A large number of children sustain fractures after relatively minor trauma and several investigators have associated these fractures to a deficient accumulation of bone during growth. This study was conducted to better characterize the skeletal phenotype associated with low-energy impact fractures of the forearm in girls. The densities of cancellous, cortical, and integral bone and the cross-sectional area were measured in the radius of 100 healthy white girls (aged 4-15 years) using computed tomography (CT); 50 girls had never fractured and 50 girls had sustained a forearm fracture within the previous month. Fractured and nonfractured groups were matched for age, height, weight, and Tanner stage of sexual development. Compared with controls, girls with fractures had, on average, 8% smaller cross-sectional area at the distal radius (1.82 ± 0.50 cm2 vs. 1.97 ± 0.42 cm2; p < 0.0001) but similar cancellous, integral, and cortical bone densities. Neither radial length nor the amount of fat or muscle at the midshaft of the radius differed between girls with and without fractures. Both study subjects and matched controls were overweight. Although mean height was at the 50th percentile, mean weight was at the 90th percentile for age-adjusted normal values. Girls who sustain forearm fractures after minor trauma have small cross-sectional dimensions of the radius and tend to be overweight. The smaller cross-sectional area confers a biomechanical disadvantage that, coupled with the greater body weight, increases the vulnerability to fracture after a fall.
APPROXIMATELY ONE in every three children suffer a fracture by the time they are 15 years of age, commonly in the distal radius.(1) These fractures generally are attributed to the increase in physical activity and consequent higher exposure to the risk of injury during childhood and adolescence. However, although significant trauma is a key factor in most fractures that occur in children, at least one-third of all pediatric fractures are not related to skating, bicycling, strenuous sports, motor-vehicle accidents, or significant trauma, suggesting that there must be other contributing factors.(2) Indeed, several investigators have suggested that bone mass, a major determinant of bone strength, is decreased in children with forearm fractures when compared with matched controls.(2–5) Using various bone measurement techniques, a reduction of 5-10% in radial bone mass has been found in children who sustain forearm fractures.
The basis for the deficient accumulation of bone mass observed in adolescents with radial fractures is not known, but, as the incidence of fractures dramatically decreases with skeletal maturity,(6–9) it is commonly thought to be related to a physiological and transient growth-related process. It has been hypothesized that the vulnerability to fractures during childhood is the consequence of a temporary decrease in bone density, which results from either an increase in unmineralized osteoid as bone formation outstrips mineral deposition or an increase in bone porosity associated with the greater bone remodeling rate of the growing skeleton.(8, 10)
The recent adaptation of quantitative computed tomography (CT) to measure accurately radial size, bone volume, and bone density has significantly improved our ability to quantify the structural components that influence bone strength.(11, 12) This study was designed to determine the structural basis for the skeletal fragility observed in children who sustain forearm fractures and to determine if this deficiency is, indeed, caused by lower bone mineralization or increased cortical porosity.
MATERIALS AND METHODS
The study subjects were healthy white girls, 6-15 years of age, 50 of whom had sustained a low-energy impact fracture of the distal forearm and 50 of whom had never had a fracture. Participants in the study group were patients treated at the Division of Orthopedics at the Children's Hospital Los Angeles. Participants in the control group were healthy volunteers from schools in the Los Angeles area who had never sustained a fracture. The investigational protocol was approved by the Children's Hospital Los Angeles Institutional Review Board, and informed consent was obtained from all subjects and their parents or legal guardians.
Eligible candidates for the study group were girls who had sustained a fracture at the distal one-third of the radius after low-energy impact trauma within 1 month of participation in this study. Children with physeal (Salter-Harris) fractures were not eligible for inclusion in this study. For the purpose of this investigation, a low-energy impact fracture was defined as that resulting from a fall with an outstretched hand from a height not greater than the child's own height while she was walking or running.(7) Girls who had sustained fractures due to motor-vehicle accidents, while roller skating, skate boarding, bicycling, skiing, or participating in any sport in which the speed one attains is higher than that of running were excluded from this study. Children whose fracture resulted from contact with any other individual including during contact sports or from falling downstairs, downhill, or from a height greater than their own also were ineligible for participation.
All participants were healthy girls without physical limitations. Girls who had any chronic illness; had been ill for more than 2 weeks during the previous 6 months; had ever been hospitalized; or had taken any medications, vitamin preparations, or calcium supplements within the previous 6 months were excluded from study. Hand dominance was documented and potential candidates underwent a physical examination by a pediatric endocrinologist to determine their height, weight, overall healthiness, and stage of sexual development. The grading system of Tanner was used, which includes assessments of the pattern of pubic hair and of breast development.(13) If discrepancies existed among criteria, greater emphasis was placed on the degree of breast development for determinations of Tanner stage. Measurements of height and weight were obtained, and thereafter, body surface area and body mass index were calculated, as previously described.(14)
Subjects with fractures were evaluated and enrolled in the study before their nonfractured counterpart. Girls with fractures were matched with girls who had never sustained a fracture for chronological age, Tanner stage, height, and weight to control for these important determinants of bone mass. For this analysis, the ages of each pair of subjects differed by less than 6 months, and neither height nor weight differed by more than 5%. Using this approach, we studied 50 unique matched pairs of girls.
Techniques and definitions of CT measurements
All CT radial measurements were obtained with the same scanner (General Electric Hilite Advantage; General Electric, Milwaukee, WI, USA) and mineral reference phantom (CT-T bone densitometry package; General Electric) for simultaneous calibration. Radial length and the locations of the scanning sites were assessed using CTs of the radius. For the purpose of this study, the length of the radius was defined as the distance between the proximal and the distal radial physes.
Measurements were obtained at the distal portion and at the midshaft (corresponding to 50% of the length) of the radius. The distal scanning site was determined by measuring the distance between the distal physeal plate and the beginning of the fracture in girls with fractures. For every pair of girls, this site was scanned in the unfractured radius of girls in the study group and in the radius matched for hand dominance of girls in the control group. At this location, the apparent density of cancellous bone (mg/cm3), the integral density of bone (mg/cm3), and the cross-sectional area (cm2) were calculated. In addition, measurements of the material density of bone (mg/cm3) were obtained at the midshaft of the unfractured radius. From this image, the cross-sectional areas (cm2) of fat and muscle tissue also were determined.
The density of cancellous bone was defined as the mean value of the CT unit of measurement (mg/cm3) at the distal radius, excluding the cortical shell. Because of the relatively small size of the trabeculae when compared with the pixel, CT values for apparent cancellous bone density are influenced not only by the amount of mineralized bone and osteoid, but also by the amount of marrow per pixel and are mainly a reflection of the number and the thickness of the trabeculae.(15) These measurements are analogous to in vitro determinations of the volumetric density of trabecular bone, which are obtained by washing the marrow from the pores of a specimen of cancellous bone, weighing it, and dividing the weight by the volume of the specimen, including the pores.(16)
The integral density of bone was defined as the mean value of the CT unit of measurement (mg/cm3) at the distal radius, including the cortical shell. This value represents an integrated measurement of bone density that includes both cortical and cancellous bone. Because of the relatively narrow rim of cortical bone in the distal radius compared with the dimensions of the CT voxel, technical limitations lead to a systematic underestimation of the density of cortical bone in the radial metaphysis. However, for this study, measurements of the density of cortical bone (mg/cm3) were obtained at the midshaft of the radius. At this site, the cortex is sufficiently thick to circumvent volume averaging errors and CT values reflect both the material density of the bone (the amount of collagen and mineral in a given volume of bone) and the porosity of cortical bone.(12) These measurements are analogous to in vitro determinations of the intrinsic mineral density of bone, which commonly are expressed as the ash weight per unit volume of bone.(17)
The CVs for repeated CT measurements of bone (the apparent density of cancellous bone, the integral density of bone, the material density of cortical bone, the radial length, and the cross-sectional area), muscle, and fat in the forearm were calculated to be between 0.6% and 2.5%.(12, 18) The time required for the procedure was 10-15 minutes and the radiation exposure was approximately 100-200 mrem (1.5 mSv) localized to the distal and midportion of the radius; the effective radiation dose was approximately 8 mrem.(19, 20)
All results are expressed as mean ± SD. The data were analyzed using the Student's t-test for paired samples, analysis of variance (ANOVA), and linear regression analysis.(21, 22) A significant level of p < 0.05 was used for all comparisons. All tests were two-sided and values of p < 0.05 indicated statistical significance for a power of 80%.
Table 1 shows the anthropometric characteristics of the 50 girls with fractures and the 50 girls without fractures. By design, there were no significant differences in developmental status among girls with and without fractures, and the mean values for age, Tanner stage, height, and weight were similar in both groups. In addition, neither body surface area nor body mass index differed between study and control groups. It should be noted that although subjects in the study and control groups were matched for weight, participants in this study were overweight when compared with age-adjusted normal percentiles for growth.(23) Although mean height for study subjects and controls was at the 50th percentile, mean weight was at the 90th percentile for the mean age-adjusted values for normal girls; overall, participants were 12 kg overweight.
Table Table 1.. Ages and Anthropometric Measurements of 50 Pairs of Girls Matched for Tanner Stage, Age, Height, and Weight
In the study group, the right forearm fractured in 29 cases and the left fractured in 21 cases. Forty-five girls fractured their radius when they fell forward, backward, or sideways on their hand/wrist while running or walking, and five girls fractured after falling from a height not greater than their own from a chair, a sofa, or a bed. The distance of the fracture from the distal physeal plate varied from 4.9% to 16.9% of the length of the radius (mean, 10.5 ± 3.5%). Significant negative correlations were observed between the age and the Tanner stage of sexual development of the child and the distance of the fracture from the physeal plate (r = −0.37 and −0.46, respectively; p < 0.001 for both). At that same distance in the nonfractured forearm, the cross-sectional area was significantly smaller in girls with fractures when compared with girls without fractures (Table 2). In all but five pairs of girls, the cross-sectional area of the radius was larger in the control subject (Fig. 1). On average, the difference in cross-sectional areas between fracture and control groups was 8%. In contrast to these findings, radial length did not differ between fractured and nonfractured girls (Table 2).
Table Table 2.. CT Measurements in the Forearm of 50 Pairs of Girls Matched for Tanner Stage, Age, Height, and Weight
Neither the density of cortical bone at the midshaft of the radius nor the densities of cancellous or integral bone at the distal radius differed between girls with and without fractures (Table 2); values for material cortical bone density were about 8 times higher than those for apparent cancellous bone density. Measurements of the material density of bone were constant and were not influenced by age, pubertal status, or any of the anthropometric measurements including CT measurements of muscle and fat in the forearm (Fig. 2). In addition, there were no differences in the forearm musculature or fat area at the midshaft of the radius between study subjects and controls (Table 2).
A large number of children sustain forearm fractures after relatively minor trauma such as falling on an outstretched hand while walking or running.(2) The results of this study indicate that in girls who sustain low-energy impact fractures of the distal radius, the cross-sectional area of the radius is smaller than in girls who have never fractured a bone. These results cannot be attributed to differences in longitudinal skeletal growth, because neither the radial length nor the heights of the girls differed between study and control groups. Additionally, because body weight, body mass index (BMI), surface area, forearm musculature, hand dominance, and Tanner stage were similar in girls with and without fractures, our findings cannot be ascribed to differences in body mass or degree of sexual development. However, it should be stressed that when compared with the established normative data for weight, girls with fractures were overweight.
The results of this study also indicated that neither the apparent density of cancellous bone nor the integral density of all bone at the distal radius differed between girls with and without fractures. Moreover, CT values for the material density of cortical bone at the midshaft of the radius were similar in all girls and were not related to a history of forearm fractures. Our findings that CT measurements of cortical bone density remain constant regardless of age do not support the notion of transient physiological changes in bone mineralization or cortical porosity during growth.
The strength of the radius depends on its cross-sectional size and the amount of bone contained in its cross-section. Because larger cross-sections have greater strength in bending and torque then smaller cross-sections with equal amounts of bone, the smaller radial size present in girls with fractures imparts a biomechanical disadvantage that limits the ultimate loading capacity of the radius.(24, 25) The findings in this study are in agreement with previous investigations showing that the distal radial cross-sectional area and the moment of inertia are strong predictors of the strength of radial specimens.(26, 27) Interestingly, in vitro data suggest that determinations of the cross-sectional geometry have better discriminatory capabilities than bone mineral density (BMD) in assessing the fragility and fracture risk of the radius.(26, 27)
Available data suggest that the small radial cross-sectional dimensions of girls with fractures may represent a very early developmental manifestation of a lifelong skeletal deficiency. Recent studies indicate that values for the cross-sectional dimensions of bones can be tracked from childhood to early adulthood and do not change percentiles during growth.(28) In this regard, Goulding et al. recently found that the low bone mass values present in girls who had sustained distal forearm fractures persisted 4 years later.(29) Additional support for the concept that a skeletal deficiency may persist throughout life are multiple reports indicating that at any age, a previous forearm fracture is an important risk factor for future fractures.(30) Postmenopausal women who suffer a forearm fracture have a 2- to 3-fold higher relative risk of future fragility fractures and similar increases in relative risk have been observed in premenopausal women.(31–33) Likewise, refracture rates of radial fractures in children have been reported to range from 5% all the way up to 33%.(29, 34) Therefore, interventions aimed at improving bone mass accretion in children who sustain fractures are not only relevant to the treatment of the existing skeletal fragility, but also can be expected to have an impact on the risk of developing osteoporosis and fractures later in life.
Like others, we found that girls with forearm fractures are, for the most part, overweight(35); an especially provocative finding when coupled with the knowledge that in the elderly, obesity is a known protector against osteoporotic fractures.(35) In this study, mean height was at the 50th percentile, while mean weight was at the 90th percentile for age-adjusted normal values; on average, these girls were 12 kg overweight. Although a small cross-sectional area may limit the ultimate loading capacity of the radius throughout life, other contributing factors for forearm fractures must vary with age. In the young, the skeletal deficiency is more likely to become apparent in obese children because of the increase in mechanical loading associated with minor falls. In the elderly, the small cross-sectional area of the radius imparts a mechanical disadvantage that becomes increasingly important as the density and loading capacity of the bone decline with age.
The reason(s) for the dramatic decrease in the incidence of radial fractures with skeletal maturity is yet to be defined. However, it is likely that fractures during childhood and adolescence result from a temporary disassociation between gains in skeletal and body mass. Although subcutaneous fat mass increases by more than three times(36) and BMI nearly doubles during the growth spurt,(37) the greatest acquisition of bone occurs later in puberty, close to the time of sexual and skeletal maturity.(38, 39) Thus, changes in skeletal size and strength during growth may not be adequate to match the very rapid increase in weight and body fat that occurs in early puberty. This disparity is even more pronounced in overweight children.
In summary, the results of this study indicate that girls who sustain forearm fractures after minor trauma have a smaller cross-sectional area of the radius than girls who do not fracture and are likely to be overweight. The limited loading capacity of the smaller radius, coupled with the extra body weight, results in greater stress within the bone and increases the vulnerability to forearm fractures during a fall. Because the cross-sectional dimensions of the bone do not cross percentiles during growth and can be tracked from childhood to early adulthood, the skeletal deficiency present in these children may persist throughout life. Therefore, children with forearm fractures may represent a large population of individuals who may be at risk of developing Colles' fractures later in life.
The authors express their gratitude to Mrs. Cara L. Wah for her comments and technical assistance with this article. This work is supported in part by a grant from The Gerber Foundation, a grant from the National Institute of Arthritis and Musculoskeletal and Skin Diseases (R01-AR4-1853-01A1), and a grant from the National Library of Medicine (1R01-LM06270-01).