefficient 0.13, group differences in pressure pain thresholds at the
P 0.22). Accordingly, only heat pain thresholds and index knee (model 3) and ipsilateral forearm (model 4)
pressure pain thresholds at the index knee and ipsilateral after adjusting for covariates. Vitamin D level was shown
forearm were examined in the mediational analyses. to significantly mediate the group differences in pressure
Testing vitamin D level as a simple mediator. We pain threshold at the index knee (percentile bootstrap
tested the indirect effects of group differences in heat 95% CI 3.4360.86) and at the ipsilateral forearm
pain thresholds and pressure pain thresholds through (percentile bootstrap 95% CI 12.4467.55) using the
vitamin D level (i.e., simple mediation). Table 3 displays percentile bootstrap 95% CI, with 5,000 resamples. As

Table 3. Vitamin D mediation of group differences in heat pain threshold at the index knee (model 1)
and ipsilateral forearm (model 2)*
Percentile
Effect Coefficient SE t P bootstrap 95% CI
Model 1
Path c 1.49 0.64 2.32 0.0229
Path a 7.39 1.85 4.00 0.0001
Path b 0.07 0.04 1.96 0.0531
Path c 0.96 0.69 1.39 0.1667
Path a b 0.51 0.28 0.011.11
Model 2
Path c 1.97 0.64 3.05 0.0030
Path a 7.39 1.85 4.00 0.0001
Path b 0.08 0.04 2.27 0.0258
Path c 1.36 0.68 1.98 0.0506
Path a b 0.60 0.29 0.091.22

* Shown are unstandardized coefficients for the mediated effect of group differences in heat pain
threshold at the index knee and ipsilateral forearm through vitamin D level, controlling for age, sex, body
mass index, and Center for Epidemiologic Studies Depression Scale score. Path c total effect of race on
heat pain threshold; path a effect of race on vitamin D level; path b direct effect of vitamin D level
on heat pain threshold; path c direct effect of race on heat pain threshold, controlling for vitamin D
level; path a b indirect effect of race on heat pain threshold through vitamin D level. 95% CI 95%
confidence interval.
A P value for the indirect effect is not provided because such a P value is contingent upon a normal
distribution of the indirect effect. Given that the product of the path a and path b coefficients is always
positively skewed, interpretation of this P value can be misleading (40).
3932 GLOVER ET AL

Table 4. Vitamin D mediation of group differences in pressure pain threshold at the index knee
(model 3) and ipsilateral forearm (model 4)*
Percentile bootstrap
Effect Coefficient SE t P 95% CI
Model 3
Path c 95.88 30.12 3.18 0.0020
Path a 7.39 1.85 4.00 0.0001
Path b 4.00 1.70 2.36 0.0205
Path c 66.31 31.93 2.08 0.0408
Path a b 29.10 14.84 3.4360.86
Model 4
Path c 60.21 30.26 1.99 0.0497
Path a 7.39 1.85 4.00 0.0001
Path b 4.95 1.68 2.96 0.0040
Path c 23.61 31.54 0.75 0.4561
Path a b 36.34 14.27 12.4467.55

* Shown are unstandardized coefficients for the mediated effect of group differences in pressure pain
threshold at the index knee and ipsilateral forearm through vitamin D level, controlling for age, sex, body
mass index, and Center for Epidemiologic Studies Depression Scale score. Path c total effect of race on
pressure pain threshold; path a effect of race on vitamin D level; path b direct effect of vitamin D level
on pressure pain threshold; path c direct effect of race on pressure pain threshold, controlling for
vitamin D level; path a b indirect effect of race on pressure pain threshold through vitamin D level.
95% CI 95% confidence interval.
A P value for the indirect effect is not provided because such a P value is contingent upon a normal
distribution of the indirect effect. Given that the product of the path a and path b coefficients is always
positively skewed, interpretation of this P value can be misleading (40).

mentioned above, these results indicate that indirect ences in responses to quantitative sensory testing in
effects through the vitamin D level are significantly black Americans (28,29,43).
different from zero. Thus, group differences in pressure Although there is an established link between low
pain thresholds are mediated by group differences in levels of vitamin D and pain due to osteomalacia (1), no
vitamin D levels. clear biologic or psychological mechanisms explain how
low levels of vitamin D may affect pain sensitivity or
DISCUSSION relate causally to other chronic pain conditions. Inter-
estingly, significant mediation was found for heat pain
In the present study, we examined the extent to threshold but not for pain tolerance. This outcome
which the vitamin D level mediated the relationship
suggests that vitamin D level may be related to pain
between race and pain in older adults with symptomatic
pathways involved in initial perception of pain but not to
knee osteoarthritis. The results revealed that group
how much pain an individual can tolerate. Pain tolerance
differences in response to experimental pain, but not
is dependent, in part, on individuals willingness to
self-reported pain, for blacks and whites are mediated
by group differences in vitamin D level. Low levels of endure noxious stimulation, which is not necessarily
vitamin D mediated the relationship between race and related to vitamin D level.
experimental pain in older adults with symptomatic knee In contrast, since there are vitamin D receptors in
osteoarthritis. The study hypothesis expected these find- nucleated cells throughout the peripheral and central
ings at the painful knee. Similar results were demon- nervous system, less than sufficient levels of vitamin D
strated at a nonpainful testing site at the forearm. Even could affect both the transmission and modulation of
in a sunny southern environment, low levels of vitamin D painful stimuli (16,21,31,44). Using a rodent model,
were endemic across the sample, with black participants Tague and colleagues (45) found that vitamin D
having more pronounced vitamin D deficiency. Black deficient rats had increased muscle innervation by noci-
participants demonstrated greater pain sensitivity in ceptors, leading to reduced pain threshold for mechan-
thermal and mechanical testing at the index knee and ical stimulation in hindlimb musculature. In vitro
the ipsilateral forearm. These findings are consistent cultures revealed that vitamin D level was inversely
with previous research demonstrating significant differ- correlated with growth of sensory neurons, leading the
VITAMIN D, RACE, AND PAIN 3933

authors to hypothesize that vitamin D deficiency may tal pain. Finally, although participants using regularly
induce muscle pain by stimulating nociceptor growth. scheduled opioids were excluded from the study, and
Furthermore, inflammatory processes contribute to in- those using opioids as needed were asked to refrain from
creased pain sensitivity among individuals with osteo- taking their medication 2 days prior to quantitative
arthritis (21,31). Sufficient levels of cellular vitamin D sensory testing, we did not account for other analgesic
have a protective effect on cell function and are be- medication use, which may have affected the results of
lieved to reduce inflammation (1,46). Thus, alterations pain testing.
in inflammation attributable to low levels of vitamin D It is unclear why race differences in vitamin D
may precipitate increased pain among individuals with level mediate group differences in experimental pain
osteoarthritis. Less than sufficient levels of vitamin D outcomes but do not mediate clinical osteoarthritis pain
may explain the increase in symptomatic osteoarthritis intensity, stiffness, or physical function on the WOMAC.
without a concurrent increase in radiographic osteo- This outcome may be related to the fact that osteoar-
arthritis (47). thritis pain intensity measures were assessed retrospec-
In this cohort, the participants CES-D score did tively (i.e., pain over the past 48 hours) and were
not correlate significantly with vitamin D level. How- therefore subject to recall bias. Alternatively, vitamin D
ever, other research has found an association between may influence certain aspects of pain processing re-
low levels of vitamin D and increased symptoms of flected by pain thresholds, while osteoarthritis-related
depression. In a 4-year study of nearly 12,600 partici- symptoms assessed by the WOMAC are likely driven by
pants, low levels of vitamin D were associated with multiple factors over and above nociceptive processes.
depression, especially in those who had had previous Moreover, a clinical threshold of vitamin D insufficiency
episodes (48). Higher levels of vitamin D were associ- may be necessary to better understand the relationship
ated with fewer symptoms of depression, even among between low levels of vitamin D, clinical knee pain, and
those with a history of depression. It is possible that total WOMAC score. In a systematic review of research
vitamin D has a direct effect on mood, since vitamin D on vitamin D and chronic pain, Straube and colleagues
supplementation in a placebo-controlled double-blind (16) did not find evidence to support a relationship
trial was shown to improve the mood and affect of in- between vitamin D and chronic pain. Results may have
dividuals diagnosed as having seasonal affective disorder been affected by methodologic considerations, such as
(49). Vitamin D receptors are ample in structures and study design weakness and limited sample size. Similarly,
cells of the brain and may contribute to overall brain in a review of 7 studies, Straube and colleagues (14)
health and enhanced nerve conduction (50). Given the found insufficient evidence to support treating chronic
substantial overlap between negative mood and chronic pain in ethnic minority patients by correcting vitamin D
pain, vitamin D insufficiency may modulate pain percep- deficiency. However, none of the studies were random-
tion through affective pathways (51). ized controlled trials, and only 2 case studies investi-
As in most research, these findings should be gated vitamin D supplementation.
interpreted in light of the study limitations. Participants Additional research is needed to strengthen these
represented a convenience sample of community- findings. High-quality observational studies and ran-
dwelling adults and older adults. The age range of par- domized controlled trials with rigorous methodologic
ticipants was 4571 years; however, the average age of control and adequate numbers of black participants are
55.8 years represents a younger cohort. Although black needed to understand the relationship between low
participants were slightly younger than whites, they levels of vitamin D, pain, and pain disparities experi-
reported higher levels of osteoarthritis-related pain. The enced by black Americans. Improving vitamin D status is
cross-sectional examination of vitamin D level and clin- inexpensive and with low risk of adverse events. Thus, if
ical and experimental pain does not permit a full under- additional research demonstrates that improving vitamin
standing of the direction of the relationship between D status lessens knee osteoarthritis pain, identifying and
vitamin D level and chronic pain. It is possible that treating vitamin D insufficiency and deficiency may
people with osteoarthritis pain spend less time outdoors improve function for older adults with osteoarthritis and
and thus have reduced opportunities for vitamin D reduce health disparities for black Americans.
synthesis. However, it is unlikely that experimental pain
affected participants vitamin D levels, thus lending ACKNOWLEDGMENTS
credence to our study model, namely, that vitamin D We would like to thank the participants and staff of the
mediates the relationship between race and experimen- Understanding Pain and Limitations in Osteoarthritic Disease
3934 GLOVER ET AL

study and the staff of the University of Florida Clinical 17. McAlindon TE, Felson DT, Zhang Y, Hannan MT, Aliabadi P,
Research Center, without whom this research would not have Weissman B, et al. Relation of dietary intake and serum levels of
been possible. vitamin D to progression of osteoarthritis of the knee among
participants in the Framingham Study. Ann Intern Med 1996;125:
3539.
AUTHOR CONTRIBUTIONS 18. Institute of Medicine. Relieving pain in America: a blueprint for
transforming prevention, care, education, and research. Washing-
All authors were involved in drafting the article or revising it ton (DC): National Academies Press; 2011.
critically for important intellectual content, and all authors approved 19. Wieland HA, Michaelis M, Kirschbaum BJ, Rudolphi KA. Osteo-
the final version to be published. Ms Glover had full access to all of the arthritisan untreatable disease? [published erratum appears in
data in the study and takes responsibility for the integrity of the data Nat Rev Drug Discov 2005;4:543]. Nat Rev Drug Discov 2005;4:
and the accuracy of the data analysis. 33144.
Study conception and design. Glover, Horgas, Riley, Staud, Bradley, 20. Centers for Disease Control and Prevention. Racial/ethnic differ-
Fillingim. ences in the prevalence and impact of doctor-diagnosed arthritis
Acquisition of data. Glover, Goodin, Kindler, King, Sibille.
United States, 2002. MMWR Morb Mortal Wkly Rep 2005;54:
Analysis and interpretation of data. Glover, Goodin, Peloquin,
11923.
Fillingim.
21. Hunter DJ, McDougall JJ, Keefe FJ. The symptoms of osteo-
arthritis and the genesis of pain. Rheum Dis Clin North Am
2008;34:62343.
REFERENCES 22. Apkarian AV, Baliki MN, Geha PY. Towards a theory of chronic
1. Holick MF. Vitamin D deficiency. N Engl J Med 2007;357:26681. pain. Prog Neurobiol 2009;87:8197.
2. Mason RS, Sequeira VB, Gordon-Thomson C. Vitamin D: the 23. Anderson KO, Green CR, Payne R. Racial and ethnic disparities
light side of sunshine. Eur J Clin Nutr 2011;65:98693. in pain: causes and consequences of unequal care. J Pain 2009;10:
3. Holick MF, Chen TC. Vitamin D deficiency: a worldwide problem 1187204.
with health consequences. Am J Clin Nutr 2008;87:1080S6S. 24. Hadjistavropoulos T, Herr K, Turk DC, Fine PG, Dworkin RH,
4. Ginde AA, Liu MC, Camargo CA Jr. Demographic differences Helme R, et al. An interdisciplinary expert consensus statement on
and trends of vitamin D insufficiency in the US population, assessment of pain in older persons. Clin J Pain 2007;23:S143.
1988-2004. Arch Intern Med 2009;169:62632. 25. Green CR, Anderson KO, Baker TA, Campbell LC, Decker S,
5. Grant WB, Peiris AN. Possible role of serum 25-hydroxyvitamin D Fillingim RB, et al. The unequal burden of pain: confronting racial
in black-white health disparities in the United States. J Am Med and ethnic disparities in pain. Pain Med 2003;4:27794.
Dir Assoc 2010;11:61728. 26. Edwards CL, Fillingim RB, Keefe F. Race, ethnicity and pain.
6. Holick MF, Binkley NC, Bischoff-Ferrari HA, Gordon CM, Pain 2001;94:1337.
Hanley DA, Heaney RP, et al. Evaluation, treatment, and preven- 27. Green CR, Hart-Johnson T. The impact of chronic pain on the
tion of vitamin D deficiency: an Endocrine Society clinical practice health of black and white men. J Natl Med Assoc 2010;102:32131.
guideline. J Clin Endocrinol Metab 2011;96:191130. 28. Rahim-Williams FB, Riley JL III, Herrera D, Campbell CM,
7. Wright NC, Chen L, Niu J, Neogi T, Javiad K, Nevitt MA, et al. Hastie BA, Fillingim RB. Ethnic identity predicts experimental
Defining physiologically normal vitamin D in African Ameri- pain sensitivity in African Americans and Hispanics. Pain 2007;
cans. Osteoporos Int 2012;23:228391. 129:17784.
8. Dawson-Hughes B. Racial/ethnic considerations in making recom- 29. Campbell CM, Edwards RR, Fillingim RB. Ethnic differences in
mendations for vitamin D for adult and elderly men and women. responses to multiple experimental pain stimuli. Pain 2005;113:
Am J Clin Nutr 2004;80:1763S6S. 206.
9. Wilkins CH, Birge SJ, Sheline YI, Morris JC. Vitamin D defi- 30. Ames BN. Low micronutrient intake may accelerate the degener-
ciency is associated with worse cognitive performance and lower ative diseases of aging through allocation of scarce micronutrients
bone density in older African Americans. J Natl Med Assoc by triage. Proc Natl Acad Sci U S A 2006;103:1758994.
2009;101:34954. 31. Lee YC, Lu B, Bathon JM, Haythornthwaite JA, Smith MT, Page
10. Harris SS, Soteriades E, Coolidge JA, Mudgal S, Dawson-Hughes GG, et al. Pain sensitivity and pain reactivity in osteoarthritis.
B. Vitamin D insufficiency and hyperparathyroidism in a low Arthritis Care Res (Hoboken) 2011;63:3207.
income, multiracial, elderly population. J Clin Endocrinol Metab 32. Heaney RP. Long-latency deficiency disease: insights from calcium
2000;85:412530. and vitamin D. Am J Clin Nutr 2003;78:9129.
11. Institute of Medicine. Dietary reference intakes for calcium and 33. Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K,
vitamin D. Washington (DC): National Academies Press; 2011. et al. Development of criteria for the classification and reporting
12. Heaney RP, Holick MF. Why the IOM recommendations for of osteoarthritis: classification of osteoarthritis of the knee. Ar-
vitamin D are deficient. J Bone Miner Res 2011;26:4557. thritis Rheum 1986;29:103949.
13. Chaganti RK, Parimi N, Cawthon P, Dam TL, Nevitt MC, Lane 34. Folstein MF, Folstein SE, McHugh PR. Mini-mental state: a
NE. Association of 25-hydroxyvitamin D with prevalent osteo- practical method for grading the cognitive state of patients for the
arthritis of the hip in elderly men: the Osteoporotic Fractures in clinician. J Psychiatr Res 1975;12:18998.
Men Study. Arthritis Rheum 2010;62:5114. 35. Radloff LS. The CES-D Scale. APM 1977;1:385401.
14. Straube S, Moore RA, Derry S, Hallier E, McQuay HJ. Vitamin D 36. Beekman AT, Deeg DJ, Van Limbeek J, Braam AW, De Vries
and chronic pain in immigrant and ethnic minority patients MZ, Van Tilburg W. Criterion validity of the Center for Epi-
investigation of the relationship and comparison with native demiologic Studies Depression scale (CES-D): results from a
Western populations. Int J Endocrinol 2010;2010:753075. community-based sample of older subjects in The Netherlands.
15. Atherton K, Berry DJ, Parsons T, Macfarlane GJ, Power C, Psychol Med 1997;27:2315.
Hypponen E. Vitamin D and chronic widespread pain in a white 37. Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW.
middle-aged British population: evidence from a cross-sectional Validation study of WOMAC: a health status instrument for
population survey. Ann Rheum Dis 2009;68:81722. measuring clinically important patient relevant outcomes to anti-
16. Straube S, Moore RA, Derry S, McQuay HJ. Vitamin D and rheumatic drug therapy in patients with osteoarthritis of the hip or
chronic pain. Pain 2009;141:103. knee. J Rheumatol 1988;15:183340.
VITAMIN D, RACE, AND PAIN 3935

38. Theiler R, Spielberger J, Bischoff HA, Bellamy N, Huber J, hypersensitivity and sensory hyperinnervation. J Neurosci 2011;
Kroesen S. Clinical evaluation of the WOMAC 3.0 OA Index in 31:1372838.
numeric rating scale format using a computerized touch screen 46. McCann JC, Ames BN. Is there convincing biological or behav-
version. Osteoarthritis Cartilage 2002;10:47981. ioral evidence linking vitamin D deficiency to brain dysfunction?
39. Hayes AF. Beyond Baron and Kenny: statistical mediation analysis FASEB J 2008;22:9821001.
in the new millennium. Commun Monogr 2009;76:40820. 47. Nguyen US, Zhang Y, Zhu Y, Niu J, Zhang B, Felson DT.
40. Preacher KJ, Hayes AF. Asymptotic and resampling strategies for Increasing prevalence of knee pain and symptomatic knee osteo-
assessing and comparing indirect effects in multiple mediator arthritis: survey and cohort data. Ann Intern Med 2011;155:
models. Behav Res Methods 2008;40:87991. 72532.
41. Baron RM, Kenny DA. The moderator-mediator variable distinc- 48. Hoang MT, Defina LF, Willis BL, Leonard DS, Weiner MF,
tion in social psychological research: conceptual, strategic, and Brown ES. Association between low serum 25-hydroxyvitamin D
statistical considerations. J Pers Soc Psychol 1986;51:117382. and depression in a large sample of healthy adults: the Cooper
42. Fritz MS, Taylor AB, MacKinnon DP. Explanation of two anom- Center Longitudinal Study. Mayo Clin Proc 2011;86:10505.
alous results in statistical mediation analysis. Multivar Behav Res 49. Lansdowne AT, Provost SC. Vitamin D3 enhances mood in
2012;47:6187. healthy subjects during winter. Psychopharmacology (Berl) 1998;
43. Hastie BA, Riley JL, Fillingim RB. Ethnic differences and re- 135:31923.
sponses to pain in healthy young adults. Pain Med 2005;6:6171. 50. Buell JS, Dawson-Hughes B. Vitamin D and neurocognitive
44. Roesel TR. Does the central nervous system play a role in vitamin dysfunction: preventing Decline? Mol Aspects Med 2008;29:
D deficiency-related chronic pain? Pain 2009;143:15960. 41522.
45. Tague SE, Clarke GL, Winter MK, McCarson KE, Wright DE, 51. Melzack R. From the gate to the neuromatrix. Pain 1999;Suppl
Smith PG. Vitamin D deficiency promotes skeletal muscle 6:S1216.