Effect of Vitamin D Supplementation on Tibial Cartilage Volume and Knee Pain Among Patients With Symptomatic Knee Osteoarthritis: A Randomized Clinical Trial

Importance  Observational studies suggest that vitamin D supplementation is associated with benefits for knee osteoarthritis, but current trial evidence is contradictory.

Objective  To compare the effects of vitamin D supplementation vs placebo on knee pain and knee cartilage volume in patients with symptomatic knee osteoarthritis and low vitamin D levels.

Design, Setting, and Participants  A multicenter randomized, double-blind, placebo-controlled clinical trial in Tasmania and Victoria, Australia. Participants with symptomatic knee osteoarthritis and low 25-hydroxyvitamin D (12.5-60 nmol/L) were enrolled from June 2010 to December 2011. The trial was completed in December 2013.

Interventions  Participants were randomly assigned to receive monthly treatment with oral vitamin D₃ (50 000 IU; n = 209) or an identical placebo (n = 204) for 2 years.

Main Outcomes and Measures  Primary outcomes were change in tibial cartilage volume (assessed using magnetic resonance imaging [MRI]) and change in the Western Ontario and McMaster Universities Arthritis Index (WOMAC) pain score (0 [no pain] to 500 [worst pain]) from baseline to month 24. Secondary outcomes were cartilage defects and bone marrow lesions (assessed using MRI).

Results  Of 413 enrolled participants (mean age, 63.2 years; 50% women), 340 (82.3%) completed the study. The level of 25-hydroxyvitamin D increased more in the vitamin D group (40.6 nmol/L) than in the placebo group (6.7 nmol/L) (P < .001) over 2 years. There were no significant differences in annual change of tibial cartilage volume (−3.4% in the vitamin D group vs −4.2% in the placebo group [between-group difference, 0.8% {95% CI, −0.2% to 1.8%}]; P = .13) or WOMAC pain score (−49.9 in the vitamin D group vs −35.1 in the placebo group [between-group difference, −14.8 {95% CI, −32.5 to 2.9}]; P = .10). There were no significant differences in change of tibiofemoral cartilage defects (0.3 in the vitamin D group vs 0.5 in the placebo group [between-group difference, −0.2 {95% CI, −0.4 to 0.1}]; P = .21) or change in tibiofemoral bone marrow lesions (−0.1 in the vitamin D group vs 0.3 in the placebo group [between-group difference, −0.5 {95% CI, −0.9 to 0.0}]; P = .06). Adverse events (≥1 per patient) occurred in 56 participants in the vitamin D group and in 37 participants in the placebo group (P = .04).

Conclusions and Relevance  Among patients with symptomatic knee osteoarthritis and low serum 25-hydroxyvitamin D levels, vitamin D supplementation, compared with placebo, did not result in significant differences in change in MRI-measured tibial cartilage volume or WOMAC knee pain score over 2 years. These findings do not support the use of vitamin D supplementation for preventing tibial cartilage loss or improving WOMAC knee pain in patients with knee osteoarthritis.

Trial Registration  clinicaltrials.gov Identifier: NCT01176344; anzctr.org.au Identifier: ACTRN12610000495022

Symptomatic knee osteoarthritis occurs among 10% of men and 13% of women aged 60 years or older.¹ Worldwide, knee osteoarthritis, together with hip osteoarthritis, are the 11th leading cause of global disability, accounting for 2.2% of total years lived with disability.² Medical costs of osteoarthritis account for 1% to 2.5% of the gross domestic product in developed countries.³ Currently there are no disease-modifying therapies for osteoarthritis; therefore, there is a need to develop cost-effective approaches to prevent the development and progression of osteoarthritis.

Vitamin D can reduce bone turnover and cartilage degradation, thus potentially preventing the development and progression of knee osteoarthritis.^4,5 Epidemiological studies showed that low serum 25-hydroxyvitamin D levels were associated with greater knee pain,^6,7 a higher prevalence of radiographic knee osteoarthritis,⁸ and higher risk of progression.^9,10 However, observational studies are subject to inherent bias and confounding factors such as physical activity and sun exposure.¹¹ In addition, 2 small existing randomized controlled trials (RCTs) reported contradictory results.^5,12 The inconsistencies are likely because of variations in inclusion criteria, outcome measures, follow-up time, and sample size.¹³ An updated systematic review called for further larger well-designed RCTs to determine whether vitamin D supplementation can slow disease progression.¹⁴ Therefore, we conducted an RCT of participants with clinically relevant inclusion criteria,¹³ to evaluate the effects of 2 years of vitamin D supplementation vs placebo on knee pain and knee cartilage volume in patients with symptomatic knee osteoarthritis combined with low 25-hydroxyvitamin D levels. Effects on other knee structural abnormalities, including cartilage defects and bone marrow lesions, were also assessed.

The Vitamin D Effect on Osteoarthritis (VIDEO) study was a randomized, double-blind, placebo-controlled trial, which was conducted between June 2010 and December 2013.¹⁵ Participants were recruited from June 2010 to December 2011 in Tasmania and Victoria, Australia, through advertisements in local media and community groups, as well as referrals from general practitioners, rheumatologists, and orthopedic surgeons. A telephone prescreening was conducted to inquire about knee pain status, comorbidities, participation in other studies, and whether the survey recipient anticipated knee or hip surgery within next 2 years. Potentially eligible participants were subsequently assessed during a clinic visit that included a physical examination, knee radiography, and assessment of serum 25-hydroxyvitamin D levels. The trial protocol appears in Supplement 1.

Inclusion and exclusion criteria were described in the published protocol.¹⁵ In brief, eligible participants were aged 50 to 79 years, had symptomatic knee osteoarthritis (assessed according to American College of Rheumatology [ACR] criteria¹⁶) for at least 6 months, and had pain of 20 to 80 mm on a 100-mm visual analog scale. In addition, participants had an ACR function class rating of I, II, or III (I [complete ability to perform usual activities of daily living] to III [ability to perform usual self-care activities but limited in vocational and avocational activities]) and relatively good health score of 0 to 2 on a 5-point Likert scale (0 [very good health] to 4 [very poor health]) according to the investigator assessment of disease status. Participants were included if their serum 25-hydroxyvitamin D levels were between 12.5 nmol/L and 60 nmol/L. Ethics approval was received from the Tasmania Health and Human Medical Research Ethics Committee (reference number H1040) and Monash University Human Research Ethics Committee (reference number CF10/1182-2010000616). Informed written consent was obtained from all participants.

Exclusion criteria included grade 3 radiographic changes according to the Altman and Gold atlas,¹⁷ severe knee pain on standing (more than 80 mm on a 100-mm visual analog scale), contraindication to magnetic resonance imaging (MRI), rheumatoid or psoriatic arthritis, lupus, cancer, severe cardiac or renal impairment, hypersensitivity to vitamin D, conditions affecting oral drug absorption, anticipated knee or hip surgery within the next 2 years, history of significant trauma of knees (eg, arthroscopy or injury to ligaments or menisci <1 year preceding the study), and history of taking vitamin D or an investigational drug within the last 30 days.

The knee that met the previously described inclusion and exclusion criteria was selected as the study knee for outcome measures. When both knees met the criteria, the study knee was defined as the one with worse pain assessed using the visual analog scale.

Participants were allocated to either the vitamin D or placebo group at a ratio of 1:1 based on computer-generated random numbers. Allocation concealment was confirmed by a central automated allocation procedure that was independent of the investigators. Treatment assignment was masked from all participants, research coordinators, and investigators and maintained until all data were collected, confirmed for accuracy, and cleaned, and statistical analyses were performed.

Participants in the treatment group were given a monthly capsule of 50 000 IU (1.25 mg) of vitamin D₃ (cholecalciferol) for 24 months (Nationwide Compounding Pharmacy).¹⁸ Participants in the control group received an identical inert placebo provided by the same company.

Knee pain was added as an additional primary end point to the protocol on June 6, 2012, after comments were received from reviewers of the methods article.¹⁵ The secondary outcomes included on clinicaltrials.gov slightly differ from the published protocol because several outcome measures were added in the substudies as secondary outcomes on clinicaltrial.gov, whereas the osteoarthritis outcomes were the focus in the published protocol.¹⁵

In this article, only the primary outcomes and a subset of secondary outcomes (as listed in the study protocol [Supplement 1]) are reported. Primary outcome measures were change in knee pain assessed using the Western Ontario and McMaster University Index of osteoarthritis (WOMAC) score¹⁹ and change in tibial cartilage volume on MRI from baseline to month 24. There were 5 prespecified secondary outcomes (cartilage defects; tibial plateau bone area; subchondral bone marrow lesion; meniscal tear and extrusion; and lower limb muscle strength), but only the outcomes for cartilage defects and bone marrow lesions on MRI are reported in this article. Post hoc analysis outcomes include 20% and 50% improvement rates in WOMAC pain score, WOMAC function and stiffness scores, visual analog scale knee pain, and the responder criteria developed by the Outcome Measures in Rheumatology Arthritis Clinical Trials-Osteoarthritis Research Society International (OMERACT-OARSI) criteria.

Knee pain was assessed at baseline and at months 3, 6, 12, and 24. Five items of WOMAC pain scale in 100-mm visual analog format were used to assess pain during walking, using stairs, in bed, sitting or lying, and standing. Items were summed to create a total pain score (range, 0-500).²⁰ Knee pain in most days of the previous month was assessed using a 100-mm visual analog scale.

The total WOMAC score indicates the sum of subscale scores including pain, stiffness, and physical function. Missing responses were managed according to the WOMAC user guide.²¹ The WOMAC pain score was considered void if more than 1 item was missing. In the event of a missing item, the remaining 4 items were averaged and then multiplied by 5.

The OARSI Standing Committee for Clinical Trials Response Criteria Initiative developed a set of responder criteria (OMERACT-OARSI) to categorize individual response to treatment as a single variable for clinical trials.²² Response using the exact OMERACT-OARSI criteria could not be directly evaluated because patient global assessment was not recorded in this trial; therefore, we used a modified OMERACT-OARSI responder definition without patient global assessment. OMERACT-OARSI responders in this study were defined as participants with (1) at least 50% improvement and an absolute change of at least 20 points in the mean WOMAC pain score or mean WOMAC function score; or (2) at least 20% improvement and an absolute change of at least 10 points in both the mean WOMAC pain score and the mean WOMAC function score.

MRI scans of the study knee were obtained according to a standardized protocol using a 1.5 T whole-body MRI unit with a commercial transmit-receive extremity coil. The sequences used for cartilage volume assessment were sagittal fat saturated (FS) T1-weighted spoiled gradient echo (GRE). Cartilage defects and bone marrow lesions were assessed using T2-weighted/proton density–weighted fast spin echo (FSE) sequences. MRIs were assessed by trained readers blinded to treatment allocation according to methods described previously.¹⁵

Cartilage volume was determined using the previously described image processing techniques.⁸ The volumes of individual cartilage plates (medial tibial and lateral tibial) were isolated by manually drawing disarticulation contours around the cartilage boundaries on a section-by-section basis then resampled by means of bilinear and cubic interpolation for final 3-dimensional rendering using OsiriX Lite imaging software (32-bit version 5.9, Pixmeo SARL). The coefficient of variation was 2.1% for medial tibia and 2.2% for lateral tibia.²³

Cartilage defects (0-4) were graded on T2-weighted images using a modified Outerbridge classification²⁴ at medial tibial, medial femoral, lateral tibial, and lateral femoral sites (described in the protocol).¹⁵ A total score was calculated as the total of subregional scores. Intraobserver reliability expressed as an intraclass correlation coefficient ranged from 0.77 to 0.94.

Bone marrow lesions, defined as discrete areas of increased signal adjacent to the subcortical bone, were measured using a modified Whole-Organ Magnetic Resonance Imaging Score (0 = none, 1 ≤ 25% of the subregion, 2 = 25%-50%, and 3 ≥ 50%).²⁵ A total score of the tibiofemoral compartment was calculated as the total of 13 subregional scores (0-39). The intraclass correlation coefficient of this bone marrow lesion ranged from 0.93 to 0.98.

Serum 25-hydroxyvitamin D was assayed at screening, month 3, and month 24 using direct competitive chemiluminescent immunoassays (DiaSorin Inc). The intraassay and interassay coefficients of variation were 3.2% and 6.0%.

Quantification of cartilage volume loss has been used to monitor the progression of knee osteoarthritis.²⁶ Previous studies reported that mean annual loss of medial tibial cartilage volume loss in patients with knee osteoarthritis was 4.5%.²⁷ Monthly intake of 50 000 IU of vitamin D would achieve serum 25-hydroxyvitamin D levels greater than 60 nmol/L²⁸ and this change was estimated to lead to an absolute reduction in medial tibial cartilage loss of 2.2% annually,⁸ which was expected to translate into a risk reduction of 44% for total knee replacement over 4 years.²⁹ Sample size calculation assumed α = .05 and β = .20 and was performed based on the Cohen formula.³⁰ We calculated that 400 participants at baseline (200 in each group), allowing 20% for dropouts, would have at least 80% power to detect a 2.2% between-group difference in medial tibial cartilage loss. For change in WOMAC pain, we anticipated a standard deviation of 70.5 on a score from 0 to 500.¹² With 400 participants, a difference between groups of 20 units on the score is detectable with 80% power.

WOMAC and visual analog scale knee pain scores were analyzed using a repeated-measures mixed model with terms for age, sex, body mass index (calculated as weight in kilograms divided by height in meters squared), treatment, month, and trial center. The correlation within the repeated measures was addressed by using an individual participant identification as a random effect. The effect of treatment was evaluated by the month × treatment interaction. In a post hoc analysis, the proportion of participants who achieved at least 20% and 50% improvement in WOMAC pain score was evaluated, which has been shown to be clinically relevant.³¹

The independent t test was used to compare annual changes in cartilage volume and absolute changes of cartilage defects and bone marrow lesions between groups. An increase in cartilage defects and bone marrow lesions was defined as a change of more than 1 unit in score. Presence of an increase in cartilage defects or bone marrow lesions was compared between the 2 groups using logistic regression.

Both intention-to-treat and per-protocol analyses were used. Per-protocol analysis was defined as achieving a 25-hydroxyvitamin D level of greater than 60 nmol/L at the month 3 visit. Multiple imputation by chained equations was used to address missing data caused by loss to follow-up and nonresponses. Imputations were performed separately for each treatment group and each outcome using baseline values, age, sex, body mass index, and serum 25-hydroxyvitamin D level. All statistical analyses were performed using Stata version 13.0 (Stata Corporation) and a 2-sided P value of .05 was considered statistically significant.

Figure 1 shows the flow of study participants. A total of 599 participants were screened for eligibility from June 5, 2010, to December 1, 2011, and 413 participants were randomly assigned to receive either vitamin D (n = 209) or placebo (n = 204). The mean age of participants was 63.2 years, 208 (50%) were women, and mean body mass index was 29.6. Participants’ demographic characteristics were comparable at baseline between 2 groups (Table 1). Seventy-three participants withdrew from the study (28 [13.4%] in the vitamin D group and 45 [22.1%] in the placebo group [P = .02]) and 340 participants (82.3%) completed the trial. There were no significant differences between participants who completed the study vs those who did not, except that among those who withdrew, more were women and had lower tibial cartilage volume (eTable in Supplement 2). Fewer participants discontinued treatment in the vitamin D group (8) than the placebo group (21). The major reason for a higher drop-out rate in the placebo group was that participants had their 25-hydroxyvitamin D levels checked by a primary care physician and started taking vitamin D after finding low vitamin D levels. Consequently, they were withdrawn from the study. All available data from the randomized participants were included in the intention-to-treat analyses.

The mean serum 25-hydroxyvitamin D level increased by 40.6 nmol/L in vitamin D group and by 6.7 nmol/L in placebo group over 2 years. Overall, 165 (79%) participants in the vitamin D group and 88 (43%) participants in the placebo group reached a 25-hydroxyvitamin D level of greater than 60 nmol/L at month 3.

Changes in WOMAC knee pain are presented in Table 2. At baseline, the mean (SD) of WOMAC pain scores were 137.9 (88.8) in the vitamin D group and 134.7 (83.4) in the placebo group (difference, 3.2 [95% CI, −13.5 to 19.8]; P = .71). Total WOMAC pain decreased over 24 months in both groups (Figure 2A). At month 24, the mean (SD) of WOMAC pain scores were 87.0 (90.1) in the vitamin D group and 97.2 (87.5) in the placebo group (difference, −10.2 [95% CI, −28.8 to 8.4]; P = .28). There was no difference in change in WOMAC pain between groups in the mixed-effect model in which all time points were included (−49.9 for the vitamin D group vs −35.1 for the placebo group; between-group difference, −14.8 [95% CI, −32.5 to 2.9]; P = .10).

Tibial cartilage volume (mean [SD]) at baseline was not different between the vitamin D group (3466 mm³ [1038]) and the placebo group (3640 mm³ [1036]) (between-group difference, −174 mm³ [95% CI of difference, −375 to 27]; P = .09). At month 24, tibial cartilage volume was also not different between the vitamin D group (3238 mm³ [989]) and the placebo group (3398 mm³ [1030]) (between-group difference, −160 mm³ [95% CI of difference, −369 to 49]; P = .13). Change in tibial cartilage volume (Table 2) was not different between the groups (−242.6 mm³ for the vitamin D group vs −301.4 mm³ for the placebo group [between-group difference, 58.8 mm³ {95% CI, −13.9 to 131.4}]; P = .11). Per-protocol analysis comparing participants who achieved a 25-hydroxyvitamin D level of greater than 60 nmol/L at their month 3 visit (n = 253) with those who did not (n = 146) (14 participants withdrew within 3 months) showed similar results (−261.9 mm³ vs −284.8 mm³ [between-group difference, 22.9 mm³ {95% CI, −51.5 to 97.3}]; P = .55).

The results for tibiofemoral cartilage defects and bone marrow lesions are shown in Table 2. The difference in cartilage defect score was not different between groups. Bone marrow lesion scores decreased in both groups and no significant difference was observed.

In post hoc analyses (Table 2), participants in the vitamin D group had statistically significant improvements in visual analog scale knee pain (Figure 2E) scores when compared with the placebo group. The vitamin D group had more improvement in the total WOMAC score (Figure 2B) and WOMAC function (Figure 2C) but not WOMAC stiffness (Figure 2D). There were 115 (64%) participants in the vitamin D group and 95 (57%) participants in the placebo group (P = .16) who achieved a 20% improvement in WOMAC knee pain score over 2 years. There were 90 (50%) participants in the vitamin D group and 65 (39%) participants in the placebo group (P = .04) who showed at least a 50% improvement in WOMAC pain score. There were more OMERACT-OARSI responders in the vitamin D group (74/209 [35%]) than the placebo group (52/204 [25%]) (P = .03). The proportion of participants who had an increase in bone marrow lesions was lower in vitamin D group (44/183 [24%]) than in the placebo group (61/175 [35%]) (P = .03).

Fifty six of the 209 participants (27%) in the vitamin D group reported at least 1 adverse event vs 37 of the 204 participants (18%) in the placebo group (Table 3). Four participants developed hypercalcemia in the vitamin D group vs 2 in the placebo group. One participant in the vitamin D group had symptoms of hyperparathyroidism (eg, muscle cramps, brittle bones, and kidney dysfunction) vs none in the placebo group. There was 1 episode of renal calculus in each group.

The purpose of this RCT was to determine whether vitamin D supplementation could reduce knee pain and cartilage loss and also prevent progression of other knee structural abnormalities in patients with knee osteoarthritis and low 25-hydroxyvitamin D levels. Results showed that even among study participants with low 25-hydroxyvitamin D, supplementation did not slow cartilage loss or improve WOMAC-assessed pain. These data suggest a lack of evidence to support vitamin D supplementation for slowing disease progression or structural change in knee osteoarthritis.

Although epidemiological studies suggest that knee osteoarthritis is more prevalent among individuals who are deficient in vitamin D, and vitamin D deficiency is associated with cartilage loss^8-10 and knee osteoarthritis symptoms,^6,7,32 the results from 2 prior RCTs were mixed. In one study, supplementation of vitamin D₃ (2000 IU/day) over 2 years showed no benefit for symptoms and structural changes in patients with knee osteoarthritis, regardless of their 25-hydroxyvitamin D levels.⁵ The other study reported a small but statistically significant benefit on symptoms in patients with vitamin D insufficiency over 1 year.¹² Both studies have limitations. The first study included patients without vitamin D deficiency who may not benefit from vitamin D supplementation and patients whose disease was too severe to respond to vitamin D treatment. Also, it had a small sample size (146 participants).¹³ The second study did not examine structural changes and had a 1-year follow-up, which may be too short to observe disease progression.³³ Our study addressed these limitations by recruiting patients without severe knee osteoarthritis with low 25-hydroxyvitamin D levels and provided follow-up for 2 years. Nonetheless, our results are largely consistent with the prior 2 trials.

Structural changes in cartilage and noncartilaginous joint tissue assessed using MRI, are now recommended outcomes for clinical trials in osteoarthritis.³⁴ An observational study showed that lower serum 25-hydroxyvitamin D levels were associated with greater cartilage volume loss over 2.7 years.⁸ In the current study, the amount of tibial cartilage volume loss in the placebo group is consistent with the findings of a previous RCT.⁵ We did not find significant effects on change in knee cartilage defects and bone marrow lesions.

Adverse effects of vitamin D use may include hypercalcemia. Although intermittent use of very high-dose vitamin D (eg, 500 000 IU/year) may not be safe,³⁵ our study suggests that a monthly regimen at a dosage of 50 000 IU is safe in elderly patients, even though the serum 25-hydroxyvitamin D levels of long-term users are at the upper limit of the normal range.³⁶

The key strength of this RCT is the inclusion and exclusion criteria. This study included only adult patients with knee osteoarthritis who had a vitamin D insufficiency—patients who may be the most likely to benefit from vitamin D supplements. We also used a predefined range of knee pain to prevent a ceiling or floor effect in the statistical analyses. Patients with late-stage knee osteoarthritis were excluded because of very little cartilage remaining; thus, any possible benefits of therapy on cartilage would be difficult to identify. By using these criteria, we studied a patient population in whom the likelihood of demonstrating an effect (if truly present) of vitamin D supplementation was maximized.

This study also had limitations. First, WOMAC pain as a second primary outcome was added during the recruitment period at the time the protocol was published. However, this change was made before the trial was completed, before any data analyses, and the original sample size had sufficient power to detect the expected difference in WOMAC pain. Second, loss to follow-up was 17.7% and was less in the vitamin D group (28 participants) than in the placebo group (45 participants) (P = .02). There were fewer participants who withdrew from their assigned intervention in the vitamin D group (n = 8) than in the placebo group (n = 21). Participants who did not adhere to their assigned intervention could be expected to have a worse outcome than those who did. Although this could bias the result toward the null, similar results were seen in per-protocol analysis, suggesting the differential dropout rate had minimal effects on our results. Last, this study did not prespecify clinical outcomes such as visual analog scale knee pain and WOMAC physical function as primary or secondary end points.

Among patients with symptomatic knee osteoarthritis and low serum 25-hydroxyvitamin D levels, vitamin D supplementation, when compared with placebo, did not result in significant differences in change in MRI-measured tibial cartilage volume or change in WOMAC knee pain score over 2 years. These findings do not support the use of vitamin D supplementation for preventing tibial cartilage loss or improving WOMAC knee pain among patients with knee osteoarthritis.

Corresponding Author: Changhai Ding, MD, PhD, Menzies Institute for Medical Research, University of Tasmania, Private Bag 23, Hobart, Tasmania, Australia 7000 (changhai.ding@utas.edu.au).

Author Contributions: Dr Ding had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Ding, Jin, and Jones contributed equally to this study.

Study concept and design: Ding, Jones, Cicuttini, Winzenberg.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Jin, Cicuttini, Zhu, Antony, Winzenberg, Blizzard, Ding.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Jin, Zhu, Antony, Wang, Blizzard, Ding.

Obtained funding: Ding, Jones, Cicuttini, Wluka, Winzenberg.

Administrative, technical, or material support: All authors.

Study supervision: Ding, Jones, Cicuttini.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Jones reports being the recipient of a National Health and Medical Research Council (NHMRC) Practitioner Fellowship. Drs Wluka and Winzenberg report being recipients of NHMRC Career Development Fellowship. Dr Ding reports being the recipient of an Australian Research Council Future Fellowship. No other disclosures were reported.

Funding/Support: This study was supported by a grant from the Australian National Health and Medical Research Council (NHMRC) (project code 605501).

Role of the Funder/Sponsor: The Australian National Health and Medical Research Council (NHMRC) had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: We thank Jodi Barling, BSN, Menzies Institute for Medical Research; Kay Nguo, PhD, Monash University; Judy Hankin, BSc(Psych), Monash University; and Alice Noone BS, Monash University, for their involvement in the coordination of this study. We also thank Rob Warren, BS, Menzies Institute for Medical Research, for measuring knee cartilage volume. These individuals were compensated in association with their respective contributions to this study.