Bone mineral density in adults with osteoarthritis or rheumatoid arthritis: a cross-sectional study of a nationwide population.


 Background

It is reported that osteoporosis commonly occurs in rheumatoid arthritis (RA), whereas there is a controversial relationship with osteoarthritis (OA). In this study, we aimed to investigate the relations between OA, RA and bone mineral density (BMD) in adults aged 20–59 years.
Methods

We merged and analyzed the database from the National Health and Nutrition Examination Survey (NHANES) 2011–2018. Arthritis status and types of arthritis were obtained from questionnaires. Lumbar BMD was measured by dual-energy X-ray absorptiometry. The associations between OA, RA and lumbar BMD were evaluated using logistic regression models. Subgroup analyses stratified by gender and race were performed. The association between disease duration of arthritis and lumbar BMD was also investigated.
Results

A total of 11094 adults aged 20–59 years were analyzed in this study. Compared with the non-arthritis group, participants with OA had a higher lumbar BMD (β = 0.023; 95% CI: 0.011–0.035), while no significant association was observed in rheumatoid arthritis (β = 0.014; 95% CI: -0.003–0.031). In the subgroup analyses stratified by gender, males with OA had a higher lumbar BMD compared with those without (β = 0.014; 95% CI: -0.003–0.031). However, this association changed in females (β = 0.007; 95% CI: -0.008–0.021). On the other hand, no significant association was observed in RA in both males and females (males: β = 0.023; 95% CI: -0.003–0.048, females: β = 0.008; 95% CI: -0.015–0.031). Moreover, the disease duration of arthritis was not associated with lumbar BMD in both OA (β=-0.0001; 95% CI: -0.0017–0.0015) and RA (β = 0.0006; 95% CI: -0.0012–0.0025).
Conclusions

The correlation between OA and lumbar BMD differed by sex. Patients with OA were more likely to have higher lumbar BMD in males, but not in females. On the other hand, no significant association was found in RA both in males and females.


Introduction
Osteoporosis (OP), osteoarthritis (OA), and rheumatoid arthritis (RA) are a series of disorders called musculoskeletal pathologies that cause intolerable pains, movement disorders, and even permanent disability [1]. With the aging of the social population, these diseases have reached about 1 in 4 adults in developed countries [2,3]. RA is an autoimmune disease of unknown etiology with a homeostatic imbalance [4]. As a biomechanical disease, the onset and development of OA are closely related to in ammatory and catabolic alterations [5]. OP is a systemic disease associated with a dramatic loss of bone mineral density (BMD). As two notable silent rheumatic diseases, OA and OP have been included on the World Health Organization disease-disabling list [1,6].
It is reported that OP commonly occurs in RA, whereas there is a controversial relationship with OA [7][8][9][10].
In this study, we therefore investigated the relations between OA, RA and BMD in adults aged 20-59 years using a population-based sample from the National Health and Nutrition Examination Survey (NHANES) database.

Study population
The NHANES program is a series of surveys focusing on health topics of US population of all ages. The data collection and analysis require special statistical analysis due to the multistage, complex clustered probability design of the survey, rather than based on a simple random sample of the US population [11].
The current study used the data from NHANES 2011-2018. The study population was restricted to adults aged 20-59 years (enrollment from NHANES 2011-2018, n = 14934). Participants without data on arthritis diagnosis (n = 23), lumbar BMD (n = 3398), or with cancer (n = 419) were excluded. Finally, a total of 11094 subjects were analyzed in this study. The ethics review board of the National Center for Health Statistics (NCHS) approved all protocols, and written informed consent was obtained from each participant [12] Arthritis A diagnosis of arthritis was based on the medical conditions questionnaires collected by NHANES during the household interviews. Arthritis was classi ed as OA, RA, psoriatic arthritis, other, and don't know (or refused) if the participants reported ever being told they had arthritis by a doctor or other health professional. We de ned the disease duration of arthritis as age in years at screening minus age when the doctor told the participants they had arthritis.

Lumbar BMD
Lumbar spine is one site typically evaluated for the assessment and treatment of osteoporosis, and BMD measurement at this site has been used as a clinical trial outcome for over three decades [13]. As the most widely accepted method, the dual-energy X-ray absorptiometry (DXA) scans provide bone measurements for lumbar spine. In NHANES program, the scans were acquired on the Hologic Discovery model A densitometers (Hologic, Inc., Bedford, Massachusetts), and were analyzed with Hologic APEX version 4.0 software. Further details of the DXA examination protocol are located on the NHANES website.
Demographic data and other questionnaire data Age, gender, race, educational level, ratio of family income to poverty, vigorous recreational activities, and smoked at least 100 cigarettes in life were recorded using questionnaires managed by intervieweradministers. Race was self-reported as Non-Hispanic White, Non-Hispanic Black, Mexican American, Other Hispanic, and Other race-including multi-racial. The description of vigorous recreational activities was based on the patient self-report described as the following question: 'Do you do any vigorous-intensity sports, tness, or recreational activities that cause large increases in breathing or heart rate like running or basketball for at least 10 minutes continuously? ' Laboratory data Biospecimens were collected for laboratory analysis to provide detailed information on the nutritional status and health of the participants. The collecting, processing, storing, and shipping of the biospecimens took place in the mobile examination center. The detailed measurement processes of blood urea nitrogen, total protein, total cholesterol, alkaline phosphatase, serum potassium, serum sodium, serum phosphorus, serum uric acid, and serum calcium can be found at the NHANES website.

Statistical analysis
To assure national representation, we used weighted analysis recommended by the analytical guideline edited by NCHS. P-value was calculated using a weighted chi-square test for categorical variables and using a weighted linear regression model for continuous variables.
The relations between arthritis and lumbar BMD were examined using multivariable logistic regression.
Three models were constructed: model 1, no covariates were adjusted; model 2, age, gender, race were adjusted; model 3, age, gender, race, educational level, body mass index, ratio of family income to poverty, vigorous recreational activities, smoked at least 100 cigarettes in life, blood urea nitrogen, total protein, total cholesterol, alkaline phosphatase, serum potassium, serum sodium, serum phosphorus, serum uric acid, and serum calcium were adjusted. Besides, following the STROBE guideline [14], we performed subgroup analyses to make better use of the data. We also performed multivariable logistic regression to explore the association between disease duration of arthritis and lumbar BMD. We performed all analyses using package R version 3.4.3 (http://www.R-project.org) and EmpowerStats software (http://www.empowerstats.com). The signi cance level was set to 0.05.

Study sample
The characteristics of the samples are presented in Table 1. The mean age of the participants with arthritis (47.99 ± 9.36 years) was signi cantly higher than the mean age of those without arthritis (37.59 ± 11.34 years). Compared with the non-arthritis group, arthritis group had a higher proportion of females (56.08% vs. 46.17%). Race, educational level, body mass index, vigorous recreational activities, smoked at least 100 cigarettes in life, blood urea nitrogen, total protein, total cholesterol, alkaline phosphatase, serum uric acid, serum sodium, and serum calcium were also signi cantly different between the two groups (p < 0.05).

Subgroup analyses
In the subgroup analyses strati ed by gender (Table 3), male adults with osteoarthritis or degenerative arthritis had a higher lumbar BMD compared with those without (β = 0.014; 95% CI: -0.003-0.031).
In the subgroup analyses strati ed by race (Table 4), non-Hispanic White adults with osteoarthritis or degenerative arthritis had a higher lumbar BMD compared with those without (β = 0.014; 95% CI: -0.003-0.031). no signi cant associations were observed in other osteoarthritis or degenerative arthritis groups or any RA groups.
Associations between disease duration of arthritis and lumbar BMD.

Discussion
The present study demonstrated that patients with OA were more likely to have higher lumbar BMD in males, but not in females. On the other hand, no signi cant association was found in RA in both males and females.
Despite years of research, the association of OP with OA is still discussed. Both diseases depend on bone metabolism and positively correlated with aging. In a study comprised 359 postmenopausal women aged 50-89 years, Povoroznyuk et al [15] found that women with a symptomatic OA had a signi cantly higher lumbar BMD compared with controls. The cross-sectional data of a Korean national survey found a negative association of lumbar BMD with the presence of knee OA [16]. A recent prospective study provided strong evidence that high femoral neck BMD is a prognostic risk factor for the development of knee and hip radiographic OA [17]. In addition, higher BMD has been shown to reduce the risk of fractures in both men and women [18,19] [26]. In this study, we found the correlation between OA and lumbar BMD differed by sex. One possible explanation is that high BMI and weight-bearing activities, which increase the risk of damage to articular cartilage leading to OA, are also bene cial to the preservation of bone mass. The strengths of this study include a population-based sample with a wide age range that is generalizable to a community population, subgroup analyses for sensitivity test, and adjustment for many potential confounders. There are also several limitations. First, due to the cross-sectional nature of this study, we were unable to elucidate the causal relationship between arthritis and BMD. More longitudinal studies investigating the causality between them are needed. Second, the diagnosis of arthritis was based on the patient's self-report that might be subject to bias. However, the consistency of self-reported arthritis and clinical con rmation has been documented [32,33]. Third, the missing information on different sites of arthritis preclude us to estimate the associations between OA, RA and BMD in speci c sites. Fourth, study results cannot be generalized to the participants with cancer because these special populations were excluded in the analyses. Finally, there might be other confounding factors we did not control for in our study. For example, glucocorticoids used for the treatment of RA were not adjusted in this study. However, it was reported that there was no signi cant association between cumulative glucocorticoid dose and BMD after adjustment of confounders [31]. The results of a population-based study also showed no signi cant difference between corticosteroid treated patients with RA and non-steroid group, indicating that the independent effect of corticosteroids on BMD is only minimal[34].
In summary, our results indicated that there is an inverse relationship between OP and the presence of OA in males, but not in females. There was no signi cant association between RA and lumbar BMD in both males and females. Our ndings may provide some important implications for better understanding the linking between OA, RA and bone health.

Declarations
Author contributions ZXZ, HTJ contributed to data collection, analysis and writing of the manuscript. PJT contributed to study design and writing of the manuscript.

Funding
This study received no funding.

Consent for publication
Not applicable.

Competing interests
The authors declare that they have no competing interests.

Ethical Statement
The ethics review board of the National Center for Health Statistics approved all NHANES protocols and written informed consents were obtained from all participants. Tables Table 1 Characteristic of study sample with and without arthritis.     Associations were adjusted for age, gender, educational level, body mass index, ratio of family income to poverty, vigorous recreational activities, smoked at least 100 cigarettes in life, blood urea nitrogen, total protein, total cholesterol, alkaline phosphatase, serum uric acid, serum sodium, serum potassium, serum phosphorus, and serum calcium.