PEDIATRICS Vol. 107 No. 6 June 2001, pp. 1387-1393

Physical Activity and Bone Measures in Young Children: The Iowa Bone Development Study

Kathleen F. Janz, EdD^*, , Trudy L. Burns, PhD^{, §,} , James C. Torner, PhD, Steven M. Levy, DDS^{, ¶}, Richard Paulos, BS^§, Marcia C. Willing, MD, and John J. Warren, DDS^¶

From the Departments of ^* Health, Leisure, and Sport Studies,  Epidemiology, ^§ Biostatistics,  Pediatrics, and ^¶ Preventive and Community Dentistry, University of Iowa, Iowa City, Iowa.

	ABSTRACT

Top Abstract Methods Results Discussion Conclusion References

Objectives.  Physical activity has a beneficial effect on bone development in circumpubertal children, although its effect on younger children is uncertain. In this cross-sectional study, we examined associations between physical activity and bone measures in 368 preschool children (mean age: 5.2 years, range: 4-6 years).

Design.  Physical activity was measured using 4-day accelerometry readings, parental report of children's usual physical activity, and parental report of children's hours of daily television viewing. Total body and site-specific bone mineral content and area bone mineral density (BMD) were measured by dual energy radiograph absorptiometry.

Results.  After adjustment for age and body size, accelerometry measures of physical activity and parental report of usual physical activity were consistently and positively associated with bone mineral content and BMD in both boys and girls (r = 0.15-0.28). Television viewing was inversely associated with hip BMD in girls (r = 0.15). The proportion of variance in bone measures explained by physical activity in linear regression models ranged from r² = 1.5% to 9.0%. In all of these models except total body BMD, at least 1 and often several of the physical activity variables entered as independent predictors. Activity variables most likely to enter the regression models were vigorous physical activity (as determined by accelerometry) and parental ranking of child's usual physical activity.

Conclusions.  Findings indicate that there are statistically significant and, perhaps important, associations between physical activity and bone measures during early childhood, well ahead of the onset of peak bone mass. This would suggest that intervention strategies to increase physical activity in young children could contribute to optimal bone development.  Key words:  accelerometry, bone density, bone mass, exercise, growth.

Strains on bone greater than needed for steady state remodeling will cause a modeling response that increases bone mass to meet the increasing load requirement.^1-4 This adaptive response occurs primarily during periods of growth and development, suggesting the importance of understanding the effects of habitual physical activity on bone accretion during childhood and adolescence.^3,5,6 Bailey et al⁷ and, more recently, Kemper⁸ have summarized pediatric study findings and concluded that persuasive evidence exists to show that physical activity, particularly weight-bearing and strength-training activity, has a positive effect on bone acquisition during the circumpubertal years. However, very few studies have examined the role of physical activity on bone development in younger children, leaving incomplete the understanding of physical activity's cumulative effect on bone acquisition, the types of activity most effective for bone adaptation in young children, and whether it is possible to increase bone acquisition during childhood through increases in physical activity.⁸ In this study, using a cohort of 368 healthy, 4- to 6-year-old children, we report cross-sectional associations between physical activity and dual energy radiograph absorptiometry (DXA) measures of bone mineral content, area bone density, and bone area total.

	METHODS

Top Abstract Methods Results Discussion Conclusion References

Study Population

Children and their parents for this study were volunteers from 890 families previously recruited and then participating in the Iowa Fluoride Study, a longitudinal study of fluoride intake and dental fluorosis. They had been recruited from 8 Iowa hospitals between 1992 and 1995 immediately postpartum.⁹ Three hundred sixty-eight children completed baseline bone and physical activity measures at the time of this report. Almost all (96%) of the cohort were white.

Skeletal Bone Density Measures

Bone measurements of the hip, lumbar spine, and total body were performed at The University of Iowa in the Clinical Research Center using a Hologic QDR-2000 DXA unit (Hologic, Inc, Waltham, MA) with a fan-beam geometry and a multiple detector array. Accuracy of the fan-beam measurement has been compared with pencil-beam results and has demonstrated high correlation (r = 0.97-0.99) in the anterioposterior spine and femoral neck regions. Reliability between scans was achieved using phantom calibration. Coefficients of variation have been found to be 1% to 2% in bone density measurements of hip, lumbar spine, and total body.^10,11 Total hip and anterioposterior view of the lumbar spine were scanned using the array mode. This measurement allowed for a very precise measure of bone density with very low radiation doses (0.59 and 0.18 mrem, respectively, for hip and lumbar spine). Bone mass (bone mineral content [BMC]), expressed in grams, reflects the absolute amount of mineral in the selected bone. Bone mineral density (BMD), expressed in g/cm², is a relative value of the amount of bone mineral divided by the area of the selected bone. Because DXA scans do not create a true volume measure (ie, g/cm³), BMD is considered an area density.

Anthropometric Measures

The height of each child was measured using a stadiometer, and weight was measured using standard physician's scale during the DXA examination visit. Both devices were routinely checked for accuracy and precision. Children were weighed and measured while wearing indoor clothes, but without shoes. Heights were recorded in tenths of centimeters and weights in tenths of kilograms.

Parental Report and Accelerometry-Determined Measures of Physical Activity

During the DXA examination visit, parents were asked to compare their child's usual physical activity level to peers using a 5-point Likert scale. Slemenda and colleagues¹² have shown that this question is associated with bone density in older children (mean 9 years old). Parents also reported their child's usual hours of television viewing per day (TV viewing). Previous research indicates that this question is a useful marker of sedentary habits (and possibly dietary habits) in children and adolescents.^13,14 In addition to parental reports, physical activity was assessed using accelerometry. At the DXA examination visit, using a standardized protocol, parents and children were shown a CSA uniaxial accelerometer (Model Number 7164, Computer Science Application, Shalimar, FL) fastened to a nylon belt. Parents were instructed to attach the belt at their child's waist (on the midaxillary line) and complete a short data recording sheet noting what times the monitor was attached and detached. Children were asked to wear the monitors for 4 consecutive days, including 1 weekend day, during 1 of the fall months (September through November). Monitors and data recording sheets were sent to parents and returned via US mail. To be included in the data analyses, we required at least 8 hours of movement counts per day and at least 2 days of data per child.

The CSA accelerometer is designed to measure normal human movement using an internal piezoceramic cantilever beam that creates a charge proportional to the magnitude of movement. Movement values are accumulated and stored over a specified time period. For our study, the time period was 1 minute. The accelerometer was small (5.0 x 3.8 x 1.5 cm), lightweight (42 g), and had a black box design. It was positioned to directly measure the acceleration of the displacement of the hip, making it particularly sensitive to monitoring weight-bearing movement. Raw data, as movement counts per minute, were stored in ASCII format and downloaded using an interface and IBM-compatible computer.

Validation studies examining this accelerometer and the construction of summary variables for intensity of movement suggest that it is a valid and reliable measure of children's physical activity.^15-20 In preschool children, movement counts have been correlated with observational methods during gymnasium free-play (r = 0.87).¹⁵ In older children and adolescents, movement counts have been correlated with energy expenditure during treadmill exercise (r = 0.86) in a laboratory setting²⁰ and with whole-day heart rate monitoring in a field setting (r = 0.58-0.63).¹⁷ The intraclass correlation coefficient for 4 days of monitoring preschool and elementary school-aged children ranges from 0.74 (95% confidence interval: 0.67-0.80) to 0.77 (0.73-0.82).^21,22 From the accelerometry raw data we constructed 2 summary variables: 1) total daily physical activity, and 2) daily minutes spent in vigorous physical activity. Total daily physical activity was calculated as the total movement counts divided by total time of measurement (min). This variable represents volume of daily physical activity. Minutes spent in vigorous physical activity represents the daily frequency of strenuous movement. It was calculated by counting the number of minutes in which movement counts were 2972. For locomotive movement in children, this cutpoint value is approximately equivalent to 6 metabolic equivalents (METs) (standard error of estimate of 1.08 METs).¹⁶ Before inferential statistical analyses, minutes spent in vigorous physical activity was adjusted for total minutes of measurement. Figure 1 presents typical data output using whole-day accelerometry and our constructed summary variables.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Data output from whole-day accelerometry.

Statistical Analysis

Gender-specific analyses were conducted to examine the distributional properties of the variables. Means and standard deviations were calculated. Student's t tests and the Mann-Whitney tests, when variables were not normally distributed, were used to examine gender differences. When necessary for the assumption of normality, measures were transformed by calculating the natural logarithm. Gender-specific Pearson product moment correlation coefficients were calculated for physical activity, anthropometric variables, and age. After adjustment for age, weight, and height, gender-specific partial correlation coefficients between physical activity variables and each of the bone measures were calculated. General linear model procedures with Tukey pairwise comparisons were used for categorical analyses of the relationship between vigorous physical activity and bone measures of the hip. Continuous values for age, weight, and height and a categorical value (quartiles) for vigorous activity were used in these models. Finally, forward stepwise multiple linear regression models were constructed using bone measures as the dependent variables. Possible outlier observations were identified and their influence on fitted model parameters was assessed using regression diagnostic approaches. The final multiple regression models included age, weight, height, and those physical activity measures with associated P values <.05.

	RESULTS

Top Abstract Methods Results Discussion Conclusion References

Descriptive

Table 1 presents a description of the children's ages (at the time of the DXA scan), body size, and bone measures. Boys did not differ significantly from girls in age, weight, or height. Bone measures were greater in boys with the exception of BMC-hip, BMD- -spine, and area total. These values did not differ by gender. A description of the children's physical activity is presented in Table 2. We collected 1421 days of useable accelerometry data. Eighty-seven percent of the cohort wore the accelerometer 4 days, 11% wore it 3 days, and 2% wore it 2 days. There was no gender-related difference in number of days per child that the monitor was worn or the number of minutes the monitor was worn each day. The accelerometry-determined measures of physical activity indicated that the boys had a greater level of total physical activity than girls (766 cts/min compared with 701 cts/min). Boys also engaged in more vigorous physical activity than did girls (32 minutes/d compared with 24 minutes/d). The mean for physical activity compared with peers was slightly above average for both boys and girls, though this variable was higher for boys than for girls (3.73 compared with 3.43). TV viewing was about the same for boys and girls (~2 hr/d) and similar to parental reported TV viewing of 4-year-old children within the Framingham Children's Study.²³

Bivariate Analysis

No physical activity measure was significantly associated with weight or height. For girls, TV viewing (logged because it was positively skewed) was associated with age (r = 0.17). Partial correlation coefficients (adjusted for age, weight, and height) between physical activity and bone measures are presented in Table 3. For both boys and girls, BMC-hip, BMC-spine, BMD-hip, and area total were moderately associated with the 2 accelerometry-determined measures of physical activity, ie, total physical activity and vigorous physical activity (r = 0.15-0.28). BMC-total body in boys and BMD-spine in girls were also moderately associated with these accelerometry-determined physical activity measures (r = 0.15-0.19). In girls, physical activity compared with peers was associated with all bone measures (r = 0.15-0.26) except BMC-total body and BMD-total body. In boys, this variable was also associated with all bone measures (r = 0.18-0.25) except BMC-spine and BMD-total body. TV viewing (log) was inversely associated with BMD-hip in girls (r = 0.15).

Categorical Analysis

General linear model procedures with Tukey pairwise comparisons were used to assess categorical relationships between vigorous physical activity and bone measures of the hip. We selected the hip for this analysis because it is the segment of the skeleton thought to be most influenced by normal patterns of physical activity (ie, weight-bearing) and the site where accelerometry would be expected to be most sensitive to movement.^2-3,8 Our results are presented in Table 4 and Fig 2 and 3. In all models, quartiles of vigorous physical activity increased the explained variance for hip bone measures (P < .05). The total explained variance (r²) in the general linear models was 51% for hip BMC in boys, 61% for hip BMC in girls, 16% hip BMD in boys, and 34% hip BMD in girls. Pairwise comparisons indicated an 11.9% greater mean BMC-hip for children in the highest quartile⁴ of vigorous physical activity when compared with children in the lowest quartile.¹ Pairwise comparisons for BMD-hip suggested a 5% greater mean BMD for quartile 4 when compared with quartile 1. However, this difference was not statistically significant in boys (P = .056).

View larger version (8K): [in this window] [in a new window]   Fig. 2.   Least squares means for hip BMC by vigorous physical activity quartiles.

View larger version (8K): [in this window] [in a new window]   Fig. 3.   Least squares means for hip BMD by vigorous activity quartiles.

Linear Regression Analysis

Results from the regression analyses are presented in Fig 4, 5, and 6. With the exception of BMD-total body, at least 1 and often several of the physical activity variables entered as independent predictors of bone measures adjusted for age, weight, and height. Combined physical activity measures explained 1.5% to 9% of the variance in the bone measures, adjusted for age and body size. The variance, explained by each individual measure of physical activity, ranged from r² = 0.5% to 5.4%. In addition, the amount of explained variance in bone measures attributed to the various physical activity measures was similar for boys and girls. The physical activity measures most likely to enter regression models were vigorous physical activity and physical activity compared with peers. For both boys and girls, these variables were significant predictors of BMC-hip, BMD-hip, BMC-total body, and area total. Vigorous physical activity also predicted BMC-spine in boys and BMD-spine in girls. Physical activity compared with peers predicted BMC-spine in girls and BMD-spine in boys. TV viewing (log) was a significant predictor of BMC-hip, BMD-hip, and BMC-total body in girls. (The association between TV viewing and bone measures was inverse in direction, ie, the girls with the lowest bone measures were, on average, the same girls watching the greatest amount of television). Finally, total physical activity predicted BMC-spine in girls.

View larger version (45K): [in this window] [in a new window]   Fig. 4.   Explained variance (%) for hip and spine BMC (g) from stepwise linear regression.

View larger version (38K): [in this window] [in a new window]   Fig. 5.   Explained variance (%) for hip and spine BMD (g/cm2) from stepwise linear regression.

View larger version (51K): [in this window] [in a new window]   Fig. 6.   Explained variance (%) for total BMC (g) and total area (cm2) from stepwise linear regression.

	DISCUSSION

Top Abstract Methods Results Discussion Conclusion References

In contrast to examining young athletes where self-selection bias and unusually high levels of activity hamper external validity, our study is one of the first to examine the effects of usual everyday activity on bone accretion in young, nonathletic children. It demonstrates that physically active young children have greater site-specific BMC, site-specific BMD, total body BMC, and total bone mineral area than their less active peers. Our combined physical activity measures explained 1.5% to 9% of the variance in individual bone measures, adjusted for age and body size. The percentage of explained variance that we report is similar to previous observational studies of older, nonathletic children and adolescents.^5,12,24 When combined with these previous reports, findings suggest that physical activity is associated with bone measures throughout childhood.

In our study, minutes spent in vigorous physical activity tended to be the activity variable most highly associated with bone measures, and the activity variable that explained the greatest amount of variance in bone measures. This finding supports the current biomedical understanding that high levels of strain to the musculoskeletal system are more important to bone development than the volume of activity.^4,25,26 Results from our categorical analysis suggest an 11.9% difference in mean hip BMC between the least and most vigorously active children. On average, results from our regression analysis suggest that a 10-minute increase in daily vigorous activity in young children would result in hip BMC increases of 200 mg and spine BMC increases of 280 mg. In our cohort, these BMC values are equivalent to 3% of the mean BMC-hip and 2% of the mean BMC-spine. The relatively modest increase of 10 additional minutes of daily vigorous physical activity requires a minimal restructuring of routines and, theoretically, is well within the reach of US children. Our findings also indicate that reducing TV viewing by girls could increase BMC. However, because both boys and girls watched approximately the same amount of TV, but TV viewing did not predict bone measures in boys, the association between TV viewing and BMC in girls may simply reflect their low level of vigorous physical activity. It may be that boys, unlike girls, are reaching some minimal threshold of physical activity so that they can "afford" to watch television. In addition, it is possible that low levels of vigorous activity in girls during early childhood contribute to gender-related differences in peak bone mass later in life.

Limitations of our research include its cross-sectional design (causality should not be assumed) and the inability of our bone measurement method (DXA) to capture volume of bone (and, therefore, our inability to fully control for bone size associated with growth). It is also possible that the 1-minute time frame used to calculate vigorous activity may have missed shorter bouts of vigorous activity²⁷ and, therefore, underestimated the magnitude of the association between vigorous activity and bone measures. On the other hand, study strengths include the use of an objective method (accelerometry) that directly measured hip displacement, and therefore, weight-bearing physical activity and a relatively large cohort of children of approximately the same age and body size. These strengths allowed us to avoid many of the validity and reliability problems associated with self-reported activity in children and the low statistical power associated with small sample sizes of varying ages.²⁸ The relative homogeneity of age and body size of our cohort also minimized bone volume issues related to DXA.⁸

	CONCLUSION

Top Abstract Methods Results Discussion Conclusion References

Our research suggests that physical activity influences bone accretion before the circumpubertal years and, therefore, before the modulating effects of increasing sex hormones. The associations between activity and bone measures seem similar overall for boys and girls and are greatest in the weight-bearing region of the skeleton (hip). The implication is that more active children will have greater bone mass even before adolescence and early adulthood when physical activity is assumed to be most influential on bone development.^36-8 In addition, our findings suggest some young children are not engaging in enough activity to optimize bone health during a time in life when we assume they are most active.²⁹ Although confirmation of our findings using a prospective design and a more diverse population is needed, the associations that we report may be important in the long run because in adults, even small increases in BMC have been shown to reduce fracture rates.³⁰ Our study raises the possibility that the sooner children become active, the greater their bone accrual and the lower their risk for osteoporotic fractures later in life.

	ACKNOWLEDGMENTS

This work was supported by the following National Institutes of Health grants: National Institutes for Dental and Craniofacial Research RO1-DE12101 and RO1-DE09551, and General Clinical Research Centers Program, National Center for Research Resources RR00059.

We thank Joan Grabin, Jennifer Tisch, Barb Simon, Cynthia Moore, Tina Craig, Brett Norell, and Chuck Dufano for their organizational efforts. We also thank Julie Gilmore, Dr Phyllis Stumbo, and Dr Stewart Trost for technical assistance.

	FOOTNOTES

Received for publication Aug 14, 2000; accepted Nov 20, 2000.

Address correspondence to Kathleen F. Janz, EdD, Department of Health, Leisure, and Sport Studies, 102 FH, University of Iowa, Iowa City, IA 52242. E-mail: kathleen-janz{at}uiowa.edu

	ABBREVIATIONS

DXA, dual energy radiograph absorptiometry; BMC, bone mineral content; BMD, bone mineral density; TV viewing, television viewing per day; METs, metabolic equivalents.

	REFERENCES

Top Abstract Methods Results Discussion Conclusion References

1. Frost HM Muscle, bone, and the Utah paradigm: a 1999 overview. Med Sci Sports Exerc 2000; 32:911-917 [Medline]
2. Grimston SK, Willows ND, Hanley DA Mechanical loading regime and its relationship to bone mineral density in children. Med Sci Sports Exerc 1993; 25:203-1210
3. Snow-Harter C, Marcus R Exercise, bone mineral density, and osteoporosis. Exerc Sport Sci Rev 1991; 19:389-418 [Medline]
4. Umemura Y, Ishiko T, Yamauchi T, Kurono M, Mashiko S. Five jumps per day increase bone mass and breaking force in rats. J Bone Miner Res. 1997;12:9:1480-1485
5. Bailey DA, McKay HA, Mirwald RL, Crocker PR, Faulkner FA A six-year longitudinal study of the relationship of physical activity to bone mineral accrual in growing children: The University of Saskatchewan Bone Mineral Accrual Study. J Bone Miner Res 1990; 14:1672-1679
6. Faulkner RA, Houston CS, Bailey DA, Drinkwater DT, McKay HA, Wilkinson AA Comparison of bone mineral content and bone mineral density between dominant and non-dominant limbs in children 8-16 years of age. Am J Hum Biol 1993; 5:491-499 [CrossRef]
7. Bailey DA, Faulkner RA, McKay HA Growth, physical activity, and bone mineral acquisition. Exerc Sport Sci Rev 1996; 24:233-266 [Medline]
8. Kemper HC Skeletal development during childhood and adolescence and the effects of physical activity. Pediatr Exerc Sci 2000; 12:198-216
9. Levy SM, Kiritsy MC, Slager SL, Warren JJ Patterns of dietary fluoride supplement use during infancy. J Public Health Dent 1998; 58:228-233 [Medline]
10. Koo WWK, Walters J, Bush AJ Technical considerations of dual-energy x-ray absorptiometry-based bone mineral measurements for pediatric studies. J Bone Miner Res 1995; 10:1998-2004 [Medline]
11. Sato M, McClintock C, Kim J, Dual-energy x-ray absorptiometry of raloxifene effects on the lumbar spine and femora of ovariectomized rats. J Bone Miner Res 1994; 9:715-724 [Medline]
12. Slemenda CW, Miller JZ, Hui SL, Reister TK, Johnston Jr CC. Role of physical activity in the development of skeletal mass in children. J Bone Miner Res. 1991;6:11:1227-1233
13. Dietz WH, Gortmaker SL Do we fatten our children at the television set? Obesity and television viewing in children and adolescents. Pediatrics 1985; 75:807-812 [Abstract/Free Full Text]
14. Wong ND, Hei TK, Qaqundah PY, Davidson DM, Bassin SL, Gold KV Television viewing and pediatric hypercholesterolemia. Pediatrics 1992; 90:75-79 [Abstract/Free Full Text]
15. Fairweather SC, Reilly JJ, Grant S, Whittaker A, Paton JY Using the Computer Science and Applications (CSA) activity monitor in preschool children. Pediatr Exerc Sci 1999; 11:413-420
16. Freedson PS, Sirard J, Debold E, Calibration of the Computer Science and Applications, Inc (CSA) accelerometer [abstract]. Med Sci Sports Exerc 1997; 29:S45
17. Janz KF Validation of the CSA accelerometer for assessing children's physical activity. Med Sci Sports Exerc 1994; 26:369-375 [Medline]
18. Janz KF, Witt JD, Mahoney LT The stability of children's physical activity as measured by accelerometry and self-report. Med Sci Sports Exerc 1995; 27:1326-1332 [Medline]
19. Janz KF, Mahoney LT Maturation, gender, and video game playing are related to physical activity intensity in adolescents: the Muscatine study. Pediatr Exerc Sci 1997; 9:353-363
20. Trost SG, Ward DS, Moorehead SM, Watson PD, Riner W, Burke JR Validity of the Computer Science and Applications (CSA) activity monitor in children. Med Sci Sports Exerc 1998; 30:629-633 [Medline]
21. Janz KF, Paulos R, Burns TL, Levy SM Measuring physical activity in pre-school children: The Iowa Bone Growth Study [abstract]. Measurement in Physical Education and Exercise Science 2000; 4:188
22. Trost SG, Pate RR, Freedson PS, Sallis JF, Taylor WC Using objective physical activity measures with youth: How many days of monitoring are needed? Med Sci Sports Exerc 2000; 32:426-431 [Medline]
23. Moore LL, Nguyen UD, Rothman KJ, Cupples LA, Ellison RC Preschool physical activity level and change in body fatness in young children: the Framingham Children's Study. Am J Epidemiol 1995; 142:982-988 [Abstract/Free Full Text]
24. Welten DC, Kemper HC, Post GB, Weight-bearing activity during youth is a more important factor for peak bone mass than calcium intake. J Bone Miner Res 1994; 9:1089-1096 [Medline]
25. Groothausen J, Siemer H, Kemper HCG, Twisk J, Welten DC Influence of peak strain on lumbar bone mineral density: an analysis of 15-year physical activity in young males and females. Pediatr Exerc Sci 1997; 9:159-173
26. Rubin CT, Lanyon LE Regulation of bone mass by mechanical strain magnitude. The effect of peak strain magnitude. Calcif Tissue Int 1984; 37:411-417
27. Welk GJ, Corbin CB, Dale D Measurement issues in the assessment of physical activity in children. Res Q Exerc Sport 2000; 71:59-73
28. Sallis JF, Saelens BE Assessment of physical activity by self-report: status, limitations, and future directions. Res Q Exerc Sport 2000; 71:1-14 [Medline]
29. Department of Health and Human Services. Physical Activity and Health: A Report of the Surgeon General. Atlanta, GA: US Department of Health and Human Services, Public Health Service, CDC, National Center for Chronic Disease Prevention and Health Promotion; 1996:81-208
30. Lewis RD, Modlesky CM Nutrition, physical activity, and bone health in women. Int J Sports Nutr 1998; 8:250-284 [Medline]