Mendelian randomization study on vitamin D levels and osteoarthritis risk: a concise report

Bergink, Arjan P; Trajanoska, Katerina; Uitterlinden, André G; van Meurs, Joyce B J

doi:10.1093/rheumatology/keaa697

Abstract

Objective

The role of vitamin D in OA is unclear and previous epidemiological studies have provided inconsistent results. We conducted a two-sample Mendelian randomization (MR) study to investigate the causal relationship between genetically determined serum vitamin D levels and hip/knee OA.

Methods

Six single-nucleotide polymorphisms (SNPs) associated with vitamin D levels in the Study of Underlying Genetic Determinants of Vitamin D and Highly Related Traits Consortium were selected as instrumental variables. Summary statistics of the SNPs effects on OA were derived from the Iceland and UK Biobank, comprising 23 877 knee OA cases, 17 151 hip OA cases and >562 000 controls. The control samples match the OA cases in age, sex and county of origin.

Results

The MR analyses showed no causal association between genetically determined vitamin D levels and knee OA [odds ratio (OR) 1.03 (95% CI 0.84, 1.26)] or hip OA [OR 1.06 (95% CI 0.83, 1.35)].

Conclusion

Genetic variations associated with low vitamin D serum levels are not associated with increased risk of hip or knee OA in community-dwelling older adults, suggesting that vitamin D levels are not causally linked to OA. It is therefore unlikely that vitamin D supplementation protects against hip or knee OA.

Mendelian randomization, genetics, vitamin D, osteoarthritis

Rheumatology key messages

Genetically determined vitamin D levels are not associated with hip or knee osteoarthritis risk.
It is therefore unlikely that vitamin D supplementation protects against hip or knee OA.

Introduction

Vitamin D status influences bone metabolism; while severe deficiency causes rickets in children and osteomalacia in adults, more mild deficiencies lead to a decrease in bone mineral density (BMD) and increased fracture risk [1]. Since OA is associated with bone characteristics like increased BMD [2] and fracture risk [3], it is hypothesized that vitamin D status influences the development of OA. However, studies on the association between vitamin D levels and OA show inconsistent results [4] .

One possible explanation for these conflicting results is that the vitamin D status of the subjects measured in the circulation at one time point does not reflect the actual vitamin D status over a longer period of time. The vitamin D levels at the time of blood draw may be influenced by factors such as sun exposure, dietary intake and vitamin D supplementation, which cannot be completely corrected for by adjustment or matching cases and controls. Another problem in studies on the association between vitamin D levels and OA is the possible confounding effect of physical activity level: lower activity could lead to less sun exposure, resulting in lower vitamin D levels.

A way to obtain more reliable estimates of the vitamin D status over a longer period of time without confounding is to consider genetic variants that determine vitamin D levels. A meta-analyses of a genome-wide association study (GWAS) in the Study of Underlying Genetic Determinants of Vitamin D and Highly Related Traits (SUNLIGHT) Consortium found common variants near genes involved in cholesterol synthesis (DHCR7), hydroxylation (CYP2R1, CYP24A1) and vitamin D transport (GC) influencing vitamin D status [5]. In an expansion of this study, two additional loci in SEC23A and AMDHD1 were found [6]. Individuals with risk alleles at these loci had substantially elevated risk for vitamin D insufficiency.

We conducted a two-sample Mendelian randomization (MR) study to examine whether there is a causal relationship between genetically increased vitamin D levels and hip/knee OA. MR is a study design in which genetic variants are used as a proxy for modifiable exposure to test the unconfounded effect of the exposure on a specific outcome.

Methods

Six single-nucleotide polymorphisms (SNPs) associated with the 25-hydroxyvitamin D [25(OH)D] level were identified from the expanded SUNLIGHT GWAS meta-analysis [6] (79 366 discovery sample and a 40 562 replication sample): rs3755967 (in GC), rs12785878 (in NADSYN1/DHCR7), rs10741657 (in CYP2R1), rs17216707 (in CYP24A1), rs10745742 (in AMDHD1) and rs8018720 (in SEC23A). Next we conducted an MR study to describe the effect of genetically lowered 25(OH)D on the odds of hip and knee OA in the Iceland and UK Biobank meta-analysis [7]. The Icelandic study included 5714 individuals with hip OA and 4672 individuals with knee OA, as well as >172 000 controls. The UK Biobank study contained 11 437 individuals with hip OA, 19 205 with knee OA and >389 000 controls, all of white British ancestry, selected from the UK Biobank resource of 500 000 British volunteers [7].

In our MR analysis, we assessed the effects of the SNPs on the risk of hip and knee OA, weighting the effect of each SNP by its effect on vitamin D. The individual estimates of the effect were then pooled using statistically efficient estimators formally analogous to those of inverse-variance weighted (IVW) meta-analysis [8]. In addition, we used weighted median (WM), an MR method that generates consistent effect estimates as long as >50% of the SNPs are valid instrumental variables [9]. The R package MendelianRandomization [10] was used to generate the IVW and WM MR effect estimates (R Foundation for Statistical Computing, Vienna, Austria).

Results

In Table 1, the association between the six SNPs associated with vitamin D levels and hip and knee OA are shown. In the Iceland population, two SNPs reached nominal significance for one of the studied OA phenotypes (rs12785878 for hip OA and rs10745742 for knee OA). In the UK Biobank population, rs17216707 was significantly associated with knee OA. The combined analysis showed only a nominal significant association for rs17216707, where the C allele (associated with a lower vitamin D level) was associated with a decreased risk for knee OA.

Table 1

Open in new tab

Vitamin D SNPs and hip/knee OA

Phenotype	SNP	Chromo some	Effect allele	Other allele	Iceland		UK Biobank		Iceland and UK Biobank combined
Phenotype	SNP	Chromo some	Effect allele	Other allele	OR	P-value	OR	P-value	OR	95% CI	P-value
Hip OA	rs3755967	4	T	C	1.027	0.316	1.016	0.283	1.019	0.993, 1.045	0.154
	rs12785878	11	G	T	1.055	0.035	0.984	0.345	1.005	0.978, 1.033	0.703
	rs10741657	11	A	G	1.009	0.721	0.982	0.201	0.989	1.013, 0.966	0.352
	rs17216707	20	C	T	0.970	0.334	1.028	0.120	1.014	0.984, 1.045	0.372
	rs10745742	12	T	C	1.002	0.935	0.984	0.249	0.988	1.012, 0.965	0.335
	rs8018720	14	G	C	0.982	0.605	1.012	0.492	1.006	0.975, 1.038	0.708
Knee OA	rs3755967	4	T	C	1.025	0.384	1.015	0.190	1.017	0.995, 1.038	0.124
	rs12785878	11	G	T	1.053	0.052	1.008	0.534	1.017	0.994, 1.040	0.159
	rs10741657	11	A	G	1.022	0.401	0.995	0.639	1.000	1.019, 0.979	0.915
	rs17216707	20	C	T	0.994	0.850	0.953	0.001	0.959	0.983, 0.935	0.001
	rs10745742	12	T	C	0.937	0.015	0.988	0.281	0.981	1.000, 0.962	0.055
	rs8018720	14	G	C	1.004	0.918	1.002	0.910	1.002	0.976, 1.028	0.887

Phenotype	SNP	Chromo some	Effect allele	Other allele	Iceland		UK Biobank		Iceland and UK Biobank combined
Phenotype	SNP	Chromo some	Effect allele	Other allele	OR	P-value	OR	P-value	OR	95% CI	P-value
Hip OA	rs3755967	4	T	C	1.027	0.316	1.016	0.283	1.019	0.993, 1.045	0.154
	rs12785878	11	G	T	1.055	0.035	0.984	0.345	1.005	0.978, 1.033	0.703
	rs10741657	11	A	G	1.009	0.721	0.982	0.201	0.989	1.013, 0.966	0.352
	rs17216707	20	C	T	0.970	0.334	1.028	0.120	1.014	0.984, 1.045	0.372
	rs10745742	12	T	C	1.002	0.935	0.984	0.249	0.988	1.012, 0.965	0.335
	rs8018720	14	G	C	0.982	0.605	1.012	0.492	1.006	0.975, 1.038	0.708
Knee OA	rs3755967	4	T	C	1.025	0.384	1.015	0.190	1.017	0.995, 1.038	0.124
	rs12785878	11	G	T	1.053	0.052	1.008	0.534	1.017	0.994, 1.040	0.159
	rs10741657	11	A	G	1.022	0.401	0.995	0.639	1.000	1.019, 0.979	0.915
	rs17216707	20	C	T	0.994	0.850	0.953	0.001	0.959	0.983, 0.935	0.001
	rs10745742	12	T	C	0.937	0.015	0.988	0.281	0.981	1.000, 0.962	0.055
	rs8018720	14	G	C	1.004	0.918	1.002	0.910	1.002	0.976, 1.028	0.887

SNP: single nucleotide polymorphism, OR: odds ratio, OA: osteoarthritis

Table 1

Open in new tab

Vitamin D SNPs and hip/knee OA

Phenotype	SNP	Chromo some	Effect allele	Other allele	Iceland		UK Biobank		Iceland and UK Biobank combined
Phenotype	SNP	Chromo some	Effect allele	Other allele	OR	P-value	OR	P-value	OR	95% CI	P-value
Hip OA	rs3755967	4	T	C	1.027	0.316	1.016	0.283	1.019	0.993, 1.045	0.154
	rs12785878	11	G	T	1.055	0.035	0.984	0.345	1.005	0.978, 1.033	0.703
	rs10741657	11	A	G	1.009	0.721	0.982	0.201	0.989	1.013, 0.966	0.352
	rs17216707	20	C	T	0.970	0.334	1.028	0.120	1.014	0.984, 1.045	0.372
	rs10745742	12	T	C	1.002	0.935	0.984	0.249	0.988	1.012, 0.965	0.335
	rs8018720	14	G	C	0.982	0.605	1.012	0.492	1.006	0.975, 1.038	0.708
Knee OA	rs3755967	4	T	C	1.025	0.384	1.015	0.190	1.017	0.995, 1.038	0.124
	rs12785878	11	G	T	1.053	0.052	1.008	0.534	1.017	0.994, 1.040	0.159
	rs10741657	11	A	G	1.022	0.401	0.995	0.639	1.000	1.019, 0.979	0.915
	rs17216707	20	C	T	0.994	0.850	0.953	0.001	0.959	0.983, 0.935	0.001
	rs10745742	12	T	C	0.937	0.015	0.988	0.281	0.981	1.000, 0.962	0.055
	rs8018720	14	G	C	1.004	0.918	1.002	0.910	1.002	0.976, 1.028	0.887

Phenotype	SNP	Chromo some	Effect allele	Other allele	Iceland		UK Biobank		Iceland and UK Biobank combined
Phenotype	SNP	Chromo some	Effect allele	Other allele	OR	P-value	OR	P-value	OR	95% CI	P-value
Hip OA	rs3755967	4	T	C	1.027	0.316	1.016	0.283	1.019	0.993, 1.045	0.154
	rs12785878	11	G	T	1.055	0.035	0.984	0.345	1.005	0.978, 1.033	0.703
	rs10741657	11	A	G	1.009	0.721	0.982	0.201	0.989	1.013, 0.966	0.352
	rs17216707	20	C	T	0.970	0.334	1.028	0.120	1.014	0.984, 1.045	0.372
	rs10745742	12	T	C	1.002	0.935	0.984	0.249	0.988	1.012, 0.965	0.335
	rs8018720	14	G	C	0.982	0.605	1.012	0.492	1.006	0.975, 1.038	0.708
Knee OA	rs3755967	4	T	C	1.025	0.384	1.015	0.190	1.017	0.995, 1.038	0.124
	rs12785878	11	G	T	1.053	0.052	1.008	0.534	1.017	0.994, 1.040	0.159
	rs10741657	11	A	G	1.022	0.401	0.995	0.639	1.000	1.019, 0.979	0.915
	rs17216707	20	C	T	0.994	0.850	0.953	0.001	0.959	0.983, 0.935	0.001
	rs10745742	12	T	C	0.937	0.015	0.988	0.281	0.981	1.000, 0.962	0.055
	rs8018720	14	G	C	1.004	0.918	1.002	0.910	1.002	0.976, 1.028	0.887

SNP: single nucleotide polymorphism, OR: odds ratio, OA: osteoarthritis

Table 2 shows the results of the MR analysis between vitamin D levels and knee and hip OA. No causal effect of the genetic risk variants on knee or hip OA were found: for knee OA the odds ratio (OR) was 1.03 (95% CI 0.84, 1.26; IVW method) and for hip OA the OR was 1.06 (95% CI 0.83, 1.35). With the weight median method, similar results were found.

Table 2

Open in new tab

Results of MR analyses

Trait	Method	Fixed model
Trait	Method	β	s.e.	OR^a (95% CI)	P-value	P-heterogeneity
Knee OA	IVW	0.032	0.103	1.03 (0.84, 1.26)	0.76	0.63
Knee OA	Weighted median	0.077	0.112	1.08 (0.87, 1.34)	0.49
Hip OA	IVW	0.059	0.124	1.06 (0.83, 1.35)	0.63	0.98
Hip OA	Weighted median	0.085	0.136	1.09 (0.83, 1.42)	0.53

Trait	Method	Fixed model
Trait	Method	β	s.e.	OR^a (95% CI)	P-value	P-heterogeneity
Knee OA	IVW	0.032	0.103	1.03 (0.84, 1.26)	0.76	0.63
Knee OA	Weighted median	0.077	0.112	1.08 (0.87, 1.34)	0.49
Hip OA	IVW	0.059	0.124	1.06 (0.83, 1.35)	0.63	0.98
Hip OA	Weighted median	0.085	0.136	1.09 (0.83, 1.42)	0.53

a

Per 1 unit increase in vitamin D (nmol/l).

Table 2

Open in new tab

Results of MR analyses

Trait	Method	Fixed model
Trait	Method	β	s.e.	OR^a (95% CI)	P-value	P-heterogeneity
Knee OA	IVW	0.032	0.103	1.03 (0.84, 1.26)	0.76	0.63
Knee OA	Weighted median	0.077	0.112	1.08 (0.87, 1.34)	0.49
Hip OA	IVW	0.059	0.124	1.06 (0.83, 1.35)	0.63	0.98
Hip OA	Weighted median	0.085	0.136	1.09 (0.83, 1.42)	0.53

Trait	Method	Fixed model
Trait	Method	β	s.e.	OR^a (95% CI)	P-value	P-heterogeneity
Knee OA	IVW	0.032	0.103	1.03 (0.84, 1.26)	0.76	0.63
Knee OA	Weighted median	0.077	0.112	1.08 (0.87, 1.34)	0.49
Hip OA	IVW	0.059	0.124	1.06 (0.83, 1.35)	0.63	0.98
Hip OA	Weighted median	0.085	0.136	1.09 (0.83, 1.42)	0.53

a

Per 1 unit increase in vitamin D (nmol/l).

Discussion

Past observational studies on the association between vitamin D serum levels and OA led to conflicting results. These inconsistencies may be explained by confounding factors associated with low vitamin D levels. Especially in cross-sectional studies, a possible association can be influenced by the fact that subjects with OA of the lower limbs are less mobile and are thus less exposed to sunlight and so will have lower vitamin D serum levels. In longitudinal studies, low vitamin D levels at baseline could also indicate low activity levels, which could be related to worsening of OA. Zhang et al. [11] demonstrated that exercise may decrease cytokine and cytokine-related protein levels in the synovial fluid and inhibit inflammatory factor–mediated cartilage degradation in knee OA patients. To eliminate the influence of lifestyle factors associated with low vitamin D levels, we conducted an MR study on vitamin D levels in relation to OA.

Our MR analyses, with six SNPs associated with vitamin D levels as instrumental variables and comprising 23 877 knee OA cases, 17 151 hip OA cases and >562 000 controls, showed no causal association between genetically determined vitamin D levels and knee or hip OA. This is in line with most randomized controlled trials conducted to study the effect of vitamin D on the development of OA, finding no evidence for benefits of supplementation [12].

There are some limitations to our study. The OA case definition originated from electronic health records and clinical evaluations and was not defined by both symptoms and radiographs, using the Kellgren and Lawrence grade to define radiographic OA, as recommended for population cohort studies [13]. We can therefore not rule out that specific radiologic features of OA, like prevalent joint space narrowing of the knee or progressive knee OA, are associated with low vitamin D levels, as shown in past studies [4]. Furthermore, the combined effect on the vitamin D levels of the six SNPs we used could be too little to detect an association with knee or hip OA. Another limitation is that our MR study was based on only six SNPs as identified in an earlier GWAS and with 7.5% explained variance [6], while the latest GWAS on 25(OH)D levels identified 143 independent SNPs explaining 13% of the variance [14]. Currently we cannot exclude that associations might have been missed because of the low explained variance but do not expect this to change when using the MR instrument based on 143 SNPs. Finally, in our MR analyses we assume that there is a linear exposure risk association, thus non-linear relations, like the threshold effect, or U-shaped relations, will not be detected.

The strength of this study is that by studying genetically predicted vitamin D levels instead of a one-time measurement of circulating vitamin D levels, a possible association with OA is not confounded by lifestyle factors such as activity level, vitamin D intake or supplementation. We conclude that the common genetic variants associated with low serum vitamin D levels are not associated with increased risk of hip or knee OA, thus the supplementation of vitamin D does not protect against hip or knee OA.

Acknowledgements

The authors are grateful to Unnur Styrkarsdottir of DeCODE genetics/Amgen, Reykjavik, Iceland, for supplying the data necessary for our analyses. A.P.B. was responsible for conception and design, analysis and interpretation of the data and drafting of the manuscript and provided statistical expertise. K.T. was responsible for analysis and interpretation of the data and provided statistical expertise. A.G.U. was responsible for conception and design; analysis and interpretation of the data; critical revision of the manuscript for important intellectual content; provision of study materials or patients; administrative, technical or logistic support and collection and assembly of data; obtained funding and provided statistical expertise. J.B.J.v.M. was responsible for conception and design; analysis and interpretation of the data; drafting of the manuscript; critical revision of the manuscript for important intellectual content; administrative, technical or logistic support and collection and assembly of data. All authors approved the final article.

Funding: None of the authors received any funding related to the writing of this article.

Disclosure statement: The authors have no conflicts of interest to disclose.

Data availability statement

The data underlying this article will be shared upon reasonable request to the corresponding author.

References

1

Lips

P.

Vitamin D deficiency and secondary hyperparathyroidism in the elderly: consequences for bone loss and fractures and therapeutic implications

.

Endocr Rev

2001

;

22

:

477

–

501

. PubMed PMID: 11493580.

2

Bergink

AP

,

Rivadeneira

F

,

Bierma-Zeinstra

SM

et al.

Are bone mineral density and fractures related to the incidence and progression of radiographic osteoarthritis of the knee, hip, and hand in elderly men and women? The Rotterdam Study

.

Arthritis Rheumatol

2019

;

71

:

361

–

9

.

3

Bergink

AP

,

van der Klift

M

,

Hofman

A

et al.

Osteoarthritis of the knee is associated with vertebral and nonvertebral fractures in the elderly: the Rotterdam Study

.

Arthritis Rheum

2003

;

49

:

648

–

57

. PubMed PMID: 14558050.

4

Bergink

AP

,

Zillikens

MC

,

Van Leeuwen

JP

et al.

25-Hydroxyvitamin D and osteoarthritis: a meta-analysis including new data

.

Semin Arthritis Rheum

2016

;

45

:

539

–

46

.

5

Wang

TJ

,

Zhang

F

,

Richards

JB

et al.

Common genetic determinants of vitamin D insufficiency: a genome-wide association study

.

Lancet

2010

;

376

:

180

–

8

.

6

Jiang

X

,

O’Reilly

PF

,

Aschard

H

et al.

Genome-wide association study in 79,366 European-ancestry individuals informs the genetic architecture of 25-hydroxyvitamin D levels

.

Nat Commun

2018

;

9

:

260

.

7

Styrkarsdottir

U

,

Lund

SH

,

Thorleifsson

G

et al.

Meta-analysis of Icelandic and UK data sets identifies missense variants in SMO, IL11, COL11A1 and 13 more new loci associated with osteoarthritis

.

Nat Genet

2018

;

50

:

1681

–

7

.

8

Burgess

S

,

Butterworth

A

,

Thompson

SG.

Mendelian randomization analysis with multiple genetic variants using summarized data

.

Genet Epidemiol

2013

;

37

:

658

–

65

.

9

Bowden

J

,

Davey Smith

G

,

Haycock

PC

,

Burgess

S.

Consistent estimation in Mendelian randomization with some invalid instruments using a weighted median estimator

.

Genet Epidemiol

2016

;

40

:

304

–

14

.

10

Yavorska

OO

,

Burgess

S.

MendelianRandomization: an R package for performing Mendelian randomization analyses using summarized data

.

Int J Epidemiol

2017

;

46

:

1734

–

9

.

11

Zhang

SL

,

Liu

HQ

,

Xu

XZ

et al.

Effects of exercise therapy on knee joint function and synovial fluid cytokine levels in patients with knee osteoarthritis

.

Mol Med Rep

2013

;

7

:

183

–

6

. PubMed PMID: 23135204. Epub 2012/11/09. Eng.

12

Hussain

S

,

Singh

A

,

Akhtar

M

,

Najmi

AK.

Vitamin D supplementation for the management of knee osteoarthritis: a systematic review of randomized controlled trials

.

Rheumatol Int

2017

;

37

:

1489

–

98

.

13

Leyland

KM

,

Gates

LS

,

Nevitt

M

et al.

Harmonising measures of knee and hip osteoarthritis in population-based cohort studies: an international study

.

Osteoarthritis Cartilage

2018

;

26

:

872

–

9

.

14

Revez

JA

,

Lin

T

,

Qiao

Z

et al.

Genome-wide association study identifies 143 loci associated with 25 hydroxyvitamin D concentration

.

Nat Commun

2020

;

11

:

1647

.

This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://dbpia.nl.go.kr/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Download all slides

Month:	Total Views:
January 2021	50
February 2021	22
March 2021	25
April 2021	21
May 2021	30
June 2021	8
July 2021	49
August 2021	48
September 2021	36
October 2021	43
November 2021	19
December 2021	33
January 2022	33
February 2022	18
March 2022	26
April 2022	43
May 2022	35
June 2022	24
July 2022	74
August 2022	85
September 2022	105
October 2022	112
November 2022	101
December 2022	111
January 2023	102
February 2023	178
March 2023	118
April 2023	138
May 2023	150
June 2023	85
July 2023	117
August 2023	99
September 2023	150
October 2023	128
November 2023	82
December 2023	61
January 2024	79
February 2024	63
March 2024	53
April 2024	62
May 2024	35
June 2024	54
July 2024	38
August 2024	33
September 2024	18
October 2024	54
November 2024	26
December 2024	27
January 2025	21
February 2025	71
March 2025	31
April 2025	30
May 2025	6

Article Contents

Mendelian randomization study on vitamin D levels and osteoarthritis risk: a concise report

Abstract

Introduction

Methods

Results

Discussion

Acknowledgements

Data availability statement

References

Comments

Citations

Views

Altmetric

Email alerts

Citing articles via

Latest

Most Read

Most Cited

Article Contents

Mendelian randomization study on vitamin D levels and osteoarthritis risk: a concise report

Abstract

Introduction

Methods

Results

Discussion

Acknowledgements

Data availability statement

References

Comments

Citations

Views

Altmetric

Email alerts

Citing articles via

Latest

Most Read

Most Cited

This Feature Is Available To Subscribers Only