Received Date : 30-Mar-2016

Revised Date : 29-Aug-2016

Accepted Date : 13-Sep-2016
Article type : Original Paper
Accepted Article
Hyperuricemia Protects Against Low Bone Mineral Density, Osteoporosis and Fractures:

A Systematic Review And Meta-Analysis

Nicola Veronese1, MD, Sara Carraro1, MD, Giulia Bano1, MD, Caterina Trevisan1, MD, Marco

Solmi2, MD, Claudio Luchini3, MD, Enzo Manzato1,4, MD, Riccardo Caccialanza5, MD, Giuseppe

Sergi1, MD , Davide Nicetto6, MD, Emanuele Cereda5, MD

1
Department of Medicine, Geriatrics Section, University of Padova, Italy.
2
Department of Neurosciences, University of Padova, Padova, Italy.
3
Department of Pathology and Diagnostics, Verona University and Hospital Trust, Verona, Italy.
4
National Research Council, Aging Branch, Institute of Neuroscience, Padova, Italy.
5
Nutrition and Dietetics Service, Fondazione IRCCS Policlinico San Matteo, Pavia, Italy.
6
Azienda Provinciale per i Servizi Sanitari (APSS) Trento, Trento, Italy

Correspondence to:

Nicola Veronese, Department of Medicine - DIMED, Geriatrics Division, University of Padova,

Via Giustiniani, 2 35128 Padova, Italy. Phone: +390498218492; Fax: +390498211218.

Email: ilmannato@gmail.com

This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process, which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
10.1111/eci.12677
This article is protected by copyright. All rights reserved.
 ABSTRACT

Background: Serum uric acid (SUA) accounts for about 50% of extracellular antioxidant activity,
Accepted Article
suggesting that hyperuricemia may have a protective role in diseases characterized by high levels of

oxidative stress, such as osteoporosis. We aimed to meta-analyze data regarding bone mineral

density (BMD), osteoporosis and fractures in people with higher SUA vs. lower SUA

concentrations.

Materials and methods: Two investigators conducted a literature search using PubMed and

Scopus, without language restrictions. Standardized mean differences (SMDs) and 95% confidence

intervals (CIs) were used for BMD; risk ratios (RRs) and adjusted odds ratios (ORs) for the

prevalence of osteoporosis. Most possible adjusted hazard ratios (HRs) were used to assess the

association between baseline SUA and incident fractures.

Results: Of 1405 initial hits, 19 studies were eligible including a total of 55,859 participants.

Subjects with higher SUA levels had significantly higher BMD values for the spine (6 studies;

SMD=0.29; 95%CI: 0.22-0.35; I2=47%), total hip (7 studies; SMD=0.29; 95%CI: 0.24-0.34;

I2=33%) and femoral neck (6 studies; SMD=0.25; 95%CI: 0.16-0.34; I2=71%). Simple correlation

analyses substantially confirmed these findings. An increase of one standard deviation in SUA

levels reduced the number of new fractures at follow-up (3 studies; HR=0.83; 95%CI: 0.74-0.92;

I2=0%). No significant differences between men and women emerged, although data about women

were limited.

Conclusions: Hyperuricemia was found independently associated with BMD, supporting a

protective role for uric acid in bone metabolism disorders.

Keywords: osteoporosis; fractures; bone mineral density; uric acid; hyperuricemia.

This article is protected by copyright. All rights reserved.

 INTRODUCTION

Uric acid is the end product of purine metabolism in humans. High serum uric acid [SUA] levels
Accepted Article
seem to be a common risk factor for several metabolic and cardiovascular diseases, namely

hypertension, coronary artery disease, diabetes and metabolic syndrome.[1–4] High SUA levels are

also the main risk factor for gout, a condition that is becoming increasingly common among the

elderly.[5,6]

SUA also accounts for about 50% of extracellular antioxidant activity[7], however, prompting the

hypothesis that hyperuricemia may have a protective role in diseases characterized by high

oxidative stress levels, such as neurodegenerative diseases.[8] Osteoporosis also seems to be

associated with high oxidative stress levels and weak antioxidant barriers.[9,10] Enhanced

osteoclastic activity observed in osteoporosis is probably responsible for increased production of

reactive oxygen species [ROS], particularly as superoxide, which is confirmed by increased levels

of serum malondialdehyde [MDA] levels. One of the most damaging effects of ROS is lipid

peroxidation, the final product of which is MDA which also served as a measure of osteoclastic

activity. Inhibition of the antioxidant enzymes activities, such as superoxide dismutase and

glutathione peroxidase, was found to increase superoxide production by the osteoclasts which is

reflected by increased levels of MDA.[11] Therefore, oxidative stress is an important mediator of

bone loss since deficiency of antioxidant vitamins has been found to be more common in the elderly

osteoporotic patients.. The association between SUA levels and osteoporosis has been explored in

animal models and in vitro, generating contrasting findings. In one study, urate crystals promoted

osteoclast growth and reduced osteoblast activity, suggesting a globally negative effect of uric acid

on bone metabolism.[12] In other experiments, administering uric acid seemed to reduce

osteoclastogenesis in a dose-dependent manner, leading to a reduction in bone resorption.[13] The

possible association between SUA and bone metabolism in humans has only been explored in

recent years, with not univocal results.

This article is protected by copyright. All rights reserved.

 Hence this systematic review and meta-analysis of studies exploring the association between SUA

levels and the following outcome measures: bone mineral density [BMD], prevalent osteoporosis
Accepted Article
and incident fractures. Since the literature reports significant differences in SUA and BMD levels

between men and women, our data analysis was stratified by gender where possible.

MATERIALS AND METHODS

Search strategy

Two investigators (GB, CT) independently conducted an electronic literature search using PubMed

and Scopus with no language restrictions, from their inception until 07/01/2015, seeking studies that

analyzed BMD, osteoporosis and fractures in relation to SUA levels. Any inconsistencies were

resolved by consensus. In PubMed, the following controlled vocabulary terms and keywords were

considered: (uric acid [Text word] OR hyperuricemia [Text word] OR urate [Text word]) AND

(osteoporosis [Mesh term] OR osteoporosis [Text word] OR osteopenia [Text word] OR bone

density [Mesh term] OR bone mineral density [Text word] OR bone mineral content [Text word]

OR bone mass [Text word] OR absorptiometry, photon [Mesh term] OR DXA [Text word] OR

DEXA [Text word] OR bone [Text word] OR fractures, bone [Mesh term] OR fracture [Text

word]). A similar search strategy was run in Scopus. Reference lists of the articles included in the

analysis and of others relevant to the topic were hand-searched to identify additional, potentially

relevant publications. Conference abstracts were also considered.

Study selection

We only considered studies that: (1) compared data on BMD, osteoporosis and fractures between

participants with higher vs. lower SUA concentrations; or (2) reported simple correlation estimates

between SUA levels and BMD. Studies reporting estimates obtained using linear regression

analyses were only considered for the descriptive analyses. Studies conducted in vitro and on

animal models, and investigations on serological data of bone metabolism were excluded.

This article is protected by copyright. All rights reserved.

 We also contacted authors to ask for further information when: 1) data could not be meta-analyzed

(because means and SDs, or equivalent values for BMD, or data on fractures or osteoporosis were
Accepted Article
lacking); and 2) other relevant information was missing (e.g. estimate boundaries). At least 4

attempts were made to contact the authors concerned.

This systematic review was conducted according to the Strengthening the Reporting of

Observational Studies in Epidemiology [STROBE] criteria[14] and the recommendations in the

Preferred Reporting Items for Systematic Reviews and Meta-Analyses [PRISMA] statement.[15]

Data extraction

Two authors (DN, CL) independently recorded data extracted from the selected studies in a

standardized Microsoft Excel spreadsheet. Any disagreement was resolved by consensus with a

third author (RC). The following information was extracted: i) characteristics of the study

population (e.g. sample size, demographics, comorbidities); ii) mean duration of follow-up [for

longitudinal studies); iii) clinical setting in which the study was performed; iv) parameters relating

to SUA levels or bone metabolism [age, body mass index, serum 25 hydroxyvitamin D [25OHD]

and parathormone [PTH] concentrations, renal function [estimated from the glomerular fraction

rate], and percentage of alcohol drinkers); v) number of adjustments; and vi) method used to

ascertain fractures and measure BMD.

Outcomes

The primary outcomes were spine, total hip and neck BMD. Simple correlation analyses were

considered as co-primary endpoints. A secondary analysis focused on the unadjusted and adjusted

risk of osteoporosis and fractures emerging from cross-sectional and longitudinal studies,

respectively.

This article is protected by copyright. All rights reserved.

 Assessment of study quality

The Newcastle-Ottawa Scale (NOS)[16] was used to assess study quality. The NOS assigns a
Accepted Article
maximum of 9 points based on three quality parameters: selection, comparability, and outcome. The

investigators solved any discrepancies by jointly re-assessing an article (NV, MS and EC).

Statistical analysis

The meta-analysis was performed using the Review Manager (RevMan) software, version 5.3 for

Windows [Cochrane Collaboration, http://ims.cochrane.org/revman]. Only outcomes addressed by

at least two studies were meta-analyzed, while single-study outcomes are reported only in the

descriptive analyses. When combining studies, the random effects model was used to account for

study heterogeneity.[17] For BMD data, means and standard deviations (SD) across quantiles of the

distribution of SUA levels were analyzed to calculate standardized mean differences [SMD]. Risk

ratios [RRs] and adjusted odds ratios [ORs] were used to combine the prevalence of osteoporosis

and the adjusted ORs for this condition, respectively. For prospective studies, the most possible

adjusted hazard ratios [HRs] were used to test the association between baseline SUA and incident

fractures. In all these analyses, the estimates referred to comparisons between the highest and

lowest SUA quantiles. Pooled Pearson’s correlation coefficients were also computed for BMD and

SUA levels. All estimates were calculated together with 95% confidence intervals [CI].

Study heterogeneity was measured using the chi-squared and I-squared statistics, assuming that a

p≤0.05 for the former and a value ≥50% for the latter indicated a significant heterogeneity.[18] For

outcomes that concerned ≥4 studies and showed a significant heterogeneity, we conducted a meta-

regression analysis to see whether some variables in the characteristics of the whole sample of

studies considered (country, sample size, percentage of women, mean age, DXA method, NOS

score), and or whether differences in age, percentage of females, proportion of alcohol drinkers,

SUA levels, BMI or eGFR between subjects with higher vs. lower SUA levels were significant

moderators. Publication bias was assessed by visually inspecting funnel plots and using the Begg-

This article is protected by copyright. All rights reserved.

 Mazumdar Kendall tau[19] and the Egger bias test.[20] Then, to account for publication bias, we

used the trim-and-fill method, based on the assumption that the effect sizes of all the studies are
Accepted Article
normally distributed around the center of a funnel plot; in the event of asymmetries, it adjusts for

the potential effect of unpublished (imputed) studies.[20] Meta-regression, simple correlations and

publication bias analyses were conducted using Comprehensive Meta-Analysis V3

[http://www.meta-analysis.com/index.php].

RESULTS

The search identified 1405 potentially eligible studies, including 278 duplicates which were

rejected. After excluding 1097 papers on the grounds of a review of their titles and abstracts

(mainly since they had not data on SUA or bone health outcomes), 30 full-text articles were

examined (removing 11 since they did not report data regarding bone health outcomes), and 19

studies were ultimately included in our qualitative investigation and 16 in our meta-analysis

(Figure 1).

Study and patient characteristics

Study and patient characteristics are summarized in Supplementary Table 1. The 19 studies

analyzed[13,21–39] concerned a total of 55,859 participants, including 21,211 women (=38.0%).

Fourteen studies had a cross-sectional design, four were prospective investigations, and one

reported on both cross-sectional and longitudinal data analyses. Among the cross-sectional studies,

3 [21,23,25] were only included in our descriptive analysis. The majority of the studies was

conducted in Asia (10 studies: 9 cross-sectional and 1 longitudinal), followed by Europe (4 studies:

3 cross-sectional and 1 longitudinal) and America (4 studies: 2 cross-sectional and 2 longitudinal),

and Oceania (1 studies: both cross-sectional and longitudinal) (Supplementary Table 1). Sixteen

studies were conducted among community-dwelling participants, 2 among outpatients and 1 in

dialysis unit.

This article is protected by copyright. All rights reserved.

 Cross-sectional and case-control studies reporting BMD parameters

The cross-sectional and case-control analyses concerned 27,509 subjects, 17,511 (63.7%) of them
Accepted Article
women. Eleven studies used a Hologic DXA to estimate BMD, 2 used a Prodigy instrument, 1 an

Osteocore, and 1 study did not provide information on the method used. The median NOS score

was 4, ranging from 2 to 9 (Supplementary Table 2).

Four[13,22,24,30]of the studies included in our analyses reported descriptive characteristics for

participants grouped by SUA level: those with higher and lower SUA levels amounted to 3,444

(2,189 women) and 3,258 (1,872 women), respectively (Supplementary Table 1).

The SMDs for spine, total hip and femoral neck BMD, comparing participants with higher vs. lower

SUA concentrations, are shown in Figure 2. Pooling data from 6 studies[13,22,24,30,32,35], 4,849

participants with higher SUA levels had significantly higher BMD values than 4,672 participants

with lower SUA levels (p<0.001). A similar result emerged for total hip BMD (7

studies[13,22,24,28,30,32,35]; p<0.001) and femoral neck BMD (6 studies[13,22,24,30,32,35];

p<0.001). No significant differences between men and women were apparent for these outcomes

(Table 1).

These findings were consistent with a meta-analysis of simple correlations between SUA levels and

BMD values (Table 2). SUA level was significantly associated with spine BMD (5

studies[13,29,33,36,37,39]; r=0.141; 95%CI: 0.099-0.182; p<0.0001; I2=76%), total hip BMD (6

studies[13,29,33,36,37,39]; r=0.120; 95%CI: 0.095-0.146; p<0.0001; I2=37%), and femoral neck

BMD (3 studies[13,33,37]; r=0.095; 95%CI: 0.077-0.113; p<0.0001; I2=0%). The association

between SUA level and spine BMD was only marginally significant in men, however (2

studies[33,36]; r=0.134; 95%CI: -0.008 to 0.272; p=0.06; I2=69%) [Table 2). One study reported

data on SUA levels and forearm BMD (1/3 of the radius and ultradistal sites in the forearm without

arteriovenous fistola) in a cohort of patients with diabetes or chronic glomerulonephritis on

hemodialysis.[25] SUA levels correlated significantly with both forearm sites, but only in the cases

This article is protected by copyright. All rights reserved.

 of chronic glomerulonephritis (R=0.34, p=0.002 for 1/3 radius; R=0.25, p=0.03 for ultradistal

site).[25]
Accepted Article
In addition to the outcomes addressed by our meta-analysis, several cross-sectional studies reported

linear regression estimates adjusted for potential confounders that substantially confirmed our

findings.[13,24,30,32,33] Two additional studies only reported linear regressions used to estimate

the association between SUA and BMD: Ishii et al.[23]] found a significant linear association

between SUA levels and spine BMD [β=0.08, p=0.049) in 615 Japanese women; and Dalbeth et al.

showed a significant association between SUA levels and total hip BMD in 2,501 American

subjects (β=0.11; 95%CI: 0.03- 0.19; p=0.0043), whose genetic urate scores were not taken into

account.[21]

Finally, Makovey et al. reported that women with higher baseline SUA levels experienced a

significantly lower loss of BMD at the spine during their follow-up.[30]

Cross-sectional and case-control studies reporting osteoporosis

As shown in Table 3, higher SUA levels were also associated with a lower presence of osteoporosis

(4 studies[13,22,32,34]; RR=0.69; 95%CI: 0.62-0.76; p<0.0001; I2=0%). Quantitative synthesis of

fully-adjusted ORs (median number of adjustments=5; range: 0-19) did not significantly alter our

findings: higher SUA levels were associated with a significant reduction in the risk of developing

osteoporosis (6 studies[13,24,32,34,37,39]; OR=0.66; 95%CI: 0.53-0.83; p<0.0001; I2=73%). These

findings were much the same in both genders (Table 3).

Longitudinal studies reporting on fractures

Five longitudinal studies[27,28,30,31,34] on fractures conducted in community-dwelling settings

followed up 28,706 participants (4,056 women [14.1%]) for a median of 5.3 years (range: 3-10),

relying on clinical assessments in four of these studies and on vertebral morphometry in one. The

median NOS score was 8, and ranged from 6 to 9 (Supplementary Table 3). All except one of

This article is protected by copyright. All rights reserved.

 these studies reported the characteristics of the samples grouped according to their higher or lower

baseline SUA concentrations (Supplementary Table 1), the two groups comprising 1,885 (963
Accepted Article
women) and 4,370 (2,410 women) participants, respectively.

The pooled adjusted HRs [median number of adjustments = 14 (range: 6-20] for the number of new

fractures during the follow-up are given in Table 4. In the analysis on the both genders together, an

increase of one SD in SUA levels corresponded to a 17% reduction in the number of new fractures

during the follow-up (3 studies[27,28,34]; HR=0.83; 95%CI: 0.74-0.92; p=0.001; I2=0%). The

results for men alone were similar, while the estimates could not be computed for women alone (no

studies available). After grouping subjects by their SUA levels and taking the group with lower

SUA levels for reference, having high SUA levels at the baseline did not appear to afford any

significant protection against incident fractures (3 studies[28,31,34]; HR=0.90; 95%CI: 0.64-1.27;

p=0.56; I2=60%), in men or women (Table 3).

Meta-regression analysis

Since some outcomes in Tables 1 and 2 had a high heterogeneity (I2>50%, p<0.05) and a sufficient

number of studies were available (at least 4 for each outcome), we ran a meta-regression analysis to

seek potential moderators of this heterogeneity. As shown in Supplementary Table 4, only study

quality was a significant moderator of lower BMD values at femoral neck sites (β=-0.10±0.02,

p<0.0001, R2=0.94). Conversely, differences in SUA between the highest and lowest quantile were

not associated with higher BMD values at spine (β=-0.0005±0.02, p=0.97, R2=0.00), total hip

(β=0.001±0.0009, p=0.31, R2=0.00) or femoral neck (β=-0.0008±0.02, p=0.97, R2=0.00).

After stratifying for median NOS score (=5), all outcomes remained significantly different between

subjects with higher versus lower SUA levels, with the exception of the total hip SMD in the

poorer-quality studies (Table 5). The SMD and R coefficients seemed to be larger in the studies of

lower quality than in those of higher quality, however (Table 5).

This article is protected by copyright. All rights reserved.

 The adjusted ORs for the presence of osteoporosis were also highly heterogeneous (I2=60%) (Table

3), whereas the unadjusted analyses were not (I2=0%), so we first checked whether the number of
Accepted Article
covariates was a significant moderator (β=-0.05±0.02, p=0.06, R2=0.43): here again, only study

quality was found to moderate higher OR values (β=-0.10±0.05, p=0.04, R2=0.14).

Publication bias

Judging from a visual inspection of funnel plots and using Egger’s test (Supplementary Table 5),

only the adjusted ORs for osteoporosis were associated with publication bias. The adjusted ORs

computed using the trim-and-fill procedure substantially confirmed our findings: genders combined,

OR=0.70 (95%CI: 0.53-0.83), men, OR=0.56 (95%CI: 0.36-0.85).

DISCUSSION

In this meta-analysis, we examined 19 observational studies involving 55,859 subjects and focusing

on parameters related to serum uric acid and bone metabolism. To the best of our knowledge, this is

the first meta-analysis to investigate the possible relationship between SUA, BMD, osteoporosis

and fractures. Cross-sectional studies suggested a strong and independent association between SUA

levels and BMD, that led to significantly lower proportions of cases and a lower risk of developing

osteoporosis (or fractures in prospective studies) for participants with higher SUA levels. When our

data were summarized descriptively, they substantially confirmed the association between high

SUA levels and high BMD. However, as most evidence relies on results from cross-sectional

studies, we do recognize that there is not enough evidence yet to substantiate causality.

Several hypotheses could be advanced to explain the significant association between high SUA

levels and a high BMD, and a consequently lower presence of osteoporosis, even after considering

potential confounders. First, there are the antioxidant properties of uric acid, which has metal-

chelating properties and is capable of scavenging superoxide, hydroxyl radicals, and singlet oxygen,

and of blocking the formation of peroxynitrite (a strong oxidant). [40] In experimental studies,

This article is protected by copyright. All rights reserved.

 oxidative stress has recently been found to down-regulate osteoblastogenesis and bone

formation.[41,42] High oxidative stress and low antioxidant marker levels seem to play a pivotal
Accepted Article
part in osteoporosis in humans too.[29] Second, SUA levels have been directly associated with

body mass index, which seems to be significantly associated with BMD values.[43] Finally, higher

SUA levels were found correlated with lower levels of bone resorption markers and 25OHD[34],

but it emerged from some studies [also included in our meta-analysis] that subjects with higher

SUA levels have a worse renal function and higher serum PTH levels, two well-known risk factors

for osteoporosis and fractures. So further studies are needed to better clarify the confounding effect

of these factors.

It is noteworthy that only some of the cross-sectional observations considered in our analysis were

heterogeneous. Study quality appeared to be the only significant moderator of our findings and

explained most of the heterogeneity, which only persisted among the lower-quality studies and for

the correlation between spine BMD and SUA levels. On the contrary other plausible moderators

(like differences in age, BMI, renal function) were not associated with higher BMD. Altogether

these findings suggest that studies with study quality is the only moderator of our findings

suggesting the existence of unmeasured confounders and that other high-quality studies are needed

for confirming our findings.

In our meta-analysis, we also explored the potential association between SUA and fractures. An

increase of one standard deviation in SUA levels at the baseline coincided with a 17% reduction in

incident fractures, but no significant protective effect of higher SUA levels came to light when

participants were grouped by higher vs. lower SUA levels. This discrepancy might have more than

one explanation. First, the analyses investigating the role of SUA based on one-SD increases in its

levels had been conducted on twice as many participants as those grouping participants by

higher/lower SUA levels, so there may be a type II error in the latter type of analysis. Second, some

methodological differences between the prospective studies need to be taken into account, including

This article is protected by copyright. All rights reserved.

 the study cohorts’ different demographic characteristics (e.g. ethnicity), the different methods used

to diagnose osteoporotic fractures, and the adoption of different exclusion/inclusion criteria.

Accepted Article
The useful findings emerging from our meta-analysis should be considered along with its

limitations. The most important one is that the evidence relies substantially on data from cross-

sectional studies, and we do recognize that there is not enough evidence yet to substantiate

causality. Nonetheless, the direction of the association was strong across all outcome measures and

the protective effect of SUA was confirmed by prospective investigations, which would support a

more convincing cause-effect relationship. However, a study included in our descriptive findings

reported that mendelian randomization is an important factor for better explaining the casual

association between higher SUA levels and BMD[21], and thus further studies using this approach

are needed. Mendelian randomization analysis, in fact, enables to investigate the possible causation

in case of suspected reverse causation. [44] It would be of interest to see in randomized controlled

trials whether lowering SUA levels coincides with worsening bone metabolism parameters. The

role of potential confounders also deserves further investigation since several studies did not take

into account their effect in the analysis. For example, 25OHD and PTH levels, or renal function

could be important mediators of the association between SUA and bone health outcomes. Finally,

we meta-analyzed studies using different heterogeneous SUA cut-offs for indicating quantiles.

However, the differences in SUA levels between the highest and lowest quintile did not appear to

explain the differences in BMD. Nonetheless we were not able to find a SUA cut-off safe for BMD

and negative outcomes associated with hyperuricemia (e.g. cardiovascular diseases). Among the

strengths of our work, on the other hand, it is worth mentioning: the large number of the studies and

participants considered in what is, to the best of our knowledge, the first meta-analysis addressing

this topic; and the fact that the prospective studies on SUA and fractures were all published in the

last two years.

This article is protected by copyright. All rights reserved.

 In conclusion, higher serum uric acid levels were found independently associated with a greater

bone mineral density, supporting a role for uric acid in bone metabolism disorders. Higher serum
Accepted Article
uric acid levels would also seem to protect against the onset of osteoporotic fractures, although

further studies are needed on this issue, particularly in women.

ACKNOWLEDGMENTS

We thank Prof. Naim Maalouf, MD, University of Texas Southwestern Medical Center, and his

colleagues for giving us additional information about their study.

Financial disclosure: Dr. Cereda reports other from NUTRICIA ITALIA, other from WUNDER,

other from AKERN, outside the submitted work. None of the authors have any financial

arrangements, organizational affiliations or other relationships that might give rise to any conflict of

interest regarding the subject matter of the submitted manuscript.

Conflict of interest: Nicola Veronese, Sara Carraro, Giulia Bano, Caterina Trevisan, Marco Solmi,

Claudio Luchini, Enzo Manzato, Riccardo Caccialanza, Giuseppe Sergi, Davide Nicetto, and

Emanuele Cereda declare that they have no conflict of interest.

Funding source: none.

Authors’ role: Nicola Veronese: conception and design of the study, draft of the paper; Sara

Carraro: acquisition of the data; Giulia Bano, Caterina Trevisan, Marco Solmi: analysis and

interpretation of data; Claudio Luchini: draft of the paper; Enzo Manzato: critical revision for

important intellectual content; Riccardo Caccialanza: acquisition of the data; Giuseppe Sergi,

Davide Nicetto: critical revision for important intellectual content; Emanuele Cereda: draft of the

This article is protected by copyright. All rights reserved.

 paper, critical revision for important intellectual content. All the Authors approved the final version

to be submitted.
Accepted Article
REFERENCES

1. Edwards NL. The role of hyperuricemia in vascular disorders [corrected] [published erratum

appears in CURR OPIN RHEUMATOL 2010 Jan;22(1):1p]. Current Opinion in

Rheumatology. 2009;21(2):132–8.

2. Holme I, Aastveit AH, Hammar N, Jungner I, Walldius G. Uric acid and risk of myocardial

infarction, stroke and congestive heart failure in 417,734 men and women in the

Apolipoprotein MOrtality RISk study (AMORIS). J Intern Med [Internet]. 2009;266(6):558–

70. Available from: http://www.ncbi.nlm.nih.gov/pubmed/19563390

3. Oda E, Kawai R, Sukumaran V, Watanabe K. Uric acid is positively associated with

metabolic syndrome but negatively associated with diabetes in Japanese men. Intern Med

[Internet]. 2009;48(20):1785–91. Available from:

https://www.jstage.jst.go.jp/article/internalmedicine/48/20/48_20_1785/_pdf

4. Dehghan A, van Hoek M, Sijbrands EJG, Hofman A, Witteman JCM. High serum uric acid

as a novel risk factor for type 2 diabetes. Diabetes care [Internet]. 2008;31(2):361–2.

Available from: http://www.ncbi.nlm.nih.gov/pubmed/17977935

5. Bolzetta F, Veronese N, Manzato E, Sergi G. Tophaceous gout in the elderly: a clinical case

review. Clinical rheumatology [Internet]. 2012;31(7):1127–32. Available from:

http://www.ncbi.nlm.nih.gov/pubmed/22367095

6. Bolzetta F, Veronese N, Manzato E, Sergi G. Chronic gout in the elderly. Aging clinical and

experimental research [Internet]. 2013;25(2):129–37. Available from:

http://www.ncbi.nlm.nih.gov/pubmed/23739897

7. Glantzounis GK, Tsimoyiannis EC, Kappas AM, Galaris DA. Uric acid and oxidative stress.

Curr Pharm Des [Internet]. 2005;11(32):4145–51. Available from:

This article is protected by copyright. All rights reserved.

 http://www.ncbi.nlm.nih.gov/pubmed/16375736

8. Chen X, Guo X, Huang R, Chen Y, Zheng Z, Shang H. Serum uric acid levels in patients
Accepted Article
with Alzheimer’s disease: A meta-analysis. PLoS ONE [Internet]. 2014;9(4):e94084.

Available from:

http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3979756&tool=pmcentrez&rend

ertype=abstract

9. Sendur OF, Turan Y, Tastaban E, Serter M. Antioxidant status in patients with osteoporosis:

A controlled study. Joint Bone Spine. 2009;76(5):514–8.

10. Wauquier F, Leotoing L, Coxam V, Guicheux J, Wittrant Y. Oxidative stress in bone

remodelling and disease. Trends in Molecular Medicine. 2009;15(10):468–77.

11. Sheweita S a, Khoshhal KI. Calcium metabolism and oxidative stress in bone fractures: role

of antioxidants. Current drug metabolism. 2007;8:519–25.

12. Chhana A, Callon KE, Pool B, Naot D, Watson M, Gamble GD, et al. Monosodium urate

monohydrate crystals inhibit osteoblast viability and function: implications for development

of bone erosion in gout. Annals of the rheumatic diseases. 2011;70(9):1684–91.

13. Ahn SH, Lee SH, Kim B-J, Lim K-H, Bae SJ, Kim EH, et al. Higher serum uric acid is

associated with higher bone mass, lower bone turnover, and lower prevalence of vertebral

fracture in healthy postmenopausal women. Osteoporosis international : a journal established

as result of cooperation between the European Foundation for Osteoporosis and the National

Osteoporosis Foundation of the USA [Internet]. 2013;24(12):2961–70. Available from:

http://www.ncbi.nlm.nih.gov/pubmed/23644878

14. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP, et al. The

Strengthening the Reporting of Observational Studies in Epidemiology (STROBE)

statement: guildelines for reporting observational studies. Journal of Clinical Epidemiology.

2008;61(4):344–9.

This article is protected by copyright. All rights reserved.

 15. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JPA, et al. The

PRISMA statement for reporting systematic reviews and meta-analyses of studies that
Accepted Article
evaluate health care interventions: explanation and elaboration. PLoS medicine [Internet].

2009;6(7):e1000100. Available from:

http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2707010&tool=pmcentrez&rend

ertype=abstract

16. Wells G, Shea B, O’Connell D, Peterson J, Welch V, Losos M, et al. The Newcastle-Ottawa

Scale (NOS) for assessing the quality if nonrandomized studies in meta-analyses. (Available

from: URL: http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp). 2012;2012.

17. DerSimonian R, Laird N. Meta-analysis in clinical trials. Controlled clinical trials.

1986;7(3):177–88.

18. Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-analysis. Statistics in

Medicine. 2002;21(11):1539–58.

19. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication

bias. Biometrics. 1994;50(4):1088–101.

20. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a

simple, graphical test. BMJ (Clinical research ed). 1997;315(September):629–34.

21. Dalbeth N, Topless R, Flynn T, Cadzow M, Bolland MJ, Merriman TR. Mendelian

randomization analysis to examine for a causal effect of urate on bone mineral density.

Journal of bone and mineral research : the official journal of the American Society for Bone

and Mineral Research. 2015;30(6):985–91.

22. Hernández JL, Olmos JM, Pariente E, Martínez J, Valero C, García-Velasco P, et al.

Metabolic syndrome and bone metabolism: the Camargo Cohort study. Menopause (New

York, NY). 2010;17(5):955–61.

23. Ishii S, Miyao M, Mizuno Y, Tanaka-Ishikawa M, Akishita M, Ouchi Y. Association

between serum uric acid and lumbar spine bone mineral density in peri- and postmenopausal

This article is protected by copyright. All rights reserved.

 Japanese women. Osteoporosis international [Internet]. 2014;25:1099–105. Available from:

http://www.ncbi.nlm.nih.gov/pubmed/24318630
Accepted Article
24. Kang KY, Hong YS, Park S-H, Ju JH. Increased serum alkaline phosphatase levels correlate

with high disease activity and low bone mineral density in patients with axial

spondyloarthritis. Seminars in arthritis and rheumatism [Internet]. 2015;1–6. Available from:

http://www.ncbi.nlm.nih.gov/pubmed/25895696

25. Kaji H, Hattori S, Sekita K, Sugimoto T, Chihara K. Factors affecting bone mineral density

in hemodialysis patients with diabetic nephropathy. Endocrine journal [Internet].

2003;50(2):127–33. Available from: http://www.ncbi.nlm.nih.gov/pubmed/12803232

26. Kim E-H, Joh H-K, Kim E-Y, Cho D-Y, Kweon H-J, Choi J-K, et al. Biochemical Markers

and Health Behavior Related with Bone Mineral Density in Adult Men. Korean Journal of

Family Medicine [Internet]. 2009;30(5):359. Available from:

http://synapse.koreamed.org/DOIx.php?id=10.4082/kjfm.2009.30.5.359

27. Kim B-J, Baek S, Ahn SH, Kim SH, Jo M-W, Bae SJ, et al. Higher serum uric acid as a

protective factor against incident osteoporotic fractures in Korean men: a longitudinal study

using the National Claim Registry. Osteoporosis International [Internet]. 2014;25(7):1837–

44. Available from: http://link.springer.com/10.1007/s00198-014-2697-2

28. Lane NE, Parimi N, Lui L-Y, Wise BL, Yao W, Lay Y-AE, et al. Association of Serum Uric

Acid and Incident Nonspine Fractures in Elderly Men: The Osteoporotic Fractures in Men

(MrOS) Study. Journal of Bone and Mineral Research [Internet]. 2014;29(7):1701–7.

Available from: http://doi.wiley.com/10.1002/jbmr.2164

29. Maggio D, Barabani M, Pierandrei M, Polidori MC, Catani M, Mecocci P, et al. Marked

Decrease in Plasma Antioxidants in Aged Osteoporotic Women: Results of a Cross-Sectional

Study. The Journal of Clinical Endocrinology & Metabolism [Internet]. 2003;88(4):1523–7.

Available from: http://press.endocrine.org/doi/abs/10.1210/jc.2002-021496

30. Makovey J, Macara M, Chen JS, Hayward CS, March L, Seibel MJ, et al. Serum uric acid

This article is protected by copyright. All rights reserved.

 plays a protective role for bone loss in peri- and postmenopausal women: A longitudinal

study. Bone [Internet]. Elsevier Inc.; 2013;52(1):400–6. Available from:

Accepted Article
http://linkinghub.elsevier.com/retrieve/pii/S8756328212013336

31. Mehta T, Bůžková P, Sarnak MJ, Chonchol M, Cauley JA, Wallace E, et al. Serum urate

levels and the risk of hip fractures: data from the Cardiovascular Health Study. Metabolism

[Internet]. 2015;64(3):438–46. Available from:

http://linkinghub.elsevier.com/retrieve/pii/S0026049514003631

32. Nabipour I, Sambrook PN, Blyth FM, Janu MR, Waite LM, Naganathan V, et al. Serum uric

acid is associated with bone health in older men: A cross-sectional population-based study.

Journal of Bone and Mineral Research [Internet]. 2011;26(5):955–64. Available from:

http://doi.wiley.com/10.1002/jbmr.286

33. Sritara C, Ongphiphadhanakul B, Chailurkit L, Yamwong S, Ratanachaiwong W, Sritara P.

Serum Uric Acid Levels in Relation to Bone-Related Phenotypes in Men and Women.

Journal of Clinical Densitometry [Internet]. Elsevier Ltd; 2013;16(3):336–40. Available

from: http://linkinghub.elsevier.com/retrieve/pii/S1094695012000704

34. Veronese N, Bolzetta F, De Rui M, Maggi S, Noale M, Zambon S, et al. Serum uric acid and

incident osteoporotic fractures in old people: The PRO.V.A study. Bone [Internet]. United

States; 2015 Oct [cited 2015 Sep 1];79:183–9. Available from:

http://www.ncbi.nlm.nih.gov/pubmed/26079996

35. Zhang D, Bobulescu IA, Maalouf NM, Adams-Huet B, Poindexter J, Park S, et al.

Relationship Between Serum Uric Acid and Bone Mineral Density in the General Population

and in Rats With Experimental Hyperuricemia. Journal of Bone and Mineral Research

[Internet]. 2015;30(6):992–9. Available from: http://doi.wiley.com/10.1002/jbmr.2430

36. Liang D-K, Bai X-J, Wu B, Han L-L, Wang X-N, Yang J, et al. Associations Between Bone

Mineral Density and Subclinical Atherosclerosis: A Cross-Sectional Study of a Chinese

Population. The Journal of Clinical Endocrinology & Metabolism [Internet].

This article is protected by copyright. All rights reserved.

 2014;99(2):469–77. Available from: http://press.endocrine.org/doi/abs/10.1210/jc.2013-2572

37. Kuyumcu ME, Yesil Y, Oztürk ZA, Cinar E, Kizilarslanoglu C, Halil M, et al. The
Accepted Article
association between homocysteine (hcy) and serum natural antioxidants in elderly bone

mineral densitometry (BMD). Archives of Gerontology and Geriatrics. 2012;55(3):739–43.

38. Jang W, Kim HJ, Li H, Jo KD, Lee MK, Song SH, et al. 1,25-Dyhydroxyvitamin D₃

attenuates rotenone-induced neurotoxicity in SH-SY5Y cells through induction of autophagy.

Biochemical and biophysical research communications [Internet]. Academic Press Inc.; 2014

Aug 15 [cited 2015 Oct 7];451(1):142–7. Available from:

http://www.scopus.com/inward/record.url?eid=2-s2.0-84906217905&partnerID=tZOtx3y1

39. Lee YJ, Hong JY, Kim SC, Joo JK, Na YJ, Lee KS. The association between oxidative stress

and bone mineral density according to menopausal status of Korean women. Obstetrics &

Gynecology Science [Internet]. 2015 Jan [cited 2015 Oct 17];58(1):46. Available from:

http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=4303752&tool=pmcentrez&rend

ertype=abstract

40. Keizman D, Ish-Shalom M, Berliner S, Maimon N, Vered Y, Artamonov I, et al. Low uric

acid levels in serum of patients with ALS: further evidence for oxidative stress? J Neurol Sci

[Internet]. 2009;285(1-2):95–9. Available from:

http://www.ncbi.nlm.nih.gov/pubmed/19552925

41. Almeida M, Han L, Martin-Millan M, O&apos;Brien C a., Manolagas SC. Oxidative stress

antagonizes Wnt signaling in osteoblast precursors by diverting beta-catenin from T cell

factor- to forkhead box O-mediated transcription. Journal of Biological Chemistry.

2007;282(37):27298–305.

42. Dalbeth N, Smith T, Nicolson B, Clark B, Callon K, Naot D, et al. Enhanced

osteoclastogenesis in patients with tophaceous gout: urate crystals promote osteoclast

development through interactions with stromal cells. Arthritis and rheumatism [Internet].

2008;58(6):1854–65. Available from: http://www.ncbi.nlm.nih.gov/pubmed/18512794

This article is protected by copyright. All rights reserved.

 43. De Laet C, Kanis J a, Odén a, Johanson H, Johnell O, Delmas P, et al. Body mass index as a

predictor of fracture risk: a meta-analysis. Osteoporosis international : a journal established

Accepted Article
as result of cooperation between the European Foundation for Osteoporosis and the National

Osteoporosis Foundation of the USA [Internet]. 2005;16(11):1330–8. Available from:

http://www.ncbi.nlm.nih.gov/pubmed/15928804

44. Lawlor DA, Harbord RM, Sterne JAC, Timpson N, Smith GD. Mendelian randomization:

Using genes as instruments for making causal inferences in epidemiology. Statistics in

Medicine. 2008;27(8):1133–63.

LEGEND TO FIGURES

Figure 1. PRISMA flow-chart.

Figure 2. Meta-analysis and pooled standardized difference in means of bone mineral density at

different anatomical sites in participants included in cross-sectional studies (comparison between

highest and lowest quantiles of serum uric acid [SUA]).

This article is protected by copyright. All rights reserved.

 Table 1. Standardized mean differences in bone mineral density parameters in participants included

in cross-sectional studies (with higher vs. lower serum uric acid levels) in gender strata.
Accepted Article
Subjects Subjects
with with SMD
N of
Parameter higher lower (95% P-value Heterogeneity
studies
SUA SUA CI)
levels levels
Spine BMD
0.25
Tau² = 0.00; Chi² = 2.27, df = 2
Men 3 1934 2028 [0.18, <0.001
(P=0.32); I² = 12 %
0.32)
0.29
Tau² = 0.00; Chi² = 1.50, df = 2
Women 3 2887 2596 [0.24, <0.001
(P=0.47); I² = 0 %
0.34)
Total hip BMD
0.26
Tau² = 0.00; Chi² = 4.42, df = 3
Men 4 2427 2465 [0.19, <0.001
(P=0.22); I² = 32 %
0.33)
0.32
Tau² = 0.00; Chi² = 0.97, df = 2
Women 3 2928 2672 [0.27, <0.001
(P=0.62); I² = 0 %
0.37)
Femoral neck BMD
0.21
Tau² = 0.00; Chi² = 2.21, df = 2 (P
Men 3 2057 2193 [0.14, <0.001
=0.33); I² = 9 %
0.27)
0.43
Tau² = 0.00; Chi² = 4.12, df = 2
Women 3 2928 2672 [0.04, <0.001
(P=0.13); I² = 52%
0.81)

Abbreviations: df, degrees of freedom; BMD, bone mineral density; SMD, Standardized Mean
Difference; SUA, serum uric acid.

This article is protected by copyright. All rights reserved.

 Table 2. Simple correlations between serum uric acid levels and bone mineral density in cross-
sectional studies.
Accepted Article
N of r
Parameter P-value Heterogeneity
studies (95% CI)
0.141
Tau² = 0.002; Chi² = 29.37, df = 7 (P
Spine BMD 5 [0.099, <0.001
<0.001); I² = 76 %
0.182)
0.134
Tau² = 0.008; Chi² = 3.265, df = 1 (P =
Men 2 [-0.008, 0.06
0.07); I² = 69%
0.272)
0.111
Tau² = 0.011; Chi² = 4.577, df = 4 (P =
Women 4 [0.087, <0.001
0.33); I² = 12 %
0.134)
0.120
Tau² = 0.00; Chi² = 12.75, df = 8 (P = 0.12);
Total hip BMD 6 [0.095, <0.001
I² = 37 %
0.146)
0.135
Tau² = 0.00; Chi² = 0.30, df = 1 (P = 0.59);
Men 2 [0.085, <0.001
I² = 0 %
0.185)
0.106
Tau² = 0.001; Chi² =9.33, df = 5 (P = 0.10);
Women 5 [0.069, <0.0001
I² =46 %
0.143)
0.095
Femoral neck Tau² = 0.00; Chi² = 1.13, df = 3 (P = 0.77); I²
4 [0.077, <0.001
BMD =0%
0.113)
Men No studies available
0.089
Tau² = 0.00; Chi² = 0.00, df = 1 (P = 0.98);
Women 2 [0.067, <0.001
I² = 0 %
0.111)

Abbreviations: df, degrees of freedom; BMD, bone mineral density; r, Pearson’s correlation
coefficient; SMD, standardized mean difference; SUA, serum uric acid.

This article is protected by copyright. All rights reserved.

 Table 3. Risk ratio and adjusted odds ratio for prevalent osteoporosis in cross-sectional studies by
serum uric acid level.
Higher SUA Lower SUA Risk
Accepted Article
levels levels ratio/
Stu P-
Odds
Parameter dies valu Heterogeneity
Participants Participants ratio
(N) e
(events) (events) (95%
CI)

0.69
Presence of <0.0 Tau² = 0.00; Chi² = 1.92, df = 3 (P =
4 [0.62,
osteoporosis 3580 (486) 3418 (661) 01 0.59); I² = 0 %
0.76)
0.74
0.00 Tau² = 0.00; Chi² = 1.51, df = 2 (P =
Men 3 [0.59,
1327 (112) 1466 (166) 9 0.47); I² = 0 %
0.93)
0.68
<0.0 Tau² = 0.00; Chi² = 0.75, df = 1 (P =
Women 2 [0.61,
2253 (374) 1952 (495) 01 0.38); I² = 0 %
0.76)
0.66
Adjusted OR <0.0 Tau² = 0.06; Chi² = 29.99, df = 8 (P
6 [0.53,
for osteoporosis 01 <0.001); I² = 73 %
0.83)
0.56
Tau² = 0.21; Chi² = 16.59, df =5 (P
Men 3 [0.36, 0.02
= 0.005); I² =70 %
0.90)
0.69
<0.0 Tau² = 0.00; Chi² = 0.62, df = 1 (P =
Women 2 [0.58,
01 0.42); I² = 0 %
0.83)

Abbreviations: df, degrees of freedom; BMD, bone mineral density; OR, odds ratio; SMD,
standardized mean difference; SUA, serum uric acid.

This article is protected by copyright. All rights reserved.

 Table 4. Association between serum uric acid levels and incident fractures (prospective studies).

N of HR
Accepted Article
N of P-
Parameter participants (95% Heterogeneity
Studies value
CI)
0.83
Per increase of one Tau² = 0.00; Chi² = 0.02, df = 3 (P
3 23658 [0.74, 0.001
standard deviation = 0.99); I² = 0 %
0.92)
0.82
Tau² = 0.00; Chi² = 0.04, df = 3 (P
Men 3 22691 [0.74, 0.001
= 0.99); I² = 0 %
0.92)
Women Only one study available
Groups with lower 0.90
Tau² = 0.09; Chi² = 10.00, df = 4 (P
SUA levels taken 3 12272 [0.64, 0.56
= 0.04); I² = 60 %
for reference 1.27)
0.72
Tau² = 0.27; Chi² = 10.81, df =3
Men 3 8576 [0.39, 0.29
(P = 0.01); I² =73 %
1.33)
1.05
Tau² = 0.00; Chi² = 0.24, df = 1 (P
Women 2 3696 [0.79, 0.74
= 0.42); I² = 0 %
1.40)

Abbreviations: df, degrees of freedom; BMD, bone mineral density; HR, hazard ratio; SMD,
standardized mean difference; SUA, serum uric acid.

This article is protected by copyright. All rights reserved.

 Table 5. Spine, total hip and femoral neck bone mineral density in cross-sectional studies stratified
by score on Newcastle-Ottawa Scale.
Accepted Article
Outcome
Effect Spine Total hip Femoral neck
Strata Analysis details
estimate BMD BMD BMD
Standardized NOS Pooled estimate, 0.37 (0.30;
mean score <5 (95%CI) 0.44)
difference P-value <0.001
Heterogeneity, I2 0% (0.72)
(P-value) 2
NOS Number of 0.20 (0.15;
score >5 studies 0.25)
No No
Pooled estimate, <0.001
heterogeneity heterogeneity
(95%CI) 0% (54%)
P-value 4
Heterogeneity, I2 <0.001
(P-value)
Number of
studies
P-valuea
Pearson’s NOS Pooled estimate, 0.17 (0.11-
correlation score <5 (95%CI) 0.23)
coefficient P-value <0.001
Heterogeneity, I2 74% (0.004)
(P-value) 3
Number of 0.11 (0.09-
NOS studies 0.13)
No No
score >5 Pooled estimate, <0.001
heterogeneity heterogeneity
(95%CI) 0% (0.64)
P-value 2
Heterogeneity, I2 0.06
(P-value)
Number of
studies
P-valuea

Abbreviations: CI: confidence interval; df, degrees of freedom; BMD, bone mineral density; NOS:
Newcastle Ottawa Scale; SMD, standardized mean difference; SUA, serum uric acid.
a
The P-value for the t-test between the two statistical analysis strata using a meta-regression
procedure. Significant results (p value <0.05) are in bold.

This article is protected by copyright. All rights reserved.

 Identification

Records identified through database Additional records identified

Accepted Article
searching through other sources
(n = 1405 ) (n = 0 )

Records excluded
(n = 1097 )

Records after duplicates removed

No data regarding bone
(n =1127 )
outcomes (n=639)
Screening

No data regarding uric

acid (n=358)
Reviews (n=89)
Records screened Case reports (n=11)
(n = 1127 )

Full-text articles assessed Full-text articles excluded,

for eligibility with reasons
(n = 30 ) (n = 11 )
Eligibility

No data about the

association between uric
Studies included in acid and bone estimates
qualitative synthesis (n=11)
(n = 19 )
Included

Studies included in
quantitative synthesis
(meta-analysis)
(n = 16 )

This article is protected by copyright. All rights reserved.

Accepted Article

This article is protected by copyright. All rights reserved.