• bone mineral;
  • bone structure;
  • bone strength;
  • children;
  • calcium;
  • neurofibromatosis 1;
  • osteopenia;
  • osteoporosis;
  • Ras pathway


  1. Top of page
  2. Abstract

People with neurofibromatosis 1 (NF1) have low bone mineralization, but the natural history and pathogenesis are poorly understood. We performed a sibling-matched case–control study of bone mineral status, morphology, and metabolism. Eighteen children with NF1 without focal bony lesions were compared to unaffected siblings and local population controls. Bone mineral content at the lumbar spine and proximal femur (dual energy X-ray absorptiometry (DXA)) was lower in children with NF1; this difference persisted after adjusting for height and weight. Peripheral quantitative computed tomography (pQCT) of the distal tibia showed that trabecular density was more severely compromised than cortical. Peripheral QCT-derived estimates of bone strength and resistance to bending and stress were poorer among children with NF1 although there was no difference in fracture frequencies. There were no differences in the size or shape of bones after adjusting for height. Differences in markers of bone turnover between cases and controls were in the directions predicted by animal studies, but did not reach statistical significance. Average serum calcium concentration was higher (although within the normal range) in children with NF1; serum 25-OH vitamin D, and PTH levels did not differ significantly between cases and controls. Children with NF1 were less mature (assessed by pubertal stage) than unaffected siblings or population controls. Children with NF1 have a generalized difference of bone metabolism that predominantly affects trabecular bone. Effects of decreased neurofibromin on bone turnover, calcium homeostasis, and pubertal development may contribute to the differences in bone mineral content observed among people with NF1. © 2013 Wiley Periodicals, Inc.


  1. Top of page
  2. Abstract

Neurofibromatosis 1 (NF1) affects 1/3,500 people [Friedman, 1999]. NF1 results from a constitutional loss of function of one of the NF1 genes, which encode a RAS pathway regulator. Common features are café-au-lait spots, axillary freckling, Lisch nodules, neurofibromas, and short stature. Focal bony manifestations (e.g., dystrophic scoliosis or tibial pseudarthrosis) are less frequent but more serious. Published series of children and adults with NF1 report lower bone mineral density (BMD) than normal [Illés et al., 2001; Kuorilehto et al., 2005; Lammert et al., 2005; Stevenson et al., 2005; Dulai et al., 2007; Yilmaz et al., 2007; Brunetti-Pierri et al., 2008; Duman et al., 2008; Tucker et al., 2009; Seitz et al., 2010; Petramala et al., 2012]. One study reported an increased lifetime fracture frequency in those with NF1 [Tucker et al., 2009].

The pathogenesis of lower bone mass (mineral content and density) and possible increase in fracture frequency in NF1 is unclear. Animal studies suggest a direct effect of decreased bone formation [Yu et al., 2005; Li et al., 2009] and increased resorption [Yang et al., 2006; Li et al., 2009]. Though some clinical investigators have measured markers of bone formation or resorption and bone-regulating hormones among people with NF1, results have been inconsistent [Illés et al., 2001; Lammert et al., 2005; Lammert et al., 2006; Brunetti-Pierri et al., 2008; Duman et al., 2008; Stevenson et al., 2008; Tucker et al., 2009; Seitz et al., 2010; Stevenson et al., 2011a, 2011b; Petramala et al., 2012].

In this study, we compared multiple measures of bone health in 18 children with NF1 (cases) and their unaffected siblings (controls). We describe a wide range of bone biomarkers and parameters of bone mass, geometry, structure, and strength in affected children and unaffected sibling controls as a means to comprehensively assess deviations from normal values. Thus, our results provide insight into the pathogenesis of NF1 bone disease.


  1. Top of page
  2. Abstract

The protocol was approved by the appropriate Institutional Research Ethics Committees.


Families known by the BC Provincial Medical Genetics Program were invited by mail to participate if they had a member fulfilling the NF1 diagnostic criteria [NIH, 1988] who had an unaffected, cohabitating biological sibling, and both were between 6 and 20 years of age. Exclusion criteria were a focal bony lesion, a chronic illness, or treatment known to influence bone health in either the child with NF1 or the unaffected sibling.

Sample size calculations were based on the results of Brunetti-Pierri et al. [2008], who studied 73 people with NF1 aged 2.8–58.9 years, unselected for skeletal problems. Mean lumbar BMD z-score was −1.38 ± 1.05. We thought the difference in our target (children, selected for lack of focal NF1 bony lesions) would likely to be smaller than in the Brunetti-Pierri study. We therefore assumed that the average decreased z-score we would observe in the NF1 patients would be −1.0 instead of −1.38 (and 0.0 in the normal siblings). We therefore needed n = 15 to achieve 85% power (alpha = 0.05, 1-sided, SD = 1.0 in both groups.) We contacted 22 families and 18 consented to participate. Reasons for non-participation were one family with chronic illness in the NF1-unaffected sibling and three families who were not interested. One child with NF1 and one unaffected sibling, closest in age, were enrolled from each family.


Each child with NF1 and his or her sibling control underwent the same assessments on the same day. All pairs, except one, were assessed between November 2006 and March 2007, fall and winter months in British Columbia. The following evaluations were performed:

Health history and physical exam

Data on all previous fractures, dietary calcium (g) and vitamin D intake (IU), parental/participant assessment of physical activity compared to peers (standardized scale: a lot less active, somewhat less active, about the same, somewhat more active, a lot more active), and sun exposure (standardized scale: Never, Seldom, Regularly, or Often) were collected using a standardized questionnaire [Kreiger et al., 1999] and published interpretation resources [Institute of Medicine, Standing Committee on the Scientific Evaluation of Dietary Reference Intakes, 1997]. Maturity was represented by pubertal stage using a questionnaire and the method of Tanner (self-assessed or completed by parents on behalf of younger children) [Duke et al., 1980]. Height (cm), body segment lengths (cm), weight (kg), and head circumference (cm) were assessed using standard techniques.

Bone mass, geometry structure, and strength

Lumbar spine, total hip, proximal femur, head, and total body bone mineral content (BMC) (g), were assessed using DXA (Hologic QDR 4500). Total body BMC was calculated as total body BMC minus head BMC. The protocol we adopted was used in previous, large-scale studies of school children residing in a similar geographic region [MacKelvie et al., 2003, 2004].

Cortical bone geometry and cortical and trabecular bone structure at the 8% and 50% sites of non-dominant tibia were assessed using peripheral quantitative computerized tomography (pQCT, XCT 3000, software version 5.50; Stratec, Pforzheim, Germany). Our protocol has been described in detail elsewhere [Macdonald et al., 2007]. Our primary variables of interest at the 8% site were trabecular vBMD (g/cm3), total vBMD (g/cm3), trabecular area (cm2), total area (cm2), section modulus (mm3), bone strength index (BSI, mm3); and at the 50% site were cortical vBMD (g/cm3), endosteal circumference (cm), periosteal circumference (cm), cortical area (cm2), total area (cm2), strength strain Index (SSI).

We calculated bone strength index (BSI, mm3) at the 8% site and estimated strength strain index (SSI, mm3) at the 50% site. BSI is an estimate of bone strength in compression and is calculated as the square of the total density and the total cross-sectional area of the load-bearing area. SSI is an estimate of bone strength in bending and is calculated as the integrated product of section modulus and cortical density. Section modulus, in turn, represents the distribution of bone mass about the neutral axis and is substantially influenced by small variations in bone size.

Laboratory measures

The marker of bone formation, serum bone-specific alkaline phosphatase (BSAP) (IU/L) (electrophoresis), calcium concentration (calcium o-cresolphthalein complexone reaction), 25-hydroxyvitamin D (25OHD) concentration (Diasorin assay, Diasorin Canada, Mississauga, Ontario) (mmol/L), and parathyroid hormone concentration (whole molecule immunoassay, Diasorin Canada) (ng/L) were assessed using by standard clinical methods. Elevated BSAP suggests increased active bone formation, as BSAP is a byproduct of osteoblast activity. Measure of bone resorption (urinary pyridinium crosslinks, deoxypyridinoline (dpd), and pyridinoline (pyd) (µmol/mol creat)), were measured from two consecutive first morning urine collections by high performance liquid chromatography as per established procedures [Stevenson et al., 2008]. Dpd and pyd are cross-linking compound of collagen fibers that can be measured as the end products of collagen breakdown. In people with increased bone disintegration, there is markedly greater dpd and pyd excretion. Dpd is more specific to bone collagen than pyd, and thus an elevation of the dpd:pyd ratio is believed to reflect a preferential increase in bone resorption, rather than a generalized collagen breakdown [Stevenson et al., 2008].

Statistical Analysis

Data were analyzed using SPSS 11.0.1. as described below.

Population reference data

Data from two large studies conducted in Vancouver region school children [MacKelvie et al., 2003, 2004; Kontulainen et al., 2005; Macdonald et al., 2005, 2007] were used as population controls. These data included age, height, weight, puberty status and bone mass, structure and strength. Data were collected using the same methods we adopted for cases and controls in our study, although we utilized a Stratec 2000 pQCT system for our study (compared with a Stratec 1000 pQCT system for data collection of younger participants in the earlier study). Results between systems were highly correlated.

z-Scores were generated for each bone mass variable of interest (total body, lumbar spine, proximal femur, and head) for each case and control by comparing the participant's actual values to predicted values. Predicted values were derived using regression models developed using same age and sex bone measures from the larger population study [MacKelvie et al., 2003, 2004]. Zero order correlation matrices were used to identify variables associated with BMC. In our initial models we included factors known to influence BMC (sex, age, height and weight, physical activity, and dietary calcium intake). Some variables were excluded in the final model based on multi-collinearity.

The same approach was used to develop standard scores for bone structure and strength variables. However, we used a different population data set from children living in a similar geographical region [Macdonald et al., 2007]. We developed regression models as described above for bone mass.

Maturity was assessed in cases and controls with respect to the Vancouver area control data. For Tanner stages 2, 3, and 4, we calculated means and standard deviations for age separately for both boys and girls from the Vancouver area children in control data set. This enabled us to assign a z-score for age for our cases and controls for their given Tanner stages. For example, Vancouver area boys at Tanner 2 have a mean age of 11.7 years, and standard deviation of 0.9 years. A child in our study in Tanner 2 who is 13.5 years old is therefore assigned a z-score of negative 2.0. Normalizing in this way enables us to compare maturity for boys and girls for a range of ages.

Comparing measures in cases versus controls

Generally, we compared the means of normally distributed continuous variables of interest in the cases and control individuals with paired t-tests. If the distributions of ordinal variables or continuous variables were not normally distributed we compared means using the Wilcoxon rank test. Associations between variables were quantified with Pearson product–moment correlation coefficients.

One-tailed tests of statistical significance were used if a particular direction was expected based on previously published data that demonstrated decreased bone mineral status and bone fragility in NF1: BMC (by DXA) and BMD (g/mm2), strength, and section modulus (by pQCT) [Illés et al., 2001; Kuorilehto et al., 2005; Lammert et al., 2005; Stevenson et al., 2005; Dulai et al., 2007; Yilmaz et al., 2007; Brunetti-Pierri et al., 2008; Duman et al., 2008; Tucker et al., 2009; Seitz et al., 2010]. Two-tailed tests were used where there were insufficient published data to support a directional hypothesis: bone structure (by pQCT), serum calcium, 25-OH vitamin D, and PTH concentrations, as well as measure of bone formation and resorption [Illés et al., 2001; Lammert et al., 2005; Brunetti-Pierri et al., 2008; Duman et al., 2008; Stevenson et al., 2008; Tucker et al., 2009; Seitz et al., 2010]. The sign test was used to compare maturity for age between cases and sibling controls.


  1. Top of page
  2. Abstract


Participants included 15 Caucasian, 2 South Asian, and 1 mixed ethnicity sibship. When all cases were compared to all controls, group means were well matched with respect to age, sex, body mass index, and various lifestyle factors known to influence bone health, including dietary intake of calcium and vitamin D, and physical activity (Table I). On the standardized sun exposure scale, 16 of 18 sibling pairs both selected the same option. In the other two sibling pairs, the child with NF1 had one higher level of sun exposure than the sibling. Past medical history (unrelated to NF1) was negative for all cases and controls; no participants were using medication.

Table I. Descriptive Characteristics of the 18 Sibling Pairs
Variable nameNF1 casesControlsP-value by paired t-test (2-tailed), except as noted
  1. SD = standard deviation.

  2. Results are reported as means (SD).

  3. a

    Kuczmarski et al. [2002].

  4. b

    The NF1 curve, rather than the normal curve, was used here for reference. This is because many cases are above the normal curve for head circumference.

  5. c

    Fisher's exact test.

Age (years)13.8 (3.3)14.0 (3.7)P = 0.77
Males: females9: 97: 11P = 0.74c
Vitamin D intake (IU/day)344 (270)314 (233)P = 0.36
Calcium intake (mg/day)1,195 (572)1,033 (612)P = 0.21
Height z-score (standardized to normal samplea)−0.53 (1.3)0.23 (1.0)P = 0.006
Weight z-score (standardized to normal samplea)0.26 (0.9)0.25 (0.7)P = 0.25
Head z-score circumference (standardized to NF1 curveb)0.66 (1.3)−0.25 (1.0)P = 0.002
Body mass index (kg/m2)0.02 (0.54)0.00 (0.94)P = 0.9

One or both members of some sib pairs fell outside the age ranges of the reference data [MacKelvie et al., 2003, 2004; Kontulainen et al., 2005; Macdonald et al., 2005, 2007]. Thus, we compared bone mass in 15 pairs of siblings and bone structure and strength in 8 sibling pairs. Only seven case–control sib pairs where both siblings were in pubertal stages 2, 3, or 4 at the time of the study were included in the maturity assessment.

Regression Models of BMC, Bone Area, and Circumference

Final regression models for BMC included sex, age, height, and weight. For girls and boys, respectively, the model accounted for 88% and 89% of the variance in total body BMC, 90% and 82% of the variance in BMC at the lumbar spine, and 78% and 80% of the variance in BMC at the proximal femur. Optimal models for bone area and circumference included the variables sex and height, and accounted for 30–40% of the variance in these variables. Adding other variables to the models described above (weight, puberty status, or age) did not significantly improve the model.


Cases were significantly shorter than sibling controls (P = 0.003). Mean z-score for height of cases was −0.53 (SD = 1.3); mean z-score for control siblings was +0.23 (SD = 1.0). There was no significant difference in upper-to-lower segment ratio [0.95 (SD = 0.075) vs. 0.93 (SD = 0.45)] or arm span-to-height ratio [1.02 (SD = 0.032) vs. 1.01 (SD = 0.045)] between cases and controls respectively. Thus, this reflects proportionate short stature in cases compared with controls.

Bone Outcomes

Total body BMC, adjusted for age, sex, height, and weight, was significantly lower in cases with NF1 than in their sibling controls (P = 0.03; Table II). There was no significant difference between cases and controls in BMC at the lumbar spine or proximal femur. Bone mineral in the head region (adjusted for age and sex) was significantly lower in children with NF1 compared with unaffected sibs (P = 0.03, Table II), this was despite the children with NF1 having larger heads as assessed by circumference.

Table II. Z-Scores for Bone Mass Variables (BMC Assessed by Dual Energy X-Ray Absorptiometry) Adjusted for Age (Year), Sex, Height (cm), Weight (kg)
MeasurementChildren with NF1—mean z-score (SD)Unaffected sibling children—mean z-score (SD)P-value by paired t-test (1-tailed)
  1. BMC, bone mineral content; SD, standard deviation.

Total body (minus head) BMC−0.39 (0.63)−0.08 (0.56)0.03
Lumbar spine BMC−0.74 (1.13)−0.36 (0.97)0.17
Proximal femur BMC−0.34 (0.44)−0.34 (0.56)0.48
Head BMC−0.55 (1.5)+0.05 (2.1)0.03

Bone Geometry, Structure, Strength, and Fracture History

Bone area and periosteal and endosteal circumferences at the 8% and 50% tibial sites were not significantly different between cases and controls (adjusted for sex and height; Table III). Cases had significantly lower trabecular (but not cortical) bone volumetric density compared with controls (adjusted for sex and height) (Table III). Those with NF1 had significantly lower bone strength at the 8% site and strength strain index at the 50% site (adjusted for sex and height) compared with unaffected sibs (P = 0.04) (Table III). There was no significant difference in lifetime fracture rate between groups: One case had severe trauma at age two years resulting in fracture of the distal phalange of the thumb, one sibling had a fracture in the growth plate of the ankle secondary to minimal trauma at age seven years, one sibling had severe trauma as a baby resulting in fracture of the clavicle.

Table III. Z-Scores for Bone Geometry, Structure, and Strength Variables at the 8% and 50% Site of the Tibia, Adjusted as Described in the Text
Bone variableCasesControl sibsP-value by paired t-test (*1-tailed or **2-tailed)
  1. vBMD, volumetric bone mineral density; SD, standard deviation.

  2. Values are mean z-score (SD).

8% site
Trabecular vBMD−1.32 (1.12)−0.19 (0.76)0.01*
Total vBMD−1.51 (0.96)−0.32 (1.06)0.015*
Trabecular area−0.44 (0.85)−0.015 (0.87)0.23**
Total area−0.46 (0.83)0.11 (0.93)0.31**
Section modulus−0.72 (0.81)0.36 (1.15)0.03*
Bone Strength Index (BSI)−1.42 (1.03)−0.16 (1.06)0.015*
50% site
Cortical vBMD−0.64 (0.97)−0.27 (0.58)0.17*
Endosteal circumference−0.20 (1.53)−0.08 (1.22)0.77**
Periosteal circumference0.19 (1.03)−0.32 (0.98)0.28**
Cortical area0.46 (0.93)−0.38 (1.22)0.09**
Total area0.08 (1.28)−0.30 (0.98)0.49**
Strength Strain Index (SSI)−0.75 (0.86)0.32 (1.16)0.035*

Serum and Urine Markers of Bone Turnover, Serum Calcium, PTH, and 25OHD Concentrations

There were no significant differences in serum and urine markers of bone formation or resorption between cases and controls (Table IV). All cases and controls had serum calcium concentrations within the normal range. However, mean values were significantly higher among cases compared with controls (P = 0.02). PTH and 25OHD concentrations did not differ between cases and control sibs. Serum 25OHD concentrations were positively associated with dietary intake of vitamin D in both cases and controls (r = 0.61 and r = 0.62, respectively, P < 0.01). Reported intakes of vitamin D and calcium did not differ between cases and controls (Table I).

Table IV. Serum and Urinary Markers of Bone Formation and Breakdown and Their Hormonal Modulators
Variable nameCasesControl sibsP-value by paired t-test (2-tailed)
  1. Differences between cases and controls tested pairwise (unless otherwise indicated) within NF1 and unaffected sibling pairs.

  2. Values are mean (SD).

Bone-specific alkaline phosphatase (IU/L)119 (60)139 (104)P = 0.50
Urinary pyridinoline (µmol/mol creat)168 (82)153 (91)P = 0.49
Urinary deoxypyridinoline (µmol/mol creat)50 (27)43 (27)P = 0.24
Ratio deoxypryidinoline/pyridinoline0.29 (0.038)0.26 (0.040)P = 0.12
Serum calcium (mmol/L)2.38 (0.064)2.34 (0.078)P = 0.018
Serum 25-OH vitamin D (nmol/L)67 (23)62 (14)P = 0.32
Serum PTH (ng/L)33 (17)26 (8)P = 0.079


In the seven sibships where case and control sibling pairs were in pubertal stage 2, 3, or 4, comparison of normalized values for age at pubertal stage showed that cases were relatively less mature than controls for their age (P = 0.004). Furthermore, all children with NF1 were less mature at the same chronological age compared with children in the general population (Fig. 1: all children with NF1 have z-score below 0).


Figure 1. Mean z-score for chronological age for each pubertal stage for controls was 0.1 reflecting close to mean pubertal stage for age. In contrast, mean z-score for cases was −1.6, indicating that maturity for cases lagged behind most others their age. All but one of the children with NF1 were delayed with respect to their unaffected sibs, as demonstrated by the positive slope of all lines connecting sib pairs.

Download figure to PowerPoint


  1. Top of page
  2. Abstract

We extend the previous literature by conducting a comprehensive assessment of bone mass, geometry, structure, strength, and markers of bone metabolism in children with NF1 as compared to healthy siblings. We compare these children to a large population of healthy children living in the same geographic region from a study that used similar methods.

The compromised total body bone mass we report in children and adolescents with NF1 compared with sib controls supports results from previous studies [Dulai et al., 2007; Stevenson et al., 2007; Yilmaz et al., 2007; Duman et al., 2008]. In addition, we noted a greater deficit in trabecular bone density, consistent with Stevenson et al. [2005, 2008] and studies of bone histology in adults with NF1 [Brunetti-Pierri et al., 2008; Seitz et al., 2010]. We also report significantly lower estimates of bone strength at both distal and mid-tibial sites in cases compared with controls, consistent with Stevenson et al.'s pQCT-based study [2005, 2008]. Lower bone strength could partially explain the reported higher lifetime fracture frequency suggested in those with NF1 [Tucker et al., 2009].

Also consistent with other reports, cases with NF1 were shorter than predicted based on family height [Szudek et al., 2000]. However, the differences in BMC we observed were not attributable to differences in tibial bone size or shape. Specifically, we did not observe diminished bone geometry after adjusting for height. Stevenson et al. [2005, 2008] had reported smaller periosteal circumferences and cortical area in cases with NF1. Discrepancies between studies may reflect that we controlled for height differences between case and control groups as well as within groups; Stevenson et al. adjusted only for height within groups. Of note, there were differences in imaging protocols: bone measures were obtained at relative distances of 4%, 38%, and 66% from the distal tibial growth plate in the Stevenson et al. [2005, 2008] study, versus the 8% and 50% sites in our study.

Mouse data provide evidence that NF1 haploinsufficiency has a primary effect on both osetoblasts and osteoclasts and their interactions with regulatory cytokines. Nf1+/− osteoprogenitors [Yu et al., 2005] and Nf1+/− osteoblasts [Li, 2009] express significantly increased osteopontin (OPN), a cytokine associated with osteoclast formation, resorption, adhesion, and migration. Neurofibromin decreases receptor activator of NF κB ligand (RANKL) expression [Elefteriou et al., 2006], and Nf1+/− osteoclasts are hyper-responsive to limiting concentrations of macrophage colony-stimulating factor, RANKL [Yang et al., 2006] and OPN [Li et al., 2009].

We did not observe significant differences in markers of bone formation or resorption between cases and controls. However, as comparisons were based on only 18 sib pairs we may have had insufficient power to demonstrate a difference (Table III); sample size was determined to address the primary outcome of bone mineralization measures. Previous studies of serum and urine markers of bone metabolism in NF1 cases reported inconsistent results with some studies reporting lower values and other no difference between cases and controls [Illés et al., 2001; Duman et al., 2008; Stevenson et al., 2008; Tucker et al., 2009; Brunetti-Pierri et al., 2008].

Our finding of a higher mean serum calcium concentration (within the normal range) was previously reported in NF1 cases, but this was in a cohort with low serum 25OHD concentrations [Seitz et al., 2010]. The discrepancy in serum calcium we observed between cases and controls was independent of differences in serum 25OHD concentrations between cases and controls. Further study is required to confirm and better understand this observation. It may be that the calcium sensing receptor is reset upward in NF1, changes similar to those observed in early stages of primary hyperparathyroidism.

Studies of Germans with NF1 reported lower serum vitamin D concentrations [Lammert et al., 2006; Tucker et al., 2009; Seitz et al., 2010] compared with controls. In contrast, the mean values for serum vitamin D in cases and controls in our study (Table IV) and that of Brunetti-Pierri et al. [2008] are not significantly different. Differences may reflect dietary intake of this nutrient (vitamin D supplementation of food is common in North America, but not in Germany), cultural practices related to sun exposure, seasonal variation, patient age, assay methodology, and/or their selection of control groups.

Delayed maturity in those with NF1 is an important finding. This finding is consistent with Virdis et al. [2003] who studied a large cohort with NF1 and reported delayed menarche in 16% of girls compared with a typical age of menarche for their mothers. A later age at take off for growth in height previously reported for adolescents with NF1 [Szudek et al., 2000] may also be a function of maturational delays in boys and girls with NF1. Other differences related to sex hormones were previously described in those with NF1. Specifically, Tucker et al. [2009] reported an earlier menopause in many women with NF1. Thus, differences in sex hormone secretion, regulation or response may explain some of the differences in bone mineralization observed in those with NF1.

Physical activity is an important determinant of bone health. Based on our measures of physical activity, the children with NF1 were no different than their control siblings. Prinzie et al. [2003] and Stevenson et al. [2007] surveyed physical activity in individuals with NF1 and revealed that scores were lower. The fact that individuals with focal bony problems were excluded from our sample may contribute to the discrepant findings between the studies. Alternatively, if there truly is a decreased activity level in children meeting our study criteria, it may be that our sample size, and/or our methodology around this measure were such that it was not appreciated.

Our study has several strengths. By assessing cohabitating sibling controls, we were more confident that differences we observed were attributable to NF1 rather than confounding environmental factors. Specifically, cases and controls shared genetic, ethnic, lifestyle, and socioeconomic backgrounds. In addition, we assessed cases and controls within a family at the same time of day and season, using the same methods. Minimizing key environmental and measurement discrepancies between cases and controls may have permitted us to detect small differences in variables such as serum calcium concentration. Our high enrollment rate of 82% indicates that participants are likely to represent other NF1 cases who attend our provincial referral center. Further, our group was not enriched for individuals with focal bony lesions as has been the case in some other studies [Illés et al., 2001; Kuorilehto et al., 2005; Lammert et al., 2005; Yilmaz et al., 2007; Brunetti-Pierri et al., 2008; Duman et al., 2008]. Finally, we adopted a comprehensive approach to more adequately assess the bone heath of children with NF1.

We also acknowledge that our study has several limitations. Our relatively small sample size was compounded by the absence of normative data needed to derive some standardized scores for our youngest and oldest participants. This limitation affects measures of BMC (15 sib pairs), bone geometry, structure and strength by pQCT (7 sib pairs), and maturity (7 sib pairs). However, the paired sibling design permitted us to use robust analytical methods to maximize statistical power.

In conclusion, we report deficits in bone mass and strength in children with NF1. This was in the absence of a vitamin D difference. At present, we do not know the long-term consequences of these deficits or whether standard therapeutic interventions that address bone health and decreased risk of fracture in other at-risk populations would be effective for younger persons with NF1. Large longitudinal and intervention studies of NF1 cases and controls would serve to identify: (i) the long-term consequences of this condition and (ii) strategies to counter the effect of NF1 on the skeletal health of young people.


  1. Top of page
  2. Abstract

We are grateful to the families for their participation, Dr. Wes Schreiber for the analysis of the alkaline phosphatases and calcium levels, and Dr. Tonya Kydland for her many contributions to the project.


  1. Top of page
  2. Abstract
  • Anonymous. 1988. National Institutes of Health Consensus Development Conference Statement: Neurofibromatosis. Neurofibromatosis 1:172178.
  • Brunetti-Pierri N, Doty SB, Hicks J, Phan K, Mendoza-Londono R, Blazo M, Tran A, Carter S, Lewis RA, Plon SE, Phillips WA, O'brian Smith E, Ellis KJ, Lee B. 2008. Generalized metabolic bone disease in Neurofibromatosis type I. Mol Genet Metab 94:105111.
  • Duke PM, Litt IF, Gross RT. 1980. Adolescents' self-assessment of sexual maturation. Pediatrics 66:918920.
  • Dulai S, Briody J, Schindeler A, North KN, Cowell CT, Little DG. 2007. Decreased bone mineral density in neurofibromatosis type 1: Results from a pediatric cohort. J Pediatr Orthop 27:472475.
  • Duman O, Ozdem S, Turkkahraman D, Olgac ND, Gungor F, Haspolat S. 2008. Bone metabolism markers and bone mineral density in children with neurofibromatosis type-1. Brain Dev 30:584588.
  • Elefteriou F, Benson MD, Sowa H, Starbuck M, Liu X, Ron D, Parada LF, Karsenty G. 2006. ATF4 mediation of NF1 functions in osteoblast reveals a nutritional basis for congenital skeletal dysplasiae. Cell Metab 4:441451.
  • Friedman JM. 1999. Epidemiology of neurofibromatosis type 1. Am J Med Genet 89:16.
  • Illés T, Hlalmai V, de Jonge T, Dubousset J. 2001. Decreased bone mineral density in neurofibromatosis-1 cases with spinal deformities. Osteoporosis Int 12:823827.
  • Institute of Medicine, Standing Committee on the Scientific Evaluation of Dietary Reference Intakes. 1997. Vitamin D. In: Dietary reference intakes for calcium, phosphorus, magnesium, vitamin D, and fluoride. Washington D.C.: National Academy Press. pp 250287.
  • Kontulainen SA, Macdonald HM, Khan KM, McKay HA. 2005. Examining bone surfaces across puberty: A 20-month pQCT trial. J Bone Miner Res 20:12021207.
  • Kreiger N, Tenenhouse A, Joseph L, Mackenzie T, Poliquin S, Brown JP, Prior JC, Rittmaster RS. 1999. The Canadian Multicentre Osteoporosis Study (CaMos): Background, rationale, methods. Can J Aging 18:376387.
  • Kuczmarski RJ, Ogden CL, Guo SS, Grummer-Strawn LM, Flegal KM, Mei Z, Wei R, Curtin LR, Roche AF, Johnson CL. 2002. 2000 CDC Growth Charts for the United States: Methods and development. Vital Health Stat 11:1190.
  • Kuorilehto T, Pöyhönen M, Bloigu R, Heikkinen J, Väänänen K, Peltonen J. 2005. Decreased bone mineral density and content in neurofibromatosis type 1: Lowest local values are located in the load-carrying parts of the body. Osteoporos Int 16:928936.
  • Lammert M, Kappler M, Mautner VF, Lammert K, Storkel S, Friedman JM, Atkins D. 2005. Decreased bone mineral density in cases with neurofibromatosis 1. Osteoporos Int 16:11611166.
  • Lammert M, Friedman JM, Roth HJ, Friedrich RE, Kluwe L, Atkins D, Schooler T, Mautner VF. 2006. Vitamin D deficiency associated with number of neurofibromas in neurofibromatosis 1. J Med Genet 43:810813.
  • Li H, Liu Y, Zhang Q, Jing Y, Chen S, Song Z, Yan J, Li Y, Wu X, Zhang X, Zhang Y, Case J, Yu M, Ingram DA, Yang FC. 2009. Ras dependent paracrine secretion of osteopontin by Nf1+/− osteoblasts promote osteoclast activation in a neurofibromatosis type I murine model. Pediatr Res 65:613618.
  • Macdonald HM, Kontulainen SA, Mackelvie-O'Brien KJ, Petit MA, Janssen P, Khan KM, McKay HA. 2005. Maturity- and sex-related changes in tibial bone geometry, strength and bone-muscle strength indices during growth: A 20-month pQCT study. Bone 36:10031011.
  • Macdonald HM, Kontulainen SA, Khan KM, McKay HA. 2007. Is a school-based physical activity intervention effective for increasing tibial bone strength in boys and girls? J Bone Miner Res 22:434446.
  • MacKelvie KJ, Khan KM, Petit MA, Janssen PA, McKay HA. 2003. A school-based exercise intervention elicits substantial bone health benefits: A 2-year randomized controlled trial in girls. Pediatrics 112:e447.
  • MacKelvie KJ, Petit MA, Khan KM, Beck TJ, McKay HA. 2004. Bone mass and structure are enhanced following a 2-year randomized controlled trial of exercise in prepubertal boys. Bone 34:755764.
  • Petramala L, Giustini S, Zinnamosca L, Marinelli C, Colangelo L, Cilenti G, Formicuccia MC, D'Erasmo E, Calvieri S, Letizia C. 2012. Bone mineral metabolism in cases with neurofibromatosis type 1 (von Recklingausen disease). Arch Dermatol Res 304:325331.
  • Prinzie P, Descheemaeker MJ, Vogels A, Cleymans T, Haselager GJ, Curfs LM, Hellinckx W, Onghena P, Legius E, van Lieshout CF, Fryns JP. 2003. Personality profiles of children and adolescents with neurofibromatosis type 1. Am J Med Genet Part A 118A:17.
  • Seitz S, Schnabel C, Busse B, Schmidt HU, Beil FT, Friedrich RE, Schinke T, Mautner VF, Amling M. 2010. High bone turnover and accumulation of osteoid in cases with neurofibromatosis 1. Osteoporos Int 21:119127.
  • Stevenson DA, Moyer-Mileur LJ, Carey JC, Quick JL, Hoff CJ, Viskochil DH. 2005. Case-control study of the muscular compartments and osseous strength in neurofibromatosis type 1 using peripheral quantitative computed tomography. J Musculoskelet Neuronal Interact 5:145149.
  • Stevenson DA, Moyer-Mileur LJ, Murray M, Slater H, Sheng X, Carey JC, Dube B, Viskochil DH. 2007. Bone mineral density in children and adolescents with neurofibromatosis type 1. J Pediatr 150:8388.
  • Stevenson DA, Schwarz EL, Viskochil DH, Moyer-Mileur LJ, Murray M, Firth SD, D'Astous JL, Carey JC, Pasquali M. 2008. Evidence of increased bone resorption in neurofibromatosis type 1 using urinary pyridinium crosslink analysis. Pediatr Res 63:697701.
  • Stevenson DA, Schwarz EL, Carey JC, Viskochil DH, Hanson H, Bauer S, Weng HY, Greene T, Reinker K, Swensen J, Chan RJ, Yang FC, Senbanjo L, Yang Z, Mao R, Pasquali M. 2011a. Bone resorption in syndromes of the Ras/MAPK pathway. Clin Genet 80:566573.
  • Stevenson DA, Viskochil DH, Carey JC, Sheng X, Murray M, Moyer-Mileur L, Shelton J, Roberts WL, Bunker AM, Hanson H, Bauer S, D'Astous JL. 2011b. Pediatric 25-hydroxyvitamin D concentrations in neurofibromatosis type 1. J Pediatr Endocrinol Metab 24:169174.
  • Szudek J, Birch P, Friedman JM. 2000. Growth in North American white children with neurofibromatosis 1 (NF1). J Med Genet 37:933938.
  • Tucker T, Schnabel C, Hartmann M, Friedrich R, Frieling I, Kruse HP, Mautner VF, Friedman JM. 2009. Bone health and fracture rate in individuals with NF1. J Med Genet 46:259265.
  • Virdis R, Street ME, Bandello MA, Tripodi C, Donadio A, Villani AR, Cagozzi L, Garavelli L, Bernasconi S. 2003. Growth and pubertal disorders in neurofibromatosis type 1. J Pediatr Endocrinol Metab 16:289292.
  • Yang FC, Chen S, Robling AG, Yu X, Nebesio TD, Yan J, Morgan T, Li X, Yuan J, Hock J, Ingram DA, Clapp DW. 2006. Hyperactivation of p21ras and PI3K cooperate to alter murine and human neurofibromatosis type 1-haploinsufficient osteoclast functions. J Clin Invest 116:28802891.
  • Yilmaz K, Ozmen M, Bora Goksan S, Eskiyurt N. 2007. Bone mineral density in children with neurofibromatosis 1. Acta Paediatr 96:12201222.
  • Yu X, Chen S, Potter OL, Murthy SM, Li J, Pulcini JM, Ohashi N, Winata T, Everett ET, Ingram D, Clapp WD, Hock JM. 2005. Neurofibromin and its inactivation of Ras are prerequisites for osteoblast functioning. Bone 36:793802.