Epidemiology of musculoskeletal injuries among sedentary and physically active adults

The health benefits of regular physical activity include reduction in the incidence and mortality associated with cardiovascular disease, diabetes, obesity, hypertension, and cancer (^21,23,28). Unfortunately, some hazards are associated with physical activity, most notably, musculoskeletal injury (^15,21,28). In light of recent federal recommendations for moderate physical activity for all U.S. adults (^21,28), health practitioners need detailed epidemiological studies to define the injury burden, to identify risk factors, and to develop targeted injury prevention programs within the physically active adult population to safely promote the health benefits of physical activity among the general public.

Data describing the types and frequencies of musculoskeletal injury among active individuals are scarce. Descriptive profiles of injuries are documented among selected physically active populations, including high school, collegiate, and professional athletes, military trainees, and recreational and elite runners (^{1,8,10,12–14,16,17,22,29,31}). To date, little information exists concerning the prevalence and types of injuries among sedentary adults and among adults engaged in a wide range of physical activities recommended for good health, such as walking, cycling, swimming, and other sports. Thus, no information currently exists describing if injuries among physically active adults are distributed similarly as those among highly skilled athletes. We also have no data that describe if injury experiences differ between men and women or between types of physical activity such as walking, jogging, and strenuous sports.

To improve the participation of the U.S. public in healthful physical activity, it is important to identify factors that deter individuals from being physically active. Decreasing activity level, stopping activity, and changing types of activity are important sources of morbidity associated with musculoskeletal injury. Fear of sustaining an injury and stopping activity because of an injury have been associated with failure to start and maintain a physically active lifestyle (^24,25). The cause, prevalence, and site of musculoskeletal injury, as well as the extent and duration of postinjury morbidity, have not been adequately described among adults with varying activity levels. In this study we describe the prevalence of all-cause and activity-related musculoskeletal injuries among a cohort of sedentary and active adults and compare the injury site, type of activity, self-treatment modes, and postinjury morbidity by sex.

MATERIALS AND METHODS

Participants.

Data were obtained from a population of adults selected from patients of the Cooper Clinic, a preventive medicine clinic in Dallas, Texas. Participants were self or employer referred to the clinic for various services such as preventive medical examinations and health, nutrition, and exercise counseling. We selected adults aged 20–85 yr who were enrolled in the Aerobics Center Longitudinal Study (ACLS), a prospective study of the health effects of physical activity and fitness at the Cooper Clinic. Baseline data for health status and behaviors, physical activity, and physical fitness levels were obtained from participants who were enrolled in the study between 1970 and 1982. After reading and signing an informed consent form, all participants underwent a physical examination by a physician, completed a personal and family health history questionnaire, had anthropometric and blood chemistry measurements taken, and completed an age-predicted, maximal treadmill graded exercise test.

Our study population included any ACLS participant with a baseline examination and who responded to a mailed follow-up survey designed to obtain extensive information about orthopedic injury history, physical activity levels, and exercise habits in 1986. Surveys were mailed to 11,972 participants. After accounting for deceased participants and undeliverable surveys, 6,315 surveys were returned for a 72% response rate. Most study participants were men (78%), Caucasian (97%), highly educated (more than 80% had a college degree), and moderately or highly fit (87%). Few participants were current smokers (6.4%).

Physical data.

Height was measured during the baseline or a subsequent clinic visit; weight was self-reported on the follow-up survey (1986). Body Mass Index (BMI) was calculated as weight kilograms/height meters (²) and classified as normal (<25) and overweight/obese (25+). Age was self-reported on the follow-up survey. Physical fitness was measured at baseline by completion of a treadmill graded exercise test using the modified Balke testing protocol under supervision of a trained technician and physician (^2,7). Details of the fitness testing procedures have been previously described (^3,4). Fitness levels were categorized from age and sex-specific quintiles of total treadmill time among the entire cohort. The low fitness group included participants in the lowest 20% (bottom quintile); the moderate and high fitness groups were categorized from the middle and top 40% of the total treadmill-time distribution.

Follow-up survey data.

Participants were sent a follow-up survey to obtain detailed information about their musculoskeletal health status including injuries and surgeries. For participants reporting an injury in the 12 months preceding the follow-up survey, each participant self-selected one reported injury as “the one most severe injury” and provided details regarding modes of self-treatment and morbidity (stopping physical activity, decreasing intensity of activity) associated with this injury. Participants also provided information about their physical activity habits and the frequency of stretching, weight training, and calisthenics exercises performed during the previous 12 months. Using the participants’ primary type of reported physical activity during the previous 12 months, we categorized their physical activity habits into four groups: sedentary, walkers, runners, and sport participants. Sport participants were those who answered “no” to the run/walk/jog question, but “yes” to a series of questions about participation in racquet sports, other strenuous sports (football, basketball, softball, etc.), cycling, or swimming during the previous 12 months. Sedentary individuals were those participants who answered “no” to all activity questions (running, walking, jogging, bicycling, swimming, racquet sports, and other strenuous sports). We classified participants as walkers and runners if they answered “yes” to the question, “Have you participated in a run/walk/jog program in the previous 12 months?” Walkers were those who reported a run/walk/jog exercise pace of 15 min/mile or slower. If the run/walk/jog pace information was missing (4–5%), participants were classified as walkers if walking comprised at least 75% of their run/walk/jog program. Runners were those who reported a run/walk/jog pace of less than 15 min/mile or who walked less than 25% of their run/walk/jog exercise session. Walkers and runners with missing data on both pace and percent of time spent in walking or running were excluded (N = 165). 1/5 of walkers and runners also reported participating in other sport activities. Of these, running and walking comprised 85–91% of their total activity time per week, and therefore for this analysis, we consider their primary physical activity to be walking or running as defined above.

Injury definition.

Participants were queried regarding their history of musculoskeletal injury for all year before the follow-up survey. We used this yearly injury history information to create two variables for each participant: previous injury history (any injury reported through 1984) and current injury (any injury reported in 1985). More detailed information was available for the current injuries including: the participant’s self-perceived cause of injury (run/walk/jog program, strenuous sport program, or other), the site of injury, and whether the injury resulted in the participant taking medication, having surgery, consulting a physician, or stopping/reducing their usual physical activity. Current injuries were classified into two groups to calculate injury prevalence. All-cause musculoskeletal injuries were defined as any self-reported muscle, tendon, bone, joint, or ligament injury; activity-related injuries were defined as any injury (as defined above) attributed by the participant to participation in a run/walk/jog or strenuous sports program that resulted in the participant taking medication, having surgery, consulting a physician, or stopping/reducing their usual physical activity.

Data analyses.

Descriptive information is presented for men and women separately by cause of injury. A graphical figure depicts the distribution of activity-related injuries by site and sex. Differences in the distribution of selected characteristics between injured and uninjured participants and between men and women were tested using the chi square test for proportions. The significance level for all analyses was set at alpha = 0.05. All statistical analyses were conducted using SAS statistical software (SAS, Cary, NC), version 6.12.

RESULTS

Men and women were not significantly different with regard to cause of injury (Table 1). 1/4 of all participants reported at least one musculoskeletal injury in the 12 months preceding the follow-up survey. Of the total all-cause musculoskeletal injuries, 82% (N = 1,052) and 84% (N = 260) of injuries among men and women, respectively, were attributed to physical activity (a run/walk/jog or strenuous sport program). The lower extremity accounted for the majority of activity-related injuries (68% for men and 65% for women).

Table 2 reports the prevalence and descriptive characteristics of all-cause and activity-related musculoskeletal injuries by sex. For both all-cause and activity-related injuries, men who were younger, more fit, more physically active, and who had previous injuries were significantly (p ≤ 0.05) more likely to be injured. For all-cause injuries, women who were younger, more physically active and had previous injuries were significantly (p ≤ 0.05) more likely to be injured. For activity-related injuries, more fit, active, and previously injured women were significantly (p ≤ 0.05) more likely to be injured. The proportion of injured participants was significantly different across physical activity groups (p ≤ 0.05) for both all-cause and activity-related musculoskeletal injuries among men and women with the exception of all-cause injuries among women. Sport participants had the highest proportion of all-cause and activity-related musculoskeletal injuries among both men and women (range: 27–31%). Among sedentary participants, 15% of men and 17% of women reported activity-related injuries. On average, approximately 67% of men and 59% of women reported being physically active at least two times per week in the previous 12 months.

Figure 1 shows the distribution of the activity-related injuries by site of injury. The knee was the site most frequently injured among men and women for both activity-related and all-cause musculoskeletal injuries (data not shown for all-cause injuries) and accounted for 19–23% of all injuries. Among all-cause musculoskeletal injuries, the back was the second and the foot the third most common injury site for both men and women. In contrast, among participants with activity-related musculoskeletal injuries, the foot was the second and the back was the third most frequently injured body site.

For the injury, the participant self-selected as the “one most severe injury” reported in the 12 months preceding the follow-up survey, the most commonly reported self-treatment mode was having an x-ray taken followed by taking medication prescribed by a physician (Table 3). 35% of injured men and 40% of injured women performed rehabilitation exercises after an injury. Women used orthotics or changed the model of their shoes more frequently than did men. The proportion of participants reporting specific self-treatments did not significantly differ between men and women.

Over 75% of men and 68% of women reporting all-cause injuries temporarily stopped their exercise program. This outcome was followed by permanently stopping their exercise program (30.5% of men and 24.4% of women reporting all-cause injuries). Significantly (p ≤ 0.05) more men than women took up another sport (21.3% of men; 13.8% of women) or missed work (21.1% of men; 13.2% of women) after sustaining an injury.

DISCUSSION

To date, the distribution of musculoskeletal injuries among adults with varying activity levels has not been described in the literature. Previous studies of injuries among select populations focused on vigorously active groups of runners, military trainees, and other elite athletes. These populations may not be representative of injuries that occur among persons with moderate or low levels of physical activity or among sedentary persons (^{10,12–14,16,17,22}). The annual injury rates of musculoskeletal injury previously reported were 35–65% among runners and 27–51% among military trainees (^{10,12,13,17,29}). In this study, we found that more than 1/4 of all participants reported a musculoskeletal injury in the 12 months preceding a mailed follow-up survey and approximately 1/3 reporting injuries permanently stopped their exercise program due to self-perceived severe injury.

Among this sample of physically active U.S. adults, we found that more than 80% of all musculoskeletal injuries reported in the follow-up survey were activity-related. This high proportion of activity-related injuries are undoubtedly attributable to the higher activity levels of this cohort. In two reports among the general European population, several investigators found that approximately 1/3 of reported musculoskeletal injuries were related to physical activity such as participation in contact, noncontact, and unorganized sports (^27,30). In one study, the age distribution of the study sample was much younger than the ACLS cohort (³⁰). The other study (²⁷), was a population-based study in which 90% of the participants were classified as sedentary; whereas, in our cohort only 33% of participants were classified as sedentary. These differences in sample characteristics makes comparison with our study difficult.

However, even persons classified as sedentary reported activity-related injuries. This may be because they may not consider themselves regularly physically active but may participate in occasional activities on weekends or at social events. This is an important group to learn more about because they may be at higher risk of musculoskeletal injury due to low strength and cardiovascular fitness levels, yet need to increase their physical activity levels to meet national health recommendations. Specific health education messages need to be developed to encourage this group of adults to become more physically active in activities that have a low risk of injury and to inform them of injury prevention strategies.

The distribution of injuries by body site among study participants was similar, but not identical, to other physically active populations. In most studies, the lower extremity is the predominate region prone to injury among all groups of physically active persons. In previous studies of injuries among collegiate and professional athletes, runners, and military personnel, 69–97% of those injured reported lower extremity injuries (^{10,12,13,16,17,22,29}). The specific site of the lower extremity injury differed depending on the population studied. For example, among military trainees, collegiate athletes, and competitive runners, the knee, foot, and ankle were injured most frequently (^{10,12,13,16,17,29}). In two general population studies, van Mechelen et al. (³⁰) (Netherlands) and Sedgwick et al. (²⁶) (Australia) reported that the ankle and spine and lower back were the most frequently injured body sites. In our study, we observed similar findings: more than 66% of injuries occurred in the lower extremity, and back injuries were listed as the second most common site of all-cause musculoskeletal injury (16%). These findings suggest that injury prevention programs designed for physically active adults should target the back and lower extremity.

Investigators of other studies (^1,8,16,31) have reported sex differences in the incidence of specific types or diagnoses of injuries, such as anterior cruciate ligament ruptures and stress fractures. These types of injuries can be very sport-specific and may not be applicable to adults engaging in recreational physical activity. In our study, men and women reported similar prevalence of all-cause and activity-related injuries. The prevalence of activity-related lower extremity injuries were also similar among men and women. However, the specific body site injured was slightly different for men and women. Among activity-related injuries, men reported a higher percentage of shoulder, eye, calf, and Achilles tendon injuries than did women. Women reported a higher percentage of foot and ankle activity-related injuries than did men. Biomechanical, hormonal, and neuromuscular factors may play a role in sex- and site-specific injury patterns and may explain the different distribution of injury of a given site between the men and women found in this study (^1,5,8,11). However, the lack of quantitative biomechanical, diagnostic, and other laboratory information in this study prevented complete analysis of the interaction between sex and site of injury.

Two previous published studies of musculoskeletal injuries from the ACLS only included injuries to selected lower extremity sites. Blair et al. (⁴) reported the incidence of knee injuries among runners as significantly higher than among nonrunners. No differences in the incidence of injury between runners and nonrunners were reported for the foot, hip, back, shoulder, or elbow. In our study, the prevalence of knee injuries was also higher among runners than nonrunners (sedentary, walkers, and sport participants) for men and women. (Data not shown.) Macera et al. (¹⁹) conducted an analysis of the relationship between physical activity and fitness levels and the risk of musculoskeletal injury to the foot, knee, hip, and back but did not show the distribution of injuries by body site. Thus, further comparisons of the results of this study with prior reports from the ACLS cohort are limited.

Although the results of this study may suggest a need for exercise cautions among adults, we should note the limitations of these findings. Survey data were self-reported and may be susceptible to recall bias (⁹). The validity of self-reported health information among adults in this sample was previously found to be excellent with 98% sensitivity and 99% specificity (³). Further, the follow-up survey had an acceptable response rate of 72%, but respondents may still be different than nonrespondents regarding lifestyle and activity habits. Macera et al. (¹⁸) investigated the patterns of nonresponse to an earlier mailed survey among the ACLS. Using baseline laboratory measurements and a personal health history, nonresponders and responders were equally healthy. However, participants reporting more positive health behaviors and more family history of chronic disease were more likely to respond to the mail survey. Assuming the response patterns reported by Macera et al. (¹⁸) are similar for the current study, participants included in this sample most likely represent health conscious adults who demonstrate behaviors consistent with good health such as engaging in physical activity and not smoking. Thus, the results of this study may not be applicable to the general population.

The measurement of physical activity and physical fitness levels are also a unique feature of this study population. Still, misclassification may exist among participants for their physical activity levels. Sedentary individuals were easily distinguished from more active participants because they reported no participation in any type of activity. Sport participants were also easily distinguished because they reported participation in at least one sport activity. Separating walkers and runners was more difficult because classification was based on a self-reported run/walk/jog pace of less than 15 min/mile to define runners and a self-reported pace of 15 min/mile or greater to define walkers. This 15 min/mile pace cut-off was used in an earlier study among the same cohort to define runners and nonrunners (⁴). Still, some joggers may have run at a pace slower than 15 min/mile, and some fast walkers may have walked at a pace faster than 15 min/mile. Despite these possible sources of misclassification, we believe that fast walkers (speed over 4.0 mph) will have injury experiences similar to runners and joggers, and that slow joggers will have injuries similar to walkers because of the varying biomechanical stresses placed on the lower extremity during movement (²⁰). Also, even though some walkers and runners also reported engaging in sport activities, the majority of their total activity time per week was attributed to walking or running. Thus, we feel these categories serve as sufficient measures of usual type of physical activity reported by each participant in the 12-month measurement period.

This study also has several strengths. It is the first comprehensive report of the descriptive epidemiology of musculoskeletal injuries among a large sample of adults with varying activity levels outside of elite athletes, military trainees, and groups of runners. The present study cohort included sedentary individuals, and the overall activity level of the cohort was higher than that of the general U.S. population of adults. 65% of the study sample engaged in physical activity at least twice per week in the 12 months immediately preceding the follow-up survey. In contrast, only 40% of the U.S. population engage in the recommended amounts of physical activity per week (²⁸). The detailed injury information in the ACLS allowed us to analyze the physical activity levels, causes of injury, and sites of injury among participants. Including data on non activity-related musculoskeletal injuries as well as activity-related injuries provides researchers with complete information regarding the total burden of musculoskeletal injury among adults. The inclusion of a significant number of women also facilitated the comparison of injury experiences by sex.

The results of this study have several implications for prevention of musculoskeletal injuries among physically active adults. First, 15–17% of sedentary adults reported an activity-related injury. Sedentary persons have lower levels of neuromuscular fitness than do fit persons, and the associated lack of neuromuscular strength and coordination may increase the potential for injury during the infrequent times they participate in recreational physical activities. Medical practitioners should be aware that even their sedentary patients may be at risk for activity-related injuries and should counsel them accordingly. Second, only 30–40% of injured participants in this study reported performing rehabilitation exercises as a result of injury, even though connective tissue, muscle, tendon, and bone strength decrease with the disuse associated with the initial inactive period after injury. The Centers for Disease Control’s National Center for Injury Prevention and Control recently published a comprehensive list of recommendations for the prevention of activity-related injuries (⁶). Although targeted toward women, these recommendations are appropriate for all adults. In addition to completing a comprehensive rehabilitation program, other selected recommendations include: beginning activity slowly for sedentary and unfit persons, gradually returning to prior activity levels following an injury, choosing low stress activities (i.e. walking), and learning to recognize the early signs of injury (pain, aching, stiffness, and swelling) and to adjust activity levels accordingly.

This study differs from clinic-based injury studies that only describe injuries among injured persons seeking medical advice. Over 83% of all injuries were related to physical activity and 66% occurred in the lower extremity with the knee listed as the most frequent site for injury. Men and women had similar injury experiences. As the U.S. Surgeon General’s office and other health organizations continue to promote increased levels of physical activity among U.S. adults, injuries resulting from moderate and vigorous exercise may increase accordingly (^21,28). We report that almost 1/4 of injured women and 1/3 of injured men permanently and 68–75% temporarily stopped their exercise programs because of injury. This finding indicates that musculoskeletal injuries can have a significant negative impact on the short and long-term activity levels of U.S. adults. Public health practitioners should be aware of the potential hazards associated with physical activity to appropriately plan and implement prevention strategies targeted toward high-risk groups. Thus, injury education and prevention programs targeted to physically active and sedentary adults may be appropriate to avoid injuries associated with a physically active lifestyle.

This study was supported by a grant awarded to the Cooper Institute for Aerobics Research, NIH NIA #AG06945.

The authors thank Dorothy Davis of the University of South Carolina Prevention Research Center and the staff of the Cooper Institute for Aerobics Research for collecting and managing the data used in this study.

Address for correspondence: Jennifer M. Hootman PhD, ATC, Centers for Disease Control and Prevention, 4770 Buford Highway, NE, Mailstop K-45, Atlanta, GA 30341; E-mail: [email protected].