Development and Reproducibility of the Bone Loading History ... : Medicine & Science in Sports & Exercise (original) (raw)
Physical activity is an important contributor to the development and maintenance of bone mineral density (BMD) (21). The childhood and adolescent years represent an important window of time to accumulate bone mineral in response to mechanical loads imparted by physical activity (13,19,29). An assessment of lifetime physical activity could play a role in identifying premenopausal women who are at risk for low BMD. This is important considering that 15-25% of BMD can be lost during the premenopausal years alone (22). Because there are no somatic symptoms associated with bone mineral loss, many premenopausal women may be at risk for low BMD, yet most will be unaware of this risk until later in life, when preventive measures are of limited utility.
To our knowledge, no historical physical activity questionnaires (PAQ) have been specifically designed to assess components of physical activity important to bone health. Most questionnaires score physical activity exclusively by energy expenditure and metabolic equivalents (METs) and do not consider the mechanical loading aspect of activity. Because bone responds to activity that is dynamic rather than static, shorter rather than longer in duration, and novel rather than typical (17,26), a bone loading unit of measure is needed to reflect the strain on bones caused by physical activity. Further, BMD changes with physical activity intervention are relatively small in magnitude and require many months of observation to detect. Therefore, questionnaires designed to assess physical activity associated with bone health should focus on variables other than those important to the cardiovascular system (i.e., long duration per activity bout with prolonged elevation of heart rate) and should reflect long-term patterns of activity. For example, force magnitude, the rate of force development experienced by the skeleton, and the long-term pattern of participation are important for understanding the bone response to physical activity (3,14).
Another limitation of many physical activity questionnaires is that they typically focus on current or recent rather than historical or lifetime physical activity. Bone mass attained in early life is believed to be the most important determinant and indicator of lifelong skeletal health (7). In addition, exercise during skeletal growth is thought to be more osteogenic than exercise after skeletal growth is complete (27). Thus, it is important to identify mechanical loads placed on the skeleton throughout life rather than only during recent time periods. Finally, it is essential for a questionnaire, particularly one focusing on historical data, to be a reliable instrument and to produce consistent results when administered on separate occasions.
The purpose of the current study was threefold. The primary aim was to develop an historical bone loading questionnaire that assesses loads applied to the skeleton using bone loading units. A secondary aim was to determine the reproducibility of the historical questionnaire. The third aim was to determine whether bone loading exposure correlated with spine and hip aBMD and whether low levels of bone loading exposure would increase the odds for low spine and hip aBMD in healthy premenopausal women with a wide range of physical activity habits.
METHODS
Questionnaire development.
The Bone Loading History Questionnaire (BLHQ) was developed with the assistance of experts in the field (1). The Classical Test Theory measurement model was used to develop the instrument (10). Instrument development included four steps to provide evidence of construct validity. The first step was to define the target construct as the historical physical activity that quantitatively and qualitatively relates to bone health. The second step was to generate items that measured the defined dimensions of the construct. The following dimensions of physical activity are recognized as important to bone health: (a) age at time of participation, (b) frequency of exposure to the activity, and (c) type of activity with respect to the forces and rate of force applied to the skeletal sites of interest (16,26).
The third step was to establish the format of the questionnaire. The first phase in developing the format was to determine which activities would be listed on the questionnaire to help women recall the activities they most regularly participated in during their lifetimes. A small sample of women (_N_= 10) aged 18-45 yr was surveyed by e-mail and asked to list the five most common activities they participated in during five historical time periods. Snowball sampling was used to generate additional respondents (24). The e-mail message asked participants to forward the survey to any female they knew in the designated age range who would be willing to respond. Using this technique, we surveyed 105 women. From these e-mail responses, we identified 35 activities in which the women most frequently participated across the five time periods. The 36 activities were primarily leisure, sport, and traditional exercise activities, and they were all included in the final edition of the BLHQ (Appendix A).
The final edition of the BLHQ consisted of five reference periods to assist participants with recall: elementary school (grades K-5), junior high school (grades 6-8), high school (grades 9-12), young adult (age 18-29), and adult (age 30-45). The questionnaire also asks for the living environment (urban, suburban, or rural) of the participant during each reference period because this may suggest a habitual level of physical activity. In addition, other questions to help characterize physical activity during childhood were asked, such as "would family and friends consider you a tom boy?"
The fourth and final step was developing the scoring of the BLHQ. Rather than scoring the activities with metabolic equivalent values (METs) to calculate energy expenditure (11,23,25), bone loading unit values for the hip and spine were developed to calculate bone loading at these two specific skeletal sites. Thirteen experts including exercise physiologists, biomechanists, and bioengineers were asked to rate the list of 36 activities on the BLHQ with regard to the load magnitude each activity placed on the hip and on the spine. Experts considered ground reaction forces and forces applied by muscle attachments when rating load magnitude for each activity. The load magnitude values provided to the experts ranged from 1 (lowest) to 3 (highest). The load magnitude ratings for the hip and for the spine provided by the individual experts were averaged to determine the load scores for each of the 36 activities listed on the BLHQ. Unique load scores for the hip and for the spine were then generated.
The bone loading units (BLU) were then calculated for each activity using the average load score and a loading rate score for both the hip and the spine. The loading rate score reflects the rate at which force is applied to the bone, and it ranged from a low value of 1, the lowest loading rate score for such activities as weight lifting and swimming, to a high value of 3 for such activities as gymnastics and basketball. A loading rate score of 3 was only assigned to activities that involved impact. Because loading rate is more important to skeletal development than the load magnitude alone, it was weighted three times that of the load magnitude (27). After weighting the loading rate score, the load score and loading rate score were added together. An example of the steps used to calculate BLU for each individual activity, at the hip or the spine, is illustrated in step 1 (Table 1). The activity-specific BLU values at the hip or the spine ranged from 4 (i.e., swimming) to 12 (gymnastics). As new activities were listed by respondents for which BLU were not already calculated, we assigned a BLU most similar to a previously determined activity (Example: soccer BLU was assigned to rugby, a newly listed activity that was similar to a previously listed activity). Table 2 shows a comparison between the newly calculated BLU and ground reaction forces from the literature as cited in Groothausen et al. (16) for common activities. A list of all activities and their individually calculated spine and hip BLU can be found in Appendix B
Steps for calculating bone loading across the reference periods.
Comparison of newly calculated hip and spine bone loading units and ground reaction forces.
A bone loading value for the hip and the spine for each activity was then calculated from the BLU and the self-reported participation characteristics (step 2, Table 1). Bone loading was calculated as the product of the BLU for the activity, the number of seasons (based on four, 3-month seasons per year), the number of years, and the frequency of participation during the specific reference period. One 3-month season was weighted as 0.25, two seasons as 0.50, three seasons as 0.75, and four seasons as 1. The scoring for seasons was adopted because meaningful and significant changes in aBMD, either with physical activity intervention or with the cessation of training, can only be detected over several months of observation (29). Therefore, individual months of participation were not considered. Frequency of participation was assigned a value of 1-4 based on the following scale: 1-3 times a month = 1, 1-2 times a week = 2, 3-5 times a week = 3, > 5 times a week = 4. Because the duration of individual bouts of activity does not need to be extensive to be beneficial to BMD (27), duration was not factored into these calculations. However, the questionnaire instructions stated that the participant should only report activities that were performed regularly.
Because there are unique BLU for the hip and the spine for each activity, bone loading was calculated individually for these two skeletal sites. After calculating a bone loading value for each activity for the hip and for the spine, the activity-specific values were added together to determine hip and spine bone loading for the five reference periods: elementary school (grades K-5), junior high school (grades 6-8), high school (grades 9-12), young adult (age 18-29), and adult (age 30-45) (Table 1, step 3). All of the bone loading values from each of the reference periods were summed (Table 1, step 4) and labeled as the total spine bone loading score or as the total hip bone loading score. All of the bone loading values from the most recent reference period (either the young adult or adult) were also summed for each individual and labeled as the recent spine bone loading score or as the recent hip bone loading score. The total spine and hip bone loading scores were used to assess reliability of the questionnaire.
The total bone loading scores for spine and hip varied considerably among the reference periods within each of the participants and also varied due to the wide age range of the participants contributing different number of years of activity data (13-40 yr). Therefore, the total bone loading score is not the best representation of bone loading exposure across the total time period assessed by the BLHQ. Thus, average annual and recent bone loading exposure was calculated individually for the spine and the hip. For each individual, average annual total bone loading exposure was calculated as the total spine or hip bone loading score, divided by the total number of years assessed by the questionnaire. For example, a 40-yr-old woman provided 35 yr of activity data, given that the questionnaire began with kindergarten, which most begin at age 5. Therefore, for a 40-yr-old woman, the bone loading exposure at the spine was calculated as her total spine bone loading score divided by 35 (Table 1, step 5).
Recent bone loading scores varied a great deal by the number of years completed in the recent reference period. Recent bone loading exposure was calculated for each individual as the sum of all spine or hip bone loading values for all activities in the most recent reference period, divided by the number of years completed in that reference period. For example, a 40-yr-old woman would have completed 10 yr in the adult reference period, which includes ages 30-45 yr. Recent spine bone loading exposure for a 40-yr-old woman was calculated as her spine bone loading score in the adult reference period divided by 10 (Table 1, step 6). Bone loading exposure and recent bone loading exposure for the hip and spine were used to assess the biological meaningfulness of the BLHQ.
Recruitment.
To evaluate the questionnaire, healthy premenopausal women (N = 80, aged 18-45 yr) were recruited from the University of Utah and the surrounding community. Recruitment consisted of making personal contacts, distributing and posting announcements, and communicating by word of mouth. The inclusion criteria were healthy adult, premenopausal women between the ages of 18-45 with a wide range of physical activity habits at the time of study recruitment. Only premenopausal women were recruited for participation because the skeletal response to physical activity intervention in postmenopausal women is less predictable and typically blunted, even when hormone replacement status is considered6. An equal number of women were recruited based on their high (N = 27), moderate (N = 26), and low (N = 27) current exercise habits assessed by their self-reported exercise frequency (high: > 3× wk−1, moderate: 2-3× wk−1, low: < 2× wk−1). Because each woman was assessed for aBMD, the following exclusion criteria were also applied: current medication use known to alter bone metabolism, history of or current oligomenorrhea or amenorrhea, history of or current eating disorder, current cigarette smoking, body mass index (BMI) greater than 30 kg·m−2, current pregnancy, and/or clinical diagnosis of osteoporosis. All women provided written informed consent, and the institutional review board of the University of Utah approved all procedures.
Questionnaire administration.
Historical physical activity was assessed using the newly developed BLHQ. For consistency, one researcher (S.D.) administered then reviewed the questionnaire after participant completion. To assess the reproducibility of the BLHQ, participants completed the questionnaire two times. The second completion was within 4-6 wk of the first assessment.
Bone mineral density and body composition assessment.
To provide evidence of the biological meaningfulness of the BLHQ, areal BMD (aBMD, g·cm−2) of the lumbar spine (L2−4) and right proximal femur (femoral neck) was assessed in each participant using dual energy x-ray absorptiometry (DXA, Hologic QDR 1000/W). A whole-body scan was also performed to assess body composition (% fat). One certified technician performed all DXA scans. Daily scans of the manufacturer-provided spine phantom for quality control revealed no machine drift (coefficient of variation, CV < 0.05%) during the study period. Short-term precision was evaluated to examine the reproducibility of aBMD and body composition outcomes. Sixteen participants volunteered to undergo two additional assessments of each skeletal site on different days, but not more than 1 wk apart. The three scans were used to calculate the following CV for the aBMD regions of interest and for body fat: 1.5% for lumbar spine, 1.6% for femoral neck, and 1.3% for fat mass. Body weight and height were measured on a traditional physician's balance scale and attached stadiometer. Participants were dressed in light clothing and stocking feet for these assessments.
Statistical analysis.
To determine recruitment related differences in age, body size, body composition, aBMD, and self-reported total spine and hip bone loading by the self-reported physical activity frequency at the time of recruitment, a one-way analysis of variance (ANOVA) was used. If significant differences were found, Tukey's HSD post hoc test was used to determine where the differences existed. To evaluate reproducibility of the questionnaire, we calculated intraclass correlation coefficients (ICC) between the first and second completion of the BLHQ. This was calculated for spine and hip bone loading within individual reference periods as well as for total spine and hip bone loading from all references periods combined. To assess the magnitude of agreement and the within-subject variability between the two repeated assessments of the questionnaire, Bland-Altman plots were created (7).
Higher BMI was correlated with higher spine and hip (femoral neck) aBMD, and higher BMI was also correlated with lower recent spine and hip bone loading exposures. Thus, partial correlation analyses were used to determine the magnitude of the correlations between spine and hip aBMD with spine and hip bone loading exposures after statistically adjusting for the confounding influence of BMI.
Lastly, logistic regression was used to determine whether women with low aBMD were more likely to report low bone loading exposures. Specifically, each of the lowest tertiles of spine and hip bone loading exposures were used to predict the lowest tertile of aBMD of the lumbar spine and femoral neck. Categorical confounding factors in this analysis included the presence or absence of mature age (> 29 yr), oral contraceptive use, low calcium intake (lowest tertile), and overweight (BMI > 25 kg·m−2). The adjusted odds ratios and 95% confidence intervals were calculated.
The natural log transformations of spine and hip bone loading exposures were used for all parametric statistical hypothesis testing because the distributions of values were skewed. The log-transformed values were used to compare the activity recruitment groups, to assess reliability, and for the correlation analysis. However, the untransformed, raw values for total bone loading and for bone loading exposure are presented in Table 3 for descriptive purposes. Statistical significance was set a priori at α = 0.05. Data are presented as means ± standard deviation (SD). All data were analyzed with SPSS software, version 13.0.
Characteristics of participants in stratified physical activity recruitment groups (N = 80).
RESULTS
The mean age of the total sample was 31 ± 7.7 yr (Table 3). The low-activity group at the time of recruitment had, on average, higher body fat percentage than those in the moderate- and high-activity groups (P = 0.0001). Self-reported total spine and hip bone loading and total spine bone loading exposure was significantly lower in the low-activity group when compared with the high-activity group. No other differences were observed among the activity groups.
Table 4 shows the intraclass correlation coefficients for reproducibility of the questionnaire. Intraclass correlation coefficients were r = 0.92, P < 0.001 for total self-reported spine bone loading and r = 0.89, P < 0.001 for total self-reported hip bone loading. The intraclass correlation coefficients for the individual reference periods ranged between 0.81 for the elementary school reference period and 0.95 for the adult reference period.
Intraclass correlation coefficient between total bone loading scores from questionnaire completion 1 and 2 of BLHQ for hip and spine.
The Bland-Altman plots were performed to determine whether systematic differences existed between the first and second administration of the questionnaire and to observe the range of within-subject variability between individual subjects. The plots showed reasonable agreement between the two reported assessments of the BLHQ over 4-6 wk (Fig. 1). There was a small bias, on average, towards higher self-reported total bone loading scores at the first versus the second assessment of the BLHQ (+64 hip, +57 spine). The magnitude of within-subject variability for total bone loading was plotted as ± 2 SD of the mean difference and ranged from −843 to +971 for total hip bone loading and from −732 to +846 for total spine bone loading (Fig. 1).
Bland-Altman plots showing the difference in total bone loading scores from repeated measures of the BLHQ plotted against the average of BLHQ scores from repeated administrations. The lines represent the mean difference score (solid) ± 2 SD (dashed).
Higher spine bone loading exposures (r = 0.338; P = 0.002) and higher hip bone loading exposures (r = 0.317; P = 0.004) were significantly correlated with higher femoral neck aBMD, after adjusting for BMI. Neither recent spine bone loading exposure nor recent hip bone loading exposure were correlated to either measure of aBMD. No measure of bone loading exposure was correlated with aBMD of the lumbar spine.
There were nonsignificant unadjusted odds for low aBMD (lowest tertile) among women reporting low total and recent bone loading exposures (lowest tertile). However, after adjusting for older ages, OC use, low calcium intake, and an overweight BMI, there were increased odds for low femoral neck aBMD among those reporting low recent hip bone loading exposure (OR = 3.62; CI = 1.09−11.95; P = 0.035). In this analysis, there was a reduced odds for low femoral neck aBMD among those with an overweight BMI (OR = 0.20; CI = 0.04−0.94; P = 0.041). In addition, after adjusting for older ages, OC use, low calcium intake, and an overweight BMI, there were increased odds for low femoral neck aBMD among those reporting low total and recent spine bone loading exposures (OR = 3.69; CI = 1.14−11.93; P = 0.029). The magnitude of the elevated adjusted odds for low femoral neck aBMD among those reporting a low recent spine bone loading exposure was identical to the adjusted odds for low femoral neck aBMD among those reporting a low recent hip bone loading exposure. This is because the same group of women who reported a recent spine bone loading exposure in the lowest tertile also reported a recent hip bone loading exposure in the lowest tertile.
DISCUSSION
We have developed a new questionnaire, based on sound measurement principles, to assess historical bone loading at the spine and the hip, which are two clinically relevant skeletal sites that are used to assess fracture risk. The primary unique feature of the BLHQ is that it quantifies skeletal loading rather than focusing on metabolic indices of activity participation. The findings indicate that the BHLQ has good test-retest reliability between two assessments that occurred 4-6 wk apart in premenopausal women (Table 4). Although there was variability in the agreement between BLHQ scores from one assessment to the next, there was only a small measurement bias toward slightly lower scores for the second assessment (Fig. 1). The BLHQ also demonstrated biological plausibility due to its associations with aBMD. After statistically adjusting for interindividual differences in BMI, total spine and hip bone loading exposures were significantly correlated with aBMD of the femoral neck. In addition, after statistically adjusting for older ages, OC use, and an overweight BMI, individuals with low femoral neck aBMD were 3.6-3.7 times as likely to report low total and recent spine bone loading and low recent hip bone loading exposures.
The BLHQ was designed to measure historical physical activity by scoring activity exclusively with bone loading units. This type of questionnaire is important because of the demonstrated benefits of the magnitude and rate of mechanical force on the skeleton (27). Both of these force elements were incorporated into the calculation of bone loading units. In addition, the questionnaire identifies the skeletal exposure to physical activity over the lifetime as well as in more recent years in premenopausal women, who are vulnerable to bone mineral losses yet who are not routinely screened for low BMD. The bone loading units have not been validated against objective measures of ground reaction or other force units in our sample, as this would have been an unrealistic task. True validation would have required measuring actual forces for all activities at all points of time that were recalled by participants. Rather, we relied on published literature (17,26,27,30), input from experts familiar with the biomechanics of activity, and reported ground reaction forces of specific activities (4,5,18,20) to develop the bone loading units. In effect, the present study used a similar process as the one used to develop the compendium of physical activities (1,2). The compendium quantified the effect of exercise intensity on the cardiovascular system, whereas the present study quantified exercise intensity with respect to the skeletal system. Application of the bone loading units to assess skeletally specific exercise intensity of two skeletal sites with the BLHQ makes it a useful alternative to caloric expenditure-based physical activity questionnaires for estimating historical bone loading. Future studies should directly compare the BLHQ with caloric expenditure-based lifetime physical activity questionnaires in their ability to identify premenopausal women who may be most at risk for developing osteopenia and osteoporosis later in life.
The BLHQ was found to have good reproducibility of bone loading exposure in premenopausal women as assessed by ICC between administration 1 and administration 2, which occurred 4-6 wk apart (r = 0.89-0.95). Recalling lifetime physical activity with an interviewer-administered questionnaire has been shown to be reliable by Friedenreich et al. (12), who developed a questionnaire to assess lifetime physical activity based on MET values in middle-aged women (mean age 61.2 yr). Their instrument covered a comparable length in time to the current study (childhood to current age), with high reproducibility for exercise/sport activities (r = 0.72). In a more recent study by Chasan-Taber and colleagues (9), reproducibility coefficients of 0.82 for total lifetime physical activity were reported with a self-administered questionnaire. Similar to other published studies (9,12), the magnitude of the ICC for the BLHQ was greater in more recent time periods than it was in earlier (elementary, junior high) time periods. Other studies also used different time periods between repeated questionnaire administrations, but reported similar results, with the time between repeated questionnaire administrations ranging from 1 wk to 1 yr (9,12).
To our knowledge, no lifetime physical activity questionnaires have used Bland-Altman plots as a method to assess reliability. Although ICC assess the reproducibility of group data, they do not capture the extent to which individual scores at the first administration agree with individual scores at the second administration. The Bland-Altman plots show the degree of response variability within the subjects (Fig. 1). In the present study, the magnitude of within-subject variability (± 2 SD) on repeated assessments of total spine and total hip bone loading on the BLHQ was the equivalent to approximately 60-70% of the mean total bone loading values for the whole group. However, on average, there was very little bias in the self-reported differences in bone loading, as the mean total hip and total spine bone loading scores were only approximately 5% lower on the second assessment as compared with the first. Because we are unaware of published data of Bland-Altman plots for other physical activity questionnaires with the intent of assessing lifetime physical activity, it is difficult to determine whether the observed within-subject variability in the repeated measures of the BLHQ is typical. However, our finding of no correlation between the difference scores and the mean scores suggests that the disagreement between repeated administrations of the BLHQ is not dependent on the magnitude of the self-reported bone loading (Fig. 1).
Participants were purposely recruited to represent a wide range of physical activity habits based primarily on the frequency of self-reported participation in all types of physical activity at the time of data collection. When the three physical activity groups were compared for demographic characteristics, the only difference among them was body fat percentage, which was highest in the low-activity group. Total spine and hip bone loading and total spine bone loading exposure were significantly different between the low- and high-activity groups. In contrast, there were no differences in aBMD among these recruitment-based physical activity groups. Therefore, we were successful at recruiting women with different levels of current physical activity and body fat percentages, but this simple grouping based on frequency alone was not adequate for detecting significant differences in aBMD (Table 3).
The final test of the BLHQ was to determine its biological plausibility. Thus, we examined whether self-reported bone loading was associated with aBMD in a sample of healthy and normally active premenopausal women. Current literature supports that BMD is influenced by exposure to mechanical loading and unloading across the lifespan in children, adolescents, and young women (19,28). This is why it is important to consider both lifetime as well as more recent physical activity when self-reported bone loading activities with aBMD. Our data support this contention in that after adjusting for BMI, higher total spine and hip bone loading exposures were significantly correlated with higher femoral neck aBMD. The magnitude of these partial correlations is also consistent with previously reported correlations between self-reported lifetime physical activity and aBMD (8). Moreover, after adjusting for older ages, OC use, and an overweight BMI, women with low femoral neck aBMD had significantly greater odds of reporting low recent spine and hip bone loading exposures. In the present study, recent spine and hip bone loading exposures were calculated for the most recent reference period, which spans multiple years rather than merely considering the past year. In fact, the average time span of the most recent reference period was approximately 7 yr for both the younger (18-29 yr) and the older (30-45 yr) women in the study. Because the skeleton is much slower to respond to physical activity than other biologic systems, a longer time frame than a single year, which is a typical time frame for many caloric expenditure-based physical activity questionnaires, may be necessary to assess the relationship between "recent" bone loading physical activity and aBMD. Taken together, these data suggest that BLHQ-derived estimates of both lifetime and more recent bone loading exposures were associated with femoral neck aBMD in premenopausal women.
From a practical perspective, the BLHQ is a self-administered questionnaire that takes about 30-40 min to complete. Women in the current study reported little difficulty completing the BLHQ. However, some simplification of the current calculation procedures may be in order. Our attempts to distinguish spine and hip bone loading as individual variables may have been unnecessary. The correlation between total spine and total hip bone loading was very high (r = 0.988; P < 0.0001), as was the correlation between recent spine and hip bone loading (r = 0.992; P < 0.0001). The data on ground reaction forces are somewhat more relevant for understanding the bone mineral response to exercise at the hip than at the spine. This is due, in part, to the attenuation of ground reaction forces that occur along the kinetic chain of the lower limb. Further, the spine is more likely than the hip to reflect metabolic and hormonal influences on the skeleton, as in the case of menstrual cycle dysfunction. Therefore, it is likely that only the calculation of hip bone loading and hip bone loading exposure is necessary.
This study has limitations. In our effort to exclude women from the sample who may have had low aBMD as a function of anything other than low bone loading, it resulted in a very homogeneous sample with respect to lumbar spine aBMD. The percentage of the interindividual range (range/minimum) × 100)) in lumbar spine aBMD was only 58%, whereas the interindividual range in femoral neck aBMD was 107%. This may be why the BLHQ was able to detect relationships of self-reported bone loading with femoral neck aBMD but not with lumbar spine aBMD. Further assessment of the BLHQ with a larger and more diverse sample with respect to bone loading and aBMD would be helpful to determine whether the absence of a relationship between self-reported bone loading and lumbar spine aBMD was a type II error. In addition, the next revision of the BLHQ should include the assessment of bone loading for the past year. This would allow a better comparison between the BLHQ and other caloric expenditure-based physical activity questionnaires, which would be especially helpful in quantifying the extent to which the BLHQ yields a more accurate prediction of aBMD. The current version of the BLHQ does not include home activities listed as physical activity options. Thus, our estimates of bone loading may be underestimated. Furthermore, the reported correlations between bone loading and aBMD may have been reduced in magnitude since it has been recently reported that home physical activity is associated with lumbar spine and femoral neck in premenopausal women (15). Self-reporting of physical activity over a lifespan is a difficult task, and we recognize that the present data include errors in recall. Lastly, the inability to completely validate the BLHQ with activity-specific force measurements is a limitation. However, the appropriate methodological and theoretical steps were followed in developing the BLHQ.
In summary, the BLHQ demonstrates that meaningful and reliable data regarding relationships between lifetime physical activity and premenopausal bone health can be obtained from a self-report questionnaire. Furthermore, the present data support the importance of sustaining bone loading physical activity across the lifespan to achieve and to maintain optimal bone health.
We would like to acknowledge Dr. Barbara Drinkwater and Dr. Gail Dalsky for their contributions to the concept and formative work in the development of the Bone Loading History Questionnaire.
This study was used to complete the requirements for the PhD degree by Shawn H. Dolan at the University of Utah.
REFERENCES
1. Ainsworth, B. E., W. L. Haskell, A. S. Leon, et al. Compendium of physical activities: classification of energy costs of human physical activities. Med. Sci. Sports Exerc. 25:71-80, 1993.
2. Ainsworth, B. E., W. L. Haskell, M. C. Whitt, et al. Compendium of physical activities: An update of activity codes and MET intensities. Med. Sci. Sports Exerc. 32:S498-S516, 2000.
3. Ainsworth, B. E., J. M. Shaw, and S. Hueglin. Methodology of activity surveys to estimate mechanical loading on bones in humans. Bone 30:787-791, 2002.
4. Anderson, D. D., B. M. Hillberry, D. Teegarden, W. R. Proulx, C. M. Weaver, and T. Yoshikawa. Biomechanical analysis of an exercise program for forces and stresses in the hip joint and femoral neck. J. Appl. Biomech. 12:292-312, 1996.
5. Bailey, D., R. Faulkner, and H. McKay. Growth, physical activity, and bone mineral acquisition. Exerc. Sport Sci. Rev. 24:233-266, 1996.
6. Bassey, E. J., and S. J. Ramsdale. Weight-bearing exercise and ground reaction forces: a 12-month randomized controlled trial of effects on bone mineral density in healthy postmenopausal women. Bone 16:469-476, 1995.
7. Bland, J. M., and D. G. Altman. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1:307-310, 1986.
8. Brahm, H., H. Mallmin, K. Michaelsson, H. Strom, and S. Ljunghall. Relationships between bone mass measurements and lifetime physical activity. Calcif. Tissue Int. 62:400-412, 1998.
9. Chasan-Taber, L., J. B. Erickson, J. W. McBride, P. C. Nasca, S. Chasan-Taber, and P. S. Freedson. Reproducibility of a self-administered lifetime physical activity questionnaire among female college alumnae. Am. J. Epidemiol. 155:282-289, 2002.
10. Crocker, L. M. Introduction to Classical and Modern Test Theory. New York: Harcourt, 1986, pp. 362-371.
11. Damilakis, J., K. Perisinakis, G. Kontakis, E. Vagios, and N. Gourtsoyiannis. Effect of lifetime occupational physical activity on indices of bone mineral status in healthy premenopausal women. Calcif. Tissue Int. 64:112-116, 1999.
12. Friedenreich, C. M., K. S. Courneya, and H. E. Bryant. The lifetime total physical activity questionnaire: development and reliability. Med. Sci. Sports Exerc. 30:266-274, 1998.
13. Fuchs, R. K., J. J. Bauer, and C. M. Snow. Jumping improves hip and lumbar spine bone mass in prepubescent children: a randomized and controlled trial. J. Bone Min. Res. 16:148-156, 2001.
14. Greendale, G. A., E. Barrett-Connor, S. Edelstein, S. Ingles, and R. Haile. Lifetime leisure exercise and osteoporosis. Am. J. Epidemiol. 141:951-959, 1995.
15. Greendale, G. A., M. H. Huang, Y. Wang, J. S. Finkelstein, M. E. Danielson, and B. Sternfeld. Sport and home physical activity are independently associated with bone density. Med. Sci. Sports Exerc. 35:506-512, 2003.
16. Groothausen, J., H. Siemer, H. C. G. Kemper, J. Twisk, and D. C. Welten. Influence of peak strain on lumbar bone mineral density: an analysis of 15-year physical activity in young males and females. Pediatr. Exerc. Sci. 9:159-173, 1997.
17. Karlsson, K. M., H. Magnusson, C. Karlsson, and E. Seemna. The duration of exercise as a regulator of bone mass. Bone 28:128-321, 2001.
18. Kemper, H. C. G., I. Bakker, J. W. R. Twisk, and W. Van Mechelen. Validation of a physical activity questionnaire to measure the effect of mechanical strain on bone mass. Bone 30:799-804, 2002.
19. Kemper, H. C. G., W. Van Mechelen, G. B. Post, J. C. Roost, and P. Lips. A fifteen-year longitudinal study in young adults on the relation of physical activity and fitness with the development of the bone mass: the Amsterdam growth and health longitudinal study. Bone 27:847-853, 2000.
20. Khan, K., H. A. McKay, H. Haapalaso, et al. Does childhood and adolescence provide a unique opportunity for exercise to strengthen the skeleton? J. Sci. Med. Sport 3:150-164, 2000.
21. Marcus, R. Role of exercise in preventing and treating osteoporosis. Rheum. Dis. Clin. North Am. 27:131-141, 2001.
22. Melton, L. J., E. J. Atkinson, M. K. O'Connor, W. M. O'Fallon, and B. L. Riggs. Determinants of bone loss from the femoral neck in women of different ages. J. Bone Min. Res. 15:24-31, 2000.
23. Nelson, M. E., M. A. Fiatarone, C. M. Morganti, I. Trice, R. A. Greenberg, and W. J. Evans. Effects of high-intensity strength training on multiple risk factors for osteoporotic fractures. JAMA 272:1909-1914, 1994
24. Neutens, J. J., and L. Rubinson. Research Techniques for the Health Sciences. Boston: Allyn and Bacon, 1997, p. 125.
25. Teegarden, D., W. R. Proulx, M. Kern, et al. Previous physical activity relates to bone mineral measures in young women. Med. Sci. Sports Exerc. 28:105-113, 1996.
26. Turner, C. H. Three rules for bone adaptation to mechanical stimuli. Bone 23:399-407, 1998.
27. Turner, C. H., and A. G. Robling. Designing exercise regimens to increase bone strength. Exerc. Sport Sci. Rev. 31:45-50, 2003.
28. Winters, K. M., and C. M. Snow. Detraining reverses positive effects of exercise on musculoskeletal system in premenopausal women. J. Bone. Min. Res. 15:2495-2503, 2002.
29. Witzke, K. A., and C. M. Snow. Effects of plyometric jumping on bone mass in adolescent girls. Med. Sci. Sports Exerc. 32:1051-1057, 2000.
30. Wolff, I., J. J. van Crooneborg, H. C. G. Kemper, P. J. Kostense, and J. W. R. Twisk. The effect of exercise training programs on bone mass: a meta-analysis of published controlled trials in pre-and postmenopausal women. Osteoporos. Int. 9:1-12, 1999.
APPENDIX A. Physical Activity History Questionnaire BLHQ-1
APPENDIX B. The index of bone loading units (BLU) for each activity.
Keywords:
BONE LOADING; PREMENOPAUSAL; AREAL BMD; PHYSICAL ACTIVITY QUESTIONNAIRE
©2006The American College of Sports Medicine