The application of DXA in pediatric medicine is a rapidly growing field. The rapidity of its growth, however, increases the likelihood of technical errors and misinterpretation of the results. One of the most important lessons to remember in pediatric densitometry is the admonition [1] from an expert in the field that “Children are not simply small adults.” An extension of this statement is that an expert in adult densitometry is not automatically an expert in pediatric densitometry.

There are several issues in pediatric densitometry that are not concerns in adult densitometry, or which are of less concern. The pediatric skeleton is constantly changing in terms of size and even shape. Ossification centers fuse in different bones at different ages. A child’s chronological age does not necessarily reflect their bone age. The onset of puberty, at whatever age it occurs in the child, has a pronounced effect on the development of the skeleton. The densitometry software in pediatric densitometry must be able to detect bone edges in the setting of lower densities than often seen in adult densitometry. Radiation safety issues are not the same. Reference databases for adults are not appropriate for use in pediatric densitometry. Similarly, the use of the T-score in pediatric densitometry is not appropriate. And finally, the diagnosis of osteoporosis should not be made on the basis of the mass or density measurement alone.

Pediatric Scan Acquisition and Analysis

The technical aspects of the performance of pediatric densitometry are not that different from those for adult densitometry. The greatest challenge in scan acquisition may indeed be keeping the child still during the scan. The shorter scan times needed by newer densitometers have helped to alleviate but not completely eliminate this problem. The manufacturer’s directions for positioning and analysis should be followed for any given type of scan. It is preferable however that the acquisition software be software that is specifically designed for a pediatric population.

Bone edge detection algorithms that are unique to each manufacturer’s DXA device enable the separation of bone from soft tissue. Edge detection algorithms designed for an adult population with an expected range of bone densities may fail when used in a pediatric population with lower body weight and lower bone mineral density (BMD). In essence, the machine may be unable to tell where the bone stops and starts. This will cause a failure in appropriate edge detection. The bone edges or bone map should be verified by the technologist during the analysis and corrected, if necessary. In a review of 34 pediatric bone density studies in which a diagnosis of osteoporosis, osteopenia, or low bone density was made, Gafni and Baron [2] found errors in bone mapping in 7 of the 34 or 21 %. After recognition of these errors, three of the seven bone densities were found to be normal, but two of the seven still could not be classified because of other errors.

Pediatric densitometry must be performed with these edge detection issues in mind. In 1993, Hologic, Inc., introduced a low density spine (LDS) software option to be used in children as well as adults with low bone density. This was an operator-selected analysis mode rather than a scan acquisition mode. The effects of low density edge detection became apparent when 100 bone density studies in children aged 2–18 years were analyzed using the LDS option as well as the standard adult analysis option [3]. When the LDS option was used, the measured bone area and bone mineral content (BMC) increased significantly. Because the bone area increased to a greater degree than the BMC, the BMD decreased an average of 8.7 % with the LDS option. Norland systems such as the XR-46™ and Excell™ utilize a dynamic filtration system that automatically adjusts the photon flux to accommodate differences in body size during scan acquisition. Systems like the GE Lunar Prodigy™ automatically select the best scan mode based on the height and weight of the patient and also employ specialized analysis algorithms for low bone density in the pediatric spine.

Radiation Safety Issues in Pediatric Densitometry

Radiation safety in densitometry in general was discussed in Chap. 5. As in adult densitometry, the overriding principle guiding radiation safety in pediatric densitometry is “as low as reasonably achievable” or ALARA. In pediatrics, however, the effective dose equivalent (HΕ)Footnote 1 cannot be assumed to be the same as in adult densitometry. Body size may have a pronounced effect on the HΕ because a smaller body size (less tissue thickness) may result in a greater dose to a specific organ and there is the potential for a proportionally greater amount of the body to be irradiated. Several studies have attempted to define HΕ in children for total body, PA lumbar spine, and proximal femur studies on both pencil-beam and fan-array densitometers [4–6]. Because the skin entrance doses are lower with pencil-beam systems than with fan-array systems, the HΕ on a pencil-beam system will also be lower, all other things being equal. On the Lunar DPX-L, a pencil-beam DXA device, the HΕ for a 5- and 10-year-old child for a PA lumbar spine study was estimated to be 0.28 and 0.20 μSv, respectively [4]. For a total body study, the HΕ values were 0.02 and 0.02 μSv. As expected, higher values for the HΕ were found by Thomas et al. [5] and Blake et al. [6] for Hologic fan-array systems. Thomas et al. [5] estimated an HΕ of 4.7 μSv for a 1-year-old for a PA lumbar spine study and values of 15.2 and 6.4 μSv for a proximal femur study in 1-year-old males and females in the fast array mode available on the QDR 4500. The difference between the males and females reflects the difference in the proportion of the gonads exposed to ionizing radiation. The effective doses calculated in this study are shown in Table 12-1. These calculations utilized the tissue weighting factors originally proposed by the ICRP in 1990 [7]. HΕ values for different Discovery/QDR 4500 scan modes were calculated by Blake et al. [6] for 5-, 10-, and 15-year-old children as well as an adult using both the ICRP60 tissue weighting factors and the newer ICRP 2007 tissue weighting factors, in which, notably, the tissue weighting factor for the gonads was decreased. These results are shown in Table 12-2. Blake et al. noted that the HΕ values were lower in the express mode, available on the Discovery, compared to the fast or array modes. All of these authors noted that the HΕ for the various scan types was very low. As was correctly pointed out by Blake et al., however, the real difference in risk is underestimated by the HΕ in children compared to adults because of the greater sensitivity of growing tissues and the longer life expectancy of the child. In keeping with the principle of ALARA, the HΕ should be kept as low as possible by using the fastest scan mode that is appropriate, shortening the scan length to accommodate the child, and by using careful technique to avoid scan restarts and repeats.

Table 12-1 Effective dose for DXA scans on a Hologic QDR 4500A densitometer as a function of age (Reproduced with permission of Elsevier from Thomas SR, et al. Effective dose of dual-energy X-ray absorptiometry scans in children as a function of age. J Clin Densitom 2005;8:415–422)
Full size table
Table 12-2 Effective doses (μSv) from spine, hip, and total body DXA examinations for different Discovery/QDR4500 scan modes (Reproduced with permission of Elsevier from Blake GM, et al. Comparison of effective dose to children and adults from dual x-ray absorptiometry examinations. Bone 2006;38:935–942)
Full size table

Bone Age

Bone age is not necessarily the same as a child’s chronological age. Bone age is a reflection of the developmental maturity of the skeleton. The presence of unfused and fused epiphyses is a reflection of developmental maturity. The epiphyses are secondary ossification centers at the ends of long bones and are responsible for longitudinal growth. The epiphyseal plate deposits cartilage which is subsequently becomes bone. Ultimately the epiphysis itself becomes engulfed in bone. Longitudinal growth stops and the epiphysis is then said to be “fused” with the rest of the bone. After fusion, the only remnant of the ossification center is a line of demarcation called the epiphyseal line. The presence of unfused epiphyses will cause a DXA image that appears bizarre to the densitometrist accustomed to adult images such as the proximal femur image in Fig. 12-1. The greater trochanter is not completely formed or fused with the rest of the proximal femur. Ossification in the greater trochanter begins around the age of 3, but is not complete until around the age of 18 [8]. Fusion of the femoral head and lesser trochanter to the proximal femur is also generally not complete until around the age of 18. The typical adult appearance of the vertebrae is not seen in young children. There are ringlike epiphyses on the upper and lower surfaces of the vertebral bodies that appear around the age of 16 but which do not fuse with the rest of the vertebral body until around the age of 25. Other regions that are commonly studied with DXA in which unfused epiphyses may be seen are the heel and the forearm. The secondary ossification center in the posterior calcaneus appears around the age of 7 and fuses at puberty. At the distal radius and ulna, secondary ossification centers appear by ages 1 and 5, respectively, although neither fuses until around the age of 20.

Fig. 12-1.
figure 00121

A DXA image of a proximal femur in a child. The greater trochanter in particular appears unusual. The greater trochanter is not fully formed and will not completely fuse to the femoral neck and shaft until approximately 18 years of age (Case Courtesy of GE Healthcare, Madison, WI).

Full size image

The state of the secondary ossification centers in the hand is used to determine bone age. There are two techniques that are traditionally used to make this determination: the Greulich and Pyle method and the Tanner and Whitehouse method [9, 10]. The Greulich and Pyle method requires a comparison of all the bones in the hand and wrist against reference X-rays for a wide range of ages. The technique has been modified in many centers to a comparison of the overall appearance of the child’s hand to a set of reference radiographs. The Tanner and Whitehouse methodFootnote 2 requires a systematic assessment of the maturity of all the bones in the hand and wrist and employs a point scoring system to determine skeletal maturity. Although studies [11, 12] have suggested that the two techniques give similar results, some authorities prefer the Tanner and Whitehouse method. Bone age is not determined from a DXA study. However, the interpretation of the bone density seen on a DXA study may well be affected by knowledge of the child’s bone age. For example, if the child’s bone age is less than their chronological age, their bone density would not be expected to be the same as their chronological peers.

Sexual Maturation Stage

Another important element in the interpretation of pediatric bone density measurements is knowledge of the level of sexual development of the child. This assessment is usually made by determining the TannerFootnote 3 stage [13]. Tanner assigned five stages to puberty for both boys and girls, with stage 1 indicating prepubertal development and stage 5 indicating mature sexual development. The five stages in girls are based on the development of the breasts and pubic hair. For boys, the stages are based on the development of the genitalia and pubic hair. Tanner stages are associated with different rates of linear growth (increase in height). In girls, the peak rate of linear growth is generally seen in Tanner stage 3 around the age of 11.5 years. In boys, the maximum rate of linear growth occurs in conjunction with Tanner stage 4 around the age of 13.5 years. In clinical practice, representative drawings are often used to allow the child to pick the body image that most closely matches their own. This is generally thought to be the least intrusive manner by which to make this determination. Parental permission is, of course, mandatory. Given that the Tanner stage represents pubertal development, it is not surprising that it is linked to skeletal maturity and rates of increases in height. This is relevant information then to the interpretation of a pediatric bone density study.

Considerations of Bone Size and Shape

The potential effect of changes in size on bone density was discussed in Chap. 6. Because the BMD obtained with DXA is a 2-dimensional areal measurement, a larger bone may have a greater BMD than a smaller bone in spite of both having identical volumetric BMDs [14, 15]. The maturation of the skeleton and increases in height will cause changes in the shape and size of the bones, making this issue particularly relevant to pediatric densitometry. In addition, children with chronic diseases are often smaller than healthy children of the same age. The interpretation of areal density in such children must be made cautiously to avoid incorrectly diagnosing a child who simply has small bones from any cause as having an abnormally low bone mass for their age. Mølgaard et al. [16] proposed a three-step method to address the potential for misdiagnosis in a pediatric population because of changes in the size or shape of the bones. The authors noted that BMD was the ratio of the BMC divided by the bone area and that if the bone area was small, the BMC would potentially be reduced. They pointed out that it was important to know whether the low BMC was the result of a small bone area. Mølgaard et al. proposed the concept of BMC adjusted for bone area which is also called “BMC for bone area.” They also proposed an assessment of bone area adjusted for height or “bone area for height” and an assessment of height adjusted for age, also called “height for age.” They suggested that this would address three potential causes of an apparent low BMD in a child: “light bones,” “narrow bones,” and “short bones.” In essence, these three parameters address the following questions: Is the height appropriate for the age or does the child have “short bones?” Is the bone size or area appropriate for the child’s height or does the child have “narrow bones?” Is the BMC appropriate for the bone area or does the child have “light bones?” The relevance of each of these findings to the health of the child may be quite different to the pediatrician. Figure 12-2 shows the plot of these three parameters on centile scalesFootnote 4 for the child whose total body bone density study is shown in Fig. 12-3. Data for the comparison of height for age comes from US Centers for Disease Control [17] growth statistics. BMC for bone area and bone area for height data are derived from the manufacturer’s pediatric reference database. In this particular example, the total body bone density Z-score shown in Fig. 12-3 is −2.2. Why might this be? The centile scales shown in Fig. 12-2 suggest that the bone area for height is only in the 24th percentile, while the height for age and BMC for bone area centiles are better. This suggests that the low BMD Z-score may be in part determined by what Mølgaard et al. [16] called “narrow bones” and not solely due to a truly low BMC.

Fig. 12-2.
figure 00122

Ancillary data provided as part of the pediatric total body study seen in Fig. 12-3. These centile line graphs provide necessary comparisons of the child’s height for their age, bone area for their height, and BMC for their bone area. This is useful in determining whether a low BMD in a child is the result of a truly decreased BMC or simply the result of bones that are smaller in size than an average child of the same age, sex, and ethnicity (Case Courtesy of GE Healthcare, Madison, WI).

Full size image
Fig. 12-3.
figure 00123

A pediatric DXA total body study performed on the GE Lunar Prodigy for a 5-year-old girl. The total body bone image is on the left and the body composition image is on the right. Note that only the Z-score is shown for the total body bone density. The bone age is also plotted on the age-regression graph and is shown as 7.5 years. This age was determined using the Tanner–Whitehouse method and inputted by the technologist prior to data acquisition (Case Courtesy of GE Healthcare, Madison, WI).

Full size image

Skeletal Development and the Use of Standard Scores in Pediatric Densitometry

As noted in Chap. 1, the standard score called the T-score in densitometry indicates the number of standard deviations above or below the average peak bone density that the patient’s bone density lies. The Z-score, on the other hand, is the standard score comparison to the bone density that is predicted for the patient’s age. Both are used in adult densitometry, although the T-score commands greater attention for its role in the diagnosis of osteoporosis based on the World Health Organization criteriaFootnote 5 as well as for fracture risk prediction. Authorities in pediatric densitometry agree, however, that the use of T-scores in children for any purpose is not appropriate. To find that a child has a bone density that is less than the average peak bone density is expected because the child may not have reached the age by which peak bone density is achieved. Such a finding then carries no particular significance and, if misinterpreted, great potential harm.

Skeletal Development

The exact age at which peak bone density is reached at any given skeletal site remains somewhat controversial. It is clear, however, that the age of peak bone density may be different between boys and girls and that the age of peak bone density may vary by skeletal site [18, 19]. The age at which peak bone density is achieved at any site may also be determined not just by the patient’s chronological age but by their bone age and pubertal status as well.

Changes in bone density in 778 Caucasian boys and girls, aged 2 to 20 years, were determined in a cross-sectional study by Zanchetta et al. [20] using DXA of the PA lumbar spine, proximal femur, and total body. In this study, BMD at the PA lumbar spine and proximal femur did not increase significantly after the age of 14 in girls. In boys, however, BMD at the spine increased throughout the age range of the study, but significant increases in proximal femur BMD were not seen in boys after the age of 16. Total body BMD in girls did not increase after the age of 16 but increased throughout the age range in boys. Nguyen et al. [18] also reported that total body peak BMD was reached earlier in girls than in boys, although at a later age than reported by Zanchetta et al. In a longitudinal study of 94 males and 92 females, aged 6–36 years with an average follow-up of 4.29 years, peak total body bone density was reached by the age of 20.8 years in females and 25.2 years in males.

Teegarden et al. [21] looked specifically at total body BMC and BMD with DXA in 247 girls and young women aged 11–32. They concluded that 99 % of peak total body BMD and total body BMC is achieved by 22.1 years and 26.2 years of age, respectively. Based on a study [19] of 300 girls and women aged 6–32, these same authors concluded that peak BMD was achieved at the PA lumbar spine by age 23, although BMC and bone area continued to increase at the spine across the age range of this study. Peak BMD at the femoral neck was reached by the age of 18.5 years.

In another very large cross-sectional study by Sabatier et al. [22], changes in BMD, BMC, and bone area were determined in 574 girls, aged 10–24 years. In this study, bone density measurements were made at the lumbar spine with DXA. Bone age was determined for girls less than age 20 using plain radiographs of the left hand and wrist with the Greulich and Pyle [9] method. Pubertal status was assessed using Tanner stages [10]. Sabiatier et al. found that the BMD and BMC in the lumbar spine increased dramatically between the bone ages of 10 and 14 or until the first year after menarche.Footnote 6 Between the bone ages of 14 and 17, the PA lumbar spine BMD and BMC continued to increase but at a slower rate. After the bone age of 17 or the fourth year after menarche, no additional significant increases in PA lumbar spine BMD or BMC were seen. The authors noted that bone age and pubertal status appeared to be more useful than chronological age in assessing skeletal status. They observed that lumbar spine BMC roughly doubled between the bone ages of 10 and 17 and that the period between the bone ages of 10 and 14 was particularly critical in the development of BMC.

The effect of pubertal state on the accumulation of bone mass had previously been observed by Theintz et al. [23]. In a study of 98 girls and 100 boys aged 9–19 years, Theintz et al. found that BMC and BMD increased rapidly between the ages of 11 and 14. The rate of increase dropped dramatically, however, after the age of 16 or 2 years after menarche. In this study, 16 appeared to be the age of peak bone mass at the lumbar spine and femoral neck in girls. These findings are very similar to those from Sabatier [22]. In boys, the increase in both BMC and BMD at the lumbar spine and proximal femur was greatest between the ages of 13 and 17. The rate of increase declined markedly after the age of 17 at both sites. No additional significant increases were seen at the femoral neck after the age of 17 in boys, although significant increases in spine BMD were still seen. Finally, in the Bone Mineral Density in Childhood Study (BMDCS), 1,554 children (761 boys, 793 girls) underwent DXA bone density testing at the PA lumbar spine, left proximal femur, nondominant forearm, and total body on the Hologic QDR 4500A, QDR 4500 W, or Delphi A [24]. Based on 3-year data from 1,442 children, the authors concluded that BMD was still increasing at all skeletal sites measured in girls aged 16 years and boys aged 17 years.

Although there are slight differences among the various studies that have attempted to determine the age of peak bone mass and density at various skeletal sites, the majority of these studies have concluded that peak bone mass is achieved at most sites by the age of 20. There is little disagreement that the overwhelming majority of peak bone density is attained by the age of 20. Increases in BMC and bone area may indeed continue, particularly at the spine, after the age of 20 [19]. This would not necessarily be reflected as an increase in BMD because BMD is the ratio of BMC to area.

The Use of Standard Scores in Pediatric Densitometry

An appreciation of the timing of peak bone density is critical to understanding why T-scores must not be used in the interpretation of bone density in children. This would be analogous to comparing the height of a 7-year-old to the height of a 35-year-old and concluding that the 7-year-old was abnormally short. Unfortunately, this misuse of the T-score is not widely appreciated. In the study cited earlier from Gafni and Baron [2] in 2002, the results of a review of 34 DXA bone density studies and the accompanying interpretations in children aged 4–17 years were reported. These children had been referred to the National Institutes of Health as possible participants in an osteoporosis treatment trial. All 34 children had a diagnosis of osteoporosis, osteopenia, or low bone mass based on the original bone density study interpretation. Gafni and Baron found that 88 % or 30 of the 34 studies had at least one error in interpretation. In 21 of the 30 studies, the T-score had been used for diagnosis even though the Z-score was also present on the printout. When the appropriate interpretation was made based on the Z-score, 12 of these 21 children actually had a normal bone density. In 5 of the 21, an accurate diagnosis could not be made because of a lack of necessary information.

Printouts of DXA studies from major DXA manufacturers in which pediatric software is used do not display a T-score. This is seen in the printout from the total body study in Fig. 12-3 and the PA lumbar spine bone density study in Fig. 12-4. Note that only Z-scores appear on the report. In the ancillary data for this study shown in Fig. 12-5, both the T-score and the % young adult comparisons are appropriately absent. This should be helpful in preventing this error in interpretation.

Fig. 12-4.
figure 00124

A pediatric DXA PA lumbar spine study performed on the GE Lunar Prodigy for a 5-year-old girl. The L1–L4 BMD is 0.617 g/cm2. The Z-score is −0.2. No T-score is provided. The use of the T-score in pediatric densitometry is not appropriate. The ancillary data for this study is seen in Fig. 12-5 (Case Courtesy of GE Healthcare, Madison, WI).

Full size image
Fig. 12-5.
figure 00125

Ancillary DXA PA lumbar spine data for the study shown in Fig. 12-4. Note that no T-score or % young adult comparison is provided (Case Courtesy of GE Healthcare, Madison, WI).

Full size image

Pediatric Reference Databases

Even with the removal of the T-score from consideration, the validity of the Z-score comparison is dependent upon the validity of the reference database. In the study from Gafni and Baron [2], the second most common error in interpretation of pediatric densitometry results was the use of a reference database that did not correctly reflect the patient’s sex or ethnicity. As noted earlier, the BMD is expected to differ between boys and girls, particularly in adolescence. Use of a pediatric database that combines both genders as though there were no expected differences in BMD will lead to erroneous interpretations [2, 25]. In particular, boys may be misclassified as having low bone mass. It is also clear from studies from Bachrach et al. [26] that the expected BMD differs among pediatric ethnic groups as well as between boys and girls. In general, she found that Blacks had a greater areal bone density than non-Blacks at the PA lumbar spine, proximal femur, and total body. For any age, the average BMD at the spine was 10 % greater in Black females and 3 % greater in Black males than in non-Blacks. There were also differences among Asian, Hispanic, and Caucasian males and females, although the differences were not as great as those seen between Blacks and non-Blacks. Among the males, spine BMD was lower in Hispanics than in Asians or Caucasians. Total hip BMD and total body BMD were greater in Caucasian males than in Hispanic or Asian males. Among the females, Asians had a lower average femoral neck and total body BMD than Hispanics and Caucasians. Bachrach and colleagues also found that gender rather than ethnicity played a significant role in the timing of increases in BMD.

In the report from Gafni and Baron [2] noted earlier, there were seven instances in which an incorrect database had been used to interpret the pediatric bone density findings. When Gafni and Baron applied the correct database to determine the Z-score, five of the seven children were no longer considered osteopenic or osteoporotic. They were, in fact, considered normal. In the other two cases, a determination could still not be made because of missing information. Gafni and Baron recommended that any reference database used to interpret pediatric bone density studies be specific for age, sex, and ethnicity. They also noted that an ideal database would consider body size and pubertal status as well.

The Bone Mineral Density in Childhood Study [24] noted earlier is an ongoing longitudinal study, from which pediatric reference databases are being created. In this study, it was clear that BMD values in childhood are not normally distributed and that specific statistical procedures are necessary to create appropriate reference databases for children. The statistical approach used is called the LMS method, previously described by Cole and Green [27]. Published reference curves from this study are currently limited by the relatively small number of Hispanic and Asian children and the length of follow-up. However, at study conclusion after 6 years, some of these limitations will be minimized. The data are specific for Hologic QDR 4500, Delphi, and Discovery systems however. Pediatric reference databases on these types of Hologic devices utilize the BMDCS reference data, supplemented by Hologic native pediatric reference data where appropriate [28].

2003, 2004, and 2007 International Society for Clinical Densitometry Guidelines for Children

In 2003, ISCD issued guidelines for the use of bone densitometry to diagnose osteoporosis in children [29]. ISCD emphasized that in males or females less than 20 years of age, the WHO criteria for the diagnosis of osteoporosis based on the measurement of BMD should not be used. ISCD stated that T-scores should not be used and should not appear in reports of bone densitometry in children; the Z-score should be used instead. The PA lumbar spine and total body were recommended as the preferred skeletal measurement sites in children. ISCD noted that while there was no agreement on standards for adjusting the BMD or BMC for bone size, pubertal stage, bone age, or body composition, any such adjustments should be clearly noted in the report. The need for an appropriate reference database for all Z-score comparisons was stressed. In 2004, the Canadian Panel of the International Society for Clinical Densitometry [30] published guidelines for the diagnosis of osteoporosis in individuals less than 20 years of age, which mirrored the guidelines published earlier that year by the parent organization.

In 2007, ISCD convened the first Pediatric Position Development Conference to specifically address issues in pediatric densitometry. The recommendations from that conference are summarized in Appendix C. Importantly, ISCD reiterated the statement from earlier guidelines that the diagnosis of osteoporosis should not be made in a child on the basis of the bone density alone [31]. According to the 2007 guidelines, such a diagnosis required the finding of a clinically significant fracture history and low bone density or low bone mineral content. Low bone mineral density or content was defined as a BMD or BMC Z-score ≤−2. DXA was the preferred bone density technique, and the PA lumbar spine and total body (without the head) were the preferred skeletal sites for measurement. The need to utilize an appropriate reference database was again emphasized.

The Specialty of Pediatric Densitometry

In addition to studies of the development of peak bone density in children and its relationship to osteoporosis in later life, there are an ever-growing number of diseases in childhood in which the bone density may be adversely affected. One primary cause of osteoporosis in children is juvenile idiopathic osteoporosis [32]. This disease is considered relatively rare. It usually occurs before puberty and is manifested by back pain, long bone fractures, and loss of height. There is often spontaneous resolution after 2–4 years, but some individuals may develop permanent disabilities. Osteogenesis imperfecta (OI) is another cause of primary osteoporosis in children. OI is also often called “brittle bone disease” much as is adult osteoporosis. OI is the result of a genetic defect in collagen synthesis that results in skeletal fragility. Signs and symptoms of OI include a bluish discoloration of the sclerae, hearing loss, short stature, and fractures. There are six or more variants of OI [33]. Type II is fatal. Type I is considered mild, while type III is the most severe, nonfatal form of OI. Bisphosphonates are being evaluated as potential therapies in both juvenile idiopathic osteoporosis and OI [34, 35]. Other genetic defects that are associated with low bone density in childhood include Turner’s syndrome, Down’s syndrome, and Klinefelter’s syndrome [33].

Secondary causes of osteoporosis or low bone mass in childhood are numerous, just as they are in adults [36]. The list includes Cushing’s syndrome, hyperthyroidism, hypopituitarism, and hypogonadism, as well as various nutritional deficiencies. Rheumatoid arthritis and inflammatory bowel disease in children are also associated with low bone density. Other diseases include sickle-cell anemia, hemophilia, and cystic fibrosis. As in adults, certain drugs such as corticosteroids and anticonvulsants may cause bone loss. As the number of childhood cancer survivors increases, the effects of antineoplastic agents on bone mass have also become a concern. As has been the case in adult densitometry, the increasing application of densitometry in the pediatric population has resulted in an increasing number of diseases now known to have an adverse effect on the skeleton of a child. This, in turn, will further increase the number of pediatric densitometry studies that are performed.

Perhaps the greatest problem faced in pediatric densitometry today is not its underutilization but, rather, its misinterpretation when performed. It must always be remembered that “children are not simply small adults” [1]. There is much more to consider in the interpretation of pediatric bone density results than the BMD itself, or even the Z-score. Excellence in adult densitometry does not automatically confer excellence in pediatric densitometry. A technologist who is cognizant of the nuances necessary for the proper performance and interpretation of pediatric densitometry can provide invaluable assistance to the interpreting physician. Technologists desiring expertise in this field are encouraged to pursue additional training, such as that offered by ISCD and publications such as the book by Sawyer et al. [37].