Low back pain (LBP) is the primary cause of activity limitation in both men and women (12) and the second most frequent reason, after upper respiratory infections, for physician visits (7). Evidence of the contribution of physical activity to the prevention and management of LBP is still inconclusive. In a review of both randomized trial and observational studies (8,14), the authors concluded that there is limited evidence that exercises to strengthen back and abdominal muscles and to improve overall fitness can decrease the incidence and duration of LBP episodes. These conclusions should be viewed cautiously as they are based on studies conducted in the workplace rather than in clinical settings or among general populations. LBP was less frequent among those participating in sporting activities in a study conducted in a clinical setting (6) and among industrial workers (23), but this relation was not established in a community-based study (13) and in another study of industrial workers (16). Campello et al. (4) suggested that the relationship between level of activity and LBP follows a U-shape curve (too little or too much activity is equally hazardous for the back). There is evidence of indirect association between LBP and sports physical activities (1), and of direct association between LBP prevalence and physical load during work (1,13). In a review of the literature regarding low-status jobs and their relationship to health risks (17), the authors claimed that physically monotonous or repetitive work, which is very common in low-status jobs, are associated with an increase in neck, shoulder, and low back problems.
Possible explanations for the inconsistent conclusions of previous studies might be differences between populations, inappropriate sampling, or the use of inappropriate measures of physical activity. Most of the above-mentioned studies addressed LBP patients in clinical settings or in workplace. Evidence stemming from the general population that represents workers and nonworkers, participants and nonparticipants in regular sporting activities, participants in various types of sporting activities, and its relationship with LBP prevalence and severity is quite rare if it exists at all. In addition, most available studies did not use a valid measure of physical activity in the context of back pain and therefore should be interpreted cautiously. In fact, studies in which physical activities were evaluated globally (17,18) failed to detect a relationship between physical activity and LBP, whereas a study (13) that used a measure that differentiated between different types of physical activities succeeded in finding such a relationship. An association between leisure time physical activities and LBP has not been demonstrated in previous studies to date (13).
The International Paris Task Force on back pain (1) confirmed that the apparently contradictory results can be partially explained by lack of uniformity in the measurements of physical activity and by the wide range of populations studied (e.g., athletes and nonathletes). Recognizing evidence for different effects of occupational and nonoccupational activities on LBP, the task force recommended the use of measurements that differentiate between three types of activity: activities of daily living, occupational activities, and recreational and sport-related activities, because these may not be associated with LBP in the same way.
Following the task force recommendations, and because of the paucity of information regarding physical activity in a general population and methodological flaws in most previous studies in this area, we chose to investigate occupational, sporting, and leisure time physical activities among all adults of a defined community (nonathletes) and their relation to LBP prevalence and severity. We hypothesized that different types of sporting activities might relate differently to LBP prevalence and therefore evaluated this potential relationship as well.
Gathering information about the relationship between the different dimensions of physical activities and LBP in a defined community, and not only among those who sought care or among subjects in working places, as was done in most previously published work in this area, may strengthen the validity of the present study results and therefore contribute important knowledge to the existing ambiguous evidence. Thus, the aim of the study was to evaluate the relationship between LBP prevalence and severity and three dimensions of physical activity in a defined community.
Setting and subjects.
A cross-sectional survey, conducted in a single town in Israel consisted of all adult inhabitants aged 22–70 (N = 3350). A list of all inhabitants obtained from the local authorities included information about age, gender, and addresses. The majority could be characterized as white-collar workers of the high and middle socioeconomic classes. Inclusion criteria were living in town at the time of the survey and being fluent in Hebrew. Those who could not provide information due to physical or cognitive medical conditions were excluded. That information was based on relatives’ reports during the survey. The list of reasons for exclusion is in the Results section. The Institutional Review Board of the Hadassah Medical Center, Jerusalem, approved the study.
The survey encompassed a 12-month period, from April 1999 to March 2000. All inhabitants were randomly assigned in 12 clusters by streets. Subjects of the first cluster were treated as a pilot study. After the pilot study, several changes were implemented in the data collection procedure in order to improve response rate. To avoid response bias, this group was not included in the main study. All inhabitants who met the inclusion criteria received, by messengers, self-administered questionnaires accompanied by a cover letter and an informed consent form. Date for collecting the questionnaires was agreed upon few days later through a telephone call from the principal investigator. This procedure enabled her to persuade potential nonrespondents to respond, to identify nonrespondents, and to ask them for reasons and LBP prevalence during the previous month.
The questionnaire consisted of two parts. The first part included questions about back pain prevalence, physical activity, smoking, working status, work satisfaction, perception of general health, and demographic information. Those who reported LBP during the previous month were asked to complete the second part of the questionnaire as well, which included questions about pain characteristics and care seeking. This group was followed up after 1 yr using similar questionnaires. The results of the longitudinal study are presented elsewhere (10).
LBP was considered as a pain between the 12th rib and the lower glutei folds that lasted at least 1 d and interfered with regular daily activity. A simple body chart illustrated the location. Prevalence of LBP was determined by answers to the questions “Did you ever/during the previous year/during the previous month experience LBP?”. The evaluated characteristics of LBP were functional disability, pain symptoms frequency and bothersomeness, pain severity, and duration. Functional disability was evaluated by the Modified Roland and Morris Disability Scale (MRMQ) (19,21) and pain symptoms by the Pain Symptoms Frequency (SFI) and Bothersomeness (SBI) Indices (21). A 10-degree pain scale registered pain severity. Duration of LBP was registered by the question: “For how long did your previous episode of pain lasted?” (optional answers: ≤ 1 month; > 1 month). Subjects were also asked about their perception of general health. A single-item self-rating instrument evaluated it with responses reported along a 100-point continuum scale (11,20).
Physical activities were evaluated by the Beacke Physical Activity Questionnaire (BPAQ) (3). This instrument provides three indices of habitual physical activity during the previous year: Occupation (OAI), Sport (SAI), and Leisure (LAI) Activity Indices. The reliability (5) and validity (24) (evaluated in a population presenting coronary risk factors) of this measure were acceptable (intraclass correlation > 0.85). Canon et al. (5) reported an indirect association between physical activity at work and direct association between physical activity at leisure and education, as was suggested before (24). Consequently, they recommended analyzing data regarding the three indices separately. The OAI is composed of sum scores of eight items regarding frequency of specific activities during work (e.g., walking, sitting, and lifting heavy things). The SAI is composed of the multiplication of three scores: level of intensity (depends on type of sport activity), time (number of activity hours per week), and proportion (number of activity months during the year). The intensity is determined by expected level of energy expenditure during a specific activity. Thus, activities with high-energy demands (e.g., soccer, bicycle riding, and jogging) received a high score, activities with moderate energy demands (e.g., swimming and walking) a moderate score, and activities with low energy demands (e.g., calisthenics and yoga) a low score. The LAI is composed of the sum scores of five items regarding frequency of activities during leisure time (e.g., cycling, going shopping or to friends on foot, and watching TV).
Other life style markers were smoking, defined by a “yes/no” answer to the question “Do you smoke on a regular basis?”, work status (employed/unemployed), and work satisfaction, evaluated by a four-category Likert-type scale (not satisfied, moderately satisfied, satisfied, or very satisfied). Demographic variables included in the study were age, gender, and education (defined by years of formal education). All measurement tools were translated into Hebrew, and their reliability was reevaluated by the authors in a preliminary study and found to be acceptable (9) (intraclass correlation > 0.89).
Independent sample t-tests were used to compare means of continuous variables (e.g., pain and disability scales, general health, age, and physical activity indices) and chi-square to compare between categorical variables (e.g., pain duration gender, and smoking)) among participants and nonparticipants in regular physical activities and between those participating in different types of sporting activities. Continuous variables were evaluated for normality by exploring their skewness and kurtosis before the analysis. Logistic regression was used to evaluate potential contributors to dichotomous dependent variables (e.g., LBP and sporting activities participation) and linear regression for continuous dependent variables (e.g., SAI, pain severity, and disability score). The relationship between physical activities and LBP measures was evaluated among those reporting LBP, in both directions, first with physical activities as the dependent variable and then with LBP prevalence and severity. All P values less than 0.05 were regarded as statistically significant. Data analysis was performed using the SPSS statistical package version 10.
Of 3350 inhabitants, 355 who were considered as the pilot study were not included, 129 were excluded, and 866 did not respond. CS survey data were not available from 129 inhabitants. The main reasons were insufficient knowledge of Hebrew (N = 59), mental or cognitive impairment (27), severe chronic illness (11), could not be contacted (e.g., mourning or living in nursing home) (20), death (5), and paraplegia (7). Finally, 2000 subjects completed the questionnaires. Data regarding nonrespondents were available from clusters 5 to 12, of whom only 42.5% provided information. The main reasons for nonresponse were lack of interest (234), lack of time (124), and protecting privacy (65). Respondents were more often females, slightly older, and reported higher frequency of LBP (28.5% vs 22.6%) than nonrespondents (Table 1).
Most participants did not smoke, were employed, and were satisfied with work. Nearly half participated in regular sporting activities, about one third experienced LBP during the previous month, and nearly half during the previous year. A total of 555 subjects who reported LBP during the previous month completed the second part of the questionnaire. More than three quarters of the LBP group had short duration pain (less than 1 month) over the previous month. Mean scores of all LBP measures indicated low or mild LBP in most cases (Table 1). Age and educational distribution of the population revealed a slightly higher rate of young and educated people than for of the general Jewish population of Israel (2) (for example, the proportion of those with over 16 yr of schooling among age group 25–44 is 18% in the general population and 35% in the study population).
Feales, unemployed, smokers, and those experiencing LBP during the previous month and year presented higher OAI (Table 2). Males and those who did not experience LBP during the previous month presented higher SAI. The SAI was not related to work status, smoking, and LBP during the previous year (Table 3). Males and those who did not smoke presented higher LAI. This was not related to work status or to LBP prevalence (Table 4).
Nearly 50% participated in regular sporting activities. This group was more educated, smoked less, was more satisfied with work, participated less in occupational activities and more in leisure activities, and presented better perception of general health and lower prevalence of LBP during the previous month and year compared with those who did not participate in regular sporting activities. In addition, those participating in sporting activities, among the LBP group, presented lower scores for all LBP measures (pain severity scale, MRMQ, SFI, and SBI) compared with those who did not. No age, gender, and work status differences were detected between the groups (results of univariate analysis;Table 5).
Regression analysis revealed similar results to those of univariate analysis. High OAI, low LAI, low perception of general health, female gender, low education, and smoking increased the risk of not participating in sporting activities. Age, work status, and work satisfaction did not contribute to sporting activities participation (Table 6). In addition, increased LBP measures (MRMQ and SFI) among those reported LBP contributed to nonparticipation in regular sporting activities (Table 7).
The majority of those who were engaged in regular sporting activities participated in moderate (56.4%), followed by high (27.7%), and by low (15.9%) energy expenditure activities. Men, younger people, and those with higher perception of general health were more likely to participate in high energy activities than women, older people, and those with low perception of general health, respectively. No differences were detected between those who experienced LBP during the previous month and year, and those who did not with regard to the type of sporting activity (Table 8).
The only contributors to LBP prevalence (dependent variable) during the previous month and year were high OAI, and low perception of general health. Increased age contributed also to LBP during the previous month. Participating in any sporting activity, type of sporting activity, LAI, work status, work satisfaction, level of education, and smoking did not contribute to LBP prevalence (Table 9). However, low SAI, low education, and low perception of general health contributed to higher score in most LBP measures. This was demonstrated with regard to pain severity scale, MRMQ, SFI, and SBI (Table 10).
This community-based study describes three dimensions of habitual physical activities and their relation to LBP prevalence and severity among adults, age 22–70, and among LBP subjects in a defined community. Whether participating in physical activities leads to less frequent LBP or to decreased severity rather than the opposite is beyond the scope of a cross-sectional survey. Therefore, data were analyzed in two directions, first with participating in physical activities as the dependent variable and then with LBP prevalence and severity as dependent variables. The relationship between physical activities and LBP was evaluated in three steps. The first step addressed the three dimensions of physical activity, the second addressed participation versus nonparticipation in sporting activities, and the third addressed type of sporting activities.
The three dimensions of physical activities yield different relationships to LBP prevalence and to participants’ sociodemographics as has been demonstrated before (1,13,17). Low OAI, high SAI, and to some extent high LAI were associated with low prevalence of LBP and with several markers of healthy lifestyle (e.g., nonsmoking and participation in sporting activities) (results of univariate analysis). The inverse relationship between OAI and LBP and between SAI and LBP was also confirmed by multivariate analysis. High OAI contributed to increased frequency of LBP and high SAI contributed to decline in all LBP measures. These results are in congruence with previous reports about low-status jobs, which stressed the relationship between high physical load during work and LBP (13,16,17) and with evidence regarding the contribution of sporting activities to less frequent LBP (6,22).
The relatively high proportion of participation in regular sporting activities (50%) can be attributed to the sociodemographic characteristics of this population (characterized as white-collar workers, and middle and high sociodemographic class) and to differences from the general population in Israel (slightly younger and more educated) (2). The results indicated some common characteristics of those who were free of LBP and those who participated in regular sporting activities. Both groups were presented healthier lifestyle (e.g., smoked less and had lower OAI) and reported higher perception of general health as compared with those who experienced LBP or did not participate in regular sporting activities. However, regression analysis revealed that although increased LBP measures contributed to decline in sporting activities participation, participating in regular sporting activities did not contribute to low LBP prevalence or to decreased severity (dependent variables were exchanged). The significance of decreased prevalence of LBP among those who participated in sporting activities can be debated. It seems that participating in regular sporting activities is an indicator of a healthy lifestyle, indicating less smoking, participating frequently and intensively in sporting activities on a regular basis, avoiding heavy occupational activities, and being more active during leisure time. These factors lead to lower back morbidity, as was reported before (16). Similarly to our results, there is previous evidence of smokers facing increased risk of suffering from LBP (15).
The role of sporting activity participation was further evaluated according to the type of sporting activity defined by level of energy expenditure. Although the sociodemographic profile of those involved in high energy sports differed from that of those involved in moderate and low energy sports (e.g., more males, younger, and presented better perception of general health and higher SAI), they did not differ from each other with regard to LBP prevalence. Therefore, our baseline assumption that different sporting activities might relate differently to LBP prevalence was not confirmed.
We employed several methodological strategies to enhance the validity of the findings. The population (70% response rate) represents a target population of all adult inhabitants, age 22–70, of a defined community. Respondents and nonrespondents were compared for LBP prevalence, age, and gender. Age differences, although statistical significant, are irrelevant in the context of LBP. Last but not least, the terms LBP and sporting activities were clearly defined and evaluated by reliable and validated measures (3,9,19,21).
The limitations of this study include the use of self-administered questionnaires, which requires a certain level of education. That should be taken into consideration especially because the proportion of low-educated participants is lower than that of the general population (10). A slightly excessive prevalence of LBP is expected in this study compared with the target population for two reasons. First, because women were more likely to respond, as was demonstrated before (3), and, second, because respondents reported a higher prevalence of LBP. Data regarding nonrespondents to the cross-sectional survey was available from less than 50% of this group. The sociodemographic characteristics of the population differ from those of the general population in Israel, and therefore any attempt to extrapolate them to other populations’ interpretations should be carried out carefully.
The results support previous evidence of indirect association between LBP and sports physical activities (1) and of direct association between LBP prevalence and physical load during work (1,13). There are several shared characteristics between those who participate in regular sport physical activities and those free of LBP. Both groups report higher perception of general health and are involved in fewer occupational physical activities. These characteristics can be considered as markers of a healthy life. Whether LBP severity is the cause of less frequent sporting activity rather the opposite effect is not clear. However, considering the overall, well-known benefits of sporting activities and our results, it seems that LBP patients, as the general population, should be encouraged to participate in regular sporting activities regardless of its type.
The study was partially supported by grants from the Israeli Sports and Physical Education Authorities, Ministry of Education, Culture, and Sport.
The College of Judea and Samaria granted publication fee.
1. Abenhaim, L., M. Rossignol, J. P. Valat, et al. The role of activity in the therapeutic management of back pain: report of the international Paris Task Force on back pain. Spine 25: 1S—33S, 2000.
2. Annual Bulletin of Statistics. Israel Central Bureau of Statistics, 1998, pp. 602.
3. Baecke, J. A., H. J. Burema, and J. E. R. Frijters. A short questionnaire for the measurement of habitual physical activity
in epidemiological studies. Am. J. Clin. Nutr. 36: 936–942, 1982.
4. Campello, M., M. Nordin, and S. Weiser. Physical exercise and low back pain (Review). Scand. J. Med. Sci. Sports 6: 63–72, 1996.
5. Canon, F., B. Levol, and F. Duforez. Assessment of physical activity
in daily life. J. Cardiovasc. Pharmacol. 25 (Suppl. 1): S28–S34, 1995.
6. Cox, J. M., and K. K. Trier. Exercise and smoking habits in patients with and without low back and leg pain. J. Manip. Physiol. Ther. 10: 239–245, 1987.
7. Hart, L. G, R. A. Deyo, and D. C. Cherkin. Physician office visits for low back pain: frequency, clinical evaluation, and treatment patterns from a U.S. Natl. Survey Spine 20: 11–19, 1995.
8. Helewa, A., C. H. Goldsmith, and H. A. Smythe. Measuring abdominal muscle weakness in patients with low back pain and matched controls: a comparison of 3 devices. J. Rheumatol. 20: 1539–1543, 1993.
9. Jacob, T., M. Baras, A. Zeev, and L. Epstein. Reliability of a set of LBP related measurement tools. Arch Phys. Med. Rehabil. 82: 735–742, 2001.
10. Jacob, T., M. Baras, A. Zeev, and L. Epstein. A longitudinal, community based study of low back pain outcomes. Spine
11. Kaplan, G., V. Barell, A. and Lusky. Subjective state of health and survival in elderly adults. J. Gerontol. 43: 114–120, 1988.
12. Kelsey, J. L., D. J. Mundt, and A. L. Golen. Epidemiology of low back pain. In: The Lumbar Spine and Back Pain, 4th Ed., M. I. V. Jayson (Ed.). Edinburgh: Churchill Livingstone, 1992, pp. 537–549.
13. Kujala, U. M., S. Taimela, T. Viljanen, et al. Physical loading and performance as predictors of back pain in healthy adults: a 5-year prospective study. Eur. J. Appl. Physiol. 73: 452–458, 1996.
14. Lahad, A., A. D. Malter, A. O. Berg, and R. A. Deyo. The effectiveness of four interventions for the prevention of low back pain. JAMA 272: 1286–1291, 1994.
15. Leboef-Yde, C., K. O. Kyvik, and N. H. Brunn. Low back pain and lifestyle. Part 1: Smoking. Information from a population based sample of 29,424 twins Spine 23: 2207–2214, 1998.
16. Leino, P. I. Does leisure time physical activity
prevent low back disorders? A prospective study of metal industry employees. Spine 18: 863–871, 1993.
17. Lundberg, U. Stress responses in low-status jobs and their relationship to health risks and musculoskeletal disorders. Ann. N. Y. Acad. Sci. 62: 162–172, 1999.
18. Natvig, B. D. Bruusgaard, and W. Eriksen. Localized low back pain and low back pain as part of widespread musculoskeletal pain: two different disorders? A cross-sectional population study. J. Rehabil. Med. 33: 21–25, 2001.
19. Patrick, D. L., R. A. Deyo, S. J. Atlas, D. E. Singer, A. Chapin, and R. B. Keller. Assessing health-related quality of life in patients with sciatica. Spine 20: 1899–1909, 1995.
20. Pijls, L. T. J., E. J. M. Feskens, and D. Kromhout. Self-rated health, mortality, and chronic diseases in elderly, the Zutphen study, 1985–1990. Am. J. Epidemiol. 138: 840–848, 1993.
21. Roland, M., and R. Morris. A study of the natural history of back pain. Part 1: Development of a reliable and sensitive measure in low back pain. Spine 8: 145–150, 1983.
22. Takemasa, R., H. Yamamoto, and T. Tani. Trunk muscle strength in and effect of trunk muscle exercises for patients with chronic low back pain: the difference in patients with and without organic lumbar lesions. Spine 20: 2522–2530, 1995.
23. Toroptsova, N. V., L. I. Benevolenskaya, A. N. Karyakin, I. L. Sergeev, and S. Erdesz. “Cross-sectional” study of low back pain among workers at an industrial enterprise in Russia. Spine 20: 328–332, 1995.
24. Washburn, R. A., and H. J. Montoye. The assessment of physical activity
by questionnaire. Am. J. Epidemiol. 123: 563–576, 1986.
Keywords:©2004The American College of Sports Medicine
PHYSICAL ACTIVITY; HEALTHY LIFESTYLE; PARTICIPATION