Various Types of Arthritis in the United States: Prevalence and Age-Related Trends From 1999 to 2014

Abstract

Objectives. To determine the prevalence trends of osteoarthritis (OA), rheumatoid arthritis (RA), and other types of arthritis in the United States from 1999 to 2014.

Methods. We analyzed data on 43 706 community-dwelling adults aged 20 years and older who participated in the 1999–2014 National Health and Nutrition Examination Surveys. We accounted for survey design and sampling weights so that estimates were nationally representative. We assessed temporal trends in age-standardized arthritis prevalence by using joinpoint regression.

Results. Age-adjusted prevalence of arthritis was 24.7% (OA = 9.7%; RA = 4.2%; other arthritis = 2.8%; “don’t know” type = 8.0%). Prevalence of OA increased from 6.6% to 14.3%, whereas RA prevalence decreased from 5.9% to 3.8%. Increase in OA prevalence was significant in both genders; in non-Hispanic Whites, non-Hispanic Blacks, and Hispanics; and in people with high socioeconomic status. Decrease in RA prevalence was more pronounced in men, non-Hispanic Blacks, and participants with low income or obesity.

Conclusions. Between 1999 and 2014, nearly one quarter of American adults reported arthritis. The prevalence of OA has more than doubled over time, whereas RA prevalence has declined.

Arthritis is a leading cause of disability and pain in the United States.¹ In 2010 to 2012, 52.5 million Americans (22.7%) aged 18 years and older reported arthritis, including osteoarthritis (OA), rheumatoid arthritis (RA), or other types, such as psoriatic arthritis.² Arthritis-related joint pain limits functional ability and quality of life.³ Osteoarthritis, the most common type of arthritis,² often affects knees, hips, and lower back.⁴ Rheumatoid arthritis, a chronic autoimmune inflammatory disease of synovial tissues and the most common type of autoimmune arthritis causes joint pain, stiffness, swelling, and decreased joint movement, leading to structural damage, deformity, and disability; 75% of those affected are women.⁵ Arthritis prevalence is expected to increase.⁶ Given the health and economic burden of arthritis, understanding prevalence trends is of substantial public health interest.³ We examined prevalence and age-related trends of various types of arthritis in the United States from 1999 to 2014.

METHODS

Data came from the 1999–2014 National Health and Nutrition Examination Surveys (NHANES).⁷ This continuous cross-sectional survey of noninstitutionalized civilians uses complex multistaging to derive a representative sample of the US population.

A total of 43 790 people aged 20 years and older participated in the 1999–2014 NHANES; we excluded 84 who declined to answer the question on arthritis or did not know whether they had ever been diagnosed with arthritis, leaving 43 706 in this study. Arthritis was defined by self-report (“Has a doctor or other health professional ever told you that you have arthritis?”) and classified as OA, RA, other, or “don’t know” on the basis of the response to the question “Which type of arthritis was it?” The survey also collected data on age, gender, race/ethnicity, and poverty–income ratio. The NHANES examiners measured weight and height and calculated body mass index (defined as weight in kilograms divided by the square of height in meters). We classified as obese participants with a body mass index greater than or equal to 30.

We combined 2-year data cycles as well as crude and calculated age-adjusted prevalences. Age standardization allowed comparisons independent of age by using age groups and weights based on the 2010 Census data (20–39 years, weight 0.3631; 40–59 years, weight 0.37629; ≥ 60 years, weight 0.26057). We assessed temporal trends in age-standardized arthritis prevalence by using Joinpoint Regression Program 4.3.1.0 (National Cancer Institute, Bethesda, MD). We allowed a maximum of 1 joinpoint and used the Bayesian information criterion method for model selection. The analysis compared models with 0 and 1 joinpoint to determine the most parsimonious model to fit the annual percent changes (APC). We performed analyses in SAS (version 9.4, SAS Institute, Cary, NC) and we considered P < .05 statistically significant. We calculated standard errors, confidence intervals, and P values in accordance with the complex survey design by using Taylor series linearization methods.

RESULTS

The proportion of the 43 706 participants with any arthritis, after we accounted for study design and applied sampling weights, was 23.7% (RA = 4.1%; OA = 9.2%; other = 2.7%; and “don’t know” = 7.7%). Most participants with RA were Hispanic or non-Hispanic Black, had low family income, and had high-school or some college education. OA was more prevalent in older (≥ 60 years) non-Hispanic White women with a high family income or a college degree, and in obese participants. Other arthritis was more common in younger (< 60 years) men.

Age-standardized prevalence of any arthritis was 24.7% and did not change significantly over time. We saw decreasing trends in participants in the “other” racial/ethnic group or those who were obese (Table A, available as a supplement to the online version of this article at http://www.ajph.org).

Age-standardized prevalence of RA was 4.2%, decreasing from 5.9% in 1999–2000 to 3.8% in 2013–2014 (APC = −5.20; P < .01; Figure 1). The decrease was pronounced in men, non-Hispanic Blacks, participants with poverty–income ratio 1 to 3, and obese participants (Table B, available as a supplement to the online version of this article at http://www.ajph.org).

FIGURE 1—

Age-Standardized Prevalence of Overall Arthritis, Rheumatoid Arthritis (RA), Osteoarthritis (OA), Other, and “Don’t Know” Type of Arthritis: United States, 1999–2014

Note. There was a significant decreasing trend in prevalence of RA (annual percent change [APC] = −5.20; P < .01) and an increasing trend in prevalence of OA (APC = 8.80; P < .01). The prevalence of participants who did not know their type of arthritis decreased (APC = −6.77; P < .05).

Age-standardized prevalence of OA was 9.7%, increasing from 6.6% in 1999–2000 to 14.3% in 2013–2014 (APC = 8.80; P < .01; Figure 1). The increase was significant in both genders; in non-Hispanic Whites, non-Hispanic Blacks, and Hispanics; in people with high socioeconomic status; and in those with a high-school diploma or higher level of education (Table C, available as a supplement to the online version of this article at http://www.ajph.org).

Age-standardized prevalence of other arthritis was 2.8%; it did not change significantly over time (Figure 1), except for a significant decrease among obese participants (Table D, available as a supplement to the online version of this article at http://www.ajph.org).

Age-standardized prevalence of participants who did not know their type of arthritis was 8.0%, decreasing from 9.9% in 1999–2000 to 5.5% in 2013–2014 (APC = −6.77; P < .05; Figure 1). The decrease was significant in women, non-Hispanic Whites, people with a low income (poverty–income ratio ≤ 1), those with a high-school level of education, and obese people (Table E, available as a supplement to the online version of this article at http://www.ajph.org). There were no significant changes in median age of arthritis diagnosis over time.

DISCUSSION

Contrary to the current literature,⁸ Hispanics and non-Hispanic Blacks were less likely than non-Hispanic Whites to have arthritis, possibly attributable to lack of adequate health care for correct diagnosis. Non-Hispanic Blacks and Hispanics are more likely than non-Hispanic Whites to use self-care or alternative or complementary medicine and less likely to receive traditional medical care. OA was more common in older White women.⁹ Age-related degeneration and hormonal changes (lower estrogen levels associated with increased risk for OA pain in women) may explain this finding.⁹

A previous study projected that by 2040, 26% of US adults will have arthritis.² If one considers the prevalence reported in the current study, this prediction seems conservative. Compared with 22.7% in 2010 to 2012,² prevalence increased to 26.3% in 2013 to 2014. In the 16-year period, age-adjusted prevalence of RA decreased overall. It is plausible that introduction of effective drugs improved treatment regimens in the past 20 years¹⁰ and decreased RA prevalence. For instance, prevalence of patients who have taken glucocorticoid therapy within 1 year after onset of RA increased. The finding may be also associated with the reduced frequency of joint replacements and other joint operations related to RA.¹¹ However, during the same period, age-adjusted prevalence of OA increased in almost all subgroups. The increase in OA with age is a consequence of cumulative exposure to risk factors and biological changes such as oxidative damage, thinning of cartilage, or muscle weakness.

The prevalence of obese people with arthritis decreased in this study. Although obesity has been recognized in the literature as a risk factor for arthritis,¹² the prevalence of obese people with all types of arthritis has decreased significantly. These findings reflect that average body mass indexes were higher in recent years than in early NHANES surveys.

Limitations

We defined arthritis by self-report of doctor diagnosis, so inaccurate recall may have biased prevalence estimates and led to potential misclassifications. However, self-report of arthritis has been reported to have acceptable validity and reliability for large epidemiological studies.²

Public Health Implications

The NHANES data regarding prevalence of various types of arthritis can be used to inform health care policy and reform and provide resources to reduce risk factors. Because of the public health burden associated with arthritis, cost-saving and effective treatments are necessary to minimize arthritis symptoms, maximize functional capacity, and reduce disability.