The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Department of the Army or the Department of Defense. All authors are employees of the United States government, and this work was prepared as part of their official duties.
The incidence of primary hip osteoarthritis in active duty US military servicemembers†
Article first published online: 30 MAR 2009
Copyright © 2009 by the American College of Rheumatology
Arthritis Care & Research
Volume 61, Issue 4, pages 468–475, 15 April 2009
How to Cite
Scher, D. L., Belmont, P. J., Mountcastle, S. and Owens, B. D. (2009), The incidence of primary hip osteoarthritis in active duty US military servicemembers. Arthritis & Rheumatism, 61: 468–475. doi: 10.1002/art.24429
- Issue published online: 30 MAR 2009
- Article first published online: 30 MAR 2009
- Manuscript Accepted: 7 JAN 2009
- Manuscript Received: 6 AUG 2008
Although multiple studies have reported the prevalence of primary hip osteoarthritis (OA), little has been reported on incidence rates of hip OA. We sought to determine the incidence rate and demographic risk factors of hip OA in an ethnically diverse and physically active population of US military servicemembers.
A query was performed using the US Defense Medical Epidemiology Database for the International Classification of Diseases, Ninth Revision, Clinical Modification code for hip OA (715.95). Multivariate Poisson regression analysis was used to estimate the rate of hip OA per 100,000 person-years, controlling for sex, race, age, rank, and service.
The overall unadjusted incidence rate of hip OA was 35 per 100,000 person-years. Women, compared with men, had a significantly increased adjusted incidence rate ratio for hip OA of 1.87 (95% confidence interval [95% CI] 1.73–2.01). The adjusted incidence rate ratio for black servicemembers when compared with white servicemembers was 1.32 (95% CI 1.23–1.41). The adjusted incidence rate ratio for the ≥40-year-old age group compared with the 20-year-old group was 22.21 (95% CI 17.54–28.14). With junior officers as the referent category, junior enlisted, senior enlisted, and senior officers rank groups had a significantly increased adjusted incidence rate ratio for hip OA. With the Air Force as the referent category, each service had a significantly increased adjusted incidence rate ratio for hip OA.
Female sex; black race; age ≥40 years; junior enlisted, senior enlisted, and senior officer rank groups; and military service in the Navy, Army, or Marines were all risk factors for hip OA.
Osteoarthritis (OA) is a progressive disease caused by destruction of articular cartilage and proliferation and remodeling of subchondral bone. Symptomatic OA is a clinical diagnosis based on pain in the presence of particular radiographic features: osteophytes, joint space narrowing, subchondral sclerosis, and cysts. It is estimated that OA affects 27 million Americans (1), accounting for 55% of all arthritis-related hospitalizations (2). OA has been found to have average direct yearly costs between $2,650 and $5,700 per person (3, 4). More than half of the societal costs of OA are lost productive work time, which has been estimated at $7.1 billion (5, 6).
Primary hip OA is defined as hip OA in which there is no preexisting anatomic abnormality or disease process. Previous large-scale studies have found the prevalence of radiographic hip OA in US men and women older than 55 years to vary by nearly a factor of 10, between a low prevalence of 2.7%, reported in the first National Health and Nutrition Examination Survey, up to a high prevalence of 25.1%, reported in the Johnston County Osteoarthritis Project (7, 8). Despite the variance in prevalence in these studies, European studies have shown that the prevalence of radiographic hip OA has remained constant throughout the past 4 decades within an urban population (9).
The majority of studies describing the frequency of hip OA have been prevalence studies. Previous studies have suggested that patients with OA have higher mortality rates than patients without OA (10, 11). It may be hypothesized that this increase in mortality in patients with symptomatic hip OA may be attributable to its correlation with obesity (12–15) or to the secondary functional and physical limitations that can exacerbate other chronic medical comorbidities, such as hypertension and cardiovascular disease. Incidence rather than prevalence studies of hip OA are unaffected by the difference in mortality rates, and incidence was therefore determined to be a more appropriate tool to study the demographic risk factors, such as sex and race.
Calculating the incidence rate of hip OA is challenging because of the difficulty in defining the point of disease onset. The incidence of hip OA can also be difficult to quantify because the structural changes of the disease occur in most people as they get older, but these changes may not be accompanied by symptoms that warrant medical attention (14). Furthermore, these incidence rates may vary, depending on whether only moderate and severe radiographic changes are studied or whether mild changes are also included.
To better understand the population-based epidemiologic characteristics of hip OA, numerous studies have sought to determine the risk factors for this disease. In particular, prevalence with respect to sex, race, age, and activity level has been extensively studied, but only 2 studies have described incidence rates for hip OA (16, 17). The incidence rates and epidemiologic variables in US Armed Forces servicemembers have not been reported (7) and are the subject of this study.
MATERIALS AND METHODS
The US Armed Forces represents a physically active population of male and female servicemembers with generally high occupational demands. These servicemembers participate in organized physical fitness training programs, and they must meet the standards of their individual service's physical fitness test and height/weight requirements semiannually. At the initial military entry examination, preexisting hip conditions are screened by history and physical examination. If there are any concerns about a preexisting hip condition, radiographs and an orthopedic consultation are obtained. Findings of any preexisting physical abnormalities, including potential causes of secondary hip OA such as developmental dysplasia of the hip, Legg-Calvé-Perthes disease, slipped capital femoral epiphysis, hip dislocation within 2 years preceding examination, or chronic OA excludes those individuals from initial entry into active duty military service.
The military maintains large medical databases, making it an excellent population in which to study musculoskeletal disorders such as hip OA. One such database is the Defense Medical Epidemiology Database (DMED), which compiles International Classification of Diseases, Ninth Revision (ICD-9) coding information for every patient encounter in a military treatment facility. This database also maintains the total number of servicemembers on active duty each year and contains patient demographic and military-specific data. The DMED is a frequently updated database that is able to track military servicemembers as they move throughout the world. The DMED also adjusts for servicemembers as they retire and as they enlist into the Armed Forces. It has been used previously to provide information on various musculoskeletal conditions (18, 19).
The DMED provides 4 types of data: demographic features, inpatient hospitalizations, ambulatory visits, and reportable events. The outpatient data in the DMED is a combination of the standard ambulatory data records extracted from the Ambulatory Data System, from the Composite Health Care System used in military treatment facilities worldwide, and from outsourced (nonmilitary) outpatient health care facilities providing care to active-duty servicemembers.
To determine the total number of patients with hip OA, we queried the ambulatory DMED system for the years 1998–2006 using the ICD-9 Clinical Modification code 715.95. We chose to include only outpatient data because hip OA is primarily an outpatient diagnosis, and the small numbers from inpatient data would not have been significant. Additionally, inpatient data is less likely to represent first time diagnosis, and its inclusion could lead to multiple counting of primary hip OA. Ambulatory encounters were limited to a first occurrence to exclude repeat coding of the same initial diagnosis for all services during the study period. We then categorized the results by sex, race, age, rank, and military service. Race data is routinely obtained from the Defense Manpower Data Center, which compiles servicemembers' self-report of race with the following options: white, black, Hispanic, Alaskan Native/American Indian, Asian/Pacific Islander, and other. The DMED classifies these categories into 3 larger groups: white, black, and other. Mixed-race individuals were classified according to self-report. The age categories used were <20 years, 20–24 years, 25–29 years, 30–34 years, 35–39 years, and ≥40 years. The rank categories used were junior enlisted (E1–E4), senior enlisted (E5–E9), junior officers (O1–O3), and senior officers (O4–O9). The military service categories used were Army, Navy, Air Force, and Marines. The DMED does not compile height or weight data on servicemembers and so this information was not available for analysis. The database was also queried for the total number of servicemembers on active duty during the study time period, and the results were categorized by sex, race, age, rank, and service. To estimate incidence, 1 exposure year was defined as 1 year that the servicemember was in the US Armed Forces.
For incident OA, the outcome measure was the unadjusted incidence rate per 100,000 person-years. We used multivariate Poisson regression to estimate the rate of hip OA per 100,000 person-years by sex, race, age, rank, and service (unadjusted rates). Additionally, using Poisson regression, we computed rate ratios for sex, using men as the referent, and controlling for differences in race, age, rank, and service between men and women (adjusted rates). Rate ratios were also calculated for race, using white as the referent category; age, using <20 years as the referent category; rank, using junior officers as the referent category; and service, using the Air Force as the referent category, all of which were adjusted for other covariates. All statistical analysis was performed using SAS software, version 9.1 (SAS, Cary, NC). This study received Institutional Review Board approval from William Beaumont Army Medical Center Department of Clinical Investigation Protocol.
A total of 4,262 cases of hip OA were documented in our population at risk of 12,096,304 person-years. The overall incidence rate of hip OA in our population was 35 cases per 100,000 person-years. The unadjusted incidence rate of hip OA was 54 cases per 100,000 person-years among women and 32 cases per 100,000 person-years among men. Women, when compared with men, had a significantly increased adjusted incidence rate ratio for hip OA of 1.87 (95% confidence interval [95% CI] 1.73–2.01). When examining the sex/age subcategories, it was found that the adjusted incidence rate ratio for women, when compared with men, was statistically significant at all age subcategories. The incidence rates for men and women by age group are listed in Table 1.
|Age group, years||Women||Men||Adjusted incidence rate (95% CI)‡|
|Cases||Person-years||Unadjusted incidence rate†||Cases||Person-years||Unadjusted incidence rate†|
The unadjusted incidence rate for hip OA per 100,000 person-years was 33 among whites, 51 among blacks, and 21 among others. Blacks, when compared with whites, had a significantly increased adjusted incidence rate ratio for hip OA of 1.32 (95% CI 1.23–1.41). When examining the racial age subcategories, it was found that the adjusted incidence rate ratio for blacks, when compared with whites, became statistically significant at the age subcategory of 25–29 years and each older subcategory thereafter. The incidence rates for race are presented in Table 2.
|Age group, years||Cases||Person-years||Unadjusted incidence rate†||Adjusted incidence rate ratio (95% CI)‡|
We found that older servicemembers had a higher incidence rate for hip OA when compared with younger servicemembers. The highest incidence rate was seen in the ≥40 years age group with an incidence rate of 140 per 100,000 person-years. The adjusted rate ratio for the ≥40 years age group compared with the <20 years age group was 22.21 (95% CI 17.54–28.14). A complete listing of the age data is shown in Table 3.
|Age group, years||Incidence rate†||Adjusted incidence rate ratio (95% CI)‡|
The unadjusted incidence rate for hip OA among the 4 rank groups was 16.6 for the junior enlisted group, 24.8 for junior officers, 46.4 for the senior enlisted group, and 112 for the senior officers per 100,000 person-years. When compared with the junior officers as the referent category, all of the other rank groups had increased adjusted incidence rate ratios for hip OA: junior enlisted group 1.90 (95% CI 1.63–2.22), senior enlisted group 1.31 (95% CI 1.16–1.48), and senior officers 1.50 (95% CI 1.31–1.73). The incidence rates for rank groups are presented in Table 4.
|Age group, years||Cases||Person-years||Unadjusted incidence rate†||Adjusted incidence rate ratio (95% CI)‡|
The unadjusted incidence rate for hip OA among the 4 services was 43.3 for the Army, 37.0 for the Navy, 26.8 for the Air Force, and 26.3 for the Marines per 100,000 person-years. Each service, when compared with the Air Force as the referent category, had a significantly increased adjusted incidence rate ratio for hip OA: Army 1.90 (95% CI 1.75–2.07), Navy 1.63 (95% CI 1.49–1.78), and Marines 1.85 (95% CI 1.64–2.09). We found that the adjusted incidence rate ratio for all services, when compared with the Air Force, was statistically significant at all the age subcategories. The incidence rates for service are presented in Table 5.
|Age group, years||Cases||Person-years||Unadjusted incidence rate†||Adjusted incidence rate ratio (95% CI)‡|
In this large-scale database study, we found that female sex; black race; age ≥40 years; junior enlisted, senior enlisted, and senior officer rank groups; and military service in the Navy, Army, or Marines were significant risk factors for the development of incident hip OA when compared with male sex, white race, age <40 years, junior enlisted rank group, and service in the Air Force. Incidence rates were determined for this unique population of individuals who are all active-duty US military servicemembers. These servicemembers are required to meet the physical induction standards of their service when joining the military, which excludes from service those individuals with certain preexisting medical conditions, e.g., potential causes of secondary hip OA such as developmental dysplasia of the hip, Legg-Calvé-Perthes disease, and slipped capital femoral epiphysis. If there are any concerns about a preexisting hip condition, radiographs and an orthopedic consultation are obtained prior to initial entry into active duty military service. Active-duty US military servicemembers continue to be physically active on a daily basis. Standard training for active-duty military servicemembers includes multiple weekly aerobic and strength-training sessions, daily physically demanding military occupational specialty training, overseas deployments, maintaining height/weight requirements, and passing a semiannual service-specific physical fitness test. A service member's inability to complete any of the aforementioned requirements could result in a medical or administrative discharge from the military.
Previous incidence studies of hip OA have reported an overall incidence rate between 56 and 88 cases per 100,000 person-years, with incidence rates for men of 47–69 cases and for women of 68–105 cases per 100,000 person-years (16, 17). The incidence rate of hip OA in our population was 35 cases per 100,000 person-years, with 32 cases for men and 54 cases for women per 100,000 person-years. Our lower overall incidence rates are likely attributable to a younger and healthier population. As seen in Table 3, there was an incidence rate of 55 for those in the 35–39 years age category, with a jump to 140 in the ≥40 years age category per 100,000 person-years.
Our results found an overall increased risk in the development of hip OA in women compared with men, with an adjusted incidence rate ratio of 1.87 (95% CI 1.73–2.01). This finding is in accordance with the results of a meta-analysis performed by Srikanth et al (20). Analyzing 2 incidence studies of hip OA (16, 17), both demonstrated a statistically significant reduction in the overall incidence rate of hip OA in men when compared with women (incidence rate ratio 0.64, 95% CI 0.48–0.86). Our study showed that at all individual age subcategories there was a significant increased adjusted incidence rate ratio when comparing women with men (Table 1). The uniform findings of an increased incidence of hip OA in women in this and the previously mentioned incidence studies need to be evaluated with respect to previous prevalence studies of hip OA. Prior prevalence studies of hip OA have reported mixed results, including an increased prevalence in women (13, 14, 21, 22) or a lack of a sex difference (8, 9, 15, 20, 23). Sex differences in incidence studies, as compared with prevalence studies, are likely to be more reliable because they are unaffected by the difference in mortality rates between men and women. The previous incidence studies of hip OA were performed among whites (16, 17). In contrast, our study is the first to our knowledge to have discovered an increased incidence of hip OA among women in a racially diverse population.
Our results found an overall increased risk in the development of hip OA in black servicemembers compared with white servicemembers, with an adjusted incidence rate ratio of 1.32 (95% CI 1.23–1.41). Currently, there are no published studies comparing the incidence rate between these 2 races because epidemiologic studies have historically looked at only the prevalence of hip OA in diverse populations around the globe. Results of a small study of nearly 200 people from South Africa reported a hip OA prevalence rate of only 2% in their black population (24). A large-scale study of a rural population in North Carolina showed the prevalence of hip OA to be nearly equal between blacks and whites, with overall prevalences of 29.9% and 26.4%, respectively (8). We acknowledge that it is difficult to interpret the data for the racial servicemember category of other because it is made up of numerous ethnicities; however, after adjusting for the covariant risk factors, the other racial servicemember category did not have a significantly different incidence rate when compared with white servicemembers. Numerous studies of Asian populations have shown that the prevalence rate of primary hip OA is small, ranging from 0–2% in the Japanese, Chinese, and Taiwanese populations (25–29). This lower age-standardized prevalence of hip OA occurs in native Asian populations compared with the US white population (28) and has been shown to translate to Asians living within the US as well (30). The most common cause of hip OA in Asian populations has been found to be secondary hip OA, caused by acetabular dysplasia, rather than primary hip OA (25–27, 31).
As anticipated, our study demonstrated a statistically significant increased risk for the development of hip OA with increasing age. The adjusted incidence rate ratio was 22.21 (95% CI 17.54–28.14) for those servicemembers ≥40 years of age compared with those servicemembers <20 years of age. Cushnaghan and Dieppe have reported that the hip was the earliest joint affected by OA for both sexes (32). It is recognized that hip OA incidence rates increase with age (9, 15–17). However, what is not clear is the extent to which activity level and occupational demands may impact the incidence.
There was a statistically significant increased risk for development of hip OA when examining the adjusted incidence rate ratio for servicemembers in the junior enlisted, senior enlisted, and senior officers rank groups when compared with the junior officers rank group. Compared with servicemembers in the Air Force, there was a statistically significant increased risk for development of hip OA when examining the adjusted incidence rate ratio for servicemembers in the Army, Navy, and Marines. In this database, rank group and military service act as a proxy for activity level. Therefore, these results emphasize the role of activity level, including occupational demands, in the development of hip OA. Those military servicemembers in the junior enlisted rank group and those servicemembers in the Army, Navy, and Marines generally have an overall higher activity and occupational demand level than the referent groups. The lowest rank group, consisting of junior enlisted servicemembers, comprises the majority of combat and combat support units, maintains superior physical readiness through structured physical training programs, and is subject to the physical rigors of repeated combat deployments. For example, servicemembers in the Army and Marines serving as infantry often deploy for combat tours and routinely conduct combat operations carrying an average load of 70 pounds. This high level of physical activity is in accordance with our finding that the adjusted incidence rate ratio was highest in the junior enlisted rank group. Although their current duties may not be as physically demanding as the enlisted ranks, the senior officer rank group has the second highest adjusted incidence rate ratio. This may be attributable to the cumulative physical activity that they encounter during the years of service required to reach their current rank level.
Several studies that have examined former recreational and professional athletes have shown a positive correlation between various sports participation and an increased prevalence of hip OA (33–41). A retrospective case–control study of regular high-mileage, long-distance male runners who were longitudinally followed for 15 years found radiologic evidence of OA consisting of subchondral sclerosis (59%) and osteophyte formation (29%) in their cohort (39). Additionally, 30% of these runners at age 42 years had symptomatic hip pain, and their long-term, high-intensity, high-mileage running was found to be statistically equivalent to age as a risk factor for the development of radiographic hip OA and symptomatic hip pain. Prolonged exposure to higher levels of recreational physical activity has been reported to have a statistically significant increase in the risk of developing radiographic symptomatic hip OA in women and men ranging from a factor of 2 to 4.5, respectively (37, 40, 41). In contrast, several surveys have found no increased risk for the development of hip OA secondary to running or increased physical activity levels (42–45).
Similarly, the prevalence of radiographic hip OA has been reported to be statistically significantly increased for farmers (12, 46, 47) and for those with jobs involving heavy manual labor (48, 49) compared with population controls, ranging from factors of 3.2–10 to 2.0, respectively. Military service is somewhat comparable to farming and heavy manual labor in that the occupational demands often require regular heavy lifting, prolonged standing, and walking over rough terrain. Both women and men who have been exposed to high physical demands either at work or during sports activities have been found to have a statistically significant increased risk for symptomatic primary hip OA of 4.3 (95% CI 1.1–11.0) and 8.5 (95% CI 4.0–17.9), respectively (41, 50). Overall, the higher incidence rate of hip OA observed in military servicemembers in the ≥40 years age group in this study is attributable to their overall higher activity and occupational demand level compared with that of the general population. Moreover, within this military population it has been observed that lower rank servicemembers and those serving in the Army, Navy, and Marines will have higher incidence rates of hip OA, most likely due to their overall higher activity level and occupational demand.
The greatest strength of this study is the large number of active-duty servicemembers who were captured in the closed military health care system and annotated in the DMED. The authors acknowledge the limitations inherent to any large database study. First, multiple physicians evaluated and coded the patient encounters, which may decrease the accuracy of the diagnosis of hip OA. The DMED is also subject to the dynamic cohort of military servicemembers that is the subject of this study. Although the DMED cannot track individuals, the total number of persons used in our statistical analysis represents the population as a whole at risk for each calendar year. Second, most studies describing hip OA have focused on radiographic findings. In this study we are uncertain to what extent the diagnoses were determined by radiographic findings alone, clinical evidence of symptomatic hip OA, or a combination of the two. As the data is based upon patient-driven clinic visits, one can assume that the data is reflective of symptomatic complaints in the context of radiographic findings. Nonradiographic methods based on pain or other symptoms may result in the overdiagnosis of hip OA. Conversely, many patients with radiographic evidence of hip OA do not have symptoms and may not seek medical care, and thus would not be captured in a database study (14). It is symptomatic hip OA that causes pain, disability, and utilization of health care resources and therefore is the subject of this study. Third, our subjects were all active-duty servicemembers who maintain characteristics that may limit the ability to generalize these results. The results can be utilized, however, in comparison with populations of physically active individuals whose activity may be through sports or occupation. Further prospective studies would be helpful to better understand the effect of high activity levels and occupational demand on the development of hip OA within an ethnically diverse population.
Dr. Scher had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study design. Scher, Belmont, Owens.
Acquisition of data. Scher, Owens.
Analysis and interpretation of data. Scher, Belmont, Mountcastle, Owens.
Manuscript preparation. Scher, Belmont, Owens.
Statistical analysis. Scher, Belmont, Mountcastle, Owens.
- 44The effect of running on the pathogenesis of osteoarthritis of the hips and knees. Clin Orthop 1987; 198: 106–9., .