Leg and ankle fractures are common in the pediatric population and generally result from sports injuries or vehicular accidents1-5. As these fractures occur in skeletally immature individuals with open physes1,3, most are treated nonoperatively with cast immobilization. The most predictable consequences of cast immobilization and subsequent weight-bearing restriction are loss of bone mineral tissue, substantial muscle atrophy, and corresponding functional limitations6-8.
Previous studies of bone changes after fractures of the tibia or ankle in the adult population have shown that bone mineral density of the ipsilateral tibia is decreased compared with that of the contralateral limb9-22. Moreover, bone loss occurs at adjacent sites, both proximal and distal to the fracture, such as the ipsilateral hip12,13,19,22-25. A substantial loss of bone mineral of up to 50% may develop in the fractured bone during the first six to twelve months after the fracture9,10,21,24 and may continue, even after the patient returns to prefracture activity24. It may take years for patients with bone loss following a fracture to recover the lost bone9-12,16,17,20,24, or they may have only partial recovery26. This situation increases the risk of a second fracture in the same limb, even from minimal trauma14,20.
We are aware of only one study in which bone loss after orthopaedic surgery was quantified prospectively in children, and most of those children had a neuromuscular impairment and none had a fracture27. A comparison of preoperative and postoperative scans showed that the bone mineral density of the distal part of the femur on the operatively treated side had decreased 16.5% on average27. A few retrospective studies have been performed to investigate the reduction of bone mineral mass after lower-limb fracture in children. All of the authors concluded that bone loss was not clinically important in the long term as the bone appeared to be recoverable during childhood12,28,29. Nevertheless, the transitory bone mineral deficit may delay rehabilitation, the child's return to school, and resumption of activities of daily living27. It also constitutes a hypothetical risk of the patient sustaining a second fracture of the same limb when restarting physical activities.
To the best of our knowledge, no prospective study has been done to investigate the relationship between lower-limb fractures and bone mineral mass measured at the time of fracture in adolescents or to explore the bone loss due to cast immobilization and weight-bearing restriction. The first objective of the present study was to determine if adolescents sustaining a lower-limb fracture had a normal or reduced bone mineral mass at the time of the fracture compared with healthy controls and normative references. The second aim was to quantify objectively the bone mineral loss at various bone sites due to cast immobilization and subsequent decreased weight-bearing.
Study Design and Subjects
We conducted a longitudinal matched case-control study that included fifty adolescents with a lower-limb fracture and fifty healthy paired controls with no history of fracture (aged ten to sixteen years). Adolescents treated for a first lower-limb fracture in the accident and emergency department of our Children's Hospital between January 2005 and December 2008 were invited to participate in the study. Injured adolescents were selected if they were admitted to the hospital as an inpatient for orthopaedic reduction or minimally invasive surgery (closed reduction and stabilization with percutaneous wires or screws) of the fracture while they were under general anesthesia. Patients who required an open reduction of the fracture with implant placement were not eligible for inclusion. Participants were required to wear a long leg cast, to be non-weight-bearing during the initial healing phase, and to agree to be followed at the orthopaedic clinic from the time of fracture to final cast removal. During the same time period, healthy adolescents were recruited as controls from among the patients’ visitors and the children of medical staff, and by advertisement at the University Hospitals of Geneva. Exclusion criteria for injured and healthy subjects were (1) a history of lower-limb fracture; (2) a history of chronic disease, including gastrointestinal disease with malabsorption; (3) an eating disorder; (4) congenital or acquired bone disease; (5) any condition limiting physical activity; (6) taking any medication known to influence bone metabolism; (7) hospitalization for more than two weeks; and (8) a history of neurological impairment.
All participants and their parents provided written consent, and the protocol was approved by the Children and Mother's Ethics Committee (protocol number 04-057, Ped 04-002). Parents who did not wish their children to continue in the follow-up study were not asked to state a reason, and the case was removed from our study.
Measurements
Anthropometric Data
Height was assessed to the nearest 0.1 cm with the patient in bare or stocking feet with use of a precision mechanical stadiometer (Holtain, Dyfed, United Kingdom),and body weight was measured to the nearest 0.1 kg with use of a calibrated beam scale (Seca, Reinach, Switzerland). Body mass index (BMI) was calculated as weight (kg)/height squared (m2). The first height measurement was obtained with the injured adolescent in a supine position and the stadiometer placed horizontally on a table. Weight was measured with the patient in single-limb stance, and the weight of the plaster was then subtracted.
Medical History
Personal and medical histories were assessed before testing. The fracture type in each injured subject was determined on radiographs according to its anatomical location. The type and modalities of treatment, particularly the duration of immobilization in a long leg cast, the period of non-weight-bearing, and the duration for which the below-knee walking cast was used, were determined and recorded by a pediatric orthopaedist (D.C.).
Bone Mineral Variables
Areal bone mineral density (g/cm2) and bone mineral content (g) were determined with the individual in the supine position by the same technician using dual x-ray absorptiometry (DXA) (Lunar Prodigy; GE Healthcare, Madison, Wisconsin). The following regions of interest were considered on anteroposterior views: total body, lumbar spine (L2-L4), total lower limb, total hip, femoral neck, Ward triangle, greater trochanter region, and midpart of the femoral shaft. The injured adolescents had the first DXA scans performed within three days after the fracture, and the scans were repeated on the day of final cast removal. The DXA scans of the matched healthy controls were done at baseline and after the same duration as used for the matched fracture case. The long-term stability of the instrument was assessed by measuring spine phantoms supplied by the manufacturer (Calibration Block Phantom for Prodigy SN: 9081) and was found to have a precision of 0.2% coefficient of variation over the duration of the study. Bilateral areal bone mineral density of the calcaneus (g/cm2) was measured in all subjects with use of a peripheral DXA (Lunar Pixi, GE Healthcare).The bone mineral density of the injured adolescents was assessed at the time of fracture and at final cast removal, with the first measure performed while the patient was under general anesthesia in the operating room. The calcaneus was measured with use of the subregion analysis program. A region of interest was positioned with the use of anatomic features of the heel. The long-term precision and stability of the Lunar Pixi instrument was assessed by measuring a calcaneal phantom before any use. Areal bone mineral density is easily measured in children and has a reproducibility equivalent to that in adults, for whom the standard deviation (SD) of repeated measures is about 0.010 g/cm2 (according to the manufacturer's literature). In children30, the mean and SD of the difference between measures were reported to be 0.0001 and 0.015 g/cm2 and the 95% limits of agreement between measures were −0.029 to +0.029 g/cm2.
Bone mineral density and bone mineral content (total body and lumbar spine) Z-scores adjusted for age and sex were derived from the Dutch references for white children described by van der Sluis et al.30. Z-scores for the bone mineral density of the calcaneus were predicted with use of the proportional model based on height alone described by Chinn et al.31.
Statistical Analysis
The study was designed to assess the effect of a fracture event on bone scan measurements. A matched-pair cohort study was conducted to compare bone parameters (bone mineral density and bone mineral density), measured with DXA, between (1) the injured adolescents and healthy matched controls at baseline, (2) the injured and uninjured legs of the injured adolescents at cast removal, and (3) the injured adolescents and healthy controls at cast removal or an equivalent time point for the controls. The calcaneal bone mineral density was considered as the primary end point, and all other DXA values (bone mineral density and bone mineral content for the entire lower limb, femoral diaphysis, entire hip, femoral neck, Ward triangle, greater trochanter, total body, and L2-L4 levels in the lumbar spine) were considered secondary.
As we found no data in the literature regarding this specific question in pediatric subjects, we took into consideration the differences in bone mineral density values between an individual's two lower extremities reported in adult studies12,13,19,22-25, in which reductions in bone mineral density and bone mineral content following tibial fracture amounted to between 20% and 50%. A sample size analysis for paired cohort studies was performed with a power of 80% and an alpha of 0.1%. On the basis of a mean bone mineral density of the calcaneus equal to 0.5 g/cm2 in the control group, an anticipated difference in means equal to 0.25 g/cm2 (a decrease of 50% between the case and control groups, or between the injured and healthy legs), and an anticipated SD equal to 0.15 g/cm2, we needed at least twelve pairs (twenty-four subjects). However, on the basis of a mean bone mineral density of the calcaneus equal to 0.5 g/cm2 in the control group, an anticipated difference in means equal to 0.10 g/cm2 (a decrease of 20% between the case and control groups), and an anticipated SD equal to 0.15 g/cm2, we needed a maximum of forty-four subject pairs (eighty-eight subjects).
Descriptive analyses, such as the arithmetic mean, SD, percentage, and medians with interquartile ranges (IQR) (25th to 75th percentiles) were used to report age, height, weight, BMI, and Z-scores in the two groups of participants (injured patients versus healthy adolescents) and to compare DXA values between the injured and uninjured lower limbs of the injured adolescents as well as between and the injured and control groups. A Shapiro-Wilk test with an alpha threshold of 5% was used to test normality of continuous variables. Paired Student t tests with an alpha threshold of 5% were used if the variables had a normal distribution. Otherwise, a paired Wilcoxon test with an alpha threshold of 5% was used to assess the difference in DXA values between the injured and uninjured lower limbs and between the cases and controls. All results were adjusted for multiple tests with use of the Bonferroni correction. Data analyses were performed with use of Stata 9.2 software (StataCorp, College Station, Texas).
Source of Funding
This work was fully supported by grants from the Swiss National Science Foundation (SNSF number 405340-104611). The content of this publication does not necessarily reflect the views of the SNSF, nor does the mention of tradenames, commercial products, or organizations imply endorsement of such by the SNSF, or by the authors. The funding source did not play a role in the investigation.
Comparison of Baseline Results Between Injured Patients and Healthy Controls
Fifty adolescents (thirty-one boys and nineteen girls ranging in age from ten to sixteen years [mean and SD,12.9 ± 1.7 years]) admitted for treatment of a lower-limb fracture were enrolled. There were forty-two fractures of the ankle and eight fractures of both the tibia and the fibula. The fracture was physeal in thirty-eight patients, metaphyseal in four, diaphyseal in four, and epiphyseal in four. Thirty-two patients were treated with closed reduction and cast immobilization, whereas eighteen patients had closed reduction and stabilization with either percutaneous wires/screws and a long leg cast. The mean duration of immobilization in the long leg cast without weight-bearing was 26.2 ± 8.1 days; thereafter, patients wore a functional below-knee cast for 23.6 ± 9.9 additional days until radiographic healing.
Forty-eight patients had bone mineral measurements obtained at all sites, and their data were included in the statistical data analysis. At baseline, no difference in age, height, body weight, or BMI was observed between the injured and healthy control groups (see Appendix). The mean Z-scores for L2-L4 bone mineral density and bone mineral content as well as for calcaneal bone mineral density were within the normal range (Z-scores > −1.0). There was no significant difference between the two groups with regard to L2-L4 bone mineral density/bone mineral content Z-scores or calcaneal bone mineral density Z-scores (Table I).
Comparison of Bone Parameters at Cast Removal
Forty-eight injured adolescents and forty-six healthy controls agreed to undergo the follow-up DXA measurements after cast removal or at an equivalent time point (controls). Patients with a lower-limb fracture were paired with healthy controls according to their sex and age (±0.5 years). It was possible to match forty-four patients with a lower-limb fracture with a healthy control. Age, height, body weight, and BMI did not differ between the injured and healthy control groups (Tables I and III).
When comparing the ipsilateral and contralateral lower limbs of the injured adolescents at the time of cast removal, we observed that bone mineral density and bone mineral content at all sites were significantly lower on the injured side. Differences were highly significant (p < 0.0001 for all comparisons) and ranged from –5.8% to –31.7% for bone mineral density and from –5.2% to 19.4% for bone mineral content (Table II). We then compared injured adolescents and healthy controls and observed a significant decrease of bone mineral density at the total hip, greater trochanter, calcaneus, and total lower limb (Table IV). However, the differences between the groups with respect to mean bone mineral content at various sites and mean bone mineral density/bone mineral content Z-scores of the total body and at L2-L4 were not significant (Table I). A significant reduction (<0.0001) of the median calcaneal bone mineral density Z-score was observed at the injured side in the fracture group (–1.03), whereas the median calcaneal bone mineral density Z-score at the uninjured side (0.234) was similar to that of the healthy controls (Table I).
Fractures during adolescence may be considered a public health problem. In this study, we aimed to determine whether adolescents sustaining a lower-limb fracture had abnormal bone mineral density at the time of fracture, and to quantify bone mineral loss due to cast immobilization and subsequent weight-bearing restriction. To our knowledge, this is the first report to demonstrate that bone mineral density and bone mineral content at various skeletal sites did not differ significantly from those of healthy controls at the time of fracture and that values remained within the normal range when compared with normative references. Second, we observed a significant reduction of bone mineral mass of up to –31.7% in the injured lower limb following cast immobilization.
Previously, it has been hypothesized that the high incidence of forearm fractures in children resulted mainly from a transient deficit in bone mineral mass coincident with longitudinal growth32,33. Other authors reported significant relationships between distal forearm fractures in children and DXA values in the lumbar spine34-36, femoral neck35,36, greater trochanter35,36, and total body36,37. However, few researchers have investigated the relationship between fractures at any skeletal site and the findings on regional or total body DXA scans. Clark at al. examined prospectively the relationship between baseline regional and total body DXA data and the occurrence of a fracture in the two years following the scan in 6213 children38. In their study, the average total-body-less-head bone mineral content was 0.6% lower in the fracture cases than it was in the healthy participants. When height, weight, and bone area were added to the logistic regression model, the odds ratio for the risk of fracture increased to 1.89. In a retrospective study, Manias et al. demonstrated that height-and-weight-adjusted total body bone mineral content and areal bone mineral density were lower in children who had recently sustained a fracture compared with fracture-free controls39. Finally, Ferrari et al. compared regional bone mineral content between forty-two girls who had sustained a fracture at any site and eighty-three who had remained fracture-free during a follow-up study40. The only difference between the groups was in bone mineral content at the radial diaphysis, which was 4.3% lower in the fracture group. Unfortunately, most of these studies had a number of potential limitations, such as the choice of controls or the fracture type studied, but the most important is that bone mineral mass was measured after the occurrence of the fracture34,35,39,41-44. In addition, most DXA scans were performed within four to six weeks after cast removal33-35,45, at least, and the observed reduction in bone mineral mass might be attributed to cast immobilization. Moreover, the fracture count and type were assessed with use of recall questionnaires; no information was obtained by radiographic evaluation.
In our study, bone mineral density/bone mineral content Z-scores at the lumbar spine as well as bone mineral density Z-scores at the calcaneus were normal at the time of the fracture. We also demonstrated that bone mineral density/bone mineral content Z-scores of the total body, lumbar spine, and uninjured calcaneus remained normal after cast immobilization. On the basis of these findings, we concluded that lower-limb fractures are not associated with low bone mineral density or bone mineral content at various sites and, therefore, common lower-limb fractures during adolescence are not due to osteopenia.
To our knowledge, the present investigation is the first to quantify prospectively bone mineral loss in adolescents who sustained a first lower-limb fracture requiring cast immobilization with subsequent restricted weight-bearing while the cast was in place. As previously shown in adults9,10,21,24, we demonstrated that bone mineral density and bone mineral content were decreased after cast removal in the injured limb when compared with the contralateral limb. Similar to what has been found in adults12,13,19,22-25, bone loss occurred not only at the level of the fracture, but also at adjacent sites both proximal and distal to the fracture, such as the ipsilateral hip or calcaneus. However, the bone mineral loss reported in adolescents (reaching up to –31.7%) was lower in magnitude than that observed in adults (–50%)24. In our patients, the calcaneus was the site with the greatest mineral decrement. Our findings suggest that the lack of mechanical loading on the affected lower limb had a profound negative effect on bone metabolism, leading to a reduction in bone mineral density and bone mineral content and, potentially, reduced bone strength. This phenomenon was limited to the injured limb; no modification was observed in the healthy lower limb or at the spine level, even if the whole body was physically inactive during the period of cast immobilization. Therefore, the decrement of bone mineral mass goes beyond the simple concept of a decrease of physical activity, and experimental data suggest that the magnitude, frequency, and gradients of strain may influence bone adaptation. Osteocytes may perceive these mechanical changes in stimuli via mechanotransduction, bone fluid flow, and shear stress effects on cytoskeletal conformation, piezoelectric fields, and streaming potentials46. Following a fracture, there is an increase in the rate of mineral turnover, both locally because of the traumatic process itself and regionally, mostly because of the associated immobility22. Bone resorption is not matched by bone formation, resulting in a remodeling imbalance and a net loss of bone. Immobilization and non-weight-bearing during the treatment of fractures also contribute to reduced stimulation, and temporary disuse and immobility of the fractured leg result in bone loss and muscle atrophy of the lower limb47.
It may take years before adults with bone mineral loss following a fracture recover the lost bone9-12,16,17,20,24, or there may be only partial recovery26. This situation increases the risk of a second fracture in the same limb, even from minimal trauma14,20. In the pediatric population, refracture is most common in young boys and suggests a possible combination of a fragile fracture union and overly strenuous physical activity. Forearm refractures usually are associated with incomplete healing of greenstick fractures or follow diaphyseal fractures48-50, whereas a second fracture may occur in adults as a result of a decrease in bone mineral mass after cast-mediated immobilization14,20. Currently, there is no evidence regarding the risk of children sustaining a second fracture after a lower-limb fracture.
In conclusion, at the time of fracture, bone mineral mass and density of the total body, lumbar spine, and calcaneus are not reduced, compared with the values in healthy subjects, in adolescents sustaining a first lower-limb fracture; this suggests that these fractures are not related to osteopenia in youth. In addition, this study quantified objectively the bone mineral loss induced by cast immobilization following a fracture of the lower limb in adolescents. As previously shown in adult populations, the decrement of bone mineral occurs not only at the level of the fracture but also at adjacent lower-limb sites both proximal and distal to the fracture. These results may encourage physicians, coaches, and physical education professionals to instruct adolescents who have sustained a lower-extremity fracture to delay their return to strenuous physical activities and competitive sports. Finally, the results of this study require further investigation to determine if the bone mineral mass will return to normal or if a definitive decrement is to be expected.
Note: The authors acknowledge the medical assistance of the staff of the Orthopaedic Unit of Geneva Children's Hospital, and the technical assistance of Cancela Giulio with the measurement of bone mineral densities.
Disclosure: One or more of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of an aspect of this work. None of the authors, or their institution(s), have had any financial relationship, in the thirty-six months prior to submission of this work, with any entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. Also, no author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.